| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1The John Rankin Laboratory of Pulmonary Medicine, Department of Population Health Sciences, and 2Department of Pediatrics, University of Wisconsin-Madison, Madison, Wisconsin
Submitted 12 January 2005 ; accepted in final form 11 July 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATERIALS AND METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
90% of maximal O2 uptake. Arterial O2 saturation decreased to 
94% during the exercise in 8 of 21 subjects, in large part as a result of a decrease in arterial PO2 (PaO2): from 93.0 ± 7.7 to 79.7 ± 4.0 Torr. A widened alveolar-to-arterial PO2 difference and the magnitude of the ventilatory response contributed approximately equally to the decrease in PaO2 during exercise. Airflow limitation and airway inflammation at baseline did not correlate with exercise gas exchange, but an exercise-induced increase in sputum histamine levels correlated with exercise PaO2 (negatively) and alveolar-to-arterial PO2 difference (positively). Mean pulmonary resistance was high during exercise (3.4 ± 1.2 cmH2O·l–1·s) and did not increase throughout exercise. Expiratory flow limitation occurred in 19 of 21 subjects, averaging 43 ± 35% of tidal volume near end exercise, and end-expiratory lung volume rose progressively to 0.25 ± 0.47 liter greater than resting end-expiratory lung volume at exhaustion. These mechanical constraints to ventilation contributed to a heterogeneous and frequently insufficient ventilatory response; arterial PCO2 was 30–47 Torr at end exercise. Thus pulmonary gas exchange is impaired during high-intensity exercise in a significant number of habitually active asthmatic subjects because of high airway resistance and, possibly, a deleterious effect of exercise-induced airway inflammation on gas exchange efficiency. exercise-induced arterial hypoxemia; airway inflammation; exercise hyperventilation; induced sputum; prolonged exercise
O2)], which were not sufficient to stress the pulmonary system's capabilities for ventilation and gas exchange.
There are several reasons to suspect greater gas exchange disturbance during exercise in asthmatic than in healthy nonasthmatic subjects. It has been shown using the multiple inert gas elimination technique that alveolar ventilation-to-perfusion (A/
) nonuniformity at rest is greater in even mildly to moderately asthmatic subjects than in healthy subjects; this often results in a considerably widened alveolar-to-arterial PO2 difference (A-aDO2) and, although less often, arterial hypoxemia (54). This gas exchange disturbance is thought to be primarily caused by an uneven distribution of alveolar ventilation (A) due to narrowed and obstructed airways (54). The airway narrowing is caused by a variety of factors that arise as a consequence of airway inflammatory mediator release, including bronchial smooth muscle contraction, mucosal edema and thickening of the airway wall, peribronchial fluid accumulation and "cuffing" of the small airways, and luminal mucus and liquid accumulation (9). Additionally, the increased airflow during exercise is thought to stimulate the release of inflammatory mediators (e.g., histamine and leukotrienes) from airway cells during or shortly after exercise in asthmatic subjects (6). Because the bronchial smooth muscle and the vascular effects of these mediators are responsible for the peripheral airway changes that likely contribute to gas exchange disturbance in asthmatic subjects, exercise may cause or exacerbate any impairment in gas exchange. Thus gas exchange during exercise in asthmatic subjects might be impaired because of airway dysfunction at rest or an airway inflammatory response caused by the exercise.
Airway resistance, which is often high in asthmatic subjects, increases the propensity for expiratory flow limitation (EFL) and dynamic hyperinflation during exercise in asthmatic compared with nonasthmatic subjects (14, 34, 52). In healthy subjects, EFL and dynamic hyperinflation have been shown to blunt the very important compensatory hyperventilation of heavy exercise (33, 40, 41), and presumably the same should hold true for asthmatic subjects. Thus high-resistance airways might predispose many asthmatic subjects to a less robust ventilatory response and arterial hypoxemia during exercise. However, bronchodilation, which is known to occur during exercise in asthmatic subjects (14, 39, 52), might help protect against excessive EFL in asthmatic subjects with airflow limitation at rest. The extent of exercise ventilatory constraint will consequently depend on the extent of baseline airflow limitation, the magnitude of exercise-induced bronchodilation, and the requirement for ventilation as dictated by the exercise workload.
