INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The prevalence of symptomatic gastroesophageal reflux (GER) is greater in asthmatic individuals than in other patient populations (1). Abnormal GER as reflected by ambulatory pH monitoring criteria has been reported in over 80% of asthmatic individuals (2). Some asthma patients experience reflux-associated respiratory symptoms (RARS), including dyspnea, wheezing, and cough (1, 3). These are often severe enough to warrant additional -agonist use (1, 3). Moreover, Irwin and coworkers found that treating GER was the most important intervention for improving patients with difficult-to-control asthma who were referred to their clinic (4).

The strong association between GER and asthma suggests a causal relationship between the two conditions. Most investigators have assumed that the relationship is due to GER triggering bronchospasm. Proposed mechanisms include microaspiration or a vagally mediated esophagobronchial reflex (5). However, despite causing respiratory symptoms, both spontaneous GER and acid perfusion (AP) of the esophagus have only minimal (6, 7) or no effect on either lung function or bronchial reactivity in asthmatic individuals (8). The importance of GER in asthma severity has been questioned (9).

The problem, then, is how best to explain the paradox of GER worsening asthma symptoms without a clinically important change in lung function. Studies of respiratory sensation in other clinical conditions suggest that respiratory effort is its most important determinant (10). In some conditions, increased ventilation causes breathlessness in the absence of changes in lung function (11). The purpose of the present study was to determine whether AP could alter ventilation and respiratory sensation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The subjects were nonasthmatic patients with normal lung function, who were referred for motility testing as part of an investigation for esophageal disease. They were willing to give informed written consent for their participation after the experimental protocol was explained to them. Instructions to the participants were as complete as possible, but specific details about the Bernstein test were withheld to avoid biasing their responses. Because anxiety can contribute to breathlessness, the patients were reassured and made as comfortable as possible prior to beginning the study. The study was approved by the Conjoint Ethics Committee of the University of Calgary and Foothills Hospital.

Ventilation was measured with a No. 2 Fleisch pneumotachometer (Gould Systems, Cleveland, OH) and a CD15 demodulator ± 2 cm H₂O pressure transducer (Validyne, Northridge, CA). Flow was integrated with a model 8815A Hewlett Packard integrator (Waltham, MA). The patients breathed through a mouthpiece attached to the pneumotachometer, and wore a noseclip. Rib cage and abdominal anteroposterior diameters were measured with pairs of magnetometers (GMG Scientific, Burlington, MA) placed at the levels of the middle of the body of the sternum and just above the umbilicus, respectively. The magnetometers were calibrated for volume by the isovolume method (12).

Gastric (Pga) and esophageal (Pes) pressures were measured with a four-channel, water-filled catheter (Mui Scientific, Mississauga, Ontario, Canada). The most distal port was 1 cm from the end of the catheter. The other three ports were 5 cm from each other, and 6, 11, and 16 cm from the distal end of the catheter, respectively. The proximal ends of three of the channels were attached to pressure transducers (Pressure/perfusion motility system TDS-4000; Sandhill, Littleston, CO). The most proximal channel was used to perfuse the esophagus with either normal saline or 0.1 N HCl.

Diaphragm electromyographic (EMG) signals were measured both with an esophageal catheter (13) and with surface electrodes. The surface electrodes were attached on the right side in the seventh, eighth, or ninth intercostal interspaces between the mid- and posterior axillary lines. The signals were contaminated with expiratory muscle activity with more anterior placement of the surface electrodes. The signals were amplified, bandpass-filtered between 100 and 2,000 Hz, rectified, and integrated, with a 100-ms time constant (1500 system with 15C01 EMG amplifiers; DISA, Denmark). All signals were recorded on an eight-channel paper recorder (7758B recording system; Hewlett-Packard Inc., Waltham, MA). The Borg scale was used as the psychophysical measure of breathlessness (14). Patients also used it to describe the severity of their throat, swallowing, chest, and stomach discomfort during each perfusion period.

ExperimentaI ProtocoI

The motility studies were done early in the morning, after an overnight fast. Patients receiving prokinetic agents, omeprazole, or H₂- receptor blockers were asked to discontinue these medications at least one 1 wk before the study began. Antacids were withheld on the day of the study. Gastrointestinal and respiratory functional inquires were conducted prior to the study. Patients were weighed, measured, and underwent spirometry.

