Effects of Progressive Resistance Training on Body Composition in Frail Older Adults: Results of a Randomized, Controlled Trial

Abstract

Background. Progressive resistance exercise training (PRT) has been shown to increase muscle strength and fat-free mass (FFM) in elderly persons. Limited information is available regarding the effects of PRT on lean and fat mass in frail elderly persons.

Methods. Ninety-one community-dwelling sedentary men and women, 78 years and older with physical frailty (defined using standardized objective criteria) were enrolled in a 9-month trial of exercise training (ET). Physical frailty was defined as having 2 of the 3 following criteria: modified Physical Performance Test score between 18 and 32, peak aerobic power between 10 and 18 ml/kg/min, or self-report of difficulty or assistance with two instrumental activities of daily living or one basic activity of daily living. Participants were randomly assigned to either a control (CTL) group that performed a low intensity home exercise program or a supervised ET group that performed 3 months of low intensity exercise and 3 months of PRT.

Results. After completion of PRT, ET participants had greater improvements than did CTL participants in maximal voluntary force production for knee extension (mean Δ +5.3 ± 13 ft/lb vs +1.1 ± 11 ft/lb, p =.05), measured using isokinetic dynamometry. Total body FFM (measured using dual energy x-ray absorptiometry) increased in the ET group, but not in the CTL group (mean Δ +0.84 ± 1.4 kg vs +0.01 ± 1.5 kg, p =.005). Total, trunk, intra-abdominal, and subcutaneous fat mass (measured using dual energy x-ray absorptiometry and ¹H-magnetic resonance imaging) did not change in response to PRT.

Conclusions. Three months of supervised PRT induced improvements in maximal voluntary thigh muscle strength and whole body FFM in frail, community-dwelling elderly women and men. This supervised exercise program may not be sufficient to reduce whole-body or intra-abdominal fat area in this population.

Age-associated losses of skeletal muscle mass and strength (1–3) and increased fat mass (4–6) are well documented, and are associated with important clinical outcomes such as mobility impairments (3,7,8), disability (2,8–10), falls (9), and fractures (3). There is an accumulating body of evidence in support of progressive resistance exercise training (PRT) as an intervention to delay or reverse sarcopenia (11–13). Numerous studies performed in healthy older adults have consistently shown that high intensity PRT induces significant increases in fat-free mass (FFM) (14–16), muscle fiber area (17,18), and muscle cross-sectional area (19–21). A few studies have shown that PRT can decrease total fat-mass (14–16) and visceral fat (22,23). The observed changes in body composition have varied between studies, and relate to the intensity and duration of the intervention, study sample characteristics and size, and measurement techniques. Limited information is available for physically frail elderly populations, who are at greatest risk for sarcopenia and related morbidity. There is some evidence suggesting that the muscle hypertrophy response to PRT may be impaired in very old individuals (19,24). The aim of this study was to evaluate the changes in FFM and fat mass in response to PRT, in frail community-dwelling elderly women and men.

Methods

Participants

Men and women 78 years and older were recruited from the community to participate in a 9-month ET study. Participants provided written informed consent, which was approved by the Institutional Review Board of the Washington University School of Medicine.

We used objective criteria to define mild to moderate physical frailty. To be eligible, volunteers had to meet at least two of three of the following criteria: 1) modified Physical Performance Test score between 18 and 32 (possible maximum score = 36), 2) report of difficulty and/or assistance with up to two instrumental activities of daily living and/or one basic activity of daily living, 3) peak aerobic power (VO_2peak) between 10 and 18 ml·kg⁻¹·min⁻¹. Our screening instruments and procedures have been described (25).

Exclusion criteria were: 1) active serious illness within the past 6 months or conditions that would contraindicate weight training; 2) cognitive impairment judged to interfere with informed consent or completion of assessments or ET; 3) neuromuscular disorders more difficult to ameliorate with exercise (severe Parkinson's disease, stroke with hemiparesis, myasthenia gravis); 4) sensory impairments judged to interfere with following instructions for testing or exercise; 5) chronic use of steroids or immunosuppressive drugs; 6) use of estrogen, androgen, or progesterone-containing compound within 12 months; 7) cigarette use within 12 months; and 8) diagnosis of cancer within 5 years (except for superficial skin cancer).

