One would think that it should be possible to quantify “severity” with great precision, but this is not yet the case. Part of the difficulty is that, unlike other illness, asthma severity is not defined solely as the extent of impairment in organ function. Rather, the definition consists of an arbitrary combination of the signs and symptoms present and the intensity of the cardiorespiratory abnormalities observed. According to consensus guidelines, a severe episode is believed to exist when one or more of the following features are present: accessory muscle activity, a paradoxical pulse exceeding 25 mm Hg, a heart rate greater than 110 beats/minute, a respiratory rate greater than 25–30 breaths/minute, a limited ability to speak, a peak expiratory flow rate (PEFR) or FEV₁ < 50% of predicted, and an arterial saturation less than 91–92% (3–6). In this paradigm, a PEFR or FEV₁ < 35% of predicted is considered by some to represent a life-threatening episode (4–6). Although such a classification seems straightforward, it has inherent limitations that interfere with its precision. For example, the components do not develop simultaneously or at unique levels of impairment, so they often relate poorly to one another (7–12). Further, their presence or absence does not predict outcomes (13–17).

Figure 1

Figure 1. Distribution of clinical and physiologic parameters in acutely ill patients with asthma.

Download Figure | Download Powerpoint

[More] [Minimize]

Using the above-cited definitions, 436 subjects (74%) had PEFR values < 50% of expected and would have been considered as having had a severe episode, whereas 47% would have been considered to have had life-threatening attacks. However, only 20 and 26%, respectively, would have met the requirements for tachycardia and tachypnea and 11% would have been judged as having significant hypoxemia. In point of fact, less than a quarter of the attacks in this group were severe enough to merit hospitalization and none put the participants at risk for death. As in other studies, 78% of the episodes cleared within 2 hours irrespective of the initial presentations (13–17). Twenty-two percent were believed to need hospitalization. Of these, 2% were admitted to the medical intensive care unit (MICU). No one required ventilatory support and there were no fatalities. These data suggest that the currently used definitions are not reflective of the events seen in practice and that it might be worthwhile to make them more evidence driven.

The distribution in Figure 1 is not unique (7, 8, 19, 20). In a large multicenter description of asthma exacerbations, the mean PEFR on presentation was 44% of predicted (20). Seventy-five percent of the participants had values less than 50% of predicted and 26% were in the life-threatening category (20). About half of the latter were discharged from the ED. Three patients died but the cause and severity were not described.

What, then, is a severe episode? Pragmatically, it is one that places patients in peril and/or requires prolonged therapy to resolve. A useful way of making that distinction relatively quickly is to incorporate the immediate response to β₂ agonists. All hospitalized patients in Figure 1, and others that we have examined (13–17), presented with low PEFR values that failed to increase to 45% of predicted or more after administration of 5 to 10 mg of nebulized albuterol (Figure 2B)

Figure 2. Cumulative dose–response relationships for albuterol in acutely ill patients with asthma. The ordinate indicates the peak expiratory flow rate (PEFR) as a percentage of predicted normal and the abscissa indicates the cumulative dose of albuterol in milligrams. (A) Response of the entire population; (B) response of those admitted and discharged. Data points represent mean values and error bars represent 1 SEM. Reprinted by permission from Reference 16.

Download Figure | Download Powerpoint

[More] [Minimize]

The incidence and prevalence of severe episodes of asthma are unknown and vary as a function of the definition employed. In addition to the clinical and physiological parameters given above, hospitalizations, admissions to intensive care units, CO₂ retention, intubation and mechanical ventilation, as well as fatal and near-fatal episodes have all been employed as manifestations of severity. On the basis of the most recent information from the National Center for Health Statistics (Hyattsville, MD) on the incidence of acute episodes and admissions (1) and assuming that hospitalization is a perfect surrogate, 4% of the attacks nationally and 21% of all episodes treated in urgent care centers would be considered severe. Intensive care admissions range from 2 to 20% (13, 21, 22), the presence of hypercarbia varies from 10 to 63% (21–23), and intubation and respiratory support fluctuates between 2 and 70% (Table 1)

Most asthma attacks are short-lived and clear with removal of the offending agent, but outside of laboratory provocations, there are no systematic data on spontaneous recovery. In controlled situations, the more intense the stimulus the longer the destabilization takes to resolve, but even at the height of an induced reaction the obstruction dissipates rapidly after use of a bronchodilator. Studies on naturally occurring episodes are less clear-cut. There is no consistent relationship between the duration of a attack and the physiological changes that ensue (23, 27), and patients with apparently similar degrees of impairment can recover at different rates (13–17). Such observations suggest that undefined qualitative or quantitative differences in pathophysiology may exist depending on the initiating event.

The reason people develop persistent complaints that bring them to medical attention is not clear. Historically, upper respiratory tract infections, discontinuation of medications, severe psychological stress, and contact with noxious immunologic or irritant stimuli are frequent antecedent events. One tends to think of the immediate development and progression of symptoms, but such events are relatively infrequent. About 13 to 14% of people in large cohorts describe complaints that were less than 3 hours in duration (20, 28). Instead, symptoms are often present for days to weeks (29–31) and patients tend to seek resolution when their rescue medicines no longer work. Typically, there is a gradual increase in complaints and the use of rescue medications over 5–7 days before the attacks develop, followed by a more rapid rise in the immediate 2- to 3-day period (31). The PEFR falls concurrently. Smaller studies have reported a prodromal increase in complaints before PEFR changed (32), but those observations are now called into question.

