This study was approved by the Pittsburgh VA subcommittee for the protection of human subjects, and all volunteers provided written informed consent before participation.

Participants with COPD were recruited from the home oxygen program at the VA Pittsburgh University Drive location. Spirometry was completed in the preceding 30 days before data collection to confirm moderate or severe COPD. Moderate COPD was defined FEV₁/FVC <70% and FEV₁ 64 to 50% predicted. Severe COPD was defined as FEV₁/FVC <70% and FEV₁ <50% predicted (19). Exclusion criteria for the COPD group was the presence of an indwelling tracheostomy tube and/or naso-gastric feeding tube, history of stroke or Parkinson's disease, active pulmonary infection, fever, exacerbated COPD, cardiac ischemia or congestive heart failure, and severe metabolic derangements. A total of 25 male patients with COPD with a mean age of 69 years (SD, 8; range, 52–87 yr) participated. Nineteen out of twenty-five participants were classified as severe COPD with a mean FEV₁ of 35% predicted (range, 14–49% predicted), and 6 out of 25 were classified as moderate COPD (mean FEV₁, 59%).

Family members of the patients with COPD, staff, and visitors to the medical center were recruited for the healthy control group. Control subjects had no history of oropharyngeal dysphagia or complaints about swallowing function, no history of neurologic disease, no history of oral or pharyngeal cancer, and no history of COPD. Spirometric testing was completed to verify a FEV₁/FVC ratio greater than 70% predicted. All potential participants completed a screening questionnaire that consisted of questions that are routinely used to identify dysphagic symptoms and risk factors for dysphagia during clinical evaluation. Any verbal or written description of dysphagic symptoms provided by the potential participant, such as coughing, choking, or the feeling of food sticking in their throat, or a positive response to any of the questions on any portion of the questionnaire excluded the participant from the study. A total of 25 healthy control subjects with a mean age of 64 years (SD 9, range 51–81) participated. There were 12 male and 13 female subjects.

For both groups, the use of medications at the time of the study that may have an effect on swallowing function, such as psychotropic drugs and barbiturates, excluded potential subjects from participation. To rule out dementia, a delayed/immediate recall ratio < 95% on the Story Retelling-Immediate/Story Retelling-Delayed subtests of the Arizona Battery for Communicative Disorders of Dementia excluded enrollment (20).

The KayPentax Swallowing Station and Swallowing Signals Lab (KayPentax, Lincoln, NJ) enabled the acquisition and recording of simultaneous respiratory and surface electromyographic (SEMG) measurements. Two auxiliary ports enabled respiratory signals to be displayed from an uncalibrated respiratory inductance plethysmograph (Respitrace; Ambulatory Monitoring Inc., Ardsley, NY) that measured changes in cross-sectional area of the rib cage and abdomen, allowing for the determination of direction of motion (inhalation or exhalation). The software takes the combined movement of chest and abdomen and calculates a summed output that is reflective of changes in Vt. Chest motion was displayed in one port, and a summed output of chest and abdominal motion (Vt) was displayed in the second port. Combined sum was used during data analysis to assure that the chest and nasal signals were pulmonary related and not artifact. Respiratory data were also obtained from the Signals Lab's calibrated nasal cannula (Micro Switch AWM200 Microbridge Mass Airflow sensor; Honeywell, Morristown, NJ) that displayed the direction, level, and duration of airflow. For the SEMG, two Ag/AgCl electrodes and a reference electrode contained in a single cohesive housing designed to detect the occurrence of the swallow were affixed under the patient's chin behind the mental symphysis according to the manufacturer's instructions. This muscle group includes the mylohyoid, anterior belly of the digastrics, and geniohyoid (21). The electrode placement and muscle group have been shown to be valid and reliable indicators of the pharyngeal swallow (22). The swallowing station software provided a rectified, integrated, and smoothed EMG signal so that a distinctive peak at the height of laryngeal elevation occurring during each swallow was clearly displayed and time-locked to the other respiratory signals. The occurrence of each swallow was determined by the simultaneous appearance of deglutitive apnea where the flow at the nasal cannula ceases and a black, flat line is visualized and electromyographically using SEMG. Figures 1 and 2 show examples of the raw data.

Figure 1. Raw data showing inhalation before the swallow.

Download Figure | Download Powerpoint

[More] [Minimize]

Figure 2. Raw data showing postswallow inhalation.

Download Figure | Download Powerpoint

[More] [Minimize]

Data were recorded while participants randomly and spontaneously swallowed a total of nine 2.5-g solid (vanilla wafer cookie) portions that were placed on a plate and 10 5-ml semi-solid (vanilla pudding) boluses that were measured out onto individual teaspoons. To simulate as natural an eating environment as possible, each subject was seated comfortably at a table as they self-fed without cueing.

