INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

PROLONGED HEAVY-INTENSITY endurance running or cycling exercise to exhaustion causes low- and high-frequency fatigue of the diaphragm in healthy subjects of varying fitness levels (1, 10, 15). We are concerned with the causes of exercise-induced diaphragm fatigue. Previous studies have shown that the diaphragmatic force output, as represented by the time integral of diaphragmatic pressure (Pdi) multiplied by breathing frequency (f), developed during high-intensity exercise, was not sufficient, by itself, to cause diaphragmatic fatigue, as tested in subjects who voluntarily increased diaphragmatic force output under resting conditions (2). So it was proposed that whole body exercise imposed extra stressors on the diaphragm that would hasten fatigue. We speculated that the increased sensitivity to diaphragm fatigue caused by whole body exercise was primarily due to the relative reduction in blood flow to the diaphragm under conditions in which the respiratory muscles would have to compete with the locomotor muscles for the available cardiac output. In this same study, we also showed that whole body exercise to exhaustion, which caused diaphragm fatigue, did not elicit fatigue in a minimally exercised muscle of the hand (2). Accordingly, we reasoned that the increased force output per se of the diaphragm, as required to produce hyperpnea during exercise, might also contribute significantly to exercise-induced diaphragm fatigue.

In the present study, we used proportional assist mechanical ventilation to partially unload the respiratory muscles during heavy endurance cycling exercise to determine the role of diaphragmatic force output per se on diaphragmatic fatigue caused by whole body exercise.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects. Seven male subjects with normal lung function and no signs of arterial hypoxemia (arterial oxygen saturation was above 92%) during the maximal oxygen uptake (O_2 max) test were recruited to participate in this study. Informed consent was obtained in writing, and the Institutional Review Board of the University of Wisconsin-Madison approved all procedures. Physical characteristics of the subjects were as follows: age = 27.1 ± 2.5 yr; height = 177.3 ± 2.4 cm; weight = 72.0 ± 2.6 kg; O_2 max = 55.4 ± 4.3 ml · kg¹ · min¹, range = 33.3 to 72.5 ml · kg¹ · min¹.

Inspiratory muscle unloading. A feedback-controlled proportional assist ventilator (PAV) was used to reduce the work of the inspiratory muscles during exercise (22). Briefly, subjects breathed through a Hans-Rudolph one-way breathing valve that was connected (on the inspiratory side) to the PAV. The PAV contains a linear motor that drives a piston (5-liter capacity) that develops pressure in proportion to inspiratory airflow and volume. The level of assist was controlled by potentiometers on the control panel of the ventilator, and there are separate controls for volume assist (elastic) and flow assist (resistance). During inspiration, the PAV makes mouth pressure positive in proportion to volume and flow, such that the proportional assist (unloading) of the respiratory muscles occurs throughout the inspiratory cycle. In practice, we set the amount of flow and volume assist at the maximal levels each subject could tolerate during heavy exercise, as determined from practice sessions before testing. During these practice sessions and during testing sessions, subjects were verbally coached to "relax" and permit the PAV to assist each inspiration as much as possible (see Fig. 1).

View larger version (16K): [in this window] [in a new window]   Fig. 1.   An example of one subject's pressure-volume loops during exercise at the same workload (midpoint of exercise time), with no assist (NA; solid line) and with the proportional assist ventilator (PAV; broken line). Work of breathing was calculated from the area contained within the pressure-volume loops (see METHODS). Pes, esophageal pressure.