The pattern of recovery of acute decompensations is bimodal (Figure 2). As presented earlier, the majority clear within 1–2 hours with protocol treatment (Figure 2). Termination after ad hoc regimens is considerably longer (13). Of those sent home after directed care, about 3% return to the ED within 24 hours because of relapses and 7% return within 1 week (13). In studies using noncare path ED regimens, recurrences over the next 1–2 weeks vary between 15 and 17% (13, 28, 33). There are no differences in relapses as a function of the speed of onset of symptoms (28).

About one-fifth to one-third of ED patients have poor short-term responses to albuterol and require admission to the hospital (13, 17, 28) (Figure 2B). The manner in which such episodes ultimately resolve is presented in Figure 3

Figure 3. Resolution of acute asthma in hospitalized patients as measured by peak expiratory flow rate (PEFR). The ordinate indicates the PEFR as a percentage of predicted normal and the abscissa shows the recovery in hours. ED represents the amount of obstruction seen in the emergency department before treatment. Admit represents the PEFR when the patients were hospitalized. The subsequent observations were made at 12, 24, and 48 hours after admission. Data points represent mean values and error bars represent 1 standard error of the mean.

Download Figure | Download Powerpoint

[More] [Minimize]

An evaluation of MICU admissions and outcomes is presented in Table 1. The material was derived from 18 investigations published over the last quarter-century and represents 112 years of observation in 2,157 patients (21, 22, 24, 26, 29, 30, 34–45). In the studies listed, about 21 patients with asthma were admitted to the MICU per year. If the first study (Scoggin and coworkers [34]) is excluded, the number falls to 13–14 patients per year. As shown, criteria for MICU care were rarely stated. Intubation and ventilatory assistance was believed to be necessary in approximately one-third of admissions, but the range was extremely wide (3–70%). The considerations employed in making this decision were given in 40% of the publications and included exhaustion, progressive hypercapnia, unconsciousness, deterioration of mental status, cardiopulmonary arrest, or the clinical assessment of the attending physician. There was no uniformity apparent in their application. The frequency of intubation also varied considerably (range, fewer than 1–29 cases per year). Fatalities occurred in 2.7% of the total admissions (59 of 2,157) and 8.1% of those intubated (range, 0–38%). Mechanical ventilation was associated with 267 complications (number per intubation, 1.3 ± 0.6; range, 0–7%). Events included hypotension, pneumothorax, pneumomediastinum, atelectasis, nosocomial pneumonia, arrhythmias, sepsis, gastrointestinal bleeding, and cerebral anoxia. There were no apparent trends over time. Research is needed to determine why such variability in treatment and outcomes exists and whether it could be altered by the use of consensus criteria.

Morbidity and potential mortality increase in asthma with socioeconomic deprivation, ethnicity, urban dwelling, and comorbid issues such as drug abuse (46–49). The relative roles that single or aggregate elements play are not yet established. Members of minority populations living in large cities bear a disproportionate burden of the disease, are admitted more frequently, and are at greater risk for death (46–49). In the year 2000, the ED visits and hospitalization rates were 125 and 220% higher, respectively, among black non-Hispanics compared with white non-Hispanics (50).

At autopsy, the classic gross anatomic features of those who die from asthma are airway narrowing, extensive plugging of the airways with mucus and inflammatory infiltrates, hyperinflation, and atelectasis (51). Microscopically, exudation of plasma proteins, mucosal and submucosal edema, hypertrophy and hyperplasia of the bronchial smooth muscle and the microvasculature, denudation of the epithelium, and thickening of the subepithelial collagen layer are the norm (51). Infiltration of the large and small airways with eosinophils, neutrophils, plasma cells, and lymphocytes is common (52). Basophils are also found in increased numbers in fatal disease (53). Qualitatively similar changes in histology are seen in biopsies from asymptomatic patients (54, 55). There is deposition of Type III and IV collagen and fibronectin by myofibroblasts beneath the basement membrane in both biopsy and postmortem specimens (56), but the fibrosis and obliteration of small airways seen in chronic bronchitis and cigarette smoking is absent (51). Some fatal episodes are associated with a high percentage of activated cytotoxic CD8⁺ T lymphocytes in the airways, possibly because of rhinovirus and respiratory syncytial viral infections (57).

Airway wall area in fatal cases has been reported to be increased between 10 and 100% over that seen in nonfatal disease and between 50 and 300% over that in nonasthmatic control subjects (58, 59). Such increases may be related to the intensity of the disease. In nonlethal illness, wall thickening is most marked in midsized and small membranous bronchi. In patients dying from asthma, the thickening is more generalized (58, 59). Aging and the duration of asthma are associated with a progressive increase in smooth muscle volume and reduction in airway lumen, but these findings are not prerequisites for fatal attacks (60). Young individuals who die of their illness do not have greater wall thickness as compared with nonasthmatic control subjects (60).

Beneath the subepithelial collagen layers in the conducting airways lies a network of elastic fibers that form discrete longitudinal bundles (61). The elastic layer may contribute to the ability of the airways to resist deformation by altering the nature of the mucosal folds that form when smooth muscle contracts. This region is thickened in asthma (56, 62), but no significant differences have been found between the number of longitudinal bundles or mucosal folds in the lungs of nonfatal and fatal cases and control subjects (63). Although changes in wall structure can potentially impact the degree of airway narrowing that follows provocative stimuli, as yet there are no data on how they influence the intensity of acute episodes or the response to treatment.