Food boluses that were not swallowed entirely on the first swallow were not analyzed because bolus volume was controlled in this experiment. Any discrepancy between nasal and plethysmography signals, such as those that can occur when the nasal cannula becomes occluded and the signal lost or when the respitrace signal exhibited electrical artifact, were excluded from the final analysis. Both groups swallowed a total of 250 pudding swallows, from which 230 of the control group (92%) and 238 (95%) from the COPD group were acceptable for analysis. From a total of 225 cookie swallows, 214 (95%) were suitable for analysis from the healthy control group and 219 (97%) from the COPD group.

Each swallow was noted by the simultaneous appearance of deglutitive apnea where nasal airflow ceases while the airway is closed and a peak in the SEMG signal when swallowing muscles are activated. These combined signals were also used to distinguish periods of breath holding and mouth breathing from deglutitive apnea because only bolus swallows gave a time-linked apnea/SEMG signal. Respiratory phase in which the swallow occurred and postswallow airflow direction (negative or inhalation vs. positive or exhalation) were determined using the combined nasal airflow and plethysmographic signals. When mouth breathing resulted in a temporary loss of nasal airflow, the signal from plethysmography was used to determine respiratory phase so long as there had been good correlation between nasal and plethysmography traces.

Swallows that occur at or near end-exhalation or the transition between exhalation and inhalation are associated with lower lung volumes. Swallows that occur at the transition between inhalation and exhalation or near end-inhalation are associated with higher Vt. Because lung volume at the time of the swallow is directly related to the amount of subglottic pressure generated during the swallow (12), the frequency of high versus low Vt was examined. To estimate Vt at the time of the swallow (high vs. low), the duration of the exhalation or inhalation portion of the respiratory phase was determined in milliseconds using the Swallowing Signals Lab software. This value was divided into four quadrants. The first and last quadrants were used to classify swallows into early or late inhalation or exhalation.

The duration of deglutitive apnea (DDA) was defined as the length of time that the nasal signal returned to baseline (zero flow), indicating central inhibition of respiratory musculature and airway closure associated with swallowing (23). The brief negative, nonrespiratory deflection that commonly appears as the pharynx begins to open postswallow was not included in DDA.

Blinding procedures were used for all measurements. Twenty percent of the data was randomly selected for reliability purposes. Intrarater agreement for dichotomous determinations (respiratory phase) ranged from a Kappa value of 0.685 to 0.739. Interrater reliability ranged from 0.648 to 0.771. For the continuous variable (DDA) intraclass correlation was 0.80 for intrarater and 0.6 for interrater measurements. Logistic regression was used to analyze breathing–swallowing data. Apnea data were not normally distributed; therefore, the nonparametric generalized linear model was used for the analysis. Statistical significance was set at P < 0.05 before data analysis. All analyses were performed with SAS version 8 software (SAS Institute Inc., Cary, NC).

Logistic regression analysis showed that participants with COPD swallowed the solid bolus (cookie) during inhalation significantly more often than normal subjects (P = 0.002). When comparing food types within each group (cookie vs. pudding), the percentage of swallows that occurred during inhalation was not significantly different in the COPD group (P = 0.277). However, healthy subjects swallowed significantly less often during inhalation when eating the cookie (P = 0.001). Table 1 is a statistical table, and Figure 3 is a graphic representation of percentages.

Figure 3. Graphic display of the percentage of swallows for each consistency that occurred during inhalation. Black bars, control subjects; gray bars, patients with chronic obstructive pulmonary disease.

Download Figure | Download Powerpoint

[More] [Minimize]

Postswallow inhalation occurred significantly more often in the COPD group with the semi-solid (pudding) when compared with control subjects (P < 0.001). Figure 4 shows a graphic representation of the percentages. Within the COPD group, postswallow inhalation also occurred significantly more often with pudding when compared with cookie (P = 0.001). There was no difference in the rate of postswallow inhalation when pudding and cookie were compared within the control group (P = 0.457). Table 2 displays the statistical table.

Figure 4. Graphic display of the percentage of swallows for each consistency that were followed by inhalation (postswallow inhalation). Black bars, control subjects; gray bars, patients with chronic obstructive pulmonary disease.

Download Figure | Download Powerpoint

[More] [Minimize]

Between-group comparisons found that the normal and COPD participants swallowed the cookie at high Vt with the same frequency (P = 0.36). However, subjects with COPD swallowed pudding at low Vt significantly more often than the normal control subjects (P < 0.001). Within-group comparison of food types revealed that the subjects with COPD also swallowed pudding at low Vt significantly more often than cookie (P = 0.006). Conversely, the control subjects swallowed cookie at low Vt significantly more often than pudding (P = 0.034). Table 3 is the statistical table, and Figure 5 shows a graphic representation of between group comparisons. Correspondingly, there were significant interactions between food and group for swallows occurring during inhalation (P = 0.030), for postswallow inhalation (P = 0.007), and for low Vt swallows (P = 0.001).

