INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We and others have shown that the diaphragm can become fatigued after high-intensity cycle exercise to exhaustion in healthy young subjects (1). Exercise-induced diaphragmatic fatigue was observed in both sedentary and physically active subjects (1, 2). Aging has profound effects on peripheral muscle function. In humans, muscle mass decreases by 25% to 30% and muscle strength declines by 30% to 40% by 60 to 70 yr of age (4). The diaphragm is not protected from the effects of aging. Diaphragmatic strength is significantly reduced in the elderly, with reductions of 13% to 25% found in two recent studies (5, 6). Studies done in vitro (7) and in vivo (8) also suggest that the diaphragm might be more fatigable in the elderly. Changes in pulmonary mechanics with aging (such as loss of elastic recoil) (9, 10) increase the work and O₂ cost of breathing during exercise (11, 12). Thus, elderly subjects should be predisposed to develop diaphragmatic fatigue after strenuous exercise. However, exercise capacity as reflected by maximum oxygen consumption (O_2max) declines progressively with age (13). Given this decline in maximum exercise capacity, there is the question of whether elderly subjects exercise intensely enough to develop diaphragmatic fatigue. The first aim of this study was to determine whether diaphragmatic fatigue occurs in the elderly after high-intensity cycle exercise to the limits of tolerance. Accordingly, we measured twitch transdiaphragmatic pressure (Pdi_tw) before and after high-intensity, constant-workload cycle exercise to exhaustion in a group of healthy elderly subjects.

Elderly subjects commonly complain of leg fatigue as their reason for stopping exercise (14). Edwards and colleagues showed many years ago that low-frequency fatigue of the quadriceps muscle can occur after cycle exercise (15). However, significant subjective symptoms of fatigue or dyspnea can occur in the absence of overt contractile fatigue (16, 17). Whether elderly subjects are capable of exercising intensely enough to induce overt contractile fatigue of the quadriceps muscle has not been determined. Therefore, the second aim of our study was to determine whether significant fatigue of the quadriceps muscles occurs after intense exercise in the healthy elderly. Measurement of quadriceps twitch force (Q_tw) during magnetic stimulation of the femoral nerve is a reliable method for detecting quadriceps fatigue (18). Accordingly, we measured quadriceps Q_tw before and after high-intensity, constant-workload cycle exercise to the limits of tolerance in the same group of healthy elderly subjects.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subjects

Ten healthy elderly subjects aged 68.3 ± 1.4 yr (mean ± SE) (and all older than 60 yr of age) volunteered for the study. Their height and weight were 1.78 ± 0.02 m and 92.2 ± 4.6 kg, respectively. The subjects' body mass index (BMI) was 28.8 ± 1.4 kg/m² (i.e., the subjects were generally overweight). Two subjects were obese with a BMI > 30 kg/m². All subjects were in good health and were free of respiratory, cardiac, or neurologic disease. All subjects had pulmonary function test results within normal limits (FEV₁: 86 ± 2% predicted; TLC: 97 ± 5% predicted; diffusing capacity for carbon monoxide [DL_CO]: 104 ± 7% predicted), an unremarkable resting electrocardiogram, and no ST-T-wave changes suggestive of ischemia during a maximal, symptom-limited incremental exercise test. The study was approved by the appropriate institutional review boards, and written informed consent was obtained from all subjects.

Exercise

An incremental, symptom-limited exercise test was performed to determine each subject's maximal work capacity. After a 3-min acclimatization period and 1 min of unloaded cycling (an accessory motor was used to rotate the flywheel of the cycle at the start of exercise to overcome inertial forces), the workload was increased by 15 W/min until the subject could no longer continue. The last workload for which a subject was able to complete the full minute of cycling was designated Wpeak. Constant-workload exercise was performed on an electronically braked cycle ergometer (CPE-2000; Medgraphics, St. Paul, MN) at 65% to 75% of Wpeak to volitional exhaustion on two separate days. Diaphragmatic function was evaluated on one study day and quadriceps function on the other. Time constraints (after exercise) and patient tolerance would not allow both measurements to be performed on the same day. The order in which diaphragmatic and quadriceps function were evaluated was randomized, with a balanced design (i.e., five subjects underwent the diaphragm studies first and five subjects underwent the quadriceps studies first). Subjects were allowed to rest for at least 2 d between exercise bouts. The subjects were then again allowed 3 min to become acclimatized to the breathing circuit, and then exercised for 1 min of unloaded cycling and 2 min at 15 W (warm-up period) before initiating exercise at 65% to 75% of Wmax. Cardiopulmonary parameters during exercise were measured as described previously (16). Immediately after exercise was discontinued, the patients were asked whether they stopped because of shortness of breath, leg discomfort, chest pain, lightheadedness, or other symptoms. If patients chose more than one symptom, we asked them to choose the symptom that they felt was most responsible for their stopping exercise.

