Endocrinology and Metabolism


Skeletal muscle loss or sarcopenia in aging has been suggested in cross-sectional studies but has not been shown in elderly subjects using appropriate measurement techniques combined with a longitudinal study design. Longitudinal skeletal muscle mass changes after age 60 yr were investigated in independently living, healthy men (n = 24) and women (n = 54; mean age 73 yr) with a mean ± SD follow-up time of 4.7 ± 2.3 yr. Measurements included regional skeletal muscle mass, four additional lean components (fat-free body mass, body cell mass, total body water, and bone mineral), and total body fat. Total appendicular skeletal muscle (TSM) mass decreased in men (−0.8 ± 1.2 kg,P = 0.002), consisting of leg skeletal muscle (LSM) loss (−0.7 ± 0.8 kg, P = 0.001) and a trend toward loss of arm skeletal muscle (ASM; −0.2 ± 0.4 kg,P = 0.06). In women, TSM mass decreased (−0.4 ± 1.2 kg, P = 0.006) and consisted of LSM loss (−0.3 ± 0.8 kg, P = 0.005) and a tendency for a loss of ASM (−0.1 ± 0.6 kg, P = 0.20). Multiple regression modeling indicates greater rates of LSM loss in men. Body weight in men at follow-up did not change significantly (−0.5 ± 3.0 kg, P = 0.44) and fat mass increased (+1.2 ± 2.4 kg, P = 0.03). Body weight and fat mass in women were nonsignificantly reduced (−0.8 ± 3.9 kg, P= 0.15 and −0.8 ± 3.5 kg, P = 0.12). These observations suggest that sarcopenia is a progressive process, particularly in elderly men, and occurs even in healthy independently living older adults who may not manifest weight loss.

  • aging
  • body composition
  • skeletal muscle mass
  • longitudinal study

senescence in both animals and humans is associated with body composition changes and related functional decline (9). Cross-sectional population studies in humans suggest that body weight (1,21, 22) and fatness increase (4) up to the sixth decade with a gradual decline thereafter in body weight (35), skeletal muscle mass (10,35, 36), and bone mineral (29,30). Frailty and sarcopenia may represent the advanced stages of progressive age-related body composition changes (6, 32). Weakness, falls, functional limitations, immobility, and osteoporotic fractures may be linked to declines in musculoskeletal mass (6, 32).

Although substantial information exists on the nature of skeletal muscle mass and related skeletal changes in old age, available information represents analyses based largely on cross-sectional studies (2, 12) that often relied on measurement methods of questionable validity and accuracy (e.g., urinary creatinine and anthropometry) or with small sample sizes (3). These reports frequently failed to exclude elderly subjects with chronic wasting diseases (37) and those who had recently dieted for weight loss. A fundamental and as yet unanswered question concerns the magnitude of longitudinal age-related skeletal muscle mass loss in the absence of clinically evident disease in weight-stable elderly subjects.

A recent cross-sectional study of healthy adults examined the interrelationships among appendicular skeletal muscle mass, age, and gender (12). Although both older women and men had less absolute amounts of skeletal muscle compared with their younger counterparts, after adjustment for height and weight, relative reductions were larger in men. A compilation by Forbes (9) of cross-sectional body composition studies from infancy to old age arrived at a similar conclusion. These observations imply that, in men, the absolute rate of decline in skeletal muscle mass with senescence exceeds that in women, but whether or not these findings apply to a longitudinally evaluated healthy and mobile elderly population remains uncertain.

The present study was designed to evaluate skeletal muscle mass loss over time in healthy community-dwelling and independently living elderly adults. Specifically, the hypothesis was that, after excluding persons with underlying disease and voluntary weight loss, there is an ongoing reduction of skeletal muscle mass. A secondary aim of the study was to explore previously reported (9, 12) gender differences in the rate of age-related skeletal muscle decline.


Subject Recruitment

Clinical status and body composition were longitudinally monitored in a cohort of healthy men and women. The primary subject recruitment pool consisted of 1,449 healthy adults who were tested between 1986 and 1994 (Fig. 1) as part of an ongoing study at the Body Composition Unit of the New York Obesity Research Center (12, 13, 38). Changes in body composition instrumentation with the replacement of dual photon absorptiometry by dual-energy X-ray absorptiometry (DEXA) occurred in May 1989. Between May 1989 and December 1994, 261 subjects over the age of 60 yr at the time of initial study had been evaluated. This group was targeted to participate in the follow-up study. The study cohort was developed by attempting to contact these 261 subjects.