Figure 5. Graphic display of the percentages of swallows for each consistency that occurred at or near end-exhalation or the transition between exhalation and inhalation (low Vt). Black bars, control subjects; gray bars, patients with chronic obstructive pulmonary disease.

Download Figure | Download Powerpoint

[More] [Minimize]

There was no difference in DDA between the groups for either consistency. In addition, there were no significant differences between the consistencies within the control group (P = 0.589). However, the DDA for pudding in the COPD group was significantly shorter than the DDA for cookie (P = 0.03).

Because swallowing during inhalation and postswallow inhalation are thought to be risk factors for aspiration, within-group comparisons of DDA between inhalation versus exhalation were made for each food type.

When the food types were combined, DDA was longer in both groups for swallows that occurred during the inhalation phase. The median DDA for control subjects was 670 milliseconds during exhalation and 801 milliseconds during inhalation (P = 0.025). The median DDA in the COPD group was 624 milliseconds during exhalation and 736 milliseconds during inhalation (P = 0.001). In the control group, the DDA for pudding swallows that occurred during inhalation were significantly longer than those that occurred during exhalation (P = 0.018). There was no difference in DDA between inhalation and exhalation for cookie (P = 0.868). In the COPD group, the DDA of swallows that occurred during the inhalation phase were significantly longer for both food types (P = 0.011 for cookie and 0.002 for pudding). Table 4 is a statistical table of these data.

In the control group, the DDA was not significantly different for cookie or pudding swallows when inhalation followed (P = 0.244 for cookie and P = 0.276 for pudding). However, in the COPD group, the DDA of pudding swallows that were followed by inhalation was significantly shorter in duration (P = 0.0411).

Additional evidence linking pulmonary compromise with disruptions in breathing and swallowing coordination can be found in the DDA values. Deglutitive apnea occurs involuntarily during each swallow when the respiratory muscles are centrally inhibited and the airway closes. DDA is most often cited to range from approximately 0.5 to 1.0 seconds (8, 10, 26, 28, 43–45). The overall mean apnea that we measured in both groups was consistent with other reports (654.00 ms in the control group and 641.00 ms in the COPD group). DDA has also been examined as a function of age, bolus volume, and gender; however, our study is the first to examine DDA in the context of respiratory phase relationships (inhalation vs. exhalation). Participants with COPD showed longer DDA for both materials when swallows occurred during inhalation, whereas the control subjects prolonged DDA with the pudding only. Prolongation of DDA could indicate the presence of a compensatory mechanism, such as providing more time for recoil forces to generate higher DPsub or more time for the bolus to traverse the pharynx and enter the esophagus. Spontaneous swallowing compensations occurring in patients with COPD were previously reported by Mokhlesi and colleagues (46).

There was no effect of DDA in the control subjects on the postswallow respiratory phase; however, a shorter DDA was measured in the subjects with COPD for pudding. This finding, when combined with the increased frequency with which these subjects swallowed at low Vt and inhaled after swallowing pudding, seems to suggest that ventilatory drive or “air hunger” can take precedence over deglutition when necessary. The increased frequency of postswallow inhalation of saliva during exacerbation reported by Shaker and colleagues demonstrates this point.

We set our inclusion/exclusion criteria so that we would capture subjects who were homogenous in terms of clinical severity. However, a limitation of our study is that we are unable to determine if disrupted breathing and swallowing patterns vary as a function of the severity of obstructive disease. A larger study that stratifies patients with COPD by severity level could address this issue. Another limitation of our study is that swallowing function was not measured. Additional study of the potential physiologic swallowing alterations that may occur at different points within the respiratory cycle or at known lung volumes may further explain the interplay of respiratory and deglutitive functions.

The study of swallowing and breathing relationships is important because oropharyngeal dysphagia is an underdiagnosed comorbidity in patients with COPD (47–49). Indeed, the presence of COPD was shown to be the most significant risk factor for aspiration pneumonia in nursing home patients (50). The direct observation of swallowing physiology in patients with COPD under fluoroscopy has revealed findings that are consistent with dysphagia, such as delayed pharyngeal response, decreased base of tongue retraction, and reduced laryngeal elevation (46). Suggested causes for oropharyngeal dysphagia in COPD are muscle fatigue of the upper aerodigestive tract, alterations in airway protective mechanisms, and anxiousness (16). A lack of reciprocity between respiration and deglutition has been suggested as one of the mechanisms that can induce swallowing impairments in patients with COPD (47, 48).

In summary, when compared with healthy control subjects, patients with COPD were more likely to swallow a bolus that required mastication during inhalation and were more likely to swallow a semisolid at a low Vt and with a shortened DDA and subsequent inhalation after the swallow. The observed impaired breathing and swallowing patterns in the patients with COPD suggests a possible explanation for the presence of dysphagia in persons who do not have neurologic illness. Unrecognized prandial aspiration before or during COPD exacerbation may also contribute to the severity of the exacerbations.