The aim of this study was to characterize the gas exchange and breathing mechanics responses to submaximal and high-intensity exercise to exhaustion in habitually active asthmatic subjects. We also sought to determine the relations among airway inflammation, breathing mechanics, and gas exchange during the exercise and hypothesized that significant relations would exist among them.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Subject Selection
Habitually active men and women (aged 18–45 yr) with a known history of asthma or suspected asthma underwent several screening sessions to determine eligibility for participation (see Screening Studies). Potential subjects taking oral or inhaled steroids were excluded from study participation. A total of 94 individuals completed at least one screening study: 27 met the inclusion criteria for participation, and 21 of these subjects completed the entire protocol. The subjects consisted of five recreational triathletes, three recreational endurance runners, one cyclist, one collegiate ultimate Frisbee player, one collegiate football player, one competitive gymnast, one collegiate club hockey player, one equestrian athlete, and seven subjects who exercised for general health.
Pulmonary Function Testing
Lung volumes and forced expiratory flow rates were determined according to American Thoracic Society recommendations (3) with a commercially available system (Jaeger). Functional residual capacity (FRC) was measured in a body plethysmograph, and total lung capacity (TLC) was calculated as the sum of FRC and the inspiratory capacity (IC). Forced oscillation (IOS; model MS-IOS, Jaeger) was performed at a fixed breathing frequency and duty cycle (12 breaths/min, duty cycle = 0.35) and was used to determine total respiratory resistance (Rrs, 5 Hz) and frequency dependence of resistance (Rrs, 5–25 Hz). Exhaled nitric oxide (eNO) was measured as described previously (50). A single-breath, breath-holding technique was used to determine diffusing capacity for carbon monoxide (45). A single-breath N2-washout test was used to calculate closing volume and the slope of phase III during expiration as described previously (48).
Sputum Induction and Processing and Inflammatory Mediator Measurements
Subjects inhaled two actuations of albuterol 15 min before sputum induction. Nebulized 3% hypertonic saline (Ultra-Neb, Devilbiss, Somerset, PA) was inhaled for 5 min with a nose clip in place. Subjects then expectorated lung sputum into a sterile specimen container that was kept on ice throughout the procedure. This protocol was repeated until an acceptable volume of sputum was collected or the forced expiratory volume in 1.0 s (FEV1.0) had fallen by >20%, in which case albuterol was administered immediately. Mucus plugs were manually selected and processed as described previously (2). The same technician processed all sputum samples and completed all total and differential cell counts for each sputum study.
Competitive enzyme immunoassay kits (Cayman Chemical, Ann Arbor, MI) were used to measure histamine in the sputum supernatant and 9
,11
-prostaglandin F2 (PGF2) in the urine.
Methacholine Challenge
Airway reactivity was determined using a five-breath dosimeter protocol with methacholine chloride (Provocholine, Methapharm, Coral Springs, FL) according to American Thoracic Society recommendations (4). After diluent inhalation, doubling doses of methacholine (0.031–16 mg/ml) diluted in saline were inhaled. After each dose, IOS and two forced vital capacity (FVC) maneuvers were performed (in that order), and the testing was concluded when FEV1.0 decreased by 
20% from the postdiluent value or after the maximum dose of 16 mg/ml was achieved.
Arterial Blood Measurements
Samples of arterial blood were drawn anaerobically over 10–20 s at rest and during exercise for measurement of arterial PO2 (PaO2), arterial PCO2 (PaCO2), and pH via a blood-gas analyzer (model ABL505, Radiometer, Copenhagen, Denmark), and oxyhemoglobin saturation (SaO2) was measured with a CO-oximeter (model OSM-3, Radiometer). Arterial blood lactate concentration was measured using an electrochemical analyzer (model 1500 Sport, Yellow Springs Instrument, Yellow Springs, OH).
Exercise Study Apparatus
The apparatus used for the exercise studies has been described previously (57). Room air (laboratory temperature = 24 ± 1.2°C, relative humidity = 47.6 ± 12.4%) was used as the inspirate for the first eight subjects enrolled in the study (Table 1). Compressed dry air from a gas cylinder (0.21 inspired O2 fraction, laboratory temperature = 24 ± 1.3°C) was used as the inspirate for the remaining 13 subjects. For each individual subject, the same inspirate was breathed at all exercise workloads during the screening and arterial blood exercise studies.
|
Before exercise, subjects performed several maximum volitional flow-volume loops (MFVLs) while standing on the treadmill and breathing on the same apparatus used during exercise. EFL was estimated as described previously (41), and end-expiratory lung volume (EELV) during exercise was determined by subtracting the exercise IC volume from TLC. Ventilatory capacity during exercise was estimated using the preexercise MFVL and the spontaneous exercise tidal flow-volume loops as described previously (34). Ventilatory capacity was thus calculated on the basis of the maximal expiratory flow rates achievable (as defined by the MFVL) at the actual operating lung and tidal volumes measured during the exercise. A nasopharyngeal 10-cm latex balloon-tipped catheter (Ackrad Laboratories, Cranford, NJ) connected by polyethylene tubing to a differential pressure transducer (Validyne) was used to measure esophageal pressure. Inspiratory pulmonary resistance (RLi) was calculated at peak inspiratory flow according to the technique of Mead and Whittenberger (21, 42).