Paired-stimuli technique. The paired-electrical stimuli technique and its reproducibility have been described in detail (3, 20). Subjects were seated in a semirecumbent position. A level of supramaximal bilateral phrenic nerve stimulation (BPNS) was determined by gradually increasing stimulus current until a plateau was reached in compound action potential (M wave) of the diaphragmatic electromyogram and diaphragmatic pressure (Pdi). During BPNS data collection, the M wave amplitude was checked to ensure that each stimulus resulted in supramaximal stimulation of the diaphragm. Two supramaximal electrical stimuli were delivered to both phrenic nerves at a constant current, and the time interval between stimuli was varied from 10 ms (100 Hz) to 100 ms (10 Hz) (20). For each stimulation protocol, five to eight repeated measures of twitch stimulation (at 1 Hz) and three to five repeated paired stimulations at 10, 20, 50, 70, and 100 Hz were obtained at functional residual capacity (FRC). The resultant esophageal pressure (Pes), gastric pressure, Pdi, and M waves from the left and right costal diaphragm were collected on computer, polygraph record (Gould Em2300) and magnetic tape (Hewlett Packard) for later analysis. Tidal and end-expiratory lung volumes were continuously monitored throughout the stimulation tests by connecting the subject to a wedge spirometer and providing the subject with visual feedback by using an oscilloscope display of lung volume.

High-frequency response. High-frequency stimulation effects are reflected in the response of the diaphragm to the second stimulus. It was assumed that the diaphragmatic response to the first stimulus was similar to a twitch response. Therefore, the single five to eight repeated twitch stimulations were ensemble averaged and then subtracted by computer from the ensemble-averaged paired responses at each frequency, and the amplitude of the resultant response (T₂) was measured. High-frequency fatigue was considered to be present if T₂ at 50, 70, or 100 Hz was different from preexercise values immediately after exercise but not different at 30 min postexercise, whereas the 10- and 20-Hz T₂ were still different from control (3, 4).

Exercise data collection. During exercise, expired gases, flow rates, volumes, Pes, gastric pressure, Pdi, and mouth pressure were monitored continuously and stored on magnetic tape and on computer for later analysis (see Ref. 10 for details). At 2- to 3-min intervals during the no-assist trial, inspiratory capacity measurements were made to estimate end-expiratory lung volume (12). Inspiratory capacity maneuvers proved to be too disruptive to breathing during the PAV trial and were not carried out. For analysis, data from 20 to 30 consecutive breaths were averaged at 3-min intervals during exercise. Pes and Pdi were integrated over the period of inspiratory flow. Means of Pdi and inspiratory time integral of Pes (Pes) were calculated over the 20- to 30-breath sample by the computer, and the results were multiplied by f and labeled the diaphragm force output and the inspiratory muscle force output, respectively. Work of breathing (Wb) was defined as the integrated area of the pressure-tidal volume (VT) loop (16). Wb multiplied by f represented the amount of work done per minute on the lungs. The use of the area within the Pes-volume loop underestimates the actual Wb by a variable amount (6). Thus our measurements provide only a conservative estimate of the total amount of work done by the respiratory muscles during exercise, with both no-assist and PAV exercise.

Exercise test protocols. The seven subjects exercised on an electromagnetically braked cycle ergometer (Elema). The first exercise test determined subjects' O_2 max by using a progressive short-term test to volitional exhaustion, as outlined previously (11). Arterial blood O₂ saturation was estimated throughout this exercise test by ear oximetry (Hewlett-Packard) to determine whether arterial hypoxemia occurred in any of our subjects.

Statistical analysis. Statistical analyses were done by using the statistical program SigmaStat (Jandel). All data are reported as means ± SE. One-way ANOVA with repeated measures was used to determine differences in mean values over the duration of the exercise and recovery period. Two-way ANOVA was used to compare the no-assist exercise to the PAV exercise at the same time points. The level of significance was set at P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Respiratory muscle force output and oxygen uptake. Exercise time to exhaustion during control and PAV conditions averaged 9.6 ± 0.6 min (range = 8.0-12.5 min), and these times were duplicated for each subject during the PAV trial. During exercise under control conditions, the force output of the diaphragm as represented by Pdi · f increased from rest to 20% of total exercise time (rest = 162 ± 0.7 cmH₂O · min; 20% total exercise time = 409.5 ± 68.2 cmH₂O · min) and thereafter leveled off or declined slightly (see Fig. 2). In contrast with during exercise with PAV, Pdi · f increased by ~117% of the rest value (rest = 160.1 ± 20 cmH₂O · min; 20% total exercise time = 191.0 ± 27 cmH₂O · min) and continued to slowly increase until ~60% of the exercise time was reached, and then it decreased slightly to the end of exercise (Fig. 2). Total inspiratory muscle force output (diaphragm plus accessory inspiratory muscles), as represented by Pes · f, increased progressively a total of almost 290% until 60% of total exercise time, when it leveled off over the remaining time. During PAV the Pes · f increased to ~200% greater than rest during the initial 40% of exercise time and then leveled off over the remainder of the exercise time.

