Impact of age on the development of fatigue during large and small muscle mass exercise

Abstract

To examine the impact of aging on neuromuscular fatigue following cycling (CYC; large active muscle mass) and single-leg knee-extension (KE; small active muscle mass) exercise, 8 young (25 ± 4 years) and older (72 ± 6 years) participants performed CYC and KE to task failure at a given relative intensity (80% of peak power output). The young also matched CYC and KE workload and duration of the old (iso-work comparison). Peripheral and central fatigue were quantified via pre-/postexercise decreases in quadriceps twitch torque (∆Q_tw, electrical femoral nerve stimulation) and voluntary activation (∆VA). Although young performed 77% and 33% more work during CYC and KE, respectively, time to task failure in both modalities was similar to the old (~9.5 min; P > 0.2). The resulting ΔQ_tw was also similar between groups (CYC ~40%, KE ~55%; P > 0.3); however, ∆VA was, in both modalities, approximately double in the young (CYC ~6%, KE ~9%; P < 0.05). While causing substantial peripheral and central fatigue in both exercise modalities in the old, ∆Q_tw in the iso-work comparison was not significant (CYC; P = 0.2), or ~50% lower (KE; P < 0.05) in the young, with no central fatigue in either modality (P > 0.4). Based on iso-work comparisons, healthy aging impairs fatigue resistance during aerobic exercise. Furthermore, comparisons of fatigue following exercise at a given relative intensity mask the age-related difference observed following exercise performed at the same workload. Finally, although active muscle mass has little influence on the age-related difference in the rate of fatigue at a given relative intensity, it substantially impacts the comparison during exercise at a given absolute intensity.

INTRODUCTION

Investigations addressing the effects of aging on the development of neuromuscular fatigue, defined as a reversible decrease in the torque- or power-generating capacity of a muscle or muscle group (20, 45), have traditionally focused on single-joint exercise performed at a given relative intensity (4). This form of exercise modality only involves a small skeletal muscle mass and is not limited by cardiopulmonary constraints (3). In contrast, exercise involving a large muscle mass, such as cycling (CYC), evokes substantial cardiopulmonary responses that are unfavorably altered with advancing age. For example, age-related cardiovascular limitations (17, 19, 50) and increased ventilatory work (30) have the potential to limit leg blood flow and oxygen delivery during exercise (16, 30), thereby exacerbating the development of neuromuscular fatigue (16, 18, 21). Consequently, by only utilizing single-joint exercise, the current literature may, unintentionally, be underestimating the influence of age on the development of fatigue by experimentally minimizing the role of age-related impairments in the cardiopulmonary system. Furthermore, the vast majority of studies have focused on comparisons at a given relative exercise intensity, an experimental approach that minimizes the impact of aging on functional aerobic capacity and potentially obscures the practical significance (i.e., activities of daily living) of age-related changes in fatigue (4). Therefore, to fully appreciate the impact of aging on the development of fatigue, it would be most appropriate to examine both small and large muscle mass exercise at a given relative and absolute intensity.

Exercise-induced neuromuscular fatigue is caused by a decrease in neural activation of the muscle (i.e., central fatigue) and/or biochemical changes at or distal to the neuromuscular junction that causes an attenuated contractile response to neural input [i.e., peripheral fatigue (7)]. Furthermore, as muscle activation is critically dependent on descending neural drive, the integrity of the motor pathway linking the brain with the exercising skeletal muscle, termed the corticospinal pathway (including the motor cortex and spinal motoneurons), is considered an important contributor to central fatigue. To this end, transcranial magnetic stimulation (TMS) can be used to monitor alterations in the excitability of the corticospinal pathway of a particular muscle during exercise (15). To decipher alterations in corticospinal excitability at the motoneuronal level, cervicomedullary stimulation (CMS) of corticospinal axons can be employed. Additionally, changes in motor cortical excitability can be examined when short-latency electromyographic (EMG) responses from TMS [motor-evoked potential (MEP)] are normalized to those evoked from CMS [cervicomedullary MEPs (CMEP) (46)]. However, such a comprehensive assessment of fatigue with advancing age, in combination with both small and large muscle mass at an absolute and relative exercise intensity, has yet to be performed.

Consequently, the aim of this study was to elucidate the impact of aging on the development of fatigue during rhythmic, aerobic exercise involving a large (CYC) and a small [single-leg knee-extension (KE)] muscle mass at both absolute and relative exercise intensities. It should be emphasized that dynamic KE is characterized by both a cardiac and pulmonary reserve capacity at maximal exercise intensity (18), which means smaller cardiopulmonary constraints on leg blood flow and O₂ delivery compared with CYC (18, 30, 39). It was therefore hypothesized that, during aerobic exercise at a given absolute intensity and for a given duration, 1) aging would exacerbate the development of fatigue, and 2) the amount of muscle mass involved in the exercise would influence the difference in end-exercise fatigue between young and old, with a greater difference occurring after CYC compared with KE. Furthermore, when normalizing for systemic age-related changes, by utilizing exercise at a given relative intensity, end-exercise neuromuscular fatigue would be unaffected by age.