Quadriceps Measurements

The femoral nerve was stimulated with a magnetic stimulator (Magstim 200; Magstim Co. Ltd., Whitland, Dyfed, Wales), using a 45-mm figure-of-eight coil (18). The technical details of this procedure have been described previously (16). Quadriceps compound motor action potentials were recorded with surface electrodes placed over the belly of the rectus femoris. In the majority of subjects, magnetic stimulation of the femoral nerve elicited a large shock artifact that obscured the compound motor action potential. However, with repositioning of the surface electrodes and grounding (before baseline measurements), compound motor action potentials were obtained before and after exercise in three patients. A minimum of eight unpotentiated twitches (Q_twu) were measured before and at 10, 30, and 60 min after exercise. After a vigorous voluntary contraction, the subsequent twitch is significantly increased in magnitude (twitch potentiation) (19). Recent studies have suggested that the potentiated twitch is more sensitive for detecting fatigue than is the unpotentiated twitch, particularly when the degree of fatigue is small (20, 21). Accordingly, we measured Q_tw at 5 s after a maximum voluntary contraction (MVC). Subjects maintained the MVC for 5 s, and visual feedback of the force signal was provided to the subject with an oscilloscope. The maneuvers were repeated a minimum of six times to obtain at least six potentiated twitches (Q_twp). However, because we had shown in a prior study that the degree of potentiation was slightly smaller after the first and, to a lesser extent, after the second MVC, we discarded the first two measurements (21). Potentiated twitches were measured before and at 10, 30, and 60 min after exercise. Potentiated twitches were always measured after measurement of the unpotentiated set of twitches had been completed.

To determine the degree to which our subjects could voluntarily activate their quadriceps muscle, twitches were measured during the last two MVC maneuvers of each set of measurements in every subject. Superimposed twitches were compared to resting potentiated twitches to determine the percent activation during the MVC maneuver (100  superimposed twitch/resting potentiated twitch × 100) (22).

To determine whether cycle exercise could elicit both high- and low-frequency fatigue, paired stimuli were applied to six subjects (23). The paired stimuli were given at interstimulus intervals of 100 ms, 50 ms, and 10 ms, corresponding to stimulation frequencies of 10 Hz, 20 Hz, and 100 Hz, respectively. The paired stimuli were given with the same figure-of-eight 45-mm coil, which was powered by two linked magnetic stimulators (BiStim Module; Magstim). Four paired stimuli were applied at each stimulation frequency before and at 10, 30, and 60 min after exercise. The paired stimuli were always applied after the Q_twu measurements but before the Q_twp measurements. The values of Q_twu were ensemble-averaged and then digitally subtracted from the ensemble-averaged paired stimuli to obtain the amplitude of the second twitch, T₂. A significant decrease in T₂ after exercise would be indicative of fatigue at that particular stimulation frequency (23). All twitches were obtained at 100% of stimulator output. Q_tw and quadriceps electromyographic data were digitized and stored on disk, using Windaq software (Dataq Instruments Inc., Akron, OH) at a sampling rate of 1,000 Hz.

Diaphragm Measurements

The phrenic nerves were subjected to both cervical (24) and anterolateral magnetic stimulation (25). Gastric, esophageal (Pes), and transdiaphragmatic (Pdi) pressure were measured with two balloon catheters, using standard techniques as previously described (1). As was done with Q_tw, pressure signals were digitized and stored on disk. In the case of one subject, the gastric balloon developed a leak during the experiment, and data from this subject were discarded.

Cervical magnetic stimulation was applied with a commercial magnetic stimulator (Magstim 200), using a circular 90-mm coil (24). The technical details of this procedure have been described previously (17).