View larger version (15K): [in this window] [in a new window]   Fig. 2.   Diaphragmatic pressure force output (Pdi · f; top), all inspiratory muscles force output (Pes · f; middle), and the Pdi/Pes ratio (bottom) during NA exercise () and PAV exercise (). Absolute exercise time averaged 9.6 ± 0.6 min (range of 8-12.5 min). Mean Pes · f and Pdi · f were significantly reduced during exercise with PAV vs. NA (P < 0.05). Note that preexercise values in Figs. 2-5 were obtained after subjects completed warm-up exercise. Values are means ± SE.

View larger version (17K): [in this window] [in a new window]   Fig. 3.   Group mean work of breathing (A) and group mean oxygen consumption (O2; B) during the whole body endurance exercise with NA () and PAV (). Both O2 and work of breathing were significantly reduced during PAV vs. NA (P < 0.05). Values are means ± SE.

Ventilatory reponses to exercise. Ventilatory responses to exercise and to PAV are shown in Figs. 4 and 5. In control, minute ventilation (E) rose by an average of 35% from minute 3 to exhaustion entirely because of an increase in f as VT peaked early and then fell slightly throughout the remainder of exercise. E/O₂ fell by 25% from beginning to end of exercise. With PAV, E was reduced to an average of 9.6 ± 3.4% less than control throughout exercise due solely to a reduction in f with no change in VT. Inspiratory time and mean inspiratory flow rate were also reduced during exercise with PAV. As shown in Fig. 4, the reduction in E during exercise with PAV paralleled the reductions in O₂ (and CO₂); hence, PAV did not affect the degree of hyperventilation (i.e., E/O₂) throughout exercise.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Minute ventilation (E; A) and the E/O2 ratio during the whole body exercise with NA () and with PAV (). Mean E was lower during PAV vs. NA (P < 0.05), whereas mean E/O2 did not drop (P < 0.05). Values are means ± SE.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Tidal volume (VT; top), breathing frequency (f; middle) and mean inspiratory flow rate (VT/TI; bottom) during NA () and PAV () exercise. Mean f and VT/TI were reduced during PAV vs. NA conditions (P < 0.05). Values are means ± SE.

Diaphragmatic fatigue. Immediately after intense, whole body exercise with no assist, the Pdi response to supramaximal twitch stimulation decreased an average of 20.3 ± 3.1% compared with control values (range = 4.2 to 30.3%; P < 0.03; Fig. 6). Twitch Pdi was not significantly changed from preexercise control after PAV exercise. In the no-assist exercise, twitch Pdi was still decreased at 60 min of recovery (11.6%) but was no longer different from preexercise control values (P > 0.05).

View larger version (21K): [in this window] [in a new window]   Fig. 6.   Diaphragmatic pressure (Pdi) (n = 7) response to a single supramaximal 1-Hz stimulation before and after (10, 30, and 60 min) whole body exercise with NA (solid bars) and with PAV (open bars). *Significantly different from preexercise control (P < 0.05).

T₂. Immediately after the no-assist exercise, T₂ amplitude was significantly reduced at 10, 20, 50, 70, and 100 Hz (P < 0.05) (Fig. 7A). Thirty minutes postexercise, 10- and 20-Hz values were still substantially reduced (32.8 and 23.9%, respectively) compared with preexercise values, whereas 50-, 70-, and 100-Hz values were not different vs. control values. These data indicate the presence of high- and low-frequency fatigue in the diaphragm immediately after whole body endurance exercise with no assist. In contrast, at all times after exercise with PAV, T₂ amplitude at all frequencies was not reduced below preexercise control values (Fig. 7B).