Baseline and Follow-up Assessments

Diet evaluation

Under the supervision of a registered dietician, participants completed 3-day food records at baseline and on completion of each exercise phase. The food intake records were analyzed for total energy intake and macronutrient content using Nutritionist IV (First Databank, San Bruno, CA). A specific diet was not prescribed, and participants were instructed not to make drastic changes in their diet. To control for variations in calcium and vitamin D intake, participants were provided with supplemental calcium and vitamin D, to adjust intake to about 1200 mg/day and 600 IU/day, respectively.

Skeletal muscle strength

Maximal voluntary muscle strength for knee extension and flexion was measured with isokinetic dynamometry, using procedures that have been described (26). One-repetition maximum (1-RM) testing was performed only for participants in the ET group, as described below.

Body composition

Total body dual energy x-ray absorptiometry (DEXA) (Hologic QDR1000/W, software version 6.2OD; Waltham, MA) was used to assess FFM and fat mass (27). Coefficients of variation in our laboratory for total FFM, trunk fat, leg lean mass, and leg fat mass are 1.8%, 2%, 5%, and 5%, respectively.

Proton magnetic resonance imaging (MRI) was used to obtain images of the abdomen at baseline and after 6 months. Serial images above and below the L4–L5 interspace were acquired on a 1.5-T superconducting magnet (Siemens, Iselin, NJ) using a T1-weighted pulse sequence. NIH Image (v1.62) analysis software (Scion Corporation, Frederick, MD) was used to identify and quantify abdominal subcutaneous and intra-abdominal fat areas. Within a region, fat was quantified using segmentation and pixel intensity thresholding to visually separate fat from other tissues. To minimize classification errors, pre- and postintervention MRI images were analyzed at the same time by a trained technician who was blinded to group assignment. The coefficients of for intra-abdominal (visceral) fat area (VAT) and subcutaneous fat area (SAT) are 6 ± 3% and 7 ± 1%, respectively.

Randomization Procedures

Study participants were randomly assigned, on completion of the baseline assessments, to ET or control (CTL) groups in a 3:2 ratio, using a computer-generated random permutation procedure (28) and a block design.

ET Program

The supervised ET program consisted of three, approximately 3-month-long, phases of ET that have been previously described in detail (25). This report focuses on muscle strength and body composition measures obtained after the second exercise phase.

Phase 1 exercise used a group format and included 22 exercises that focused on improving flexibility, balance, coordination, speed of reaction and, to a modest extent, strength that has been described previously (29). Phase 2 added PRT. After familiarization with the equipment, 1-RM voluntary strength was measured on each of six different exercises (knee extension, knee flexion, seated bench press, seated row, leg press, biceps curl), which were performed bilaterally on a Hoist weightlifting machine (Hoist Fitness Systems, San Diego, CA). Initially, the participants performed 1–2 sets of 6–8 repetitions of each exercise at 65% of their 1-RM. Our goal was for the participants to progress their workload to 3 sets of 8–12 repetitions performed at 85%–100% of the initial 1-RM, although not all participants were able to achieve this goal. Measurements of 1-RM were repeated at monthly intervals so that workloads could be progressed. The participants also continued to perform a shortened version of the Phase 1 exercises. Phase 2 exercise sessions took 60–90 minutes to complete, including rest periods.

Participants were required to attend exercise sessions 3 times/week and complete 36 sessions of each exercise phase before follow-up assessments and progression to the next phase of ET.

Home Exercise Program (CTL Group)

This program included 9 of the 22 exercises included in Phase I of the ET program, and focused primarily on flexibility (29). CTL participants attended a monthly exercise class at our exercise facility, and used a calendar to self-monitor adherence to the exercises. They were asked to perform the exercises at home 2–3 times/week. Follow-up testing was performed at the end of each 3-month interval.

