American Journal of Respiratory and Critical Care Medicine

This Joint Statement of the American Thoracic Society (ATS), and the European Respiratory Society (ERS) was adopted by the ATS Board of Directors, March 2001 and by the ERS Executive Committee, June 2001

  • Introduction, 520

  • 1. Tests of Overall Respiratory Function

  • G. John Gibson, William Whitelaw, Nikolaos Siafakas

    • Static Lung Volumes, 521

    • Dynamic Spirometry and Maximum Flow, 521

    • Maximum Voluntary Ventilation, 522

    • Arterial Blood Gases: Awake, 522

    • Measurements during Sleep, 523

    • Tests of Respiratory Control, 524

    • Carbon Monoxide Transfer, 525

    • Exercise Testing, 526

    • Conclusion, 526

  • 2. Tests of Respiratory Muscle Strength

  • Malcolm Green, Jeremy Road, Gary C. Sieck, Thomas Similowski

    • Pressure Measurements, 528

    • Devices for Measuring Pressures, 528

    • Techniques for Pressure Measurement, 530

    • Volitional Tests of Respiratory Muscle Strength, 531

    • Pressures Obtained via Phrenic Nerve Stimulation, 535

    • Abdominal Muscle Stimulation, 542

    • Conclusion, 542

  • 3. Electrophysiologic Techniques for the Assessment of Respiratory Muscle Function

  • Thomas K. Aldrich, Christer Sinderby, David K. McKenzie, Marc Estenne, Simon C. Gandevia

    • Electromyography, 548

    • Stimulation Tests, 554

    • Conclusion, 556

    • Summary, 557

  • 4. Tests of Respiratory Muscle Endurance

  • Thomas Clanton, Peter M. Calverly, Bartolome R. Celli

    • Measures of Respiratory Muscle Activity Used in Endurance Testing, 559

    • Ventilatory Endurance Tests, 562

    • Endurance to External Loads, 564

    • Endurance of the Diaphragm, 568

    • Conclusion, 569

  • 5. Assessment of Respiratory Muscle Fatigue

  • Gerald S. Supinski, Jean Will Fitting, François Bellemare

    • Types of Fatigue, 571

    • Tests of Respiratory Muscle Fatigue, 572

    • Conclusion, 578

  • 6. Assessment of Chest Wall Function

  • Stephen H. Loring, Andre de Troyer, Alex E. Grassino

    • Pressures in the Chest Wall, 580

    • Assessment of the Properties of the Relaxed Human Chest Wall: Rahn Diagram, 580

    • Assessment of the Function of the Active Chest Wall: Campbell Diagram, 581

    • Estimation of Ventilation Based on Chest Wall Motion: Konno-Mead Diagram, 582

    • Devices Used to Monitor Breathing: Pneumograph, Magnetometer, and Respiratory Inductive Plethysmograph, 583

    • Optical Devices Used to Measure Chest Wall Motion, 584

    • Inferring Respiratory Muscle Contribution to Breathing from Chest Wall Motion, 584

    • Inferring Respiratory Muscle Contribution to Breathing from the Esophageal–Gastric Pressure Relationship: Macklem Diagram, 585

    • Inferring Respiratory Muscle Contribution to Breathing from Pressure–Volume Relationships, 585

    • Inferring Diaphragm Activation and Electromechanical Effectiveness from EMG, 585

    • Conclusion, 586

  • 7. Imaging Respiratory Muscle Function

  • Neil B. Pride, Joseph R. Rodarte

    • Transmission Radiography, 588

    • Ultrasound, 589

    • Volumetric Imaging, 591

    • Nuclear Medicine, 591

    • Summary, 591

  • 8. Tests of Upper Airway Function

  • Neil J. Douglas, Samuel T. Kuna

    • Electromyography, 593

    • Upper Airway Resistance, 594

    • Indirect Laryngoscopy, 596

    • Fiberoptic Imaging, 596

    • Computed Tomographic Scanning, 596

    • Magnetic Resonance Imaging, 597

    • Acoustic Reflection, 597

    • Flow–Volume Loops, 597

    • Polysomnography, 597

    • Muscle Biopsy, 598

    • Strength, Fatigue, and Endurance of Upper Airway Muscles, 598

    • Site of Pharyngeal Airway Closure during Sleep, 598

    • Conclusion, 598

  • 9. Tests of Respiratory Muscle Function in Children

  • Claude Gaultier, Julian Allen, Sandra England

    • Physiology of the Developing Respiratory Pump, 601

    • Tests of Respiratory Function, 601

    • Conclusion, 607

  • 10. Assessment of Respiratory Muscle Function in the Intensive Care Unit

  • Martin J. Tobin, Laurent Brochard, Andrea Rossi

    • Breathing Pattern, 610

    • Lung Volumes, 611

    • Pressure Measurements, 611

    • Prediction of Weaning, 617

    • Conclusion, 619

Over the last 25 years, great efforts have been made to develop techniques to assess respiratory muscle function. Research output in this area has progressively increased, with the number of peer reviewed articles published on respiratory muscle function having increased remarkably during the 1995–2000 period compared with 1980–1985.

This official joint statement represents the work of an expert ATS/ERS committee, which reviewed the merits of currently known techniques available to evaluate respiratory muscle function. The statement consists of 10 sections, each addressing a major aspect of muscle function or a particular field of application. Each section addresses the rationale for the techniques, their scientific basis, the equipment required, and, when pertinent, provides values obtained in healthy subjects or in patients. Some of the techniques reviewed in this statement have thus far been used primarily in clinical research and their full potential has not yet been established; however, they are mentioned for the purpose of stimulating their further development.

Through continued efforts in the area of respiratory muscle testing, it is anticipated that there will be further enhancement of diagnostic and treatment capabilities in specialties such as intensive care, sleep medicine, pediatrics, neurology, rehabilitation, sports medicine, speech therapy, and respiratory medicine.

Routine measurements of respiratory function, that is, volumes, flows, and indices of gas exchange, are nonspecific in relation to diagnosis but give useful indirect information about respiratory muscle performance. On occasion, the presence of respiratory muscle dysfunction is first suspected from the pattern of conventional respiratory function tests. More frequently, they are of use in assessing the severity, functional consequences, and progress of patients with recognized muscle weakness.

Rationale and Scientific Basis

The most frequently noted abnormality of lung volumes in patients with respiratory muscle weakness is a reduction in vital capacity (VC). The pattern of abnormality of other subdivisions of lung volume is less consistent. Residual volume (RV) is usually normal or increased, the latter particularly with marked expiratory weakness (1). Consequently, total lung capacity (TLC) is less markedly reduced than VC, and the RV/TLC and FRC/TLC ratios are often increased without necessarily implying airway obstruction.

The VC is limited by weakness of both the inspiratory muscles, preventing full inflation, and expiratory muscles, inhibiting full expiration. In addition to the direct effect of loss of muscle force, reductions in compliance of both the lungs (2) and chest wall (3) also contribute to the reduction of VC in patients with chronic respiratory muscle weakness. In severe weakness, the TLC and VC relate more closely to lung compliance than to the distending force (4, 5) (Figure 1)

. The mechanism of reduced lung compliance is unclear. Contrary to earlier suggestions, it is probably not simply due to widespread microatelectasis (6). Static lung volumes may also be affected in some patients by coexistent lung or airway disease. Vital capacity, thus, reflects the combined effect of weakness and the static mechanical load on the respiratory muscles.

In mild respiratory muscle weakness, VC is less sensitive than maximum respiratory pressures. However, the curvilinear relation between VC and maximum inspiratory pressure (5) (Figure 2)

implies that, in more advanced disease, marked reductions in VC can occur with relatively small changes in maximum pressures.

In patients with isolated or disproportionate bilateral diaphragmatic weakness or paralysis, the VC shows a marked fall in the supine compared with the erect posture because of the action of gravitational forces on the abdominal contents. In some patients, this postural fall may exceed 50%. In most normal subjects, VC in the supine position is 5–10% less than when upright (7) and a fall of 30% or more is generally associated with severe diaphragmatic weakness (8).

Methodology and Equipment

Recommendations and requirements for the measurement of VC and other lung volumes are covered in detail elsewhere (9, 10).


VC has excellent standardization, high reproducibility and well-established reference values. It is easily performed, widely available, and economical. It is quite sensitive for assessing progress in moderate to severe respiratory muscle weakness. The rate of decline has been shown to predict survival in both amyotrophic lateral sclerosis (11) and Duchenne muscular dystrophy (12).


VC has poor specificity for the diagnosis of respiratory muscle weakness. In mild weakness, it is generally less sensitive to changes than are maximum pressures (13).


Serial measurements of VC should be routine in monitoring progress of patients with acute and chronic respiratory muscle weakness.

Measurement of postural change of VC gives a simple index of weakness of the diaphragm relative to the other inspiratory muscles.

Rationale and Scientific Basis

Airway resistance is normal in uncomplicated respiratory muscle weakness (14). Airway function may appear to be supernormal when volume-corrected indices such as FEV1/VC or specific airway conductance are used (2).

The maximum expiratory and maximum inspiratory flow–volume curves characteristically show a reduction in those flows that are most effort dependent, that is, maximum expiratory flow at large lung volumes (including peak expiratory flow) and maximum inspiratory flow at all lung volumes (2, 5) (Figure 3)

. The descending limb of the maximum expiratory flow–volume curve may suggest supernormal expiratory flow when this is related to absolute volume (2, 3). With severe expiratory weakness, an abrupt fall in maximum expiratory flow is seen immediately before RV is reached (1). In health the FEV1 is usually less than the forced inspiratory volume in 1 second. Reversal of this ratio is seen with upper (extrathoracic) airway obstruction, as well as in respiratory muscle weakness, and may give a pointer to these diagnoses during routine testing.

The effect of coughing can be visualized on the maximum expiratory flow–volume curve in healthy subjects as a transient flow exceeding the maximum achieved during forced expiration. The absence of such supramaximal flow transients during coughing presumably results in impaired clearance of airway secretions and is associated with more severe expiratory muscle weakness (15). Even with quadriplegia, however, some patients can generate an active positive pleural pressure in expiration (16). This can allow them to achieve the pressure required for flow limitation through most of expiration so that FEV1 may still be reliable as an index of airway function. Impaired maximal flow in some neuromuscular diseases may also reflect poor coordination of the respiratory muscles rather than decreased force per se.

Oscillations of maximum expiratory and/or inspiratory flow—the so-called sawtooth appearance—are seen particularly when the upper airway muscles are weak and in patients with extrapyramidal disorders (17) (Figure 4)


Methodology and Equipment

Recommendations and requirements for maximum flow–volume curves are covered in detail elsewhere (9, 10).


Maximum flow–volume curves are easily performed, widely available, and economical. Peak expiratory flow can be obtained with simple portable devices.


Intersubject variability is greater than for VC. Reference values for V·emax at standard percentages of FVC may present problems of interpretation.


Visual inspection may suggest the likelihood of weakness.

The sawtooth appearance in an appropriate context may suggest weakness or dyscoordination of upper airway muscles. However, this appearance is nonspecific and is seen also in some subjects with obstructive sleep apnea, nonapneic snoring, and thermal injury of the upper airway.

Rationale and Scientific Basis

The maximum voluntary ventilation was formerly recommended as a more specific test for muscle weakness than volume measurements but, in practice, the proportionate reduction is usually similar to that of VC (18, 19). Disproportionate reductions may be seen in Parkinson's disease (20), in which the ability to perform frequent alternating movements is impaired.

Methodology and Equipment

Recommendations and requirements are covered elsewhere (10).


No advantages are perceived in most situations.


The test depends on motivation and is tiring for the subject.


Maximum voluntary ventilation is not generally recommended for patients with known or suspected respiratory muscle weakness but may be helpful in the assessment and monitoring of patients with extrapyramidal disorders.

Rationale and Scientific Basis

In chronic muscle weakness, even when quite severe, PaO2 and the alveolar–arterial Po2 difference are usually only mildly abnormal (2, 21). In acute muscle weakness, PaO2 may be more markedly reduced, but the picture may be complicated by atelectasis or respiratory infection (22).

With mild weakness, PaCO2 is usually less than normal (19, 22), implying alveolar hyperventilation. In the absence of primary pulmonary disease, daytime hypercapnia is unlikely unless respiratory muscle strength is reduced to < 40% of predicted and VC is reduced to < 50% of predicted (19) (Figures 5 and 6)

. Elevation of venous bicarbonate concentration occasionally gives an important clue to otherwise unsuspected hypercapnia. Patients with muscle weakness are less able than normal subjects to compensate for minor changes in respiratory function. If hypercapnia is established or incipient, even minor infections may cause a further rise in PaCO2, as also may injudicious use of sedative drugs or uncontrolled oxygen.


Arterial blood gases assess the major functional consequence of respiratory muscle weakness. In patients with Duchenne muscular dystrophy, hypercapnia has been shown to predict shorter survival (12).


Definitely abnormal arterial blood gases usually imply late and severe impairment of respiratory muscles and therefore their measurement is neither sensitive nor specific. Daytime values may underestimate the severity of abnormal gas exchange.


Measurement of arterial blood gases is routinely performed to assess the consequences of respiratory muscle weakness.

Rationale and Scientific Basis

Patients with moderate or severe respiratory muscle weakness characteristically show dips in oxygen saturation (SaO2) related to periods of rapid eye movement (REM) sleep (23, 24) (Figure 7)

. The episodic desaturation is usually due to hypopnea and less often to apnea and is associated particularly with phasic REM sleep, when brief periods of rapid, irregular eye movements are accompanied by reduced activity of skeletal muscles (24) (Figure 8) . The hypopneas and/or apneas may appear to be either “central” (Figure 8) or “obstructive,” or sometimes a mixture of both. The precise pattern of such events depends on the relative activation of the respiratory pump and upper airway dilator muscles (24). Obstructive apneas are more likely in weak patients who are also overweight (25). In patients with severe respiratory muscle weakness, some apneas that appear to be central may in fact be obstructive, incorrect classification being due to failure of external sensors to detect chest wall movements of reduced amplitude (26).

Hypercapnia in patients with slowly progressive weakness probably develops first during sleep. Continuous monitoring during sleep (e.g., with a transcutaneous Pco2 electrode) shows a gradual rise in Pco2 during REM sleep (23) (Figure 7). Consequently, PaCO2 measured shortly after waking is more likely to be elevated than values obtained later in the day. Symptoms of nocturnal hypoventilation include morning headaches, daytime sleepiness, and lack of energy. Similar symptoms can also result from sleep disruption associated with frequent apneas and hypopneas, even in the absence of persistent hypercapnia. Daytime somnolence is particularly common in patients with myotonic dystrophy. However, even though sleep hypopnea and apnea are frequently found in this condition, they appear not to explain the sleepiness of most patients with myotonic dystrophy (27).

The timescale of progression from nocturnal to persistent diurnal hypercapnia in patients with chronic respiratory muscle weakness is not known.


Polysomnographic techniques are described in detail elsewhere (28). To assess whether upper airway narrowing is a contributing cause of apneas or hypopneas may require use of a supraglottic or esophageal pressure sensor. Interpretation of recordings obtained by inductance plethysmography or other devices that measure rib cage and abdominal expansion is problematic in patients with quadriplegic or diaphragm paralysis. It is essential to check the polarity of the tracings and to compare phase relations awake and asleep.

Reliability of the devices for monitoring Pco2 in sleep is currently doubtful and requires more study.


Overnight oximetry is simple to perform.

Nocturnal measurements are more sensitive for detection of abnormal pulmonary gas exchange than daytime blood gases.


Polysomnography is labor-intensive and relatively expensive. Current evidence suggests that nocturnal hypoxemia is a less good prognostic indicator than either vital capacity or awake PaCO2 (12, 29).


The role of sleep measurements in patients with respiratory muscle weakness is currently uncertain. Polysomnography may be useful in patients with daytime sleepiness and suspected nocturnal hypoventilation, perhaps especially if awake PaCO2 is borderline or only mildly elevated.

Marked REM-related desaturation is seen occasionally in patients with relatively normal daytime SaO2 (26). More typically, however, the severity of nocturnal desaturation is predictable from daytime measurements, with more marked desaturation in patients with lower daytime PaO2, higher PaCO2, and lower VC (23) (Figure 9)


Sleep studies should be performed in all patients for whom nocturnal ventilatory support is being considered. On occasion, the finding of frequent hypopneas and/or apneas that are predominantly obstructive will suggest a trial of treatment with nasal continuous positive airway pressure. More frequently, however, in patients with respiratory muscle weakness, bilevel pressure support or another method of noninvasive intermittent positive pressure ventilation will be the treatment of choice. Because there is no evidence that treatment of abnormalities of gas exchange per se during sleep is beneficial, currently there is no indication for widespread application of polysomnography in the absence of relevant symptoms.

Rationale and Scientific Basis

The respiratory control system may be considered to have three functional components: (1) sensory receptors that provide information about the status of the respiratory system (only chemoreceptors that measure arterial Pco2, Po2, and pH are usually considered or tested, but there are many other sensory inputs of importance); (2) the central integrating circuits; and (3) the motor output to the respiratory muscles. The tests available are stimulus response tests, in which a receptor is stimulated and the motor output or a downstream mechanical effect of motor output, is measured. It is important to recognize that these tests are generally unable to separate the three functional components of the control system.

Minute ventilation and arterial Pco2 are maintained at normal levels even with quite marked weakness of the respiratory muscles, implying that the control system compensates for the weakness by driving the respiratory muscles harder than normal. The mechanism by which the control system identifies muscle weakness and adjusts its motor output is unknown. The increased motor output is difficult to appreciate because it succeeds in generating only normal pressures, volumes, and flows. It is most readily apparent when accessory muscles or abdominal muscles are more active than normal during quiet breathing. If phasic contraction of scalenes, sternocleidomastoids, pectoral muscles, or abdominal muscles can be palpated, it is safe to conclude that respiratory motor output is above normal.

When respiratory muscles are chronically severely weak and arterial Pco2 begins to rise, two explanations are possible. The muscles may be so weak that they cannot continually generate sufficient alveolar ventilation. Otherwise, an abnormality of the ventilatory control system may be allowing the Pco2 to rise even though the muscles themselves are quite capable of keeping it normal. A gradual shift in the Pco2 “set point” of the controller does seem to occur in some patients with muscle disease, as it does in some cases of sleep apnea and chronic obstructive pulmonary disease.

Laboratory tests of overall respiration that have been used to try to assess the control system include inhalation of hypercapnic or hypoxic gas mixtures to stimulate chemoreceptors, with measurements of ventilation or occlusion pressure to assess motor output, and sleep studies to monitor behavior of the control system during sleep.

In patients with weak muscles, interpretation of slopes of conventional ventilatory curves is clouded for several reasons.

• The output of the controller is abnormally high when ventilation is normal. The controller may therefore be on the nonlinear part of its normal response curve.

• The high motor neuron output cannot be measured directly and its mechanical effect (e.g., ventilation) is reduced in the presence of weakness.

• The response will become flat if ventilation nears the limit of respiratory muscle endurance and that limit may be only a short distance above resting ventilation.

Abnormal central control of respiration is well documented in bulbar poliomyelitis and other conditions affecting the central nervous system, presumably because of direct involvement of medullary respiratory centers. It has been suggested that certain muscle diseases are also associated with primary abnormalities of central respiratory control; these conditions include myotonic dystrophy, acid maltase deficiency, and other congenital myopathies. Impaired ventilatory responses to CO2 and/or hypoxia have frequently been described, but in many cases, respiratory muscle function was assessed inadequately. In myotonic dystrophy it has been shown that the relations between hypercapnia and both maximum respiratory pressures and VC are similar to those in nonmyotonic diseases (30).

Occlusion pressure is the pressure generated in the airway (and by inference the pressure generated in the pleural space) by contraction of inspiratory muscles when the airway has been occluded at end expiration. It was introduced to separate hypoventilation due to high pulmonary resistance or elastance from hypoventilation due to a failure of the respiratory pump apparatus (i.e., the muscles, passive components of the chest wall, and the control system) (31, 32). Occlusion pressure amplitude does not directly assess either the degree of muscle weakness or the degree of neuronal adjustment to the weakness. P0.1 is the pressure generated in the first 100 milliseconds of inspiration against an occluded airway. Its timing is such that it is not influenced by the conscious response to occlusion and as an index of ventilatory drive it has the advantage over ventilation of being independent of the mechanical properties of the lung (31). It is, however, dependent on the contractile state and function of the respiratory muscles and consequently on the lung volume at which it is measured. For example, because of the length–tension relationship of the muscles, a reduced value for a given neural output would be expected with pulmonary hyperinflation and an elevated FRC. On the other hand, if inspiration starts below equilibrium lung volume the value of P0.1 recorded depends on relaxation of the expiratory muscles.

Values of P0.1 are around 1 cm H2O in normal subjects at rest, around 3 cm H2O in patients with stable chronic obstructive pulmonary disease, and may be 10 cm H2O or more in acute respiratory failure due to chronic obstructive pulmonary disease or acute respiratory distress syndrome. Such values reflect a high ventilatory drive consequent on a greatly increased mechanical load. Some, although not all, studies have suggested that in patients with chronic obstructive pulmonary disease receiving ventilatory support values greater than 4–6 cm H2O are associated with failure to wean (33).

In patients with weak muscles, resting P0.1 tends to be normal or slightly increased (34). In the model of acute respiratory muscle weakness provided by partial curarization of healthy subjects, the slope of P0.1 response to CO2 is increased even though the ventilatory response is reduced (35). However, in patients with chronic weakness the ventilatory and P0.1 slopes are both diminished (even though resting P0.1 is normal or increased). Hence, a reduced response in such individuals does not necessarily imply impaired ventilatory drive (30).

Methodology and Equipment

For assessment of ventilatory responses to hypercapnia or hypoxia (36), the subject inhales a gas mixture that causes a change in either arterial Po2 or Pco2. A plot of Po2 (or Pco2) against ventilation (or, for Po2 response the algebraic constants describing a hyperbola) are compared with normal values. The induced change in blood gases may be continuous (rebreathing methods) or a few discrete points (steady state methods). Usually Pco2 is held constant while Po2 is changed and vice versa. Standard methods are available for measuring ventilatory responses during rebreathing (37, 38).

Steady state or quasi-steady state tests (39) are done simply by having the subject inhale a prepared mixture of gases, usually for 5 minutes (40). Judgments about the safety of inducing hypoxemia or acidosis are made clinically for individual patients. In chronically hypoxemic patients, transient responses to inhalation of pure oxygen may be useful and are safe (40).

For the measurement of P0.1, it is essential to close the airway exactly at the point of zero flow. This is usually done by separating the inspiratory and expiratory lines with one-way valves and then closing the inspiratory line while the subject is exhaling. Conscious subjects must be unable to anticipate occlusions, which must be done silently and unexpectedly. Obstruction can be simply performed by inflating a balloon within the lumen of the inspired line or by closing a valve. A sensitive transducer and timer are used to record pressure at 0.1 second.


A completely flat ventilatory response may identify defective chemoreceptor or brainstem function, but lesser abnormalities are difficult to interpret.

Occlusion pressure (P0.1) is relatively easy to measure. Marked discrepancies between occlusion pressure and minute ventilation point to a lung disease causing substantial increase in airway resistance or lung elastance. Usually, however, such a problem is clinically evident and better evaluated by spirometry.


Indices of ventilatory control have a wide normal range and are subject to overinterpretation.

Occlusion pressures in general, and P0.1 in particular, are difficult to interpret without additional measurements of mechanics and control events through the whole respiratory cycle, which are usually not available. P0.1 is a valid index of neural output only at FRC. Breath-to-breath scatter in the data requires averaging of many breaths to obtain precise results. The theoretical issues regarding measurement and interpretation have been reviewed (41).

Clinical Applications

These tests are seldom used in routine clinical assessment of stable patients. In acute respiratory failure, mouth occlusion pressure during unstimulated breathing may be of value in assessing respiratory drive and the likelihood of successful weaning.

Occlusion pressure has no proven clinical value in respiratory muscle disease but may occasionally be helpful by pointing to an unsuspected mechanical problem.

If a patient is known to have a mixed problem of muscle weakness and a lung disease (e.g., polymyositis plus interstitial pulmonary fibrosis) and the response of the controller to CO2 or O2 is being studied, P0.1 can be measured in conjunction with ventilation as the response and may be a more reliable way of comparing the result with normal values.

Rationale and Scientific Basis

Single-breath CO diffusing capacity (transfer factor) (DlCO) in patients with muscle weakness is usually normal or mildly reduced. Reduction is due to inability to achieve full distension of the lungs at TLC and consequent failure to expose all the alveolar surface to carbon monoxide. As with other extrapulmonary causes of lung volume restriction, the transfer coefficient (Kco) is often supernormal.


The measurement is easily performed and well standardized.


A reduced DlCO is a nonspecific finding (but if accompanied by elevation of Kco it suggests extrapulmonary volume restriction). Any effects of respiratory muscle weakness on the measurements are indirect.

Clinical Applications

The pattern of normal or mildly reduced DlCO and raised Kco directs attention to extrapulmonary conditions, that is, respiratory muscle weakness, pleural disease or rib cage abnormalities. Otherwise, the main role of measurement of CO uptake is in the recognition or exclusion of coexistent lung disease.


In many patients with muscle weakness, exercise is limited, and therefore, maximum oxygen consumption is reduced because of weakness of the leg muscles rather than cardiorespiratory factors. The limited available data suggest that the relation of workload to oxygen consumption is normal, as also are indices of submaximal exercise performance (42).


Formal testing allows confirmation and quantification of exercise incapacity and may aid elucidation of its mechanism.


Exercise is limited by weakness of nonrespiratory muscles in many patients with neuromuscular disease. Exercise testing is poorly standardized in this patient population.

Clinical Applications

Exercise testing may help determine the main factor(s) limiting exercise capacity, especially if related or coexistent cardiac or pulmonary disease is present or suspected.


This Section of the Statement has explored the usefulness of analyzing the results of pulmonary function tests to infer alterations in respiratory muscle function. Some such inferences are as follows:

  1. Respiratory muscle weakness reduces VC.

  2. Expiratory muscle weakness can increase RV.

  3. Reduction in chest wall and lung compliance, as a consequence of muscle weakness, reduces lung volumes, notably VC.

  4. A fall in VC in the supine position, compared with when upright, suggests severe diaphragm weakness or paralysis.

  5. With respiratory muscle weakness the maximal expiratory and inspiratory flow–volume loops show a reduction in effort-dependent flows (peak flows) and a sharp fall in end-expiratory flow.

  6. Reduced maximal flows in neuromuscular disease may reflect poor respiratory muscle coordination.

  7. Maximum inspiratory and expiratory flow–volume curves showing sawtooth oscillations are seen when the upper airway muscles are weak and also in patients with extrapyramidal disorders (e.g., Parkinson's disease).

  8. PaO2 and PaCO2 are affected by muscle weakness. Mild weakness causes slight hypoxemia and hypocapnia; severe weakness causes hypercapnia, but only when strength is < 40% predicted. A raised bicarbonate level may suggest muscle weakness.

  9. Respiratory muscle weakness may cause desaturation and hypercapnia during REM sleep.

  10. CO transfer (DlCO) in patients with muscle weakness is normal or mildly reduced but, as with other causes of extrapulmonary lung volume restriction, the transfer coefficient (Kco) is often raised.

1. Kreitzer SM, Saunders NA, Tyler HR, Ingram RH. Respiratory muscle function in amyotrophic lateral sclerosis. Am Rev Respir Dis 1978;117: 437–447.
2. Gibson GJ, Pride NB, Newsom Davis J, Loh C. Pulmonary mechanics in patients with respiratory muscle weakness. Am Rev Respir Dis 1977; 115:389–395.
3. Estenne M, Heilporn A, Delhez L, Yernault J-C, De Troyer A. Chest wall stiffness in patients with chronic respiratory muscle weakness. Am Rev Respir Dis 1983;128:1002–1007.
4. Gibson GJ, Pride NB. Lung mechanics in diaphragmatic paralysis. Am Rev Respir Dis 1979;119:119–120.
5. De Troyer A, Borenstein S, Cordier R. Analysis of lung volume restriction in patients with respiratory muscle weakness. Thorax 1980;35:603–610.
6. Estenne M, Gevenois PA, Kinnear W, Soudon P, Heilporn A, De Troyer A. Lung volume restriction in patients with chronic respiratory muscle weakness: the role of microatelectasis. Thorax 1993;48:698–701.
7. Allen SM, Hunt B, Green M. Fall in vital capacity with weakness. Br J Dis Chest 1985;79:267–271.
8. Laroche CM, Carroll N, Moxham J, Green M. Clinical significance of severe isolated diaphragm weakness. Am Rev Respir Dis 1988;138:862–866.
9. Quanjer PH. Standardised lung function testing. Eur Respir J 1993; 6(Suppl 16):3S–102S.
10. American Thoracic Society. Standardisation of spirometry: 1987 update. Am Rev Respir Dis 1987;136:1285–1298.
11. Fallat RJ, Jewitt B, Bass M, Kamm B, Norris FH. Spirometry in amyotrophic lateral sclerosis. Arch Neurol 1979;36:74–80.
12. Phillips M, Smith PEM, Carroll N, Edwards RHT, Calverley PMA. Does nocturnal oxygen desaturation predict survival in childhood onset muscular dystrophy? Thorax 1997;52:A18.
13. Black LF, Hyatt RE. Maximal static respiratory pressures in generalised neuromuscular disease. Am Rev Respir Dis 1971;103:641–650.
14. Wesseling G, Quaedvlieg FCM, Wouters EFM. Oscillatory mechanics of the respiratory system in neuromuscular disease. Chest 1992;102:1752–1757.
15. Polkey MI, Lyall RA, Green M, Leigh PN, Moxham J. Expiratory muscle function in amyotrophic lateral sclerosis. Am J Respir Crit Care Med 1998;158:734–741.
16. Estenne M, van Muylem A, Gorini M, Kinnear W, Heilporn A, de Troyer A. Effects of abdominal strapping on forced expiration in tetraplegic patients. Am J Respir Crit Care Med 1998;157:95–98.
17. Vincken WG, Cosio MG. Flow oscillations on the flow–volume loop: clinical and physiological implications. Eur Respir J 1989;2:543–549.
18. Serisier DE, Mastaglia FL, Gibson GJ. Respiratory muscle function and ventilatory control. I. In patients with motor neurone disease; II. In patients with myotonic dystrophy. Q J Med 1982;51:205–226.
19. Braun NMT, Arora NS, Rochester DF. Respiratory muscle and pulmonary function in polymyositis and other proximal myopathies. Thorax 1983;38:616–623.
20. Tzelepis GE, McCool FD, Friedman JH, Hoppin FG. Respiratory muscle dysfunction in Parkinson's disease. Am Rev Respir Dis 1988;138: 266–271.
21. Lane DJ, Hazleman B, Nichols PJR. Late onset respiratory failure in patients with previous poliomyelitis. Q J Med 1974;43:551–568.
22. Harrison BDW, Collins JV, Brown KGE, Clark TJH. Respiratory failure in neuromuscular diseases. Thorax 1971;26:579–584.
23. Bye PTP, Ellis ER, Issa FG, Donnelly PM, Sullivan CE. Respiratory failure and sleep in neuromuscular disease. Thorax 1990;45:241–247.
24. White JES, Drinnan MJ, Smithson AJ, Griffiths CJ, Gibson GJ. Respiratory muscle activity and oxygenation during sleep in patients with muscle weakness. Eur Respir J 1995;8:807–814.
25. Labanowski M, Schmidt-Nowara W, Guilleminault C. Sleep and neuromuscular disease: frequency of sleep-disordered breathing in a neuromuscular disease clinic population. Neurology 1996;47:1173–1180.
26. Smith PEM, Calverley PMA, Edwards RHT. Hypoxaemia during sleep in Duchenne muscular dystrophy. Am Rev Respir Dis 1988;137:884–888.
27. Gilmartin JJ, Cooper BG, Griffiths DJ, Walls TJ, Veale D, Stone TN, Osselton JW, Hudgson P, Gibson GJ. Breathing during sleep in patients with myotonic dystrophy and non-myotonic respiratory muscle weakness. Q J Med 1991;78:21–31.
28. American Thoracic Society. Indications and standards for cardio-pulmonary sleep studies. Am Rev Respir Dis 1989;139:559–568.
29. Gay PC, Westbrook PR, Daube JR, Litchy WJ, Windebank AJ, Iverson R. Effects of alterations in pulmonary function and sleep variables on survival in patients with amyotrophic lateral sclerosis. Mayo Clin Proc 1991;66:686–694.
30. Gibson GJ, Gilmartin JJ, Veale D, Walls TJ, Serisier DE. Respiratory muscle function in neuromuscular disease. In: Jones NL, Killian KJ, editors. Breathlessness. Hamilton, Canada: CME; 1992. p. 66–71.
31. Whitelaw WA, Derenne JP, Milic-Emili J. Occlusion pressure as a measure of respiratory center output in conscious man. Respir Physiol 1975;23:181–199.
32. Matthews AW, Howell JBL. The rate of isometric inspiratory pressure development as a measure of responsiveness to carbon dioxide in man. Clin Sci Mol Med 1975;49:57–68.
33. Sassoon CSH, Te TT, Mahutte CK, Light RW. Airway occlusion pressure: an important indicator for successful weaning in patients with chronic obstructive pulmonary disease. Am Rev Respir Dis 1987;135:107–113.
34. Baydur A. Respiratory muscle strength and control of ventilation in patients with neuromuscular disease. Chest 1991;99:330–338.
35. Holle RHO, Schoene RB, Pavlin EJ. Effect of respiratory muscle weakness on P0.1 induced by partial curarization. J Appl Physiol 1984;57: 1150–1157.
36. Cherniack NS, Dempsey J, Fencl V, Fitzgerald RS, Lourenco RV, Rebuck AS, Rigg J, Severinghaus JW, Weil JW, Whitelaw WA, et al. Workshop on assessment of respiratory control in humans. I. Methods of measurement of ventilatory responses to hypoxia and hypercapnia. Am Rev Respir Dis 1977;115:177–181.
37. Read DJC. A clinical method for assessing the ventilatory response to CO2. Australas Ann Med 1967;16:20.
38. Rebuck AS, Campbell EJM. A clinical method for assessing the ventilatory response to hypoxia. Am Rev Respir Dis 1974;109:345.
39. Weil JW, Byrne-Quinn E, Sodal JE, Filey GF, Grover RF. Acquired attenuation of chemoreceptor function in chronically hypoxic man at altitude. J Clin Invest 1971;50:186.
40. Cunningham DJC, Cormack RS, O'Riordan JLH, Jukes MGM, Lloyd BB. An arrangement for studying the respiratory effects in man of various factors. Q J Exp Physiol 1957;42:294.
41. Whitelaw WA, Derenne J-P. Airway occlusion pressure. J Appl Physiol 1995;74:1475–1483.
42. Carroll JE, Hagberg JM, Brooks MH, Shumate JB. Bicycle ergometry and gas exchange measurements in neuromuscular disease. Arch Neurol 1979;36:457–461.

For Abbreviations see page 547.


Muscles have two functions: to develop force and to shorten. In the respiratory system, force is usually estimated as pressure and shortening as lung volume change or displacement of chest wall structures. Thus, quantitative characterization of the respiratory muscles has usually relied on measurements of volumes, displacements, pressures, and the rates of change of these variables with time.

Several important considerations have to be kept in mind:

  1. Pressures at a given point are usually measured as a difference from barometric pressure.

  2. Pressures measured at a point are taken to be representative of the pressure in that space. Differences in pressure at different locations in normal subjects can arise from two causes: gravity and shear stress (1). Gravity causes vertical pressure gradients related to the density of the contents of the space. In the thorax this gradient is 0.2 cm H2O · cm−1 height and is related to lung density. In the abdomen, this gradient is nearly 1 cm H2O · cm−1 height. Pressure fluctuations are usually little affected by gravitational gradients. Deformation of shape-stable organs can cause local variations in pressure, such as those that occur when the diaphragm displaces the liver during a large forceful diaphragmatic contraction (2). Pleural pressure may not be uniform in patients with disordered lung architecture, particularly emphysema. The schematic drawing in Figure 1

    shows relationships between pressures and intervening respiratory structures and equipment.

  3. Pressure differences across structures are usually the relevant “pressures” for characterizing those structures. Table 1

    Download Figure |

    TABLE 1. Pressures for basic respiratory mechanics

    lists pressures measured at a point and pressure differences across structures, which are usually taken in a direction such that positive pressure differences inflate the structure or lung.

  4. A pressure difference between two points is always the pressure difference across two or more structures or groups of structures. For example, the pressure difference between the pleural space and the body surface in a breathing person is both the trans-chest wall (transthoracic) and the transpulmonary pressure.

The relationship between pressure and force is complex. For example, thoracic geometry plays a major role in the efficiency of the conversion of force into pressure. The latter also depends on the mechanical characteristics of the rib cage and abdominal wall with which respiratory muscles interact: a stiffer rib cage better resists distortion and therefore allows more pressure to be produced by the diaphragm for a given level of force (3). It follows that pressures should be viewed as indices of global respiratory muscle “output” rather than as direct measures of their “contractile properties.” Phonomyography could in future provide information related to force (4, 5) (see also sections on fatigue).

To test respiratory muscle properties, pressures can be measured either during voluntary maneuvers (see subsequent section) or during involuntary contractions, notably in response to phrenic nerve stimulation (see subsequent section). In the former, the synergistic action of several inspiratory or expiratory muscle groups is tested. In the latter, the pressure developed is specific to the contracting muscle(s).

The purpose of this article is to describe the methodology used to measure the various pressures for the assessment of respiratory muscle strength.


A comprehensive review of the techniques for measurement of pressures in respiratory physiology and of the associated problems was presented by Milic-Emili (6) in 1984.

Pressure Transducers

As for most pressure measurements of respiratory events, a frequency response flat up to 10–15 Hz is adequate to measure both dynamic and static pressures related to contractions of respiratory muscles. The frequency response of a transducer can be much altered by the characteristics of the systems attached to it, including balloons, tubing, and interconnecting fittings (7) (see subsequent section). Thus, testing the response characteristics of any transducer with the specific connectors and fittings that are to be used to make the measurements of pressure is highly recommended (7).

When differential pressure transducers are used, care must be taken that their two sides have identical frequency responses. Calibration is best made with water manometers. Electrical calibration is acceptable, but should be checked regularly with a water manometer.

The required range and sensitivity of the transducers depends on the test in question. Phrenic nerve stimulation in disease may develop pressures as low as a few centimeters of water, whereas maximal static maneuvers in healthy subjects can be associated with positive and negative pressures exceeding 200 cm H2O. It may be possible to use a single type of transducer for all respiratory muscles tests, provided that it is sufficiently sensitive, with a resolution of approximately 0.5 cm H2O and a range ± 200 cm H2O. Pressure differences between two points can be measured directly with two catheters connected to a single differential pressure transducer.

Excellent pressure transducers, with such characteristics, are commercially available, including devices based on a metal “membrane.” More recently, other types of transducer that provide good results (e.g., piezoelectric transducers) have been made available at lower cost.

Probes for “Internal” Pressures
Balloon catheter systems.

The balloon catheter system is the most widely used method for recording esophageal pressure (Pes, Poes; see Appendix for a list of abbreviations) as a reflection of pleural pressure (Ppl), and gastric pressure (Pga) as a reflection of abdominal pressure (Pab) (8). Air-containing latex balloons are sealed over catheters, which in turn transmit pressures to the transducers. Single- and double-balloon catheter systems are commercially available, but can be made in-house at low cost. Double-balloon catheters associated with an electromyograph (EMG) electrode have been used (911). When choosing or preparing a balloon catheter system, careful attention must be given to its physical characteristics. Indeed, the volume of the balloon, its volume–pressure characteristics, and the dimensions of the catheter can influence the measurement of pressure and introduce major errors. Standardization has been proposed (12).

For the measurement of Pes, good results have been provided by latex balloons 5–10 cm long, 3.5–5 cm in perimeter, and with a thin wall (8, 13, 14). For accurate transmission of pressure, air should be introduced into the balloon until it is fully distended to smooth out folds, and then most of the air removed so that a volume is retained at which the rubber is unstretched without distending the esophagus significantly. A volume of 0.5 ml is adequate for balloons with these characteristics. The volume displacement coefficient of the balloon catheter–transducer system should be measured, particularly if the balloon will measure positive pressures, to ensure that the pressure level to be measured does not completely empty the balloon into the catheter and transducer. Thus, if high positive pressures are to be measured (e.g., for Pes during maximal expiratory maneuvers) a volume of 0.5 ml may be inadequate (6). Balloon volumes should be checked repeatedly during measurements.

For the measurement of Pga, balloon volume is less crucial and measurements can be made with a balloon volume of 1–2 ml, given that this remains within the range of volume over which the rubber is unstretched. If studies of relatively long duration are planned, the walls of the gastric balloon should be thicker than those of esophageal balloons to increase resilience to gastric secretions.

Respiratory muscle studies can involve dynamic maneuvers with high rates of change in pressure (e.g., sniffs and twitches) resulting in a significant risk of a damped signal if the frequency response of the measuring system is inadequate, as may occur if the internal diameter of the catheter is too small or the gas volume too large. Polyethylene catheters with an internal diameter 1.4–1.7 mm and 70–100 cm in length provide, when associated with adequate transducers, an appropriate frequency response (6).