Medications
Subjects were instructed to refrain from using short-acting -agonists within 12 h and any long-acting -agonists, antihistamines, leukotriene modifiers, and sodium cromoglycate within 48 h of each study. Table 2 includes a list of medications for each subject enrolled in the study. Additionally, subjects were instructed to refrain from ingesting any food, beverages, or other products containing caffeine for 8 h before all studies.
|
Up to three separate screening studies were completed to determine eligibility for participation. Eligible subjects were required to demonstrate at least one of four inclusion criteria: 1) 12% increase in FEV1.0, from the baseline value, after inhalation of two actuations of albuterol, 2) demonstration of exercise-induced bronchospasm (EIB), defined as 10% decrease in FEV1.0 after an incremental exercise test to exhaustion, 3) 2% eosinophils (as percentage of white blood cells) in the cellular phase of the induced sputum, and 4) 4.0 mg/ml methacholine causing a 20% decrease in FEV1.0 (PC20).
Eligible subjects completed an exercise study designed to evaluate breathing mechanics and gas exchange during submaximal and high-intensity exercise to exhaustion.
Exercise Study Protocol
Subjects voided their bladder immediately on arrival at the laboratory. Fluid ingestion was standardized and included a total of 800 ml of water distributed in four servings (150–250 ml per serving) throughout the study. Preexercise lung function tests, including spirometry, IOS, FRC and TLC, and eNO, were completed. A 20-gauge arterial catheter was inserted percutaneously into a radial artery under local 1% lidocaine anesthesia, and a nasopharyngeal temperature probe (Mallinckrodt Medical, St. Louis, MO) and balloon-tipped catheter were placed intranasally into the lower one-third of the esophagus. After the resting lung function tests and catheter placements, a baseline urine sample was collected. Immediately before the exercise, three MFVL and repeated IC maneuvers were performed while the subjects stood on the treadmill.
Subjects initially exercised at two submaximal workloads for 3 min each (sub-1 and sub-2), and blood samples were collected and IC maneuvers were performed during the final 30 s of each workload. After a brief (3–5 min) rest, exercise was resumed and performed to exhaustion at a constant speed and grade (hereafter referred to as the prolonged exercise bout) that elicited a metabolic rate of 90% maximal O2 (O2 max, total exercise time = 11.2 ± 3.5 min). Arterial blood was collected every 2 min, beginning at minute 1, and also at exhaustion. IC maneuvers were performed at similar time points. At 40 min after completion of the prolonged exercise bout, another exercise test to exhaustion was performed; the details and results from this bout are presented elsewhere (27). A urine sample and induced sputum sample were collected 45 and 60 min after the final exercise test to exhaustion, respectively. The sputum results from the initial screening study were used as the baseline (i.e., preexercise) values with which to compare the postexercise sputum sample.
Data Analyses
Repeated-measures ANOVA containing one within-subject factor (measurement time) and one between-subjects factor (high- or low-SaO2 group; see RESULTS) was used to compare group mean values. If a significant main effect (group or time) or interaction (group x time) was observed, Tukey's post hoc test was used to determine where the significant differences existed. Relations between variables were determined using Pearson's correlation coefficients. Statistical significance was set at P = 0.05. Unless otherwise stated, values are means ± SD.
|
RESULTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Individual results from the screening studies are shown in Table 1. On average, subjects met 2.8 of the 4 inclusion criteria. Overall, EIB was the most prevalent of the four criteria (18 of 21 subjects), whereas 2% eosinophils in the sputum was the least prevalent (11 of 21 subjects). The PC20 criterion and percent change in FEV1.0 after albuterol inhalation were equally prevalent among the subjects. The severity of airway dysfunction and reactivity, as determined from one test, did not always correlate with the severity measured with a different test. Notably, in subjects 6, 17, and 21, EIB was significant but the PC20 was >16.0 mg/ml. On the basis of results from the screening studies, we have classified the asthma in the present group of subjects as mild to moderate (43a).