Statistical Analysis

Individuals that provided DEXA data at the 3-month and 6-month time points (before and after Phase 2) were included in this analysis, including participants who discontinued the interventions that provided such data. Between-group comparisons of continuous variables were performed using t tests, or Wilcoxon's test as a nonparametric alternative. Chi-square tests or Fisher's exact tests were used for between-group comparisons of categorical variables. Analysis of covariance (ANCOVA) was used to evaluate variables that were measured at only two time points, using the 6-month value as the dependent variable and the 3-month value as covariate (or baseline for MRI measures). Associations between continuous variables were analyzed with Pearson's correlation coefficient. Statistical significance was defined as α level ≤ 0.05. Data were analyzed using SAS statistical software (SAS Institute, Inc., Cary, NC).

Results

Recruitment

Four hundred forty-four individuals (152 men; 292 women) underwent the pre-enrollment evaluations (Figure 1). One hundred sixty-five were excluded from participation: 113 did not meet the selection criteria (79 too fit; 16 too frail; 67 women taking hormone-replacement therapy), and 70 had medical exclusions. Sixty-seven women elected to enroll in a concurrent study of exercise combined with hormone-replacement therapy, and 93 individuals declined participation.

Of the 119 individuals enrolled (69 ET, 50 CTL), 91 completed Phase 2 and provided complete DEXA data at 3 and 6 months after baseline (Figure 1). Seven ET participants discontinued the intervention by month 6 but provided follow-up data, and 20 participants (15 ET, 5 CTL) dropped out before 6 months and did not provide follow-up data. Eight participants (1 ET, 7 CTL) completed Phase 2 but did not provide complete DEXA data. Reasons for dropout are described in Figure 1. Two individuals in the ET group developed or exacerbated existing soft tissue injuries in the shoulder regions and dropped out of the study. There were no other adverse events. In comparison to participants included in the analysis, those who failed to provide follow-up data were older (85 vs 83 years, p =.01), had a lower modified Physical Performance Test score (25 vs 29, p =.004), and had a lower VO_2peak (13.3 ml·kg⁻¹·min⁻¹ vs 15.4 ml·kg⁻¹·min⁻¹, p =.0004), but there were no significant differences in initial DEXA measures. Of the 91 participants who provided DEXA follow-up data, 54 also provided MRI data at baseline and 6 months.

Participant Characteristics

The mean age of participants in both groups was 83 ± 4 years, with a trend toward a lower percentage of Caucasians in the ET group (p =.06). There were no other differences in the baseline characteristics of the study groups (Table 1).

Compliance with Protocol

For the ET group, the time to complete Phase 2 was 140 ± 41 days (median = 128 days), whereas for the CTL group it was 100 ± 25 days (median = 98 days) (p <.0001). Time to complete Phase 1 and 2 was 277 ± 62 days (median = 259 days) for the ET group, and 241 ± 68 days (median = 224 days) for the CTL group (p =.01). ET participants exercised an average of 2.2 ± 0.3 days/week. Average workloads for four representative resistance exercises over the course of Phase 2 are presented in Table 2. The seven ET participants who discontinued Phase 2 and provided follow-up data completed 28.4 ± 2.9 sessions (range 23–31).

ET and Skeletal Muscle Strength

We observed significant group differences in the change in knee extension (Table 3). For the ET group, there was also a significant increase in all 1-RM measures (Table 4). The relative increase in 1-RM for the ET group was 17% (leg flexion) to 43% (leg extension) above the pretraining value.

ET and Body Composition

FFM increased more in the ET group than in the CTL group (p =.005) (Table 5). Regional DEXA measurements indicated greater increases in leg lean mass in the ET group than in the CTL group (p =.03 for left leg, p =.06 for right leg). All between-group differences in change remained significant after controlling for sex. For ET participants, mean change in FFM between 3 and 6 months was correlated with mean change in 1-RM for knee extension (r = 0.43, p =.004). The change in isokinetic knee extension torque for the ET group was correlated with total weight lifted during months 3–6 (r = 0.30, p =.04). Changes in trunk or leg fat mass were not different between the groups.

