Effect of Respiratory Muscle Fatigue on Breathing Pattern During Incremental Exercise 1- 3
M. JEFFERY MADOR and FREDERIC A. ACEVEDO Introduction SUMMARY Weexamined the breethlng pattern during Incremental exercise before and after Induc-
Recently the effect of respiratory mustion of inspiratory muscle fatigue. Our aim was to determine whether Induction of fatigue alters cle fatigue on the subsequent breaththe ventilatory response to exercise and In particular whether such changes are most apparent ing pattern was examined. In both an at high levels of exercise when minute ventilation and thus Inspiratory load are greatest. A group anesthetized (1) and awake animal prepof 10 healthy SUbjects was studied on a cycle ergometer. Fatigue was achieved by having the subject aration (2) rapid shallow breathing was breathe against an Inspiratory threshold load that required the subject to generste 80% of the predetermined maximal mouth pressure to Initiate airflow. Breathing pattern, oxygen consumption (VO,), observed following induction of fatigue. mouth occlusion pressure (p•.,), and a visual analog scale (VAS)for respiratory effort were obtained Similarly, in a recent study in human subfor 3 min at rest and at 25, 50, 75, and 100% of the subject's maximal work load (Wmax) as determined jects Gallagher and colleagues (3) obby preliminary testing. Exercise was performed on two separate occasions, once Immediately after served rapid shallow breathing immediInduction of fatigue and the other as a control. Induction of fatigue had no effect on resting breathately following induction of fatigue. In Ing and only minimal effects at the lower work loads (25 and 50% Wmax). At the higher work loads contrast, we have recently examined the (75 and 100% Wmax) Induction of fatigue significantly altered the pattern of breathing during exerbreathing pattern in healthy human subcise. At 75% of Wmax the respiratory frequency (f) Increased from 22.5 :!: 4.4 (SO) during control jects before and after induction of either to 27.0 :!: 6.7 breathS/min (p < 0.02) following Induction of fatigue; tidal volume was not significantly diaphragmatic fatigue or fatigue of all altered, 2.15 :!: 0.65 versus 2.24 :!: 0.74 L during control. The Increase In f was due to reductions the inspiratory muscles (4). We did not in both Inspiratory and expiratory time because fractional Inspiratory time remained unchanged. Mouth occlusion pressure and the VASfor respiratory effort also were Increased following Induction observe any alterations in resting breathof fatigue, 12.9 :!: 4.0 versus 10.7 :!: 4.4 em H,O (p < 0.02) and 54 :!: 16 versus 41 :!: 17 mm"(p ing pattern following induction of either < 0.05). In contrast, Induction of fatigue hed no effect on Vo" CO, production, or heart rate. These diaphragmatic or global inspiratory musresults Indicate that Induction of respiratory muscla fatigue can alter the response to hlgh.lntenslty cle fatigue. These results suggested that exercise, producing rapid but not shallow breathing and Increasing neuromuscular output (p•..) fatigue alone in the absence of an inand respiratory sensation (VAS) for a given work load. AM REV RESPIR DIS 1991; 143:462-468 creased inspiratory load may not be sufficient to produce alterations in breathing pattern. We hypothesized that both fatigue and an increased inspiratory load kg), and height 174 ± 10 em (range 155 to was determined from the ECG. The electrimay be necessary for the development 185em). All subjects had prelimary spirome- cal signals from the recording amplifiers were of alterations in breathing pattern fol- try performed on an Ohio spirometer (Ohio connected to a microprocessor, which perlowing the induction of respiratory mus- Medical Products, Madison, WI) with mea- formed analog-to-digital conversionsand calcle fatigue in awake healthy human sub- surement of FEV" FVC, the FEV./FVC ra- culated average values of VI, VT, f, 0, uptake jects. To test this hypothesis we examined tio, and 12 s maximal voluntary ventilation (Vo,), CO, production (Veo,), and fc every the breathing patern before and after in- (MVV). The FVC was 5.22 ± 1.33 L (per- 30 s. The values were continuously printed duction of fatigue during both resting centage of predicted, 102 ± 11070); FEV" 4.11 and simultaneously stored on disk. The inbreathing (fatigue alone) and incremen- ± 0.95 L (percentage of predicted, 96 ± spiratory flow signal was recorded on a strip tal leg exercise(fatigue plus increased in- 10%), and FEV./FVC ratio, 79 ± 6%. The chart recorder (Gould Inc., Cleveland, OH), spiratory load). Incremental leg exercise MVV was 177 ± 44 Llmin (percentage of and inspiratory time (11), expiratory time (Th), predicted, 104 ± 11%). The study was apwas used as a stimulus to augment min- proved by the appropriate institutional review ute ventilation and therefore inspiratory boards, and informed consent was obtained (Receivedin originalform December 29, 1989 and load. The breathing pattern was exam- from all subjects. in revised form September 21, 1990) ined at 25, 50, 75, and 100010 of the subApparatus. The subjects breathed through ject's maximal work load (as determined a two-way non-rebreathing valve of low reFrom the Division of Pulmonary Medicine, by preliminary testing) to assess the mag- sistance and deadspace (Model 2600; Hans State University of New York at Buffalo, and the nitude of increase in inspiratory load re- Rudolph, Kansas City, MO). Inspiratory flow Veteran's Administration Medical Center, Buffaquired to produce alterations in breath- was measured with a pneumotachograph lo, New York. 2 Supported by grants from the American Lung ing pattern followinginduction of fatigue. (Model 3813; Hans Rudolph) and a differen1