Bilateral anterolateral magnetic stimulation was applied with two 45-mm figure-of-eight coils, each of which was connected to a magnetic stimulator (Magstim 200) (25). The two magnetic stimulators were connected in such a way as to allow simultaneous stimulation of the phrenic nerves on each side of the diaphragm. The coils were placed at the posterior border of the sternomastoid muscle at the level of the cricoid cartilage. When the optimal position of the coil was located, the position was marked and used for all subsequent twitches.

To obtain potentiated twitches, subjects performed a maximal Mueller maneuver for 5 s against an occluded airway. Visual feedback of the esophageal pressure curve was provided. Potentiated diaphragmatic twitch Pdi_twp was measured 5 s after the Mueller maneuver was completed. The technique (cervical or anterolateral stimulation) that produced the largest unpotentiated twitch was used for the Pdi_twp measurements. Five subjects were stimulated through the anterolateral and four subjects through the cervical approach. Because lung volume can affect twitch amplitude, all twitches were recorded at the same end-expiratory lung volume (FRC) as determined by end-expiratory pressure. Before nerve stimulation, the mouthpiece (Model 4200C; Hans Rudolph, Kansas City, MO) was occluded via a mouth shutter to prevent any change in volume during nerve stimulation. All twitches were obtained at 100% of stimulator output.

Eight unpotentiated twitches for each stimulation technique (cervical and anterolateral) and six potentiated twitches were measured before and at 15, 30, and 60 min after exercise. Potentiated twitches were always measured after measurement of the unpotentiated sets of twitches had been completed. The order of cervical and anterolateral magnetic stimulation was randomized.

A minimum of eight maximal sniff maneuvers were performed before and immediately after exercise. Visual feedback of the Pes signal was provided to the subject. The best effort was chosen for analysis. During exercise, peak and mean inspiratory swings were calculated on a breath-to-breath basis from the digitized pressure tracings. The beginning and end of inspiration were determined from the inspiratory flow signal, and mean pressures were calculated for this time period. Baseline Pes and Pdi were taken as the respective pressures immediately preceding the inspiratory upswing. Diaphragmatic and esophageal pressure-time integrals (PTP) were calculated as the product of mean pressure, inspiratory time, and respiratory rate. Tension-time indices were calculated from both the esophageal and diaphragmatic pressure curves as the product of mean pressure/maximal sniff pressure (Pdi and Pes) × inspiratory time/total breath duration.

Lactate Measurements

For lactate measurements, a Teflon catheter was inserted into a vein in the dorsum of the hand. To arterialize the blood, the dorsum of the hand was heated until skin temperature exceeded 42° C for at least 2 min before sampling (26). To clear catheter and tubing deadspace, the first 2 ml of blood was discarded. Blood samples were obtained before and at 5, 15, and 30 min after exercise. Lactate measurements were made during the quadriceps study day.

Data Analysis

Changes in variables over time were analyzed by repeated measures analysis of variance (ANOVA). If the ANOVA gave a significant result, specific times were compared with the paired t test with Bonferroni's correction. Data are expressed as mean ± SE. For each individual subject, a persistent decrease in Q_tw of > 15% after exercise was considered potentially indicative of quadriceps fatigue. In our laboratory, the average within-subject, between-trial coefficient of variation (CV) for Q_tw was 8% (21). We reasoned that any difference from baseline that was twice the average between-trial variability was unlikely to be due to change alone. Similarly, a persistent decrease in Pdi_tw of > 15% after exercise was considered potentially indicative of diaphragmatic fatigue (the average within-subject, between-trial CV for Pdi_tw was 6% to 8%) (27, 28).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cardiopulmonary Parameters During Exercise

During the maximal incremental exercise test, the subjects exercised to a peak workload of 108 ± 10 W. O_2peak was 1.46 ± 0.23 L/min, or 71 ± 3% of the predicted value. Peak heart rate (HR) was 140 ± 6 beats/min, or 92 ± 4% of the predicted maximum. Peak exercise ventilation was 56.2 ± 4.4 L/min, or 50.4 ± 4.5% of the 12-s MVV. The respiratory quotient at end-exercise was 1.16 ± 0.04. An anaerobic threshold was clearly reached in nine of the ten subjects at a O₂ of 51.8 ± 2.5% of the predicted maximum. Nine subjects stopped exercise because of leg discomfort and one subject stopped exercise because of a dry mouth.