The catheter should be reasonably stiff, with a series of holes arranged in a spiral pattern over the entire portion of the catheter covered by the balloon, because the gas in the balloon tends to shift to the point where the pressure surrounding it is most negative, i.e., the top of the balloon in upright subjects.

Liquid-filled catheters.

Fluid-filled catheter systems have been employed, mainly in neonates and small animals for study of respiratory mechanics. Their advantage is that the transmission of pressure involving a noncompressible fluid (usually water) gives a high-frequency response. The catheters can, thus, be thinner than for balloons, theoretically reducing discomfort. An important practical difficulty is the need for regular flushing of the catheter, to avoid plugging of distal holes and to keep the catheter–manometer system free of air bubbles, which may dampen the measured pressure. Another drawback is that while the gas bubble in the balloon migrates to the point where the pressure is least (which is thought to minimize artifacts in the esophagus and to locate pressure at the surface of the gastric air bubble in the stomach) in a liquid-filled catheter, pressure is always measured at the end of the catheter, which may not be the optimal site. Respiratory muscle studies in adult humans with this technique are limited or not described, and its place in this context is probably limited.

Catheter-mounted microtransducers.

Catheter-mounted microtransducers, often referred to as Millar catheters (15, 16), have a level of performance comparable to that of balloon catheters (17, 18). Their management during long studies is probably easier, with a lower risk of technical problems (e.g., leaking balloons), and they may be easier to tolerate for the subject. Their frequency response is high, which may eliminate the phase lag sometimes seen with balloon catheters during extremely rapid pressure changes. However, catheter-mounted microtransducers record pressure at a single focused point so that the measured Pes may not be as representative of Ppl as balloon catheters, which sample pressure at the point where it is most negative. They are also much more expensive than balloon catheter systems, and may be difficult to sterilize and reuse with confidence.

Other systems.

Other systems exist to measure pressures in humans, including fiberoptic sensors. Fiberoptic sensors have long been used for measurement of intracerebral pressures in neurosurgery (19) (for review, see Yellowlees [20] and Shapiro and coworkers [21]). They are probably adequate to measure respiratory pressures (22), and may offer advantages over other devices, including decreased chance of false measurements due to occlusion with water or mucus, less chance of kinking, and, possibly, more rapid response to pressure changes. This remains to be precisely established, and, apparently, no study of fiberoptic systems in respiratory muscle tests is available.

Devices for Measurement of Airway Opening Pressure

Air-filled catheter systems are commonly used to measure pressures in airways and at the mouth. Airway opening pressure (Pao) is usually sampled from a side tap (lateral pressure) in a mouthpiece (Pmo), tracheal tube (Ptr), face mask (Pmask), or from a nostril plug (Pnas) (23). For nasal pressure to reflect airway pressure there must be free communication between the nostrils and mouth, with nasal flows. If Pao is measured from a side tap of a mouthpiece or a tracheal tube during a maneuver that involves gas flow, the cross-section of the device through which the subject breathes must be large enough to avoid measurement errors due to the Bernoulli effect (24). In some cases, Pao serves to estimate alveolar pressure (Pa, Palv) during dynamic respiratory efforts made against an obstructed airway (e.g., mouth pressure response to phrenic nerve stimulation). For Pao to reflect Pa accurately the transmission of pressure from the alveoli to the airway has to be very fast. The time constant of transmission is the product of the flow resistance offered by the airways (Raw) and the compliance of the extrathoracic airways (Cuaw) including the mouth, cheeks, and equipment. In practice the internal volume of the measuring equipment (mouthpiece, face mask, tracheal tube) contributes negligibly to the time constant (6), but should be minimized in patients with an already increased time constant, such as patients with chronic obstructive pulmonary disease (COPD). The compliance of the cheeks can be minimized by holding them rigid with the hands.

Esophageal, Gastric, and Transdiaphragmatic Pressures
Scientific basis.

Transdiaphragmatic pressure (Pdi) is defined as the difference between Ppl and Pab (13) and, in practice, is generally equated to the difference between Pes and Pga, so that Pdi = Pga − Pes (where Pes is usually, but not always, negative). This is contrary to most pressures across a structure, which are taken at a direction such that positive pressures inflate (e.g., positive transpulmonary pressures inflate the lung). For this reason Pdi is also sometimes defined as Pdi = Pes − Pga. As the diaphragm is the only muscle in which contraction simultaneously lowers Pes and increases Pga, an increase in Pdi is, in principle, the result of diaphragmatic contraction unless there is passive stretching. An inspiratory effort produced with a completely passive unstretched diaphragm is associated with a negative change in Pes and Pga but no change in Pdi. This assumes that changes in Pes or Pga induced by mechanisms other than diaphragm contraction are uniformly transmitted across the diaphragm from one compartment to the other. This is probably true when the diaphragm is relaxed (6, 13) at functional residual capacity (FRC), but may be modified when the diaphragm is stretched, as at low lung volumes.


Pes and Pga are most often measured by passing a pair of probes, generally balloon catheters (see previous passages), through the nose, following local anesthesia of the nasal mucosa and pharynx. Their position is usually assessed by asking the subject to perform sharp sniff maneuvers while monitoring the signal on an oscilloscope or computer screen. A simple technique is to advance both probes well into the stomach, as judged by a positive deflection during a sniff and then to withdraw one of them until the sniff-related pressure deflection first becomes negative, indicating that the balloon has entered the esophagus. It is then withdrawn a further 10 cm. The validity of the Pes measurement can be checked by matching Pes to Pao during static Mueller (inspiratory) maneuvers (the dynamic occlusion test) (6, 12, 14). Displacement of balloons is minimized by taping the catheters to the nose. The distance between the nostril and the tip of the balloons varies with the size of the subject, but is usually 35–40 cm for Pes and 50–60 cm for Pga in adults.

Placing the probes becomes more difficult when the subject cannot perform voluntary inspiration (e.g., with anesthetized patients, diaphragmatic paralysis, cognitive impairment, or muscle incoordination). The pressure signals during a swallow can then be useful: A balloon is positioned in the esophagus if swallowing is associated with a slow, powerful rise in pressure, whereas if this does not occur the balloon is likely to be in the stomach. Measurement of balloon distance from the nostril can be a useful indication of its position.

It is advisable to measure Pes and Pga separately by using two pressure transducers, with Pdi derived from a third differential pressure transducer or reconstructed electronically offline. This allows the investigator to monitor balloon position and detect confounding events such as esophageal spasms, as well as recording the three pressures independently. Resting Pga is usually positive with respect to atmosphere due to hydrostatic pressure in the abdomen. For respiratory muscle measurements Pga is conventionally taken as zero at resting end expiration.


Pdi is specific for diaphragm contraction (see previous passages). Separate measurements of Pes and Pga provide information on the components of this contraction and Pes on the inspiratory driving pressure (Pes/Pdi ratio).


The procedures require the subject's co-operation and occasionally untrained healthy volunteers can fail to increase Pdi because of lack of coordination, in the absence of any diaphragmatic abnormality (25). This is, however, unusual during the inspiratory phase of quiet breathing at rest. The measurements are mildly uncomfortable, both initially (when swallowing the catheters) and during studies. However, the discomfort of swallowing a thin catheter is small compared with other established medical procedures and scarcely “invasive.” Good-quality equipment and adequate practice minimize the discomfort, but some skill is necessary and passing the probes can be time-consuming. Particular care must be taken in patients with impaired swallowing, as well as esophageal diseases, or disorders at the level of the gastroesophageal sphincter.

Mouth Pressure and Nostril Pressure
Scientific basis.

Pmo is easy to measure and changes may give a reasonable approximation of change in alveolar pressure and thus Pes, providing there is relatively little pressure loss down the airways, or across the lungs. This may be realistic with normal lungs, particularly when changes in lung volume are small, but is unlikely to be fulfilled in patients with severe lung or airway disease. When used in combination with voluntary static and dynamic maneuvers at FRC, Pmo provides a global index of the action of synergistic respiratory muscles. When the diaphragm contracts in isolation against a closed airway, as with phrenic nerve stimulation, Pmo may be a useful reflection of Pdi.

Pnas is also easy to measure (see Volitional Tests of Respiratory Muscle Strength) but has the same caveats as Pmo.


Pmo is measured at the side port of a mouthpiece. It should be possible to occlude the mouthpiece at the distal end and a small leak should be incorporated to prevent glottic closure during inspiratory or expiratory maneuvers (26). The type of mouthpiece used can significantly influence the results (27). The issue of the lung volume at which Pmo should be measured during static efforts is addressed in the section on volitional tests (see subsequent section), and the various maneuvers that can be used to obtain useful Pmo data during phrenic nerve stimulation are described in the section on phrenic nerve stimulation (see subsequent section).

Pnas is measured with a polyethylene catheter held in one nostril by a soft, hand-fashioned occluding plug; respiratory maneuvers are performed through the contralateral nostril (23).

A standard mouthpiece for Pmo, or a nasal plug (custom made or commercially available) for Pnas, and one pressure transducer are required. Portable Pmo devices (28) are useful for screening and bedside studies.

Advantages of mouth pressure and nasal sniff pressure.

The main advantage of Pmo and Pnas are their simplicity and ease of use, both for the operator and for the subject.

Disadvantages of mouth pressure and nasal sniff pressure.

The measurement of Pmo does not allow the investigator to discriminate between weakness of the different respiratory muscles. When Pmo or Pnas is used as a substitute for Pes during dynamic maneuvers (sniff test, phrenic nerve stimulation), glottic closure or airway characteristics may prevent adequate equilibration.


The principal advantage of volitional tests is that they give an estimate of inspiratory or expiratory muscle strength, are simple to perform, and are well tolerated by patients. Passage of balloon catheter systems into the esophagus and/or stomach is not usually required. However, it can be difficult to ensure that the subject is making a truly maximal effort. Although normal subjects can potentially activate peripheral and respiratory muscles fully during voluntary efforts (29), even experienced physiologists cannot always do this reliably for respiratory efforts (30) and naive subjects have even greater difficulty (31). Thus, it is hard to be certain whether low mouth pressure measurements truly represent reduced strength, or merely reduced neural activation. Indeed, there may be some activation of agonist muscles simultaneously (32). However, in practice a normal result can be of value in precluding clinical weakness.

Maximal Static Inspiratory and Expiratory Pressure
Scientific basis.

Measurement of the maximum static inspiratory pressure that a subject can generate at the mouth (Pimax) or the maximum static expiratory pressure (Pemax) is a simple way to gauge inspiratory and expiratory muscle strength. The pressure measured during these maneuvers reflects the pressure developed by the respiratory muscles (Pmus), plus the passive elastic recoil pressure of the respiratory system including the lung and chest wall (Prs) (Figure 2

[33]). At FRC, Prs is zero so that Pmo represents Pmus. However, at residual volume (RV), where Pimax is usually measured, Prs may be as much as −30 cm H2O, and thus makes a significant contribution to Pimax of up to 30% (or more if Pmus is decreased). Similarly, Pemax is measured at total lung capacity (TLC), where Prs can be up to +40 cm H2O. Clinical measures and normal values of Pimax and Pemax do not conventionally subtract the respiratory system recoil.

The mouth pressures recorded during these maneuvers are assumed to reflect respiratory muscle strength (Pmus) if Prs is subtracted. However, maximum muscle strength in skeletal muscles is the force developed under isometric conditions with a muscle at its optimal length. In generating pressures during respiratory maneuvers, muscle shortening (or lengthening) may occur, with changes in force–velocity and force–length relationships (3436). The relationship between the tension (force) generated by a respiratory muscle (strength) and the pressure produced in the thorax or mouth is complex. The diaphragm is both a curved structure and acts as a piston so that the pressure or force per unit area output is only indirectly related to muscle tension. In addition, the mechanical linkage of each individual respiratory muscle within the chest wall and with other inspiratory or expiratory muscles influences the net pressure produced. Thus, even though activation may be maximal, the pressure produced is derived from a complex set of interactions within and between muscles and the chest wall and its contents. Nevertheless, it is the pressure developed by the inspiratory muscles that drives ventilation and, in spite of the many assumptions, these measures can usefully reflect global respiratory muscle strength for clinical evaluation as well as physiological studies. Thus, when respiratory muscle weakness occurs, the Pimax can be more sensitive than the VC because the relationship between VC and Pimax is curvilinear (37), so that decreases in respiratory muscle strength occur before decreases in lung volume can be identified. On the other hand, between- and within-individual variation in muscle strength is considerably greater than that for vital capacity. Between-individual variability may reflect the large variations in strength in normal individuals.

Because of the force–length relationship and the varying contribution of Prs, Pimax and Pemax vary markedly with lung volume (38). Subjects find it easier to maximize their inspiratory efforts at low lung volumes and expiratory efforts at high volumes; therefore, by convention and to standardize measurement, Pimax is measured at or close to RV and Pemax at or close to TLC. In some laboratories Pimax and Pemax are measured at FRC, and this may be more accurate for certain research studies, but in this case the lung volume should be specifically stated (39). In patients with abnormally high lung volumes (e.g., patients with COPD), a low Pimax may partly reflect the shortened inspiratory muscle fiber length associated with increased lung volume at RV rather than reduced inspiratory muscle strength (Figure 3)

. Furthermore, hyperinflation is often associated with intrinsic positive end-expiratory pressure (PEEPi), so inspiratory efforts start from a negative airway pressure. Thus, if Pimax is measured as the maximal negative airway pressure, it will underestimate the actual pressure generated by the inspiratory muscles. Optimally, under such circumstances, Pimax should be measured as the total negative deflection of the occluded airway pressure during the inspiratory effort, including the effort required to draw down PEEPi.


A number of authors have reported normal values for Pimax and Pemax (see Table 2

TABLE 2. Reference normal ranges for pemax and pimax*




Source (Ref.)

Mouthpiece Design
10623.4 ± 4.512.7 ± 3.140 Tube
6022.8 ± 4.112.1 ± 2.126 Tube
8021.2 ± 4.412.4 ± 2.741 Tube
32515.1 ± 8.011.1 ± 3.542Flanged
8014.4 ± 3.310.4 ± 3.043Flanged
4613.7 ± 3.710.3 ± 2.544Flanged
9416.1 ± 2.99.6 ± 2.440 Tube
6014.9 ± 2.68.5 ± 1.526 Tube
12113.5 ± 6.78.9 ± 2.441 Tube
4809.2 ± 3.27.0 ± 2.642Flanged
879.1 ± 1.67.2 ± 2.143Flanged
8.7 ± 2.3
6.9 ± 2.3

*Values represent kilopascals (1 kPa = 10.19 cm H2O), mean ± SD.

Reprinted by permission from Reference 27.

Definition of abbreviations: Pemax = maximum static expiratory pressure; Pimax = maximum static inspiratory pressure.

[26, 4044]). The variation between these results presumably indicates differences between the groups studied and the way in which the tests were performed and measured. Here, we propose a standardized approach to test performance and measurement.

Flanged mouthpieces are readily available in pulmonary function laboratories and although they give values somewhat lower than those obtained with a rubber tube mouthpiece, the differences are not usually material in a clinical setting (27). These mouthpieces are also easier for patients to use, especially those with neuromuscular weakness. The flanged mouthpiece can be attached to a short, rigid tube with a three-way tap or valve system to allow normal breathing followed by either a maximum inspiratory or expiratory maneuver (Figure 4)

. For research studies it may be preferable to use a rubber tube as mouthpiece (26). However, this has to be held tightly around the lips, to prevent leaks. This can be difficult for patients and naive subjects particularly at high pressures, leading to significant pressure losses. The system requires a small leak (approximately 2-mm internal diameter [id] and 20–30 mm in length) to prevent glottic closure during the Pimax maneuver and to reduce the use of buccal muscles during the Pemax maneuver. The inspiratory and expiratory pressure must be maintained, ideally for at least 1.5 seconds, so that the maximum pressure sustained for 1 second can be recorded. The peak pressure may be higher than the 1 second of sustained pressure but is believed to be less reproducible.

Historically, the aneroid manometer was used to measure the pressure but this is not recommended as the analog signal on the dial can be difficult to read accurately and pressure transients are difficult to eliminate. Mercury should be avoided for safety reasons. A recording system should be used to collect the pressure data and display it in analog form (strip chart recorder), or it can be digitized and displayed for measurement (28) or the 1-second average computed (Figure 5)

. The pressure transducers should be calibrated regularly against a fluid manometer with baseline pressure equal to atmospheric pressure.

The test should be performed by an experienced operator, who should strongly urge subjects to make maximum inspiratory (Mueller maneuver) and expiratory (Valsalva maneuver) efforts at or near RV and TLC, respectively. Subjects are normally seated and noseclips are not required. Because this is an unfamiliar maneuver, careful instruction and encouraged motivation are essential. Subjects often need coaching to prevent air leaks around the mouthpiece and to support the cheeks during the expiratory efforts, and this may be helped by having them pinch their lips around the mouthpiece. Once the operator is satisfied, the maximum value of three maneuvers that vary by less than 20% is recorded. Less variability may be necessary in a research setting, but even low variability may not guarantee that maximal efforts have been made (45).


The pressures measured at the mouth during maximum inspiratory or expiratory maneuvers are widely used specific tests of respiratory muscle strength. Normal values are available for adults, children, and the elderly. The tests are not complicated to perform and are well tolerated by patients. The recent development of hand-held pressure meters means the technique may be easily used at the bedside (28).


These tests are volitional and require full cooperation. Accordingly, a low result may be due to lack of motivation and does not necessarily indicate reduced inspiratory or expiratory muscle strength.

Normal values and applications.

The recorded values of Pimax and Pemax may be compared with published normal values (Table 2). The values that most closely reflect the protocol described here with a flanged mouthpiece, are those obtained by Wilson and coworkers (43). Normal values for the elderly (4648) and children (43, 4951) have been reported. The normal ranges are wide (Table 2), so that values in the lower quarter of the normal range are compatible both with normal strength and with mild or moderate weakness. However, a Pimax of −80 cm H2O usually excludes clinically important inspiratory muscle weakness. Values less negative than this are difficult to interpret and in such circumstances it would be appropriate to undertake more detailed studies. A normal Pemax with a low Pimax suggests isolated diaphragmatic weakness.

Regional measurements.

Static respiratory muscle pressures generated against a closed airway can be recorded from balloon catheter systems passed into the esophagus (see Techniques for Pressure Measurements) to measure Pes as a reflection of Ppl or into the stomach where Pga can be used to reflect Pab. Esophageal pressure does not include lung elastic recoil pressure but does include chest wall recoil pressure. The main indication for balloon catheter measurements of maximum respiratory muscle pressures is to estimate the strength of the separate muscle groups, notably the diaphragm (from Pdi), or to measure strength when the patient is unable to maintain a proper seal around the mouthpiece.

With the balloon catheters in place, various maneuvers can be used to assess global inspiratory muscle or diaphragm strength. These tests are usually performed at FRC. In the Mueller (maximal inspiratory) maneuver the diaphragm and inspiratory muscles are contracted with the aim of creating the biggest negative thoracic pressure without regard to abdominal pressure. However, this usually does not generate maximum Pdi (25, 52). As an alternative, the subject may perform an expulsive maneuver, wherein the individual is requested to “bear down as for defecation” and simultaneously superimposes a Mueller maneuver. When given visual feedback, this complex maneuver can be mastered by trained subjects to give the largest values of Pdi (up to 240 cm H2O or more) (53). It may reflect nearly maximal neural activation of the diaphragm, perhaps with fiber lengthening (52, 54). However, the technique is difficult for naive subjects and in the clinical setting (55). Twitch occlusion studies have confirmed that such maneuvers can produce maximal neural activation of the diaphragm (56).

Advantages and disadvantages for regional measurements.

The measurement of maximum static transdiaphragmatic pressure, Pi,di,max, produced during the described maneuvers, can provide specific information about maximal diaphragm strength. However, these tests require passage of balloon catheters and the necessary coordination is difficult for naive subjects and patients. There are limited normal data. It is difficult to control for muscle (fiber) length, and for velocity of shortening. This test is recommended only as a research tool or in respiratory muscle function laboratories with specialized expertise.

Sniff Tests
Scientific basis.

A sniff is a short, sharp voluntary inspiratory maneuver performed through one or both unoccluded nostrils. It involves contraction of the diaphragm and other inspiratory muscles. To be useful as a test of respiratory muscle strength, sniffs need to be maximal, which is relatively easy for most willing subjects, but may require some practice.

The sniff was described in 1927 as a radiological test of diaphragm paralysis because, in normal subjects, it was associated with crisp diaphragm descent during inspiration (57, 58). Esau and coworkers (59) suggested that a short, sharp sniff would approximate the diaphragm contraction elicited by a brief stimulation of the phrenic nerves (59, 60). Miller and coworkers (61) showed that normal subjects generated greater Pdi during maximal sniffs than during maximal static inspiratory efforts, perhaps because the maneuver achieves rapid, fully coordinated recruitment of the inspiratory muscles (62). The detailed respiratory mechanics of this dynamic maneuver have been little studied, but numerous studies using the sniff in normal subjects and patients have found it to be a robust measure. The nose appears to act as a Starling resistor, so that nasal flow is low and largely independent of driving pressure, Pes (63). Pdi measured during a sniff (Pdi,sn,max) reflects diaphragm strength and Pes reflects the integrated pressure of the inspiratory muscles on the lungs (Figure 6)


More recently it has been suggested that pressures measured in the mouth, nasopharynx, or one nostril give a clinically useful approximation to esophageal pressure during sniffs (64, 65). Because these measurements do not require the passage of esophageal or gastric balloons, they are easier for operator and subject. However, pressure transmission may be impaired, particularly when there is significant disease of the lungs (66).


For measurement of maximal sniff pressures, patients are encouraged to make maximum efforts. Sniffs can be achieved only when one or both nostrils are unoccluded, to allow the passage of air. An occluded sniff may be called a “gasp,” and is more difficult for subjects to perform reproducibly. Subjects should be instructed to sit or stand comfortably, and to make sniffs using maximal effort starting from relaxed end expiration. Detailed instruction on how to perform the maneuver is not necessary, and may be counterproductive. However, subjects should be exhorted to make maximal efforts, with a rest between sniffs. Most subjects achieve a plateau of pressure values within 5–10 attempts.

Transdiaphragmatic pressure during sniff.

Esophageal and gastric balloons are passed by the usual technique. Transdiaphragmatic pressure during sniff (Pdi,sn) is reasonably reproducible within normal subjects, although there is wide variability between subjects (61) (Table 3)

TABLE 3. Transdiaphragmatic pressures during maximal static respiratory efforts and maximal sniffs


 (cm H2O)

 (cm H2O)


Definition of abbreviations: Pi,di,max = maximum static transdiaphragmatic pressure; Pdi,sn = transdiaphragmatic pressure during sniff.

Reprinted by permission from Reference 61.

. The values tend to be as large, or larger, than Pdi during maximum static inspiratory efforts (61).

Esophageal pressure during sniff.

The methodology is as for Pdi,sn but with the passage of an esophageal balloon alone.

Nasal sniff pressure.

Pressure is measured by wedging a catheter in one nostril. Various techniques for wedging are available including foam, rubber bungs, and dental impression molding. The subject sniffs through the contralateral unobstructed nostril. The pressure in the obstructed nostril reflects the pressure in the nasopharynx, which is a reasonable indication of alveolar pressure. This in turn approximates esophageal pressure, particularly if the lungs are normal with a mean Pnas/Pes ratio of 0.92 (64, 66). In COPD, nasal sniff pressure (Pnas,sn) tends to underestimate esophageal pressure during sniff (Pes,sn) but can complement Pimax in excluding weakness clinically (67).

Mouth and nasopharyngeal pressures can also be measured, and also reflect alveolar pressure, but are less easy for the subject than nasal pressure and have no significant advantages (65) (Figure 7)


The sniff is a dynamic maneuver and so a pressure measurement system is required with a frequency response of > 10 Hz. This can be achieved by a standard balloon catheter system with a suitable pressure transducer. Use of catheter-mounted transducers has also been described (18).


The sniff is easily performed by most subjects and patients and requires little practice. It is relatively reproducible and has a smaller range of normal values than mouth pressures (61, 62). It is a useful voluntary test for evaluating diaphragm strength in the clinical setting (55), giving equal or greater pressures than maximal static efforts (61, 68).

It is possible to achieve greater transdiaphragmatic pressures by certain maneuvers, such as the modified Mueller maneuver (see previous passages), in highly trained and well-motivated subjects. This may be important in physiological studies but is not usually clinically relevant.

Sniff nasal pressure is technically simple.


The pressures measured during a sniff may be less than maximal static values because of shortening of the inspiratory muscles (pressure–velocity relationship) (69). The average volume change during a sniff is approximately 500 ml with some gas rarefaction, which may be somewhat more than the volume by which gas expands during a static maneuver against a closed airway (63).

Sniffs are difficult or impossible if there is upper airway distortion, and particularly if the nose is completely obstructed. It may be difficult to pass the balloons if there is severe bulbar weakness, but this would not preclude measurement of Pnas.

Sniffs are voluntary maneuvers and, therefore, poorly motivated subjects may perform submaximal efforts. These can often, but not always, be detected as variability tends to be greater than for maximal maneuvers.

The sniff generates rapid pressure changes, so measurement requires a catheter system and transducer with a higher frequency response (see previous passages) than for static maneuvers.


Maximal sniffs have been widely used and validated as reproducible and reliable tests of diaphragm or global inspiratory muscle function. This test can, therefore, be used in research studies, although care must be taken with reproducibility, as in any voluntary maneuver.

The sniff maneuver is a useful part of the clinical evaluation of respiratory muscle strength and correlates well with response to therapy (Figure 6) (68). The normal values are shown in Table 3 (61). There is a wide normal range, reflecting the wide range of normal muscle strength in different individuals. In clinical practice Pdi,sn,max values greater than 100 cm H2O in males and 80 cm H2O in females are unlikely to be associated with clinically significant diaphragm weakness (55). Values of maximal sniff Pes or Pnas numerically greater than −70 cm H2O (males) or −60 cm H2O (females) are also unlikely to be associated with significant inspiratory muscle weakness. However, these reflect the integrated pressure of all the inspiratory muscles and it is possible that there could be a degree of weakness of one or more of these muscles that would not be detected at this level.

Cough Tests
Scientific basis.

Measurement of pressure during cough is of interest because the main expiratory muscles, the abdominal muscles, are also those used in cough. Reduced cough pressure is of clinical importance because it may predispose to chest infections. Also, in some patients technical difficulties preclude measurement of mouth pressure during a static maximal expiratory maneuver (Pemax); in these patients measurement of maximal cough pressures is an alternative measurement technique.

A normal cough consists of four phases: inspiratory, compressive, expulsive, and relaxation (70). For cough to be a measurement of expiratory muscle strength, two aspects require consideration: how standard is the maneuver and what should be measured.

Expiratory muscle strength is influenced by lung volume (71). It is normal to take a deep breath before a maximal cough. Thus, although no instructions are given concerning the magnitude of inspiration, the actual lung volume is probably relatively constant for a given individual during serial measurements. Compression requires a functional glottis; in some disorders, for example bulbar type amyotrophic lateral sclerosis, this may not be present (72). A glottis that does not open immediately causes an uncomfortable choking sensation and makes measurement of mouth pressure (during either a Pemax maneuver or a cough) difficult, but would not exclude obtaining useful measurements from a gastric or esophageal balloon (73).

Theoretically, the peak cough pressure could be measured at the mouth, esophagus, or stomach, but measurements at the mouth have not been reported. Pdi is generated during cough (74) and forced expiration (75) so that Pes,co is always less than Pga,co (76). This Pdi can be substantial in a few subjects (76); it may be due to active contraction of the diaphragm, or passive stretching. In general, gastric pressure can be viewed as a reasonable measure of abdominal muscle strength. The surface EMG from abdominal muscle also varies with cough intensity (77), but difficulties in standardizing it between measurements make this relatively impractical as a measurement of strength.


After passage and positioning of appropriate balloon catheters the subject is asked to cough as forcefully as possible. Visual feedback seems to be helpful, as with the diaphragm (53). Peak pressures are measured between the baseline at relaxed end-expiratory lung volume, and peak pressure (Figure 8)


Voluntary coughs are usually initiated from above FRC. Lung volume may need to be controlled or measured, although this may not be necessary for clinical measurements.

It would, theoretically, be possible to measure the pressure generation during an induced (e.g., with citric acid) cough in patients unable to cough voluntarily, for example, those in intensive care units.

Cough pressures can be large, so it is necessary to have an adequate volume of air in the balloon, to avoid compression of the gas bubble in the balloon and its displacement into the catheter–transducer system.

Normal ranges.

No normal data exist from large studies. Black and Hyatt (78) found Pemax and cough esophageal pressures comparable in a subgroup of their subjects. Other authors have found lower values in patients with COPD (79). The mean maximal Pga,co in a small group of normal subjects 20–75 years of age was 230 cm H2O for men and 166 cm H2O for women, with lower limits of 160 cm H2O for men and 120 cm H2O for women (76).

Scientific Basis

The diaphragm is innervated exclusively by the phrenic nerve and thus phrenic nerve stimulation (PNS) provides a specific means to investigate the diaphragm independent of other inspiratory muscles. Indices that may not be specific for diaphragm contraction when measured during voluntary maneuvers do relate to the diaphragm when they are derived from PNS. Examples include surface recordings of the electromyogram of the costal diaphragm (EMGdi), Pmo or Pes, and phonomyogram (PMGdi). The other major feature of PNS is that it eliminates the influence of the central nervous system. PNS can give important information about the mechanical function of the diaphragm, namely, about how the force of contraction is transformed into pressure, and can be used to confirm whether a contraction is maximal. PNS superimposed on naturally occurring or voluntary contractions (twitch occlusion principle; see subsequent section) can provide an objective estimate of the maximal voluntary pressure that the diaphragm can produce.


Since 1980, PNS has been investigated as a technique for elucidating mechanical aspects of diaphragm function. The 1989 National Heart, Lung, and Blood Institute (NHLBI) workshop on respiratory muscle fatigue identified it as one of the most promising techniques in this field (80).

The stimulus to the nerve, which can be an externally applied electric field or secondary currents surrounding a magnetic field (see Magnetic Stimulation), elicits synchronized activation of motor units and subsequent muscle contraction. The effects of phrenic nerve stimulation can be studied both electrophysiologically (see Stimulation Tests in Section 3 of this Statement) and mechanically (this Section).

Four main PNS techniques have been used, mostly in healthy volunteers, less often in patients. Two of them, needle stimulation (81, 82) and implanted wire stimulation (83), are invasive, with the risk of hematoma and phrenic nerve damage. Needle stimulation is not now recommended, and is not further discussed. Implanted wire stimulation is probably safer, and may be a convenient means to obtain repeated twitches over long periods of time. The two others techniques, transcutaneous electrical PNS (ES) and magnetic stimulation (MS), have been more extensively studied and have minimal side effects.

Subjects in Phrenic Nerve Stimulation

For laboratory and clinical studies subjects should be sitting in a comfortable chair. Headrests may be helpful for ES.

Lung volume.

A major condition for evaluating the pressure response to PNS is adequate relaxation of the respiratory muscles at FRC, or within approximately 500 ml (84).


Most studies have been performed with seated subjects, but twitch transdiaphragmatic pressure (Pdi,tw) seems to be little altered by posture in contrast to sniff transdiaphragmatic pressure (Pdi,sn) and static Pdi (56, 65, 84, 85). This feature could be useful in an intensive care unit (ICU) setting.

Abdominal binding.

Abdominal binding has little effect on voluntary Pdi values (86), but markedly increases Pdi,tw (56, 65, 84, 87). The rationale for binding the abdomen during PNS is to make the contraction closer to isometric than with a compliant abdominal wall; this may be appropriate for physiological studies (56, 88, 89). In a clinical setting it is difficult to standardize a binding technique and so the abdomen is usually unbound.

Transcutaneous Electrical Phrenic Nerve Stimulation
Scientific basis.

An externally applied electrical field induces depolarization of phrenic nerve fibers, at mid-distance between the electrode and the cathode. If the stimulus is intense enough, all fibers are activated synchronously, giving predictable and reproducible results (see Stimulation Techniques in Section 3 of this Statement).


For bilateral ES, the operator should stand behind the seated subject. The skin in the stimulated region is degreased and mildly abraded to decrease its electrical impedance, allowing lower current intensities. Monopolar or bipolar electrodes can be chosen (Figure 9)

. For monopolar electrodes, the anode is usually taped on the skin below the clavicle medially, and the cathode is held in the hand. Monopolar electrodes probably make it easier to find the nerve because a greater number of spots can be tested. However, because the electrical field is less focused than with bipolar electrodes, it may be more difficult to avoid costimulation of the adjacent sternocleidomastoid muscle or the brachial plexus. Bipolar electrodes are more specific but are also slightly more difficult to use. Various models of bipolar electrode are commercially available. They generally include felt tips 5 mm in diameter with an interelectrode distance of approximately 2 cm. When performing stimulation, the tips of the electrode should be soaked in saline and the cathode should be proximal.

The phrenic nerve is usually located beneath the posterior border of the sternocleidomastoid muscle, at the level of the cricoid cartilage (Figure 9). It is easier to locate and to isolate from the brachial plexus in subjects with long and slender necks. A simple technique to locate the nerve is to set the stimulator on repetitive stimulation mode, e.g., at a frequency of 1 Hz, with a relatively low intensity, and to try various sites. Identification of the correct site may be aided by careful observation of the abdomen; it will therefore be desirable to remove the shirt. Once the nerve is identified, the operator marks the spot and the orientation of the electrode. Current intensity is then increased, while monitoring the EMG to obtain supramaximal stimulation (see subsequent paragraphs), which is generally achieved with 30- to 50-mA shocks. This procedure is performed on each side separately, before applying ES bilaterally. It is then advisable to reconfirm that stimulation is supramaximal. It is also possible to judge supramaximality from the plateau of pressure response, but changes can be more difficult to interpret unless the plateau procedure is repeated to check that there is no loss of supramaximality.


A constant-current stimulator capable of delivering square wave shocks of 0.1-millisecond duration and of modulable intensity is used, which should include two synchronized outputs to allow bilateral stimulation. Two triggered EMG amplifiers and a display should be available, so that the muscle action potential (M wave) can be checked online by the operator. Several manufacturers provide complete machines that offer a wide panel of sophisticated stimulation and EMG acquisition options. Stimulators and amplifiers can also be bought separately.


If skillfully performed, ES generates a “pure” diaphragmatic contraction. The corresponding output is thus representative of diaphragm properties alone. ES can be reproducible in skilled hands.


The stimulus intensities required to achieve supramaximal stimulation can be uncomfortable. From the technical point of view, there are several difficulties. First, maintaining optimal contact between the stimulating electrode and the nerve can be difficult. It may be necessary to impose a significant degree of pressure on the soft tissues of the neck, which can be painful for the subject and awkward for the operator, particularly in obese or old subjects, or those with hypertrophy of neck muscles. Skin-taped stimulating electrodes have been used in healthy volunteers (90), but they probably do not guarantee reproducible results in all settings. Neck- and electrode-stabilizing devices have been proposed (10, 85) that can be effective, but are cumbersome. Second, it is sometimes impossible to dissociate PNS from brachial plexus stimulation, particularly at high current intensities. This can be a source of discomfort for the subject, and can theoretically modify the characteristics of the rib cage, with which the contraction of the diaphragm interacts. Third, it can be impossible to locate the nerve, or to do so easily enough to obtain reliable supramaximal stimulation. Because of these difficulties, maintaining a constant symmetric maximal stimulus may need repetitive ES, which in itself can increase twitch pressure by potentiation or the staircase phenomenon (91, 92) (see also Twitch Potentiation).

The technical expertise required for effective ES may, thus, be a source of variability in research studies and limit its use in the clinical field, particularly in demanding settings such as the ICU or exercise.

Magnetic Stimulation
Scientific basis.

The ability of magnetic fields to stimulate nervous structures has long been known (93). Magnetic stimulation creates intense and brief magnetic fields, which, unlike electric currents, are only mildly attenuated by natural barriers such as skin and bone. They can therefore reach deep nervous structures, where stimulation is produced in situ by the electrical fields induced by the rapidly changing magnetic fields (Figure 10)

. The mechanisms of neural response to magnetic stimulation are different from those of the response to electrical stimulation (9497), and therefore the results obtained with the two techniques may have different interpretations. Nevertheless, magnetic stimulation has the advantage of being relatively painless and is thus easily applicable in the clinical setting. Several review articles were discussed by Chokroverty (98).

During the last 10 years, magnetic stimulation has been extensively used to stimulate the central nervous system in conscious humans (99). From the respiratory point of view, magnetic stimulation applied to the cervical spine (CMS) elicits a bilateral diaphragm contraction (100). The coil is centered over the spinous process of the seventh vertebral body (C7), but this does not mean that the seventh roots are stimulated. Depolarization of a nervous structure by magnetic stimulation requires that the stimulating current and the nerve share a common pathway: centering a circular 90-mm coil around C7 would, thus, generally stimulate the third to fifth cervical roots (101103), depolarized in their intraforaminal segment (104, 105). Although it is generally believed that CMS provokes diaphragm contraction through the stimulation of cervical roots, it has been suggested that the C7 CMS magnetic field may reach the phrenic nerves anteriorly, through the neck, and thus stimulate the phrenic nerve trunk at a point more distal than with ES (106) (Figure 11)


Cervical magnetic stimulation also stimulates other elements of the cervical roots and nearby nerves, thus causing some contraction of neck and upper rib cage muscles, as well as diaphragm (100, 107109) (see Comparison between Transcutaneous Electrical Phrenic Nerve Stimulation and Cervical Magnetic Stimulation).


The subject, comfortably seated in a chair, is asked to bend the neck forward slightly. The coil is applied to the back of the neck, its midline coinciding with the axis of the vertebral column (Figure 11). Optimal results are generally obtained with the coil centered around the spinous process of the seventh cervical vertebra (C7), but slightly higher and lower positions should be tried with monitoring of pressure or EMG, although care may be required to obtain satisfactory surface EMG signals. The optimal coil position may vary with the size and neck morphology of the subject. Stimulation intensity is generally set to the maximal output of the stimulator (see Supramaximal Stimulation).

Several manufacturers provide magnetic stimulators suitable for CMS. High-powered machines, capable of producing magnetic fields of 2–2.5 T with a medium-size circular coil, should be used. Doughnut-shaped coils 90 mm in diameter are particularly suitable for generating bilateral diaphragmatic contractions for the measurement of twitch pressure (100, 107110).


Cervical magnetic stimulation provides easy bilateral PNS. It is not painful: The subject simply perceives a contraction of neck muscles that provoke an extension movement, and a hiccup-like sensation. Any operator reasonably familiar with medical or physiological tests can obtain reliable results after a brief period of training. The number of stimulations applied during a given CMS session is often lower than with ES, which reduces the risk of potentiation and the staircase phenomenon (see previous passages, and Twitch Potentiation, subsequent section). Location of the nerves is technically easier with CMS. The risk of falsely low results due to difficulty in locating the nerves and other technical problems is lower with CMS than with ES. In addition, because of its sites of action (cervical roots of the phrenic nerve [101, 103] or intramediastinal segment of the phrenic nerve [106]), CMS can activate diaphragm fibers innervated by an accessory or ectopic phrenic nerve (111) that would not be accessed by ES (112).


Cervical magnetic stimulation lacks the specificity of ES for the diaphragm, because of coactivation of muscles innervated by cervical roots or the brachial plexus. Interpretation of Pdi,tw by CMS is, therefore, not exactly the same as that of Pdi,tw by ES (see subsequent section). Confirming supramaximal stimulation can sometimes be difficult (see subsequent section). Obtaining a reliable EMGdi signal is difficult with CMS, but technical solutions are available (use of shielded EMG cables, transient muting of the EMG amplifiers, etc.) (113) and modern EMG recorders are now designed to support magnetic stimulation. A reasonable distance should be maintained between the stimulating coil and credit cards, computer disks, and the like.

Other Magnetic Stimulation Techniques
Focal magnetic phrenic nerve stimulation.

Small figure of eight- shaped magnetic coils can be used for focal stimulation (focal MS) of the phrenic nerve in the neck unilaterally or bilaterally, at the same point as stimulation by ES (114, 115). Bilateral focal MS is easily applied by an operator standing in front of the subject and gives values for Pdi,tw and the Pga,tw/Pes,tw ratio that are close to ES (115), thus avoiding any problems associated with stimulating upper trunk muscles (see subsequent section). This technique could make bilateral PNS easier in supine patients, as in the ICU, because for ES the operator has to stand behind the patient and for CMS the subject's neck has to lie over the coil, which may be uncomfortable and impede optimal positioning. Unilateral focal MS may allow assessment of the mechanical properties of one hemidiaphragm alone (87, 114, 116).

Anterior magnetic stimulation.

The possibility of evoking a bilateral diaphragm EMG response through anterior magnetic stimulation (antMS) with a 90-mm circular coil similar to that used for CMS, placed flat over the upper part of the sternum, has been described (106). Anterior magnetic stimulation is potentially a simple technique also applicable to supine subjects in difficult settings, but pressure responses remain to be evaluated.