The physiologic and clinical features of asthma are believed to derive from an interaction between the resident and infiltrating inflammatory cells (64–67). Those thought to play critical roles are mast cells, basophils, eosinophils, lymphocytes, and the airway epithelium. The parts played by macrophages and other cellular constituents are incompletely defined. Neutrophils are likely important but there are less data. Neutrophils and interleukin (IL)-8 levels are higher in the bronchoalveolar lavage fluid obtained from intubated patients with asthma and in sputum from patients with severe disease (68). They may be recruited by injury to the epithelium and their elastases are potent mucin secretagogues that can facilitate plasma exudation (69).

Eosinophils are prominent components of the cellular influx. Interleukin-5 stimulates the release of these cells into the circulation and extends their survival. Once activated, eosinophils are a rich source of leukotrienes, granular proteins (major basic protein and eosinophilic cationic protein), and oxygen-derived free radicals that are capable of destroying the epithelium. Besides resulting in a loss of barrier and secretory function, such damage elicits the production of chemotactic cytokines, leading to further inflammation. Chemokines and cytokines likely involved in eosinophil recruitment and transmigration from the capillaries such as monocyte chemotactic protein, RANTES, macrophage inflammatory peptide, and IL-5 are higher in patients with severe disease (70, 71). The role of the eosinophil in establishing and maintaining airway hyperresponsiveness is undergoing reevaluation. Studies using antibodies against IL-5 show a disassociation between inflammatory and physiologic events after an antigen challenge, and blood and sputum eosinophilia (72).

Normally lymphocytes represent 10 to 20% of the total cells recovered by bronchoalveolar lavage, but they are prominent cells in biopsies in asthma (52). In postmortem material, increased numbers of lymphocytes are distributed uniformly in the large and small airways of patients with asthma independent of disease severity and nearly all of them are activated T lymphocytes (52, 66, 73). T lymphocytes are defined by the profiles of the cytokines released. Helper T Type 1 cells are characterized by the production of IL-2 and interferon γ and are involved in developing cell immunity. Helper T Type 2 cells generate IL-4 and IL-5 and play a prominent role in humoral immunity. Interleukin-4 initiates the production of IgE and IL-5 influences eosinophils as described (66). After antigen challenge, there is a differential increase in helper T Type 2 cells and a downregulation of the helper T Type 1 series (66). Such an imbalance has been postulated to play an important part in allergic asthma, but it is not yet clear how it is involved in nonatopic illness. The manner in which the cellular constituents interact to initiate and/or sustain acute episodes also is not yet established.

Controlled laboratory provocations produce a variety of inflammatory mediators in blood and airway fluids (64), but there is incomplete information as to what transpires during spontaneously occurring acute episodes. Leukotrienes, proinflammatory mediators, metalloproteinases, activated T lymphocytes, eosinophils, neutrophils, and their associated cytokines have been reported in symptomatic patients with asthma, as have chemokines such as monocyte chemotactic protein-1, macrophage inflammatory peptide, RANTES, and IL-5 (66, 67, 71, 74). In some studies (75) high levels of eotaxin have been observed. This compound is produced by epithelial cells after stimulation by cytokines and recruits eosinophils. There is no information yet as to whether the inflammatory compounds associated with acute episodes are quantitatively or qualitatively specific to the initiating stimulus or if, and how, their levels impact the response to treatment.

Most patients suffering from acute asthma request therapy with a constellation of complaints consisting of dyspnea, cough, and wheezing (7, 13–17, 19). The first two can also occur as isolated entities (76). No sign or symptom is uniformly present. Dyspnea is absent in 17 to 18% of cases (19) and wheezing is absent in 5% (7, 13–17, 19). The physical signs that are encountered are tachypnea, tachycardia, wheeze, hyperinflation, accessory muscle use, pulsus paradoxus, diaphoresis, cyanosis, and obtundation (7, 13–17, 19). The use of accessory muscles is observed in about 30% of cases at presentation, a paradoxical pulse in 15–20%, sweating in 12%, and cyanosis in less than 1% (7, 8, 13–17, 19). There are no data on the prevalence of cough, but it is a prominent prodrome (77). Changes in mental status are appropriately viewed as heralding an impending catastrophe (3–6), but there are no data on how frequently they occur.

Sweating, the use of accessory muscles, a paradoxical pulse, and the inability to communicate in full sentences are all associated with the presence of substantial airway narrowing (18, 27, 78). It is critical to appreciate that the absence of such signs does not guarantee an insignificant episode. It is well established that patients with marked bronchial narrowing can have minimal symptoms (79).

Wheezing is a poor indicator of functional impairment (27). It often increases as the obstruction resolves and the patient's ability to move air improves. Rather, it is the absence of wheezing and/or the nature of the sounds that are important. A quiet chest in a dyspneic or obtunded patient with asthma is a serious event. In addition, although not formerly codified, clinical experience indicates that the combination of high-pitched wheezing with poor air movement is associated with marked bronchial narrowing.