METHODS

Participants

Sixteen healthy, recreationally active men [8 young (25 ± 4 years) and 8 old (72 ± 6 years)] volunteered for this study. Subject characteristics are presented in Table 1. Young and old participants were well matched for physical activity, anthropometric measures, and hematologic characteristics. Written, informed consent was obtained from all participants before their inclusion in the study. The Institutional Review Boards of the University of Utah and the Salt Lake City Veterans Affairs Medical Center approved all protocols.

Table 1.

Descriptive characteristics for young and old participants

	Young	Old
Age, yr	25 ± 4	70 ± 6*
Height, cm	179 ± 8	176 ± 5
Mass, kg	81 ± 11	82 ± 15
BMI, kg/m2	25 ± 2	27 ± 4
Quadriceps muscle mass, kg	2.0 ± 0.3	1.9 ± 0.6
Physical activity
Sedentary	1,202 ± 57	1,190 ± 75
Light	141 ± 58	104 ± 37
Moderate	59 ± 28	66 ± 25
Vigorous	11 ± 13	7 ± 13
Steps/day	8,716 ± 1,106	8,932 ± 3,471
Cycling V̇o2max, ml·kg−1·min−1	44.9 ± 2.3	27.9 ± 1.8*
Hematologic characteristics
RBC, M/µl	5.3 ± 0.3	5.0 ± 0.3
Hemoglobin, g/dl	15.8 ± 1.2	15.6 ± 0.7
Hematocrit, %	46 ± 3	46 ± 4
Cholesterol, mg/dl	160 ± 18	179 ± 30
Triglycerides, mg/dl	80 ± 32	70 ± 21
HDL, mg/dl	44 ± 8	49 ± 6
LDL, mg/dl	101 ± 17	109 ± 35

Values are means ± SD. BMI, body mass index; HDL, high-density lipoprotein; LDL, low-density lipoprotein; RBC, red blood cells; V̇o_2max, maximal oxygen consumption.

^*Significant difference between groups, P < 0.05.

Experimental Procedures

All participants completed protocol A before protocol B (see below). Across six preliminary visits (three per protocol), participants were familiarized with all experimental testing procedures. Each participant then performed the protocol-specific experimental trial. The young then returned for an additional visit in which they matched the absolute exercise intensity and duration as the old. This equated to a total of 8 visits for the old and 10 visits for the young. All exercise sessions followed a standardized, task-specific warm-up and were separated by at least 24 h. Quadriceps fatigue was quantified by the pre- to postexercise decrease in neuromuscular function (see Neuromuscular quadriceps function and torque acquisition). To quantify exercise-induced changes in corticospinal excitability, participants received three stimulation sets before and immediately after exercise. Each stimulation set consisted of three TMS, one CMS, and one femoral nerve stimulation (FNS). The stimulations were randomized and triggered during a constant quadriceps contraction equating to 20% of the EMG activity recorded during preexercise maximal voluntary contractions (MVCs) (Fig. 1).

Fig. 1.

Schematic illustration of the exercise protocols. Neuromuscular and corticospinal data were collected before and after both single-leg knee-extensor (A) and cycling (B) exercise. C, cervicomedullary stimulation; EMG, electromyographic; F, femoral nerve stimulation; MVC, maximal voluntary contraction; T, transcranial magnetic stimulation; 20% EMG contraction, 20% of the EMG associated with the preexercise MVC.

Protocol A: dynamic KE.

The experimental protocol is depicted in Fig. 1A. Before all baseline neuromuscular measures and exercise sessions, participants performed a 1-min unloaded warm-up on a custom-made knee-extensor ergometer. The first preliminary visit consisted of a maximal incremental knee-extensor exercise test (range of motion: 90°–170° of knee angle). Subjects maintained a consistent cadence at 60 revolutions/min throughout all testing procedures. Peak power output (W_peak) and oxygen consumption (V̇o_2peak) were determined by increasing the workload from an unloaded state by 5–10 W/min until task failure (revolutions/min below 55 for >5 s, despite verbal encouragement). W_peak represents the workload during the last stage completed, whereas V̇o_2peak represents the average O₂ consumption during the last minute of exercise. During the second and third visit, participants performed familiarization trials at 80% W_peak to task failure (T_lim). During the fourth visit, considered the experimental day, participants performed the actual T_lim trial where exercise-induced quadriceps fatigue and changes in corticospinal excitability were assessed. For comparisons of fatigue following exercise performed at the same absolute workload and for the same duration (i.e., iso-work), the young participants returned for a fifth visit to match the power output (Watts) and exercise time of the old participants on a person-by-person basis, following the W_peak rank order. Thus, for example, the young participant with the highest W_peak matched the performance (power output and exercise time) of the old participant with the highest W_peak. Neuromuscular function and corticospinal excitability were assessed before and after exercise.

Protocol B: CYC exercise.

The experimental protocol is depicted in Fig. 1B. Before all baseline neuromuscular measures and exercise sessions, participants performed a warm-up on a cycle ergometer (Velotron, Elite model, Racer Mate, Seattle, WA) consisting of 3 min of exercise at 25, 50, and 75 W (9 min total). Subjects were asked to maintain a consistent cadence at 70 rpm throughout all testing procedures. During the first visit, all participants performed an incremental CYC test (20 W + 25 W/min) to task failure (revolutions/min below 65 for >5 s, despite verbal encouragement). W_peak and V̇o_2peak were determined as described above. The goals of visits 2–5 for this protocol were similar, but with the different exercise modality, to those in protocol A (see above).