Diaphragmatic Response to Phrenic Nerve Stimulation

Diaphragmatic response to PNS would ideally be measured as work or power. However, length changes and velocity of shortening are invariably ignored and the focus has been on assessing diaphragm force development, either by measurement of pressure, or sound (phonomyography).

Pressure responses to PNS (twitches) are widely used to study diaphragm contraction. They depend not only on diaphragm properties, but also on the load the diaphragm acts against, as for any muscle. This load depends on the mechanical characteristics of the rib cage and abdominal wall (see Techniques for Pressure Measurements).

Measurement of the sounds created by muscle contraction can now be quantified by phonomyography. At present this is a research tool, but it has considerable promise as it is technically easy and noninvasive (117).

Transdiaphragmatic pressure.

In response to PNS, Pdi rapidly rises to a peak, and then decreases exponentially to its baseline value (Figure 12)

to give a characteristic Pdi,tw. The time-to-peak and its first time derivative (rate of rise, dp/dt) depend on several factors, including previous muscle shortening (e.g., increase in lung volume). The amplitude of the twitch reflects the transformation of diaphragm force into pressure and depends on diaphragm strength and contractile properties as well as rib cage and abdominal wall compliance. The dynamics of relaxation of the twitch can be described by the time necessary to reach a Pdi value of 50% of the peak (half-relaxation time) or by the time constant (τ) of an exponential fitted to the pressure–time relationship (87, 118) (see Relaxation Rate in Section 5 of this Statement).

Mouth pressure.

Provided that the diaphragm contracts in isolation and that the corresponding change in alveolar pressure is adequately transmitted to the airways opening, Pmo can in theory reflect diaphragm contraction. In response to PNS, a twitch-shaped negative Pmo swing is seen (Pmo,tw) (Figure 12). In healthy subjects, Pmo,tw closely matches Pes,tw at different lung volumes and correlates with Pdi,tw (119). Conversely, in patients with COPD, Pmo,tw at relaxed FRC is damped and time lagged with respect to Pes,tw, due to an increased airway time constant (120). Both in normal subjects (110) and in patients with diaphragm weakness (121) there is adequate matching of Pmo,tw and Pes,tw with CMS, when precautions are taken to prevent glottic closure. In normal subjects, Pmo,tw measured during CMS is generally more negative than 11 cm H2O (110), depending on lung gas volume (Vl).

Although there are not yet enough data to propose a precise technique to measure Pmo,tw, the following recommendations seem reasonable. Stimulation should be attempted at relaxed FRC when respiratory system recoil pressure is ordinarily zero. When PEEPi is present, the twitch amplitude should be measured starting at the PEEPi level. If a supramaximal stimulation is obtained, the data can be retained, if time to peak tension is normal. If it is prolonged with low amplitude and prolonged relaxation, then abnormal pressure transmission (probably due to glottic closure) should be first suspected, and stimulation should be repeated during a mild expiratory effort against an occluded mouthpiece.


The principal advantage of Pmo,tw is its simplicity and ease of use. Thus, portable Pmo or Psn devices (28) could probably be adapted for combination with CMS as a simple screening test.


The main disadvantage of Pmo,tw is in ensuring the adequacy of pressure transmission from the alveoli to the mouth, particularly in airway obstruction. It seems that glottic closure, which may prevent change in Pmo, is particularly frequent with CMS, although it can occur with ES as well. A technique such as a mild inspiratory effort at FRC (120) or an expiratory effort (110) not only makes the procedure more complicated, but it also changes the meaning of the observed results. Indeed, central nervous system activation and lung volume influence the pressure response to PNS (see Confounding Factors). With CMS, Pmo,tw can be the product of diaphragm contraction but also of neck muscle contraction. The importance of this confounding factor remains to be studied in detail (see Confounding Factors).

Esophageal Pressure

The measurement of esophageal pressure alone in response to PNS provides an intermediate approach between the measurement of Pdi and Pmo. Measuring Pes,tw is less simple than measuring Pmo,tw, but the problem of incomplete pressure equilibration due to glottic closure or airways time constant does not exist. Normal values for Pes,tw are by definition similar to values for Pmo,tw.

Comparison between Transcutaneous Electrical Phrenic Nerve Stimulation and Cervical Magnetic Stimulation
Practical aspects.

The practical aspects of ES and CMS are described in previous sections. In summary:

  1. ES is the original method for generating an isolated contraction of the diaphragm, but for the operator is difficult to master.

  2. CMS is easier and faster to apply, with a lower risk of false results due to technical problems. It is better tolerated by the subject. There is some cocontraction of the upper rib cage and neck muscles, stiffening the rib cage, so that Pdi may be greater than with ES.

Physiological aspects.

The two techniques are comparable with respect to reproducibility (similar within-occasion and between-occasion coefficients of variation) (100, 107) and the time characteristics of the Pdi twitches are very close. When supramaximal bilateral ES and CMS are compared in the same subjects (107109) Pdi,tw values measured during CMS appear consistently higher by approximately 20–25% than Pdi,tw measured during ES, the difference being accounted for by more negative Pes values.

Conclusion and perspectives.

Cervical magnetic stimulation and ES do not provide the same physiological information, so that normal values for PNS-related pressures (see Applications and Perspectives) depend on the technique used. This also means that the results of CMS and ES may be affected differently in diseases affecting both diaphragm and rib cage muscles. The possibility of focused phrenic nerve magnetic stimulation in the neck with small figure-of-eight coils bilaterally (114, 115) will probably reconcile the physiological need for pure diaphragm contraction and the clinical need for simplicity. CMS and other techniques of magnetic stimulation of the phrenic nerve, being easier to use than ES, should then overcome the limitations of ES for large series, exercise, ICU, or intraoperative studies.

Confounding Factors
Supramaximal stimulation.

To provide valid information about the maximal strength of the whole diaphragm, PNS should be bilateral and supramaximal.

Although PNS should be supramaximal if Ptw is to accurately reflect maximal diaphragm mechanical output, this is not necessary when PNS is used to study phrenic conduction time (see Section 3 of this Statement). Failure to achieve supramaximality leads to underestimation of diaphragm mechanical output, overestimation of central drive when using twitch occlusion (see Particular Techniques), and variability. The “gold standard” to achieve a reasonable degree of certainty about the supramaximality of PNS is to establish a recruitment curve using the EMGdi, and stimulation is maximal when there is no further increase in EMGdi in response to an increase in intensity. Increasing stimulus intensity by a further 10–20% provides a reasonable safety margin that compensates for slight changes in the quality of the stimulus. With CMS, it can be difficult to reach a clear plateau in amplitude of the action potential with intensity, but the peak-to-peak amplitudes of the action potentials produced by CMS at the maximal intensity of stimulation with a 2.5-T magnet are not different from the peak-to-peak amplitudes produced by supramaximal bilateral ES (109). The increasing power of stimulators may help reduce this problem in the future.

Lung volume.

Lung volume has a major influence on the ability of the diaphragm to produce pressure, during voluntary static or dynamic maneuvers (13, 36, 122128) and in response to PNS (83, 84, 117, 119, 129134). This is a result of the inverse relationships of length with force in skeletal muscles, and of lung volume with diaphragm length (36, 69, 125). Pdi,tw decreases as lung volume increases, with a prominent reduction in Pes,tw that is close to zero at TLC (83, 129131, 134136) (Figure 13)

. Isovolume changes in rib cage and abdominal configuration also influence Pdi,tw (83, 84, 125).

A long-standing increase in lung volume tends to be compensated for by adaptive mechanisms at the level of the sarcomere, known as “length adaptation” in animals (137141) and probably in humans (131, 142). Thus, the observations made during acute changes in lung volume may not be as relevant to chronic hyperinflation.

How sensitive to changes in lung volume is the pressure response to PNS? Between FRC and TLC, Pdi,tw and Pmo,tw decrease by approximately 3%/100 ml (83, 84, 119, 130, 132), and between RV and FRC by approximately 5%/100 ml (83, 128). These changes appear to be reduced if care is taken to avoid potentiation (129) and may be less in the elderly (131, 134).

Lung volume, and if possible rib cage/abdominal configuration, should be carefully controlled when assessing PNS pressures in research settings. When PNS is repeated in patients with labile lung volumes, FRC should be measured on the day of the study. Assessing lung volume may be less crucial for clinical assessment, recognizing that a change in Pdi,tw or Pmo,tw can reflect changes in diaphragm properties, or lung volume, or both.

Twitch potentiation.

A transient increase in the contractility of a skeletal muscle follows its contraction. This phenomenon is called potentiation (143, 144).

The possibility of twitch potentiation should be taken into account when interpreting studies involving PNS. A period of quiet breathing, e.g., 15 min, should be allowed before recording diaphragm twitches, particularly if maximal maneuvers or sniffs are performed beforehand (145147).

Hypertrophy of neck muscles.

When there is a bilateral paralysis, Pes,tw and Pdi,tw during ES are zero. With CMS, coactivated neck muscles can, theoretically, produce some degree of Pes,tw during CMS. This effect may be small in most subjects (148), but could be larger in patients with hypertrophied inspiratory neck muscles (149).

Particular Techniques
Twitch occlusion

Scientific basis. The degree of activation of a skeletal muscle during voluntary efforts can be assessed by use of the twitch occlusion technique. Twitches produced by electrical stimulation of a parent nerve are superimposed on a voluntary isometric contraction of the muscle (150). The amplitude of the twitch decreases linearly as the strength of the underlying contraction increases. When the activation is maximal it completely suppresses the twitch (150153) (Figure 14)

. It is usual to compare the amplitude of the interpolated twitch with that of the fully potentiated twitch performed on the relaxed muscle.

Normal subjects can voluntarily produce maximal diaphragmatic contraction during inspiratory and expiratory efforts as judged by twitch occlusion (56, 154, 155), as can patients with COPD (131). However, interpolation can be complicated by the effects of series compliance on twitch height (156).

Main results and applications.

Superimposed Pdi,tw decreases linearly with the degree of underlying voluntary Pdi (Pdi,vol) (Figure 14) according to:

The y intercept (a) of Equation 1 closely matches the value of Pdi,tw obtained at relaxed FRC, and its x intercept closely matches Pdi,max. The a/b ratio corresponds to the diaphragm twitch-to-tetanus ratio, which is 0.20–0.25 in most studies, close to the twitch-to-tetanus ratio of skeletal muscles in mammals (157), giving credibility to the technique.

From the twitch occlusion technique, it has been possible to:

  1. Assess the degree of central activation associated with voluntary diaphragm contraction (88, 158).

  2. Discriminate between the central and peripheral components of diaphragm fatigue or weakness (88, 159, 160) (see Types of Fatigue in Section 5 of this Statement).

  3. Detect central inhibition of the drive to breathe (131, 161, 162).

  4. Estimate Pdi,max from submaximal efforts (56).


The technique by which bilateral supramaximal PNS is to be performed should be that established for ES (56, 88, 120, 131, 154, 155, 160, 161, 163167) or CMS (108). A visual feedback of Pdi is provided to the subject and the operator on an oscilloscope or computer screen. A standard procedure is as follows. First, Pdi,max is determined. Then, the subject is asked to produce, in a stepwise manner, fractions of Pdi,max previously marked on the screen of the oscilloscope. While a given step is briefly sustained, one or a few stimuli are delivered so that data for Equation 1 can be built up. To assess the degree of central activation a single stimulus over any inspiratory effort followed by a stimulus in relaxed condition may be sufficient (88).


Twitch occlusion allows an investigator to separate central from peripheral weakness or fatigue of the diaphragm (80).


The twitch occlusion technique is restricted by the difficulties inherent in ES and Pdi measurement. The consistency of the stimulus delivered to the nerves becomes increasingly difficult to maintain as neck muscles are recruited during intense inspiratory maneuvers, although CMS (109) or Pmo (120) could help. When graded voluntary Pdi maneuvers are performed, neck muscles often stay relaxed during low-intensity diaphragm contractions, but are coactivated during high-intensity contractions (70% of Pdi,max and above) (56), which may decrease rib cage distortability. During twitch occlusion from relaxed conditions to maximal effort, diaphragm contraction will interact with a distortable rib cage at low Pdi,vol values, and with a much stiffer rib cage at high Pdi,vol values, complicating the results. This problem may disappear with CMS, because it tends to stabilize the upper rib cage at all levels of Pdi,vol (109). The impact of twitch potentiation on the twitch occlusion procedure is not known. Theoretically, potentiation of Pdi,tw would lead to an underestimate of the central component.

Force–frequency curves.

When stimulation frequency increases, the mechanical output of the stimulated muscle increases up to a plateau (tetanic contraction). Force–frequency curves contain much information about skeletal muscle contractility, and reflect the type and severity of muscle fatigue.

Human diaphragm force–frequency (or, rather, pressure–frequency) characteristics in vivo have been described with unilateral PNS (81, 168, 169) and bilateral PNS (116, 170172), in a limited number of highly motivated subjects. Despite their interest, data remain scarce, because the studies are difficult. First, sequences of supramaximal stimuli are painful and can barely be tolerated, even unilaterally. Second, maintaining adequate sustained supramaximal PNS is demanding, especially as uncomfortable spread of the stimulus to the brachial plexus becomes almost unavoidable with repeated stimuli. Therefore, the use of pressure–frequency curves appears limited and cannot reasonably be proposed for clinical purposes or even for patient-based research. An alternative is to generate a surrogate force–frequency curve using paired phrenic nerve stimuli (173). This approach is acceptable to naive elderly subjects (174) and patients (175).

Normal Values

Values for Pdi,tw or Pmo,tw in normal subjects can be found in more than 40 published studies, irrespective of the technique used (10, 56, 65, 8385, 8890, 100, 107, 110, 115, 116, 118120, 129, 131, 132, 134, 145, 146, 155, 160, 161, 165167, 170173, 175190) (Table 4)

TABLE 4. Twitch transdiaphragmatic pressures measured in a variety of studies

First Author

 (cm H2O, mean)

Bilateral Supramaximal ES: Healthy Young Volunteers
Bellemare (56)34.6
Aubier (118)34.4Needle stimulation
Hubmayr (83)31.4Implanted wire stimulation
Mier (179)8–3315 control subjects referred for suspected diaphragm
weakness; significant overlap between groups; Pdi,tw was
discriminant only for diagnosis of severe diaphragm weakness
Mier (84)24.4Supine position
Gandevia (155)28.0
Mier (82)19.4–29.33 subjects; comparison between transcutaneous electrical
stimulation and needle stimulation
Mador (90)28.9
Wragg (107)29.7Comparison CMS–ES
Eastwood (85)21–283 subjects; validation of a cervical apparatus aiming at
standardizing PNS for repeated studies: data reproducible
among 25 repeated studies
Laghi (108)32.3Comparison CMS–ES
Similowski (109)23.3Comparison CMS–ES
Bilateral Supramaximal ES: Older Healthy Volunteers
Similowski (131)19.8–31.76 subjects aged 50–72 yr, serving as control subjects in a study of
diaphragm properties in hyperinflated patients with COPD
Mancini (182)18.86 subjects aged 50 ± 8 yr, serving as control subjects in a study of
diaphragm function in chronic heart failure
Cervical Magnetic Stimulation: Healthy Young Volunteers
Similowski (100)33.4First-generation stimulator; supramaximality of PNS not
always certain
Wragg (107)36.5Comparison CM–ES
Wragg (146)36.5Study of diaphragm twitch potentiation; values in the table are
the unpotentiated ones
Laghi (184)38.912 subjects
Hamnegard (129, 187)17–34
Similowski (109)27.5Comparison CM–ES
Laghi (108)37.7Comparison CM–ES
Cervical Magnetic Stimulation: Older Healthy Volunteers
Polkey (190)25.4
Patients with COPD
Similowski (131)10–26Bilateral ES
Wanke (183)15.2Bilateral ES
Polkey (134, 190)

Definition of abbreviations: CMS = cervical magnetic stimulation; COPD = chronic obstructive pulmonary disease; ES = electrical stimulation; Pdi,tw = twitch transdiaphragmatic pressure; PNS = phrenic nerve stimulation.

Values represent means or ranges.

. Normal young (20–35 years) male individuals provide the vast majority of the data.

Bearing in mind that the technique used and various methodological factors can exert a major influence on the results, it seems possible to propose the following, for bilateral PNS.

Amplitude of bilateral twitch transdiaphragmatic pressure.

Assuming a correct technique, at FRC, Pdi,tw with ES should be above 15 cm H2O and Pdi,tw with CMS should be above 20 cm H2O.

A Pdi,tw below 15 cm H2O, whatever the PNS technique, should raise high suspicion of diaphragm dysfunction (110, 114, 131, 179, 187, 190196).

Influence of age, sex, and other characteristics.

As with voluntary pressures (26, 128, 197), normal Pdi,tw tends to be lower in women than in men and tends to decrease with age (120, 131, 182, 190). The influence of other factors such as height, race, fitness, etc., is unknown, although data suggest that baseline Pdi,tw in highly fit subjects is not different from normal subjects (185), but with a better resistance to maximal exercise-induced fatigue.

Other characteristics of the twitch transdiaphragmatic pressure response to phrenic nerve stimulation.

The ratio of Pes,tw to Pdi,tw in normal individuals is usually between 0.35 and 0.55.

Amplitude of twitch transdiaphragmatic pressure.

Available data are still scarce, but a value of Pmo,tw during CMS less negative than −11 cm H2O should probably prompt more detailed investigation (110). Pmo,tw during ES tends to be lower, and a value more negative than −8 cm H2O with this technique is probably normal (119).

Applications and Perspectives

The use of PNS-derived pressures to study the mechanical action of the diaphragm assumes that stimulation is bilateral and supramaximal, and that care is taken to control for lung volume and twitch potentiation.

Research applications.

Bilateral ES is an important tool in human diaphragm research, used to study the properties of the diaphragm and the mechanisms of its fatigue independently of volitional influences. The twitch occlusion technique, as the only means to separate the peripheral and central components of diaphragm dysfunction, is also an important, if complex, research tool. The use of paired stimuli could facilitate research on the frequency characteristics of diaphragm fatigue (174).

Cervical magnetic stimulation provides slightly different information from ES, mainly because of rib cage muscle coactivation. However, its simplicity for the subject and the operator makes it a valuable tool for clinical research, especially in difficult settings, such as the ICU, or when repeated studies are needed, such as the evaluation of therapeutic interventions.

Clinical applications.

There are virtually no contraindications to ES and MS, but their use requires some precautions. MS should not be performed in patients with a cardiac pacemaker. Patients with orthopedic implants, even metallic ones, can be studied with MS (e.g., candidates for phrenic pacing after high cervical cord injury) (198) after sufficient time for consolidation.

Clinical use.

Pressure responses to PNS play an important part in the clinical evaluation of inspiratory muscle function. Diaphragmatic weakness can be established and quantified with PNS, particularly when voluntary maneuvers are equivocal. The use of MS with Pmo may be an even simpler nonvolitional measure.

The preoperative assessment of patients for possible phrenic pacing after high cervical cord lesions constitutes a particular application. PNS provides information on phrenic nerve conduction time, which is important in making the decision to undertake pacing (198). It also provides an estimate of the degree of diaphragm disuse atrophy, which is an important determinant of the reconditioning strategy (199).

Given that methodological precautions are rigorously respected, PNS is one of the more reproducible respiratory muscle tests, and this makes it suitable for follow-up studies (see Phrenic Nerve Stimulation in Section 3 of this Statement).


The abdominal muscles are major contributors to respiration, both through their expiratory action on the rib cage and their mechanical linkage with the diaphragm. Their function can be explored by voluntary maneuvers (see Volitional Tests of Respiratory Muscle Strength), with the usual limitations of subjects understanding and cooperation. A nonvolitional test giving information on abdominal muscle mechanical output would, therefore, be of interest.

Contraction of the abdominal muscles can be provoked by their direct electrical stimulation. This technique has been used in humans to study the action of various abdominal muscles on the rib cage (200), and in a study of diaphragm maximal voluntary activation (155). From a therapeutic point of view, its use has been considered to enhance cough in patients with cervical cord injury (201). However, direct electrical stimulation is painful and supramaximality is difficult to achieve. It is also difficult to activate all muscles groups at once.

Contraction of the abdominal muscles can also be provoked by stimulation of their parent nerves and roots (202). Theoretically, this allows supramaximal stimulation (and therefore reproducibility) and, if the stimulus is widespread enough, simultaneous activation of all abdominal muscles. Magnetic stimulation over the vertebral column at the level of the eighth to tenth thoracic vertebra could provide an easy-to-use nonvolitional assessment of abdominal muscle strength and fatigue (202, 203).


The purpose of this Section is to describe the tests used to assess respiratory muscle strength. To test strength, pressures can be measured either during voluntary maneuvers or during involuntary contractions, particularly in response to phrenic nerve stimulation.

A. Volitional Tests of Respiratory Muscle Strength: Volitional tests are often simple for patients to perform, but it can be difficult to be certain that a maximum effort has been made. This can lead to difficulty in the interpretation of low results.

1. Maximum Static Inspiratory and Expiratory Pressure

i. Pimax and Pemax are commonly used, clinically useful measurements. Some individuals have difficulty with the technique and the interpretation of low results can be problematic.

ii. Maximum static transdiaphragmatic pressure (Pi,di,max) provides specific information on diaphragm strength, but can be a difficult maneuver in naive subjects and patients. Pi,di,max has a wide normal range and has limited usefulness in clinical practice.

2. Maximum Sniff Pressures: Maximum sniff efforts can be achieved by patients with little practice; sniff pressures are reproducible and have a narrower normal range than static mouth pressures or Pi,di,max. Sniff esophageal pressure assesses global inspiratory muscle strength and sniff Pdi is a clinically useful measure of diaphragm strength. Sniff nasal pressure provides a useful noninvasive measure of inspiratory muscle strength and has been validated in patients with neuromuscular disease.

3. Maximum Cough Pressure: Cough gastric pressure is measured as an index of abdominal muscle strength. Pga,co is a useful test to supplement Pemax, particularly in patients unable to perform the Pemax maneuver reliably. To date, few data are available for normal values of Pga,co.

B. Nonvolitional Tests of Respiratory Muscle Strength

1. Phrenic Nerve Stimulation: Phrenic nerve stimulation is specific for the diaphragm and is not influenced by the central nervous system.

i. Electrical phrenic nerve stimulation (ES) can achieve selective supramaximal stimulation of the diaphragm, but requires considerable skill, is sometimes uncomfortable for patients, and is difficult to achieve in some clinical settings (e.g., the ICU).

ii. Magnetic phrenic nerve stimulation (MS) is technically easier for the operator and less uncomfortable for the patients. Cervical magnetic stimulation (CMS) elicits a bilateral diaphragm contraction. CMS is less specific than ES, and coactivates muscles innervated by the brachial plexus. Achieving and confirming supramaximal nerve stimulation can be difficult, and recording the diaphragm EMG can pose problems. Unilateral anterior–lateral MS is more specific than CMS, and results are similar to ES. Unilateral MS allows the investigation of hemidiaphragm function, and bilateral anterior–lateral MS reliably achieves supramaximal stimulation.

iii. Twitch transdiaphragmatic pressure (Pdi,tw) provides an index of diaphragm, or hemidiaphragm, strength. Normal values are available and Pdi,tw is a useful clinical measurement. Twitch mouth pressure (Pmo,tw) can provide a noninvasive measure of diaphragm strength, but inadequate transmission of alveolar pressure to the mouth, in patients with airway obstruction or when there is glottic closure, is a substantial practical problem limiting clinical application.

iv. Pdi,tw is a research technique useful to assess the degree of activation of the diaphragm during voluntary efforts (the technique of twitch occlusion) and Pdi,max can be estimated from submaximal efforts.

2. Abdominal Muscle Stimulation: Magnetic stimulation over the eighth to tenth thoracic vertebrae posteriorally, and the recording of twitch gastric pressure, provide a clinically applicable nonvolitional test of abdominal muscle strength. Data on normal values for twitch gastric pressure are limited.


P Pressure (usually mouth pressure if site not specified, as Pimax)

L Length

V Volume

EMG Electromyogram

PMG Phonomyogram

Sites and Modifiers

a, alv Alveolar

ab Abdomen

abw Abdominal wall

ao Airway opening

aw Airway

bs Body surface

di Diaphragm, transdiaphragmatic

es, oes Esophageal

g, ga Gastric

ia Intercostal/accessory muscles

mo Mouth

mus Muscle

nas Nostril, nasal

np Nasopharynx

ph Phrenic

pl Pleural

rc Rib cage

rs Respiratory system (lung and chest wall)

w, cw Chest wall

ant Anterior

post Posterior

e Expiratory

i Inspiratory


co Cough

max Maximal

sn Sniff

tw Twitch

vol Voluntary

Stimulation Descriptors

ELS Electrical stimulation

MS Magnetic stimulation

Ant MS Anterior magnetic stimulation

CMS Cervical magnetic stimulation

fMS Focal magnetic stimulation

PNS Phrenic nerve stimulation

EMG Electromyogram

PMG Phonomyogram

M-wave EMG response to PNS

Lung Volumes

FRC Functional residual capacity

RV Residual volume

TLC Total lung capacity

VC Vital capacity


CNS Central nervous system

HFF High-frequency fatigue

LFF Low-frequency fatigue

MVC Maximal voluntary contractions

SMVC Submaximal voluntary contractions


Give measure, then site, then maneuver, then descriptor. Thus:

Pes,tw,CMS = twitch esophageal pressure following cervical magnetic stimulation

Pe,mo,max = maximal expiratory mouth pressure (usually abbreviated to Pemax)