During constant-load exercise, the subjects exercised at 75 ± 7 W, or approximately 70.8 ± 3.9% of the Wpeak measured during the incremental test. Subjects exercised for 21.9 ± 1.8 min during the quadriceps fatigue studies and for 17.7 ± 1.9 min during the diaphragmatic fatigue studies. The difference in endurance times was significant at p < 0.001 (probably reflecting that the subjects exercised with esophageal and gastric balloons in place during the diaphragm fatigue studies). Cardiopulmonary parameters during the two constant-load exercise studies are shown in Table 1. No other significant differences were observed between the two exercise tests. In both studies, subjects exercised very intensely, reaching peak O₂ values and peak heart rates that exceeded those obtained during the maximal incremental exercise test.

Quadriceps Muscle Measurements

At baseline, Q_twu averaged 8.5 ± 0.7 kg. Q_twu (expressed as a percentage of the baseline value) before and after exercise is shown in Figure 1. Q_twu fell significantly to 64.0 ± 6.3% of the baseline value at 10 min after exercise (p < 0.0003), and remained significantly decreased at 30 min after exercise. Q_twp at baseline averaged 12.7 ± 1.3 kg, which was 49 ± 7% (p < 0.0004) larger than Q_twu. Q_twp fell significantly, to 63.5 ± 2.7% of the baseline value at 10 min after exercise (p < 0.0001), and remained significantly decreased at 30 and 60 min after exercise (Figure 1). Q_twu and Q_twp before and at 10 min after exercise in each subject are shown in Figure 2. Q_twu fell by > 15% in nine of the 10 subjects, whereas Q_twp fell by > 15% in all 10 subjects.

View larger version (22K): [in this window] [in a new window]   Figure 1.   Unpotentiated (open squares) and potentiated (closed squares) Qtw and MVC force (open triangles) expressed as percentages of baseline values before and 10, 30, and 60 min after exercise (mean ± SE, n = 10). * Significant difference from baseline value. Unpotentiated and potentiated Qtw fell significantly after exercise and remained significantly decreased at 30 min after exercise.

View larger version (23K): [in this window] [in a new window]   Figure 2.   Unpotentiated and potentiated Qtw before and 10 min after exercise in each subject (n = 10). Subjects who showed at least a 15% reduction in Qtw after exercise, which was potentially indicative of contractile fatigue of the quadriceps muscle, are represented by closed circles, whereas subjects who did not meet this criterion are represented by open circles. Qtwu fell by > 15% after exercise in nine of the 10 subjects, whereas Qtwp fell by > 15% in all 10 subjects.

At baseline, MVC averaged 42.6 ± 3.3 kg. Twitches superimposed upon the MVC maneuver in the fresh state (baseline) averaged 25.9 ± 4.0% of the resting potentiated twitch value, indicating that subjects were not able to fully activate their quadriceps muscles during the MVC maneuver. No significant differences from baseline were observed in the degree of activation of the quadriceps muscle at any time after exercise (i.e., subjects achieved the same degree of activation of the quadriceps muscle in the fatigued state as they did in the fresh state). MVC was not significantly different from its baseline value at any time after exercise (94.1 ± 3.7% of the baseline value at 10 min after exercise, 91.0 ± 3.9% at 30 min after exercise, and 87 ± 4.9% at 60 min after exercise).

In three subjects, compound motor action potentials not obscured by stimulus artifact were obtained before and after exercise. Compound motor action potential amplitude was not significantly different from baseline at any time after exercise. At baseline, the paired twitch values averaged 13.0 ± 1.1 kg at 10 Hz, 16.6 ± 2.0 kg at 20 Hz, and 17.9 ± 2.9 kg at 100 Hz. At baseline, T₂ averaged 8.7 ± 0.9 kg at 10 Hz, 10.1 ± 1.3 kg at 20 Hz, and 11.6 ± 2.0 kg at 100 Hz. T₂ at 10 Hz, 20 Hz, and 100 Hz before and after exercise is shown in Figure 3. T₂ was significantly decreased from baseline at 10 and 30 min after exercise, at 10 Hz and 20 Hz, respectively, but not at 100 Hz. The decrease in amplitude of T₂ after exercise was significantly greater at 10 Hz and 20 Hz than at 100 Hz (ANOVA, p < 0.0001 and p < 0.003, respectively).