One of the more important contributors to the symptomatic response is the degree of hyperinflation. The intensity of breathlessness rises and falls directly with changes in end-expiratory volumes (80). Simple increases in the work of breathing do not seem to be the cause because reducing them with inspiratory positive pressure has a limited impact on induced sensations (81). In contrast, compensating for the inspiratory threshold load (see subsequent text) with continuous positive-pressure breathing materially improves the subject's discomfort. Other elements are also involved, however. For the same degree of resistance and changes in end-expiratory volumes, histamine produces more dyspnea in normal people than does breathing with an external resistive load (82). Because this effect is blocked by inhaled lidocaine, vagal irritant receptors may be involved.

Some individuals cannot associate the magnitude of their obstruction with symptom intensity and, as a result, may delay seeking treatment until their pulmonary reserves are exhausted (83, 84). In the original description, 10% of the patients did not report any symptoms in the face of severe airflow limitation (84). Subsequently other authors have failed to find correlations between dyspnea ratings and peak flow measurements in as many as 60% of patients with asthma (85). It remains to be determined whether this is due to poor perception or other factors. As an example of the first point, children with a decreased ability to detect external resistant loads tend to have histories of near-fatal attacks (86). In addition, some (87, 88), but not all (89), investigations indicate that patients with asthma with a blunted perception of dyspnea may have more severe and potentially deadly episodes. They also have more ED visits, hospitalizations, and a greater death rate (88).

Impaired perception is associated with inflammatory markers in the sputum such as eosinophils and related proteins (90), but conflicting data exist on the effects of inhaled steroids (91, 92). Neither short- nor long-acting β₂ agonists appear to influence the ability to sense obstruction (93, 94).

The sine qua non of an episode of acute asthma is a nonuniform, reversible increase in airway resistance that results in diminished flow rates, premature airway closure, hyperinflation of the lungs and thorax, increased work of breathing, changes in elastic recoil, and frequency-dependent behavior (7–9, 19, 27, 95–100). In addition, there is abnormal distribution of ventilation and perfusion and altered arterial blood gases (9, 23, 96, 101–108). Symptomatic patients with marked obstruction and significant hyperinflation can demonstrate electrocardiographic evidence of pulmonary hypertension, right ventricular strain, hypotension, and poor peripheral perfusion (99, 109).

The abnormalities in pulmonary mechanics can be substantial. In large data sets, the FEV₁ or PEFR of patients entering acute care units average about 40% of predicted (7, 13–17, 19, 27). Substantial air trapping is common clinically, but the intensity of hyperinflation has been quantitated only in small studies (27). Residual volume (RV) can approach 400% of normal and functional residual capacity (FRC) can be double the expected values (27). During most attacks, total lung capacity (TLC) is close to predicted (27, 97).

The hyperinflation of acute asthma plays an important role in pathophysiology. As the resistive work of breathing rises, the inspiratory muscles must generate more force to offset the increase. Augmentations in lung volume help constricted bronchi remain open, thereby attenuating the lumenal reductions. Some of this happens by dynamic airway closure and some by reflexly narrowing the glottis and using it to dynamically slow or control respiration (110). As the lungs enlarge, however, it becomes progressively more difficult to inflate them further and the elastic work of breathing rises. This can be offset somewhat by reversibly shifting the pressure–volume relationship of the parenchyma upward and to the left, thereby reducing the force necessary to initiate respiration (95, 97, 98). The frequency with which this occurs is unknown but it tends to develop in severe attacks that are associated with increases in TLC (95, 97, 98).

The muscles of respiration can be compromised because hyperinflation increases the inspiratory threshold load that must be overcome (83). During the initial contraction of the inspiratory muscles, the volume of air in the thorax does not change until sufficient pressure is generated to override the elastic recoil of the respiratory system. This so-called auto-PEEP (positive end-expiratory pressure) rises as dynamic hyperinflation increases. As the thorax and lungs increase in volume, length–tension relationships shorten and the strength of contraction eventually diminishes. The radius of curvature of the diaphragm may also decrease until it is placed at a mechanical disadvantage. At some point, deflation can no longer remain passive and the expiratory muscles come into play. Ultimately, the accessory muscles of respiration also become active. The frequency of frank respiratory muscle fatigue in acute asthma is unknown, but is probably low. In an investigation, no evidence of muscle weakness or fatigue could be found in acutely ill patients even though their mean FEV₁ values were less than 50% of predicted (111).

There can be significant cardiac consequences associated with large rapid changes in lung volumes. Blood flow through the great veins in the chest becomes limited; pulmonary vascular resistance rises, presumably because of alveolar capillary compression; the pre- and afterload of the left ventricle increases; and cardiac output falls (18, 99, 109). With gas trapping, the negative swings in pleural pressure, decreases in stroke volume during inspiration, and high ventilatory frequencies all contribute to the pathogenesis of a paradoxical pulse (112). A clinical point of note is that changing the pattern of breathing may alter the intensity of the signs associated with hyperinflation. Deep breaths accentuate both the need for accessory muscle use and the intensity of pulsus paradoxicus.

Tachypnea rapidly develops in response to sudden bronchial narrowing, perhaps through the activation of irritant receptors (23). Because of the nonuniform distribution of the obstruction, there are regional changes in the time constants of the lungs (product of resistance times compliance) (96). As a result, the incoming air takes the path of least resistance and a relatively small number of alveolar units receives the majority of the tidal volume, resulting in maldistribution of ventilation (96). Regional hypoxemia then causes pulmonary blood flow to shift to the better ventilating units, but it cannot fully adapt, so ventilation–perfusion mismatches and alterations in arterial blood gases develop (96, 107, 108).