Subject Characteristics
Age, height, and body mass of the subjects were 25.5 ± 6.0 yr, 172.9 ± 7.7 cm, and 72.8 ± 11.8 kg, respectively. O2 max, pulmonary function, medications, and atopy status for each subject are shown in Table 2. O2 max was on average 119% of the age-predicted value (P < 0.05 vs. predicted). Overall, pulmonary function was slightly but significantly compromised relative to that of healthy age- and height-matched control subjects [P < 0.05 for FEV1.0 and forced expiratory flow at 25–75% of vital capacity (FEF25–75) vs. predicted]. On average, subjects were mildly airflow limited according to the American Thoracic Society criteria for assessing an obstructive abnormality (3). There were exceptions, however: in subjects 3, 5, and 10, FEV1.0/FVC was <65% and FEF25–75 was 29–53% of predicted. TLC (6.58 ± 1.02 liters) and FRC (3.08 ± 0.49 liters) were 102 and 100% of predicted normal (3), respectively, and diffusing capacity of the lung for carbon monoxide (33.4 ± 7.1 ml·min–1·Torr–1) averaged 88% of the predicted value [P < 0.05 vs. predicted (38)]. Mean phase III slope from the single-breath N2 washout averaged 1.19 ± 0.50% N2/l (range 0.61–2.35% N2/l), and slopes were >2.0 SD above the normal range (1.97 ± 0.26% N2/l) in 4 of 21 subjects (49). eNO was elevated (52.4 ± 35.6 ppb) compared with previous data in normal subjects, in which eNO measured using the same methodology was 25–35 ppb (50, 57).
Exercise Responses
Exercise-induced oxyhemoglobin desaturation.
Subjects were divided into two groups on the basis of nadir SaO2 during the prolonged bout (15): those that maintained >94% SaO2 [mean ± SD, 95.6 ± 1.0% (Hi-SaO2), n = 13] and those in which SaO2 decreased to 94% [mean ± SD, 92.4 ± 1.2%, range 90.1–94.0% (Lo-SaO2), n = 8; Fig. 1A]. The Lo-SaO2 group exhibited significant and sustained decreases in PaO2 beginning at sub-2 and at all times during the prolonged bout (Fig. 1B). In the Hi-SaO2 group, PaO2 initially decreased 6.6 Torr at minute 1 of the prolonged bout, but thereafter it rose progressively so that it was slightly greater than the resting value by the end of exercise. Both groups exhibited a variable but progressive metabolic acidosis, and arterial blood lactate rose similarly in both groups during the exercise (Fig. 2).
|
|
Metabolic rate and ventilation.
There were no significant differences in O2 between the groups during the exercise bouts (Table 3). During prolonged exercise, there was a progressive 34% rise in minute ventilation (E) from minute 1 to end exercise, and this increase was mediated solely by an increased breathing frequency in both groups. The dead space-to-tidal volume ratio decreased from rest during exercise but rose significantly over the final half of prolonged exercise. The ventilatory equivalent for CO2 production (E/CO2) also rose over time during prolonged exercise and was significantly lower in the Lo- than in the Hi-SaO2 group from minute 5 to the end of exercise.
|
E/CO2 was significantly correlated with PaCO2 during prolonged exercise (Fig. 4), indicating that the primary determinant of PaCO2 was the magnitude of the total ventilatory response. Additionally, the magnitudes of the A-aDO2 and PaCO2 were significantly correlated to PaO2 (Fig. 5). Five subjects showed no hyperventilatory response to prolonged heavy-intensity exercise, i.e., PaCO2 of 40 Torr or greater at exercise termination.
|
|
|
Breathing mechanics. Pulmonary resistance was similar in the two groups before the exercise (Fig. 6), and although group means for RLi increased slightly from baseline during all three workloads, there were no major group mean time-dependent changes during the prolonged bout. During the prolonged bout, in six subjects RLi was decreased at minute 1 compared with baseline (–0.9 ± 0.7 cmH2O·l–1·s), whereas in most of the subjects (n = 14), RLi increased compared with baseline (+1.4 ± 1.2 cmH2O·l–1·s). After minute 1 of exercise, RLi remained unchanged or decreased in 19 subjects (average change from minute 1 = –0.3 ± 0.6 cmH2O·l–1·s) and increased by 2.0 cmH2O·l–1·s at end exercise in 1 subject. There was a significant positive correlation between RLi and PaCO2 at the end of the prolonged bout (r = 0.45, P = 0.05).
|
E expressed as a percentage of ventilatory capacity reached 84 and 97% during the prolonged bout in the Hi- and Lo-SaO2 subjects, respectively. EELV was similar in the Hi- and Lo-SaO2 groups throughout the exercise bouts, averaging 51% of TLC at rest, decreasing slightly during sub-1, and then increasing to resting levels during sub-2. During the prolonged bout, EELV increased progressively to 0.27 liter greater than resting FRC at the final measurement in both groups. There were no significant correlations between EFL or E expressed as a percentage of ventilatory capacity and EELV/TLC, E/CO2, or PaCO2 during the exercise bouts.