Methods Experiment I Subjects. A group of 10 healthy subjects, 8 males and 2 females, volunteered for this study. Their age was 29.6 ± 4.5 yr (range 23 to 36 yr), weight 69 ± 14 kg (range 48 to 93
462
tial pressure transducer (Model MP-45; Validyne Corp., Northridge, CA). Tidal volume (VT) was obtained by integration of the flow signal. Expired gas waspassed through a mixing chamber and analyzed for oxygen (0,) and carbon dioxide (CO,) by a paramagnetic 0, analyzer (ADC) and an infrared CO, analyzer (ADC), respectively. The heart rate (fc)
Association of New York and the American Heart Association of New York, an institutional. grant from the State University of New York at Buffalo, and by Veterans Administration Medical Research Funds. 3 Correspondence and requests for reprints should be addressed to M. J. Mader, Division of Pulmonary Medicine, Buffalo 'fA Medical Center, 3495 Bailey Ave., Buffalo, NY 14215.
EFFECT OF RESPIRATORY MUSCLE FATIGUE ON BREATHING PATTERN DURING EXERCISE
and fractional inspiratory time (TI/TIot) were measured from this tracing. Expired CO 2was sampled at the mouthpiece, analyzed by asecond infrared CO 2analyzer (LB-2 Medical Gas Analyzer; Beckman Instruments, Fullerton, CA), and recorded breath by breath on the strip chart recorder. End-tidal CO 2 (FETe02) was measured from this tracing. In three subjects secretions in the FETe02 line precluded measurement of FETe02 during heavy exercise so that complete FETe02 measurements were obtained in only seven subjects. Mouth occlusion pressure (Pe.i) was measured in a standard fashion (5) using a commercial inflatable balloon system (Model 9300; Hans Rudolph) attached to the inspiratory limb of the breathing circuit. A pneumatic hand switch, which controlled balloon inflation and deflation, was used to occlude the inspiratory line of the circuit during expiration and to maintain this occlusion for the first 0.25 to 0.30 s of the subsequent inspiration. The pressure was measured at the mouth with a differential pressure transducer (Model MP-45;Validyne)and recorded on a strip chart recorder (Gould) run at a paper speed of 125 mm/s. The Pm developed 100 ms after the start of inspiration (P 0.1) was calculated from the pressure tracing. Approximately three to five occlusion pressures were obtained during the second minute of exerciseat each work load. The sensation of respiratory effort was assessed with a visual analog scale (VAS) (6) consisting of a vertical straight line 100 mm in length. Subjects were instructed to make a mark on the line indicating the intensity of their sensation of respiratory effort at that particular point in time. The subjects were specifically instructed to scale their "effort to breathe" and to disregard any other sensations associated with whole-body exercise. VASmeasurements were obtained during the second minute of exercise at each work load . . Exercise testing. Subjects were studied on three separate occasions. On the first day a preliminary incremental exercisetest was performed on an electronically braked cycle ergometer (Rodby Electronik, Enhorna, Sweden) to determine the maximum working capacity (Wmax) for each subject. For this exercise test after 1 min of unloading cycling the work load was increased every minute by 25 W until the subject was unable to continue. The last work load for which the subject was able to complete the full minute of cycling was designated Wmax. On the second and third experimental days the subjects exercised on the same cycle ergometer at 25, 50, 75, and 100010 of the predetermined Wmax. Subjects exercised for 3 min at each work load. For these experiments the subjects were allowed 2 to 3 min to acclimatize to the breathing circuit, and measurements of resting breathing were obtained for an additional 3 min before initiation of exercise. Thus exercise was initiated 5 to 6 min after termination of loaded breathing. Exercise was performed following induction of inspiratory muscle fatigue on one experimental day; on the other day exercise
alone was performed (control day). To prevent an order effect from introducing systematic bias into the results the control exercise bout was performed on Day 2 on half the subjects and on Day 3 on the other half. All exercise tests were separated by at least 2 days. Measurements of breathing pattern, V02, Ve02, fc and FETe02 were analyzed for the third minute of exercise at each work load. Sincethe measurement of P 0.1 and VASmight affect the breathing pattern, these measurements were obtained during the second minute of exercise at each work load. Inspiratory loading. The subject's maximal mouth pressure (Pm max) was measured while performing a maximum inspiratory effort against an occluded airway near residual volume (7). To prevent glottic closure a small leak was produced by insertion of an 18-gaugeneedle in the mouthpiece. The Pm max maneuvers were repeated until three reproducible measurements sustained for at least 1 s were recorded. The highest value obtained was used for analysis. Following completion of the Pmmaxmaneuvers, the subjects were instructed to breathe against an inspiratory threshold load. Threshold loading was performed using the technique of Nickerson and Keens (8). With this technique, initiation of airflow