The typical blood gas abnormalities seen in acute asthma consist of a combination of hypoxemia, hypocapnia, and respiratory alkalosis (9, 23, 101–105). Generally, the more severe the obstruction, the lower the arterial oxygen tension (Pa_O2); however, with one exception (r = 0.73) the correlation coefficients in the literature are not high (r = 0.36 to 0.41) (9, 23, 101–105). Typically, most asthma attacks are not associated with marked hypoxemia and arterial desaturation (Figure 1) (9, 23, 101–105). The average Pa_O2 is about 69 mm Hg at sea level and has been reasonably consistent from study to study (23, 103). Unlike chronic obstructive lung disease, oxygen tensions less than 50 mm Hg are infrequent and were seen in just 8% of the patients in the seminal investigations (23, 101–105). Values less than 40 mm Hg are even more uncommon and have been reported in less than 2% of cases (9, 23, 101–105). This is the likely reason why cyanosis is such an infrequent occurrence. Resolution of the hypoxemia is slow. The ventilation–perfusion mismatching can last for weeks after a single episode (104) because of persistent peripheral airway obstruction (96, 100).

Alveolar hypoventilation can develop de novo from extensive airway plugging and overwhelming obstruction, pharmacologic depression of respiration, ventilatory fatigue from breathing against a large mechanical load, or from an abnormality in ventilatory control mechanisms. However, patients generally hyperventilate, causing a reduction in the Pa_CO2 (mean, 34 mm Hg at sea level) (23, 102). Retention of CO₂ is observed in about 10% of cases requesting emergency care (23). When it occurs, the FEV₁ is typically 20% of predicted or less (23). Even then, the elevation is modest, averaging from 10 to 15 mm Hg over normal (i.e., 50–55 mm Hg) at sea level (23, 101–104). Again, unlike chronic obstructive lung disease, Pa_CO2 values between 60 and 70 mm Hg are infrequent (less than 7% of patients). Values exceeding 70 mm Hg are extremely rare (0.4%) unless ventilation is actively depressed. Normocarbia tends to occur in about 15 to 20% of cases. Because it is frequently associated with a marked reduction in FEV₁ (usually less than 25% of predicted), it should be viewed as impending respiratory failure and treated as such (9, 23, 101–104).

The majority of patients have respiratory alkalosis (9, 23, 101–104). The prevalence of respiratory acidosis follows that of hypercarbia. With extreme airflow limitation, metabolic acidosis may be seen (23). Several mechanisms are potentially involved. If cardiac output is compromised, hypoxia of the peripheral tissues may cause lactic acidosis to develop or worsen (23, 106). In addition, increased oxygen consumption by the respiratory muscles may also contribute. It may also be induced with the aggressive administration of nonselective sympathomimetics (113).

Other conditions that can be confused with asthma include chronic obstructive lung disease, bronchiectasis, endobronchial lesions, foreign bodies, extra- or intrathoracic narrowing of the trachea, cardiogenic and noncardiogenic pulmonary edema, pneumonia, and pulmonary emboli. Misdiagnoses are present in about 1% of chart reviews of general asthma admissions in our institution and have been reported in as many as 8 to 25% of intensive care unit admissions (21, 25).

One particularly troubling syndrome is glottic dysfunction (114, 115). This condition is a conversion reaction with paradoxical adduction of the vocal cords and closely mimicks asthma. The obstruction can be of sufficient magnitude to cause hypoxemia and hypercapnia (114). Unlike asthma, however, the blood gas pattern is one of central alveolar hypoventilation. One helpful physical finding is that the sounds of air movement over the neck either halt abruptly during inspiration or expiration, or there is continuous high-pitched wheezing and/or stridulous breathing.

The diagnosis is definitively established by visualization of the glottis when the patient is symptomatic. If not performed then, the characteristic dysfunctional movement of the cords will be absent and one must ultimately rely on provocational studies to produce it. Because it is possible that asthma and glottic dysfunction can coexist, it is essential that both the extra- and intrathoracic airways be evaluated. The author's laboratory uses fiberoptic techniques to monitor the movement of the cords, followed by a ventilation scan of the chest. Because the intrathoracic airways are not involved in glottic dysfunction, the distribution of ventilation is normal; it is maldistributed with asthma (96, 107). Because of the prolonged duration of the ventilation–perfusion abnormalities (104), the scan need not be performed at the height of the patient's induced complaints. Others examine the movement of the cords during provocations, produce temporary unilateral vocal cord paralysis to reduce upper airway obstruction, and then evaluate lower airway function (115). Visualization and/or ventilation scans can also be performed in the ED when the need arises.

Acute treatment lies in recognition of the syndrome and its essentially benign nature. When in doubt about the diagnosis, the best course is to treat the patient for asthma and evaluate for laryngeal problems later. The prevalence is unknown.