|
Baseline Rrs 5–25 Hz was higher in the Lo-SaO2 group before exercise (2.2 ± 1.7 and. 0.8 ± 1.3 cmH2O·l–1·s for Lo- and Hi-SaO2 subjects, respectively, P = 0.04), but there were no other significant differences in baseline lung function between the groups. There were no significant correlations between baseline FEV1.0, FEV1.0/FVC, FEF25–75, or Rrs at 5 Hz and the four measures of gas exchange (i.e., PaO2, PaCO2, A-aDO2, and SaO2) during sub-1, sub-2, or the prolonged bout. However, baseline Rrs at 5–25 Hz was significantly correlated with PaO2 (r = –0.53, P = 0.01), PaCO2 (r = 0.53, P = 0.01), and SaO2 (r = –0.47, P = 0.03) at the end of the prolonged bout.
Airway Inflammation
Eosinophils at baseline amounted to 6% of the white blood cells in the sputum and did not change after exercise (Table 5). After exercise, urinary 9,11-PGF2 was significantly increased by 117 ± 207% from the baseline value when all subjects were combined and analyzed as one group. Sputum histamine after exercise was increased above baseline in the Lo-SaO2 subjects (37.8 ± 27.1 vs. 61.4 ± 47.3 ng/ml, P = 0.04), but not in the Hi-SaO2 group (34.3 ± 30.5 vs. 39.1 ± 26.6 ng/ml, P = 0.75). When all subjects were analyzed as one group, postexercise sputum histamine was significantly correlated with PaO2 (r = –0.60, P = 0.007) and the A-aDO2 (r = 0.54, P = 0.02) measured at the end of the prolonged bout. Additionally, the change in histamine after exercise (i.e., the difference between postexercise and baseline) was significantly correlated with the changes in PaO2 (r = –0.68, P = 0.002) and the A-aDO2 (r = 0.50, P = 0.03) between rest and the end of the prolonged bout.
|
Figures 7 and 8 contrast the exercise results for two subjects with similar O2 max and MFVLs but with different ventilatory responses and arterial blood gases during the exercise. Figure 7 shows a subject with a high RLi at rest and a scooped expiratory limb of the MFVL (forced expiratory flow at 50% of vital capacity = 46% predicted). Consequently, EFL occurred at a very moderate increase in E, and substantial increases in EELV ensued. Pulmonary resistance decreased by 1.5 cmH2O·l–1·s during sub-1 and early during the prolonged bout but increased to the resting value during sub-2 and across time during the prolonged bout. Presumably at least in part because of the high resistive and elastic work of breathing, PaCO2 rose progressively during the exercise, and concomitant decreases in PaO2 and SaO2 ensued. Finally, the unchanged expiratory limb of the immediate-postexercise MFVL shows the lack of a significant exercise-induced bronchodilatory response.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Gas exchange and breathing mechanics were assessed in 21 habitually active asthmatic subjects during high-intensity, constant-work rate treadmill exercise to exhaustion. Significant decreases in SaO2 during the exercise were noted in 8 of the 21 subjects due in part to a decreased PaO2, which in turn was caused by a widened A-aDO2 and an insufficient ventilatory response. Ventilation was constrained during the exercise, as suggested by extensive EFL, a high fractional utilization of the maximum ventilatory capacity, pronounced dynamic hyperinflation, and a variable but high pulmonary resistance. These mechanical constraints likely contributed to the highly heterogeneous exercise ventilatory response, with several subjects exhibiting a rising PaCO2 or frank CO2 retention (>40 Torr PaCO2) during exercise. Sputum histamine and urinary 9,11-PGF2 were increased after exercise, providing evidence for an airway inflammatory response to the exercise stimulus. Furthermore, the postexercise sputum histamine levels were significantly correlated with several measures of gas exchange during exercise.
Phenotypic Diversity Among Subjects
We used four different tests of airway function to evaluate our subjects for airway pathology consistent with the diagnosis of asthma. Our subjects showed variable degrees of airway hyperresponsiveness, airflow limitation, and airway inflammation (Table 1). Thus the results from the inclusion criteria studies highlight the heterogeneity of bronchial asthma and the lack of association between different tests of airway function in asthmatic subjects.
Also, subjects 6, 17, and 21 had a PC20 >16 mg/ml. This is important given the notion that airway hyperresponsiveness to direct-acting stimuli is necessary for the diagnosis of asthma. The alternative view is that asthma is more akin to a syndrome with multiple phenotypes and indicators of airway pathology (55). Asthma is a complex and heterogeneous condition, and in our view it is overly simplistic to ascribe the diagnosis of asthma on the basis of one outcome measure. Furthermore, we sought to study a group of subjects with a diverse range of airway characteristics consistent with the diagnosis of bronchial asthma.