The magnetometry discs and surface EMG electrodes were then attached. The four-channel pressure catheter was then inserted, followed by the esophageal EMG catheter. These catheters were inserted transnasally, with the aid of lubricant but without topical anesthetic. The patients were then placed in the supine position and remained in that position through the motility and Bernstein tests (15). The experimental setup is shown in Figure 1.

View larger version (37K): [in this window] [in a new window]   Figure 1.   Experimental setup. Diagram shows leads and catheters in place. Saline and 0.1 N HCl were infused into the lower esophagus through the motility catheter. Two of the other ports in the motility catheter were used to measure esophageal and gastric pressure. Abbreviations: rc = rib cage magnetometer coil; ab = abdominal magnetometer coil; Pes, Pga = locations of ports in motility catheter used for measurement of esophageal and gastric pressures, respectively; Es = surface EMG electrode; Edi = esophageal EMG catheter electrode.

The motility study took 30 to 45 min. Once it was completed, the pressure catheter was placed across the lower esophageal sphincter (LES), so that the distal port recorded gastric pressure and the third port (11 cm from the catheter tip) recorded Pes. The most proximal port of the pressure catheter was used to perfuse the esophagus with normal saline or 0.1 N HCl at a rate of 5 ml/min. The esophageal EMG catheter was then positioned to optimize the inspiratory diaphragm EMG signal. This position was usually within 5 cm of the midpoint of the LES. The motility study allowed accurate determination of the location of the LES. The signals were then recalibrated and the isovolume maneuver repeated.

To provide a stable baseline, the patients breathed through the mouthpiece attached to the pneumotachometer, with the noseclip in place, for a minimum of 5 min before the beginning of the motility study. In sequence, saline, acid, and saline were perfused for 5-min periods each. The patients were neither told when the infusions were begun nor which solution they were receiving. The signals were recorded continuously. At the end of each infusion period, the patients were asked to rate their throat, swallowing, chest, and stomach discomfort, as well as respiratory sensation with the Borg scale. The motility nurse questioned the patients separately about their symptoms during acid perfusion of the esophagus, and determined whether the Bernstein test was positive (BP) or negative (BN) according to whether acid reproduced their typical heartburn pain.

Statistics

Two-way analysis of variance (ANOVA) was used to compare the results of different continuous variables in the three infusion periods. As a follow-up to the two-way ANOVA, Tukey's studentized range test was used to compare the different parameters between periods. The relationship between changes in symptoms and E was explored with linear regression analysis. Two-sample t tests were used to compare changes from baseline in the BP versus the BN patients. Data were expressed as mean ± SD. The minimum level of statistical significance accepted was p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

All but one of the 25 patients in the study were referred for evaluation of suspected esophageal disease. Presenting symptoms included dysphagia, atypical chest pain, poorly controlled pyrosis despite intensive medical antireflux therapy, or motility testing was ordered as part of the preoperative evaluation for esophageal surgery. One patient was referred for evaluation of intractable chronic cough. This patient had normal pulmonary function and methacholine challenge tests. The characteristics of the study patients are given in Table 1. The BP and BN groups were similar with regard to age, gender ratio, and body mass index (BMI). Spirometry gave normal results in both groups.

Respiratory Symptom and Discomfort Scores

Eleven of the Bernstein tests were negative and 14 were positive. The numeric scores of throat, swallowing, chest, and stomach discomfort ratings and respiratory symptoms are shown in Table 2. The changes in discomfort ratings during acid perfusion (AP) were all greater in the BP group, but the difference only reached statistical significance for chest discomfort (D). The patients felt that throat discomfort was primarily related to the presence of the catheters. Most of the symptomatic patients described the other discomforts as pressure or burning sensations resembling their heartburn pains. The average change in breathing sensation during AP was greater in the BP patients, but the difference was not statistically significant (Table 2). Most of the symptomatic patients described the sensation as an awareness of their breathing being greater but not unpleasant. The two patients whose ventilation increased to the greatest extent described the unpleasant sensation of shortness of breath or an inability to get enough air.

Ventilation

The E of the entire group was greater during AP than at baseline (Table 3). The increase in E was due to an increase in respiratory rate (RR). Neither tidal volume (VT) nor its rib cage (Vrc) or abdominal (Vab) components changed during AP (Table 3). Comparison of the BP with the BN patients demonstrated that the change in E was greater in the BP patients (7.0 ± 1.6 to 9.4 ± 2.6 L/min, versus 7.2 ± 1.5 to 7.4 ± 1.5 L/min, p = 0.002) (Figure 2).