requires development of a level of pressure sufficient to lift a plunger, which in the fully dependent position occludes the orifice of an otherwise sealed cylinder. Weights were added to the plunger so that the subject was required to generate 80% of the Pmmax to initiate airflow. Expiration was unloaded and the subjects were allowed to choose their own breathing patterns. The TI/TIot adapted by the subjects during loaded breathing was 0.54 ± 0.09 (range 0.41 to 0.66). The endurance time was calculated as the time from the start ofthe run until the subject was no longer able to generate the target pressure in a squarewave fashion. In this experiment endurance time was 13.7 ± 3.0 min. When the subject was unable to generate the target pressure for five consecutive inspiratory efforts, all the inspiratory muscles including the diaphragm were considered fatigued. Following the fatigue run the subjects performed three Pmmax maneuvers, and the highest value was selected for comparison to baseline. The subjects were then immediately transferred to the cycle ergometer and exercise was initiated as described previously. Experiment II To obtain additional evidencethat fatigue was present following inspiratory threshold loading we restudied three of the subjects. During these experiments gastric (pga) and esophageal (Pes) pressures weremeasured by means of two thin-walled latex balloons, one positioned in the stomach and the other in the middle third of the esophagus (9). The balloons were attached to polyethylene tubing 100 em in length with an internal diameter of 1.67mm and an external diameter of 2.42 mm. The gastric balloon was connected to two pressure transducers (MP-45; Validyne).
463
The esophageal balloon was connected to the opposite side of one of the transducers to which the gastric balloon was connected and to a third pressure transducer. The first transducer measured Pga, the second Pdi, and the third measured Pes. The subject's Pdi max was recorded whileperforming a combined Muller expulsive maneuver against an occluded airway at functional residual capacity (10). As with the Pm max maneuvers a small leak to prevent glottic closure was produced by insertion of an 18-gauge needle in the mouthpiece. Pga and Pes were displayed on a storage oscilloscope (Gould) to aid the subject in reaching Pdi max. At least three reproducible measurements were obtained in each subject. The subject's Pesmax was then recorded during a maximal inspiratory maneuver. The subject then breathed against the inspiratory threshold load at 80% of Pesmax in a fashion identical to that of Experiment I (fatigue run). As in Experiment I, when the subject was unable to generate the target pressure for five consecutive inspiratory efforts all the inspiratory muscles including the diaphragm were considered fatigued. Following threshold loading Pdi max and PeSmax were recorded immediately and at 5, 10,and 15min. During the fatigue run the plateau of each square wave for Pes and Pdi was measured breath to breath for 1 min at the beginning, during the middle, and during the penultimate minute of the fatigue run to estimate mean Pes (pes) and mean Pdi (Pdi). Duty cycle(TI/TIot) for the diaphragm and for the rib cage muscles was measured directly from the Pdi andPes tracings, respectively. We then calculated a tension-time index for the diaphragm (Pdi/Pdi max x TI/TIot) (11) and for the rib cage muscles (Pes/Pesmax x TI/TIot) (12). Multiple fatigue runs were performed in each subject (four runs in Subject 1 and three runs in Subjects 2 and 3). Each fatigue run was performed on a separate day. In three experiments (one in Subject 1 and two in Subject 3) the rate of relaxation during a voluntary sniff maneuver was examined before and after the fatigue run. In these experiments a sniff maneuver consisted of a short, sharp inspiratory effort against an occluded airway while the subject wore a nose clip as orginally described by Esau and colleagues (13). Subjects were educated in this technique during an initial trial period via a visual presentation of the sniff tracings. The criteria of Levy and coworkers (14) for the acceptability of curves for later analysis was employed. Before performing the sniff maneuvers the response time of each balloon catheter-recorder system was measured by placing the balloon catheter in a pressurized larger balloon and bursting the latter with a hot needle to create a square-wave fall in pressure (pop test) (15). The 10 to 90% rise time (Tr) was found to be 0.015 s. During each sniff Pdi and Pes wererecorded on a strip chart recorder run at a paper speed of 125 mm/s. The pressure signals were also digitized at 200 Hz and stored on disk. The maximum relaxation rate
464
MADOR AND ACEVEDO
(MRR) wasmeasured as the peak rate of pressure decay (dP/dt max) during the sniff. Since the MRR is pressure dependent it was normalized by dividing by ~P (plateau pressure - baseline pressure) to permit comparison of curves of different peak pressures (13). The time constant (tau) was also calculated by replotting the pressure signals on a logarithmic scale. This yielded a straight line overthe lower 70070 of the curve, indicating that the pressure decay was monoexponential. The time constant of this exponential portion is equal to the reciprocal of the slope of the line. In addition we measured the correlation coefficient of the individual exponential regression, and the r value had to be greater than 0.98 for a measure of tau to be accepted (16).