Exercise-Induced Arterial Hypoxemia
In the present study, the decreased SaO2 during prolonged heavy-intensity exercise was due almost equally to a decreased PaO2 and a temperature- and pH-induced shift in the oxyhemoglobin dissociation curve. Similar to previous findings in healthy young adults, the exercise PaO2 was highly related to the efficiency of gas exchange, as measured with the A-aDO2, and to the magnitude of the ventilatory response during the exercise, as revealed by the PaCO2 (Fig. 5).
An important difference between the present study and previous studies on gas exchange during exercise in asthmatic subjects (see the introduction) is that our subjects were active and, therefore, capable of working at moderately high work rates and at a high percentage of their O2 max for several minutes. Thus the exercise was at a sufficiently high metabolic requirement to stress the capacity of the cardiopulmonary system for gas exchange and O2 transport. Previous studies of asthmatic subjects exercising at lower exercise intensities have shown preservation of, or increased, PaO2 during exercise (5, 12, 19, 20, 24, 35).
In the present group of subjects, the exercise-induced arterial hypoxemia (EIAH) occurred at lower exercise metabolic rates and in subjects with lower O2 max than previously reported in healthy nonasthmatic individuals. In health, significant EIAH occurs only in a minority of subjects of moderate-to-high fitness who are able to achieve high metabolic rates for several minutes (15, 57); it is infrequent in men and women with a O2 max of less than 60 and 50 ml·kg–1·min–1, respectively (30, 57). In contrast, 50% of our asthmatic subjects (2 men and 2 women) in the Lo-SaO2 group had a O2 max of <42 ml·kg–1·min–1 (range 39.2–41.6) and were exercising at a mean metabolic rate of only 36.1 ± 3.8 ml·kg–1·min–1 at the end of the prolonged bout. Furthermore, 10 of 21 subjects had a 10-Torr decrease in PaO2 during the prolonged exercise bout, with this initial drop persisting to the end of the bout in 7 of the 10 subjects. In contrast, Wasserman et al. (56) found no reduction in PaO2 in 10 healthy but untrained men who performed constant-load cycle exercise to exhaustion at an exercise intensity similar to that used in our prolonged bout.
Baseline Pulmonary Function and Arterial Blood Gas Status During Exercise
Because of the potential for narrowed airways to alter the ventilation distribution and minimize the hyperventilation of heavy exercise, it is reasonable to hypothesize that the level of baseline airflow limitation might provide some prediction of arterial blood gas status during exercise in asthmatic subjects. However, for a given level of baseline airflow limitation in our subjects, exercise PaO2 and PaCO2 spanned a wide range of values. For example, for eight subjects, baseline FEV1.0/FVC was 0.65–0.70 (and FEF25–75 was 55–60% predicted), but PaO2 at end exercise was 71–97 Torr, PaCO2 was 33–47 Torr, and SaO2 was 90–96%. Thus several subjects with compromised baseline lung function and significant EFL during exercise also hyperventilated normally and were able to maintain normal PaO2 during exercise.
The finding that resting lung function was not predictive of the adequacy of pulmonary gas exchange during high-intensity exercise in our subjects with asthma reflects the complexity of multiple factors operating during exercise. On the one hand, exercise causes progressive reductions in mixed venous O2 content, increases in cardiac output, and demands large increases in A, all of which should compromise arterial blood gas status in subjects with already impaired airway and gas exchange function. On the other hand, exercise promotes bronchodilation, which should improve the capacity for ventilation and also improve A/ distribution. However, the magnitude of any exercise-induced bronchodilation is likely to be variable in asthmatic subjects, depending in part on the level of irreversible airflow limitation. For example, subjects 15, 18, 20, and 21 showed some amount of baseline airflow limitation that was only minimally improved after -agonist inhalation (Tables 1 and 2).
Breathing Mechanics During Exercise
Pulmonary resistance showed large variability among subjects during exercise: RLi was normal in several subjects (2 cmH2O·l–1·s) but increased in others (>2–3 cmH2O·l–1·s) relative to values previously reported in healthy subjects (8, 21). The slight group mean increases in RLi during exercise are largely explainable by the increased inspiratory flow rates (peak inspiratory flow = 0.7 ± 0.2 l/s at rest vs. 4.7 ± 1.0 l/s during prolonged exercise bout) and the nonlinearity of the pressure-flow relationship (8, 22). On the basis of previous studies that determined transpulmonary pressure at different inspiratory flow rates in otherwise resting humans (22, 42), we calculated that the increased inspiratory flow rates, per se, caused a 1.0 cmH2O·l–1·s increase in mean RLi at minute 1 of the prolonged exercise bout. This accounts for all the observed increase in RLi in 10 of the 14 subjects who demonstrated an increase relative to rest at this time. The increased RLi in the remaining four subjects (+2.1 ± 1.2 cmH2O·l–1·s vs. baseline after correction for increased flow rates) was beyond that attributable to increased inspiratory flow rate, per se, indicating a decrease in airway caliber during exercise in these subjects.