View larger version (32K): [in this window] [in a new window]   Figure 2.   V· E of individual patients during saline and acid perfusion periods. Patients are grouped according to whether they had a positive or negative Bernstein test result (see text). V· E for the Bernstein-positive patients was 7.0 ± 1.6, 9.4 ± 2.6, and 7.8 ± 2.5 L/ min in the first saline perfusion, acid perfusion, and second perfusion periods, respectively. The V· E for the Bernstein-negative patients was 7.2 ± 1.5, 7.4 ± 1.5, and 6.6 ± 1.2 L/min in the first saline perfusion, acid perfusion, and second saline perfusion periods, respectively.

There was a linear correlation between the increase in E and the change in chest discomfort (D) score during AP (E = 0.7 + 0.8 D; r = 0.67; p < 0.01) (Figure 3). The increased swallowing discomfort also correlated with the change in E (E = 1.1 + 0.8 swallowing discomfort; r = 0.50; p < 0.05). The change in E and the increases in chest and swallowing discomfort correlated with the change in respiratory sensation (B) (B = 0.2 + 0.4 E; r = 0.70; p < 0.001; B = 0.1 + 0.4 D; r = 0.62; p = 0.01; B = 0.3 + 0.5 swallowing discomfort; r = 0.55; p = 0.03).

View larger version (14K): [in this window] [in a new window]   Figure 3.   Relationship between the change in V· E and change in chest discomfort (D) rating during acid perfusion of the esophagus.

Pressure and Electromyographic Signals

Tidal swings in Pes and gastric pressure (Pga) did not change during AP (Table 3). Neither did the ratio of esophageal to gastric tidal pressure swings (Pes/Pga). There was a correlation between the product of Pes and RR and the change in respiratory sensation (r = 0.53, p < 0.01). Neither the surface (Es) nor the esophageal (Edi) electromyographic signal magnitude changed during AP.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

To avoid the controversy surrounding the effects of GER on airway tone in asthmatic individuals, nonasthmatic subjects with normal lung function were chosen for this study. The main finding of the study is that E increases during AP. The change in E was greater in the BP patients, and the severity of chest discomfort correlated with the increase in E. Although the literature on the subject is not extensive, several reports have addressed the relationship between pain and ventilation. Sarton and colleagues (16) found that pain increased E in normal volunteers. Both Bourke (17) and Borgbjerg and associates (18) found that painful stimuli enhanced the ventilatory response to CO₂. Our finding that ventilation increased in response to pain during AP of the esophagus is consistent with these findings (16).

The changes in breathing sensation correlated with the increase in E during AP. The BP patients reported an increase in respiratory sensation even though the changes in E were relatively small. Most of the BP patients described the respiratory sensation during AP as a greater awareness of breathing effort that was not unpleasant. Despite having normal lung function, two patients were breathless during AP. Adams and coworkers determined the change in E required to make normal subjects breathless (19). Even though the magnitude of the ventilation increase was consistent for each individual, they found a wide range between subjects. An increase of only 12 L/min caused breathlessness in some of their normal subjects. Two of the 14 BP patients in our study described the respiratory sensation during AP as breathlessness, and felt that it was unpleasant. Their E doubled during AP, an increase of approximately 6.5 L/min. This increase was proportionately similar to that described by Adams and coworkers since our breathless subjects were relatively small women and most of their patients were average-sized males (19). Our subjects were naive patients, relatively obese, and were studied while supine. Anxiety and the factors mentioned previously may have contributed to their symptoms. Discomfort related to the esophageal catheters and monitoring equipment was similar in the saline and AP periods, and should not have contributed to the differences in respiratory sensation.

Limitations of the Study

Ideally, pH would have been measured continuously to ensure that esophageal pH was neutral during the saline perfusion periods and low during AP. Some subjects may have experienced spontaneous GER during the control saline period. In some, E and discomfort scores were higher during the second saline period than during the first, suggesting the possibility of incomplete esophageal clearing of acid after AP. In others, E was lower in the second saline period, possibly to compensate for the greater ventilation during AP. We decided not to monitor pH, since it would have required a third probe, adding both to the patients' discomfort and the time needed to complete the study. Moreover, pH monitoring would have added substantially to the cost of the study.