Pg,co = gastric pressure during a cough

1. Loring SH, Yoshino K, Kimball WR, Barnas GM. Gravitational and shear-associated pressure gradients in the abdomen. J Appl Physiol 1994;77:1375–1382.
2. Loring SH, Kurachek SC, Wohl ME. Diaphragmatic excursion after pleural sclerosis. Chest 1989;95:374–378.
3. Chihara K, Kenyon CM, Macklem PT. Human rib cage distortability. J Appl Physiol 1996;81:437–447.
4. Maton B, Petitjean M, Cnockaert JC. Phonomyogram and electromyogram relationships with isometric force reinvestigated in man. Eur J Appl Physiol 1990;60:194–201.
5. Petitjean M, Bellemare F. Phonomyogram of the diaphragm during unilateral and bilateral phrenic nerve stimulation and changes with fatigue. Muscle Nerve 1994;17:1201–1209.
6. Milic-Emili J. Techniques in respiratory physiology. II. Measurements of pressures in respiratory physiology. In: Otis AB, editor. Techniques in the life sciences: physiology, Vol. P4/II. New York: Elsevier; 1984. p. 1–22, 412.
7. Jackson AC, Vinegar A. A technique for measuring frequency response of pressure, volume, and flow transducers. J Appl Physiol 1979;47: 462–467.
8. Milic-Emili J, Mead JJ, Turner JM, Glauser EM. Improved technique for estimating pleural pressure from esophageal balloons. J Appl Physiol 1964;19:207–211.
9. Onal E, Lopata M, Ginzburg AS, O'Connor TD. Diaphragmatic EMG and transdiaphragmatic pressure measurements with a single catheter. Am Rev Respir Dis 1981;124:563–565.
10. McKenzie DK, Gandevia SC. Phrenic nerve conduction times and twitch pressures of the human diaphragm. J Appl Physiol 1985;58:1496–1504.
11. Javaheri S, Vinegar A, Smith J, Donovan E. Use of a modified Swan-Ganz pacing catheter for measuring Pdi and diaphragmatic EMG. Pflugers Arch 1987;408:642–645.
12. Yernault J. Lung mechanics. I. Lung elasticity. In: Quanjer PH, editor. Standardized lung function testing, Chapter 4. Bull Eur Physiopathol Respir 1983;19(Suppl 5):28–32.
13. Agostoni E, Rahn H. Abdominal and thoracic pressures at different lung volumes. J Appl Physiol 1960;15:1087–1092.
14. Baydur, A., Behrakis PK, Zin WA, Jaeger M, Milic-Emili J. A simple method for assessing the validity of the esophageal balloon technique. Am Rev Respir Dis 1982;126:788–791.
15. Jacobs R, Killam H, Barefoot C, Millar H. Human application of a catheter with tip-mounted pressure and flow transducers. Rev Surg 1972;29:149–152.
16. Millar HD, Baker LE. A stable ultraminiature catheter-tip pressure transducer. Med Biol Eng 1973;11:86–89.
17. Gilbert R, Peppi D, Auchincloss JH Jr. Measurement of transdiaphragmatic pressure with a single gastric-esophageal probe. J Appl Physiol 1979;47:628–630.
18. Evans SA, Watson L, Cowley AJ, Johnston ID, Kinnear WJ. Normal range for transdiaphragmatic pressures during sniffs with catheter mounted transducers. Thorax 1993;48:750–753.
19. Wald A, Post K, Ransohoff J, Hass W, Epstein F. A new technique for monitoring epidural intracranial pressure. Med Instrum 1977;11:352–354.
20. Yellowlees IH. Fibreoptic sensors in clinical measurement. Br J Anaesth 1991;67:100–105.
21. Shapiro S, Bowman R, Callahan J, Wolfla C. The fiberoptic intraparenchymal cerebral pressure monitor in 244 patients. Surg Neurol 1996; 45:278–282.
22. Koska J, Kelley E, Banner MJ, Blanch P. Evaluation of a fiberoptic system for airway pressure monitoring. J Clin Monit 1994;10:247–250.
23. Heritier F, Rahm F, Pasche P, Fitting JW. Sniff nasal inspiratory pressure: a noninvasive assessment of inspiratory muscle strength. Am J Respir Crit Care Med 1994;150:1678–1683.
24. Mead J. Mechanical properties of the lung. Physiol Rev 1961;41:281–330.
25. De Troyer A, Estenne M. Limitations of measurement of transdiaphragmatic pressure in detecting diaphragmatic weakness. Thorax 1981;36:169–174.
26. Black L, Hyatt R. Maximal respiratory pressures: normal values and relationship to age and sex. Am Rev Respir Dis 1969;99:696–702.
27. Koulouris N, Mulvey DA, Laroche CM, Green M, Moxham J. Comparison of two different mouthpieces for the measurement of Pimax and Pemax in normal and weak subjects. Eur Respir J 1988;1:863–867.
28. Hamnegard CH, Wragg S, Kyroussis D, Aquilina R, Moxham J, Green M. Portable measurement of maximum mouth pressures. Eur Respir J 1994;7:398–401.
29. Gandevia SC, McKenzie DK. Activation of the human diaphragm during maximal static efforts. J Physiol 1985;367:45–56.
30. Bigland-Ritchie BR, Furbush FH, Gandevia SC, Thomas CK. Voluntary discharge frequencies of human motor neurones at different muscle lengths. Muscle Nerve 1992;15:130–137.
31. Allen GM, Gandevia SC, McKenzie DK. Reliability of measurements of muscle strength and voluntary activation using twitch interpolation. Muscle Nerve 1995;18:593–600.
32. De Troyer A, Legrand A, Gevenois P-A, Wilson TA. Mechanical advantage of the human parasternal intercostal and triangularis sterni muscles. J Physiol 1998;513:915–925.
33. Agostoni E, Mead J. Statics of the respiratory system. In: Fenn WO and Rahn H, editors. Handbook of physiology: respiration, Vol. 1, Section 3. Washington DC: American Physiology Society; 1964. p. 387–409.
34. Farkas GA, Rochester DF. Functional characteristics of canine costal and crural diaphragm. J Appl Physiol 1988;65:2253–2260.
35. Road J, Newman S, Derenne JP, Grassino A. In vivo length–force relationship of canine diaphragm. J Appl Physiol 1986;60:63–70.
36. Braun NMT, Arora NS, Rochester DF. Force–length relationship of the normal human diaphragm. J Appl Physiol 1982;53:405–412.
37. DeTroyer A, Borenstein S, Cordier R. Analysis of lung volume restriction in patients with respiratory muscle weakness. Thorax 1980;35:603–610.
38. Black LF, Hyatt RE. Maximal static respiratory pressures in generalised neuromuscular disease. Am Rev Respir Dis 1971;103:641–650.
39. Rochester DF. Tests of respiratory muscle function. Clin Chest Med 1988;9:249–261.
40. Rinqvist T. The ventilatory capacity in healthy subjects: an analysis of causal factors with special reference to the respiratory forces. Scand J Clin Lab Invest 1966;18:8–170.
41. Rochester DF, Arora NS. Respiratory muscle failure. Med Clin North Am 1983;67:573–598.
42. Leech JA, Ghezzo H, Stevens D, Becklake MR. Respiratory pressures and function in young adults. Am Rev Respir Dis 1983;128:17–23.
43. Wilson SH, Cooke NT, Edwards RHT, Spiro SG. Predicted normal values for maximal respiratory pressures in Caucasian adults and children. Thorax 1984;39:535–538.
44. Vincken W, Ghezzo H, Cosio MG. Maximal static respiratory pressure in adults: normal values and their relationship to determinants of respiratory function. Bull Eur Physiopathol Respir 1987;23:435–439.
45. Aldrich TK, Spiro P. Maximal inspiratory pressure: does reproducibility indicate full effort? Thorax 1995;50:40–43.
46. Enright PL, Kronmal RA, Monolio TA, Schenker MB, Hyatt RE. Respiratory muscle strength in the elderly. Am J Respir Crit Care Med 1994;149:430–438.
47. McElvany G, Blackie S, Morrison NJ, Wilcox GP, Fairbarn MS, Pardy RL. Maximal static respiratory pressures in the normal elderly. Am Rev Respir Dis 1989;139:277–281.
48. Enright PL, Kronmal RA, Higgins M, Schenker M, Haponik EF. Spirometry reference values for women and men ages 65–85. Cardiovascular Health Study. Am Rev Respir Dis 1993;147:125–133.
49. Gaultier C, Zinman R. Maximal static pressures in healthy children. Respir Physiol 1983;51:45–61.
50. Wagener JS, Hibbert ME, Landau LI. Maximal respiratory pressures in children. Am Rev Respir Dis 1984;129:873–875.
51. Cook CD, Mead J, Orzalesi M. Static volume–pressure characteristics of the respiratory system during maximal efforts. J Appl Physiol 1964;19:1016–1022.
52. Hershenon MA, Kikuchi Y, Loring SH. Relative strengths of the chest wall muscles. J Appl Physiol 1988;65:852–862.
53. Laporta D, Grassino A. Assessment of transdiaphragmatic pressure in humans. J Appl Physiol 1985;58:1469–1476.
54. Hillman DR, Markos J, Finucane KE. Effect of abdominal compression on maximum transdiaphragmatic pressure. J Appl Physiol 1990; 68:2296–2304.
55. Mier-Jedrzejowicz A, Brophy C, Moxham J, Green M. Assessment of diaphragm weakness. Am Rev Respir Dis 1988;137:877–883.
56. Bellemare F, Bigland-Ritchie B. Assessment of human diaphragm strength and activation using phrenic nerve stimulation. Respir Physiol 1984;58:263–277.
57. Alexander C. Diaphragm movements and the diagnosis of diaphragmatic paralysis. Clin Radiol 1966;17:79–83.
58. Hitzenberger K. Das Zwerchfell in gedunden und beranben Zustand. Vienna, Austria: Springer; 1927.
59. Esau SA, Bye PT, Pardy RL. Changes in rate of relaxation of sniffs with diaphragmatic fatigue in humans. J Appl Physiol 1983;55:731–735.
60. Esau SA, Bellemare F, Grassino A, Permutt S, Roussos C, Pardy RL. Changes in relaxation rate with diaphragmatic fatigue in humans. J Appl Physiol 1983;54:1353–1360.
61. Miller JM, Moxham J, Green M. The maximal sniff in the assessment of diaphragm function in man. Clin Sci 1985;69:91–96.
62. Laroche CM, Mier AK, Moxham J, Green M. The value of sniff esophageal pressures in the assessment of global inspiratory muscle strength. Am Rev Respir Dis 1988;138:598–603.
63. Pertuze J, Watson A, Pride N. Limitation of maximum inspiratory flow through the mouth [abstract]. Clin Respir Physiol 1987;23:34S.
64. Heritier F, Rahm F, Pasche P, Fitting J-W. Sniff nasal pressure: a non-invasive assessment of inspiratory muscle strength. Am J Respir Crit Care Med 1994;150:1678–1683.
65. Koulouris N, Vianna LG, Mulvey DA, Green M, Moxham J. Maximal relaxation rates of esophageal, nose, and mouth pressures during a sniff reflect inspiratory muscle fatigue. Am Rev Respir Dis 1989;139: 1213–1217.
66. Uldry C, Fitting J. Influence of airway obstruction on sniff nasal inspiratory pressure. Am J Respir Crit Care Med 1995;151:A414.
67. Uldry C, Janssens JP, de Meralt B, Fitting JW. Sniff nasal inspiratory pressure in patients with chronic obstructive pulmonary disease. Eur Respir J 1997;10:1292–1296.
68. Polkey MI, Green M, Moxham J. Measurement of respiratory muscle strength. Thorax 1995;50:1131–1135.
69. Gandevia SC, Gorman RB, McKenzie DK, Southon FC. Dynamic changes in human diaphragm length: maximal inspiratory and expulsive efforts studied with sequential radiography. J Physiol 1992;457:167–176.
70. Bouros D, Siafakas N, Green M. Cough: physiological and pathophysiological considerations. In: Roussos C, editor. The thorax. New York: Marcel Dekker; 1995. p. 1335–1354.
71. Rahn H, Otis AB, Chadwick LE, Fenn WO. The pressure–volume diagram of the thorax and lung. Am J Physiol 1946;146:161–178.
72. Krietzer S, Saunders M, Tyler HR, Ingram RH. Respiratory muscle function in amyotrophic lateral sclerosis. Am Rev Respir Dis 1978; 117:437–447.
73. Polkey MI, Lyall RA, Green M, Leigh PN, Moxham J. Expiratory muscle function in amyotrophic lateral sclerosis. Am J Respir Crit Care Med 1998;158:734–741.
74. Coryllos PN. Action of the diaphragm in cough. Am J Med Sci 1937; 194:523–535.
75. De Troyer A, Sampson M, Sigrist S, Kelly S. How the abdominal muscles act on the rib cage. J Appl Physiol 1983;58:1438–1443.
76. Kyroussis D, Polkey MI, Hughes PD, Fleming TA, Wood CN, Mills GH, Hamnegard C-H, Green M, Moxham J. Abdominal muscle strength measured by gastric pressure during maximal cough. Thorax 1996;51(Suppl 3):A45.
77. Cox ID, Osman RCA, Hughes DTD, Empey DW. The abdominal electromyogram as an objective measure of cough intensity. Thorax 1983; 38:222.
78. Black LF, Hyatt RE. Maximal respiratory pressures: normal values and relationships to age and sex. Am Rev Respir Dis 1969;99:696–702.
79. Loudon RG, Shaw GB. Mechanics of cough in normal subjects and in patients with obstructive respiratory disease. Am Rev Respir Dis 1967;96:666–677.
80. National Heart, Lung, and Blood Institute. Respiratory muscle fatigue. NHLBI Workshop. Am Rev Respir Dis 1990;142:474–480.
81. Aubier M, Farkas G, De Troyer A, Mozes R, Roussos C. Detection of diaphragmatic fatigue in man by phrenic stimulation. J Appl Physiol 1981;50:538–544.
82. Mier A, Brophy C. Measurement of twitch transdiaphragmatic pressure: surface versus needle electrode stimulation. Thorax 1991;46:669–670.
83. Hubmayr RD, Litchy WJ, Gay PC, Nelson SB. Transdiaphragmatic twitch pressure: effects of lung volume and chest wall shape. Am Rev Respir Dis 1989;139:647–652
84. Mier A, Brophy C, Moxham J, Green M. Influence of lung volume and rib cage configuration on transdiaphragmatic pressure during phrenic nerve stimulation in man. Respir Physiol 1990;80:193–202.
85. Eastwood PR, Panizza JA, Hillman DR, Finucane KE. Application of a cervical stimulating apparatus for bilateral transcutaneous phrenic nerve stimulation. J Appl Physiol 1995;79:632–637.
86. Koulouris N, Mulvey DA, Laroche CM, Goldstone J, Moxham J, Green M. The effect of posture and abdominal binding on respiratory pressures. Eur Respir J 1989;2:961–965.
87. Wilcox PG, Eisen A, Wiggs BJ, Pardy RL. Diaphragmatic relaxation rate after voluntary contractions and uni- and bilateral phrenic stimulation. J Appl Physiol 1988;65:675–682.
88. Bellemare F, Bigland-Ritchie B. Central components of diaphragmatic fatigue assessed by phrenic nerve stimulation. J Appl Physiol 1987; 62:1307–1316.
89. Yan S, Gauthier AP, Similowski T, Faltus R, Macklem PT, Bellemare F. Force–frequency relationships of in vivo human and in vitro rat diaphragm using paired stimuli. Eur Respir J 1993;6:211–218.
90. Mador MJ, Magalang UJ, Rodis A, Kufel TJ. Diaphragmatic fatigue after exercise in healthy human subjects. Am Rev Respir Dis 1993;148: 1571–1575.
91. Desmedt JE, Hainaut K. Kinetics of myofilament activation in potentiated contraction: staircase phenomenon in human skeletal muscle. Nature 1968;217:529–532.
92. Van Lunteren E, Vafaie H. Force potentiation in respiratory muscles: comparison of diaphragm and sternohyoid. Am J Physiol 1993;264: R1095–R1100.
93. D'Arsonval A. Production des courants de haute fréquence et de grande intensité; leurs effets physiologiques. C R Soc Biol 1893;45: 122–124.
94. Olney RK, So YT, Goodin DS, Aminoff MJ. A comparison of magnetic and electrical stimulation of peripheral nerves. Muscle Nerve 1990;13:957–963.
95. Maccabee PJ, Amassian VE, Cracco RQ, Eberle LP, Rudell AP. Mechanisms of peripheral nervous system stimulation using the magnetic coil. Electroencephalogr Clin Neurophysiol Suppl 1991;43:344–361.
96. Schmid UD, Walker G, Schmid-Sigron J, Hess CW. Transcutaneous magnetic and electrical stimulation over the cervical spine: excitation of plexus roots rather than spinal roots. Electroencephalogr Clin Neurophysiol Suppl 1991;43:369–384.
97. Ono S, Oishi M, Du CM, Takasu T. Magnetic stimulation of peripheral nerves: comparison of magnetic stimulation with electrical stimulation. Electromyogr Clin Neurophysiol 1995;35:317–320.
98. Chokroverty S. Magnetic stimulation in clinical neurophysiology. London: Butterworth; 1990.
99. Barker AT, Jalinous R, Freeston IL. Noninvasive magnetic stimulation of human motor cortex. Lancet 1985;8437:1106–1107.
100. Similowski T, Fleury B, Launois S, Cathala HP, Bouche P, Derenne JP. Cervical magnetic stimulation: a new painless method for bilateral phrenic nerve stimulation in conscious humans. J Appl Physiol 1989; 67:1311–1318.
101. Chokroverty S, Shah S, Chokroverty M, Deutsch A, Belsh J. Percutaneous magnetic coil stimulation of the phrenic nerve roots and trunk. Electroencephalogr Clin Neurophysiol 1995;97:369–374.
102. Chen R, Collins S, Remtulla H, Parkes A, Bolton CF. Phrenic nerve conduction study in normal subjects. Muscle Nerve 1995;18:330–335.
103. Zifko U, Remtulla H, Power K, Harker L, Bolton CF. Transcortical and cervical magnetic stimulation with recording of the diaphragm. Muscle Nerve 1996;19:614–620.
104. Maccabee PJ, Amassian VE, Eberle LP, Rudell AP, Cracco RQ, Lai KS, Somasundarum M. Measurement of the electric field induced into inhomogeneous volume conductors by magnetic coils: application to human spinal neurogeometry. Electroencephalogr Clin Neurophysiol 1991;81:224–237.
105. Mills KR, McLeod C, Sheffy J, Loh L. The optimal current direction for excitation of human cervical motor roots with a double coil magnetic stimulator. Electroencephalogr Clin Neurophysiol 1993;89: 138–144.
106. Similowski T, Mehiri S, Attali V, Duguet A, Straus C, Derenne J-P. Comparison of magnetic and electrical phrenic nerve stimulation in assessment of phrenic nerve conduction time. J Appl Physiol 1997; 82:1190–1199.
107. Wragg S, Aquilina R, Moran J, Ridding M, Hamnegard C, Fearn T, Green M, Moxham J. Comparison of cervical magnetic stimulation and bilateral percutaneous electrical stimulation of the phrenic nerves in normal subjects. Eur Respir J 1994;7:1788–1792.
108. Laghi F, Harrison MJ, Tobin MJ. Comparison of magnetic and electrical phrenic nerve stimulation in assessment of diaphragmatic contractility. J Appl Physiol 1996;80:1731–1742.
109. Similowski T, Duguet A, Straus C, Boisteanu D, Attali V, Derenne J-P. Assessment of the voluntary activation of the diaphragm in man using cervical and cortical magnetic stimulation. Eur Respir J 1996;9: 1224–1231.
110. Hamnegard CH, Wragg S, Kyroussis D, Mills G, Bake B, Green M, Moxham J. Mouth pressure in response to magnetic stimulation of the phrenic nerves. Thorax 1995;50:620–624.
111. Rajanna MJ. Anatomical and surgical considerations of the phrenic and accessory phrenic nerves. J Int Coll Surgeons 1947;60:42–53.
112. Sarnoff SJ, Sarnoff LC, Whittenberger JL. Electrophrenic respiration. VII. The motor point of the phrenic nerve in relation to external stimulation. Surg Gynecol Obstet 1951;93:190–196.
113. Luo YM, Polkey MI, Johnson LC, Lyall RA, Harris ML, Green M, Moxham J. Diaphragm EMG measured by cervical magnetic and electrical phrenic nerve stimulation. J Appl Physiol 1998;85:2089–2099.
114. Mills GH, Kyroussis D, Hamnegard CH, Wragg S, Moxham J, Green M. Unilateral magnetic stimulation of the phrenic nerve. Thorax 1995;50:1162–1172.
115. Mills G, Kyroussis D, Hamnegard C, Polkey M, Green M, Moxham J. Bilateral magnetic stimulation of the phrenic nerves from an anterolateral approach. Am J Respir Crit Care Med 1996;154:1099–1105.
116. Bellemare F, Bigland-Ritchie B, Woods JJ. Contractile properties of the human diaphragm in vivo. J Appl Physiol 1986;61:1153–1161.
117. Petitjean M, Ripart J, Couture J, Bellemare F. Effects of lung volume and fatigue on evoked diaphragmatic phonomyogram in normal subjects. Thorax 1996;51:705–710.
118. Aubier M, Murciano D, Lecocguic Y, Viires N, Pariente R. Bilateral phrenic stimulation: a simple technique to assess diaphragmatic fatigue in humans. J Appl Physiol 1985;58:58–64.
119. Yan S, Gauthier AP, Similowski T, Macklem PT, Bellemare F. Evaluation of human diaphragm contractility using mouth pressure twitches. Am Rev Respir Dis 1992;145:1064–1069.
120. Similowski T, Gauthier AP, Yan S, Macklem PT, Bellemare F. Assessment of diaphragm function using mouth pressure twitches in chronic obstructive pulmonary disease patients. Am Rev Respir Dis 1993;147:850–856.
121. Hughes PD, Polkey MI, Kyroussis D, Hamnegard C-H, Moxham J, Green M. Measurement of sniff nasal and diaphragm twitch mouth pressure in patients. Thorax 1998;53:96–100.
122. Rahn H, Otis AB, Chadwick L, Fenn O. The pressure volume diagram of the thorax and lung. Am J Physiol 1946;146:161–178.
123. Byrd RB, Hyatt RE. Maximal respiratory pressures in chronic obstructive lung disease. Am Rev Respir Dis 1968;98:848–856.
124. Decramer M, Demedts M, Rochette F, Billiet L. Maximal transrespiratory pressures in obstructive lung disease. Bull Eur Physiopathol Respir 1980;16:479–490.
125. Loring SH, Mead J, Griscom NT. Dependence of diaphragmatic length on lung volume and thoracoabdominal configuration. J Appl Physiol 1985;59:1961–1970.
126. Heijdra YF, Dekhuijzen PN, van Herwaarden CL, Folgering HT. Effects of body position, hyperinflation, and blood gas tensions on maximal respiratory pressures in patients with chronic obstructive pulmonary disease. Thorax 1994;49:453–458.
127. Wanke T, Merkle M, Zifko U, Formanek D, Lahrmann H, Grisold W, Zwick H. The effect of aminophylline on the force–length characteristics of the diaphragm. Am J Respir Crit Care Med 1994;149:1545–1549.
128. Tolep K, Higgins N, Muza S, Criner G, Kelsen SG. Comparison of diaphragm strength between healthy adult elderly and young men. Am J Respir Crit Care Med 1995;152:677–682.
129. Hamnegard CH, Wragg S, Mills G, Kyroussis D, Road J, Daskos G, Bake B, Moxham J, Green M. The effect of lung volume on transdiaphragmatic pressure. Eur Respir J 1995;8:1532–1536.
130. Smith J, Bellemare F. Effect of lung volume on in vivo contraction characteristics of human diaphragm. J Appl Physiol 1987;62:1893–1900.
131. Similowski T, Yan S, Gauthier AP, Macklem PT, Bellemare F. Contractile properties of the human diaphragm during chronic hyperinflation. N Engl J Med 1991;325:917–923.
132. Yan S, Similowski T, Gauthier AP, Macklem PT, Bellemare F. Effect of fatigue on diaphragmatic function at different lung volumes. J Appl Physiol 1992;72:1064–1067.
133. Gauthier AP, Yan S, Sliwinski P, Macklem PT. Effects of fatigue, fiber length, and aminophylline on human diaphragm contractility. Am J Respir Crit Care Med 1995;152:204–210.
134. Polkey MI, Kyroussis D, Keilty SE, Hamnegard CH, Mills GH, Green M, Moxham J. Exhaustive treadmill exercise does not reduce twitch transdiaphragmatic pressure in patients with COPD. Am J Respir Crit Care Med 1995;152:959–964.
135. Gauthier AP, Verbanck S, Estenne M, Segebarth C, Macklem PT, Paiva M. Three-dimensional reconstruction of the in vivo human diaphragm shape at different lung volumes. J Appl Physiol 1994;76:495–506.
136. Loring SH. Three-dimensional reconstruction of the in vivo human diaphragm shape at different lung volumes [editorial]. J Appl Physiol 1994;76:493–494.
137. Tabary JC, Tabary C, Tardieu C, Tardieu G, Goldspink G. Physiological and structural changes in the catís soleus muscle due to immobilization at different lengths by plaster casts. J Physiol 1972;224:231–244.
138. Farkas GA, Roussos C. Diaphragm in emphysematous hamsters: sarcomere adaptability. J Appl Physiol 1983;54:1635–1640.
139. Farkas GA, Roussos C. Adaptability of the hamster diaphragm to exercise and/or emphysema. J Appl Physiol 1983;53:1263–1272.
140. Oliven A, Supinski G, Kelsen SG. Functional adaptation of diaphragm to chronic hyperinflation in emphysematous hamsters. J Appl Physiol 1986;60:225–231.
141. Matano T, Tamai K, Kurokawa T. Adaptation of skeletal muscle in limb lengthening: a light diffraction study on the sarcomere length in situ. J Orthop Res 1994;12:193–196.
142. Prezant DJ, Aldrich TK, Karpel JP, Lynn RI. Adaptations in the diaphragm's in vitro force–length relationship in patients on continuous ambulatory peritoneal dialysis. Am Rev Respir Dis 1990;141:1342–1349.
143. Wilson DF, Skirboll LR. Basis for posttetanic potentiation at the mammalian neuromuscular junction. Am J Physiol 1974;227:92–95.
144. Vandervoort AA, Quinlan J, McComas AJ. Twitch potentiation after voluntary contraction. Exp Neurol 1983;81:141–152.
145. Mador MJ, Magalang UJ, Kufel TJ. Twitch potentiation following voluntary diaphragmatic contraction. Am J Respir Crit Care Med 1994; 149:739–743.
146. Wragg S, Hamnegard C, Road J, Kyroussis D, Moran J, Green M, Moxham J. Potentiation of diaphragmatic twitch after voluntary contraction in normal subjects. Thorax 1994;49:1234–1237.
147. Desmedt JE, Hainaut K. Modifications des propriétés contractiles du muscle strié au cours de la stimulation électrique répétée de son nerf moteur chez l'homme normal. C R Acad Sci Paris 1967;264:363–366.
148. Mills G, Kyroussis D, Hamnegard C, Wragg S, Moxham J, Green M. Cervical magnetic stimulation of the phrenic nerves in bilateral diaphragm paralysis. Am J Respir Crit Care Med 1997;155:1565–1569.
149. Attali V, Mehiri S, Straus C, Salachas F, Meininger V, Derenne J-P, Similowski T. Influence of neck muscles hypertrophy on mouth pressure response to cervical magnetic stimulation [abstract]. Am J Respir Crit Care Med 1996;153:A786.
150. Merton PA. Voluntary strength and fatigue. J Physiol 1954;67:553–564.
151. Bellanger AY, McComas AJ. Extent of motor unit activation during effort. J Appl Physiol 1981;51:1131–1135.
152. Behm DG, St-Pierre MM, Perez D. Muscle inactivation: assessment of interpolated twitch technique. J Appl Physiol 1996;81:2267–2273.
153. Rutherford OM, Jones DA, Newham DJ. Clinical and experimental application of the percutaneous twitch superimposition technique for the study of human muscle activation. J Neurol Neurosurg Psychiatr 1986;49:1288–1291.
154. Gandevia SC, McKenzie DK. Human diaphragmatic endurance during different maximal respiratory efforts. J Physiol 1988;395:625–638.
155. Gandevia SC, McKenzie DK, Plassman BL. Activation of human respiratory muscles during different voluntary manoeuvres. J Physiol 1990; 428:387–403.
156. Loring SH, Hershensen MB. Effects of series compliance on twitches superimposed on voluntary contractions. J Appl Physiol 1992;73: 516–521.
157. Close RI. Dynamic properties of mammalian skeletal muscles. Physiol Rev 1972;52:129–197.
158. Ferguson GT. Use of twitch pressures to assess diaphragmatic function and central drive. J Appl Physiol 1994;77:1705–1715.
159. Bigland-Ritchie B, Jones DA, Hosking GP, Edwards RHT. Central and peripheral fatigue in sustained maximum voluntary contractions of human quadriceps muscle. Clin Sci Mol Med 1978;54:609–614.
160. McKenzie DK, Bigland-Ritchie B, Gorman RB, Gandevia SC. Central and peripheral fatigue of human diaphragm and limb muscles assessed by twitch interpolation. J Physiol 1992;454:643–656.
161. Allen GM, Hickie I, Gandevia SC, McKenzie DK. Impaired voluntary drive to breathe: a possible link between depression and unexplained ventilatory failure in asthmatic patients. Thorax 1994;49:881–884.
162. Lourenço RV, Miranda JM. Drive and performance of the ventilatory apparatus in chronic obstructive lung disease. N Engl J Med 1968; 279:53–59.
163. McKenzie DK, Gandevia SC. Phrenic nerve conduction times and twitch pressures of the human diaphragm. J Appl Physiol 1985;58:1496–1504.
164. Hershenson MB, Kikuchi Y, Loring SH. Relative strengths of the chest wall muscles. J Appl Physiol 1988;65:852–862.
165. McKenzie DK, Plassman BL, Gandevia SC. Maximal activation of the human diaphragm but not inspiratory intercostal muscles during static inspiratory efforts. Neurosci Lett 1988;89:63–68.
166. Levy RD, Nava S, Gibbons L, Bellemare F. Aminophylline and human diaphragm strength in vivo. J Appl Physiol 1990;68:2591–2596.
167. Allen GM, McKenzie DK, Gandevia SC, Bass S. Reduced voluntary drive to breathe in asthmatic subjects. Respir Physiol 1993;93:29–40.
168. Moxham J, Morris AJ, Spiro SG, Edwards RH, Green M. Contractile properties and fatigue of the diaphragm in man. Thorax 1981;36: 164–168.
169. Bai TR, Rabinovitch BJ, Pardy RL. Near-maximal voluntary hyperpnea and ventilatory muscle function. J Appl Physiol 1984;57:1742–1748.
170. Johnson BD, Babcock MA, Suman OE, Dempsey JA. Exercise-induced diaphragmatic fatigue in healthy humans. J Physiol 1993;460:385–405.
171. Babcock MA, Johnson BD, Pegelow DF, Suman OE, Griffin D, Dempsey JA. Hypoxic effects on exercise-induced diaphragmatic fatigue in normal healthy humans. J Appl Physiol 1995;78:82–92.
172. Babcock MA, Pegelow DF, McClaran SR, Suman OE, Dempsey JA. Contribution of diaphragmatic power output to exercise-induced diaphragm fatigue. J Appl Physiol 1995;78:1710–1719.
173. Yan S, Gauthier AP, Similowski T, Faltus R, Macklem PT, Bellemare F. Force–frequency relationships of in vivo human and in vitro rat diaphragm using paired stimuli. Eur Respir J 1993;6:211–218.
174. Polkey MI, Kyroussis D, Mills GH, Hughes PD, Moxham J, Green M. Paired phrenic nerve stimuli for the detection of diaphragm fatigue in humans. Eur Respir J 1997;10:1859–1864.
175. Hughes PD, Polkey MI, Harrus ML, Coats AJ, Moxham J, Green M. Diaphragm strength in chronic heart failure. Am J Respir Crit Care Med 1999;160:529–534.
176. Dureuil B, Viires N, Cantineau JP, Aubier M, Desmonts JM. Diaphragmatic contractility after upper abdominal surgery. J Appl Physiol 1986;61:1775–1780.
177. Derrington MC, Hindocha N. Measurement of evoked diaphragm twitch strength during anaesthesia: adaptation and evaluation of an existing technique. Br J Anaesth 1988;61:270–278.
178. Similowski T, Fleury B, Launois S, Cathala HP, Bouche P, Derenne JP. Stimulation magnétique cervicale (SMC). Une nouvelle méthode de stimulation phrénique bilatérale utilisable en clinique. Rev Mal Respir 1988;5:609–614.
179. Mier A, Brophy C, Moxham J, Green M. Twitch pressures in the assessment of diaphragm weakness. Thorax 1989;44:990–996.
180. Geddes LA, Mouchawar G, Bourland JD, Nyenhuis J. Inspiration produced by bilateral electromagnetic, cervical phrenic nerve stimulation in man. IEEE Trans Biomed Eng 1991;38:1047–1048.
181. McKenzie DK, Gandevia SC. Recovery from fatigue of human diaphragm and limb muscles. Respir Physiol 1991;84:49–60.
182. Mancini DM, Henson D, LaManca J, Levine S. Respiratory muscle function and dyspnea in patients with chronic congestive heart failure. Circulation 1992;86:909–918.
183. Wanke T, Merkle M, Formanek D, Zifko U, Wieselthaler G, Zwick H, Klepetko W, Burghuber OC. Effect of lung transplantation on diaphragmatic function in patients with chronic obstructive pulmonary disease. Thorax 1994;49:459–464.
184. Laghi F, D'Alfonso N, Tobin MJ. Pattern of recovery from diaphragmatic fatigue over 24 hours. J Appl Physiol 1995;79:539–546.
185. Babcock MA, Pegelow DF, Johnson BD, Dempsey JA. Aerobic fitness effects on exercise-induced low-frequency diaphragm fatigue. J Appl Physiol 1996;81:2156–2164.
186. Hamnegard CH, Wragg S, Kyroussis D, Mills GH, Polkey MI, Moran J, Road J, Bake B, Green M, Moxham J. Diaphragm fatigue following maximal ventilation in man. Eur Respir J 1996;9:241–247.
187. Hamnegard CH, Wragg SD, Mills GH, Polkey MI, Bake B, Moxham J, Green M. Clinical assessment of diaphragm strength by cervical magnetic stimulation of the phrenic nerves. Thorax 1996;51:1239–1242.
188. Mador JM, Rodis A, Diaz J. Diaphragmatic fatigue following voluntary hyperpnea. Am J Respir Crit Care Med 1996;154:63–67.
189. Mador M, Dahuja M. Mechanisms for diaphragmatic fatigue following high-intensity leg exercise. Am J Respir Crit Care Med 1996;154:1484–1489.
190. Polkey M, Kyroussis D, Hamnegard C-H, Mills G, Green M, Moxham J. Diaphragm strength in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1996;154:1310–1317.
191. Aubier M, Murciano D, Lecocguic Y, Viires N, Jacquens Y, Squara P, Pariente R. Effect of hypophosphatemia on diaphragmatic contractility in patients with acute respiratory failure. N Engl J Med 1985; 313:420–424.
192. Aubier M, Murciano D, Viires N, Lebargy F, Curran Y, Seta JP, Pariente R. Effects of digoxin on diaphragmatic strength generation in patients with chronic obstructive pulmonary disease during acute respiratory failure. Am Rev Respir Dis 1987;135:544–548.
193. Laroche CM, Cairns T, Moxham J, Green M. Hypothyroidism presenting with respiratory muscle weakness. Am Rev Respir Dis 1988;138: 472–474.
194. Laroche CM, Carroll N, Moxham J, Green M. Clinical significance of severe isolated diaphragm weakness. Am Rev Respir Dis 1988;138: 862–866.
195. Aubier M, Murciano D, Menu Y, Boczkowski J, Mal H, Pariente R. Dopamine effects on diaphragmatic strength during acute respiratory failure in chronic obstructive pulmonary disease. Ann Intern Med 1989;110:17–23.
196. Murciano D, Rigaud D, Pingleton S, Armengaud MH, Melchior JC, Aubier M. Diaphragmatic function in severely malnourished patients with anorexia nervosa. Effects of renutrition. Am J Respir Crit Care Med 1994;150:1569–1574.
197. Hsiun-ing C, Kuo CS. Relationhip between respiratory muscle function and age, sex, and other factors. J Appl Physiol 1989;66:943–948.
198. Similowski T, Straus C, Attali V, Duguet A, Jourdain B, Derenne J-P. Assessment of the motor pathway to the diaphragm using cortical and cervical magnetic stimulation in the decision making process of phrenic pacing. Chest 1996;110:1551–1557.
199. Nochomovitz ML, Hopkins M, Brodkey J, Montenegro H, Mortimer JT, Cherniack NS. Conditioning of the diaphragm with phrenic nerve stimulation after prolonged disuse. Am Rev Respir Dis 1984; 130:685–688.
200. Mier A, Brophy C, Estenne M, Moxham J, Green M, De Troyer A. Action of abdominal muscles on rib cage in humans. J Appl Physiol 1985;58:1438–1443.
201. Linder SH. Functional electrical stimulation to enhance cough in quadriplegia. Chest 1993;103:166–169.
202. Kyroussis D, Polkey MI, Mills GH, Hughes PD, Moxham J, Green M. Stimulation of cough in man by magnetic stimulation of the thoracic nerve roots. Am J Respir Crit Care Med 1997;156:1696–1699.
203. Kyroussis D, Mills GH, Polkey MI, Hamnegaard C-H, Koulouris N, Green M, Moxham J. Abdominal muscle fatigue after maximal ventilation in humans. J Appl Physiol 1996;81:1477–1483.

Respiratory muscle contraction depends on electrical activation of the muscles. Influenced both by involuntary and voluntary inputs, the electrical impulses originate in the respiratory neurons of the brainstem, are carried via motor nerves, transmit through neuromuscular junctions, and propagate throughout muscle membranes. Failure at any of these sites can result in dyscoordination and reversible or irreversible muscle weakness. The task of electrophysiologic tests is to assess the integrity of the respiratory neuromotor apparatus.

There are two main types of electrophysiologic tests of respiratory muscle function: electromyography and stimulation tests. These tests are interrelated, and they are related also to tests of mechanical action of the respiratory muscles, described elsewhere in this Statement.


Electromyography (EMG) is the art of describing myoelectric signals (1), the electrical manifestations of the excitation process elicited by action potentials propagating along muscle fiber membranes. The EMG signal is detected with electrodes, and then amplified, filtered, and displayed on a screen or digitized to facilitate further analysis. Electromyography of respiratory muscles can be used to assess the level and pattern of their activation, so as to detect and diagnose neuromuscular pathology, and, when coupled with tests of mechanical function, to assess the efficacy of the muscles' contractile function (see Electromechanical Effectiveness in Section 6 of this Statement).

Scientific Basis
Single fiber action potential.

Depolarization of a muscle fiber membrane, caused by the flow of ions across the sarcolemma, generates an electric field outside the muscle fiber, which can be detected by extracellular recording electrodes as voltage changes over time; this voltage transient is known as the action potential. Although the transmembrane potential changes generated by depolarization of a given fiber are always identical in shape and amplitude (the “all or nothing” phenomenon), the shape and amplitude of a recorded action potential depend on factors such as the orientation of the recording electrodes with respect to the active muscle fibers, the distance between the muscle fibers and the electrodes, the filtering properties of the electrodes, and the muscle fiber action potential conduction velocity (2).

In humans, muscle fiber conduction velocity ranges from 2 to 6 m/second (3), depending on passive and active components of the muscle fiber membrane. The passive components (cable properties) include capacitance per unit length (proportional to fiber circumference) and internal resistance (inversely proportional to the square of the fiber diameter). The active components (membrane excitability) depend on ion gradients across the membrane and properties of the ion gating channels, which in turn are influenced by electric field strength, temperature, and chemical milieu (especially pH and Ca2+ concentrations). Muscle fiber conduction velocity has been shown to vary with fiber diameter (4), temperature (3), electrolyte gradients across the cell membrane (5), pH (5), and fatigue (1, 6).

Single motor unit action potential.

Each motor unit is composed of a number of individual muscle fibers innervated by a single anterior horn cell. All individual fibers within a motor unit are activated almost simultaneously. The amplitude and shape of the resulting motor unit action potential (MUAP) are influenced not only by all the factors that can affect single fiber action potentials, but also by such factors as the number of fibers within the motor unit, the spatial dispersion of motor unit fibers, differences in length of the motor neuron terminal axons, and possibly fiber-to-fiber differences in action potential conduction velocity (2, 7).

Summation of motor unit signals.

Compound muscle action potentials (CMAPs) represent the summated electrical activity generated by all motor units synchronously activated by nerve stimulation (see subsequent passages). The observed CMAP is influenced by the number of activated motor units, their synchronization, the shape of individual MUAPs, and cancellation of opposite phase potentials.

The interference pattern EMG results from the temporal and spatial summation of asynchronously firing MUAP trains during spontaneous muscle contractions, when individual MUAPs can no longer be distinguished (8). The observed interference pattern EMG is thus a function of the number of active motor units, their firing rates and synchronization, the shapes of their individual MUAPs (in turn dependent on all the factors listed here previously), and cancellation of opposite phase potentials (2, 8).

EMG Equipment
Recording electrodes.

Electromyography signals can be detected as the difference between the signal from an electrode placed on or in the muscle under investigation (active electrode) and the signal from another electrode placed in an electrically silent region (“indifferent” or reference electrode). This electrode arrangement is usually referred to as a monopolar electrode. When the two electrodes connected to a differential amplifier are positioned on or in the same muscle under investigation, the electrode arrangement is usually referred to as bipolar. Optimal EMG signals depend on the use of electrodes with appropriate configuration and fixed geometry, on maintenance of electrode position relative to the muscle, on alignment of the electrodes with respect to fiber direction, on selection of sites with relatively low density of motor end plates, and on avoidance of signal disturbances.

Electrodes can be placed on the skin overlying the neck, the chest wall muscles, or the area of apposition of the diaphragm to the chest wall; they can be swallowed into the esophagus to measure crural diaphragm EMG; or they can be inserted into the respiratory muscle of interest, using needle, wire, or hook electrodes. Selection of an appropriate electrode system requires consideration of advantages and disadvantages specific to the technique and the context of the study (Table 1)

TABLE 1. Types of recording electrodes for respiratory muscle electromyograms

Type of Electrode


Surface electrodes
Chest wallNoninvasiveCross-talk
Large volume samplingVariable filtering
EsophagealLess influenced by body habitusDiscomfort
Less cross-talkUnreliable in diaphragm hernia
Intramuscular electrodesLess influenced by cross-talkDiscomfort
Single motor unit recordings possibleDifficult to place
Small pneumothorax risk

Possible sampling error

Surface electrodes.

Surface electrodes have been used to measure activity of diaphragm, intercostal, scalene, abdominal, and accessory muscles. After the skin is shaved, cleaned, and dried, electrodes are placed over or as close as possible to the muscle to be investigated and are secured with optimal skin contact. The placement is determined by palpation and by the investigator's knowledge of respiratory muscle anatomy, but no standards exist for electrode design or positioning. Furthermore, there is no consensus on methods either to maintain electrode orientation with respect to muscle fibers and innervation zones or to control for influences of variable muscle-to-electrode distance (as, e.g., with variations in the amount of subcutaneous fat), or for cross-talk from adjacent muscles.

Advantages of surface electrodes are their noninvasive nature and their ability to sample a large number of motor units. For many individual nondiaphragm respiratory muscles, however, their proximity to one another and to nonrespiratory trunk muscles makes surface electrode recordings unreliable. Examples include cross-talk from scalenes and platysma in recordings of sternocleidomastoids (9), from external oblique and pectoralis in recordings of intercostal muscles, and among various abdominal muscles (10, 11). Furthermore, variations in interindividual body habitus, for example, subcutaneous fat tissue or deformity of the chest wall, produce variable muscle-to-electrode filtering effects.

Esophageal electrodes.

Esophageal electrodes are metal electrodes mounted on a catheter, which is inserted via the nose or mouth and positioned with the electrode rings at the level of the crural diaphragm. In adults, the motor innervation zone of the crural diaphragm lies 1–3 cm cephalad to the gastroesophageal junction, with the left side approximately 1 cm cephalad to the right (12).

Esophageal electrode catheters are often equipped with gastric balloons and weights on the proximal end, which anchor them in position when outward traction pulls the gastric balloon snugly against the gastroesophageal junction. This feature may limit motion of the electrodes with respect to the esophagus, but it does not prevent either diaphragm movement relative to the electrodes, or the resulting artifacts (1316). Motion artifacts are minimized in semistatic contractions.

A more reliable method to reduce electrode positioning artifacts during dynamic maneuvers is continuous optimization of diaphragm-electrode positioning. An electrode array of eight rings mounted 1 cm apart on a catheter is introduced and adjusted to provide optimal EMG activity from the central pair of electrodes (17, 18) (Figure 1A)

. A computer program samples all electrodes continuously, and selects the pair closest to the crural diaphragm at any point in the respiratory cycle. Application of a double subtraction technique using the difference between signals obtained from each electrode pair caudal and cranial to the crural diaphragm (19) further enhances the signal-to-noise ratio.

Advantages of esophageal recordings are that they are not influenced by obesity and that, when used with or combined into the same catheter as esophageal and gastric pressure monitors (12), they enable simultaneous recordings of diaphragm EMG and transdiaphragmatic pressure. Disadvantages include the discomfort and the remote risks of regurgitation, aspiration, and vagally mediated bradycardia associated with their placement. Diaphragmatic hernia may be a source of error in esophageal recordings. In theory, the EMG from esophageal electrodes may not be representative of the diaphragm as a whole, because it samples only the crural portions of the diaphragm. However, although crural–costal dissociation during breathing has been demonstrated in animals (20, 21), it appears not to be a problem in humans (2224).

Intramuscular electrodes: advantages.

Intramuscular electrodes provide relatively selective recordings from nondiaphragm respiratory muscles (911, 2530) with sufficient discrimination of individual motor unit activity to allow evidence of denervation or myopathy to be detected. Whitelaw and Feroah (30) have provided a detailed description of a safe method to record single motor unit activity in intercostal muscles with wire electrodes. A number of investigators have demonstrated techniques for placement of monopolar or bipolar needle electrodes in the human diaphragm, either by a medial subcostal approach (31, 32) or by a lower intercostal approach (33, 34) (Figure 2)

, sometimes assisted by real-time ultrasound. Fine wire electrodes for single motor unit recordings have also been implanted in the right hemidiaphragms of humans (13, 35). Although the flexible nature of wire electrodes makes them relatively stable during volume changes of up to 1.5 L around the FRC, artifactual changes in recording conditions occur with larger volume changes (13). Wire implants are often used in upper airway muscle studies (see Electromyography in Section 8 of this Statement).

As far as advantages and disadvantages are concerned, intramuscular electrodes are optimal for analysis of action potentials in the assessment of myopathic changes and for comparing single motor unit firing frequencies among different respiratory muscles and clinical conditions (13) (Figures 3 and 4)

. Cross-talk is less of a problem with implanted than surface electrodes, although it is not completely eliminated. Implanted electrodes are difficult to place, however, and are potentially less useful than surface or esophageal recordings for quantifying global respiratory muscle activity. For intercostal recordings and especially for triangularis sterni and diaphragm recordings (26), there is a small risk of pneumothorax. As with any skin-penetrating technique, there are risks of bleeding and bruising. With the use of sterile disposable needles or wires, there is no practical risk of transmission of infectious disease.

EMG signal processing.

The preferred amplifier design for detection of weak myoelectric signals is a differential amplifier, which amplifies the difference between two paired inputs and thereby eliminates signals, such as 50 or 60 Hz from power lines, that have common influences on both outputs (common mode rejection). Input impedance is high to minimize loss of voltage across the two active electrodes. Bandpass filtering is usually employed, with a high-pass filter to remove the signal's direct current component and a low-pass filter set below the data acquisition frequency to avoid distortion (by aliasing) of the signal caused by undersampling. Filter settings will vary, depending on the type of electrode and the application; for surface or esophageal recordings, a band width of 10 to 1,000 Hz is often used.

For simple measurements of motor latencies or CMAP amplitudes, the data display can be as simple as a storage oscilloscope equipped with a camera. A more versatile system, essential for frequency domain measurements, is a computer with an analog-to-digital converter. Because in digitally sampled signals only spectral components of frequencies lower than half the sampling frequency can be observed (the Nyqvist theorem), the sampling frequency should be chosen accordingly.

To provide an analog signal proportional to “average” EMG activity at any point in time, many investigators subject the raw EMG signal to rectification and “leaky” integration. Unless sophisticated analog gating and/or filtering techniques are employed, however, the effects of multiple artifacts detailed in the following sections limit this technique to providing only a crude estimation of the level of muscle activation.

Data Analysis
Time domain EMG analysis.

Time domain EMG analysis represents the electrical activity of the muscles as a function of time. At low levels of contraction, isolated MUAPs can be distinguished and analyzed by such time domain indices as signal amplitude and different types of integrated EMG, including full wave rectified and averaged signal (FRA) and root mean square of the signal (RMS) (36), all of which increase as a function of the number of motor fibers in the unit. Frequency-related indices, such as rise time and duration, can also be measured (36, 37); these are influenced more by action potential conduction velocity and the width of the innervation zone than by the size of the motor unit.

With increasing levels of contraction, more motor units are recruited, and the firing rate increases. The resulting interference pattern EMG can be analyzed by indices such as amplitude, RMS, and FRA, and by frequency-related indices such as zero crossing distance and turn-point distance (36).

Frequency domain analysis.

Frequency domain analysis is a technique to express EMG power as a function of frequency. EMG power is intimately related to its frequency characteristics (8), and frequency characteristics in turn are related to muscle membrane conduction velocity, to filtering properties of the electrodes, to muscle–electrode distances, and to noise (36). Frequency domain analysis greatly simplifies the evaluation of all of these factors, which are difficult to evaluate in the time domain.

The relative contributions of high and low frequencies to EMG signals can be estimated crudely by splitting the signal, filtering the signal with different bandpass filters, integrating the output of each bandpass filter, and comparing the two integrals, to yield a ratio of high-frequency power (e.g., 130–250 Hz) to low-frequency power (e.g., 30–50 Hz), the H/L ratio (38). More useful information can be obtained by power spectral analysis, for which a “window” of time domain EMG data is digitized and subjected to a computerized fast Fourier transform. The fast Fourier transform components are squared and their products are calculated, giving the power spectrum, which graphs the power of the signal as a function of frequency (Figure 5)

. To avoid artifactual overestimation of high-frequency content, it is necessary to condition the window of data by tapering its amplitude at the beginning and end of the window or by replacing all data before the first and after the last zero crossing with zeros. These “shaping” processes lead to a slight underestimation of high-frequency content of the EMG (39).

Numerical quantification of the power spectrum is possible by calculating the spectral moments (36). Spectral moments (M) of order n are defined as:

where P is power density, f is frequency, i is the index over which the power density product is summed, i = 0 is the direct current component, and imax is the index associated with the highest frequency in the spectrum.

The spectral moment of zero order (M0) and the root mean square (RMS = M01/2/p, where p is the number of points in the sample) are indices of total EMG power. Theoretically, RMS reflects the force output of the muscle (8). However, both M0 and RMS are influenced by a number of other parameters, especially conduction velocity (8). In applications in which the power spectrum is expected to shift, as in fatigue, the first-order spectral moment (M1) may be a more useful index of muscle activation, because it is not affected by changes in action potential conduction velocity (36). Quantification of the distribution of power in the spectrum can be obtained by calculating the center frequency (fc = M1/M0), also known as the mean or centroid frequency. Figure 5 shows how EMG signal power spectra are influenced by interelectrode distance.

Filter effects.
Muscle-to-electrode distance.

Increasing muscle-to-electrode distance results in reduced signal amplitude with relatively larger attenuation of high- than low-frequency power, so that the relative distribution of power in the spectrum is altered and fc is reduced (1, 1417, 40) (Figure 6)


Interelectrode distance and alignment with respect to fiber direction (bipolar arrangement).

Both interelectrode distance and the orientation of the electrodes with respect to fiber direction can affect power spectra (1, 18, 41). As shown in Figure 5, reductions in interelectrode distance reduce signal power, with relatively higher attenuation of low-frequency components. With bipolar recordings of EMG, electrode orientation with respect to fiber direction and interelectrode distance should be standardized, and investigators should recognize that these relationships are likely to change with muscle contraction.

Signal disturbances.
Electrode motion-induced artifacts.

Movement of the electrode or a change in the pressure on the electrode results in a redistribution of the charges in the resistive–capacitive interface between the electrode and the tissues, causing relatively large amplitude artifacts with low frequency (mostly below 20–25 Hz) (39, 42). Most of the motion artifact can be filtered out with high-pass filters. However, some EMG power occurs at frequencies below 25 Hz, and even the most efficient filter attenuates some of the power above its cutoff frequency. Thus, high-pass filtration inevitably leads to loss of low-frequency power from the EMG signal. When power spectrum analysis is applied, the power related to electrode movement can be reduced (42) or replaced by an extrapolation of the diaphragm EMG power to those frequencies (39).


Background noise, signals of unidentifiable origin, can be assumed to have constant power density over the frequency region of interest in EMG recordings. On the basis of this assumption, the noise component of the signal-to-noise ratio is usually estimated from power density values obtained in the uppermost frequency range of the EMG power spectrum (39, 43, 44). Noise originating from modern electrophysiologic instruments is not usually a problem, but ancillary equipment, such as pressure or motion sensors, pumps, or respirators, may introduce noise in various frequency ranges.

Disturbances from power lines.

Sinusoidal alternating currents (ACs) originating from power lines can usually be reduced to a negligible level by proper shielding of electrode cables and connections, by using amplifiers with high common mode rejection ratio, and by connecting all instruments to the same ground point. Residual power line disturbances can be filtered out with “notch filters,” or, alternatively, the power of the affected frequency and its harmonics can be excluded in power spectral analysis.

Potential artifactual influences of physiologic origin.
Cross-talk signals.

These are signals originating from muscles other than the muscle being investigated. The best-described source of cross-talk is electrical activity from the heart (EKG). Esophageal recordings of diaphragm EMG are particularly susceptible to cardiac cross-talk, which provides a signal with some 10 times the power of the diaphragm EMG but with a much lower fc (39, 42). For measurements of stimulation-elicited CMAPs, EKG artifacts are relatively easy to avoid by triggering the stimulus from the QRS complex with an appropriate delay. Alternatively, any record visibly superimposed on the QRS complex can simply be discarded.

For time domain analyses over long periods of time or for power spectral analyses, EKG artifacts pose a significant problem. Heart rates often change, as, for example, with inspiration and with exercise. The frequency content of the EKG is lower than that of EMG activity, but there is considerable overlap (39, 43). Thus, if the high-pass filter is set high enough to eliminate most of the EKG activity, it will also eliminate much of the EMG activity. When EMG frequency content might be shifting toward lower frequencies, as during increasing muscle effort, excessive high-pass filtration will result in spurious reductions in measured RMS. Methods to eliminate the cardiac activity from esophageal recordings of diaphragm EMG by template subtraction (45) have been proposed. However, because the amplitude and shape of the electrocardiogram change with changes in lung volume, the safest method to avoid cardiac influences on diaphragm EMG signals still appears to be selection of signal segments that are free of cardiac activity (39).

Another source of cross-talk in esophageal recordings of diaphragm EMG is esophageal peristalsis (39). The esophageal peristalsis signals show a relatively large amplitude and relatively low-frequency content. In general, signal segments contaminated by esophageal peristalsis, usually easily identified as strong, slow esophageal pressure waves, should be excluded from analysis.

Other examples of cross-talk sources include contamination of diaphragm EMG data recorded from electrodes on the lower rib cage with abdominal or intercostal muscle activity and contamination of intercostal muscle EMGs by interference from pectorals, abdominals, or diaphragm (46).

Innervation zones.

Electrodes positioned close to or over an innervation zone (a region with a high density of motor end plates) produce complex interference patterns, because the action potentials elicited by firing of an individual motor unit may propagate in opposite directions relative to the electrodes (2, 47, 48). Characteristically, there is reduced total power and increased high-frequency content (49). This effect is maximal when bipolar electrodes are oriented parallel to fiber direction; esophageal electrodes, which are arranged approximately perpendicular to fibers, are less susceptible (18).

Influence of changes in muscle length or chest wall configuration.

Changes in chest wall configuration systematically affect the amplitude and frequency content of evoked CMAPs measured with chest surface or esophageal electrodes (13, 50). With spontaneous EMG activity, artifact-free signals recorded from esophageal electrodes are not systematically affected by changes in chest wall configuration (17, 18, 25, 50, 51). Because conduction velocity is not significantly affected by muscle length (52), most of the chest wall configuration effects on EMG are probably due to muscle-to-electrode distance or orientation changes.