After minute 1 of prolonged exercise, all but one subject exhibited an unchanged or decreased RLi across time, despite the increasing inspiratory flow rates, strongly suggesting an exercise-induced bronchodilation. These findings are consistent with the bronchodilatory effect of exercise previously reported in most asthmatic subjects (14, 39, 51, 53). In one subject, a marked increase in RLi was noted during the prolonged bout, a finding similar to that reported by Beck et al. (7), in which a subject experienced marked increases in airway resistance during exercise. Thus the airways in a small minority of asthmatic subjects are especially hyperresponsive to the exercise stimulus, despite the strong bronchodilatory influences during exercise.
Our subjects were expiratory flow limited and utilized a large fraction of their ventilatory capacity during exercise. Significant dynamic hyperinflation ensued, and mean EELV increased to well above resting FRC during the prolonged bout. These findings are in general agreement with those from previous studies in asthmatic subjects (34, 52, 53). Moreover, in the present group of subjects, EFL occurred during sub-1 in seven subjects, at a E of only 62 l/min, and during sub-2 in 10 subjects, at a E of 78 l/min. Conversely, in nonasthmatic young adult men and women, EFL does not normally occur until E approaches or exceeds 120 and 100 l/min, respectively (33, 40, 41). Thus asthmatic subjects experience EFL at lower exercise work rates and to a greater degree than their nonasthmatic counterparts.
Heterogeneous Hyperventilatory Response During Exercise
In the present study, PaCO2 spanned a wide range of values at all exercise workloads. During submaximal exercise, the mean values for PaCO2 were consistent with results from the literature that reported arterial blood gases during treadmill exercise in asthmatic (5, 12) and nonasthmatic (29, 57) subjects. In contrast, during high-intensity exercise (i.e., prolonged bout), several subjects displayed a normal exercise hyperventilation (<35 Torr PaCO2), whereas others showed a high or progressively rising PaCO2 across time. Furthermore, four subjects had frank CO2 retention (41.2–46.6 Torr PaCO2) at end exercise. The high prevalence of an insufficient hyperventilatory response in our subjects contrasts with previous studies in asthmatic subjects who showed a normal or even greater hyperventilation during exercise than healthy control subjects (5, 12, 19, 24, 28, 35). The moderate exercise intensities used in these studies, however, required only modest increases in A, which did not stress the capacity of the pulmonary system to generate airflow. However, one of five subjects in a previous study did retain CO2 (45 Torr) during moderate-intensity treadmill exercise (5).
What are the causes of the heterogeneous ventilatory response during exercise in the present group of asthmatic subjects? First, it is clear that the failure to adequately reduce PaCO2 was due to an inadequate increase in E, rather than an excessive dead space-to-tidal volume ratio (Fig. 4). Ventilation during high-intensity exercise is determined by the net effect of two competing influences: 1) the magnitude of and sensitivity to multiple feedforward and chemoreceptor- and locomotor-linked feedback inputs to the brain stem acting to stimulate ventilation out of proportion to CO2 production (45), and 2) the constraints on ventilation imposed by the mechanical characteristics of the extra- and intrathoracic airways and chest wall, in combination with the capacity for pressure development by the respiratory muscles. In this regard, many of the asthmatic subjects in this study had a high airway resistance with a compromised capacity for expiratory airflow (i.e., high flow-resistive work of breathing) and a high elastic work of breathing and shortened diaphragm as a consequence of dynamic hyperinflation during the exercise. EFL is also associated with an increased EELV and attenuates the exercise hyperventilation during high-intensity exercise in healthy, young highly fit men (33, 41) and women (40), in physically active older adults (31), and in subjects with chronic airflow limitation (44). Importantly, however, the extent of mechanical constraint and dynamic hyperinflation in our subjects was markedly greater than that seen in even highly fit athletes exercising at higher metabolic rates with far greater ventilatory requirements.