View larger version (31K): [in this window] [in a new window]   Fig. 7.   Resultant response at 10, 20, 50, 70, and 100 Hz for control and postexercise. Sampling times are shown for NA (A) and for the PAV exercise protocol (B) at 10, 30, and 60 min postexercise. *Significant difference vs. preexercise control at 10 min postexercise (P < 0.05). #Significant difference vs. preexercise control at 30 min postexercise (P < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Dual cause of exercise-induced diaphragmatic fatigue. Our present findings and those from our laboratory's previous study (2) point to two general causes of exercise-induced diaphragmatic fatigue. First, the effects of whole body exercise per se on the threshold of diaphragmatic force output required for fatigue was previously shown to be significant and substantial (2). As mentioned above, this whole body exercise fatigue factor is likely due to less blood flow availability to the diaphragm during exercise (vs. voluntary hyperpnea at rest) in the face of high blood flow demands by the locomotor muscles. Second, the combination of past and present findings shows that the work of the diaphragm incurred during high-intensity whole body exercise, although not sufficient to cause fatigue when mimicked voluntarily under resting conditions, was critical to causing diaphragmatic fatigue in the presence of whole body exercise. These findings are consistent with the additional observation that the effects of exhaustive high-intensity whole body exercise per se were not sufficient to elicit fatigue in nonexercising muscles of the hand (2, 14). Thus we postulate that the development of diaphragmatic fatigue during exercise is a function of the relationship between the magnitude of diaphragmatic work incurred and the adequacy of its blood supply; the less the latter is available to the diaphragm, the less work is required to produce its fatigue. In healthy subjects of widely varying fitness levels (1), an imbalance of force output by the diaphragm vs. blood flow and/or oxygen transport availability to the diaphragm appears to occur during exhaustive high-intensity endurance exercise (10, 15).

Consequences of exercise-induced diaphragmatic fatigue. By contrasting control and unloaded conditions during exercise, the present study speaks indirectly to the consequences of exercise-induced diaphragmatic fatigue on the regulation of ventilation and of respiratory muscle recruitment. First, f increases substantially over time during high-intensity exercise. Many mechanisms have been implicated, and some researchers have speculated that this tachypnea may be in part related to respiratory muscle fatigue (5). We observed E and f to increase over time (at a fixed workload) in both control and unloaded conditions; however, f was reduced during the latter half of the unloaded trials, relative to control. The fact that the lower f occurred during the ventilator-assisted runs suggests that the prevention of diaphragmatic fatigue may have been involved. However, this lower f also occurs simultaneously with a reduced E; when VT was examined at a given E (21) over the entire time course of the exercise, we found no consistent effect of the unloading on VT or f. Second, we also found that exercise E was significantly reduced during unloading, but this does not necessarily imply an effect of unloading and/or diaphragm fatigue on ventilatory control because the reduction in E matched the reduction in O₂ (and CO₂). Thus E/CO₂ was unchanged. In addition to this metabolic effect of respiratory muscle unloading during heavy exercise, there is also the strong possibility of a behavioral effect on the regulation of breathing as a result of the foreign sensation of accompanying positive-pressure mechanical ventilation. These extraneous effects may explain at least part of the disparity in reported findings on ventilatory control with respiratory muscle unloading during exercise (5, 7, 8, 13). In short, we believe the multiple effects of mechanical unloading during heavy exercise preclude us from interpreting these data solely in terms of a consequence of diaphragmatic fatigue.

Summary and relevance. In summary, present findings have shown that the magnitude of force output required by the diaphragm during high-intensity whole body exercise is a significant determinant of exercise-induced diaphragmatic fatigue. The present data also suggest that most of the time-dependent fall in tidal Pdi, which occurs over the time course of heavy exercise, may be at least in part attributable to exercise-induced diaphragm fatigue.