General Instrumentation

Physical activity level.

Once instructed on proper operating procedures, participants wore an accelerometer (GTIM, Actigraph, Pensacola, FL) for 10 consecutive days. Daily physical activity was assessed both as steps per day and physical activity counts per minute. The counts per minute was subsequently categorized into sedentary, low, moderate, and vigorous intensities using commercially available software (Actilife, Actigraph).

Determination of quadriceps muscle mass and hematology.

Before any exercise, quadriceps muscle volume was estimated by anthropometric measurements, as described by Jones and Person (24), while applying a recommended correction factor for the determination of quadriceps muscle mass (3). Additionally, blood was collected from the antecubital vein to assess blood lipids, fasting glucose, and hemoglobin.

Electromyography.

The EMG signals were recorded by surface electrodes (Ag-AgCl, 10-mm diameter) placed on the muscle belly and tendon of the vastus lateralis (VL) in a monopolar configuration. Prior to electrode placement, the skin was lightly abraded with fine sandpaper and cleaned with an alcohol swab. EMG signals were amplified 1,000 times (Neurolog Systems, Digitimer, Welwyn Garden City, Hertfordshire, UK), band-pass filtered (20–1,000 Hz; NL-844, Digitimer), and analog to digitally converted at a sampling rate of 2,000 Hz using a 16-bit Micro 1401 mk-II and Spike 2 data collection software (Cambridge Electronic Design, Cambridgeshire, UK) running custom written scripts. The right-knee angle was monitored continuously during dynamic knee-extensor (string potentiometer) and CYC exercise (crank encoder with a timing belt). During a brief (~45 s) bout of exercise at the experimental workload, the knee angle at which the peak EMG burst of the VL occurs (i.e., the centering point) was determined. For EMG analysis during exercise, 10 s of rectified EMG waveforms were overlaid around the crank angle centering point and averaged (48). EMG was averaged over a 100-ms period, i.e., 50 ms before and 50 ms after the centering point, to provide the average knee extensor EMG at each time point (48). EMG was normalized to that obtained during the first minute of exercise.

Neuromuscular quadriceps function and torque acquisition.

To examine exercise-induced fatigue, measures of neuromuscular quadriceps function of the right leg were assessed before and immediately after exercise. For the assessment of neuromuscular function, participants were seated upright in a custom-made chair with their hip and knee flexed at 120° and 90°, respectively. The time between the end of exercise and the beginning of the postexercise assessment of neuromuscular function was similar between KE and CYC exercise (KE: young, 38 ± 13 s; old, 38 ± 8 s; CYC: young, 35 ± 7 s; old, 42 ± 14 s; P > 0.20 for all comparisons). A noncompliant cuff was attached to a calibrated load cell (MLP 300, Transducer Techniques, Temecula, CA) 2–3 cm superior to the lateral malleolus. Cuff position remained constant throughout each session. The assessment of quadriceps function included three MVCs, over which an FNS was performed. If muscular activation was suboptimal during the MVC, the FNS would evoke a superimposed twitch. After each MVC, another FNS was initiated to evoke a resting potentiated quadriceps twitch (Q_tw). Voluntary quadriceps activation (VA) was calculated as VA = [1 − (superimposed twitch/Q_tw)] × 100 (6, 29). Exercise-induced peripheral and central fatigue were quantified as the percent reduction of the averaged Q_tw and VA, respectively, from before to after exercise. Furthermore, the maximal rate of torque development (MRTD, calculated as the highest positive derivative of the torque during a 10-ms interval) and peak relaxation rate (PRR, the highest negative derivative of the torque during a 10-ms interval) were analyzed for each Q_tw.

Femoral nerve stimulation.

In all protocols, the motor nerve was stimulated with the anode placed between the greater trochanter and the iliac crest and the cathode placed over the femoral nerve in the femoral triangle. Optimal position (i.e., greatest twitch force) for the stimulating electrode was determined by delivering low-intensity single-pulse stimuli (200-µs pulse width; 100–150 mA) via a movable cathode probe and a constant current stimulator (voltage range: 100–400 V; model DS7AH, Digitimer). Once located, a self-adhesive 3-cm round cloth cathode was fixed and remained in this position for the remainder of the session. Thereafter, stimulation intensity was increased by 20-mA increments until the size of the evoked twitch and compound muscle action potential (M-wave) demonstrated no further increase (i.e., area of maximal M-wave, M_max) at rest, and this was further confirmed during a 50% MVC. Stimulation intensity was set at 130% of M_max intensity and kept constant throughout the testing session (pooled mean; young: 322 ± 72 mA; old: 416 ± 137 mA).

Cervicomedullary stimulation.

The corticospinal pathway was electrically stimulated at the level of the cervicomedullary junction. Self-adhesive electrodes (3-cm round cloth) were placed on the grooves behind the mastoid processes with the cathode placed on the left (100-µs pulse width, D-185 mark IIa, Digitimer) (47). Stimulator intensity was set to achieve a CMEP of ~30% M_max during a constant quadriceps contraction at an intensity corresponding to 20% of the EMG obtained during a quadriceps MVC (young: 458 ± 144 V; old: 495 ± 108 V). This procedure provides room for either an exercise-induced increase or decline in corticospinal responses (41). In protocol B, only four of the eight old individuals tolerated CMS.