The critical component of asthma care is a rapid, specific assessment of the patient's illness. It is helpful to have a checklist to follow to make certain that important elements are not missed. It is also essential to monitor progress with objective testing. The paramount historical features are previous hospitalizations, intensive care admissions, hypercapnia, intubation, and assisted ventilation. These are discussed in detail in the next section. Other factors that need to be noted are the kinds of medications used and comorbid conditions (e.g., heart disease, substance abuse, previous barotrauma, and atelectasis) that may predispose the patient to complications. The physical signs of value are listed above. Auscultation of the chest is mandatory to determine the presence, type, and nature of the breath sounds, the quality of aeration, and the symmetry of thoracic movement. Rales are not part of asthma and their presence should alert the caregiver that other diseases might be at the root of the patient's problem. In addition to examining the thorax, evaluation of the head and neck is important to detect conditions such as angioedema, upper airway obstruction, and barotrauma.

Objective measures of airflow limitation are necessary to confirm the presence of bronchial narrowing, evaluate its magnitude, and determine the adequacy of gas exchange. Physicians' subjective estimates of obstruction, including those of chest specialists, are often grossly inaccurate (116) and cannot be relied on. Indices derived from a forced exhalation, such as PEFR or FEV₁, coupled with measures of pulse oximetry are usually adequate (13). PEFR values less than 30% of predicted are particularly noteworthy and, as discussed, airflow limitation of this magnitude is often slow to resolve. The use of arterial saturation to monitor gas exchange rather than arterial blood gases is based on the close correlation between the two. In the absence of oxygen therapy, arterial desaturation and hypercarbia occur concurrently and develop only with severe obstruction or ventilatory depression (9, 23, 101–104).

Arterial saturation values of 90% or more are uncommonly associated with problems. Typically, the higher the Sa_O2, the lower the incidence of unwanted occurrences and the less likely the presence of CO₂ retention (117–119). Only 3 of 72 cases with Sa_O2 values greater than 90% in one investigation had unrecognized respiratory failure (117). On the downside, it is well established that saturation monitoring does not detect alveolar hypoventilation during O₂ breathing and, therefore, a high Sa_O2 can promote a false sense of security in patients with CO₂ retention (120, 121). Fortunately, this event is infrequent and has been reported less than 1% of the time in acute asthma (117–119); nonetheless, the possibility must be kept in mind concerning all patients receiving supplemental O₂. Saturation monitoring can also be misleading in the presence of met- or carboxyhemoglobinemia, nail polish, and increased skin pigmentation (121). When in doubt, it is best to analyze arterial blood gases.

Patients in whom measures of arterial blood gases are necessary are those with a pretreatment Sa_O2 of less than 90%; anyone in whom the Sa_O2 falls during observation; and those whose PEFR does not improve to 40 to 45% of predicted, or worsens, after treatment (13–17). Arterial blood gases should also be obtained in anyone admitted to hospital. Patients with hyper- or normocapnia and persistent respiratory and/or metabolic acidosis early in the course of their episode require follow-up assessments after receiving adequate doses of β₂ agonists (13–17). The presence of any of these elements requires continuous monitoring in an environment that can provide immediate ventilatory support.

The reasons why patients die in hospital of a potentially reversible disease have not been elucidated. Certainly features such as the presence of overwhelming obstruction, comorbid conditions, nosocomial infections, and barotrauma contribute, but they occur relatively infrequently. Other potential explanations include gaps between accepted standards of care and actual performance. Surveys of practice patterns and patient behaviors in fatal cases indicate that failures by caregivers to conduct thorough evaluations and recognize warning signs are arguably the most common causes of adverse outcomes in both ambulatory and hospital settings (122–135). For patients, items of importance include a lack of understanding or misinterpretation of the seriousness of their symptoms, poor compliance with (or refusal of) medical advice, and denial (122–124, 126–133). In addition, analysis of ED management has found patchy adherence to recommended standards. Key elements such as global assessment of the severity of the obstruction, the use of objective measures of pulmonary mechanics, or examination of gas exchange are sometimes ignored (20, 136–138). In a multicenter French study, arterial blood gases were obtained only in 67% of patients deemed to have life-threatening attacks (20). In other investigations, they were drawn just 17% of the time (136). Similar patterns of omissions have also been reported in hospitalized patients (138), but improve when deficiencies are pointed out (139). As long as the need for objective assessments continues to be ignored, patients will remain in needless jeopardy.

It is debatable whether the existence of asthma per se shortens longevity. Earlier studies have concluded not (140), but more recent epidemiologic data suggest that patients with asthma die at a higher rate than contemporaries (141). Examination of this work, however, shows that asthma, per se, rarely killed the participants (23 asthma fatalities in a cohort of 13,540). Instead, the excess deaths occurred from conditions that have no readily apparent relationship to asthma such as malignancies, cardiovascular disease, and accidents.

The overall mortality statistics for asthma for the twentieth century in the United States have been flat (142) and the death rate has hovered at about 2.0 per 100,000 (143). In the last part of the twentieth century, there was an upward trend and deaths rose from 1.3 to 1.9 per 100,000 (144). Although this is a large percentage change, the absolute number is still quite small. The most recent statistics indicate that 5,438 patients with asthma died out of 16 to 17 million individuals at risk (1). By way of comparison, the fatalities from heart disease over the same period were more than two orders of magnitude greater (number of deaths, 729,974; rate, 194.6/100,000) and those for cancer were 99 times higher. About five times more people died from suicide than asthma (145).