Data Analysis Data were analyzed by analysis of variance and paired t tests (two-tailed) when appropriate. The results are expressed as the mean ± standard deviation unless otherwise stated, and p < 0.05 was considered significant. Results
Experiment I Pm max fell following induction of fatigue from a baseline value of 146 ± 21 to 121 ± 13 em H 20 (p < 0.01). Every subject was able to complete all stages of exerciseduring the control day. In contrast, following induction of fatigue four subjects were unable to exercise for the full 3 min at 1000/0 of Wmax. These subjects exercised for only 1.33 ± 0.13 min at 100% of Wmax following induction of fatigue. Following induction of fatigue there were no significant differences in the V02, VC02, or fc compared to control values at rest or at 25,50, 75, and 100% ofWmax (figure 1).The maximal V02attained during control exercise was 2.93 ± 0.68 Llmin (43.4 ± 7.0 ml/kg) compared with 2.80 ± 0.70 L/min (41.1 ± 5.7 ml/kg) followinginduction of fatigue. For the subgroup of four subjects who stopped exercise prematurely following
the induction of fatigue, however,the V02 during exercise at 100% of Wmax tended to be lower following induction of fatigue, 2.40 ± 0.81 compared with 2.70 ± 0.97 L/min during control (p < 0.10). In contrast, for the other six subjects V02 was 3.08 ± 0.46 during control compared with 3.07 ± 0.51 Llmin following induction of fatigue (p = NS). To allow comparison of the breathing pattern at an equivalent V02we decided to compare the breathing pattern during the final minute of exercise at 100% of Wmax following induction of fatigue with that obtained during the equivalent time period (isotime) during control. During resting breathing no significant differences in breathing pattern or P O• l were observed following induction of fatigue compared with control. At 25% of Wmax a small increase in P O• l was observed following induction of fatigue 4.0 ± 1.9 compared with 3.3 ± 1.6cmH20 during control (p < 0.02). At 50% of Wmax both VT/TI and VI were increased following induction of fatigue, 1.37 ± 0.20 L/s and 36.64 ± 7.54 L/min versus 1.15 ± 0.26 Lis (p < 0.006) and 33.77 ± 7.62 L/min (p < 0.006) during control, respectively. Alterations in breathing pattern following induction of fatigue became more apparent at the higher work loads (table 1 and figure 2). At 75% of Wmax the VIwas increased following induction of fatigue, 58.05 ± 13.78 compared with 50.40 ± 13.29 L/min (p < 0.04) during control. This increase in VI was due to an increase in f from 22.5 ± 4.4 during control to 27.0 ± 6.7 breaths/ min (p < 0.02) following induction of fatigue, but VT remained unchanged, 2.24 ± 0.74 during control versus 2.15 ± 0.65 L (p = NS) following induction of
fatigue. Interestingly, the magnitude of the fall in Pm max following induction of fatigue tended to correlate with the degree of tachypnea elicited by inspiratory muscle fatigue (r = 0.62, p < 0.10). The increase in f was due to a decrease in both Th from 1.39 ± 0.38 during control to 1.20 ± 0.37 s (p < 0.02) following induction of fatigue and Tr from 1.25 ±0.27 s during control to 1.04 ± 0.22s (p < 0.(03) following induction of fatigue. Accordingly Tt/Ttot remained unchanged, 0.47 ± 0.03 versus 0.46 ± 0.03 following induction of fatigue. Vr/Tt and P O• l both increased following induction of fatigue, 2.07 ± 0.46 L/min and 12.9 ±4.0 em H 20 versus 1.78 ± 0.43 Llmin (p
M. JEFFERY MADOR and FREDERIC A. ACEVEDO Introduction SUMMARY Weexamined the breethlng pattern during Incremental exercise before and after Induc-
Recently the effect of respiratory mustion of inspiratory muscle fatigue. Our aim was to determine whether Induction of fatigue alters cle fatigue on the subsequent breaththe ventilatory response to exercise and In particular whether such changes are most apparent ing pattern was examined. In both an at high levels of exercise when minute ventilation and thus Inspiratory load are greatest. A group anesthetized (1) and awake animal prepof 10 healthy SUbjects was studied on a cycle ergometer. Fatigue was achieved by having the subject aration (2) rapid shallow breathing was breathe against an Inspiratory threshold load that required the subject to generste 80% of the predetermined maximal mouth pressure to Initiate airflow. Breathing pattern, oxygen consumption (VO,), observed following induction of fatigue. mouth occlusion pressure (p•.,), and a visual analog scale (VAS)for respiratory effort were obtained Similarly, in a recent study in human subfor 3 min at rest and at 25, 50, 75, and 100% of the subject's maximal work load (Wmax) as determined jects Gallagher and colleagues (3) obby preliminary testing. Exercise was performed on two separate occasions, once Immediately after served rapid shallow breathing immediInduction of fatigue and the other as a control. Induction of fatigue had no effect on resting breathately following induction of fatigue. In Ing and only minimal effects