Transcranial magnetic stimulation.

Stimuli were delivered over the motor cortex using a concave double-cone coil (Magstim 200; Magstim, Whitland, UK) to elicit an MEP in the quadriceps during all protocols. Optimal positioning of the TMS coil (~2–3 cm to the left of the vertex with a posterior to anterior direction of current flow in the motor cortex) was determined before the experiment and marked on the scalp for accurate placement throughout the study. To test a similar portion of the motoneuron pool, the stimulator intensity was set to evoke MEPs of similar size to the CMEP obtained during quadriceps contractions at an intensity corresponding to 20% of the EMG obtained during a quadriceps MVC (41) (young: 56 ± 13% of stimulator output; old: 53 ± 11% of stimulator output). In those individuals who did not tolerate CMS, TMS intensity was set to evoke an MEP of ~30% M_max.

Protocol Specific Instrumentation

Protocol A: dynamic KE.

ventilation, pulmonary gas exchange, cardiovascular, and perceived exertion measures.

Ventilation and pulmonary gas exchange were measured with a metabolic cart (Innocor, Innovision, Odense, Denmark). Heart rate (HR) was measured via the R-R interval derived from a 12-lead electrocardiogram acquired with a data acquisition system (Spike 2, Cambridge Electronic Design). Mean arterial pressure (MAP) was determined via a Finometer utilizing finger photoplethysmography (Finapres Medical Systems, Amsterdam, The Netherlands). Stroke volume was estimated by the beat-by-beat assessment of the pressure waveform using the Modelflow method (Beatscope, version 1.1, Finapress Medical Systems). Cardiac output (CO) was calculated as the product of HR and stroke volume. Systemic vascular conductance was calculated as the quotient of CO and MAP. Participant’s rating of perceived exertion was obtained at the end of each minute during exercise using Borg’s Category Ratio 10 scale (9).

leg blood flow.

Common femoral artery blood flow (FBF) measurements were collected at rest and during exercise (for visits 1 and 5 of protocol A) utilizing a Logic 7 ultrasound (General Electric Medical Systems, Milwaukee, WI), as previously described (50). Leg vascular conductance (LVC) was calculated as FBF/MAP.

Protocol B: CYC exercise.

ventilation, pulmonary gas exchange, cardiovascular, and perceived exertion measures.

Ventilation, pulmonary gas exchange, HR, and rating of perceived exertion were measured as described in protocol A. Based on the time-to-task-failure during the final practice trial, CO was measured at 90% of the T_lim completion time using an inert gas rebreathing technique (Innocor, Innovision) (35).

Data Analysis

All data were stored and analyzed offline using Spike2 data acquisition software. The area for each evoked response (MEP, CMEP, and M_max) was measured and averaged over the three stimulation sets. To account for potential changes within the peripheral axons and/or the muscle, MEPs and CMEPs were normalized to M_max. To isolate alterations in the excitability of the motor cortex, MEP was normalized to CMEP (28). All data are reported as means ± SD.

Statistical Analysis

The overall statistical model for this study was a three-way ANOVA comparing neuromuscular function over time (pre- vs. postexercise), between groups (young vs. old), and between modalities (KE vs. CYC). However, to directly test the hypotheses regarding the development of fatigue, the following a priori planned comparisons were made: 1) pre- versus postexercise neuromuscular function in the young and old for each modality (KE and CYC) and exercise intensity (absolute and relative), 2) the exercise-induced change in neuromuscular function between young and old, and 3) the exercise-induced change in neuromuscular function between exercise modalities. For comparison 1, paired t-tests were performed. To control the Type I error rate, within a family of comparisons, we adjusted the alpha using the Holm-Bonferroni correction. Comparisons 2 and 3 were analyzed using two-way mixed model ANOVAs for the exercise-induced change in neuromuscular function. For the remaining variables, two-way mixed model ANOVAs (group by workload and group by time) were performed for the comparison of FBF during the maximal KE test and T_lim trials, respectively. If an ANOVA indicated a main effect, a Holm-Sidak post hoc test was performed to identify the differences. Additionally, Student’s t-tests were employed to compare descriptive characteristics, pre- and postchanges in twitch mechanics (MRTD, PRR), corticospinal excitability (M_max, MEP, CMEP, and MEP/CMEP), and end-exercise (last minute) metabolic and ventilatory data. α was set at 0.05.

RESULTS

Baseline Neuromuscular Function

Baseline neuromuscular function was similar between KE and CYC visits in the young (MVC: 256 ± 46 Nm and 266 ± 49 Nm, P = 0.55; Q_tw: 68 ± 19 Nm and 74 ± 9 Nm, P = 0.35; VA: 92 ± 4% and 93 ± 3%, P = 0.66; MRTD: 1,410 ± 241 Nm·s and 1,316 ± 116 Nm·s, P = 0.34; PRR: 519 ± 56 Nm·s and 528 ± 81 Nm·s, P = 0.82) and old (MVC: 159 ± 21 Nm and 178 ± 31 Nm, P = 0.23; Q_tw: 47 ± 7 Nm and 52 ± 9 Nm, P = 0.36; VA: 95 ± 5% and 94 ± 3%, P = 0.69; MRTD: 965 ± 267 Nm/s and 980 ± 240 Nm/s, P = 0.88; PRR: 384 ± 105 Nm·s and 389 ± 155 Nm·s, P = 0.92). Although VA was not different between groups (P = 0.12), baseline MVC, Q_tw, MRTD, and PRR were all diminished in the old (P < 0.05).