As with morbidity, large metropolitan areas with sizeable urban minority populations have driven the mortality statistics in the United States (46–49). African-Americans have a death rate more than 200% higher than that of their white contemporaries (50). The reasons for the greater risk are unclear and issues such as ethnic and racial discrepancies in disease severity, availability of care, educational disparities, cultural dissimilarity in the manner in which the disease is perceived or treatment sought, and differences in pharmacogenetics are under investigation (46–49, 146). Thus far, no clear element has emerged. Poverty, per se, does not appear to be the issue. Increased death rates from asthma are not found in states such as Mississippi, Louisiana, and West Virginia, which have the lowest per capita income (147).

Death from acute episodes is uncommon and is reported in less than 0.1% of patients in large series (13–17, 20, 21, 24). Surveys of fatalities suggest that the vast majority occur suddenly and at places such as home or work and imply that little time, or opportunity, exists to forestall the course (125, 128, 132). This is statistically true but somewhat misleading. Of the patients reported thus far, slightly less than one-half died in hospital (123–128, 130, 131, 133, 135, 148, 149) and in about 85% the duration of the final episode was 12 hours or more, giving more than ample opportunity for treatment (122–124, 127, 128, 130, 132–135).

The quintessential phenotype of rapidly fatal disease is “catastrophic,” “brittle,” or “sudden asphyxic” asthma (38, 150, 151). This is a dramatic but rare happening. One study reported only 9 cases from a total of 1,169 consecutive admissions to hospital (152). In the index series, 10 cases were seen in a decade, but only 5 of them suffered respiratory arrest (38). Affected individuals show rapid decompensations, extreme hypercapnia, combined metabolic and respiratory acidosis, silent chests, and a faster rate of improvement than control subjects. Much of this pattern, save for the speed of onset and recovery, mirrors other patients with life-threatening illness (26, 30, 88, 153, 154). At autopsy, the airways are often devoid of inspissated secretions (155) and contain more neutrophils and eosinophils in the submucosa (150, 156). In one investigation, there was greater infiltration of the bronchioles with eosinophils and mononuclear cells, more lumenal obstruction, and more smooth muscle thickness than in control subjects without a history of respiratory disease (157). Moreover, the adjacent muscular pulmonary arteries did not display the morphologic features of chronic hypoxia. In others, there were inflammatory infiltrates in both proximal and distal lung tissues that, in contrast to stable asthma, contained large numbers of activated lymphocytes and eosinophils (158). These findings strongly suggest a unique entity, but there are relatively few numbers of observations to formulate unequivocal conclusions (58, 156–158).

It is assumed that there are far more “near-fatal” episodes of asthma than deaths, but there are no data as to how many. Here, too, a lack of uniform diagnostic criteria impacts the statistics. Near deaths have been variously defined singularly or in combination as the need for intubation and mechanical ventilation, respiratory arrest, respiratory failure, respiratory acidosis without intubation, two or more hospitalizations despite administration of oral corticosteroids, two or more episodes of pneumothorax or pneumomediastinum associated with status asthmaticus, altered consciousness, a Pa_CO2 of 45 mm Hg, a Pa_CO2 of 50 mm Hg, a pH of less than 7.35, or a pH of less than 7.20 (22, 25, 26, 29, 30, 39, 44, 88, 153, 159–162). Few would quarrel that these occurrences proclaim severe attacks, but it does not necessarily follow that they represent potential fatalities. For instance, hypercarbia and respiratory acidosis occur 10 to 20 times more frequently than the need for ventilatory assistance and do not necessarily imply an ominous prognosis (23, 101–104). Elevated initial values for Pa_CO2 fall and respiratory acidosis dissipates with appropriate therapy without the risk of death or the need for intubation (9, 21, 23, 24, 101–104).

An assessment of the magnitude of the problem of near-fatal asthma is also hampered by a lack of information regarding the size of the denominator in the “at risk ratio.” Considering that most publications provide cases that have been gathered over years (Table 1) (25, 26, 29, 30, 39, 44, 88, 153, 159–162), the denominator may, or may not, be large. In the data in Table 1, an average of 0.32% of MICU patients annually were in a high-risk category as judged by the need for intubation and ventilatory support. In 2 reports involving 2,655 admissions for asthma seen in a 20-year period, 135 patients (5.1% of the total; mean, 0.25%/year) would be considered as having “near-death” experiences (21, 24). Of these, only 48 of 2,655 or 1.8% were intubated. Hence, just 0.09% of patients with asthma presenting per year in those institutions might have died (21, 24). To reach firm conclusions about the prevalence of near-death experiences, prospective multicenter investigations using well-defined diagnostic criteria are needed.

The identification of individuals at high risk for a fatal attack of asthma is a challenging problem, particularly when it is a first-time occurrence. Statistically, such patients are likely to have repeated hospitalizations, multiple emergency room visits, and previous intubations and/or ventilatory assistance (22, 24, 26, 29, 44, 124, 129, 132, 162–167). A history of hypercapnia with one episode is particularly important because it tends to recur with others (24, 25, 30, 39, 154). They also have an increased prevalence of poorly controlled disease, take more medicine, and/or have more nocturnal symptoms than do case–control subjects (22, 24, 25, 29, 44, 88, 153, 162). In hospital, such individuals may show unstable airways with large diurnal swings in lung function (152). Often, PEFR measures in the morning are considerably smaller than in the evening. It is still debatable whether they have adequate access to health care, increased airway responsiveness or an abnormality in their ability to perceive dyspnea (22, 24, 25, 88, 113, 118, 153, 161, 162). They do not appear to be different in terms of the extent of pulmonary function defects (29, 88, 153).