at the lower work loads (25 and 50% Wmax). At the higher work loads contrast, we have recently examined the (75 and 100% Wmax) Induction of fatigue significantly altered the pattern of breathing during exerbreathing pattern in healthy human subcise. At 75% of Wmax the respiratory frequency (f) Increased from 22.5 :!: 4.4 (SO) during control jects before and after induction of either to 27.0 :!: 6.7 breathS/min (p < 0.02) following Induction of fatigue; tidal volume was not significantly diaphragmatic fatigue or fatigue of all altered, 2.15 :!: 0.65 versus 2.24 :!: 0.74 L during control. The Increase In f was due to reductions the inspiratory muscles (4). We did not in both Inspiratory and expiratory time because fractional Inspiratory time remained unchanged. Mouth occlusion pressure and the VASfor respiratory effort also were Increased following Induction observe any alterations in resting breathof fatigue, 12.9 :!: 4.0 versus 10.7 :!: 4.4 em H,O (p < 0.02) and 54 :!: 16 versus 41 :!: 17 mm"(p ing pattern following induction of either < 0.05). In contrast, Induction of fatigue hed no effect on Vo" CO, production, or heart rate. These diaphragmatic or global inspiratory musresults Indicate that Induction of respiratory muscla fatigue can alter the response to hlgh.lntenslty cle fatigue. These results suggested that exercise, producing rapid but not shallow breathing and Increasing neuromuscular output (p•..) fatigue alone in the absence of an inand respiratory sensation (VAS) for a given work load. AM REV RESPIR DIS 1991; 143:462-468 creased inspiratory load may not be sufficient to produce alterations in breathing pattern. We hypothesized that both fatigue and an increased inspiratory load kg), and height 174 ± 10 em (range 155 to was determined from the ECG. The electrimay be necessary for the development 185em). All subjects had prelimary spirome- cal signals from the recording amplifiers were of alterations in breathing pattern fol- try performed on an Ohio spirometer (Ohio connected to a microprocessor, which perlowing the induction of respiratory mus- Medical Products, Madison, WI) with mea- formed analog-to-digital conversionsand calcle fatigue in awake healthy human sub- surement of FEV" FVC, the FEV./FVC ra- culated average values of VI, VT, f, 0, uptake jects. To test this hypothesis we examined tio, and 12 s maximal voluntary ventilation (Vo,), CO, production (Veo,), and fc every the breathing patern before and after in- (MVV). The FVC was 5.22 ± 1.33 L (per- 30 s. The values were continuously printed duction of fatigue during both resting centage of predicted, 102 ± 11070); FEV" 4.11 and simultaneously stored on disk. The inbreathing (fatigue alone) and incremen- ± 0.95 L (percentage of predicted, 96 ± spiratory flow signal was recorded on a strip tal leg exercise(fatigue plus increased in- 10%), and FEV./FVC ratio, 79 ± 6%. The chart recorder (Gould Inc., Cleveland, OH), spiratory load). Incremental leg exercise MVV was 177 ± 44 Llmin (percentage of and inspiratory time (11), expiratory time (Th), predicted, 104 ± 11%). The study was apwas used as a stimulus to augment min- proved by the appropriate institutional review ute ventilation and therefore inspiratory boards, and informed consent was obtained (Receivedin originalform December 29, 1989 and load. The breathing pattern was exam- from all subjects. in revised form September 21, 1990) ined at 25, 50, 75, and 100010 of the subApparatus. The subjects breathed through ject's maximal work load (as determined a two-way non-rebreathing valve of low reFrom the Division of Pulmonary Medicine, by preliminary testing) to assess the mag- sistance and deadspace (Model 2600; Hans State University of New York at Buffalo, and the nitude of increase in inspiratory load re- Rudolph, Kansas City, MO). Inspiratory flow Veteran's Administration Medical Center, Buffaquired to produce alterations in breath- was measured with a pneumotachograph lo, New York. 2 Supported by grants from the American Lung ing pattern followinginduction of fatigue. (Model 3813; Hans Rudolph) and a differen1
Methods Experiment I Subjects. A group of 10 healthy subjects, 8 males and 2 females, volunteered for this study. Their age was 29.6 ± 4.5 yr (range 23 to 36 yr), weight 69 ± 14 kg (range 48 to 93
462
tial pressure transducer (Model MP-45; Validyne Corp., Northridge, CA). Tidal volume (VT) was obtained by integration of the flow signal. Expired gas waspassed through a mixing chamber and analyzed for oxygen (0,) and carbon dioxide (CO,) by a paramagnetic 0, analyzer (ADC) and an infrared CO, analyzer (ADC), respectively. The heart rate (fc)
Association of New York and the American Heart Association of New York, an institutional. grant from the State University of New York at Buffalo, and by Veterans Administration Medical Research Funds. 3 Correspondence and requests for reprints should be addressed to M. J. Mader, Division of Pulmonary Medicine, Buffalo 'fA Medical Center, 3495 Bailey Ave., Buffalo, NY 14215.