We also emphasize the finding that several subjects with substantial EFL maintained a reduced EELV and also showed the typical normal hyperventilation of heavy sustained exercise (Fig. 8). The factors accounting for the heterogeneous ventilatory response in the face of similar degrees of constraint in this group of asthmatic subjects are unknown. Interindividual differences in sensitivity to the multiple chemoreceptor- and locomotor-linked stimuli to breathe may explain some of the difference (26). In health, there is also variation in the physiological response to EFL, in terms of its effects on operating lung volumes and on the ventilatory response (33). We can only conclude that EFL and a high E expressed as a percentage of ventilatory capacity were present in all subjects who lacked a sufficient hyperventilation during the exercise; however, EFL, per se, did not always result in a blunted ventilatory response during exercise.
Airway Inflammation
There is good reason to postulate a relation between airway inflammation and gas exchange during exercise (see the introduction). Exercise caused an increase in urinary 9,11-PGF2 when all subjects were analyzed as one group and an increase in sputum histamine in the Lo-SaO2 group only, providing evidence for mast cell degranulation during or shortly after exercise in asthmatic subjects. Other investigators also showed an increase in urinary 9,11-PGF2 after exercise in asthmatic subjects (43, 46). The osmotic hypothesis of EIB stipulates that inflammatory mediators be released into the airway wall during or shortly after exercise (6); however, previous studies are inconsistent regarding the airway inflammatory response to exercise in asthmatic subjects (10, 13, 23, 31, 36, 43, 46). Some of the discrepancy is probably due to the difficulty in accurately detecting changes in inflammatory mediators in the airway lumen or mucosal layer, differences in metabolism of the mediators and the time interval between exercise and collection of sputum or bronchoalveolar lavage, and differences in exercise work rate between the separate studies.
Does airway inflammatory mediator release in response to exercise in asthmatic subjects affect gas exchange during the provocative exercise bout? On the one hand, we found significant correlations between postexercise sputum histamine and A-aDO2 and PaO2 during exercise and also between the exercise-induced change in histamine and the differences in PaO2 and A-aDO2 between end exercise and rest. Thus exercise PaO2 was lower and A-aDO2 was wider in subjects with increased sputum histamine after exercise. If released during exercise, the effects of these mediators on the bronchial vasculature and smooth muscle may have caused a worsened ventilation distribution and perpetuated the development of arterial hypoxemia, even in the absence of reductions in large airway caliber during exercise. Indeed, inhalation of the inflammatory mediators histamine, leukotriene D4, and platelet-activating factor causes an increase in resting A-aDO2 and a decrease in PaO2 in asthmatic subjects (1, 17, 18).
On the other hand, the lack of a time-dependent increase in A-aDO2 or decrease in PaO2 beyond minute 1 of exercise (Figs. 1 and 3) suggests that the inflammatory mediator release did not occur during but only after the exercise [the conventional but untested hypothesis (6)] or that any mediator release during exercise does not negatively impact pulmonary gas exchange. One of the principal effects of inflammatory mediator release is bronchial smooth muscle contraction, but the bronchoprotective effect of exercise in asthmatic subjects (51) might prevent this contraction. Other compensatory mechanisms that would help prevent deterioration in gas exchange in the face of inflammatory mediator release include regional hypoxic pulmonary vasoconstriction, collateral ventilation, and the increase in overall A/ during exercise. Constant-load, high-intensity exercise in nonasthmatic subjects also does not result in any time-dependent decreases in PaO2 or increases in A-aDO2 beyond minute 1 of exercise (29, 57), and acute administration of a cocktail of anti-inflammatory agents before exercise in trained young nonasthmatic subjects had no effect on gas exchange during exercise (58). Future studies that acutely block or chronically treat the airway inflammation of asthma are a necessary next step to rigorously determine the effects of airway inflammation on exercise gas exchange.
Conclusions
Habitually active, mildly to moderately asthmatic subjects appear to develop exercise-induced arterial hypoxemia at lower metabolic rates than their nonasthmatic counterparts. This is related in large part to an insufficient exercise ventilatory response, likely as a consequence of substantial mechanical constraints for ventilation at only moderate levels of E. Additionally, we found evidence supporting a role for airway inflammatory mediator release in the widened A-aDO2 and decreased PaO2 during exercise in asthmatic subjects.
|
GRANTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
|
REFERENCES |
|---|
|
TOPABSTRACTMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
and 
, respectively) at rest, during submaximal exercise (sub-1 and sub-2), and during the prolonged exercise bout to exhaustion. A: SaO2. B: arterial PO2 (PaO2). For individual data, solid lines represent Hi-SaO2 subjects and dashed lines represent Lo-SaO2 subjects. Values are means ± SD. *Significantly different from baseline (P < 0.05). 
Significant difference between Hi- and Lo-SaO2 groups (P < 0.05).