Influence of changes in muscle temperature.

Muscle temperature increases during exercise, because of the increase in blood flow to the muscle (53, 54), and metabolic heat production (55). Because the propagation velocity of the muscle fiber action potential is correlated with temperature (56, 57), EMG frequency content must also be temperature dependent. No method for control of or correction for temperature in recordings of respiratory muscles has yet been described.


See Table 2

TABLE 2. Applications for respiratory muscle electromyograms

Type of Test


Needle EMGDenervation↓MUAPs, fibrillation potentials, and
positive sharp waves
Demyelination↓MUAPs, without potentials
Chronic denervation↓No., ↑size of MUAPs
MyotoniaMyotonic discharges
MyopathyShort, polyphasic potentials
Interference pattern EMG signalParalysis or dyscoordinationRespiratory muscle activation pattern
Quantification of neural driveChanges in FRA or RMS
Efficiency of contractionΔPdi/ΔEdi

Spectral analysis

Definition of abbreviations: ΔPdi/ΔEdi = ratio of tidal respiratory change in transdiaphragmatic pressure to tidal respiratory change in integrated diaphragm EMG; EMG = electromyography; FRA = full wave rectified and averaged signal; MUAP = motor unit action potential; RMS = root mean square of the signal.


Single fiber and motor unit analysis.

Single fiber and motor unit signal analyses are useful for diagnosis of nerve or muscle pathology. For the diagnosis of neuromuscular disease, limb muscles are more commonly studied than respiratory muscles, because they are more readily accessible. A number of investigators have demonstrated the usefulness of needle electromyography of the diaphragm for the diagnosis of neuromuscular disease, particularly neuropathic processes such as Guillain–Barré syndrome, lower motor neuron involvement with spinal cord injury, and polyneuropathy of critical illness (31, 32, 34).

Bolton (34) has pointed out that the relatively high-frequency, low-amplitude potentials of the normal diaphragm are often difficult to differentiate from myopathic potentials. Nevertheless, several neuromuscular diseases present primarily with respiratory muscle weakness; as experience is gained with single fiber and motor unit analysis of respiratory muscles, these techniques applied to respiratory muscles may provide the earliest evidence of the neuro- or myopathic process.

Interference pattern signal.

The interference pattern EMG (raw EMG from surface electrodes) of respiratory muscles is useful for the determination of the timing and level of muscle activation during respiratory activities. Thus, EMGs can help to determine which of the many respiratory muscles are active in various phases of respiration, in various body positions, in various states of consciousness, and in various clinical conditions (see Electromyography in Section 8 of this Statement). Specifically, the absence of voluntary or involuntary EMG activity can be used as evidence of paralysis of specific respiratory muscles. EMGs can also help to quantify the respiratory muscle activation responses to loaded breathing and to CO2-stimulated breathing or to monitor and control mechanical ventilation (58, 59). Furthermore, when related to pressure or force developed by respiratory muscles, EMGs can help to assess the electromechanical “efficiency” of respiratory muscle function (see Electromechanical Effectiveness in Section 6 of this Statement).

Interindividual comparisons of absolute FRA or RMS values do not meaningfully reflect respiratory drive, because of varying filtering influences of electrode placement relative to the contracting muscles and/or anatomic differences between subjects, for example, the amount and type of interlaying tissue. However, changes in these indices in response to interventions such as changes in inspired CO2 concentration, loaded breathing, exercise, states of consciousness, drugs, or other influences, reflect changing motor output of the central nervous system (CNS) to respiratory muscles. Because some respiratory muscles are silent during quiet breathing, it is not practical to normalize respiratory muscle EMG activity to that observed during resting tidal breathing. It is often more practical to normalize EMG activity to that observed during sniff inhalations or maximal inspiratory efforts (22).

The interference pattern EMG of respiratory muscles may also be useful for the assessment of respiratory muscle fatigue (see Electromyography in Section 5 of this Statement). Localized muscle fatigue is accompanied by a reduction in muscle fiber action potential conduction velocity (1), which is reflected in the diaphragm EMG power spectrum as a shift toward lower frequencies (41, 52). This “spectral shift” is most commonly quantified as a reduction in fc. Spectral shifts have been detected in healthy subjects during inspiratory resistive breathing (38, 60), in patients with weak inspiratory muscles constrained to breathe with prolonged inspiratory duty cycles (61) or with exertion-induced inspiratory muscle overload (62), and in patients with chronic obstructive pulmonary disease during exertion (43), and they are associated with changes in respiratory effort sensation (63). The EMG frequency spectrum is often influenced by other factors, such as the signal-to-noise ratio, electrode position, and recruitment of muscles that potentially contribute to cross-talk. Technical factors affecting the EMG power spectrum must always be taken into account before any physiologic interpretations of a spectral shift can be made (1, 1519, 4148).


Peripheral nerve, spinal, or cortical stimulation, either by implanted electrodes (for peripheral nerves) or by externally applied electric or magnetic fields, elicit relatively synchronized activation of motor units at reproducible and predictable levels. The resulting compound action potentials and subsequent muscle contraction allow for measurement of the efficiency of neural and neuromuscular transmission. The muscle responses to stimulation are discussed in Phrenic Nerve Stimulation in Section 2 of this Statement.

Scientific Basis

For practical purposes, human respiratory motor nerves are not accessible over a sufficient length to permit phrenic or other respiratory nerve conduction velocities to be measured. However, motor latency can be measured. Included in the latency are the times required for (1) initiation of action potentials in the axons, (2) rapid saltatory conduction through myelinated axons, (3) slow conduction along the thinner terminal twigs, and (4) chemical transmission across the neuromuscular junctions.

Nerve stimulation.

Stimulation of respiratory nerves can be accomplished with electrical or magnetic stimulators (Table 3)

TABLE 3. Types of respiratory muscle stimulation

Type of Stimulation


ImplantedLess uncomfortable during stimulationDifficult to place
Precise control of stimuli
Requires little preparationDifficult to maintain contact
during voluntary effort
MagneticDoes not require contact with skinExpensive and cumbersome
PainlessRelatively unselective

Difficult to stimulate repetitively

Precise location of stimulus uncertain
   (so latency measurements may be
. The former are less expensive, less cumbersome, more rugged, and more precisely controllable; the latter are easier to apply and less painful for patients. For electrical stimulation, the phrenic nerve is stimulated transcutaneously by surface electrodes at the posterior border of the sternomastoid, or with implanted needle, wire, or hook electrodes. The phrenic nerve(s) can also be stimulated magnetically, via a dorsal cervical approach, which stimulates the lower cervical nerve roots; via an anterior presternal approach, which stimulates both phrenic nerves; or, using one or two figure-of-eight coils, unilaterally or bilaterally over the anterior neck to stimulate one or both phrenic nerves.

Several other respiratory nerves and muscles can also be stimulated, either transcutaneously or by needle or wire electrodes. Pradhan and Taly (28) have demonstrated a technique for stimulating lower intercostal nerves via probe electrodes for latency measurements. The ventral roots of intercostal nerves have been stimulated either by high-voltage stimulation over the spine (6466) or by surgically implanted wire electrodes (67). The rectus abdominus and oblique muscles can also be stimulated with large surface area electrodes (68), and abdominal muscle stimulation has been shown to be effective enough to facilitate cough in tetraplegic patients with impaired expiratory muscle function (66). The abdominal muscles can be activated by magnetic stimulation of nerve roots at the level of T10.

In most cases, it is important to be sure that the delivered stimulus is strong enough to activate maximally all motor units in the muscle of interest. To that end, elicited CMAP amplitude is measured as a function of stimulus intensity, and, for subsequent measurements, stimulus intensity is set to be supramaximal, 20–50% above that required for maximal response. During long experiments, it should be regularly verified that stimulus intensity is supramaximal.

Cortical stimulation.

It is now possible to stimulate cortical and subcortical neural pathways in human subjects, using high-voltage (up to 1,500 V) electrical stimulators (69) and magnetic stimulators (up to 3.0 T) (70). Electrical stimulation of the cortex requires saline-soaked gauze pads (or silver–silver chloride electrodes) on the scalp with the anode positioned over the relevant region of the motor cortex. For the diaphragm and intercostal muscles, the optimal site is close to the midline at or just anterior to the vertex (71). The cathode can either be a single electrode 6–8 cm anterior to the anode, a ring of electrodes placed around the scalp, or an electrode several centimeters lateral to the anode at the vertex. Commercially available magnetic stimulation coils consist of several circular coils in a single housing (usually > 10 cm in diameter) positioned (usually tangentially) with their edges close to the scalp region of interest. Positioning of large circular coils slightly anterior or posterior to the vertex will routinely activate corticofugal output from the medial parts of both hemispheres, including major inspiratory and expiratory muscles in normal subjects, and structures as lateral as the hand area are also activated. More focal stimulation can be achieved with double or “butterfly” coils. There are several nonstandard coils, but their properties must be carefully determined before their clinical utility can be assessed.

To obtain minimal latencies for response to transcranial stimulation, and thereby to avoid coming to the erroneous conclusion that there is a deficit in corticospinal conduction, it is best to record them during voluntary (or reflex) contractions producing at least 10% of maximal force. Responses can be obtained without a background contraction, but the latencies are more variable.

The presence of cardiac or other implanted electrical devices, including pacing wires and pulmonary artery catheters equipped with thermisters, is an absolute contraindication for magnetic stimulation and for high-voltage electrical stimulation of cranial or spinal structures. Epilepsy and known intracortical pathology are relative contraindications for cranial stimulation, as is the presence of intracranial clips.

Although transcranial stimulation has now been used for well over a decade with few reports of side effects, two issues deserve mention. First, there is a risk of induction of a seizure, particularly in those with pre-existing cortical pathology (such as a recent cerebrovascular accident). Although seizures have not been observed with repeated single stimuli, greater precautions are necessary for the newer rapid-rate stimulators (72, 73). Second, the use of some but not all transcranial magnetic stimulators has caused sustained elevations in auditory threshold, presumably due to the brief, but high-intensity “click” produced by the stimulus passing into the coil.

Data Analysis

Nerve stimulation is essential for measurements of nerve conduction velocity or latency. Furthermore, the electromyographic signal elicited by supramaximal nerve stimulation (CMAP) provides a different perspective than spontaneous EMG, with two distinct advantages: assurance of maximal activation, and a generally higher signal-to-noise ratio. CMAPs are detected, usually with surface electrodes applied over the costal margin, and the time between triggering of the stimulus and detection of the elicited CMAP is recorded. In most cases, CMAP amplitude and/or area are also recorded.

Latencies can also be measured after cortical stimulation (Figure 7)

. The “central” conduction time (CCT) is an indirect estimate of the time taken for the descending volley to travel from the motor cortex to the relevant motoneurons. The CCT can be measured by subtraction of the peripheral conduction time (estimated by stimulation over the relevant spinal segment or root) from the total conduction time, from stimulus to onset of the motor-evoked potential (MEP). The measured CCT, however, is not a true measurement of central conduction velocities, because the exact site and timing of the relevant descending corticofugal volleys show some variability and because stimulation over the spinal cord activates the motor axons at a variable distance from the motoneurons. The MEP observed in EMG recordings of most human trunk and neck muscles, including diaphragm, intercostals, scalene, and abdominal muscles (64), has an onset latency consistent with a rapidly conducting oligosynaptic pathway. There is no evidence that this MEP involves a contribution from bulbopontine respiratory neurons.

The MEP can also be assessed in terms of amplitude (and dispersion). Transcranial electrical stimulation via an anode over the appropriate scalp site evokes a direct corticospinal volley (D-wave) followed by a series of indirect trans-synaptically evoked corticospinal volleys (I-waves; interval between volleys, 0.8–1.0 milliseconds). With increased intensity of electrical stimulation, the site of activation along the corticospinal path is located deeper within the brain, reaching the level of the pyramidal decussation in the medulla with strong stimuli (74). The size of the D-wave increases with increased stimulus intensity, as does the size of I-waves. With transcranial magnetic stimulation, the precise corticospinal response depends on the location of the evoked currents. Hence, there are differences in the magnetic activation of the corticospinal output from the hand areas (lateral region of the primary motor cortical strip) compared with the leg areas (within the medial edge of the motor cortex) with different directions of stimulus current. With high-power nonfocal coils (which have the greatest diagnostic utility) transcranial magnetic stimuli evoke not only I-waves but also small D-waves (75). There is probably more trial-to-trial variability in the components of the descending volley to magnetic stimulation than to electrical stimulation.

During voluntary contraction, the minimal latency of responses is reduced, and the MEPs are increased in size compared with relaxation (Figure 8)

, reflecting increased excitability not only at the spinal level (76) but also at the motor cortex (77). For the diaphragm, this effect has been documented during both volitional efforts (35) and CO2-driven hyperpnea (78).

Interpretation of the results of motor cortical stimulation must be made with caution for several reasons. First, stimulation over the motor cortex activates both excitatory and inhibitory structures within the cortex. Second, a single stimulus as brief as 50 milliseconds evokes multiple descending corticospinal volleys. Third, the evoked motor response is elicited preferentially in α motoneurons that are close to firing threshold as a net result of voluntary, involuntary, and reflex inputs. Finally, change in MEP can be said to reflect a change in cortical physiology only if the excitability of the spinal cord has been proven to remain constant.

When cortical stimulation is delivered during voluntary effort, the MEP is followed by a period of near silence in the EMG recorded from the relevant muscle. The latter part of the silent period is due to inhibition of motor cortical output (7981). The duration of the silent period grows with increasing intensity of cortical stimulation but does not vary greatly with the relative strength of voluntary contraction.


Normal phrenic nerve/diaphragm latencies, elicited by electrical stimulation at the neck, have been reported to average 6–8 milliseconds in adults (12, 16, 82, 83), with lower values in children (see Table 4)

TABLE 4. Applications of respiratory muscle stimulation tests

Stimulation Test

Normal Values (in Adults)

Abnormal in:
Phrenic nerve/diaphragm latency6–8 msDemyelinating polyneuropathies, e.g.,
Guillain–Barré syndrome
Elicited CMAPDepends on recording electrodePhrenic nerve palsy, neuromuscular
transmission defect, polyneuropathy
of sepsis, myopathies
Central conduction time, total
conduction time minus
phrenic nerve latency
∼ 4 ms
Multiple sclerosis

Definition of abbreviation: CMAP = compound muscle action potential.

. Because the right phrenic nerve is shorter than the left, latency is slightly shorter on the right side. CMAP amplitudes, recorded from chest wall surface electrodes, average 500–800 mV. Normal values for magnetic stimulation-induced phrenic nerve/diaphragm latency or for intercostal nerve/muscle latency have not been fully established (84, 85).

Phrenic nerve/diaphragm latencies are abnormally slow in demyelinating polyneuropathies, notably Guillain-Barré syndrome. They are usually nearly normal, but are associated with markedly depressed CMAP amplitude, in traumatic neuropathies, such as postcardiac surgery phrenic nerve palsy or the polyneuropathy of critical illness. In myasthenia gravis, a reversible decrement in diaphragm CMAP can be elicited by repetitive phrenic nerve stimulation.

CCTs from cortex to phrenic motoneurons are approximately 4 milliseconds in normal adults (35). Normal values for D- and I-wave amplitudes and the effects of disease are not yet established, but it is apparent that I-wave amplitude is reduced by general anesthetic agents.

Transcranial stimulation to determine CCT has been used in the assessment of a range of upper motor neuron disorders and particularly in the assessment of patients with possible CNS demyelination, including multiple sclerosis. It can be applied specifically to the respiratory muscles, or, more commonly, to muscles in both upper and lower limbs.

Compound muscle action potentials (CMAP) in response to electrical or magnetic nerve or cortical stimulation can also provide useful information. Lack of a CMAP after nerve stimulation is an indication of paralysis, with the lesion located proximal to or at the neuromuscular junction. Lack of a CMAP in response to cortical stimulation when a CMAP is elicited by phrenic nerve stimulation has been used to identify good candidates for phrenic nerve pacing.

Changes in CMAP amplitude, especially as compared with changes in elicited muscle twitch strength (such as phrenic stimulation-induced diaphragm CMAP as compared with transdiaphragmatic twitch pressure [Pdi,tw]) can be used as evidence for or against the development of neural or neuromuscular transmission defects (when both Pdi,tw and CMAP decrease) or contractile defects (when Pdi,tw decreases but CMAP does not) (86).


In a manner analogous to the use of the electrocardiogram to assess cardiac function, electrophysiological tests of respiratory muscles—respiratory muscle motor latencies and electromyography—can be used to assess (1) respiratory drive, respiratory muscle coordination, and the level of activation of individual muscles; (2) the presence of neural and neuromuscular pathology; and (3) the apparent efficacy of the contractile function of the muscles, when used in conjunction with measurements of contractile force. The special challenges presented by data analysis complexity and by a host of potential artifacts lead to the need for great care in the application of EMG techniques to respiratory muscles. Nevertheless, neurophysiological tests are emerging as indispensable components of the respiratory muscle physiologist's arsenal.


This Section of the Statement has described available electrophysiologic tests, the functions of which are to assess the integrity of the respiratory neuromotor apparatus. These electrophysiologic tests are technically complex and require considerable expertise.

There are two main types of test: electromyography (EMG) and stimulation tests.

Type 1: EMG.

For the respiratory muscles the EMG can be used to assess the level and pattern of activation, to detect and diagnose neuromuscular pathology, and when combined with tests of mechanical function to assess the efficacy of contraction.

The EMG can be recorded with surface electrodes (for diaphragm, intercostal, scalene, abdominal, and accessory muscles) or an esophageal electrode (for the crural diaphragm). Surface electrodes are noninvasive and sample a large number of motor units, but contamination (cross-talk) from other muscles is a substantial problem, as is the effect of body size and shape on signal amplitude. Esophageal electrodes provide more specific information, but the technique is invasive and complex.

Surface and esophageal electrodes can record the interference pattern EMG (raw EMG) of the respiratory muscles and are useful to determine the timing and level of respiratory muscle activation during breathing, at rest, and under load. Frequency domain analysis of the EMG is used, as a research tool, to investigate respiratory muscle fatigue (discussed in Section 5 of this Statement).

Intramuscular electrodes can be used to record, relatively selectively, from the diaphragm and intercostal muscles. Motor neuron firing frequency can be measured and neuromuscular disorders diagnosed. However, the techniques are invasive and technically difficult.

Type 2: Stimulation Tests.

Stimulation tests measure the efficiency of neural and neuromuscular transmission.

Nerve stimulation can be achieved with electrical or magnetic stimulators. Electrical stimulation is inexpensive and relatively selective, but is uncomfortable and can be technically difficult. Magnetic stimulation is easier to achieve and less uncomfortable, but can be less selective and is expensive.

Most commonly the phrenic nerves are stimulated and the diaphragm EMG elicited, for the measurement of phrenic nerve/diaphragm latencies and CMAP amplitudes. Latencies are prolonged in some neuromuscular disorders (e.g., demyelination) and CMAP is reduced in amplitude (e.g., traumatic damage to the phrenic nerves).

Cortical stimulation is most commonly performed with a magnetic stimulator, and permits the measurement of central conduction times (CCT) for limb muscles and diaphragm. CCT can be prolonged as, for example, in multiple sclerosis. Cortical stimulation is not selective, and the application of the technique to the respiratory system is a highly specialized skill.

1. Lindström L, Magnusson R. Interpretation of myoelectric power spectra: a model and its applications. Proc IEEE 1977;65:653–662.
2. Lindström L. A model describing the power spectrum of myoelectric signals. I. Single fiber signal. Res Lab Med Electr, Göteborg, Sweden. 1973;5:73.
3. Stålberg, E. Propagation velocity in human muscle fibers in situ. Acta Physiol Scand 1966;70(Suppl 287):2–112.
4. Håkansson CH. Conduction velocity and amplitude of the action potential as related to circumference in the isolated frog muscle. Acta Physiol Scand 1956;37:14–34.
5. Juel C. Muscle action potential propagation velocity changes during activity. Muscle Nerve 1988;11:714–719.
6. Metzger JM, Fitts RH. Fatigue from high- and low-frequency muscle stimulation: role of sarcolemma action potentials. Exp Neurol 1986;93:320–333.
7. Broman H, Lindström L. A model describing the power spectrum of myoelectric signals. II. Motor unit signals. Res Lab Med Electr, Göteborg, Sweden. 1974;8:74.
8. Lindström L, Broman H. A model describing the power spectrum of myoelectric signals. III. Summation of motor unit signals. Res Lab Med Electr, Göteborg, Sweden. 1974;9:74.
9. De Troyer A, Peche R, Yernault JC, Estenne M. Neck muscle activity in patients with severe chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1994;150:41–47.
10. De Troyer A, Estenne M, Ninane V, Van Gansbeke D, Gorini M. Transversus abdominis muscle function in humans. J Appl Physiol 1990;68: 1010–1016.
11. De Troyer A, Farkas GA. Linkage between parasternals and external intercostals during resting breathing. J Appl Physiol 1990;69:509–516.
12. McKenzie DK, Gandevia SC. Phrenic nerve conduction times and twitch pressures of the human diaphragm. J Appl Physiol 1985;58:1496–1504.
13. Gandevia SC, McKenzie DK. Human diaphragmatic EMG: changes with lung volume and posture during supramaximal phrenic nerve stimulation. J Appl Physiol 1986;60:1420–1428.
14. Grassino AE, Whitelaw WA, Milic-Emili J. Influence of lung volume and electrode position on electromyography of the diaphragm. J Appl Physiol 1976;40:971–975.
15. Kim MJ, Druz WS, Danon J, Machnach W, Sharp JT. Effects of lung volume and electrode position on esophageal diaphragmatic EMG. J Appl Physiol 1978;45:392–398.
16. Önal E, Lopata M, Ginzburg AS, O'Connor TD. Diaphragmatic EMG and transdiaphragmatic pressure measurements with a single catheter. Am Rev Respir Dis 1981;124:563–565.
17. Beck J, Sinderby C, Weinberg J, Grassino AE. Effects of muscle-to-electrode distance on the human diaphragm electromyogram. J Appl Physiol 1995;79:975–985.
18. Beck J, Sinderby C, Lindstrôm L, Grassino A. Influence of bipolar electrode positioning on measurements of human crural diaphragm EMG. J Appl Physiol 1996;81:1434–1439.
19. Sinderby CA, Beck JC, Lindström LH, Grassino AE. Enhancement of signal quality in esophageal recordings of diaphragm EMG by the double subtraction technique. J Appl Physiol 1997;82:520–530.
20. Van Lunteran E, Haxhiu MA, Cherniak NS, Goldman M. Differential costal and crural diaphragm compensation for posture changes. J Appl Physiol 1985;58:1895–1900.
21. Sprung J, Deschamps C, Hubmayer RD, Walters B, Rodarte J. In vivo regional diaphragm function in dogs. J Appl Physiol 1989;67:655–662.
22. Sinderby C, Beck J, Weinberg J, Spahija J, Grassino A. Voluntary activation of the human diaphragm in health and disease. J Appl Physiol 1998;85:2146–2158.
23. Beck J, Sinderby C, Lindstrom L, Grassino A. Effects of lung volume on diaphragm EMG signal strength during voluntary contractions. J Appl Physiol 1998;85:1123–1134.
24. Beck J, Gottfried SB, Navalesi P, Skrobik Y, Comtois N, Rossini M, Sinderby C. Electrical activity of the diaphragm during pressure support ventilation in acute respiratory failure. Am J Respir Crit Care Med 2001;164:419–424.
25. Delhez L. Electrical responses of the human diaphragm to the electrical stimulation of the phrenic nerve. Electromyogr Clin Neurophysiol 1975;15:359–372.
26. Estenne M, Ninane V, De Troyer A. Triangularis sterni muscle use during eupnea in humans: effect of posture. Respir Physiol 1988;74:151–162.
27. Estenne M, Zocchi L, Ward M, Macklem PT. Chest wall motion and expiratory muscle use during phonation in normal humans. J Appl Physiol 1990;68:2075–2082.
28. Pradhan S, Taly A. Intercostal nerve conduction study in man. J Neurol Neurosurg Psychiatry 1989;52:763–766.
29. Whitelaw WA, Markham DR. Electrode for selective recording of electromyograms from intercostal muscles. J Appl Physiol 1989;67:2125–2128.
30. Whitelaw WA, Feroah T. Patterns of intercostal muscle activity in humans. J Appl Physiol 1989;67:2087–2094.
31. Saadeh PB, Crisafulli CF, Sosner J, Wolf E. Needle electromyography of the diaphragm: a new technique. Muscle Nerve 1993;16:15–20.
32. Silverman JL, Rodriguez AA. Needle electromyographic evaluation of the diaphragm. Electromyogr Clin Neurophysiol 1994;34:509–511.
33. De Troyer A, Leeper J, McKenzie D, Gandevia SC. Neural drive to the diaphragm in patients with severe COPD. Am J Respir Crit Care Med 1997;155:1335–1340.
34. Bolton CF. Clinical neurophysiology of the respiratory system. Muscle Nerve 1993;16:809–818.
35. Gandevia SC, Rothwell JC. Activation of the human diaphragm from the motor cortex. J Physiol 1987;384:109–118.
36. Lindström L, Petersen I. Power spectrum analysis of EMG signals and its applications. In: Desmedt JE, editor. Progress in clinical neurophysiology. Vol. 10. Computer-aided electromyography. Basel, Switzerland: S. Karger; 1983. p. 1–51.
37. Daube JR. Electrophysiologic studies in the diagnosis and prognosis of motor neuron diseases. Neurol Clin 1985;3:473–493.
38. Gross D, Grassino A, Ross WRD, Macklem PT. Electromyogram pattern of diaphragm fatigue. J Appl Physiol 1979;46:1–7.
39. Sinderby C, Lindström L, Grassino AE. Automatic assessment of electromyogram quality. J Appl Physiol 1995;79:1803–1815.
40. Daubenspeck JA, Leiter JC, McGovern JF, Knuth SL, Kobylarz EJ. Diaphragmatic electromyography using a multiple electrode array. J Appl Physiol 1989;67:1525–1534.
41. Sinderby CA, Comtois AS, Thomson RG, Grassino AE. Influence of the bipolar electrode transfer function on the electromyogram power spectrum. Muscle Nerve 1996;19:290–301.
42. Schweitzer TW, Fitzgerald JW, Bowden JA, Lynn-Davies P. Spectral analysis of human diaphragm electromyogram. J Appl Physiol 1979;46:152–165.
43. Aldrich TK, Adams JM, Arora NS, Rochester DF. Power spectral analysis of the diaphragm electromyogram. J Appl Physiol 1983;54:1579–1584.
44. Arvidsson A, Grassino A, Lindström L. Automatic selection of uncontaminated electromyogram as applied to respiratory muscle fatigue. J Appl Physiol 1984;56:568–575.
45. Levine S, Gillen J, Weiser P, Gillen M, Kwatny E. Description and validation of an ECG removal procedure for EMGdi power spectrum analysis. J Appl Physiol 1986;60:1073–1081.
46. Sinderby C, Friberg S, Comtois N, Grassino A. Chest wall muscle cross-talk in the canine costal diaphragm electromyogram. J Appl Physiol 1996;81:2312–2327.
47. Basmajian JV, DeLuca CJ. Muscles alive: their functions revealed by electromyography. Baltimore, MD: Williams & Wilkins; 1985.
48. Desmedt JE. Methodes d'etude de la fonction neuromusculaire chez l'homme: myogramme isometrique, electromyogramme d'excitation et topographie de l'innervation terminale. Acta Neurol Psychiatr Belg 1958;58:977–1017.
49. Roy SH, DeLuca CJ, Schneider J. Effects of electrode location on myoelectric conduction velocity and median frequency estimates. J Appl Physiol 1986;61:1510–1517.
50. Beck J, Sinderby C, Lindström L, Grassino A. Diaphragm interference pattern EMG and compound muscle action potentials: effects of chest wall configuration. J Appl Physiol 1997;82:520–530.
51. Weinberg J, Sinderby C, Sullivan L, Grassino A, Lindström L. Evaluation of diaphragm electromyogram contamination during progressive inspiratory maneuvers in humans. Electromyogr Clin Neurophysiol 1997;37:143.
52. Sinderby C, Lindström L, Comtois N, Grassino AE. Effects of diaphragm shortening on the mean action potential conduction velocity. J Physiol 1996;490:207–214.
53. Clarke RSJ, Hellon RFR, Lind AR. The duration of sustained contractions of the human forearm at different muscle temperatures. J Physiol 1958;143:454–463.
54. Humphreys PW, Lind AR. The blood flow through active and inactive muscles of the forearm during sustained handgrip contractions. J Physiol 1963;166:120–135.
55. Edwards RHT, Hill DK, Jones DA. Heat production and chemical changes during isometric contractions of the human quadriceps muscle. J Physiol 1975;251:303–315.
56. Fink R, Luttgau HD. An evaluation of membrane constants and potassium conductance in metabolically exhausted fibers. J Physiol 1976;263:215–239.
57. Roberts DV. Simultaneous measurement of propagation velocity of action potential and contraction wave in frog striated muscle. J Physiol 1969;147:62–63.
58. Sinderby C, Navalesi P, Beck J, Skrobik Y, Comtois N, Friberg S, Gott- fried SB, Lindström L. Neural control of mechanical ventilation in respiratory failure. Nat Med 1999;5:1433–1436.
59. Parthasarathy S, Jubran A, Tobin MJ. Assessment of neural inspiratory time in ventilator-supported patients. Am J Respir Crit Care Med 2000;162:546–552.
60. Bellemare F, Grassino AE. Evaluation of human diaphragmatic fatigue. J Appl Physiol 1982;53:1196–1206.
61. Bellemare F, Grassino A. Force reserve of the diaphragm in patients with chronic obstructive pulmonary disease. J Appl Physiol 1983;55:8–15.
62. Sinderby C, Weinberg J, Sullivan L, Lindström L, Grassino A. Electromyographical evidence for exercise-induced fatigue in patients with chronic cervical cord injury or prior polio infection. Spinal Cord 1996;34:594–601.
63. Sinderby C, Spahija J, Beck J. Changes in respiratory effort sensation over time are linked to the frequency content of diaphragm electrical activity. Am J Respir Crit Care Med 2001;163:1637–1641.
64. Gandevia SC, Plassman BL. Responses in human intercostal and truncal muscles to motor cortical and spinal stimulation. Respir Physiol 1988; 73:325–338.
65. Lance JW, Drummond PD, Gandevia SC, Morris JG. Harlequin syndrome: the sudden onset of unilateral flushing and sweating. J Neurol Neurosurg Psychiatr 1988;15:635–642.
66. Linder SH. Functional electrical stimulation to enhance cough in quadriplegia. Chest 1993;103:166–169.
67. DiMarco AF, Altose MD, Cropp A, Durand D. Activation of the inspiratory intercostal muscles by electrical stimulation of the spinal cord. Am Rev Respir Dis 1987;136:1385–1390.
68. Mier A, Brophy C, Estenne M, Moxham J, Green M, DeTroyer A. Action of abdominal muscles on rib cage in humans. J Appl Physiol 1985;58:1438–1443.
69. Merton PA, Morton HB. Stimulation of the cerebral cortex in the intact human subject. Nature 1980;285:227.
70. Barker AT, Jalinous R, Freeston IL. Noninvasive stimulation of the human motor cortex. Lancet 1985;1:1106–1107.
71. Maskill D, Murphy K, Mier A, Owen M, Guz A. Motor cortical representation of the diaphragm in man. J Physiol 1991;443:105–121.
72. Chokroverty S, Hening W, Wright D, Wolczak T, Goldberg J, Burger R, Belsh J, Patel B, Flynn D, Shah S. Magnetic brain stimulation: safety studies. Electroencephalogr Clin Neurophysiol 1995;97:36–42.
73. Pascual-Leone A, Houser CM, Reese K, Shotland LI, Grafman J, Sato S, Valls-Sole J, Brasil-Neto JP, Wassermann EM, Cohen LG. Safety of rapid-rate transcranial magnetic stimulation in normal volunteers. Electroencephalogr Clin Neurophysiol 1993;89:120–130.
74. Burke D, Hicks RG, Stephen JP. Corticospinal volleys evoked by anodal and cathodal stimulation of the human motor cortex. J Physiol 1990; 425:283–299.
75. Burke D, Hicks R, Gandevia SC, Stephen J, Woodforth I, Crawford M. Direct comparison of corticospinal volleys in human subjects to transcranial magnetic and electrical stimulation. J Physiol 1993;470:383–393.
76. Rothwell JC, Thompson PD, Day BL, Boyd S, Marsden CD. Stimulation of the human motor cortex through the scalp. Exp Physiol 1991;76:159–200.
77. Baker SN, Olivier E, Lemon RN. Task related variation in corticospinal output evoked by transcranial magnetic stimulation in the macaque monkey. J Physiol 1995;488:795–801.
78. Murphy K, Mier A, Adams L, Guz A. Putative cerebral cortical involvement in the ventilatory response to inhaled CO2 in conscious man. J Physiol 1990;420:1–18.
79. Inghilleri M, Berardelli A, Cruccu G, Manfredi M. Silent period evoked by transcranial stimulation of the human cortex and cervicomedullary junction. J Physiol 1993;466:521–534.
80. Roick H, von Giesen HJ, Benecke R. On the origin of the postexcitatory inhibition seen after transcranial magnetic brain stimulation in awake human subjects. Exp Brain Res 1993;94:489–498.
81. Taylor JL, Butler JE, Allen GM, Gandevia SC. Changes in motor cortical excitability during human muscle fatigue. J Physiol 1996;490:519–528.
82. Chen R, Collins S, Remtulla H, Parkes A, Bolton CF. Phrenic nerve conduction study in normal subjects. Muscle Nerve 1995;18:330–335.
83. Moorthy SS, Markand ON, Mahomed Y, Brown JW. Electrophysiologic evaluation of phrenic nerves in severe respiratory insufficiency requiring mechanical ventilation. Chest 1985;88:211–214.
84. Luo YM, Polkey MI, Johnson LC, Lyall RA, Harris ML, Green M, Moxham J. Diaphragm EMG measured by cervical magnetic and electrical phrenic nerve stimulation. J Appl Physiol 1998;85:2089–2099.
85. Luo YM, Johnson LC, Polkey MI, Harris ML, Lyall RA, Green M, Moxham J. Diaphragm electromyogram measured with unilateral magnetic stimulation. Eur Respir J 1999;13:385–390.
86. Aldrich TK. Transmission fatigue of the rabbit diaphragm. Respir Physiol 1987;69:307–319.

Muscle endurance is the ability to sustain a specific muscular task over time. It is a highly integrated and complex quality of a muscle or a group of muscles that is related to its resistance to fatigue. To a large extent, any measurement of endurance is task specific because different tasks result in varying recruitment patterns of motor units and synergistic muscle groups, each with varying endurance qualities. The wide variety of techniques that have been developed to measure endurance of the respiratory muscles differ largely on the type of task that is being performed. For each specific task, an endurance curve can be generated by plotting task intensity versus the time it can be sustained. Task failure is an event defined by the inability to continue performing the required task (Figure 1)

. At high levels of intensity, a task can be performed for only a few repetitions. As the intensity is decreased, each task can be endured for a longer time until a level can be sustained for an indefinite period (i.e., hours). The latter is referred to as the maximum sustainable task or load. Another estimate of endurance involves performing incremental increases in task intensity for a given time period until a peak intensity is identified, which is the maximum that can be maintained for a finite period of time (Figure 1). This intensity is not sustainable but may also be used to reflect endurance properties.

Although respiratory muscle strength and endurance appear to be closely linked in many conditions (13), there are numerous examples in which endurance would not be accurately predicted from estimates of maximum pressures or maximum ventilatory capacity. Furthermore, the characteristics of endurance curves for a given muscle may change with training, disuse, drug treatment, and so on. For example, in heart failure patients (4) or in normal subjects (5) following certain respiratory muscle training protocols, larger relative effects are seen on endurance compared with strength. Some patients with asthma show inherent elevations in endurance properties as a fraction of strength (6), as do patients with cystic fibrosis (7), suggesting these patients naturally train for endurance during periods of airway obstruction. In contrast, patients with chronic obstructive pulmonary disease (COPD) (8) or patients receiving acute steroid treatment (9) show marked reductions in endurance properties relative to strength. Therefore, endurance measurements can be useful in some clinical and investigative settings for evaluating patient populations and responses to treatment and rehabilitation.


Many different kinds of tasks have been used to quantify the endurance properties of the respiratory muscles. Most often, endurance has been defined in terms of the ability to sustain a level of minute ventilation (ventilatory endurance) or a level of inspiratory and sometimes expiratory pressure. However, these simple measures often present limitations to evaluating the effect of the load on the respiratory muscles. From a muscle energetics viewpoint, the energy requirements of a working muscle (and therefore a rough estimate of its level of activation) are determined largely by the tension developed over time (i.e., tension–time product) and the rate of mechanical work being performed (W·) (10, 11).

Pressure–Time Product

Refer to Pressure Measurements in Section 2 of this Statement for specific techniques for the measurement of pressure at the airway opening and esophageal, gastric, and transdiaphragmatic pressures. The pressure–time product (PTP) is the integration of respiratory pressure over time (i.e., ∫ Pdt). It is common to express PTP over a 1-minute interval (i.e., units = pressure × time; e.g., cm H2O × minutes). The integration process can be performed by most medical amplifiers or digital computers, much like flow is integrated to obtain minute ventilation. If such techniques are used, assurances must be made that expiratory pressures during the expiratory phase, or inspiratory pressures generated due to chest wall elastic recoil, are excluded from the analysis of inspiratory PTP.

A common expression of the PTP is the mean pressure generated over an entire breath cycle (Pa) in Equation 1, in which

For example, if PTP is measured for a single breath period, then the sampling period would be total breath period (Ttot). A signal averaging circuit (available on most medical amplifiers for determining mean vascular pressure) can often be used to measure Pa directly, online. These are usually composed of “leaky integrators” with time constants of approximately 20 seconds. The analysis can also be done by digital computer or mechanical devices (12).

The Pa value calculated in Equation 1 can be measured at the mouth or airway opening if one wishes to estimate the average pressure generated by all the respiratory muscles working against an external load (i.e., Pamo). Alternatively, it can be measured using: transpulmonary pressure (Pal) for measurements of activity of the chest wall and its muscles against the lung and airways (2); transdiaphragmatic pressure (Padi) for the activity of the diaphragm alone (1); or total respiratory muscle pressure (Pamus) for activity of the synergic respiratory muscles against the lung and rib cage (13).

When Pa is normalized to a fraction of the maximum inspiratory pressure available, it is referred to as the pressure–time index (PTI). For example, for measurements of pressure at the mouth or airway opening, Equation 2 is

where PTImo is the pressure–time index measured at the mouth and Pi,max is the maximum inspiratory pressure that can be generated at the mouth or airway opening (usually obtained for a period exceeding 1 second). For the PTI for the diaphragm (PTIdi), maximal transdiaphragmatic pressure (Pdi,max) is substituted for Pi,max and Padi is substituted for Pamo. Refer to Section 2 of this Statement for techniques of measuring Pi,max and Pdi,max. Traditionally, the term tension–time index (TTI) has been applied to this measurement (1). From a physiologic viewpoint, TTI is the “ideal” variable, which is deterministic for a large number of relevant factors in muscle physiology, including muscle energetics and blood flow. However, for most experimental and clinical measurements for the respiratory system, the transduction of muscle tension into respiratory pressures is not straightforward. Therefore, to avoid misinterpretation of the data, it is recommended that PTI be substituted for TTI when pressure comprises the measured variable (see Section 5 of this Statement).


Under conditions of relatively constant ventilation, respiratory muscle endurance (1), blood flow (14), and changes in oxygen consumption of the respiratory system (V·o2,rs) (15, 16) have been shown to be significantly correlated to changes in PTI (Figure 2)

. Furthermore, PTI is a parameter that describes the pressure-generating activity of the muscles, independent of a specific breathing rhythm, breathing frequency, or type of load within the experimental limits tested (1). Normalizing to maximum pressure can also be useful as a measure of the amount of pressure “reserve” utilized during contraction. For example, most normal subjects can sustain a PTIdi of up to approximately 0.18 (1) and a PTI for the chest wall muscles and the synergic inspiratory muscles of up to approximately 0.3 (2). These “critical” PTI values may be useful in estimating whether the muscles are undergoing contractions that are “likely” to lead to a loss of force, or fatigue (17, 18). However, critical PTI should be used with considerable caution, as it is highly likely that the critical PTI may vary somewhat across various pathological conditions. This has not been studied extensively. In addition, in clinical situations there is often some uncertainty regarding the accuracy of measurements of Pi,max or Pdi,max used to calculate PTI (see Volitional Tests of Respiratory Muscle Strength in Section 2 of this Statement).