Fig. 1.

Flow chart of enrollment in the study. IWL, intentional weight loss; WR, weight regain; DHR, double hip replacement; EE, endurance exerciser.

Contact was made with 170 (65%) of the 261 elderly subjects. Of the 170 subjects contacted, 100 (59%) agreed to return for evaluation during the years 1995 to 1997, a follow-up time ranging from 1 to 9 yr in men and 1 to 10 yr in women. The remaining 70 subjects (41%) were not willing to participate or were excluded due to health reasons (e.g., self-reported cancer, cerebrovascular disease, nonambulatory physical disability, or serious chronic illness). Subjects with high blood pressure adequately controlled with medications and with nonclinically significant conditions such as mild osteoarthritis were not excluded.

During follow-up, each subject completed a medical examination that included screening blood tests after an overnight fast. Blood samples were sent to a commercial laboratory (Corning Clinical Laboratories and Quest Diagnostics, Teterboro, NJ) for analysis. Of the 100 medically evaluated subjects, of whom 71 were women and 29 were men, 22 were excluded for the following reasons: one male was an outlier with regard to age (baseline age of 107 yr, far beyond the age range of remaining subjects); intentional diet-related weight loss (>10 kg;n = 6); relapse from previous diet with weight gain between baseline and follow-up evaluations (>5 kg; n = 5); diabetes mellitus based on self report of physician-diagnosed disease (n = 5); double hip replacement (n = 1); endurance exerciser (bicycler, runner, mountain climber × 10 yr; n = 1); or cancer (n = 3). A total of 78 healthy, weight stable subjects, 54 women and 24 men, were included in the statistical analyses.

The study was approved by the Institutional Review Board of St. Luke's-Roosevelt Hospital, and subjects gave written consent before participation.

Body Composition

Subjects reported in a fasted state to the Body Composition Core Laboratory in the morning. They then donned hospital gowns and foam slippers. Body weight and height were measured to the nearest 0.1 kg (Weight Tronix, New York, NY) and 0.5 cm (Holtain Stadiometer, Crosswell, Wales), respectively.

Skeletal muscle and four related lean components (i.e., fat-free body mass, body cell mass, total body water, and bone mineral mass) along with total body fat were monitored over time. Skeletal muscle compartments of the arms and legs were quantified using regional DEXA-measured lean soft tissue. Total appendicular skeletal muscle was calculated as the sum of arm and leg lean soft tissue (14). Fat was measured by DEXA, and fat-free body mass was then calculated as the difference between body weight and total body fat. Total body bone mineral mass was also evaluated using DEXA. Total body water, which usually represents 70–75% of fat-free body mass (39), was measured using tritiated water (3H2O). Body cell mass was derived from total body potassium as measured with a 4-pi 40K whole body counter (23).


Body composition was measured with whole body DEXA (Lunar Radiation, Madison, WI) operated in slow scan mode. Software version 3.6 was used to analyze all of the DEXA scans. The system software provided estimates of the following three compartments: fat, lean soft tissue, and bone mineral for the whole body and specific regions. Fat-free body mass was calculated as the sum of lean soft tissue and bone mineral.

Regional measurements were made separately for the legs and arms (14). With the use of anatomic landmarks, the legs (i.e., soft tissue extending from a computer-generated line drawn through and perpendicular to the axis of the femoral neck and angled with the pelvic brim to the phalange tips) and arms (i.e., soft tissue extending from the center of the arm socket to phalange tips) were first isolated on the skeletal DEXA anterior view planogram. System software then provided estimates of leg and arm fat, lean soft tissue, and bone mineral mass. The fat and bone mineral-free portion of the arms and legs was assumed to represent leg and arm appendicular skeletal muscle mass along with a small and relatively constant amount of skin and underlying connective tissue (14). Repeated daily measurements over 5 days in four subjects showed a coefficient of variation (CV) of 1.7% for leg, 2.0% for arm, and 2.6% for total appendicular skeletal muscle.