Fifty-four participants (20 CTL; 34 ET) provided MRI data at baseline and 6 months. Baseline VAT was 179 ± 85 cm² for the CTL group (215 ± 92 cm² for men; 142 ± 62 cm² for women), and 195 ± 104 cm² for the ET group (240 ± 97 cm² for men; 145 ± 90 cm² for women) (group difference, p =.53). Baseline SATs were 167 ± 48 cm² (CTL) and 192 ± 99 cm² (ET) (p =.73). Between baseline and 6 months, the change in VAT was −3.8 ± 29 cm² for the CTL group and −7.0 ± 43 cm² for the ET group (p =.69) (median percent change −0.6% vs −7.2%, respectively); the change in SAT was −2.1 ± 28 cm² for the CTL group and −3.9 ± 25 cm² for the ET group (p =.72) (median percent change −2.5% vs −4.6%, respectively).

Total energy intake at 3 months was 1912 ± 384 kcal for the ET group and 1766 ± 356 kcal for the CTL group (p =.20); at 6 months it was 1932 ± 402 kcal for the ET group and 1925 ± 392 kcal for the CTL group (p =.92; p =.29 for group change between 0 and 6 months; p =.74 for change between 3 and 6 months). Total body weight at 6 months was 77 ± 16 kg for the ET group and 71 ± 14 kg for the CTL group (p =.92; p =.29 for group change between 0 and 6 months).

Discussion

These findings confirm that supervised PRT induces greater increases in lean mass and muscle strength than home-based exercise in frail women and men 78 years old and older. We have previously reported that these gains translate to improvements in functional performance and self-reports of disability (25). Depending on the exercise and gender, 1-RM strength increased between 17% and 43% above baseline. These changes are not as large as those observed in previous 12-week PRT studies in elderly participants (17,19,30). However, those trials studied healthier participants (17,30) or frail institutionalized elderly persons who performed a more limited set of PRT exercises (19). Our findings are limited to, and have better applicability to, frail community-dwelling elderly individuals with multiple comorbidities who are willing to participate in a PRT program.

Our supervised ET protocol was generally well tolerated, with few adverse events directly related to exercise. It is easily adaptable to fitness centers and outpatient physical therapy settings, many of which now have PRT facilities. Our goal was for participants to perform PRT exercises at 85% of their 1-RM, but many participants were unable to achieve that. Also, many participants required longer than 6 months to complete Phases 1–2 of the ET intervention due to acute illness, weather, and holidays. Replication of our ET program in community-based settings will require personal trainers and physical therapists with sufficient training to adapt the exercises to physical impairments common in frail elderly people (e.g., from arthritis, inactivity, or previous injuries), and prescribe appropriate and realistic exercise intensities for this population.

In agreement with prior reports, the magnitude of the increases in total body and leg lean mass was less than the observed strength gains. The change in lean mass was, however, related to the change in 1-RM strength. The absolute and relative increases in lean mass are consistent with previous 12-week studies of PRT in elderly persons (19). Our findings support the notion that PRT can stimulate skeletal muscle hypertrophy even in physically frail elderly persons, but the increases in strength observed over this relatively short time period are due largely to factors such as neural recruitment mechanisms (19,31,32).