When the level of ventilation increases at a constant PTP, the V·o2,rs is increased and endurance is reduced (15, 19). For example, in Figure 3A

, when a subject is inspiring with a constant PTP (individual isopleths), increasing flow rates result in markedly increased oxygen consumption of the respirator system (V·o2,rs). Furthermore, when PTP is kept constant, increasing mechanical work rates of the respiratory system W·rs result in reduced inspiratory muscle endurance (Figure 3B). Therefore, when the tasks involve high levels of ventilation, as may occur during exercise, during ventilatory endurance measurements, or in patients with high or changing ventilatory requirements, the various measures of pressure over time (i.e., PTP, Pa, and PTI) become less predictive as global measures of the activity or endurance of the muscles. Under these conditions, the mechanical work rate (W·rs), discussed below, begins to take on a greater significance (19). As shown in Figure 3C, when ventilation is allowed to vary over a wide range of PTP, W·rs becomes highly predictive of the V·o2,rs and therefore the energy utilization of the respiratory muscles.

Another illustration of these points is that the critical PTI for the respiratory muscles working synergically can vary from 0.12 to 0.4, depending on the particular pattern of ventilation, particularly when inspiratory flows and timing are varied over a wide range (13, 20). Nevertheless, under most testing conditions, when ventilation remains relatively low and constant, and duty cycle is kept within a range that is normally seen during spontaneous ventilation (i.e., 0.3–0.5), measures of PTP (alternatively, PTI or Pa) are still the most predictive global measure of respiratory muscle activity available.

Work Rate of the Respiratory System

Generally, the ventilatory work rate (power output) of the respiratory system (W·rs) refers to mechanical work performed by the respiratory muscles against the lungs and chest wall during ventilation. It is calculated as the integration of the appropriate measures of pressure × volume (see Assessment of the Function of the Active Chest Wall: Campbell Diagram in Section 6 of this Statement). In this discussion, we will use W·rs to also include the work rate performed by the respiratory system against any external loading device. Work rate is expressed in joules per minute (1 J = 1 kPa · 1 L; 1 kPa = 10.2 cm H2O). The complete measurement of work of breathing against the lung and chest wall, for both inspiration and expiration, is complex, largely because components involving movement and distortion of the chest wall are difficult to quantify without relatively sophisticated analyses. However, in many cases, measuring the work performed against an external load (W·ext) may provide sufficient information for purposes of respiratory muscle endurance testing.

If a subject is breathing against an external load and ventilation remains near spontaneous levels during loading, the rate of work performed against the lung and chest wall remains relatively unchanged from normal breathing. Therefore, any “changes” in W·rs can be attributed largely to changes in the work performed against the external load, or W·ext. For example, if a subject were breathing against an inspiratory resistive load, W·ext would be directly proportional to changes in Pamo because (Equation 3)

where V·i = inspiratory minute ventilation. Equation 3 emphasizes one of the reasons why measures of the pressure–time product are so powerful in predicting endurance and changes in energy consumption during external loading. If V·i stays constant, changes in Pamo become the sole determinant of changes in W·ext.

The use of Equation 3 eliminates the necessity of performing complex integrations of individual pressure–volume loops for each breath, which are required for more sophisticated estimates of the total W·rs, discussed below. Therefore, it is possible to measure changes in W·ext, online, with digital or electronic multiplication of Pamo and V·i.

An additional component of W·ext occurs from gas compression (expiration) or decompression (inspiration) when large pressures are generated in the airways (15). During inspiratory loading, this gas decompression can account for as much as 0.4 L of displaced tidal volume in normal subjects, elevating the work of breathing by as much as 50%. If thoracic volume is measured in a volume displacement box, this additional volume is measured directly. However, it can also be calculated. Appropriate equations for adjustment of V·i for gas decompression depend on the nature of the loading device (15, 19, 21). For example, if a threshold loading device is used, in which inspiratory pressure generation is approximately constant, the inspired minute ventilation can be adjusted appropriately by adding the decompression volume calculated in the following way:

where ΔVt,i is the additional inspiratory tidal volume in liters (due to gas decompression) that must be added for each breath in the calculation of inspiratory minute ventilation (V·i) in Equation 3. Vt,i is the inspired tidal volume (before gas decompression); FRC is the functional residual capacity in liters (measured independently); Pbs is body surface pressure (usually atmospheric); PH2O is water vapor pressure at body temperature; and Pmo is the threshold loading pressure at the end of an inspiration. Of course, all pressures in Equation 4 must be of the same units (e.g., mm Hg or kPa). For measures of ventilatory endurance, or when there are changing levels of ventilation, a significant portion of the work being performed by the respiratory muscles is done against the resistive and elastic properties of the lung and chest wall. Therefore, accurate estimates of total W·rs must include these measurements. The work rate against the lung and chest wall is most often obtained by the Campbell method (22), which requires the use of an esophageal balloon for estimating pleural pressure and measurement of a relaxation–pressure–volume curve for the lung and chest wall. The original Campbell method (22) is somewhat tedious to apply practically for routine clinical endurance measurements. Equipment is now available to perform the calculations automatically by computer; but even with computerized techniques, examination of the breath-by-breath pressure–volume loops is required. For relevant discussions of the appropriate use of the Campbell method and the Campbell pressure–volume diagram, refer to reviews (2325).


As discussed previously, when ventilatory flow rate increases, total W·rs becomes an increasingly important determinant of both energy consumption of the muscles and endurance time (Figure 3). For ventilatory endurance testing, measurements of W·rs overcome the problems of variability in lung and chest wall impedance between subjects and in the same subjects over time. Such changes in lung mechanics are inevitable in patients who may have wide diurnal variations and fluctuations over more extended time periods. Therefore, measurement of W·rs may be necessary to draw appropriate conclusions regarding the endurance properties in various patient groups. To a large extent, these studies have yet to be systematically performed.

Whether W·rs or Pa should be chosen as the primary global measure of respiratory muscle activity for endurance testing cannot be stated with certainty at this time. It would be ideal if a comprehensive relationship between W·rs, Pa, V·o2,rs, and endurance for the respiratory muscles could be derived for all loading conditions. From an energetics standpoint, the relationship between them is roughly described for the inspiratory muscles by Equation 5:

where Ers is the efficiency of the inspiratory muscles and Pamus is the mean respiratory muscle pressure per breath (15). Equation 5 suggests that if one knew Ers in a given subject, as well as V·i, the energetics and presumably the endurance of the respiratory muscles could be predicted. Unfortunately, Ers is not particularly constant at different relative velocities of muscle shortening (24) or at differing ventilations, depending on the way breaths are performed (21), making this ideal difficult to obtain.


The largest disadvantage of monitoring W·rs during endurance measurements is the complexity of its accurate measurement and analysis. This is not true, however, for the component of W·rs that comes from W·ext. Furthermore, after decades of studies regarding the work of breathing, there are portions of chest wall movement and distortion that remain elusive and difficult to quantify under loading conditions. As shown in Figure 4

, distortions of the chest wall are commonly seen as an adaptive response to external loading (26). Distortions are also seen during maximum ventilatory maneuvers (27). Finally, the simple measurement of the relaxation pressure–volume curve is not easy to obtain in many patients because of the requisite for complete muscle relaxation (28, 29).

Finally, one component of Pa that may be important in determining endurance characteristics, and that is not directly related to W·rs, involves the influence of developed pressure on blood flow during contraction. For example, as Padi increases, blood flow to the diaphragm is limited, presumably by the relationships between tissue pressure and vascular conductance (14, 30) (Figure 2B). Because sustainable task intensities may in part reflect a balance of energy utilization and supply, it is likely that the influence of Pa on muscle perfusion has an independent effect on endurance that cannot be fully accounted for by its mathematical contribution to W·rs or its energetic contribution to V·o2,rs.


The goal of ventilatory endurance testing is to define the maximum sustainable ventilation (MSV), usually expressed as a fraction of maximal voluntary ventilation (MVV). The time duration needed to define “sustainable” is a topic of some controversy and varies with the specific technique described below. As shown in Table 1

TABLE 1. Predicted values for maximum sustainable ventilation/maximum voluntary ventilation

Author (Ref.)

Subject Age

No. of Subjects

Subject Sex
 (No. F/M)

Keens (7)26 ± 21616/060 ± 8
Keens (7)31 ± 114 0/1462 ± 9
Leith and Bradley (5)31 ± 312 1/1180 ± 6
Bai (39)31 ± 35 0/575 ± 4
Belman and67 ± 42514/963 ± 11
Gaesser (69)(estimate)
Mancini (4)
50 ± 1
55 ± 9

*Results represent means ± SD.

Definition of abbreviations: F = female; M = male; MSV = maximum sustainable ventilation; MVV = maximum voluntary ventilation.

, normal subjects can sustain ventilations ranging from 60 to 80% of MVV. Therefore, with submaximal exercise, it is probably rare that any normal individuals ever exceed their MSV, because maximum exercise ventilations average approximately 61 ± 14% of MVV in the normal population (31). In some athletes, ventilation is maintained near the sustainable level of sedentary subjects. For example, elite cross-country skiers can sustain ventilation averages during exercise in excess of 100 L/minute, or approximately 61% of their predicted MVV for periods of 30 to 85 minutes (32), with little or no evidence of fatigue. However, the baseline MVV in these athletes is frequently elevated above normal, and unlike normal subjects, they can sustain 86–90% of MVV for 4 minutes, presumably because of their extreme conditioning. In a clinical setting, the measurement of ventilatory endurance takes on a much greater importance because patients with chronic lung disease or perhaps heart failure (4) may progress to a condition in which exercise is limited by their ability to sustain ventilation. The ventilatory endurance test is a measure of both inspiratory and expiratory muscle endurance.


Early techniques for measuring MSV required repeated trials of MVV with gradually decreasing levels of ventilation, until an MSV could be determined (33). These have generally been found to be exhaustive and time-consuming, rendering them largely impractical for most clinical investigations. However, more recent methods have been developed that make the procedure more practical to perform, requiring only 10–25 minutes/test (4, 34, 35).

For all measurements of MSV in obstructed patients, it is recommended that the test be preceded by administration of a nebulized bronchodilator. This may be particularly useful if ventilatory endurance is to be repeated at different times, for example, before and after rehabilitation, to reduce inherent variability in airway resistance.

The test begins with the routine measurement of a 12-second MVV, using the same equipment employed for the MSV test. Protocols for technique and reproducibility of MVV, which meet American Thoracic Society (ATS) criteria, are available (36, 37). Accurate MVV measurements are critically important for interpretation of MSV. There are two primary techniques for acquiring MSV, the maximum effort technique and the maximum incremental technique, as discussed below.

The maximum effort technique requires subjects to target a ventilation of approximately 70–90% of their MVV (7, 34), using visual feedback from a spirometer or an oscilloscope (Figure 5)

. Sometimes, one or two short practice trials are used to determine the starting target ventilation. During the first 2–5 minutes, the target ventilation is adjusted up or down to a level slightly lower than the subject's maximum effort. The subject is then continually encouraged to meet the target for the next 8 minutes. There are some studies that have described measuring only a 4-minute MVV as an indicator of endurance (32). Although a potentially useful and practical approach, insufficient data are available to evaluate whether this provides a sufficient estimate of sustainable ventilation. In all studies, it is necessary to control end-tidal carbon dioxide (PetCO2) during the test, usually by adjustment of the carbon dioxide fraction (Fco2) in the rebreathing dead space. The average ventilation achieved over the last minute is considered to be the MSV. It is not routine to measure the W·rs or V·o2,rs, but there are rational advantages in doing so, as discussed below.

There is no standardized equipment available for measuring ventilatory endurance. However, the system used should have the following capabilities: (1) provide for maintaining isocapnia during hyperpneic maneuvers; (2) have a low impedance to air flow, which meets accepted standards for spirometry such as ATS criteria (e.g., < 2.5 cm H2O/L/second to 14 L/second) (36); (3) provide reasonable humidification of the inspired air; and (4) provide real-time visual feedback of ventilation. Mechanical systems that approach these criteria have been described in the literature (5, 7, 34, 38), one of which is illustrated in Figure 5. Care must be taken if pneumotachographs are used for ventilation measurements, to ensure that they are linear over the range of measured flows, that their electronic drift is compensated for, and that they do not contribute significantly to the resistance of the system. If the pneumotachograph is in the patient line, appropriate compensations should be made for changes in gas viscosity due to supplemental oxygen if it is used.

The maximum incremental technique is a newer procedure for obtaining an estimate of MSV. It uses 10% incremental increases in target ventilation, every 3 minutes, beginning at 20% of MVV until the subject cannot sustain the target ventilation for the last 3-minute period (4, 35) (MSV is calculated from the last 10 breaths of the last minute of the highest target ventilation). This technique, which resembles an incremental exercise test, was demonstrated to result in MSV measurements nearly identical to those that could be attained by traditional approaches, and was well tolerated by subjects (4).

The importance of sustaining a maximum ventilation for the three or more minutes at the end of the test should be emphasized. Presumably, during this period, fatigue of the respiratory muscles is progressing rapidly because it is a period of maximum effort following a relatively long period of “near-maximum effort.” Presumably this results in a decay of ventilation to a near sustainable level.

Normal Values

Normal values for MSV, by any method, have not been systematically obtained over a wide population, and results vary considerably between laboratories (Table 1). The large differences in predicted values may be due in part to variations in technique, particularly with respect to impedances of the mechanical measuring devices. The system impedance can have substantial effects on the total W·rs at high ventilations. In addition, there are differences in the populations studied and what was defined as sustainable. It is recommended that when publishing reports of ventilatory endurance, the value for the impedance of the measuring device be stated. Until more complete population standards and uniform equipment and techniques are available, each laboratory is advised to establish its own population standards.

Results for MSV should be reported as a fraction of measured MVV (MSV/MVV%) and either as an absolute value (L/minute) or as a fraction of predicted MVV (MSV/MVV% pred). The latter, which has not been used routinely, provides a normalization of the absolute sustainable value to the patient's age, height, and sex, independent of inherent lung or respiratory muscle function.


There are a number of advantages to measuring MSV as an indicator of respiratory muscle endurance, the most important of which is its close resemblance to the task performed during exercise. It therefore provides clinically relevant data that can be related to function. Second, it is probably a measure of both inspiratory and expiratory muscle endurance because in normal subjects there appear to be decrements in both inspiratory and expiratory function after MSV maneuvers (39). Finally, maximum ventilatory maneuvers result in evidence of diaphragm fatigue (3941). Interestingly, this does not appear to be true for patients with COPD (42).


The disadvantages of using MSV as an indicator of endurance are related to the difficulty in estimating the relative contribution of lung and chest wall mechanics to the measurement. MVV measurements are highly susceptible to relatively small changes in flow resistance, the effects of which are amplified exponentially as ventilation increases (24). Therefore, the load on the respiratory muscles is not uniform across patients or even in the same patients over time.

This is of considerable importance in COPD or other obstructive lung diseases in which day-to-day and diurnal variations in airway mechanics are common. Second, the wide variety of strategies utilized in a given patient to perform MVV-like maneuvers leaves many potential sources of variance between subjects. For example, in patients with COPD, effective use of the expiratory muscles is often limited during elevated ventilations (compared with normal subjects) because of early maximum flow limitation. This is accompanied by hyperinflation and shorter inspiratory muscle lengths with a greater proportional burden on the inspiratory muscles than would be seen in normal subjects. In patients with COPD, measurements of MSV/MVV% as an indicator of respiratory muscle endurance have suggested excellent ventilatory muscle endurance relative to strength, as compared with control subjects (34). However, because their mechanical abnormalities have greater relative influence at higher ventilations, the denominator of the MSV/MVV fraction may be artificially low in these patients, giving the impression that endurance properties are normal. When external resistive loading techniques are utilized, which reduce the contribution of lung and chest wall mechanics as a factor in the measurement, it is found that the endurance capacity of the respiratory system is relatively low in the COPD population compared with normal subjects (8).

The problem of the contribution of the inherent impedance of the respiratory system could be overcome by careful measurement of W·rs during the test. As shown in Figure 6

, redrawn from the experiments of Tenney and Reese (33), a strikingly different view of endurance can be seen when work rate or power output is quantified. Although this subject could sustain approximately 68% of his MVV, he could sustain only approximately 30% of maximum W·rs. The investigators also showed that despite experimental alterations in pulmonary impedance, the W·rs-versus-Tlim relationship did not change appreciably (33).

In summary, ventilatory endurance testing can be useful as a functional measurement, particularly in the setting of rehabilitation or other forms of treatment. Results should be viewed with some understanding of the test's limitations with respect to separation of muscle properties versus intrinsic mechanical properties of the lungs and chest wall. This limitation could potentially be overcome by measuring W·rs, although this may not be practical in many clinical settings. The most promising approach appears to be the incremental method. However, full support of the technique awaits verification in other laboratories.


When an external mechanical load is applied to the airway opening, the respiratory muscles must generate an additional pressure to overcome the impedance of the load. The external load can be one of several types: (1) a flow resistive load, in which the pressure required of the muscles is dependent on the flow rate across the resistance. Flow resistive loads can be linear or nonlinear depending on whether they produce laminar or turbulent flow; (2) elastic loads, in which the pressure required of the muscles is dependent on lung volume. The higher the tidal volume, the higher the pressure required. Such loads are flow independent; (3) threshold loads, in which a finite pressure is required to open a valve that allows flow to occur. Therefore, the pressure required of the muscles at the airway opening is relatively constant, independent of both volume and flow. Threshold loads result in contractions that are similar to isotonic contractions; or (4) isoflow loads, in which the flow rate and therefore, the rate of inflation is held constant and the pressure generated against the flow is a measured output variable. Isoflow loads are similar in concept to “isokinetic” contractions of limb muscles, in which velocity of shortening is held constant.

To conduct a respiratory muscle endurance test with an external load requires setting the task that the subject must perform against the load. For example, the subject may be asked to breathe normally or to breathe with a set breathing pattern or with a specific muscle configuration. Different ways of contracting against the load result in markedly different measures of endurance, reemphasizing the importance of the concept of task specificity.

The advantage of using externally applied loads is that it is much easier to control the relevant variables during the test. It is even possible to design tests that are specific to the diaphragm (1) or the rib cage muscles (2, 43). Generally, these tests require large developed pressures against normal or relatively modest changes in ventilatory requirements. Such conditions are similar to those of weight lifting, with relatively low velocities of shortening. In contrast, measures of ventilatory endurance, described previously, are more like activities of running with large velocities of shortening and participation by a large number of synergic muscle groups. Interestingly, measurements of endurance to high inspiratory resistive loads appear to be more a reflection of rib cage muscle endurance than diaphragm endurance (44). Therefore, the exact extent to which measurements of endurance to high external loads apply to ventilatory endurance or to clinically relevant conditions such as exercise has not been well defined.

A large number of devices and techniques have been developed to measure endurance of external loads. The most common is the use of orifice-type flow resistance applied to the inspiratory circuit (20, 45). Excellent studies can be performed with flow resistances, but because the pressure load seen by the respiratory muscles depends on the developed flow, the technique requires visual feedback of some form of ventilation, preferably the flow rate. Therefore, for practical reasons, flow resistances have largely been replaced in most clinical laboratories by threshold loading devices or other techniques discussed below. The techniques below have generally been used to measure inspiratory muscle endurance.

Maximum Sustainable Threshold Loading

Nickerson and Keens (46) developed a method in which endurance times are measured in response to gradually decreasing threshold pressures, starting near Pi,max. They described one of the first threshold loading devices that was relatively flow independent. The test usually begins with a careful measurement of Pi,max. Sequential Tlim measurements are then made, beginning at approximately 90% of Pi,max and decreasing in increments of 5%. Subjects are allowed to rest between each measurement for approximately 10 times the length of Tlim. No attempt is made to control the breathing pattern. Task failure is determined at each load by the inability to maintain ventilation against the load, resulting in the subject coming off the mouthpiece. Other definitions of task failure define a point at which a subject is unable to generate the threshold pressure or a target flow for three consecutive breaths (47). The first pressure that can be sustained for more than 10 minutes is considered the sustainable inspiratory pressure (SIP). The SIP is determined by averaging the pressures over the last 20 breaths.

The original Nickerson and Keens (46) threshold loading device has never been available commercially, but is made of a simple plunger, with leaded rings added to the inside of the chamber to weight the valve. A more modern version is illustrated in Figure 7

. There is a linear relationship between increases in weight and the pressure required to lift the plunger. The original device used a plunger, seated with a 1-in. O ring onto a 45° surface (46). Larger O rings result in more flow independence but less stability. Even small changes in the size of the contact circumference and the precision of the seating can have large effects on the weight/pressure relationship. Therefore, each homemade valve requires independent additional supports for the plunger, which improve its stability (48), and the use of standardized, commercially available valve mechanisms (nondisposable positive end-expiratory pressure valves), which improve the pressure–flow characteristics (Figure 7) (49). Some commercially available spring-loaded threshold valves do not have the pressure range necessary for testing endurance in most patients.

Normal values.

As with most respiratory muscle endurance testing techniques, normal values have not yet been developed. For example, the influence of stature, age, and sex is not described and the numbers of subjects have been low. Nickerson and Keens (46) tested 15 normal individuals ranging from 5 to 75 years of age. The 12 adults could maintain a mean ± SD SIP of 82 ± 22 cm H2O, or 71 ± 10% of Pi,max. On a second trial, in 12 subjects, both Pi,max and SIP increased by approximately 10%, while the relationship of SIP/Pi,max remained constant. Somewhat different results were found by Martyn and coworkers (50) when using the method of Nickerson and Keens (46). They found that the SIP was 52 ± 6% of Pi,max on the first trial. However, when subjects were asked to repeat the loads that they had previously failed, they were able to increase their SIP to 77 ± 6% of Pi,max (50).


The attraction of the technique of Nickerson and Keens (46) has been that it provides a method for evaluating global respiratory muscle endurance in a one-session test, much like a pulmonary function test. There were no previous studies that defined a technique to establish sustainable pressure in a practical setting. Furthermore, the test is noninvasive and is tolerated relatively well, the equipment required is inexpensive and does not require a great deal of training or coordination for the subject, and the results are relatively independent of the mechanics of breathing because minute ventilation increases minimally.


It is clear that subjects will adjust their breathing pattern as they attempt to breathe against any kind of large mechanical load, and they will learn to do this over time (51). This effect may have been underestimated by Nickerson and Keens (46) as discussed by Martyn and coworkers (50). Relatively small changes in duty cycle (52), inspiratory flow rate (20, 52), and tidal volume (13) can have relatively large effects on endurance measures. Therefore, it would seem appropriate to control the pattern of contraction against the load during the test. However, it is likely that a naive subject will be able to achieve longer Tlim values when allowed to breathe spontaneously. Artificially imposing a breathing pattern may not be appropriate for body size, vital capacity, or CO2 production. Furthermore, chest wall configuration, and therefore respiratory muscle recruitment, are quite different when inspiring against “target” respiratory patterns, when agonists and antagonists are simultaneously recruited (52), as compared with spontaneous or maximum uncontrolled inspirations (13). Nevertheless, the effects of the pattern of contraction and recruitment on Tlim result in an inherent measurement variability between subjects and in the same subject over time (50). It is likely that this problem could be overcome to some extent by measuring Pamo and W·ext, because they are likely to be the most dominant determinants of Tlim, regardless of the pattern of breathing. However, this has not been measured systematically in available clinical studies using maximum sustainable threshold loading.

Having subjects begin with endurance trials at the highest pressure loads can be exhausting, uncomfortable, and time-consuming for the patient. The test generally requires a minimum of 2 hours, as was originally described (46).

Maximum Incremental Threshold Loading

The incremental threshold loading technique was described in the late 1980s (8, 50, 53). In concept, it was designed to resemble a Bruce protocol, which is popular for incremental, whole body exercise testing. Before the study, the patient's Pi,max is measured by standard techniques (see Pressure Measurements in Section 2 of this Statement). The subjects inspire from a threshold valve, as described previously, beginning with initial threshold pressures of approximately 30–40% of Pi,max. The threshold pressure is then increased by a unit of weight (e.g., 100 g) added to the outside of the valve, resulting in a change in pressure of approximately 5–10% of Pi,max, until the load cannot be tolerated for 2 minutes. The maximum inspiratory mouth pressure that can be tolerated for the full 2-minute interval is considered the peak pressure (Ppeak) (8). This technique has generally been applied to the measure of inspiratory muscle endurance.

Normal values.

Normal values obtained by this technique have not been thoroughly described. Small sets are displayed in Table 2

TABLE 2. Predicted values for incremental threshold loading*

Author (Ref.)


Subjects No.



Notes on End Point
Martyn (50)33 ± 214 (9/5) 88 ± 10NRPpeak determined from highest
load over 1 min
McElvaney (53)31 ± 510 (5/5)Trial 1:∼ 0.22 ± 0.07Highest Ppeak tolerated for
 84 ± 17∼ 0.25 ± 0.08full 2 min
 Trial 2:
 87 ± 21
Morrison (56)67 ± 48 (5/3) 80 ± 17∼ 0.32 ± 0.12Highest Ppeak tolerated for
full 2 min
Eastwood (57)30 (28–41)7 (5/2)Trial 1:NRHighest Ppeak tolerated for 30 s
∼ 75 ± 20 0.26 ± 0.11
Trial 3:

∼ 94 ± 21

*Results represent means ± SD.

Pi,max was measured at residual volume.

Mean (range).

Definition of abbreviations: F = female; M = male; NR = not reported; Pi,max = maximum inspiratory pressure; Ppeak = peak threshold pressure achieved during incremental loading under conditions stated in Notes on End Point column; PTIpeak = peak pressure–time index achieved during incremental loading under conditions stated in Notes on End Point column.

. Results are relatively consistent between laboratories, particularly for the peak PTI that can be achieved in the last stage (PTIpeak, 0.25 to 0.32). Older subjects appear to demonstrate less initial strength (Pi,max) but an ability to achieve higher relative PTIpeak values (8), which may be consistent with chest wall adaptations to aging (54).


The incremental threshold test holds strong appeal as a measure of inspiratory muscle function because it is well tolerated and provides a clear outcome variable that is somewhat easier to define than sustainable efforts. Furthermore, it appears to be sensitive to disease states and clinical treatment (3, 8, 9, 55). The test has been described by the original authors as more reproducible than the technique of Nickerson and Keens (46); as being tolerated well by naive subjects, who give results similar to those of trained subjects; and having an overall outcome that is little affected by the breathing pattern (56). However, subsequent testing by Eastwood and coworkers (57), using similar techniques, found that normal subjects demonstrate a considerable learning effect from the first to the third trial.

An interesting observation is that a peak W·ext reaches its highest value and then falls precipitously before attaining Ppeak, while oxygen consumption and pressure development are still rising (50). This means that efficiency is falling during the final stages of the test. It is promising that peak W·ext may be a useful measurement of the capacity of the muscles of the chest wall to perform external work, a value that may be as important clinically as the measure of endurance. A second interesting finding is that some subjects are able to achieve higher Ppeak values during this test than they can attain during Pi,max maneuvers (50, 53, 57), a phenomenon attributed to the fact that some subjects are unable to maximally activate their inspiratory muscles during Pi,max testing (58).


Unfortunately, the extent to which the results from incremental tests represent endurance or strength is not entirely clear. Strictly speaking, it is not a test that has been proven to be a direct measure of endurance, just as an incremental exercise test is not generally considered an endurance test. However, when individuals are asked to attempt to sustain the maximum threshold pressure previously reached with the incremental method, the average Tlim they can maintain is approximately 6 minutes (53). This suggests that although the maximum incremental threshold pressure is not sustainable, it is certainly approaching the asymptote of a typical inspiratory muscle endurance curve. Another suggestion that the threshold test is approximating an endurance measurement is the fact that the maximum PTImo in the last stage is approximately 0.22–0.32 (50, 57), which is similar to sustainable pressures described for the rib cage muscles (2) and the inspiratory muscles working in synergy in a normal range of duty cycles and low flow (13).

Of some concern is the tendency for hypoventilation and desaturation during the test (57). Although modest desaturation is unlikely to affect the measurement appreciably in normal subjects (57, 59), hypercapnia may contribute to a loss of function, unrelated to endurance characteristics (60, 61). Finally, as the intensity of the load increases, subjects consistently decrease end-expiratory lung volume to maximize the length and configuration of their inspiratory muscles (57). This is something of a disadvantage for testing, because the capacity of the muscles to contract against the load is changing during the test. However, it is likely that such changes in configuration are typical of patient responses to many types of high inspiratory mechanical or ventilatory loads and is not a problem unique to incremental loading. Finally, the recruitment patterns of the respiratory muscles may vary during incremental loading and may not totally reflect the endurance characteristics of breathing against constant submaximal loads.

Repeated Maximum Inspiratory Pressures

McKenzie and Gandevia (6, 62, 63) have developed a technique that uses 18 repeated Pi,max maneuvers. The test begins with measurement of Pi,max and practice efforts using visual feedback of airway opening pressure. Three different breathing patterns have been described (6, 62). The most practical appears to be a series of 18 Pi,max contractions lasting 10 seconds each, with 5 seconds of rest between contractions (duty cycle = 0.67) (62). A similar approach has been used to measure expiratory and limb muscle endurance (6). The only equipment required is a manometer for measuring airway opening pressure. This technique has been generally used to measure inspiratory muscle endurance.

Normal values.

In normal young subjects (n = 12), with a duty cycle of 0.67, the average inspiratory mouth pressure attained in the last contraction is 87 ± 3% of Pi,max (mean ± SD) (6). The PTImo at this point is approximately 0.58. Using a similar protocol, but with a slower frequency, pressures dropped to approximately 77% (6). Interestingly, when the duty cycle is reduced to 0.5, no drop in pressure generation is observed across the 18 contractions in normal subjects (6). The PTImo is then 0.25, which may be just below the threshold for fatigue for the rib cage muscles (2, 43).


The technique provides a measurement that is entirely independent of lung and chest wall mechanics, as well as mechanical work of breathing, making it potentially useful for understanding endurance properties of the respiratory muscles without interference from chest wall or lung mechanics. It appears to be sensitive to the influence of lung disease (6), is simple to perform, and lends itself to the potential for a pulmonary function testing environment.


Potential disadvantages include the fact that the endurance characteristics may reflect the anaerobic capacity of the muscles to sustain force, rather than aerobic endurance, because it is likely that blood flow is largely occluded to the muscles during the prolonged contractions. It also does not appear that in 18 contractions a sustainable level of pressure is fully attained (62). Furthermore, patients with severe lung disease may find it difficult to perform such extended maximum inspiratory maneuvers without discomfort or dyspnea. This technique has yet to be independently tested in patients.

Maximum Sustainable Isoflow

The isoflow method allows subjects to inspire with Pi,max against a device that provides a constant inspiratory flow rate to the lungs (14, 64) (Figure 8)

. In this way, it resembles the repeated Pi,max technique but the lungs are inflated and the inspiratory muscles are allowed to shorten at a relatively constant rate. The method was modeled after isokinetic testing devices commonly used in limb muscle evaluation. Visual feedback of inspiratory pressure, over time, is provided from an oscilloscope. Breathing pattern is generally set such that the subjects hyperventilate during the test. The inspiratory airflow is humidified, and PetCO2 is maintained at eucapnia with supplemental CO2. For routine measurements, inspiratory flow is maintained, at approximately 1 L/second, inspiratory time at 1.5 seconds and total breath period at 3.5 seconds (duty cycle = 0.42). Many other breathing patterns have been used with this technique (14, 64, 65); however, for normal subjects this pattern has been shown to be well tolerated. Subjects continue to inspire maximally with each breath for 10 minutes. Airway opening pressures generally decline exponentially during this period until a “sustainable” pressure is obtained (Figure 8). Using curve-fitting techniques, it has been shown that sustainable pressures in normal subjects can be calculated within 5% with only 5 minutes of endurance testing (64).

To roughly calculate the additional inspiratory pressure used to overcome lung and chest wall impedance, the isoflow apparatus can be modified to inflate the subject's lungs during complete relaxation (13). This additional positive pressure can be added to active inspiratory pressures developed during each breath to estimate the total inspiratory muscle pressure (Pmus).

The isoflow apparatus consists of a large and well-regulated pressure source providing inspiratory flow across an extremely high resistance (13, 64). The pressure drop across the resistance is so high (8,000 to 14,000 cm H2O) that any additional inspiratory pressures developed by the subject at the airway opening have negligible effects on flow rate. Flow is initiated by negative mouth pressures of −2 to −3 cm H2O and turned off at +2 to +3 cm H2O by an electrical triggering circuit. Subjects are protected from the high-pressure source by breathing from a nonrebreathing valve, which will ensure that flow bypasses the mouth if there is no active inspiration. End-tidal CO2 is monitored continuously, and additional CO2 is bled into the inspiratory line to maintain PetCO2 at eucapnia. This technique has been used primarily to measure inspiratory muscle endurance.

Normal values.

Normal values have not been well described over a wide range of subjects. However, in 15 normal subjects (8 males and 7 females; age, 26 ± 6 years) with breathing patterns described above, the peak airway pressure dropped to 70 ± 7% of their initial pressures (measured with inspiratory flow of 1 L/second), and 61 ± 12% of Pi,max by the end of 10 minutes of repeated contractions (64). The sustainable PTI with the pattern of contraction described above was 0.18 ± 0.04. There is a small but significant training effect between the first and fourth trials with the procedure (64).


The isoflow technique has the advantage that most of the important parameters influencing respiratory endurance measurements are controlled. For example, PetCO2 (and therefore arterial oxygen saturation), breath timing, inspiratory flow, and tidal volume are fixed. Furthermore, lung and chest wall mechanics can be accounted for at a first approximation (13). An additional strength is the fact that it is possible to measure the inspiratory muscle strength under similar conditions used in the endurance test. This avoids the difficulty of comparing pressure measurements under static contractions (Pi,max) with contractions under dynamic conditions, where changes in length and velocity of contraction affect pressure development (13, 64). Furthermore, because subjects are performing maximal contractions, the fatigue process develops rapidly and the sustainable pressures can be obtained in a few minutes of testing. The decay of inspiratory pressure over time is an additional variable that can be helpful in distinguishing effects on the fatigue process, independent of sustainable pressure development (61). The test is noninvasive and is tolerated well by naive subjects.


As yet, the isoflow technique has not been used to test patient populations and therefore its utility has not been determined in the clinical setting. It has, however, been shown to be useful for studying mechanisms of fatigue (13, 59, 61, 65). One potential problem with applying the technique on patients is the difficulty with imposing the same breathing pattern used on normal subjects. For example, normal subjects have relatively high ventilatory requirements during the test to assure maintenance of ETCO2, whereas patients with lung disease may not be able to physically perform such high levels of ventilation. Furthermore, the method also depends on subject cooperation, and one cannot be certain of the relative contributions of the rib cage or the diaphragm during contractions. Finally, the equipment used for the isoflow technique is not available commercially, although it is not particularly expensive to create from basic equipment in most physiology laboratories.

In summary, at the present time, for clinical applications, the most promising and practical technique for evaluating the endurance qualities of the global inspiratory muscles against external loads appears to be the incremental threshold loading technique, originally described by Martyn and coworkers (50) and refined by later studies (8, 53, 56, 57). However, because of the uncertainties with regard to its specificity for endurance, we recommend that the term “maximum incremental performance” be used to describe its outcome measures until further information is available. The usefulness of the technique will be advanced by careful quantification of Pamo, W·rs, and V·o2,rs during the tests and by consistent recording of the maximum values that can be maintained for the full 2-minute increments. Further work needs to be done to define predicted values in the normal population.

Other methods described here are of considerable value under experimental conditions and should be considered as options, particularly for specific experimental designs where careful control of the variables is critical. Whenever possible, with any of these techniques, the goal should be to define the sustainable level of Pa and W·rs. Further studies in patient populations will be required to determine the comparative usefulness of these techniques in a clinical environment.

All of the external loading techniques are more likely to reflect the endurance qualities of the rib cage muscles as compared with the diaphragm (44, 57, 66). This should be kept in mind with regard to their clinical implications.


To specifically load the diaphragm requires the subject to attempt to generate a target transdiaphragmatic pressure rather than a target mouth pressure. This is because it is possible to generate inspiratory pressures at the airway opening by the rib cage muscles in the absence of significant diaphragm contribution. Furthermore, pressure development by the diaphragm is not only distributed against the rib cage but is used to contract against the abdominal contents as well.


Roussos and coworkers (45) tested for the maximal sustainable, transdiaphragmatic load. Subjects sustained a given Pdi until they could no longer reach the target pressure. There were no requirements on breathing frequency or duty cycle. They found that approximately 40% of Pdi,max could be sustained for 60–90 minutes. A higher Pdi lasted for a shorter time.

A more precise technique for measurement of diaphragm endurance in humans was developed by Bellemare and Grassino (1, 67). Subjects are instrumented with an esophageal and gastric balloon as described in Pressure Measurements in Section 2 of this Statement. Maximum transdiaphragmatic pressure is determined and then subjects proceed to inspire through a variable inspiratory flow resistance with a set breathing pattern by watching an oscilloscope (Figure 9)

. Two target pressures are displayed on the oscilloscope screen: Pdi and gastric pressure (Pga). Tidal volume and duty cycle (Ti/T tot) are monitored as well. The subject generates a target Pdi by actively inspiring against a variable resistance and a target Pga of approximately 50% of Pdi. Runs with Pdi of 0.2 to 0.8 and Ti/T tot of 0.2 to 0.7 were tested. The product of Pdi/Pdi,max × Ti/T tot was found to be the best predictor for endurance. Values of 0.15 to 0.18 or smaller could be sustained for more than 1 hour. Values above 0.18 were sustainable for shorter periods. This work proposed the concept of PTIdi as a consistent index to predict development of fatigue and failure. Accomplishing this requires active contraction of the abdominal muscles during inspiration, resulting in the diaphragm generating approximately equal but opposite pressures against the abdomen and rib cage. The choice of using a Pga of 50% of Pdi was arbitrary and may not have been a critical variable. A procedure for quickly determining a sustainable pressure load has not been developed. Bellemare and Grassino (1, 67) have used this technique primarily to understand the determinants of diaphragm fatigue by performing repeated endurance trials with varying Pdi targets and did not intend its use as a clinical test.

The equipment required for this technique is shown in Figure 9. It is possible that the use of a threshold resistance may be of benefit for controlling esophageal pressure (Pes) during inspirations rather than an orifice-type resistance, thus eliminating the need for controlling both Pdi and abdominal pressure (Pab) with visual feedback.

Normal Values

Measurements of diaphragm endurance have not been routinely performed in a large number of normal subjects. However, Bellemare and Grassino (1) described the normal sustainable PTIdi to be in the range of 0.15–0.18 when tidal volume remained approximately 0.75 L (Figure 2A). Therefore, at a duty cycle of approximately 0.4, normal subjects can sustain approximately 40–50% of Pdi,max.


The technique described by Bellemare and Grassino (1) is the only method that has been identified for specifically loading the diaphragm and ensuring that it fatigues during endurance measurements. By limiting tidal volume and contracting the abdominal muscles (i.e., diaphragm antagonists), changes in diaphragm shortening with contraction are reduced to a minimum. This means that the energetics of diaphragm contraction are uniquely dependent on Pdi and not W·di because the diaphragm is doing little mechanical work against the rib cage or abdomen.


Because of the need for invasive instrumentation and the requirement of the subject to coordinate a rather unnatural pattern of abdominal and thoracic expansion, the technique has not been applied extensively to the clinical environment. PTIdi was measured in patients being weaned from a ventilator. It shows that patients developing a PTIdi higher than 0.18 failed the weaning trial. The same patients were tested at a later date with favorable evolution and had a PTIdi below 0.15, and they could be weaned (68). One potential complication lies in the fact that diaphragm blood flow may be determined in part by the relative negative or positive pressures on its surface. For example, Buchler and coworkers (30) demonstrated that blood flow is obstructed more by high positive abdominal pressures than by similar Pdi values obtained by negative pleural pressures. This would suggest that by contracting the abdominal antagonist muscles simultaneously with inspiration, there may be reduced blood flow to the diaphragm, resulting in a greater probability of fatigue and a lower endurance.

In summary, for a specific measure of diaphragm endurance, the technique of Bellemare and Grassino (1) remains the only method currently available. The methodology may become more accessible in the clinical environment as techniques develop for rapidly attaining a measure of sustainable PTIdi, using incremental or maximum effort approaches similar to those described for global inspiratory muscle function. This technique remains in the domain of clinical research. Few studies are available.


This Section of the Statement has explored the available techniques to assess respiratory muscle endurance. The measurements and techniques include the following:

  1. Pressure–time product (PTP): The integration of inspiratory pressure swing over time. Pressure can be esophageal, Pdi, or mouth pressure (if an external resistance is added to the circuit). PTP of the expiratory muscles can also be measured. If pressure is normalized to a fraction of the maximum pressure, the value obtained is the pressure time index (PTI). A PTI of 0.15–0.18 is the upper limit that can be sustained indefinitely by the diaphragm in humans and as high as 0.3 for the rib cage muscles. The PTI thresholds are lower if inspiratory flow is high.

  2. Work of breathing: Calculated by the integration of pressure on tidal volume, measures work against an external inspiratory or expiratory load and is a useful test for measuring endurance as a function of workload. Values of work of breathing relate well to oxygen consumption over a wide range of ventilations. This measurement is limited to respiratory research and could benefit from computerized equipment to facilitate measurement and analysis in the clinical setting.