Body composition estimates were all acquired on the same DEXA system, and software version 3.6 was used for data analysis at baseline and follow-up. An anthropomorphic spine phantom made up of calcium hydroxyapatite embedded in a 17.5 × 15 × 17.5-cm block was scanned for quality control each morning before subject evaluation. The phantom was also scanned immediately before and after all DEXA system manufacturer maintenance visits. The measured phantom bone mineral density was stable throughout the study period at 1.166–1.196 g/cm2. Ethanol and water bottles (8-liter volume), simulating fat and fat-free soft tissues, respectively, were scanned as soft tissue quality control markers monthly. The range in measuredR values over the study period was 1.255–1.258 (CV = 0.127%) and 1.367–1.371 (CV = 0.103%) for ethanol and water, respectively.

Whole Body Counting

The St. Luke's 4-pi whole body counter was used to measure40K (23). The 40K raw counts accumulated over 9 min were adjusted for self-absorption of the 1.46 meV γ-ray emission by a 42K calibration equation based on body size (26). The within-subject CV in our laboratory for 40K counting is 4% (25). Total body potassium, in millimoles, was calculated as 40K/0.0118.

Body cell mass was calculated from total body potassium as body cell mass (kg) = 0.00833 × TBK (mmol). This equation is based on two assumptions, that the average intracellular potassium content is ∼3 mmol/g of nitrogen and that the nitrogen content is 0.04 g/g wet tissue (19). There is good support for these assumptions in weight-stable healthy women and men across a wide age spectrum (5, 18).

Counting efficiency was measured continuously throughout the study by evaluating calibration standards in the form of 10-lb (∼4 kg) plastic bottles containing potassium chloride (5,000 dpm of 40K) in distilled water. At the beginning of each study day, bottles approximating the body size and weight of each subject were counted. The between-day CV for 40K counting of calibration standards over the study period was 2.4%.

Radiolabeled Water Dilution

A blood sample was taken before and 3 h after subjects received 200 μCi (7.4 mBq) of 3H2O (24). Total body water volume was estimated as the3H2O dilution space multiplied by 0.96, to correct for nonaqueous hydrogen exchange (28). Total body water in kilograms was calculated as the product of total body water volume and density at 37°C (i.e., 0.994 g/cm3). The within-subject CV for repeated 3H2O dilution volume estimates is 1.5% (28). An 8-liter plastic bottle containing an accurately measured 6-liter quantity of tap water was used as a daily quality control phantom, with tracer dosage proportional to individual study subject body weight. The CV for repeated phantom measurements was 1.3% over the study interval.

Statistical Methods

Power calculation.

The power to test the main hypothesis was based on the change in total appendicular skeletal muscle of 0.4 kg/decade and 0.8 kg/decade for women and men respectively, in the cross-sectional cohort reported by Gallagher et al. (12), using the same methodology. Repeated DEXA measures scanned over 5 days on four subjects indicated a between-day SD for total appendicular skeletal muscle of 0.53 kg. With the use of these estimates of change and variability, the power to detect a change from baseline of 0.2 kg in women was 0.83, with a sample size of 50. The power to detect a change from baseline of 0.4 kg in men was 0.97, with a sample size of 25.

Data analysis.

The hypothesis was tested that, after excluding persons with underlying disease and voluntary weight loss, there is an ongoing erosion of appendicular skeletal muscle mass in aging men and women. The statistical significance of longitudinal changes in total appendicular skeletal muscle mass was tested within each gender using pairedt-tests. Pearson correlation coefficients were used to quantify bivariate relationships between change in appendicular skeletal muscle and corresponding changes in other lean components, fat-free body mass, total body water, body cell mass, and bone mineral.

The relationships between changes over time in appendicular skeletal muscle and baseline appendicular skeletal muscle, age, gender, and follow-up interval were investigated using multiple regression analysis. The observed change in appendicular skeletal muscle was set as the dependent variable, and baseline appendicular skeletal muscle, age, gender, and follow-up interval were added to the regression model as independent variables. In the analyses, two α-tailed levels of 0.05 were considered statistically significant.