Previous investigators have reported that PRT reduces whole-body fat mass (14–16) and regional fat mass (22,23,33) in elderly men and women. Campbell (14) observed decreases in whole-body fat mass, using hydrodensitometry, in healthy 56- to 80-year-old men and women, in response to 12 weeks of PRT and a controlled energy diet. Treuth and colleagues (23) observed decreases in intra-abdominal fat mass using single-slice computed tomography scans, but not total fat mass (hydrodensitometry) in healthy elderly women after 16 weeks of PRT. In a separate study of elderly men, using a more intensive 16-week PRT protocol and a diet that restricted changes in caloric intake, they observed significant decreases in total and trunk fat mass by DEXA (16). Hunter and colleagues (22) observed significant reductions in percent fat and fat mass in 26 healthy, nonobese, 61- to 77-year-old women and men, and intra-abdominal fat mass by computed tomography only in the women, after 25 weeks of high intensity PRT. We did not observe greater reductions in total or regional fat mass in the ET versus the CTL group. Although we monitored the participants' dietary intake periodically, and did not observe changes over time, we did not rigorously control the participants' energy intake. It is possible that the participants in our study either underreported their energy intake or increased it episodically in response to the energy demands of ET. In addition, ET participants did not exercise at as high an intensity as did participants in prior studies in which fat mass decreased. However, studies of the effects of exercise on abdominal adiposity have not established a clear dose-response effect (34). Baseline abdominal MRI measurements for our participants were above the threshold associated with increased risk for the metabolic syndrome (e.g., VAT > 100 cm²) (35,36), and the observed changes were not of sufficient magnitude for risk reduction. It is possible that PRT combined with aerobic training and/or decreased energy intake is necessary in frail elderly people to achieve clinically significant reductions in abdominal adiposity. Further study is needed to clarify the optimal PRT regimen for reducing fat mass in this population, and to determine if PRT-induced fat losses confer metabolic or cardiovascular benefits in physically frail elderly persons relative to their baseline adiposity.

Our study had several limitations. First, for purposes of recruitment and retention, our CTL group performed low intensity home exercise, which could have maintained their strength and FFM more than it would in a sedentary control group. Therefore, we may have underestimated the magnitude of the changes in strength and body composition in response to PRT. Second, our sample size may have been too small, and our power inadequate, to detect significant changes in fat mass, and it is possible that our findings represent a Type II error. Third, the variability in MRI measures of fat area may have limited our ability to detect changes. Fourth, the 3-month PRT program may have been too short or of insufficient intensity to induce changes in FFM and fat mass comparable to those in studies in healthier populations.

Conclusion

In frail, community-dwelling elderly men and women, low to moderate intensity PRT induced greater increments in total and regional FFM, and isokinetic muscle strength, but no changes in fat mass, when compared to a home-based, low intensity exercise program. Additional studies are needed to confirm our findings and identify factors that augment the beneficial effects of PRT on body composition in frail elderly persons and related effects on functional performance and disability.

Decision Editor: John E. Morley, MB, BCh

Figure 1.

Results of randomization and participant flow

Table 1.

Baseline Characteristics of the Sample (N = 91).

Variable	Control (N = 38) Mean ± SD	ET (N = 53) Mean ± SD	p Value
Age	83 ± 4	83 ± 3	.89
Female (%)	55%	53%	.82
Education
Not high school graduate	21%	18%	.97
High school graduate	13%	12%
Some college	29%	29%
College graduate	37%	41%
Ethnic origin (% Caucasian)	95%	81%	.06
Short Blessed Test score	3 ± 3	2 ± 2	.18a
GDS score	1 ± 2	2 ± 2	.43a
No. of routine medications	4 ± 2	4 ± 2	.76a
Total modified PPT score	29 ± 4	29 ± 4	.65
FSQ score	27 ± 4	27 ± 4	.78
BMI (kg/m2)	26 ± 4	27 ± 5	.43
Body weight (kg)	71 ± 14	77 ± 16	.17
Waist circumference (cm)	90 ± 11	91 ± 14	.81
VO2peak (ml/kg/min)	16 ± 2	15 ± 3	.64
Total intake (kcal)	1864 ± 379	1885 ± 334	.79
Protein % intake	16 ± 3%	15 ± 3%	.48
Carbohydrate % intake	53 ± 6%	52 ± 6%	.36
Fat % intake	30 ± 6%	31 ± 5%	.38
Medical history (%)
Degenerative joint disease	76%	75%	.93
Hypertension	45%	53%	.45
Coronary artery disease	24%	23%	.85
Joint replacement	19%	15%	.66
Atrial fibrillation	16%	12%	.61
Diabetes	11%	8%	.72a
Congestive heart failure	5%	0	.17a
Previous cigarette use	47%	45%	.84
Takes lipid-lowering agents	11%	15%	.75

Notes: ^aDenotes that p value is based on Wilcoxon's test or Fisher's exact test.