Protocol A: Dynamic Knee Extension

Incremental exercise test.

Knee-extensor W_peak (68 ± 8 W and 55 ± 6 W, P < 0.05) and peak FBF (4,217 ± 538 ml/min vs 3,577 ± 630 ml/min, P < 0.05) were higher in the young than the old group. Additionally, although there was a significant effect of time (P < 0.05), there was no group effect (P = 0.13). Specifically, there was no difference between young and old subjects across unloaded (~1,460 ml/min), 10 W (~1,830 ml/min), 20 W (~2,300 ml/min), or 40 W (~2,910 ml/min) of KE exercise.

Relative exercise intensity.

Calculated coefficients of variation for the time-to-task-failure across visits 2–4 were 11% and 8% in the young and old participants, respectively. Although time to task failure was similar between young and old participants (10.8 ± 2.8 min and 10.1 ± 2.8 min, respectively; P = 0.57), the young completed more work during the task (36 ± 12 kJ vs. 27 ± 10 kJ, P < 0.05). VL-EMG increased similarly (~90% from first minute) throughout the exercise in both groups (P = 0.61). There was a significant group (P < 0.05) and time (P < 0.05) effect for FBF and LVC (Fig. 2). Other cardiovascular and metabolic data are documented in Table 2. Although both MVC and VA fell significantly more in the young compared with the old participants, Q_tw declined to a similar extent in both groups (P = 0.69, Fig. 3). In addition, the exercise-induced decreases in MRTD (young: 52 ± 18% vs. old: 49 ± 20%; P = 0.69) and PRR (young: 50 ± 18% vs. old: 40 ± 11%; P = 0.20) were similar in both groups.

Fig. 2.

Femoral blood flow (A and B) and leg vascular conductance (C and D) in young and old participants during rhythmic knee extensor exercise performed at the same relative (A and C) or absolute (B and D) exercise intensity. Data are means ± SD.

Table 2.

Cardiopulmonary and metabolic responses during the final minute of exercise

	Young	Old	Young Matched
Power output, W
KE	54 ± 8*	45 ± 6	45 ± 6
CYC	249 ± 57*	155 ± 23	155 ± 23
Exercise time, min
KE	10.8 ± 2.8	10.1 ± 2.8	10.0 ± 2.8
CYC	9.1 ± 1.7	8.4 ± 1.4	8.4 ± 1.4
V̇e, l/min
KE	84 ± 13	77 ± 20	41 ± 6*
CYC	164 ± 21*	124 ± 23	73 ± 14*
O2, l/min
KE	1.55 ± 0.25	1.46 ± 0.16	1.12 ± 0.20*
CYC	3.68 ± 0.70*	2.38 ± 0.42	2.51 ± 0.45
V̇o2, kg·ml−1·min−1
KE	20.8 ± 3.1	18.3 ± 2.2	15.1 ± 0.3*
CYC	45.5 ± 5.7*	29.3 ± 5.1	31.0 ± 7.0
V̇co2, l/min
KE	1.86 ± 0.24*	1.57 ± 0.17	1.13 ± 0.20*
CYC	4.15 ± 0.79*	2.59 ± 0.48	2.38 ± 0.45
RER
KE	1.22 ± 0.11*	1.08 ± 0.11	1.01 ± 0.0.9
CYC	1.13 ± 0.08	1.10 ± 0.11	0.94 ± 0.06*
V̇e/V̇o2
KE	51 ± 7	50 ± 12	33 ± 3*
CYC	43 ± 3*	51 ± 11	27 ± 3*
V̇e/V̇co2
KE	42 ± 4	46 ± 9	33 ± 3
CYC	38 ± 3*	46 ± 8	29 ± 3*
HR, beats/min
KE	138 ± 18	129 ± 11	112 ± 11*
CYC	183 ± 8*	150 ± 14	141 ± 16
SV, ml/beat
KE	77 ± 27	83 ± 33	91 ± 30
CYC	114 ± 17*	92 ± 17	122 ± 17*
CO, l/min
KE	10.5 ± 3.7	10.4 ± 3.4	10.0 ± 2.9
CYC	20.3 ± 3.4	15.0 ± 2.5	17.1 ± 2.8
MAP, mmHg
KE	148 ± 19	147 ± 27	129 ± 14*
CYC
SVC, ml·min−1·mmHg−1
KE	72 ± 30	75 ± 33	79 ± 26
CYC
RPE
KE	10.0 ± 0.0	9.9 ± 0.3	7.0 ± 1.1*
CYC	9.9 ± 0.3	9.9 ± 0.3	5.1 ± 2.5*

Data are means ± SD. CO, cardiac output; CYC, cycling exercise protocol; HR, heart rate; KE, single-leg dynamic protocol; MAP, mean arterial pressure; RER, respiratory exchange ratio; RPE, rating of perceived exerction; SV, stroke volume; SVC, systemic vascular conductance; V̇E, minute ventilation; V̇e/V̇co₂, ventilatory equivalent for CO₂; V̇e/V̇o₂, ventilatory equivalent for O₂; V̇co₂, carbon dioxide production; V̇o₂, oxygen consumption.