Undoubtedly, the mouthpiece and noseclip adversely affected patient comfort and ventilation. Milic-Emili and coworkers showed that wearing a noseclip and breathing through a mouthpiece, even one with a small dead space, increased E (20). However, we felt it important to accurately measure ventilation in this study. The greater E caused by the mouthpiece may have increased the patients' breathlessness scores. However, the effect should have been similar in the acid and saline periods.

Some patients had hiatal hernias, which made placement and maintenance of the esophageal catheters, especially the EMG probe, difficult.

The subjects were patients referred for evaluation of suspected esophageal disease, and had a high prevalence of esophagitis. Patients with esophagitis may be more symptomatic during AP than normal individuals. Their ventilatory response to chest discomfort might not accurately reflect that in normal or asthmatic subjects.

Effect of Gastroesophageal Reflux on Pulmonary Function

Studies of the effects of GER on asthma present an interesting paradox. Despite the strong association between the two conditions, GER has not been shown to worsen lung function or bronchial reactivity consistently (8). Either GER affects only some asthmatics (21), or its effects are minor and unlikely to be clinically significant (7). The small reductions in peak expiratory flow rate (PEFR) reported during AP can be explained by other mechanisms than a change in lung function or airway tone (7, 22). The PEF maneuver is effort dependent, and chest discomfort during AP may limit the ability of a patient to make a maximal effort (23). This may account for some of the conflicting reports on the effects of GER in the literature. Despite the inability to consistently demonstrate that spontaneous GER or AP increases airway tone, asthmatic individuals complain of RARS. Cough, dyspnea, wheezing, and -agonist use occur in asthmatic individuals in association with GER (1, 3). The present study demonstrates how GER may cause dyspnea without worsening airway obstruction. Increased ventilation during episodes of GER may cause even more breathlessness in patients with airway obstruction. The correlation between the severity of GER and asthma symptoms reported by Goodall and colleagues (6) is consistent with our findings.

Two other groups have reported dyspnea caused by GER in patients with normal lung function and bronchial reactivity. Both Pratter and Depaso showed that GER was the cause of otherwise unexplained dyspnea in patients referred to their clinics for evaluation (24, 25). Ambulatory pH monitoring confirmed the temporal association between GER and dyspnea (25). Moreover, dyspnea resolved with successful antireflux therapy (25).

Diaphragm Function During AP of the Esophagus

The diaphragm contributes to maintenance of LES tone, and relaxation of the crural fibers facilitates the entry of a food bolus into the stomach (26). Stimulation of mechanoreceptors by esophageal distension causes reflex relaxation of the crural diaphragm (27). It also causes diaphragm inhibition in humans (28). Acid has been reported to stimulate these esophageal mechanoreceptors (29). Therefore, it is reasonable to speculate that AP of the esophagus might cause diaphragm inhibition and, in turn, breathlessness. However, neither the Edi nor Es decreased during AP in the present study. Moreover, neither Vrc/Vab nor Pes/Pga changed during AP, suggesting that clinically significant diaphragm inhibition did not occur.

Reflux as a Cause of Breathlessness in Other ClinicaI Conditions

This study raises questions about the etiology of dyspnea in other clinical conditions associated with GER. Although there is not as much data as with asthma, the reported prevalence of GER and hiatal hernia in other respiratory conditions, including pulmonary fibrosis and cystic fibrosis, is high (30, 31). GER may contribute to breathlessness in these conditions. Patients with heart disease take a variety of medications that are smooth-muscle relaxants. By facilitating GER, these medications may contribute to dyspnea in cardiac patients. Reflux-induced changes in E may be responsible for unexplained dyspnea in otherwise healthy patients (24, 25).

In summary, AP of the esophagus increases E and respiratory sensation. The increase in E correlates with the severity of chest discomfort during AP. There was no evidence in the present study that AP inhibits diaphragm function. Discomfort-induced changes in E offer a logical explanation for the paradox of GER causing asthma symptoms without changing lung function, and explain how GER may cause dyspnea in patients with normal lung function and bronchial reactivity (24, 25). One would expect that GER and the associated increase in E would have a greater effect in patients with respiratory disease. In asthmatic individuals dyspnea associated with GER could be explained by increased E even in the absence of worsening bronchospasm. Our results also explain the apparent paradox that medical and surgical antireflux therapy improve asthma symptoms and reduce medication requirements without improving lung function or bronchial reactivity (32).