View larger version (18K): [in this window] [in a new window]   Figure 3.   Qtw elicited by second stimulus (T2 response) at 10 Hz, 20 Hz, and 100 Hz before (open circles) and 10 min (closed circles), 30 min (open triangles), and 60 min (closed triangles) after exercise. Mean ± SE, n = 6. * Significant difference from baseline value. T2 was significantly decreased after exercise at 10 Hz and 20 Hz, but not at 100 Hz.

Diaphragm Measurements

At baseline, Pdi_tw was 16.0 ± 1.3 cm H₂O during cervical magnetic stimulation. Pdi_twp was 25.7 ± 2.7 cm H₂O, which was 65.6 ± 16.7% larger than the values for unpotentiated twitch (Pdi_twu). Values for unpotentiated and potentiated Pdi_tw (n = 9) before and after exercise are shown in Figure 4. Neither potentiated nor unpotentiated Pdi_tw were significantly different from the respective baseline values at any time after exercise (Figure 4). The gastric and esophageal components of the twitch were also not significantly different from baseline at any time after exercise. None of the individual subjects had a persistent reduction of  15% in either unpotentiated or potentiated Pdi_tw. Neither Pdi nor Pes during a maximal sniff maneuver were significantly altered immediately after exercise. Sniff Pdi was 115 ± 6 cm H₂O before exercise and 113 ± 6 cm H₂O immediately after exercise (p = NS). Sniff Pes was 95 ± 9 cm H₂O before exercise and 94 ± 10 cm H₂O immediately after exercise (p = NS).

View larger version (24K): [in this window] [in a new window]   Figure 4.   Pditwu during cervical (closed squares) and bilateral anterolateral (closed circles) magnetic stimulation, and Pditwp (closed triangles) before and at 10, 30, and 60 min after exercise. Data are expressed as percentages of baseline values (mean ± SE, n = 9). Neither potentiated nor unpotentiated Pditw was significantly different from the respective baseline value at any time after exercise.

Breathing Pattern During Exercise

At end exercise, mean Pes and mean Pdi were 11.7 ± 1.1 cm H₂O and 11.5 ± 1.1 cm H₂O, respectively. PT Pes and PT Pdi were 316 ± 29 and 311 ± 28 cm H₂O · s · min¹, respectively. Esophageal tension time (Tes_t) and diaphragmatic tension time were 0.057 ± 0.005 and 0.044 ± 0.004. Peak inspiratory flow was 2.6 ± 0.2 L/min.

Lactate Measurement

Lactate levels before and after exercise are shown in Figure 5. Serum lactate levels increased from 1.8 ± 0.1 mEq/L at baseline to 6.8 ± 0.6 mEq/L at 5 min after exercise (p < 0.0001).

View larger version (15K): [in this window] [in a new window]   Figure 5.   Lactate levels before and 5, 15, and 30 min after exercise, (mean ± SE, n = 10). *Significant difference from baseline. Lactate levels increased significantly after exercise.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The major finding of this study was that significant contractile fatigue of the quadriceps muscle occurred after high-intensity cycle exercise to exhaustion, whereas diaphragmatic fatigue did not in a group of healthy elderly subjects.

Quadriceps Function and Exercise: Critique of Methods

Magnetic stimulation of the femoral nerve was used in this study to detect fatigue of the quadriceps muscle. This technique has recently been studied and validated (18), and offers considerable advantages over previous stimulation techniques.

One limitation with magnetic stimulation of the femoral nerve is that the compound motor action potential is often obscured by a large stimulus artifact. We were able to obtain satisfactory compound motor action potentials for three of our 10 subjects. Compound motor action potential amplitude was similar to its baseline value at all times after exercise in every subject. Thus, there was no evidence of transmission fatigue. The reproducibility of Q_twu in the absence of compound motor action potentials was 7.5 ± 0.5%, and for Q_twp it was 5.6 ± 0.9% (21).

Quadriceps Fatigue and Exercise

Significant quadriceps fatigue was observed after high-intensity, constant-load exercise. As mentioned earlier, there was no evidence of transmission fatigue (unchanged compound motor action potential), indicating that the reduction in Q_tw represented contractile fatigue. The fatigue was long lasting: Q_tw remained significantly decreased at 1 h after exercise. In addition, T₂ was significantly decreased from its baseline value after exercise, at 10 Hz and 20 Hz, but not at 100 Hz (Figure 3). The decrease in amplitude of T₂ after exercise was significantly greater at 10 Hz and 20 Hz than at 100 Hz (Figure 3). We interpret these findings as indicating that high-frequency fatigue did not occur after exercise or that recovery had occurred by 10 min when we made our first measurements (recovery from high-frequency fatigue can occur very quickly). In contrast, the significant sustained reduction in T₂ at 10 Hz and 20 Hz supports our twitch data indicating that low-frequency fatigue did occur after exercise. Edwards and colleagues achieved similar results with direct submaximal stimulation of the quadriceps in a small number of studies (n = 2) (15).