EFFECT OF RESPIRATORY MUSCLE FATIGUE ON BREATHING PATTERN DURING EXERCISE
and fractional inspiratory time (TI/TIot) were measured from this tracing. Expired CO 2was sampled at the mouthpiece, analyzed by asecond infrared CO 2analyzer (LB-2 Medical Gas Analyzer; Beckman Instruments, Fullerton, CA), and recorded breath by breath on the strip chart recorder. End-tidal CO 2 (FETe02) was measured from this tracing. In three subjects secretions in the FETe02 line precluded measurement of FETe02 during heavy exercise so that complete FETe02 measurements were obtained in only seven subjects. Mouth occlusion pressure (Pe.i) was measured in a standard fashion (5) using a commercial inflatable balloon system (Model 9300; Hans Rudolph) attached to the inspiratory limb of the breathing circuit. A pneumatic hand switch, which controlled balloon inflation and deflation, was used to occlude the inspiratory line of the circuit during expiration and to maintain this occlusion for the first 0.25 to 0.30 s of the subsequent inspiration. The pressure was measured at the mouth with a differential pressure transducer (Model MP-45;Validyne)and recorded on a strip chart recorder (Gould) run at a paper speed of 125 mm/s. The Pm developed 100 ms after the start of inspiration (P 0.1) was calculated from the pressure tracing. Approximately three to five occlusion pressures were obtained during the second minute of exerciseat each work load. The sensation of respiratory effort was assessed with a visual analog scale (VAS) (6) consisting of a vertical straight line 100 mm in length. Subjects were instructed to make a mark on the line indicating the intensity of their sensation of respiratory effort at that particular point in time. The subjects were specifically instructed to scale their "effort to breathe" and to disregard any other sensations associated with whole-body exercise. VASmeasurements were obtained during the second minute of exercise at each work load . . Exercise testing. Subjects were studied on three separate occasions. On the first day a preliminary incremental exercisetest was performed on an electronically braked cycle ergometer (Rodby Electronik, Enhorna, Sweden) to determine the maximum working capacity (Wmax) for each subject. For this exercise test after 1 min of unloading cycling the work load was increased every minute by 25 W until the subject was unable to continue. The last work load for which the subject was able to complete the full minute of cycling was designated Wmax. On the second and third experimental days the subjects exercised on the same cycle ergometer at 25, 50, 75, and 100010 of the predetermined Wmax. Subjects exercised for 3 min at each work load. For these experiments the subjects were allowed 2 to 3 min to acclimatize to the breathing circuit, and measurements of resting breathing were obtained for an additional 3 min before initiation of exercise. Thus exercise was initiated 5 to 6 min after termination of loaded breathing. Exercise was performed following induction of inspiratory muscle fatigue on one experimental day; on the other day exercise
alone was performed (control day). To prevent an order effect from introducing systematic bias into the results the control exercise bout was performed on Day 2 on half the subjects and on Day 3 on the other half. All exercise tests were separated by at least 2 days. Measurements of breathing pattern, V02, Ve02, fc and FETe02 were analyzed for the third minute of exercise at each work load. Sincethe measurement of P 0.1 and VASmight affect the breathing pattern, these measurements were obtained during the second minute of exercise at each work load. Inspiratory loading. The subject's maximal mouth pressure (Pm max) was measured while performing a maximum inspiratory effort against an occluded airway near residual volume (7). To prevent glottic closure a small leak was produced by insertion of an 18-gaugeneedle in the mouthpiece. The Pm max maneuvers were repeated until three reproducible measurements sustained for at least 1 s were recorded. The highest value obtained was used for analysis. Following completion of the Pmmaxmaneuvers, the subjects were instructed to breathe against an inspiratory threshold load. Threshold loading was performed using the technique of Nickerson and Keens (8). With this technique, initiation of airflow requires development of a level of pressure sufficient to lift a plunger, which in the fully dependent position occludes the orifice of an otherwise sealed cylinder. Weights were added to the plunger so that the subject was required to generate 80% of the Pmmax to initiate airflow. Expiration was unloaded and the subjects were allowed to choose their own breathing patterns. The TI/TIot adapted by the subjects during loaded breathing was 0.54 ± 0.09 (range 0.41 to 0.66). The endurance time was calculated as the time from the start ofthe run until