^*Significantly different from old; P < 0.05.

Fig. 3.

Pre- to postchange in neuromuscular function induced by single-leg knee extensor exercise (KE, open bars) and cycling exercise (CYC, filled bars) performed at a given relative (left) or absolute (right) exercise intensity. MVC, maximal voluntary contraction; VA, voluntary activation. Data are presented as means ± SD. #Not significantly changed from preexercise; *significantly different from old.

Although M_max was greater in the young (P < 0.05), pre- and postexercise M_max were not different in both the young (41.5 ± 0.4 μV·s vs. 40.9 ± 0.4 μV·s, P = 0.78) and the old (31.8 ± 1.4 μV·s vs. 31.6 ± 1.4 μV·s, P = 0.98). Furthermore, normalized MEP (36 ± 16% M_max vs. 39 ± 19% M_max, P = 0.68), normalized CMEP (31 ± 17% M_max vs. 34 ± 23% M_max, P = 0.75), and MEP/CMEP (130 ± 67% vs. 131 ± 49%, P = 0.99) were unchanged from baseline after exercise in the young. Similarly, normalized MEP (36 ± 13% M_max vs. 37 ± 14% M_max, P = 0.90), normalized CMEP (25 ± 10% M_max vs. 23 ± 8% M_max, P = 0.64), and MEP/CMEP (131 ± 18% vs. 147 ± 33%, P = 0.87) were unchanged from baseline after exercise in the old and were not different from the young (P > 0.11).

Absolute exercise intensity.

By design, KE exercise time and work performed were matched between groups. VL EMG in the young remained unchanged throughout the exercise (first minute to last minute, P = 0.11) while the signal increased by ~100% from the first minute to end exercise (P < 0.05) in the old. Although there was an age by time interaction effect for FBF (P < 0.05), there was no group (P = 0.59) or interaction (P = 0.31) effects for LVC (Fig. 2). Other cardiovascular and metabolic data are documented in Table 2. There was a group effect for MVC, Q_tw, and VA (P < 0.05, Fig. 3), with the post hoc analysis indicating that end-exercise fatigue in the young was less than that of the old.

M_max was unaltered by exercise in the young (43.6 ± 0.4 μV·s vs. 43.4 ± 0.3 μV·s, P = 0.91) and still greater than that of the old (P < 0.05). Furthermore, normalized MEP (36 ± 16% M_max vs. 36 ± 13% M_max, P = 0.95), normalized CMEP (34 ± 10% M_max vs. 33 ± 10% M_max, P = 0.94), and MEP/CMEP (94 ± 26% vs. 99 ± 27%, P = 0.69) were unchanged from pre- to postexercise in the young and not different from the old (P > 0.11).

Protocol B: CYC Exercise

Incremental exercise test.

W_peak was higher in the young compared with the old (309 ± 71 W and 195 ± 28 W, respectively; P < 0.05). Additionally, V̇o_2peak was significantly greater in the young compared with the old (3.64 ± 0.73 l/min vs. 2.26 ± 0.39 l/min, respectively; P < 0.05).

Relative exercise intensity.

Calculated coefficients of variation reflect between-day variability for time to task failure in both the young (10%) and old (8%) participants. Although time to task failure was similar between young and old (9.1 ± 1.7 min vs. 8.4 ± 1.4 min, respectively; P = 0.44), the young completed more total work during the exercise (140 ± 67 kJ vs. 79 ± 21 kJ, respectively; P < 0.05). VL-EMG increased similarly (~30% from the first minute) throughout the exercise in both groups (P = 0.68). Other cardiovascular and metabolic data are shown in Table 2. Although both MVC and VA fell significantly more in the young compared with the old participants, Q_tw declined to a similar extent in both groups (P = 0.69, Fig. 3). In addition, the exercise-induced decreases in MRTD (young: 39 ± 27% vs. old: 39 ± 26%; P = 0.99) and PRR (young: 39 ± 25% vs. old: 34 ± 37%; P = 0.72) were similar in both groups.

Although M_max was greater in the young (P < 0.05), pre- and postexercise M_max were not different in both the young (40.6 ± 0.5 μV·s vs. 38.9 ± 0.7 μV·s, P = 0.58) and the old (34.0 ± 0.9 μV·s vs. 31.4 ± 0.5 μV·s, P = 0.50). Furthermore, normalized MEP (26 ± 9% M_max vs. 27 ± 9% M_max, P = 0.81), CMEP (25 ± 11% M_max vs. 23 ± 6% M_max, P = 0.71), and MEP/CMEP (102 ± 24% vs. 117 ± 41%, P = 0.41) were unchanged from baseline after exercise in the young. Similarly, normalized MEP (29 ± 10% M_max vs. 30 ± 18% M_max, P = 0.80), normalized CMEP (26 ± 8% M_max vs. 27 ± 9% M_max, P = 0.88), and MEP/CMEP (102 ± 6% vs. 103 ± 17%, P = 0.94) were unchanged from baseline after exercise in the old and were not different from the young (P > 0.33).

Absolute exercise intensity.