The MVC obtained in our elderly subjects, of 42.6 ± 3.3 kg, was significantly less than the average MVC observed in younger subjects in our laboratory, which was 71.7 ± 4.9 kg (p < 0.0001). A portion of this difference was due to central factors, since our elderly subjects achieved less complete activation of their quadriceps muscle during the MVC maneuver than did the younger subjects (75.3 ± 4.0% versus 86 ± 2.5%, p < 0.025). However, even when the MVC values were corrected for this difference in activation, the MVC in our elderly subjects was, as expected, markedly smaller (68.9%) than that observed in the younger subjects.

In animal studies, senescent rats have a lower hindlimb muscle oxidative capacity than do young adult animals (29). Similarly, the percentage of type II fibers decreases with age (4). The ability to sustain power (a measure of fatigability) is also decreased with aging (4). Many of these changes are due to a reduction in physical activity rather than to aging per se. If the reduction in spontaneous activity that occurs in senescent rats is prevented by regular treadmill exercise, most of the changes just described are prevented (30). Our subjects were quite sedentary, and might therefore have been especially likely to demonstrate decrements in limb muscle endurance.

In our study, Q_tw fell significantly after exercise, whereas MVC was not significantly different from its baseline value at 10 min after exercise (Figure 1). As outlined earlier, our subjects developed primarily low-frequency fatigue after exercise. The MVC maneuver is a high-frequency maneuver, and MVC will obviously not be a good index for detecting low-frequency fatigue.

In our elderly subjects, Q_tw fell to 64.0 ± 6.3% of the baseline value at 10 min after exercise. This decrease in Q_tw was not significantly different from that observed in a group of 13 younger subjects (in whom Q_tw fell to 54.8 ± 4.2% of its baseline value at 10 min after exercise) who performed cycle exercise at similar relative intensities (as a percentage of Wpeak) and to exhaustion as did our elderly subjects (31). Clearly, however, the absolute workload was markedly smaller in our elderly subjects (75.4 ± 6.8 W, compared with 167 ± 9.0 W) because of their reduced Wpeak. Nevertheless, the induction of similar degrees of fatigue at similar relative work intensities in our elderly subjects and a prior group of younger subjects suggests that quadriceps fatigability (when normalized for reductions in maximal capacity) is not dramatically enhanced in the elderly.

Diaphragmatic Function and Exercise: Critique of Methods

Cervical and bilateral anterolateral magnetic stimulation were used to detect diaphragmatic fatigue in this study. Both techniques have been successfully used to detect diaphragmatic fatigue in prior studies (17, 20, 25, 27, 28). In our laboratory, both techniques are highly reproducible. Our within-subject, between-trial (within a single day) CV in the absence of compound motor action potentials is 6.4 ± 0.5% for cervical magnetic stimulation and 5.1 ± 1.0% for anterolateral magnetic stimulation. Similar results have been obtained by other investigators (25).

Diaphragmatic Fatigue and Exercise

Pdi_tw was not significantly different from baseline at any time after exhausting high-intensity exercise in our group of healthy but relatively sedentary elderly subjects. Furthermore, we measured Pdi_twp, which may be a more sensitive index for detecting fatigue when the degree of fatigue is small (20, 21). Again, Pdi_twp was not significantly different from its baseline value at any time after exercise. We can therefore be confident that diaphragmatic fatigue did not occur.

Why did diaphragmatic fatigue not occur in our elderly subjects despite our own and others' observation of diaphragmatic fatigue after similar exercise in young adults (1)? One possibility is that our elderly subjects did not push themselves, and stopped exercise prematurely. However, at end-exercise, O₂peak was 109.3 ± 5.6% of the O₂peak obtained during the incremental maximal exercise test. Peak HR was 98.5 ± 1.6% of the predicted maximum, the respiratory quotient at end-exercise was 1.16 ± 0.04, lactate levels after exercise were 6.8 ± 0.6 mEq/L and our subjects developed significant contractile fatigue of their quadriceps msucle. These results show that our subjects had reached their physiologic limits when they stopped exercise.