ET = exercise training group; SD = standard deviation; GDS = Geriatric Depression Scale; PPT = Physical Performance Test; FSQ = Functional Status Questionnaire; BMI = body mass index; VO_2peak = peak aerobic power.

Table 2.

PRT Workloads for ET Participants (N = 53).

Exercise	Portion of PRT Exercises Completed	Total Weight Lifted (Mean ± SD)	% Initial 1-RM (Mean ± SD)
Leg press	Week 1	171 ± 63	62 ± 6
	33%	219 ± 67	80 ± 16
	67%	230 ± 73	86 ± 20
	Final week	299 ± 92	87 ± 18
Knee flexion	Week 1	138 ± 45	62 ± 7
	33%	175 ± 47	78 ± 12
	67%	191 ± 47	87 ± 14
	Final week	243 ± 74	88 ± 14
Knee extension	Week 1	127 ± 65	60 ± 10
	33%	158 ± 70	75 ± 18
	67%	165 ± 67	83 ± 27
	Final week	221 ± 94	89 ± 26
Seated row	Week 1	106 ± 41	63 ± 7
	33%	132 ± 42	78 ± 13
	67%	137 ± 51	82 ± 16
	Final week	178 ± 74	81 ± 16

Notes: Portion of PRT exercises completed refers to the 1st week (first 3 sessions), the 3 sessions in which the participant was 33% finished with Phase 2, the 3 sessions in which the participant was 67% finished with Phase 2, and the final 3 sessions of Phase 2 training. This was done to “normalize” for individual differences in the amount of time required to complete the PRT (Phase 2) training (See Results). Total weight lifted for each exercise was calculated by multiplying the weight lifted times the number of repetitions summed over all sets.

PRT = progressive resistance training; ET = exercise training; 1-RM = one-repetition maximum; SD = standard deviation.

Table 3.

Changes in Muscle Strength Measured Using Cybex Isokinetic Dynamometry.

Variable	Interval	Control (N = 36) Mean ± SD	ET (N = 49) Mean ± SD	p Value
Knee extension 60°/s (ft/lb)	3-mo	59 ± 17	66 ± 19	.06
	6-mo	60 ± 17	72 ± 21
	Change	1.08 ± 11	5.31 ± 13	.05
Knee flexion 60°/s (ft/lb)	3-mo	42 ± 14	45 ± 13	.26
	6-mo	44 ± 14	49 ± 15
	Change	2.11 ± 7	3.21 ± 8	.39

Notes: Change is calculated by subtracting the 3-month value from the 6-month value. The first p value is based on an unpaired t test that compares the 3-month values. The second p value is based on analysis of covariance (ANCOVA) where the 6-month value is the dependent variable and the 3-month value is a covariate.

SD = standard deviation; ET = exercise training group.

Table 4.

Changes in Bilateral 1-RM Strength Measures for ET Participants.

Variable	Interval	All ET (N = 50) Mean ± SD	Males (N = 23) Mean ± SD	Females (N = 27) Mean ± SD	p Value Testing Change*	ANCOVA p Value for Sex**
Leg flexion (lb)	Before resistance training	76 ± 21	92 ± 18	62 ± 10
	After resistance training	88 ± 24	107 ± 19	72 ± 13
	Change	12 ± 10	14 ± 11	10 ± 8	.0001	0.11
	% Increase	17 ± 13	17 ± 13	17 ± 14
Leg extension (lb)	Before resistance	71 ± 30	83 ± 31	58 ± 22
	After resistance training	95 ± 39	108 ± 44	81 ± 27
	Change	24 ± 32	25 ± 39	22 ± 22	.0001	0.37
	% Increase	43 ± 52	39 ± 52	47 ± 53
Leg press (lb)	Before resistance training	93 ± 28	107 ± 22	81 ± 26
	After strength training	116 ± 35	129 ± 29	105 ± 37
	Change	23 ± 20	22 ± 21	24 ± 20	.0001	0.51
	% Increase	27 ± 21	23 ± 21	30 ± 21
Seated row	Before resistance training	60 ± 21	75 ± 16	46 ± 14
	After strength training	70 ± 25	88 ± 20	53 ± 16
	Change	10 ± 10	14 ± 10	7 ± 9	.0001	0.09
	% Increase	17 ± 18	19 ± 17	16 ± 19