the subject was no longer able to generate the target pressure in a squarewave fashion. In this experiment endurance time was 13.7 ± 3.0 min. When the subject was unable to generate the target pressure for five consecutive inspiratory efforts, all the inspiratory muscles including the diaphragm were considered fatigued. Following the fatigue run the subjects performed three Pmmax maneuvers, and the highest value was selected for comparison to baseline. The subjects were then immediately transferred to the cycle ergometer and exercise was initiated as described previously. Experiment II To obtain additional evidencethat fatigue was present following inspiratory threshold loading we restudied three of the subjects. During these experiments gastric (pga) and esophageal (Pes) pressures weremeasured by means of two thin-walled latex balloons, one positioned in the stomach and the other in the middle third of the esophagus (9). The balloons were attached to polyethylene tubing 100 em in length with an internal diameter of 1.67mm and an external diameter of 2.42 mm. The gastric balloon was connected to two pressure transducers (MP-45; Validyne).
463
The esophageal balloon was connected to the opposite side of one of the transducers to which the gastric balloon was connected and to a third pressure transducer. The first transducer measured Pga, the second Pdi, and the third measured Pes. The subject's Pdi max was recorded whileperforming a combined Muller expulsive maneuver against an occluded airway at functional residual capacity (10). As with the Pm max maneuvers a small leak to prevent glottic closure was produced by insertion of an 18-gauge needle in the mouthpiece. Pga and Pes were displayed on a storage oscilloscope (Gould) to aid the subject in reaching Pdi max. At least three reproducible measurements were obtained in each subject. The subject's Pesmax was then recorded during a maximal inspiratory maneuver. The subject then breathed against the inspiratory threshold load at 80% of Pesmax in a fashion identical to that of Experiment I (fatigue run). As in Experiment I, when the subject was unable to generate the target pressure for five consecutive inspiratory efforts all the inspiratory muscles including the diaphragm were considered fatigued. Following threshold loading Pdi max and PeSmax were recorded immediately and at 5, 10,and 15min. During the fatigue run the plateau of each square wave for Pes and Pdi was measured breath to breath for 1 min at the beginning, during the middle, and during the penultimate minute of the fatigue run to estimate mean Pes (pes) and mean Pdi (Pdi). Duty cycle(TI/TIot) for the diaphragm and for the rib cage muscles was measured directly from the Pdi andPes tracings, respectively. We then calculated a tension-time index for the diaphragm (Pdi/Pdi max x TI/TIot) (11) and for the rib cage muscles (Pes/Pesmax x TI/TIot) (12). Multiple fatigue runs were performed in each subject (four runs in Subject 1 and three runs in Subjects 2 and 3). Each fatigue run was performed on a separate day. In three experiments (one in Subject 1 and two in Subject 3) the rate of relaxation during a voluntary sniff maneuver was examined before and after the fatigue run. In these experiments a sniff maneuver consisted of a short, sharp inspiratory effort against an occluded airway while the subject wore a nose clip as orginally described by Esau and colleagues (13). Subjects were educated in this technique during an initial trial period via a visual presentation of the sniff tracings. The criteria of Levy and coworkers (14) for the acceptability of curves for later analysis was employed. Before performing the sniff maneuvers the response time of each balloon catheter-recorder system was measured by placing the balloon catheter in a pressurized larger balloon and bursting the latter with a hot needle to create a square-wave fall in pressure (pop test) (15). The 10 to 90% rise time (Tr) was found to be 0.015 s. During each sniff Pdi and Pes wererecorded on a strip chart recorder run at a paper speed of 125 mm/s. The pressure signals were also digitized at 200 Hz and stored on disk. The maximum relaxation rate
464
MADOR AND ACEVEDO
(MRR) wasmeasured as the peak rate of pressure decay (dP/dt max) during the sniff. Since the MRR is pressure dependent it was normalized by dividing by ~P (plateau pressure - baseline pressure) to permit comparison of curves of different peak pressures (13). The time constant (tau) was also calculated by replotting the pressure signals on a logarithmic scale. This yielded a straight line overthe lower 70070 of the curve, indicating that the pressure decay was monoexponential. The time constant of this exponential portion is equal to the reciprocal of the slope of the line. In addition we measured the correlation coefficient of the individual exponential regression, and the r value had to be greater than 0.98 for a measure of tau to be accepted (16).