By design, CYC exercise time and work performed were matched between groups. Cardiovascular and metabolic data are documented in Table 2. VL-EMG in the young remained unchanged throughout exercise (P = 0.21) but increased by 33 ± 16% (P < 0.05) in the old. Although both MVC and Q_tw decreased more in the old (P < 0.05), VA was not different between groups (P = 0.11; Fig. 3). Importantly, various neuromuscular function indices remained unchanged from before to after the exercise in the young (P > 0.2).

M_max was unaltered by exercise in the young (41.9 ± 0.5 μV·s vs. 40.6 ± 0.7 μV·s, P = 0.73) and still greater than that of the old (P < 0.05). Furthermore, normalized MEP (24 ± 4% M_max vs. 26 ± 7% M_max, P = 0.90), normalized CMEP (24 ± 8% M_max vs. 26 ± 9% M_max, P = 0.56), and MEP/CMEP (104 ± 20% vs. 108 ± 44%, P = 0.84) were unchanged from pre- to postexercise in the young and not different from the old (P > 0.14).

Impact of exercise modality on fatigue.

There was a main effect in both groups for exercise modality (P < 0.05), with a larger decline in MVC, Q_tw, and VA at task failure following KE compared with CYC exercise. Although the old demonstrated substantial peripheral and central fatigue in both exercise modalities (Fig. 3), when the young repeated the KE performance of the old (i.e., iso-work comparison), the exercise-induced decreases in MVC and Q_tw were significant, but only ~50% of that exhibited by the old (P < 0.05). However, when the young repeated the CYC performance of the old (i.e., iso-work comparison), MVC and Q_tw were not altered by the exercise (P = 0.2, Fig. 3). Furthermore, the iso-work trial did not alter VA in either modality in the young (P > 0.4).

DISCUSSION

This study was designed to compare the development of fatigue during large and small muscle mass exercise in physically active young and old individuals. The primary conclusion of this investigation was that, based on iso-work comparisons, aging impairs fatigue resistance during aerobic exercise which may be masked by comparisons of fatigue following exercise at a given relative intensity. This conclusion is based upon two observations: First, central and peripheral fatigue following exercise at a given workload and duration (i.e., iso-work) was significantly greater in the old compared with their younger counterparts. Second, and in contrast to the first observation, at a given relative intensity, which fails to recognize the greater amount of work performed by the young, there is little difference in fatigue resistance between young and old. Indeed, this latter approach offsets or, in the case of VA, even reverses the interpretation of the impact of aging on the development of fatigue. Finally, these findings also suggest that the amount of active muscle mass can influence the magnitude of the age-related difference in end-exercise neuromuscular fatigue, such that the compromised fatigue resistance in the old is more pronounced following large muscle mass than small muscle mass exercise.

Aging Exacerbates the Development of Fatigue During Iso-work Aerobic Exercise

The comparison of fatigue between young and old participants following exercise at a given power output and duration (i.e., iso-work) revealed an exaggerated development of both central and peripheral fatigue during KE and CYC exercise in the old (Fig. 3). For the comparisons during KE, age-related differences in muscle O₂ delivery, a key determinant of the development of fatigue during exercise (1), might be excluded as a factor contributing to this discrepancy. Specifically, both the hyperemic response to KE and [Hb], and, therefore, likely O₂ delivery (2, 18), was comparable between the groups (Fig. 2, Table 1). Although this similarity in exercise-induced leg blood flow in the young and old contradicts some studies (17, 43) but confirms other (34) prior studies, this was not completely unexpected, as the participants were, by design, engaged in regular physical activity, a factor that can prevent the lower hyperemic response to exercise often documented the elderly (5, 31). However, leg blood flow was not measured during the CYC exercise. Therefore, age-related differences in O₂ delivery cannot be ruled out, as the exercise-induced hyperemia may have been lower in the elderly (37), potentially contributing to the greater fatigue during this exercise modality in the old.

Aging decreases CYC efficiency (40), increases the metabolic cost of muscle contractions (25, 26), and shifts the breakpoint for accelerated intramuscular perturbations to an earlier absolute work rate (11), which together likely account for the faster accumulation of metabolites (i.e., Pi and H⁺) at a given absolute workload in the old (13, 49). Since these metabolites compromise skeletal muscle cross-bridge function (8, 14, 32) similarly in young and old (44), the greater accumulation in peripheral fatigue observed in the old is likely attributed to an exaggerated metabolic perturbation during the iso-work exercise compared with the young. The larger exercise-induced intramuscular metabolic perturbation in the old may also have contributed to the greater central fatigue, which is known to be influenced by the central projection of metabosensitive group III/IV muscle afferents (42).

Normalizing for Aerobic Capacity Masks the Impact of Aging on the Development of Fatigue

Although the iso-work comparisons clearly reveal a negative impact of age on fatigue resistance, the observations following exercise at an intensity normalized for the age-related decline in maximal aerobic capacity (i.e., the same relative intensity) distort this conclusion. Specifically, time to task failure during both CYC and KE performed at the same relative intensity was nearly identical between young and old participants, and the resulting degree of end-exercise peripheral fatigue was similar between the two age groups (Fig. 3). Furthermore, central fatigue and consequently the exercise-induced decrease in MVC were actually greater in the young following both exercise modalities (Fig. 3). These observations are, at least at first sight, deceiving, insinuating that aging may not compromise the development of fatigue during aerobic exercise. However, it is critical to recognize that these observations are based on exercise of the same duration performed at different workloads. Specifically, compared with the old, young individuals completed nearly 80% more work during CYC and over 30% more work during KE. Considering these substantial differences in work performed between young and old, the similar degree of peripheral fatigue actually supports, although indirectly, the conclusion based on the iso-work comparison and further emphasizes the negative impact of aging on fatigue.