Changes in respiratory mechanics with aging increase the work of breathing at any given workload (9). However, Wpeak steadily declines with age (13). The relative changes in these two factors could determine whether diaphragmatic fatigue occurs in the elderly. As expected, our elderly subjects had a significantly lower Wpeak than did our previously studied younger subjects, who displayed diaphragmatic fatigue after exercise (1, 2) (108 ± 10 W versus 216 ± 38 W in young sedentary subjects) (p < 0.001).

The pressure-time product and the tension-time index provide an estimate of inspiratory muscle work during exercise. Peak PTPes and Tes_t in our elderly subjects were significantly smaller than the values observed in young sedentary subjects who exhibited diaphragmatic fatigue after exercise (316 ± 29 cm H₂ O · s · min¹ versus 585 ± 52 cm H₂ O · s · min¹, respectively, p < 0.001; and 0.057 ± 0.005 versus 0.083 ± 0.009, respectively, p < 0.02) (1), and the differences were even greater when our elderly subjects were compared with a more active group of young subjects who also developed diaphragmatic fatigue after exercise (316 ± 29 cm H₂O · s · min¹ versus 804 ± 109 cm H₂O · s · min¹, respectively, p < 0.001; and 0.057 ± 0.005 versus 0.114 ± 0.008, respectively, p < 0.005) (2). Thus, our elderly subjects may not have developed diaphragmatic fatigue after exercise because the age-related decrease in Wpeak did not allow them to perform enough inspiratory muscle work to develop diaphragmatic fatigue after exercise. During voluntary hyperpnea, diaphragmatic fatigue typically occurs when the pressure-time product is > 550 to 600 cm H₂ O · s · min¹ (32). During exercise, however, diaphragmatic fatigue can develop at much lower pressure-time products (1, 32). It is believed that exercise-induced diaphragmatic fatigue depends not only on the amount of diaphragmatic work performed during exercise, but also on an unspecified exercise-dependent factor (32). This exercise-dependent factor was presumably also not of sufficient magnitude to induce diaphragmatic fatigue in our elderly subjects. In elderly patients with chronic obstructive pulmonary disease (COPD), diaphragmatic fatigue occurs very uncommonly after endurance exercise to symptomatic limits, despite relatively high pressure-time products (17, 33). However, it has recently been shown that structural adaptations occur in the diaphragm in response to COPD that increase its fatigue resistance (increase in the proportion of type 1 fibers, increase in mitochondrial content, and reduction in sarcomere length [34-36]).

With senescence, most humans (and animals) reduce their physical activity. In elderly subjects who are highly physically fit, the reduction in Wpeak will be less than typically observed, whereas the changes in respiratory mechanics that occur with aging are unaltered (11). Thus, highly fit elderly persons may be capable of doing sufficient work to produce diaphragmatic fatigue after exercise, although this remains to be shown. However, our results clearly show that the typical elderly subject does not develop diaphragmatic fatigue after exercise.

The effects of aging on respiratory muscle endurance have not been extensively investigated. In a human study, the pressure-time integral reached at task failure during threshold loading decreased with age (8). However, respiratory muscle strength also decreases with aging. Endurance time at equivalent target pressures (normalized for differences in strength) have been measured but not reported. In vitro studies have yielded conflicting results: increased fatigability with aging in the golden hamster (7) and no change in fatigability with aging in the Fischer rat (37). The fatigue paradigms, however, were different in the two studies. The oxidative capacity of the rat diaphragm does not appear to change with senescence, whereas it decreases in hindlimb muscles (29). Fiber types change minimally in the diaphragm with aging (37). Thus, aging appears to have a greater effect on the endurance properties of limb muscles than on those of the diaphragm. The diaphragm may be relatively resistant to the effects of aging on endurance because it is chronically active and because age-related changes in pulmonary mechanics increase its baseline level of activity. This difference in the effects of aging on endurance may help explain why our elderly subjects experienced quadriceps muscle but not diaphragmatic fatigue.

In conclusion, significant overt contractile fatigue of the quadriceps muscle follows endurance cycle exercise to exhaustion in healthy but sedentary elderly subjects, whereas diaphragmatic fatigue does not. Fatigue of the quadriceps muscle may be a major factor limiting exercise in the elderly.