Data Analysis Data were analyzed by analysis of variance and paired t tests (two-tailed) when appropriate. The results are expressed as the mean ± standard deviation unless otherwise stated, and p < 0.05 was considered significant. Results
Experiment I Pm max fell following induction of fatigue from a baseline value of 146 ± 21 to 121 ± 13 em H 20 (p < 0.01). Every subject was able to complete all stages of exerciseduring the control day. In contrast, following induction of fatigue four subjects were unable to exercise for the full 3 min at 1000/0 of Wmax. These subjects exercised for only 1.33 ± 0.13 min at 100% of Wmax following induction of fatigue. Following induction of fatigue there were no significant differences in the V02, VC02, or fc compared to control values at rest or at 25,50, 75, and 100% ofWmax (figure 1).The maximal V02attained during control exercise was 2.93 ± 0.68 Llmin (43.4 ± 7.0 ml/kg) compared with 2.80 ± 0.70 L/min (41.1 ± 5.7 ml/kg) followinginduction of fatigue. For the subgroup of four subjects who stopped exercise prematurely following
the induction of fatigue, however,the V02 during exercise at 100% of Wmax tended to be lower following induction of fatigue, 2.40 ± 0.81 compared with 2.70 ± 0.97 L/min during control (p < 0.10). In contrast, for the other six subjects V02 was 3.08 ± 0.46 during control compared with 3.07 ± 0.51 Llmin following induction of fatigue (p = NS). To allow comparison of the breathing pattern at an equivalent V02we decided to compare the breathing pattern during the final minute of exercise at 100% of Wmax following induction of fatigue with that obtained during the equivalent time period (isotime) during control. During resting breathing no significant differences in breathing pattern or P O• l were observed following induction of fatigue compared with control. At 25% of Wmax a small increase in P O• l was observed following induction of fatigue 4.0 ± 1.9 compared with 3.3 ± 1.6cmH20 during control (p < 0.02). At 50% of Wmax both VT/TI and VI were increased following induction of fatigue, 1.37 ± 0.20 L/s and 36.64 ± 7.54 L/min versus 1.15 ± 0.26 Lis (p < 0.006) and 33.77 ± 7.62 L/min (p < 0.006) during control, respectively. Alterations in breathing pattern following induction of fatigue became more apparent at the higher work loads (table 1 and figure 2). At 75% of Wmax the VIwas increased following induction of fatigue, 58.05 ± 13.78 compared with 50.40 ± 13.29 L/min (p < 0.04) during control. This increase in VI was due to an increase in f from 22.5 ± 4.4 during control to 27.0 ± 6.7 breaths/ min (p < 0.02) following induction of fatigue, but VT remained unchanged, 2.24 ± 0.74 during control versus 2.15 ± 0.65 L (p = NS) following induction of
fatigue. Interestingly, the magnitude of the fall in Pm max following induction of fatigue tended to correlate with the degree of tachypnea elicited by inspiratory muscle fatigue (r = 0.62, p < 0.10). The increase in f was due to a decrease in both Th from 1.39 ± 0.38 during control to 1.20 ± 0.37 s (p < 0.02) following induction of fatigue and Tr from 1.25 ±0.27 s during control to 1.04 ± 0.22s (p < 0.(03) following induction of fatigue. Accordingly Tt/Ttot remained unchanged, 0.47 ± 0.03 versus 0.46 ± 0.03 following induction of fatigue. Vr/Tt and P O• l both increased following induction of fatigue, 2.07 ± 0.46 L/min and 12.9 ±4.0 em H 20 versus 1.78 ± 0.43 Llmin (p