Despite similar peripheral fatigue and consequently similar metabolic perturbation in both groups at a given relative exercise intensity, the old participants incurred less central fatigue. This difference is potentially explained by the influence of aging on two key determinants of central fatigue, namely, neural feedback mediated by group III/IV muscle afferents and neural feedforward (i.e., corollary discharge linked with central motor drive) (23). Specifically, aging decreases the central impact of metabosensitive group III/IV muscle afferents, resulting in a decreased perception of fatigue (22) and likely lower central fatigue in response to a given level of intramuscular metabolic perturbation. Furthermore, central motor drive, which rises proportionally with increases in workload (48), was, given the substantially lower power output at a given relative exercise intensity during both modalities in the old, presumably lower in the old and may have additionally contributed to the diminished central fatigue exhibited by the old.

Active Muscle Mass Magnifies the Impact of Aging on the Development of Fatigue

The comparison of neuromuscular fatigue between the two age groups following CYC and KE at a given workload indicates that the amount of active muscle mass, which is larger during CYC, influences the magnitude of the difference in the development of fatigue between young and old. This is reflected by the observation that both KE and CYC caused a large decrease in Q_tw and MVC in the old, whereas in the young, fatigue following KE was only approximately half of that exhibited by the old and not present following CYC (i.e., no significant pre- to postexercise changes in neuromuscular function; Fig. 3). These findings suggest that the difference in the development of fatigue between young and old individuals is greater following exercise involving a large muscle mass compared with exercise recruiting only a small muscle mass. Interestingly, this is not the case when comparisons of fatigue are made following exercise performed at the same relative intensity (i.e., normalized for aerobic capacity). In this scenario, the magnitude of the difference in fatigue between young and old is similar following CYC and KE (Fig. 3).

The larger difference in the rate of fatigue development between young and old during CYC compared with KE may be related to the negative impact of aging on the cardiopulmonary system (17, 19, 30, 50) and associated limitations in terms of leg blood flow and O₂ delivery (16, 21). As the cardiopulmonary system is more stressed/engaged during large compared with small muscle mass exercise (18, 39), the age-related impairment of the cardiopulmonary system and subsequent limitations to leg blood flow and fatigue resistance (16, 21) may, therefore, be more apparent during CYC compared with KE at a given workload. This hypothesis is indirectly supported by the fact that leg blood flow during KE was similar in the two age groups (Fig. 2, C and D). However, as noted above, leg blood flow was not quantified during CYC and may have been attenuated (37), or unchanged, in the old (36).

Impact of Aging on Corticospinal Excitability During Exercise

The transmission of neural drive from higher brain areas to contracting muscle occurs predominately through the corticospinal pathway (10). Exercise-induced alterations in the corticospinal excitability, and therefore the efficacy of this motor pathway to relay neural signals, have the potential to influence skeletal muscle activation and central fatigue (27, 42). Interestingly, although not unanimously agreed upon (12), aging has been associated with a diminished corticospinal excitability, potentially influenced by age-related alterations at the motor cortical (33) and motoneuronal level (38). However, the lack of a change in MEPs and CMEPs during exercise suggests that fatiguing aerobic exercise tasks used in the study do not affect the excitability of the corticospinal pathway in either age group and, therefore, confirms earlier observations based on nonfatiguing exercise (12). Consequently, aging may not alter the influence of aerobic exercise on corticospinal excitability, excluding the efficacy of the motor pathway as a potential contributor to the exaggerated development of central fatigue during exercise at a given workload in old individuals.

Perspectives and Significance

Healthy aging impairs fatigue resistance during aerobic exercise at a given workload. Importantly, the comparison of fatigue following iso-time exercise performed at an intensity normalized for age-specific aerobic capacity (i.e., given relative intensity) distorts the impact of aging on the development of fatigue during dynamic exercise. Finally, muscle mass recruited during exercise has little influence on the age-related difference in fatigue when compared at a given relative intensity but may play a role in absolute exercise intensity comparisons.

GRANTS

This study was supported by National Heart, Lung, and Blood Institute Grant HL-116579 (to M. Amann).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

J.C.W. and M.A. conceived and designed research; J.C.W., T.J.H., T.S.T., S.K.S., R.S.G., and A.D.N. performed experiments; J.C.W. analyzed data; J.C.W., T.J.H., T.S.T., S.K.S., and M.A. interpreted results of experiments; J.C.W. prepared figures; J.C.W. and M.A. drafted manuscript; J.C.W., T.J.H., T.S.T., S.K.S., R.S.G., A.D.N., C.J.M., R.S.R., and M.A. edited and revised manuscript; J.C.W., T.J.H., T.S.T., S.K.S., R.S.G., A.D.N., C.J.M., R.S.R., and M.A. approved final version of manuscript.