Notes: Means and SD values are reported for selected 1-RM strength measures before and after progressive resistance training for ET participants. Change is computed by subtracting the before-training value from the after-training value. Percent increase is computed by dividing the change score by the pretraining value and multiplying by 100.

*p values test for a significant change between time points, based on paired t tests.

**p values test for a difference in change for men and women and is based on ANCOVA where the after-training value is the dependent variable and the before-training value is the covariate.

1-RM = one-repetition maximum; ET = exercise training; SD = standard deviation; ANCOVA = analysis of covariance.

Table 5.

Whole-Body and Regional Fat-Free Mass and Fat Mass Measured by Dual Energy X-Ray Absorptiometry (DEXA).

Variable	Interval	Control (N = 38) Mean ± SD	ET (N = 53) Mean ± SD	p Value
Total fat-free mass (kg)	3 months	46.2 ± 9.6	47.3 ± 9.8	.60
	6 months	46.2 ± 9.3	48.1 ± 9.7
	Change	0.0 ± 1.5	0.8 ± 1.4	.005
Percent body fat (%)	3 months	32.9 ± 8.6	34.4 ± 9.4	.45
	6 months	32.5 ± 8.6	33.4 ± 9.1
	Change	−0.4 ± 1.9	−1.0 ± 1.7	.22
Trunk fat (kg)	3 months	11.4 ± 5.4	12.2 ± 6.3	.56
	6 months	11.0 ± 5.3	11.9 ± 6.4
	Change	−0.4 ± 1.0	−0.3 ± 1.3	.60
Right leg lean mass (kg)	3 months	6.6 ± 1.6	6.9 ± 1.7	.36
	6 months	6.6 ± 1.6	7.0 ± 1.7
	Change	0.0 ± 0.3	0.1 ± 0.4	.06
Right leg fat mass (kg)	3 months	4.0 ± 1.4	4.6 ± 2.1	.09
	6 months	4.0 ± 1.4	4.6 ± 2.1
	Change	0.04 ± 0.2	−0.04 ± 0.3	.13
Left leg lean mass (kg)	3 months	6.5 ± 1.5	6.7 ± 1.8	.64
	6 months	6.5 ± 1.4	6.8 ± 1.7
	Change	−0.1 ± 0.4	0.1 ± 0.3	.03
Left leg fat mass (kg)	3 months	3.9 ± 1.4	4.6 ± 2.0	.07
	6 months	4.0 ± 1.5	4.6 ± 2.0
	Change	0.1 ± 0.3	−0.1 ± 0.3	.11

Notes: Change is calculated by subtracting the 3-month value from the 6-month value. The first p value is based on an unpaired t test that compares the 3-month values. The second p value is based on analysis of covariance (ANCOVA) where the 6-month value is the dependent variable and the 3-month value is a covariate.

SD = standard deviation; ET = exercise training group.

This work was supported by National Institutes of Health (NIH) Claude Pepper Older Americans Independence Center (OAIC) grant P01-AG13629 and by NIH General Clinical Research Center grant 5-M01 RR00036. We thank Debbie Bronder, Jil Yarasheski, and all the staff of the OAIC and the General Clinical Research Center for technical assistance with this project.

Author notes

1Department of Internal Medicine, 2Program in Physical Therapy, and 3Division of Biostatistics, Claude Pepper Older Americans Independence Center, Washington University School of Medicine, St. Louis, Missouri.