  3. Ventilatory endurance tests: Maximal sustainable ventilation (MSV) expressed as a percentage of 12 seconds of maximum voluntary ventilation. Two techniques are available to determine MSV: the maximum effort technique (the subject seeks to sustain ventilation at a target level of 70–90% MVV for 8 minutes) and the maximum incremental technique (starting at 20% MVV, the target ventilation is increased by 10% every 3 minutes). There are limited normal data for MSV and these show considerable variability. Each laboratory should develop its own normal values. MSV can be difficult to interpret (e.g., in patients with COPD). The incremental technique may prove to be of value in the clinical setting. To date, most studies of ventilatory endurance have been undertaken within a research context.

  4. Endurance of external loads applied to the airway: The external load can be resistive (the pressure required depends on flow), elastic (pressure depends on tidal volume), threshold (finite pressure required to open the valve, which is independent of flow and volume), or an isoflow load (flow rate held constant). The most widely used technique is that of threshold loading. Either the maximum sustainable threshold load or the maximum incremental threshold load can be measured. The incremental threshold loading test, which uses the same principles as an incremental exercise test, is the most commonly undertaken, and there are limited normal data available. It is not clear to what extent the test reflects respiratory muscle strength rather than endurance.

  5. Repeated maximum inspiratory pressures: In this test the subjects undertake 18 repeated Pi,max maneuvers, each effort lasting 10 seconds with a 5-second rest between contractions. Pressure drops to 87% of Pi,max in young normal subjects over the run. Equipment is simple, and only a manometer and stopwatch are required. Few data from studies in patients are available.

  6. Maximal sustainable isoflow: In this test the subject breathes against a high impedence (air tank) providing a constant flow (1 L/second). The subject develops maximal pressure at a Ti/Ttot of 0.40. Maximum pressure declines exponentially to a sustainable level of 61%, yielding a PTI of 0.18. This technique has not yet been tested in large populations. It has potential as a method of training the inspiratory muscles as well as documenting their endurance.

  7. Endurance of the diaphragm: This has been studied in normal subjects by measuring Pdi and Ti/T tot, which were kept constant by following a pattern of pressure and timing displayed on an oscilloscope. A PTI of 0.20–0.30 resulted in task failure at an earlier time. The technique was developed as a physiologic study designed to measure the use of TTIdi as a parameter to evaluate the development of diaphragm fatigue.

Of the tests of ventilatory endurance available, the most promising, in a clinical context, appears to be the maximum incremental ventilation test. To specifically assess the endurance of the inspiratory muscles, relatively independently of lung and chest wall mechanisms, the incremental threshold loading test appears to be most useful.

1. Bellemare F, Grassino A. Effect of pressure and timing of contraction on human diaphragm fatigue. J Appl Physiol 1982;53:1190–1195.
2. Zocchi L, Fitting JW, Majani U, Fracchia C, Rampulla C, Grassino A. Effect of pressure and timing of contraction on human rib cage muscle fatigue. Am Rev Respir Dis 1993;147:857–864.
3. Schulz L, Nagaraja HN, Rague N, Drake J, Diaz PT. Respiratory muscle dysfunction associated with human immunodeficiency virus infection. Am J Respir Crit Care Med 1997;155:1080–1084.
4. Mancini DM, Henson D, LaManca J, Levine S. Evidence of reduced respiratory muscle endurance in patients with heart failure. J Am Coll Cardiol 1994;24:972–981.
5. Leith DE, Bradley M. Ventilatory muscle strength and endurance training. J Appl Physiol 1976;41:508–516.
6. McKenzie DK, Gandevia SC. Strength and endurance of inspiratory, expiratory and limb muscles in asthma. Am Rev Respir Dis 1986;134:999–1004.
7. Keens TG, Krastins IRB, Wannamaker EM, Levison H, Crozier DN, Bryan AC. Ventilatory muscle endurance training in normal subjects and patients with cystic fibrosis. Am Rev Respir Dis 1977;116:853–860.
8. Morrison NJ, Richardson DPT, Dunn L, Pardy RL. Respiratory muscle performance in normal elderly subjects and patients with COPD. Chest 1989;95:90–94.
9. Weiner I, Azgad Y, Weiner M. Inspiratory muscle training during treatment with corticosteroids in humans. Chest 1995;107:1041–1044.
10. Homsher E, Kean CJ. Skeletal muscle energetics and metabolism. Annu Rev Physiol 1978;40:93–131.
11. Rall JA. Sense and nonsense about the Fenn effect. Am J Physiol 1982;242:H1–H6.
12. McCool FD, Leith DE. Mean airway opening pressure as an index of inspiratory muscle task intensity. J Appl Physiol 1986;60:304–306.
13. Clanton TL, Ameredes BT, Thomson DB, Julian MW. Sustainable inspiratory pressures over varying flows, volumes, and duty cycles. J Appl Physiol 1990;69:1875–1882.
14. Bellemare F, Wight D, Lavigne CM, Grassino A. Effect of tension and timing of contraction on blood flow of the diaphragm. J Appl Physiol 1986;54:1597–1606.
15. Collett PW, Perry C, Engel LA. Pressure–time product, flow, and oxygen cost of resistive breathing in humans. J Appl Physiol 1985;58:1263–1272.
16. Field S, Sanci S, Grassino A. Respiratory muscle oxygen consumption estimated by the diaphragm pressure–time index. J Appl Physiol 1984; 57:44–51.
17. Begin P, Grassino A. Inspiratory muscle dysfunction and chronic hypercapnia in chronic obstructive disease. Am Rev Respir Dis 1991;143: 905–912.
18. Bellemare F, Grassino A. Force reserve of the diaphragm in patients with chronic obstructive pulmonary disease. J Appl Physiol 1983;55:8–15.
19. Dodd DS, Kelly S, Collett PW, Engel LA. Pressure–time product, work rate, and endurance during resistive breathing in humans. J Appl Physiol 1988;64:1397–1404.
20. McCool FD, McCann DR, Leith DE, Hoppin FG. Pressure–flow effects on endurance of inspiratory muscles. J Appl Physiol 1986;60:299–303.
21. Cala SJ, Edyvean J, Rynn M, Engel LA. O2 cost of breathing: ventilatory vs. pressure loads. J Appl Physiol 1997;73:1720–1727.
22. Campbell EJM. The respiratory muscles and the mechanics of breathing. Chicago, IL: Year Book Publishers; 1958.
23. Banner MJ, Jaeger MJ, Kirby RR. Components of the work of breathing and implications for monitoring ventilator-dependent patients. Crit Care Med 1994;22:515–523.
24. Roussos C, Campbell EJM. Respiratory muscle energetics. In: Fishman AP, Macklem PT, Mead J, editors. Handbook of physiology, Section 3: The respiratory system. Vol. III: Mechanics of breathing, Part 2. Bethesda, MD: American Physiological Society; 1986. p. 481–509.
25. Roussos C, Zakynthinos S. Respiratory muscle energetics. In: Roussos C, editor. The thorax. New York: Marcel Dekker; 1997. p. 681–749.
26. Tobin MJ, Perez W, Guenther SM, Lodato RF, Dantzker DR. Does rib cage–abdominal paradox signify respiratory muscle fatigue? J Appl Physiol 1987;63:851–860.
27. Goldman MD, Grimby G, Mead J. Mechanical work of breathing derived from rib cage and abdominal V–P partitioning. J Appl Physiol 1976;41:752–763.
28. De Troyer A, Bastenier-Geens J. Effects of neuromuscular blockade on respiratory mechanics in conscious man. J Appl Physiol 1979;47:1162–1168.
29. Estenne M, Heilporn A, Delhez L, Yernault JC, De Troyer A. Chest wall stiffness in patients with chronic respiratory muscle weakness. Am Rev Respir Dis 1983;128:1002–1007.
30. Buchler B, Magder S, Katsardis H, Jammes Y, Roussos C. Effects of pleural pressure and abdominal pressure on diaphragmatic blood flow. J Appl Physiol 1985;58:691–697.
31. Blackie SP, Fairbar MS, McElvaney NG, Wilcox PG, Morrison NJ, Pardy RL. Normal values and ranges for ventilation and breathing pattern at maximal exercise. Chest 1991;100:136–142.
32. Anholm JD, Johnson RL, Ramanathan M. Changes in cardiac output during sustained maximal ventilation in humans. J Appl Physiol 1987; 63:181–187.
33. Tenney SM, Reese RE. The ability to sustain great breathing efforts. Respir Physiol 1968;5:187–201.
34. Belman MJ, Mittman C. Ventilatory muscle training improves exercise capacity in chronic obstructive pulmonary disease patients. Am Rev Respir Dis 1980;121:273–280.
35. Mancini DM, Henson D, La Manca J, Donchez L, Levine S. Benefit of selective respiratory muscle training on exercise capacity in patients with chronic congestive heart failure. Circulation 1995;91:320–329.
36. American Thoracic Society. Standardization of spirometry: 1994 update. Am J Respir Crit Care Med 1995;152:1107–1136.
37. Dillard TA, Piantadosi S, Rajagopal DR. Prediction of ventilation at maximal exercise in chronic air-flow obstruction. Am Rev Respir Dis 1985;132:230–235.
38. Levine S, Weiser P, Gillen J. Evaluation of a ventilatory muscle endurance training program in the rehabilitation of patients with chronic obstructive pulmonary disease. Am Rev Respir Dis 1986;133:400–406.
39. Bai TR, Rabinovitch BJ, Pardy RL. Near-maximal voluntary hyperpnea and ventilatory muscle function. J Appl Physiol 1984;57:1742–1748.
40. Hamnegard CH, Wragg SD, Kyroussis D, Mills GH, Polkey MI, Moran J, Road JD, Bake B, Green M, Moxham J. Diaphragm fatigue following maximal ventilation in man. Eur Respir J 1996;9:241–247.
41. Mador JM, Rodis A, Diaz J. Diaphragmatic fatigue following voluntary hyperpnea. Am J Respir Crit Care Med 1996;154:63–67.
42. Polkey MI, Kyroussis D, Hamnegard CH, Mills GH, Hughes PD, Green M, Moxham J. Diaphragm performance during maximal voluntary ventilation in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1997;155:642–648.
43. Fitting JW, Bradley TD, Easton PA, Lincoln MJ, Goldman MD, Grassino A. Dissociation between diaphragmatic and rib cage muscle fatigue. J Appl Physiol 1988;64:959–965.
44. Hershenson MB, Kikuchi Y, Tzelepis GE, McCool D. Preferential fatigue of the rib cage muscles during inspiratory resistive loaded ventilation. J Appl Physiol 1997;66:750–754.
45. Roussos C, Gross D, Macklem PT. Fatigue of inspiratory muscles and their synergic behavior. J Appl Physiol 1979;46:897–904.
46. Nickerson BG, Keens TG. Measuring ventilatory muscle endurance in humans as sustainable inspiratory pressure. J Appl Physiol 1982;52: 768–772.
47. Clanton TL, Dixon GF, Drake J, Gadek JE. Effects of swim training on lung volumes and inspiratory muscle conditioning. J Appl Physiol 1987; 62:39–46.
48. Clanton TL, Dixon G, Drake J, Gadek JE. Inspiratory muscle conditioning using a threshold loading device. Chest 1985;87:62–66.
49. Eastwood PR, Hillman DR. A threshold loading device for testing of inspiratory muscle performance. Eur Respir J 1995;8:463–466.
50. Martyn JB, Moreno RH, Pare PD, Pardy RL. Measurement of inspiratory muscle performance with incremental threshold loading. Am Rev Respir Dis 1987;135:919–923.
51. Belman MJ, Scott GT, Lewis MI. Resistive breathing training in patients with chronic obstructive pulmonary disease. Chest 1986;90:662–669.
52. Clanton TL, Dixon GF, Drake J, Gadek JE. Effects of breathing pattern on inspiratory muscle endurance in humans. J Appl Physiol 1985;59: 1834–1841.
53. McElvaney G, Fairbarn MS, Wilcox PG, Pardy RL. Comparison of two-minute incremental threshold loading and maximal loading as measures of respiratory muscle endurance. Am Rev Respir Dis 1989;96:557–563.
54. Tolep K, Kelsen SG. Effect of aging on respiratory skeletal muscles. Clin Chest Med 1993;14:363–378.
55. Weiner P, Azgad Y, Ganam R. Inspiratory muscle training combined with general exercise reconditioning in patients with COPD. Chest 1992;102:1351–1356.
56. Morrison NJ, Fairbarn MS, Pardy RL. The effect of breathing frequency on inspiratory muscle endurance during incremental threshold loading. Chest 1989;96:85–88.
57. Eastwood PR, Hillman DR, Finucane KE. Ventilatory responses to inspiratory threshold loading and role of muscle fatigue in task failure. J Appl Physiol 1994;76:185–195.
58. McKenzie DK, Plassman BL, Gandevia SC. Maximal activation of the human diaphragm but not inspiratory intercostal muscles during static inspiratory efforts. Neurosci Lett 1988;89:63–68.
59. Ameredes BT, Clanton TL. Hyperoxia and moderate hypoxia fail to affect inspiratory muscle fatigue in humans. J Appl Physiol 1989;66:894–900.
60. Juan G, Claverley P, Talamo C, Schnader J, Roussos C. Effect of carbon dioxide on diaphragmatic function in humans. N Engl J Med 1984;310: 874–879.
61. Ameredes BT, Clanton TL. Accelerated decay of inspiratory pressure during hypercapnic endurance trials in humans. J Appl Physiol 1988; 65:728–735.
62. McKenzie DK, Gandevia SC. Influence of muscle length on human inspiratory and limb muscle endurance. Respir Physiol 1987;67:171–182.
63. Gandevia SC, McKenzie DK, Neering IR. Endurance properties of respiratory and limb muscles. Respir Physiol 1983;53:47–61.
64. Clanton TL, Ameredes BT. Fatigue of the inspiratory muscle pump in humans: an isoflow approach. J Appl Physiol 1988;64:1693–1699.
65. Clanton TL, Hartman E, Julian MW. Preservation of sustainable inspiratory muscle pressure at increased end-expiratory lung volume. Am Rev Respir Dis 1992;147:385–391.
66. Mador MJ, Rodis A, Magalang UJ, Ameen K. Comparison of cervical magnetic and transcutaneous phrenic nerve stimulation before and after threshold loading. Am J Respir Crit Care Med 1996;154:448–453.
67. Bellemare F, Grassino A. Evaluation of human diaphragm fatigue. J Appl Physiol 1982;53:1196–1206.
68. Vassilakopoulos T, Zakynthinos S, Roussos C. The tension–time index and the frequency/tidal volume ratio are the major pathophysiological determinants of weaning failure and success. Am J Respir Crit Care Med 1998;158:378–385.
69. Belman MJ, Gaesser GA. Ventilatory muscle training in the elderly. J Appl Physiol 1988;64:899–905.

The purpose of this Section is to outline a potential diagnostic strategy to assess the development of respiratory muscle fatigue in humans. At the outset, it is important to note that there is evidence suggesting that human respiratory muscle fatigue may develop in pathophysiological states associated with the development of respiratory failure (i.e., respiratory loading induced by lung disease) (1). It is also important to recognize that fatigue is defined as a loss of the capability to generate skeletal muscle force and/or velocity that is accompanied by recovery during rest (2). As a result, a single measurement of force is inadequate to detect fatigue: rather, muscle force generating or shortening capability must be demonstrated to fall during serial measurements over time. It could further be argued that a demonstration that force subsequently rises if muscle contraction is stopped and a rest period is provided would be necessary to fully satisfy the definition of fatigue and to exclude the possibility that a given fall in force did not represent muscle injury (the latter condition, by definition, does not improve with short periods of rest). It follows that muscle “fatigue” can therefore be distinguished from muscle weakness (i.e., a reduction in the level of force generation at a given point in time) and muscle injury (i.e., a slowly reversible or irreversible decrement in muscle contractility).

This Section is divided into a brief discussion of (1) the present theories regarding the genesis of muscle fatigue and the different types of muscle fatigue, and (2) a review of the various tests available with the potential to detect the development of fatigue in normal subjects and patients.


On an operational level, it has proven convenient to classify fatigue into different types, with these different forms of fatigue representing different biophysical mechanisms of fatigue development and with each type having different physiological characteristics (3). Several such classification schemes are possible, but a widely employed convention is to classify fatigue as either (1) central fatigue, (2) peripheral high-frequency fatigue, or (3) peripheral low-frequency fatigue. We review each of these types of fatigue in turn.

Central Fatigue

“Central fatigue” refers to a condition in which muscle force generation during sustained or repetitive contraction becomes limited owing to a decline in motoneuronal output. Central fatigue is judged to be present when a truly maximum voluntary effort produces less force than one generated by direct electrical stimulation.

A number of experiments have suggested that a form of central diaphragmatic “fatigue” may develop during respiratory loading (311). A study by Bellemare and Bigland-Ritchie (5) provided evidence that such a phenomenon can occur during the application of external resistive loads to normal human subjects. This study measured transdiaphragmatic pressure generation over time before, during, and after inspiratory resistive loading, and employed superimposed electrical phrenic stimulation at various times during the experiment to determine whether subjects were capable of fully “activating” the diaphragm. This approach makes use of the observation that it is possible for well-motivated individuals to fully activate rested skeletal muscle during volitional contractions when making a maximal effort (i.e., superimposed electrical stimulation of muscle during such maximal maneuvers does not result in an increase in force generation above that achieved volitionally) (3) (see Twitch Occlusion in Section 2 of this Statement). Although achievement of such “maximal” activation is difficult, and usually cannot be achieved with every attempted contraction even in motivated individuals, one study found that all research subjects could achieve at least one maximal contraction of a limb muscle (6). As a result, evidence that maximal activation of a given muscle during a maximal volitional effort can never be achieved after a period of exercise (i.e., the superimposed electrical stimulation can always evoke an increase in force generation) constitutes evidence of “central” fatigue.

At the start of the study by Bellemare and Bigland-Ritchie, no superimposed force could be detected during the imposition of electrical stimuli on maximum volitional efforts, indicating that these subjects were capable of maximally activating the diaphragm before respiratory loading. During the course of loading, however, the extent to which the diaphragm could be activated decreased progressively. Evidence of the development of central diaphragmatic fatigue during repeated maximal and submaximal diaphragmatic contractions has also been reported in a study by McKenzie and coworkers (7).

Other work has suggested that “central fatigue” may be the result of a decrease in central respiratory motor outflow in response to opioid elaboration in the central nervous system, with the latter generated, in turn, as a consequence of the stress of loaded breathing (4, 8, 9). In support of this concept, Santiago and coworkers (8) have shown that naloxone restores the load compensatory reflex in patients with chronic obstructive pulmonary disease in whom it is initially absent. Subsequently, this group demonstrated that resistive loading in unanesthetized goats resulted in a progressive reduction in tidal volume, which was partially reversed by administration of naloxone (4).

In keeping with the results of these animal studies, an experiment involving patients with asthma found that naloxone pretreatment alters the response to methacholine challenge. In these individuals, in whom methacholine induced severe reductions in FEV1, naloxone pretreatment resulted in an increased breathing frequency, occlusion pressure, and mean inspiratory flow rate when compared with saline pretreatment (10). It has been postulated that similar central limitations of respiratory motor outflow may occur in patients with diseases that chronically load the respiratory system, contributing to the development of chronic hypercapnia.

High-Frequency Peripheral Fatigue

Central fatigue is a failure to generate force as a result of a reduction in motor output from the central nervous system. Peripheral fatigue refers to failure at the neuromuscular junction or distal to this structure and is judged to be present when muscle force output or velocity falls in response to direct electrical stimulation. Peripheral fatigue can result because of alterations in the neuromuscular junction, changes in propagation of the action potentials along the sarcolemmal membrane or into the t-tubules, changes in excitation–contraction coupling, or because of other alterations within the muscle cell (e.g., alterations in metabolism, changes in contractile proteins). Peripheral fatigue can be further classified into high-frequency and low-frequency fatigue on the basis of the shape of the postfatigue muscle force–frequency relationship. If fatigue results in depression of the forces generated by a muscle in response to high-frequency electrical stimulation (e.g., in humans, 50–100 Hz) then high-frequency fatigue is said to be present, whereas a reduction in force generation in response to low-frequency stimuli (i.e., 1–20 Hz) is taken as an indication of low-frequency fatigue. Studies have suggested that loss of force at low frequencies represents an impairment of muscle excitation–contraction coupling (i.e., a reduction in contractile protein activation in response to a given nonimpaired sarcolemmal action potential) (12). Reduction in high-frequency force generation is thought to indicate either an alteration in neuromuscular junction transmission, a reduction in sarcolemmal membrane excitability, or a reduction in action potential propagation into the t-tubular system (13, 14). Low-frequency fatigue can occur in isolation, but high-frequency fatigue is invariably associated with some alterations in muscle force generation at lower frequencies.

High-frequency fatigue has been demonstrated in the diaphragms of normal humans after a trial of high-intensity inspiratory resistive loading (this was demonstrated using high-frequency electrical phrenic stimulation) (15). In this study high-frequency fatigue resolves extremely quickly after cessation of strenuous muscle contractions (i.e., after removal of the inspiratory resistive load; see Figure 1)


Low-Frequency Peripheral Fatigue

In the presence of pure low-frequency fatigue, force generation in response to high-frequency stimulation is unimpaired, indicating that the contractile proteins are capable of generating maximal force provided that sufficient calcium is released by the sarcoplasmic reticulum (SR). As a result, impaired force generation at submaximal frequencies of stimulation may represent either a reduced level of calcium availability due to alterations in SR function or a reduction in the calcium sensitivity of the myofilaments at submaximal calcium concentrations. Both changes have been demonstrated experimentally (16, 17). Reduced myofilament calcium sensitivity can be produced experimentally by increasing hydrogen and phosphate ions (18). The explanation for impaired calcium release by the SR during contractions is less well understood, and a number of theories have been proposed to account for this phenomenon (16, 1921).

Low-frequency fatigue has been demonstrated in the diaphragm and sternocleidomastoid muscles of normal subjects breathing against high resistive loads (17, 22). Low-frequency fatigue has also been shown to develop in the diaphragm of normal subjects asked to sustain maximum voluntary ventilation for 2 minutes (23).

Implications of Different Types of Fatigue for Diagnosis

Although it is convenient to discuss the characteristics of central, peripheral high-frequency, and peripheral low-frequency fatigue separately, it is likely that these various phenomena do not occur in isolation during muscle activation. All three processes may be operating simultaneously when the respiratory muscles confront an excessive workload, with the relative importance of each depending on the duration of respiratory loading and other physiological variables (i.e., arterial pressure, arterial blood gas concentrations, nutritional state). Whereas all three processes may participate in the acute response to loading, both central and high-frequency fatigue resolve rapidly once fatiguing levels of muscle contraction cease, and only low-frequency fatigue is likely to persist over minutes to hours.

Because muscle fatigue is a complex phenomenon, a test that is well suited to detect one form of fatigue may be incapable of detecting another. Moreover, the necessity to make serial measurements of an index of muscle force generation over time to detect fatigue is a particularly difficult endeavor for the respiratory system, because a large number of variables (i.e., lung volume, thoracoabdominal configuration, muscle interaction) can vary over time. All these factors can influence the relationship between muscle force and pressure generation.

As an example, consider the utility of measuring maximum inspiratory pressure (Pi,max) serially to detect respiratory muscle fatigue. This parameter is highly effort dependent, and time-dependent reductions could represent lack of motivation, central fatigue, peripheral high-frequency fatigue, or simply an alteration in lung volume and a resultant mechanical change in the transduction of muscle force into pressure. In addition, failure of the Pi,max to change does not exclude the development of fatigue, because this test would not be suitable to detect low-frequency fatigue. As a result, one must keep in mind the potential limitations of a given test for the detection of muscle fatigue. Most tests are suitable for detecting the presence of only one component of muscle fatigue, and complete characterization of fatigue requires a complex series of assessments.

Breathing Pattern: Tidal Volume and Breathing Frequency

Rapid shallow breathing, characterized by high breathing frequency and low tidal volume, commonly develops in progressive respiratory failure or in unsuccessful attempts to wean from mechanical ventilation. These conditions are associated with an increased ventilatory load and/or a reduced respiratory muscle capacity and may therefore potentially lead to respiratory muscle fatigue (see Prediction of Weaning in Section 10 of this Statement).

Methodology and equipment.

Breathing frequency can be easily counted at the bedside and should be included in standard monitoring. Tidal volume measurements of intubated patients are commonly accomplished with the flow sensors built into modern ventilatory equipment and can be displayed on a breath-by-breath basis by these machines. Volume measurements can also be made with Wright respirometers and other spirometric devices via a mouthpiece in nonintubated patients, albeit mouthpiece placement can artifactually alter tidal volumes and respiratory patterns. To avoid such artifacts, it is possible to noninvasively monitor tidal volume by respiratory inductance plethysmography. Use of this and similar methods is described in Devices Used to Monitor Breathing: Pneumograph, Magnetometer, and Respiratory Inductive Plethysmograph in Section 6 of this Statement.


Monitoring changes in breathing frequency and tidal volume is simple and noninvasive.


The relationship between fatigue and breathing pattern is complex. Moreover, rapid shallow breathing is most likely a reflex response to an increase in the respiratory workload (24) and not the consequence of respiratory muscle fatigue per se (25). Thus, although rapid shallow breathing may accompany respiratory muscle fatigue (2), it cannot be considered a specific marker of fatigue.


Monitoring breathing frequency and tidal volume represents a part of the routine respiratory surveillance of patients, but these parameters should not be used as specific indicators of the development of respiratory muscle fatigue (see Breathing Pattern in Section 10 of this Statement).

Thoracoabdominal Motion

The analysis of breathing movements gives some insight into the level of recruitment and function of the respiratory muscles (in particular of the diaphragm, the rib cage inspiratory muscles, and the abdominal muscles). Two unusual patterns of muscle recruitment may be observed in healthy subjects subjected to fatiguing inspiratory loads (26). The first is an increased variability in compartmental contribution to tidal volume, with breaths characterized by clear rib cage predominance alternating with other breaths in which abdominal motion predominates. This pattern reflects alternatively predominant recruitment of the inspiratory rib cage muscles and of the diaphragm. Because fatigue may develop separately in the diaphragm and in the inspiratory rib cage muscles (27), such alternation may represent a way to postpone respiratory muscle failure. The second pattern is frank paradoxical movement of one compartment, generally the abdomen, that is, an inward movement of the abdominal wall during inspiration. Abdominal paradox indicates weak, absent, or inefficient contraction of the diaphragm. These two patterns may also be observed in patients showing signs of diaphragmatic fatigue during weaning trials from mechanical ventilation (2) (see Figure 5 in Section 6 of this Statement).


Most anomalies of thoracoabdominal motion can be detected by visual inspection by a trained observer. This assessment is facilitated by placing the patient in a recumbent position and conducting a visual inspection for several minutes. Quantitative measurements of rib cage–abdominal motion can also be performed (see Estimation of Ventilation Based on Chest Wall Motion: Konno–Mead Diagram in Section 6 of this Statement).


Visual inspection provides a simple bedside means of detecting alterations in respiratory muscle use.


The abnormal patterns of thoracoabdominal motion described above are not specific for respiratory muscle fatigue. Indeed, respiratory alternans and abdominal paradox can appear immediately after the institution of loaded breathing and these abnormalities do not appear to become accentuated with the development of fatigue. Furthermore, these patterns can also occur, albeit to a lesser degree, during the application of low, nonfatiguing respiratory loads (28). Thus, abnormal thoracoabdominal motion should be viewed as reflecting an increased ventilatory load, which in itself may or may not induce respiratory muscle fatigue.


Analysis of thoracoabdominal motion is most useful to detect either specific forms of respiratory muscle dysfunction (e.g., diaphragmatic paresis) and/or an increase in the ventilatory load. This assessment is, therefore, of some routine clinical use, but lacks specificity for detecting respiratory muscle fatigue (see Breathing Pattern in Section 10 and Devices Used to Monitor Breathing: Pneumograph, Magnetometer, and Respiratory Inductive Plethysmograph in Section 6 of this Statement).

Pressure–Time Index of Inspiratory Muscles

In several investigations, skeletal muscle fatigue was found to occur when a muscle generated more than 15% of its maximal force during sustained contraction. This work has led to the concept that a fatigue threshold exists, with fatigue occurring only when the level of pressure-time generated exceeds this threshold level. Additional work on this concept has shown that the fatigue threshold is higher during intermittent contractions and depends on the relative duration of contraction and relaxation. The same holds true for the inspiratory muscles of subjects submitted to external inspiratory loads. The pressure–time index of the diaphragm is defined as

where Pdi is the mean transdiaphragmatic pressure generated per breath, Pdi,max is maximal transdiaphragmatic pressure, Ti is inspiratory time, and Ttot is total breath time. When breathing is accomplished predominantly with the diaphragm, the critical PTdi is 0.15–0.18. Below this threshold, breathing can be sustained for more than 1 hour without evidence of fatigue. Above this threshold, task failure occurs after a time limit that is inversely related to PTdi (29) (see Figure 2) . In most situations in which inspiratory loads are applied, the spontaneous breathing pattern is characterized by predominant recruitment of inspiratory rib cage muscles other than the diaphragm, resulting in augmented rib cage expansion and abdominal paradox. Under these circumstances, the pressure–time index of the inspiratory rib cage muscles is defined as
where Ppl is mean pleural pressure generated per breath and Ppl,max is maximal pleural pressure (equivalent to MIP). With this breathing pattern, the critical PTrc is 0.30. Above this threshold, task failure occurs after a time limit that is inversely related to PTrc (30).

Methodology and equipment.

The measurement of esophageal pressure is required to compute PTrc, and the measurement of both esophageal and gastric pressures is required to compute PTdi. This is most commonly performed with balloon-catheter systems, as described in Pressure Measurements in Section 2 and in Figures 8 and 9 in Section 4 of this Statement.


In principle, PTdi and PTrc characterize the operational conditions of the inspiratory muscles with respect to their fatigue threshold. These indices may allow the assessment of the risk of fatigue before actual task failure occurs.


The critical values of PTdi and PTrc were established in healthy subjects breathing against external loads. The critical thresholds may be different in clinical circumstances, in which a number of pathological factors (e.g., levels of tissue perfusion, presence of hypoxemia) may influence muscle performance. Second, pressure–time index assessment is dependent on accurate measurement of maximal muscle pressure generating capacity (i.e., the Pdi,max of the diaphragm, Ppl,max for the inspiratory rib cage musculature), which is often difficult in patients. Third, shortening velocity of muscle fibers strongly influences muscle energetics and the metabolic consequences of contraction. As a result, the critical PTdi and PTrc that can be tolerated are also a function of inspiratory flow patterns, with lower values for these parameters at high inspiratory flows. Fourth, some clinical conditions (malnutrition, steroid myopathy) change muscle fiber populations, altering the relationship between muscle strength and fatigability. Conditions that result in a shift to a greater concentration of slow fibers in muscle may well result in a better tolerance of a given absolute level of the pressure–time index and an increase in the critical pressure–time index for the diaphragm and rib cage muscles. The critical pressure–time index in most patients remains to be measured.


The pressure–time index should be considered a conceptual framework within which to gauge the level of muscle function rather than an instrument for the clinical diagnosis of fatigue. Fatigue thresholds have been reported in patients with chronic obstructive pulmonary disease (31, 32); however, it remains largely untested in other pathologies.

Volitional Maximal Pressures

Skeletal muscle fatigue has been defined as a loss of capacity to develop force in response to a load that is reversible by rest (12). In accordance with this definition, respiratory muscle fatigue can, potentially, be documented by measuring a decrease in volitional maximal respiratory pressures, with demonstration of recovery with rest. As a consequence, to detect fatigue of the inspiratory muscles, one could measure either maximal static inspiratory pressure, maximal transdiaphragmatic pressure, or maximal sniff pressure.

Maximal inspiratory pressure.

Methodology and equipment. Pi,max is measured at the mouth as described in Volitional Tests of Respiratory Muscle Strength in Section 2 of this Statement.

Advantages. The measurement of Pi,max is noninvasive. Fatigue of inspiratory muscles as a whole has been documented by a transient fall of Pi,max after breathing against external loads (12), maximal voluntary hyperpnea (33), marathon running (34, 35), or labor (36).

Disadvantages. The major limitation of Pi,max as a test of fatigue is a lack of specificity, which is due to its total dependence on the subject's maximal volitional effort. Although Pi,max is reliable in highly motivated subjects, a maximal volitional effort cannot be obtained with certainty in patients. Another drawback of Pi,max is a potential lack of sensitivity for fatigue. Because a maximal static effort is associated with a high neuronal firing frequency, it reflects mainly high-frequency fatigue and may be a poor indicator of long-lasting low-frequency fatigue (see Maximal Static Inspiratory and Expiratory Pressure in Section 2 of this Statement).

Applications. Measurement of Pi,max can be used to detect inspiratory muscle fatigue in motivated volunteers, but has limited use in patients for this purpose because of difficulties in ensuring maximality of effort. Twitch interpolation techniques provide a potential means of solving this latter problem (see Twitch Occlusion in Section 2 and see Section 1 of this Statement).

Maximal transdiaphragmatic pressure.

Methodology. Pdi,max, measured with balloon-catheter systems, is described in Techniques for Pressure Measurements in Section 2 of this Statement).

Advantages. This test measures specifically the strength of the diaphragm. Diaphragmatic fatigue has been documented by a transient fall in Pdi,max after breathing with diaphragm emphasis against external loads (6), voluntary hyperpnea (33), or high-intensity exercise (37).

Disadvantages. The measurement of Pdi,max is invasive and, moreover, shares the same type of limitations as Pi,max for detecting fatigue. Because the physical maneuver that patients must carry out for reliable measurement of Pdi,max is even more complex than that required to measure the Pi,max, it cannot be recommended for assessing diaphragmatic fatigue in clinical settings.

Applications. Pdi,max can be used to detect inspiratory muscle fatigue in motivated volunteers, but probably should not be used for this purpose in patients.

Maximal sniff pressures.

Methodology. Diaphragmatic strength can be assessed by maximal sniff Pdi (38), and global inspiratory muscle strength by maximal sniff esophageal pressure (Pes) (39) or maximal sniff nasal inspiratory pressure (SNIP) (40, 41), as described in Sniff Tests in Section 2 of this Statement.

Advantages. The sniff is a volitional maneuver that is easily performed by almost all subjects and patients. The SNIP is noninvasive and often yields higher pressures than the Pi,max (39).

Disadvantages. The potential usefulness of measuring a fall of maximal sniff pressures to detect inspiratory muscle fatigue in patients remains to be established.

Applications. Although maximal sniff pressures are of established value for measuring inspiratory muscle strength, it is a test still under development for documenting fatigue.

Relaxation Rate

On cessation of contraction, skeletal muscles relax at a rate determined by their relative proportion of fast and slow fibers. When muscles fatigue, their relaxation rate declines as a result of a slower uptake of calcium previously released from the sarcoplasmic reticulum. During various types of intermittent contractions, the rate of decay of Pes and of Pdi reflects the relaxation rate of inspiratory muscles and of the diaphragm, respectively. When fatigue is induced by breathing against external loads, the inspiratory muscle relaxation rate falls early and then stabilizes, following a time course similar to that of the change in electromyogram (EMG) power spectrum. Thus, relaxation rates typically decline before the occurrence of muscle failure at about the same rate as the center frequency of EMG does. On cessation of loading, the relaxation rate recovers quickly and reaches baseline values within 5 to 10 minutes (42, 43).


The relaxation rate of Pes or Pdi can be measured during intermittent contractions against loads (42), during sniffs with airway occlusion (43) or without airway occlusion (44), and during phrenic nerve stimulation (43). The most useful and simple maneuver is the unoccluded sniff, which is easy to perform for most subjects and provides large and consistent changes in relaxation rate after fatigue (45).

Standard balloon-catheter systems are used to measure Pes and gastric pressure (Pga), from which Pdi is obtained. The maximal relaxation rate (MRR) of Pes or Pdi is calculated as the first derivative of pressure with respect to time (dP/dt) over the first half of the relaxation curve. This is obtained by drawing a tangent to the steepest portion of the pressure curve. Because the MRR increases with the amplitude of the pressure swing, it is usual to normalize the MRR and to express it as a percentage of the pressure fall in 10 milliseconds (42, 45). When the natural logarithm of pressure is plotted as a function of time, a straight line appears over the lower 60–70% portion, indicating a monoexponential decay. The reciprocal of the slope of this line represents the time constant (τ) of this exponential decay, which may be used as another measure of muscle relaxation, usually expressed in milliseconds (42, 45) (see Figure 3)

. Thus, a slower muscle relaxation is documented by a decline in the MRR and by an increase in τ. Inspiratory muscle relaxation rate can be assessed in a less invasive manner by measuring the MRR of nasopharyngeal or mouth pressure during sniffs with balloons positioned in these locations (44). An entirely noninvasive measure of inspiratory muscle MRR can be obtained by using SNIP (46). These less invasive techniques have been validated only in normal subjects. Transmission of brief pressure swings from the alveoli to the upper airways is likely to be dampened in patients with abnormal lung mechanics.


The pressure measurement system required for sniffs is described in Section 2 of this Statement. Analysis of the MRR is best done with a computer program (46).


The measurement of inspiratory muscle relaxation rate is relatively simple and requires minimal cooperation from subjects. The sniff maneuver is easily performed by most subjects and patients and does not need to be perfectly “maximal,” provided that the MRR is expressed as percentage pressure fall per 10 milliseconds. Sniffs should, however, be performed as near to the maximal as possible, because the MRR is effort dependent below 60% of maximal pressure (47). The muscle relaxation rate slows at an early stage during fatiguing contractions and may therefore indicate that inspiratory muscle fatigue is incipient.


The relationship between changes in relaxation rate and force loss during fatigue is not understood. For instance, the degree of force loss does not correlate with the changes in relaxation rate and therefore cannot be inferred from this parameter (45, 47). The rapid recovery of this index with rest also poses practical problems of measurement in clinical settings. The range of normal values for the MRR and τ is wide, with overlap between fresh and fatigued states. Serial measurements are thus required to detect the onset of inspiratory muscle fatigue in an individual (45). Finally, some clinical conditions (e.g., asthma) have been reported to elicit activation of inspiratory muscles during expiration (postinspiratory inspiratory muscle activity). Under such circumstances, persistent activation of some muscles or portions thereof would be expected to alter measured relaxation rates, distorting the relationship of alterations in the MRR to cellular events and to the development of muscle fatigue.


Measurement of the relaxation rate can be used with confidence as an early sign of fatigue in subjects subjected to high external inspiratory loads (4245) or to high-level hyperpnea (46). The inspiratory muscle relaxation rate can also be used to detect fatiguing contractions during exercise in patients with chronic obstructive pulmonary disease (48) or during weaning trials from mechanical ventilation (49). Because interpretation of changes in relaxation rate is not straightforward, this test can be considered useful only for clinical research.

Time domain analysis.

Rationale. For the respiratory muscles, as for any other skeletal muscles, a nearly linear relationship may be found between the pressure and the electrical activity they generate. The slope is related to force and the length of the diaphragm. As a result, for a given muscle length, a decrease in the ratio of respiratory muscle pressure to the integrated electromyographic activity of the muscle generating that pressure should, in theory, indicate a decrease in muscle contractility and the development of fatigue. Furthermore, it has been suggested that a decrease in this ratio indicates an alteration in excitation–contraction coupling (50). This point is also discussed in Inferring Diaphragm Activation and Electromechanical Effectiveness from EMG in Section 6 of this Statement.

Methodology. The methodology required for electromyographic assessment is reviewed in detail in EMG Equipment and Data Analysis in Section 3 of this Statement.

Advantages. In theory, this is a useful approach for separating changes in pressure-generating capacity caused by neural or neuromuscular transmission factors from changes caused by peripheral muscular factors (51). One potential major advantage of this test is the possibility to detect fatigue during spontaneous breathing (52), because no special efforts are required by patients.

Disadvantages. For this index to be valid, other factors affecting respiratory muscle contractility, such as muscle length, chest wall configuration, or lung volume, should be controlled or kept constant (53). The applicability of this particular method to the respiratory system is limited by the difficulty of recording the activity of all the muscles involved in normal or augmented breathing that contribute to the measured pressure. Their relative contribution to the generated pressure is known to change during fatigue development (6, 28), and a reliable recording of the activity of a selective respiratory muscle group is regarded as difficult by some (54). Section 3 of this Statement offers a more optimistic view. In practice, the diaphragm, the neck accessory muscles, and the abdominal muscles are most amenable to this form of testing because their electrical activity can be more easily recorded without interference from other muscles and their force production (sternomastoid) or pressure output (diaphragm and abdominals) can also be recorded in relative isolation.

When interpreting results, one must also recognize that the relationships between integrated EMG activity of the respiratory muscles and the pressure they generate may not be perfectly linear (49).

If special precautions are not taken, EMG signals (particularly those recorded from the diaphragm with an esophageal electrode) can be subject to artifactual changes caused by variations in lung volume or chest wall configuration (55). Luckily, reports provide techniques that exclude many of the artifacts associated with esophageal diaphragmatic recording. Specifically, work by Sinderby, Grassino, and others provides a means of recording and analyzing electromyographic signals so as to exclude electrocardiogram (ECG), electrode motion, noise, and esophageal peristalsis artifacts (5659). This latter work has also shown that it is possible, by using a multielectrode array, to reliably measure the diaphragm EMG amplitude and power spectrum in such a way that these variables are not affected by chest wall configuration and/or diaphragm length (56). Note, however, that even if EMG activity can be accurately recorded respiratory muscle pressures must also be reliably assessed for the pressure-to-integrated EMG ratio to be meaningful.