Despite continuing investigative effort, the risk factors predicting fatal asthma have not yet been identified from the standpoint of specificity and selectivity. They occur too frequently in the general asthmatic population and not often enough in those who are potentially in peril. In a large cross-sectional survey of patients with asthma seeking acute care in the Cleveland area, 20% gave histories of two or three visits to the ED per year, 8% had one or more hospitalizations, 3% had one or more admissions to the MICU, and 1% had been intubated previously (Figure 4)

Figure 4. Frequency of occurrence in “high-risk” indices in acute asthma. (A) Data from a cross-sectional survey of multiple hospitals in the Cleveland area; (B) data from a multicenter trial in France (20).

Download Figure | Download Powerpoint

[More] [Minimize]

A number of potentially important genes in asthma have been identified such as those regulating cytokines, transcription factors, receptors, lipid mediator generation, and immune processing (168). Preliminary data suggest that polymorphisms in the IL-4 gene may contribute to low lung function (169) or may be risk factors for life-threatening disease (170). Confirmation has not yet been forthcoming. Other observations suggest that patients harboring a mutant genotype for the production of the leukotrienes LTB₄ through LTE₄ might not respond to antileukotriene treatment (171). The prospect of the existence of genetically dysfunctional β₂-adrenergic receptors contributing to asthma morbidity and mortality has also been investigated (172–175). People taking albuterol on a regular basis and who were homozygous for arginine (Arg) at codon 16 had a small deterioration in PEFR. When this same population took the medication on an as-needed basis, the effect disappeared (173). It should be noted, however, that there was no change in FEV₁ under any circumstances, making the observations difficult to interpret. A corollary study from this data pool found that the Arg/Arg allele might confer a risk for more severe asthma attacks (174). Again, the effect was small and more data are needed before rigorous conclusions can be reached. In other investigations, no genotypic differences have been found between β₂-adrenergic receptor haplotypes and polymorphisms of the tumor necrosis factor-α or angiotensin-converting enzyme genes in patients with fatal and nonfatal asthma (175).

Ventilatory assistance can be lifesaving. Both noninvasive and invasive techniques are available (154, 283). Noninvasive facemask ventilation may offer short-term support for some subjects with hypercapnic respiratory failure who can cooperate with their care and are able to protect their airways (284). Its applicability, however, is limited by its poor patient acceptance. One study reported that 3 hours of bilevel positive airway pressure added to aerosolized albuterol and IP in the ED improved lung function and decreased admissions relative to control (285). However, the patients were not in respiratory failure and neither the pharmacologic regimens, nor the discharge decisions, were rigorously governed by protocol.

As indicated in Table 1, about 30% of patients are believed to require invasive ventilation, but the range varies widely. The generally accepted indications are progressive CO₂ retention, obtundation, and impending cardiopulmonary collapse (154, 283). The mere presence of hypercapnia is not sufficient (154). The goal of ventilatory support is to maintain adequate gas exchange until bronchodilators and CSs relieve the airflow obstruction. This usually entails sedation, and possibly paralysis, as well as strategies to minimize dynamic hyperinflation (154, 283). Sedation is accomplished with benzodiazepines combined with opioids or propofol (286–288). Even though ketamine has bronchodilator properties, its psychotropic and sympathomimetic effects limit its use (289, 290). It may be necessary to supplement sedation with neuromuscular blockade with pancuronium, vicuronium, atracurium, or cisatracurium (154, 283). The first two are eliminated through the liver and kidneys; thus, their pharmacokinetics change with impairment in these organs (154). This is not a problem with the others. All of the paralytics can be associated with myopathy, which is worsened by concomitant use of CSs and aminoglycoside antibiotics (283, 291–293). In one study, 76% of the patients receiving a paralytic and CS had elevations in creatinine phosphokinase levels and 36% had overt myopathy (293). The muscle weakness is not seen with steroids alone and the risk increases with the duration of paralysis (293).

As pointed out earlier, dynamic hyperinflation (auto-PEEP) has profound physiological effects. It rises directly with minute ventilation and can compromise cardiac output by reducing venous return (99, 109, 112). The institution of positive-pressure ventilation in an already hyperinflated thorax can markedly worsen hemodynamics and cause abrupt falls in blood pressure including cardiac collapse. Because the airways are heterogeneously narrowed, the less involved parts of the lungs may undergo regional overextension when exposed to high inflation pressures and rupture (283, 294). For these reasons, ventilatory strategies that provide the longest possible expiratory time are desired so that dynamic lung inflation is minimized. This goal is accomplished by combining the smallest tidal volume with the slowest inspiratory rate and fastest inspiratory time to keep a static end-expiratory pressure (plateau pressure) of less than 30 cm H₂O (295). Levels in this range are uncommonly associated with unwanted effects. High inspiratory flows increase peak pressure and can theoretically raise the risk of barotrauma (154, 296). Some studies, however, have not found a relationship between such pressures and complications (296). Specific details on ventilatory procedures and strategies can be found in reviews (154, 297).

Approaches designed to reduce auto-PEEP often result in hypoventilation (35). The resulting hypercapnia is well tolerated as long as it develops slowly and the Pa_CO2 remains at 90 mm Hg or less (35, 297, 298). When necessary, the pH can be defended pharmacologically (35, 298).

Despite concerns about increasing mortality, most patients survive acute episodes (Table 6)