Applications. The theoretical value of time domain electromyographic assessment in patients is that it can provide a means of determining whether observed reductions in respiratory muscle pressure-generating capacity are due to alterations in action potential transmission or to intrinsic alterations in peripheral muscle function (i.e., alterations in excitation–contraction coupling or contractile protein myofilament function). Although this technique is principally of value for experimental applications at the present time, ongoing efforts are being made to improve the reliability of this form of testing, and this type of measurement may assume a broader clinical role in the future.

Frequency domain analysis.

Rationale. Frequency domain analysis of EMG signals from the respiratory muscles has been proposed as a test to detect the occurrence of respiratory muscle fatigue in humans (60), because the power spectrum of skin surface-recorded EMG signals typically shifts to lower frequencies during fatiguing contractions (see Figure 4)


Several indices of the power spectrum have been used for this purpose, including an assessment of the “center” or “centroid” frequency of the power spectrum and a “power ratio” of a high-frequency band over a low-frequency one. Both of these indices appear to decrease with fatigue and increase with recovery. With appropriate instrumentation, these analyses can be obtained “on line” in spontaneously breathing subjects or patients. Shifts in the EMG power spectrum indicative of diaphragmatic fatigue have been documented during severe whole body exercise (61) and during loaded breathing in normal subjects, in female patients during delivery (39), as well as in ventilator-dependent patients having weaning problems (2).

Methodology. The methodology required for electromyographic assessment is reviewed in detail in EMG Equipment and Data Analysis in Section 3 of this Statement.

Advantages. Studies of normal subjects have shown a good correlation between EMG power spectrum shifts and force or pressure losses at high stimulation frequencies (high-frequency fatigue) but not with force loss at low stimulation frequencies (low-frequency fatigue) (62). High-frequency fatigue is typically associated with failure at the neuromuscular junction or at the sarcolemma. In line with these predictions, a good correspondence has been found for the human diaphragm between the rate of power spectrum shift, the pressure–time product (63), and the changes of the shape of the action potential wave form measured during phrenic nerve stimulation.

Disadvantages. The etiology of power spectral shifts with fatigue is still controversial. Possible mechanisms include a slowing of muscle fiber conduction velocity, a widening of the action potential waveform, a decrease in motor unit discharge rate, or synchronization of motor units firing (64). None of these can be directly linked to a fatiguing process at the sarcomere level. Power spectrum shifts are therefore related to central motor control, or reflex pathways, or changes in electrolyte or metabolite concentrations within the muscles.

In addition, power spectrum shifts are rapidly reversed on rest or with reduced activity even though the muscle may remain in a fatigued state. Power spectrum analysis of an EMG, therefore, cannot provide an indication as to the state of the contractile system or the excitation–contraction coupling process, or how these may change with fatigue. As mentioned previously, because of the close association of these indices with neural or sarcolemmal events, power spectral shifts recover quickly with rest (within 5 minutes). As a result, these indices can be markedly affected by the breathing pattern (63) and the breathing strategy employed (29), which in turn may cause a high breath-to-breath variability.

Applications. Because of the problems listed above, this test cannot be taken as a reliable global index of the development of muscle fatigue. The principal utility of this test is that demonstration of an EMG spectral power shift in a working muscle may provide some clue to the development of an alteration in neuromuscular transmission.

Muscle Responses to External Stimulation
Pressure–frequency relationships.

Rationale. Fatigue, defined as a decrease in the pressure- or force-generating capacity of a muscle under loaded conditions, can be most specifically detected by recording the pressure– or force–frequency curve of that muscle in response to artificial motor nerve stimulation. Of the respiratory muscles, the diaphragm (16, 65) and the sternomastoid (66) muscles are most amenable to this form of testing. Low-frequency fatigue has been documented for these muscles in normal subjects during loaded breathing (16) as well as during intense exercise (67).

Methodology. The methodology required for phrenic nerve stimulation and sternomastoid stimulation is reviewed in detail in Section 2 of this Statement.

Advantages. This technique overcomes many of the difficulties associated with volitional or spontaneous breathing efforts. Indeed, the responses are not complicated by possible variations in the level of effort expended. The response of a particular muscle can also be studied in isolation, free from the activity of other muscles. Changes in the shape of the pressure– or force–frequency curves also give indications as to the underlying mechanism of fatigue. For example, a decreased pressure or force at high stimulation frequencies may be indicative of impairment at the neuromuscular junction or at the sarcolemma, whereas a decreased force or pressure at low stimulation frequencies may suggest a possible impairment of excitation–contraction coupling.

Disadvantages. This is a difficult test to perform. Tetanic stimulation can also be painful and it may be necessary to anesthetize the skin near the electrodes. To overcome this problem, partial pressure–frequency curves may be constructed by using twin pulses and by varying the intervals between the pulses (68, 69). These are better tolerated than tetanic stimulation and can provide comparable information regarding the presence of high- and low-frequency fatigue. Because of a large intersubject variability in the responses to artificial stimulation, fatigue can be reliably detected by these techniques only when a subject serves as his/her own control. A possible exception concerns the ratio of the force or pressure developed at a low stimulation frequency (i.e., 20 Hz) over that at a high stimulation frequency (i.e., 100 Hz), for which some critical value may be recognized below which low-frequency fatigue may be said to be present (16, 65, 66, 68).

Application. Although this test provides a means of directly detecting the development of muscle fatigue, applicability of this approach is limited by (1) patient discomfort associated with high-frequency stimulation, (2) equipment expense and complexity, and (3) the need to carefully control for variation in body position, lung volume, and the electrode–nerve interface. Advances in magnetic stimulation techniques may allow a variation of this form of testing to reach more widespread clinical application in the future, but this test is currently limited to research applications.

Single twitch stimulation.

Rationale. As an alternative to tetanic or twin pulse stimulation, recording of muscle twitches in response to single nerve shocks can be employed to detect the presence of low-frequency fatigue (68). Twitch responses are much easier to obtain but are more variable than tetanic responses and are subject to additional variations caused by phenomena such as twitch potentiation (70).

Methods and equipment. The methodology required for phrenic nerve stimulation is reviewed in Section 2 of this Statement.

Advantages. This technique is nonvolitional, eliminating concerns about patient effort in the interpretation of obtained results. In addition, because only single twitches are evoked when employing this technique, much less patient discomfort is involved when compared with that produced by construction of the force–frequency relationship. Because a single shock is, by definition, a “low-frequency” stimulus, this approach also provides a means of detecting the development of low-frequency fatigue, whereas measurement of maximal volitionally produced pressure does not. In addition, the magnitude of the compound action potential evoked during single twitches can be measured and monitored over time; correlation of this assessment with force generation over time may provide a means of detecting alterations in neuromuscular transmission.

Disadvantages. The application of these tests has been largely limited to scientific investigations in normal subjects and patients with lung (71) or neuromuscular diseases (72). Their application in clinical settings is more difficult, partly because of the factors already mentioned but also because of the many pieces of equipment that are required to perform these tests. In the case of the diaphragm, these difficulties are compounded by the necessity of stimulating the phrenic nerves bilaterally. In addition, it is critical that supramaximality be attained during electrical stimulation for this test to be useful. Unless electrical current is sufficiently high (i.e., increased to a level 150% of that required to initially attain a maximal signal), alterations in the compound action potential over time may simply reflect local, axonal changes. The introduction of magnetic stimulation (73) and of phonomyography (74) may help overcome some of these difficulties.

Applications. Although this technique holds promise, more work is needed before this test can be used in the clinical arena. Because of the technical problems detailed above, it is difficult to use present stimulation techniques to accurately evaluate the time course of fatigue, and this approach is currently better suited for evaluation of the effect of an intervention or treatment on muscle function. Once standardization of this technique is achieved, it probably offers the greatest promise to provide an objective index of the development of muscle fatigue.


This Section of the Statement has reviewed the complex process of muscle fatigue and has discussed the available direct and indirect measurements relevant to the assessment of fatigue of the respiratory muscles.

Although a variety of measures and indices have been employed to assess the development of respiratory muscle fatigue in research, there is no well described technique that has been successfully developed and tested to permit precise identification of respiratory muscle fatigue in the clinical setting. Of the tests reviewed, breathing pattern analysis and measurement of thoracoabdominal motion are nonspecific indices that do not directly measure fatigue. Analysis of the pressure–time index provides a useful conceptual framework, but not a specific test of fatigue.

In the research environment, serial measurement of maximal voluntary respiratory pressures, assessment of maximum relaxation rates, frequency domain EMG analysis, and measurement of respiratory muscle pressures in response to electrical or magnetic nerve stimulation are all techniques that can be used to assess the evolution of respiratory muscle fatigue. Of these techniques, serial measurement of respiratory muscle pressure generation in response to electrical or magnetic stimulation is arguably the best technique to directly assess the development of respiratory muscle fatigue at the present time, and it offers the greatest promise for future development into an objective test of respiratory muscle fatigue in the clinical arena.

1. Cohen C, Zagelbaum G, Gross D, Roussos C, Macklem PT. Clinical manifestations of inspiratory muscle fatigue. Am J Med 1982;73: 308–316.
2. Respiratory Muscle Fatigue Workshop Group. Respiratory muscle fatigue: NHLBI Workshop Summary. Am Rev Respir Dis 1990;142:474–480.
3. Bellemare F, Bigland-Ritchie B. Assessment of human diaphragm strength and activation using phrenic nerve stimulation. Respir Physiol 1984;58:263–277.
4. Scardella A, Santiago T, Edelman N. Naloxone alters the early response to an inspiratory flow-resistive load. J Appl Physiol 1989;67: 1747–1753.
5. Bellemare F, Bigland-Ritchie B. Central components of fatigue assessed by phrenic nerve stimulation. J Appl Physiol 1987;62:1307–1316.
6. Allen GM, Gandevia SC, McKenzie DK. Reliability of measurements of muscle strength and voluntary activation using twitch interpolation. Muscle Nerve 1995;18:593–600.
7. McKenzie DK, Bigland-Ritchie B, Gorman RB, Gandevia SC. Central and peripheral fatigue of human diaphragm and limb muscles assessed by twitch interpolation. J Physiol (Lond) 1992;454:643–656.
8. Santiago T, Remolina C, Scoles V, Edelman N. Endorphins and control of breathing: ability of naloxone to restore the impaired flow-resistive load compensation in chronic obstructive pulmonary disease. N Engl J Med 1981;304:1190–1195.
9. Petrozzino JJ, Scardella A, Santiago T, Edelman N. Dichloracetate blocks endogenous opioid effects during inspiratory flow-resistive loading. J Appl Physiol 1992;72:590–596.
10. Bellofiore S, DiMaria G, Privitera S, Sapienza S, Milic-Emili J, Misgtretta A. Endogenous opioids modulate the increase in ventilatory output and dyspnea during severe acute bronchoconstriction. Am Rev Respir Dis 1990;142:812–816.
11. Adams JM, Farkas GA, Rochester DF. Vagal afferents, diaphragm fatigue, and inspiratory resistance in anesthetized dogs. J Appl Physiol 1988;64:2279–2286.
12. Aldrich TK. Respiratory muscle fatigue. Clin Chest Med 1988;9:225–236.
13. Aldrich TK. Transmission failure of the rabbit diaphragm. Respir Physiol 1987;69:307–319.
14. Bazzy AR, Donelly DF. Diaphragmatic fatigue during loaded breathing: role of neuromuscular transmission. J Appl Physiol 1993;74: 1679–1683.
15. Aubier M, Farkas G, De Troyer A, Mozes R, Roussos C. Detection of diaphragmatic fatigue in man by phrenic stimulation. J Appl Physiol 1981;50:538–544.
16. Westerblad H, Allen DG. Changes of myoplasmic calcium concentration during fatigue in single mouse muscle fibers. J Gen Physiol 1991; 98:615–635.
17. Westerblad H, Lannergren J, Allen DG. Fatigue of striated muscles: metabolic aspects. In: Roussos C, editor. The thorax, 2nd ed. New York: Marcel Dekker; 1995.
18. Nosek TM, Leal-Cardoso JH, McLaughlin M, Godt RE. Inhibitory influence of phosphate and arsenate on contraction of skinned skeletal and cardiac muscle. Am J Physiol 1990;259:C933–C939.
19. Jones DA. Muscle fatigue due to changes beyond the neuromuscular junction. In: Porter R, Whelan J, editors. Human muscle fatigue: physiological mechanisms. London: Pitman Medical; 1981. p. 178–190.
20. Shindoh C, DiMarco A, Thomas A, Manubay P, Supinski G. Effect of N-acetylcysteine on diaphragm fatigue. J Appl Physiol 1990;68:2107–2113.
21. Anzueto A, Andrade FH, Maxwell LC, Levine SM, Lawrence RA, Gibbons WJ, Jenkinson SG. Resistive breathing activates the glutathione redox cycle and impairs performance of the rat diaphragm. J Appl Physiol 1992;72:529–534.
22. Moxham J, Wiles CM, Newham DD, Edwards RHT. Contractile function and fatigue. In: Porter R, Whelan J, editors. Human muscle fatigue: physiological mechanisms. London: Pitman Medical; 1981. p. 197–205.
23. Wragg S, Aquilina R, Moran J, Hanmegerd C, Green M, Moxham J. Diaphragm fatigue following maximum ventilation in man. Am Rev Respir Dis 1992;145:A147.
24. Tobin MJ, Perez W, Guenther SM, Semmes BJ, Mador MJ, Allen SJ, Lodato RF, Dantzker DR. The pattern of breathing during successful and unsuccessful trials of weaning from mechanical ventilation. Am Rev Respir Dis 1986;134:1111–1118.
25. Mador MJ, Tobin MJ. The effect of inspiratory muscle fatigue on breathing pattern and ventilatory response to CO2. J Physiol (Lond) 1992;455:17–32.
26. Roussos C, Fixley M, Gross D, Macklem PT. Fatigue of inspiratory muscles and their synergic behavior. J Appl Physiol 1979;46:897–904.
27. Fitting JW, Bradley TD, Easton PA, Lincoln MJ, Goldman MD, Grassino A. Dissociation between diaphragmatic and rib cage muscle fatigue. J Appl Physiol 1988;64:959–965.
28. Tobin MJ, Perez W, Guenther SM, Lodato RF, Dantzker DR. Does rib cage–abdominal paradox signify respiratory muscle fatigue? J Appl Physiol 1987;63:851–860.
29. Bellemare F, Grassino A. Effect of pressure and timing of contraction on human diaphragm fatigue. J Appl Physiol 1982;53:1190–1195.
30. Zocchi L, Fitting JW, Majani U, Fracchia C, Rampulla C, Grassino A. Effect of pressure and timing of contraction on human rib cage muscle fatigue. Am Rev Respir Dis 1993;147:857–864.
31. Bellemare F, Grassino A. Force reserve of the diaphragm in COPD patients. Appl Physiol 1983;55:8–15.
32. Vassilakopoulos T, Zakynthinos S, Roussos C. The tension–time index and frequency–tidal volume ratio are the major pathophysiologic determinants of weaning failure and success. Am J Respir Crit Care Med 1988;158:378–385.
33. Bai TR, Rabinovitch J, Pardy RL. Near-maximal voluntary hyperpnea and ventilatory muscle function. J Appl Physiol 1984;57:1742–1748.
34. Loke J, Mahler DA, Virgulto JA. Respiratory muscle fatigue after marathon running. J Appl Physiol 1982;52:821–824.
35. Chevrolet JC, Tschopp JM, Blanc Y, Rochat T, Junod AF. Alterations in inspiratory and leg muscle force and recovery pattern after a marathon. Med Sci Sports Exer 1993;25:501–507.
36. Nava S, Zanotti E, Ambrosino N, Fracchia C, Scarabelli C, Rampulla C. Evidence of acute diaphragmatic fatigue in a “natural” condition: the diaphragm during labor. Am Rev Respir Dis 1992;146:1226–1230.
37. Bye PTP, Esau SA, Walley KR, Macklem PT, Pardy RL. Ventilatory muscles during exercise in air and oxygen in normal men. J Appl Physiol 1984;56:464–471.
38. Miller JM, Moxham J, Green M. The maximal sniff in the assessment of diaphragm function in man. Clin Sci 1985;69:91–96.
39. Laroche CM, Mier AK, Moxham J, Green M. The value of sniff esophageal pressures in the assessment of global inspiratory muscle strength. Am Rev Respir Dis 1988;138:598–603.
40. Héritier F, Rahm F, Pasche P, Fitting JW. Sniff nasal inspiratory pressure: a noninvasive assessment of inspiratory muscle strength. Am J Respir Crit Care Med 1994;150:1678–1683.
41. Uldry C, Fitting JW. Maximal values of sniff nasal inspiratory pressure in healthy subjects. Thorax 1995;50:371–375.
42. Esau SA, Bellemare F, Grassino A, Permutt S, Roussos C, Pardy RL. Changes in relaxation rate with diaphragmatic fatigue in humans. J Appl Physiol 1983;54:1353–1360.
43. Esau SA, Bye PTP, Pardy RL. Changes in rate of relaxation of sniffs with diaphragmatic fatigue in humans. J Appl Physiol 1983;55:731–735.
44. Koulouris N, Vianna LG, Mulvey DA, Green M, Moxham J. Maximal relaxation rates of esophageal, nose, and mouth pressures during a sniff reflect inspiratory muscle fatigue. Am Rev Respir Dis 1989;193: 1213–1217.
45. Mador MJ, Kufel TJ. Effect of inspiratory muscle fatigue on inspiratory muscle relaxation rates in healthy subjects. Chest 1992;102:1767–1773.
46. Kyroussis D, Mills G, Hamnegard CH, Wragg S, Road J, Green M, Moxham J. Inspiratory muscle relaxation rate assessed from sniff nasal pressure. Thorax 1994;49:1127–1133.
47. Mulvey DA, Koulouris NG, Elliott MW, Moxham J, Green M. Maximal relaxation rate of inspiratory muscle can be effort-dependent and reflect the activation of fast-twitch fibers. Am Rev Respir Dis 1991;144:803–806.
48. Kyroussis D, Polkey MI, Keilty SEJ, Mills GH, Hamnegard CH, Moxham J, Green M. Exhaustive exercise slows inspiratory muscle relaxation rate in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 1996;153:787–793.
49. Goldstone JC, Green M, Moxham J. Maximum relaxation rate of the diaphragm during weaning from mechanical ventilation. Thorax 1994;49:54–60.
50. Lippold OCJ, Redfearn JWT, Vuco J. The electromyography of fatigue. Ergonomics 1960;3:121–131.
51. Stephens JA, Taylor A. Fatigue of maintained voluntary maximal contraction in man. J Physiol (Lond) 1972;220:1–18.
52. Aubier M, Trippenbach T, Roussos C. Respiratory muscle fatigue during cardiogenic shock. J Appl Physiol 1981;51:499–508.
53. Grassino E, Goldman MD, Mead J, Sears TA. Mechanics of the human diaphragm during voluntary contractions: statics. J Appl Physiol 1978; 44:829–839.
54. McKenzie DK, Gandevia SC. Electrical assessment of respiratory muscles. In: Roussos C, editor. Lung biology in health and disease, 2nd ed., Part B. Vol. 85: The thorax. New York: Marcel Dekker; 1996. p. 1029–1048.
55. Gandevia SC, McKenzie DK. Human diaphragmatic EMG: changes with lung volume and posture during supramaximal phrenic stimulation. J Appl Physiol 1986;60:1420–1428.
56. Beck J, Sinderby C, Weinberg J, Grassino A. Effects of muscle-to-electrode distance on the human diaphragm electromyogram. J Appl Physiol 1995;79:975–985.
57. Sinderby C, Lindstrom L, Comtois N, Grassino A. Effects of diaphragm shortening on the mean action potential conduction velocity in canines. J Physiol (Lond) 1996;490:207–214.
58. Sinderby CA, Comtois A, Thomson R, Grassino AE. Influence of the bipolar electrode transfer function on the electromyogram power spectrum. Muscle Nerve 1996;19:290–301.
59. Sinderby C, Lindstrom L, Grassino AE. Automatic assessment of electromyogram quality. J Appl Physiol 1995;79:1803–1815.
60. Gross D, Grassino A, Ross WRD, Macklem PT. Electromyogram pattern of diaphragmatic fatigue. J Appl Physiol 1979;46:1–7.
61. Pardy RL, Bye PTP. Diaphragmatic fatigue in normoxia and hyperoxia. J Appl Physiol 1985;58:738–742.
62. Moxham J, Edwards RHT, Aubier M, De Troyer A, Farkas G, Macklem PT, Roussos C. Changes in EMG power spectrum (high/low ratio) with force fatigue in man. J Appl Physiol 1982;53:1094–1099.
63. Bellemare F, Grassino A. Evaluation of human diaphragm fatigue. J Appl Physiol 1982;53:1196–1206.
64. DeLuca CJ. Myoelectric manifestations of localized muscular fatigue in humans. Crit Rev Biomed Eng 1984;11:251–279.
65. Moxham J, Morris AJR, Spiro SG, Edwards RHT, Green M. Contractile properties and fatigue of the diaphragm in man. Thorax 1981;36:164–168.
66. Moxham J, Wiles CM, Newham D, Edwards RHT. Sternomastoid muscle function and fatigue in man. Clin Sci Mol Med 1980;59:463–468.
67. Johnson BD, Babcock MA, Sumanand OE, Dempsey JA. Exercise induced diaphragmatic fatigue in healthy humans. J Appl Physiol 1993;460:385–405.
68. Yan S, Gauthier AP, Similowski T, Faltus R, Macklem PT, Bellemare F. Force–frequency relationships of in vivo human and in vitro diaphragm using paired stimuli. Eur Respir J 1993;6:211–218.
69. Polkey MI, Kyroussis D, Hamnegard CH, Hughes PD, Rafferty GF, Moxham J, Green M. Paired phrenic nerve stimuli for the detection of diaphragm faigue in humans, Eur Respir J 1997;10:1859–1864.
70. Wragg S, Hamnegard C, Road J, Kyroussis D, Moran J, Green M. Potentiation of diaphragmatic twitch after voluntary contraction in normal subjects. Thorax 1994;49:1234–1237.
71. Similowski T, Yan S, Gauthier AP, Macklem PT, Bellemare F. Contractile properties of the human diaphragm during chronic hyperinflation. N Engl J Med 1991;325:917–923.
72. Mier-Jedrzejowicz A, Brophy C, Moxham J, Green M. Assessment of diaphragm weakness. Am Rev Respir Dis 1988;137:877–883.
73. Similowski T, Fleury B, Launois S, Cathala HP, Bouche P, Derenne JP. Cervical magnetic stimulation: a new painless method for bilateral phrenic nerve stimulation in conscious humans. J Appl Physiol 1989; 67:1311–1318.
74. Petitjean M, Bellemare F. Phonomyogram of the diaphragm during unilateral and bilateral phrenic nerve stimulation and changes with fatigue. Muscle Nerve 1994;17:1201–1209.

Mechanical properties and functions of the chest wall can be assessed by measurements of lung volume displacement, chest wall motion, and pressures measured at various locations in the chest wall. Respiratory muscle activation can be further characterized by electromyography (EMG). Techniques of pressure measurement are presented in Section 2 of this Statement. The framework for interpreting pressures in the chest wall is presented in this article.


In respiratory mechanics, it is important to distinguish between the two uses of the word “pressure.” In one case it denotes a pressure measured at a given location, as in “pleural pressure.” In the other case it denotes a difference in pressure between two points, usually on opposite sides of a structure, such as “transpulmonary pressure,” defined as the difference between pressure at the airway opening (Pao) and pressure in the pleural space (Ppl). Pressures are usually measured relative to barometric pressure (i.e., they are taken to be zero when they are equal to barometric pressure).

Pressures at a point are usually assumed to be representative of the pressure in that space (see Figure 1

in Section 2 of this Statement). This simplification must be qualified when variations of pressure within a space are to be expected (1). In particular, gravity causes vertical gradients in pressure related to the density of the semisolid or liquid contents of a space: in the thorax, this gradient is approximately 0.2 cm H2O/cm height and is affected by lung density; in the abdomen, the gradient is nearly 1 cm H2O/cm height. Temporal fluctuations in pressure, as in tidal breathing, are little affected by gravitational gradients. Shear stress resulting from the deformation of elastic, shape-stable organs can cause local variations in pressure, such as those that occur just below the diaphragm when it displaces the liver during a large forceful contraction (2).

Pressure differences across structures, as opposed to pressures measured at a point, are relevant for characterizing those structures. The schematic drawing in Figure 1 of Section 1 of this Statement shows relationships among locations where pressures can be measured (within circles) and intervening respiratory structures and equipment (within rectangles). Pleural and abdominal pressures are usually estimated by measuring esophageal and gastric pressures (Pes and Pga), respectively. Table 1 and Figure 1 in Section 2 of this Statement list pressures measured at a point and pressure differences across structures. These differences are usually taken in a direction such that positive pressure differences expand the structure (e.g., lung). An exception to this rule is transdiaphragmatic pressure (Pdi), which has been defined both as pleural pressure minus abdominal pressure, Pdi = Ppl − Pab, and as its reverse, Pdi = Pab − Ppl. The complicating effects of gravity must be considered when pleural pressure is estimated from esophageal pressure (Pes) and abdominal pressure is estimated from gastric pressure (Pga). When the diaphragm itself is completely relaxed and the actual pressure difference across the diaphragm is nil, the measured transdiaphragmatic pressure has a minimum value, usually approximately 10 cm H2O, which is attributed mostly to the gravitational hydrostatic difference between esophageal and gastric pressures. This hydrostatic transdiaphragmatic pressure, which changes only slightly with breathing (3), is usually subtracted from reported measurements of Pdi.

A pressure difference between two points may characterize two or more different structures or groups of structures. For example, the pressure difference between the pleural space and the body surface in a spontaneously breathing person is both the transthoracic (transchest wall pressure, Pcw) and the negative of transpulmonary pressure (−Pl).

Scientific Basis

Pressure differences across viscoelastic, plastoelastic structures such as the lungs and chest wall depend on the structure's volume, volume history, and rate of change of volume. Accordingly, pressure differences across respiratory structures are often represented as characteristic pressure–volume (PV) curves (4). For the relaxed respiratory system, transpulmonary, transthoracic, and transrespiratory pressures are usually plotted against lung volume in a Rahn diagram (Figure 1). The PV characteristics shown in Figure 1 are of a relaxed subject slowly inflated or deflated by a pressure source at the airway opening. All the passive structures show an increase in volume with an increase in the pressure difference across them. When two pneumatic structures are in series, for example, the lung and the chest wall, the pressure difference across both structures (the transrespiratory pressure) is the sum of the pressure differences across each, and the volume displacements of the whole are equal to the volume displacements of each part. The PV curve can be locally described by the volume at a given pressure and the slope (compliance) at that point. The compliance of a passive structure at a given volume is the ratio of volume change to pressure change (i.e., the slope of the characteristic PV curve at that volume).


Lung volume displacements and pressures are measured as described in Sections 1 and 2 of this Statement. The following properties of the chest wall and lung are found by analysis of Rahn diagrams:

  1. In trained subjects, the static PV characteristic of lung compliance is obtained from transpulmonary pressure (Pl = Pao − Pes) during an interrupted exhalation from total lung capacity (TLC) to residual volume (RV) with the glottis held open. The quasistatic deflation curve is similar, and is measured during a slow exhalation (expiring at flows less than 0.3 L/second). For subjects who cannot satisfactorily perform these maneuvers, Pl can be measured during intermittent airway occlusions, 2–4 seconds long, in an expiration from TLC to RV, or during interrupted deflation of the respiratory system with a supersyringe, valve, or other device.

  2. The PV characteristic of the relaxed chest wall (Pcw) is obtained from esophageal pressure (Pes) during a slow relaxed exhalation through pursed lips or other high resistance from TLC to functional residual capacity (FRC) and during passive inhalation with relaxation against an intermittently occluded airway between RV and FRC. Alternatively, it can be measured during exhalation from TLC to RV with periodic airway occlusions with relaxation. Relaxation above FRC, however, may be difficult for untrained subjects, and relaxation below FRC can usually be achieved only by highly trained subjects. Normal compliance at FRC is approximately 0.2 L/cm H2O.

  3. The PV characteristic of the relaxed respiratory system is obtained by plotting Pao versus Vl during the maneuvers described above (Prs = Pao − Pbs [body surface pressure]). Normal values are approximately 0.1 L/cm H2O at volumes of 40−60% of the vital capacity (VC).


The Rahn diagram is useful for describing the elastic properties of passive systems. Each curve reflects the pressure difference developed by this structure for a range of volumes. Static compliance determined from these curves can be used for diagnosis. For example, the compliance of the lung is decreased in interstitial lung disease and increased in emphysema. Chest wall compliance is decreased in ankylosing spondylitis and obesity.


Whereas the lungs' PV characteristic (i.e., the plot of elastic recoil pressure of the lung, Pl,el, versus lung volume) is relatively easy to obtain in untrained subjects, the elastic recoil pressure of the chest wall (Pcw,el versus VL) is difficult to measure because it requires complete relaxation of the respiratory muscles at various lung volumes. Relaxation can be monitored via surface EMG of the chest wall. Failure to relax the respiratory muscles is also revealed when repeated measurements of passive PV curves of the chest wall are not identical. In subjects who can achieve relaxation at FRC but not at volumes above FRC, a relaxation curve can be approximated by extending a line from the relaxation point, assuming a normal chest wall compliance of 0.1 L/cm H2O. The chest wall compliance can also be estimated in untrained subjects by the weighted spirometer technique (5). These estimations may, however, be unreliable in hyperinflated patients (e.g., in chronic obstructive pulmonary disease) who never reach a true (static) relaxation volume even when they are relaxed during exhalation. This is a well-established test, although it is seldom used for clinical diagnosis.

Scientific Basis

To evaluate respiratory muscle action, a Campbell diagram, in which lung volume on the ordinate is plotted against pleural pressure on the abscissa, can be constructed. In this diagram, pleural pressure has differing significance depending on the maneuver. Consider a subject who is slowly inflated and deflated passively by a syringe connected to the airway while the respiratory muscles are relaxed (passive inflation; Figure 2)

. The pleural pressure, which is equal to transthoracic pressure, rises and falls, describing the characteristic PV curve for the relaxed chest wall, which is the same as that in the Rahn diagram (Figure 1). Alternatively, during active slow inhalation and exhalation with an open glottis, the pleural pressure (in this case equal to transpulmonary pressure with a negative sign) becomes more subatmospheric as the lungs inflate (active inflation; Figure 2), describing the lungs' characteristic PV curve, which appears as a mirror image of the lungs' curve in Figure 1. In the Campbell diagram the two curves intersect at relaxation volume at a pleural pressure of about −5 cm H2O. The intersection represents the equal and opposite elastic recoils of the lung and chest wall.

During inhalation, the pleural pressure is the pressure across the active chest wall. The pressure generated by the inspiratory muscles is simply the pressure difference between the active chest wall characteristic and the relaxed chest wall characteristic. Work done by the inspiratory muscles,

is represented by the hatched area in Figure 2. The horizontally hatched area represents the work done to overcome elastic recoil of the lung and chest wall. Additional pressure is necessary to overcome airway resistance and lung tissue resistance; this work is shown with vertical hatching. Total work is therefore the sum of elastic and resistive work per inhalation, and is usually multiplied by breathing frequency and expressed as g · cm/ml. Work of breathing was found to average 2.2 ± 0.92 g · cm/ml at a respiration frequency of 15 breaths/minute (6), and was independent of age or sex. This is a test of great physiological interest and is widely used in research. It is seldom used for clinical evaluations. Reference 6 gives a full account of the complexities of how the work is done by the coordination of the various respiratory muscles.

Figure 3

shows a Campbell diagram with the addition of maximal static inspiratory pressure (MIP) and maximal expiratory pressure (MEP) during efforts against an occlusion (outer dashed lines). The MIP is greatest (most subatmospheric) at low lung volumes, whereas the MEP is greatest (most positive) at high lung volumes, largely because of the length–tension characteristics of the inspiratory and expiratory muscles. At high lung volume, the diaphragm and other inspiratory muscles are shorter, whereas expiratory muscles are longer (4). The inner dashed and dotted line in Figure 3 indicates the pleural pressure required to balance the elastic recoil of the lungs. TLC is at the intersection of these lines, where maximal inspiratory pleural pressure is balanced by the lungs' elastic recoil. The innermost loop represents resting breathing at FRC, as is shown for inspiration only in Figure 3.

The pressure–volume relationship during a maximal forced inspiration and expiration are shown as the inner solid line loop in Figure 3. Pressures at every volume are reduced from maximal static pressures because the muscle is shortening. The loss of maximal inspiratory pressure (the difference in pressure between the dashed and solid lines at a given lung volume in Figure 3) was estimated to be approximately 7%/L/second of flow at a volume of FRC + 1 L (7). The muscle pressure at this volume represents the maximal capacity (Pcap) of the inspiratory muscle to generate pressure while shortening maximally. During submaximal exercise, for example, healthy subjects may require esophageal pressures in the range of −30 cm H2O to ventilate the lungs, or approximately 40% of Pcap. When pressures in that range are achieved while breathing for several minutes with a duty cycle of 0.5, muscle fatigue may result (see Section 5 of this Statement). At higher levels of ventilation (maximal voluntary ventilation), at which peak flow can reach up to 10 L/second, there is a decrease in maximal inspiratory muscle pressure within 15–20 seconds, attributed to fatigue.


Measurements of pressure and volume are described in Sections 1 and 2 of this Statement.


The following properties of the chest wall and lung are found by analysis of Campbell diagrams:

  1. The quasistatic (or static) PV characteristic of the lung is obtained from esophageal pressure (Pes = −Pl,el) during a slow (or halting) inhalation and exhalation with the glottis held open.

  2. The PV characteristic of the relaxed chest wall is described above as in the Rahn diagram.

  3. Pressure generated by respiratory muscles (Pmus) is the horizontal distance (i.e., change in pressure) between the relaxation characteristic of the passive chest wall and the active one.

  4. The Campbell diagram is also used to depict the maximal static inspiratory pressure (MIP) and maximal static expiratory pressure (MEP) measured with the airway occluded. The values of MIP in healthy young males are shown as the dashed outer loop in Figure 3. Section 2 of this Statement discusses the technique used to perform the MIP test and gives normal values in health and disease.


The Campbell diagram is a convenient tool for calculating the elastic and resistive work of inspiratory and expiratory muscles. The relationship between work and oxygen consumption makes it possible to calculate the efficiency of respiratory muscles (see Figure 3 in Section 4 of this Statement). The difference between maximal static inspiratory pressures and peak actual pleural pressure measured during breathing indicates the muscle force reserve, an index that helps assess the likelihood of fatigue. This is a well-established method, often used in clinical research.


To infer respiratory muscle action, the pleural pressure must be referred to the PV curve of the “relaxed” chest wall as in the example above. For example, a given positive pleural pressure may reflect either expiratory muscle activity at FRC or slight inspiratory muscle activity at a high lung volume. Furthermore, difficulties in measuring the relaxation characteristic in untrained subjects can make estimates of Pmus uncertain. Although the Campbell diagram is easy to plot, its interpretation requires some practice.

Scientific Basis

Quantitative measurements of respiratory system motion are usually based on measurements of lung volume and displacements of chest wall structures, including the abdominal wall. Because the tissues of the chest wall are essentially incompressible, volume changes (displacements) of the chest wall surfaces are nearly equal to volume changes of the lungs and can be used to “noninvasively” estimate lung volume, without the need for a mouthpiece, mask, or other connection to the airway. (Changes in intrathoracic blood volume cause differences between chest wall and lung volume displacements, but these are usually negligible.) The relevance of noninvasive measurements of ventilation is shown in Breathing Patterns in Section 10 of this Statement.

During quiet breathing in a subject at rest, the surface of the chest wall moves in a predictable way as lung volume increases and decreases. Various dimensions of the chest wall have been measured to estimate changes in chest wall (or lung) volume. The “chest pneumogram” is a record of changes in thoracic circumference that provides a qualitative measure of ventilation. However, because the major compartments of the chest wall, the rib cage and abdomen, move independently in most conscious subjects, measuring overall chest wall displacement accurately usually requires two or more simultaneous measurements of displacement.

Konno and Mead (8) established that chest wall volume change could be assessed by measuring displacements of rib cage and abdominal surfaces. In their subjects, who were standing still, a single anteroposterior diameter sufficed to indicate motion of the rib cage; this was also true of the abdomen. Thus, the chest wall can be described as having two principal degrees of freedom of motion: Overall chest wall displacements can be specified by knowing the rib cage displacement and the abdominal displacement. The two-compartment model of the chest wall introduced by Konno and Mead and the plot of rib cage displacement versus abdominal displacement (the Konno–Mead diagram) have been used in numerous studies.

Figure 4

shows a Konno–Mead diagram of a subject breathing quietly and performing an “isovolume maneuver.” During quiet inspiration, the rib cage and abdomen move out synchronously, following the rib cage–abdomen relaxation characteristic. In the isovolume maneuver, the subject voluntarily shifts volume between rib cage and abdominal compartments by contracting and relaxing abdominal muscles with the glottis closed. Because lung volume is constant, the decrease in abdominal volume (i.e., the volume displaced by inward movement of the abdominal wall) must be equal to the increase in rib cage volume; two isovolume maneuvers performed at known lung volumes allow calibration of rib cage and abdominal displacements in terms of lung volume change.

Most methods of measurement of thoracoabdominal displacement require that the subject maintain a constant posture, as spinal flexion affects the relationships among lung volume, rib cage, and abdominal displacements (9). Therefore, postural change should, as a rule, be minimized. However, it is possible to measure spinal flexion in addition to rib cage and abdominal displacements to estimate lung volume in subjects whose posture changes (10). This method is widely used in clinical research to assess tidal volume and the relative displacement of rib cage–diaphragm.

Methodology: Pneumograph

Chest pneumographs are strain gauges that measure thoracic circumference. They can be constructed simply from a bellows that generates pressure when stretched, or an elastic tube containing mercury that changes electrical resistance when stretched. They are used principally for qualitative measurements, for example, breathing frequency.


They are inexpensive, and useful as qualitative measures of ventilation.


These devices are not accurate in subjects who are moving or whose breathing movements are not stereotypical.

Methodology: Respiratory Magnetometer

The first device to be used routinely for quantitative measurements of chest wall displacements was the respiratory magnetometer (11), which uses pairs of small electromagnetic coils fixed to the skin to measure anteroposterior or other diameters of rib cage or abdomen.


These electronic calipers are precise, accurate, and consistent, allowing repeated measurements in an individual over many days (12). Also, because they measure diameters (intercoil distance), they are useful for documenting distortions of the chest wall shape as during asthma attacks (13) or forceful respiratory efforts (14). They are also useful for measuring ventilation at rest.


Conversely, when thoracic shape distortions occur, the respiratory magnetometer becomes less accurate in measuring overall chest wall volume. Whereas during quiet breathing, one rib cage and one abdominal signal are adequate to describe chest wall displacements and lung volume changes, during forceful or unusual respiratory efforts, changes occur in the ellipticity of the rib cage cross-section and in the relative displacements of cephalic and caudal regions of the rib cage, making a single diameter inadequate for assessment of volume displacement. In such circumstances, the addition of a transverse diameter measurement can improve volumetric accuracy (13).

In all applications, the orientation of the electromagnetic coils should be controlled to keep the axes of paired coils parallel to each other to avoid errors caused by pitch and yaw. For transverse diameter measurements, in which the sides of the rib cage are not parallel, the coils may be fixed to lightweight calipers and held against the body wall by adhesive tape to maintain coil orientation (13). Displacements of the integument and soft tissues can introduce artifactual signals in obese subjects at rest and in all subjects during running, moving in bed, and so on, and thus the devices are less useful in these cases.

Methodology: Respiratory Inductive Plethysmograph

Another device that has gained acceptance since its introduction more than 15 years ago is the respiratory inductive plethysmograph (RIP), which uses two elastic bands that surround the rib cage and abdomen to provide a signal that varies with cross-sectional area. The RIP can measure ventilation within about ± 5% compared with the spirometer, and can reveal the relative contributions of rib cage and abdomen to breathing. This device is widely used to monitor ventilation.


The RIP is easy to use. Because the belts encircle a large part of the rib cage and abdomen, integumental mobility and cross-sectional shape distortions are less problematic than with magnetometers, and therefore the RIP is often used during sleep or exercise. The RIP is more accurate than magnetometers in estimating lung volume change (15), perhaps because its signal varies with cross-sectional area and samples a larger part of the moving chest wall than do magnetometers.


The RIP measures changes relative to an unknown baseline, and its accuracy depends on lung volume calibration at the time of study (see below). In addition, movement of the belts or changing body position can affect calibration.

Calibration of RIP and magnetometers.

Numerous methods for calibrating the magnetometer and RIP signals for estimating lung volume displacement have been described. All methods determine coefficients for rib cage and abdominal signals, and most require simultaneous spirometric measurements of lung volume change during periods of