Longitudinal changes in other measured body compartments and baseline gender, demographic, and body composition differences were tested for statistical significance using paired and unpaired t-tests. These descriptive analyses were considered borderline or statistically significant with P < 0.10 and P < 0.05, respectively. Several subjects were, on occasion, unable to complete all of the evaluations due to technical and personal reasons, and the reduction in sample size is identified in results. Three of the 54 women were currently receiving estrogen replacement therapy. Analysis of the data with and without the subjects on estrogen replacement gave similar results, and these subjects have therefore been retained in the analyses. Data were analyzed using SAS statistical software version 6.12 (SAS Institute, Cary, NC), and group results are presented as means ± SD in text and as means ± SE in Figs.1-4.


Subject Characteristics

The baseline and follow-up subject descriptive characteristics are summarized in Table 1. Men and women were similar in age at baseline (72.9 ± 5.5 and 70.2 ± 7.8 yr, range 60–90 yr) and follow-up (77.8 ± 6.5 and 74.8 ± 7.8 yr, range 62–96 yr). Men were heavier (P = 0.001) and taller (P = 0.001) than women, and there were no significant gender differences in body mass index or in the percentage of subjects who were currently smoking. The mean follow-up interval for the study population was 4.7 ± 2.3 yr and was similar (P = 0.64) for men (4.9 ± 2.4 yr) and women (4.6 ± 2.3 yr).

View this table:
Table 1.

Subject characteristics at time of baseline and follow-up evaluations

Longitudinal Body Composition Changes

The changes observed in body composition compartments from baseline to follow-up evaluation, expressed as absolute values and as annualized rates, are reported in Table 2for men and in Table 3 for women.

View this table:
Table 2.

Results of longitudinal body composition studies in men

View this table:
Table 3.

Results of longitudinal body composition studies in women


Total appendicular skeletal muscle mass decreased significantly in the men both as an absolute change (−0.8 ± 1.2 kg, P= 0.002) and as a percentage of baseline skeletal muscle mass (−3.3 ± 4.5%, P = 0.001). The reduction in men of total appendicular skeletal muscle mass consisted of a significant lowering of leg (−0.7 ± 0.8 kg, P = 0.001) and a borderline significant lowering of arm (−0.2 ± 0.4 kg,P = 0.06) skeletal muscle mass. The annual rate of total appendicular skeletal muscle mass change in the men was −0.2 ± 0.5 kg/yr.

The absolute change in leg skeletal muscle mass as a function of age is presented in Fig. 2 for each male subject. From the total sample of 24 men, 21 (87.5%) showed a decrease or no change (Δ <0. 5 kg) in leg appendicular skeletal muscle from baseline to follow-up. The remaining 3 men (15%) showed an increase (Δ ≥0. 5 kg) in leg appendicular skeletal muscle.

Fig. 2.

Leg skeletal muscle (kg) at baseline and follow-up in individual male subjects vs. age.

There was also a significant reduction in men of fat-free body mass (−1.4 ± 1.9 kg, P = 0.002) and one of its main components, total body water (−1.4 ± 2.9 kg, P = 0.04). Other fat-free body mass components such as body cell mass and bone mineral mass also decreased over time in men, although the changes were either borderline or not statistically significant (−0.8 ± 2.0 kg, P = 0.08 and −0.023 ± 0.122 kg,P = 0.38).

The relative loss of bone mineral in men, −0.7 ± 4.2%, was substantially less than that observed for skeletal muscle (−3.3%). Accordingly, there was a significant decrease from baseline in the ratio of appendicular skeletal muscle mass to bone mineral (16.1 ± 3.2 to 15.6 ± 4.1 kg/kg, P = 0.011).

In contrast to significant lean losses, there was no significant change in body weight (−0.5 ± 3.0 kg, P = 0.44), and fat mass increased significantly (+1.2 ± 2.4 kg,P = 0.03).

The changes in body weight and body composition components in men are summarized in Fig. 3. The observed 0.5-kg decrease in body weight was mainly reflective of a combined decrease in fat-free body mass (−1.4 kg) and an increase in fat mass (+1.2 kg). Of the observed decrease in fat-free body mass, 15% was arm (−0.2 kg) and 54% was leg (−0.7 kg) skeletal muscle. Therefore, arm and leg skeletal muscle combined accounted for 69% of fat-free mass loss. The observed change in bone mineral mass accounted for <1% (−0.023 kg) of fat-free mass loss.

Fig. 3.

Changes (±SE) over study period in body weight and body composition components in men. FFM, fat-free mass; TSM, total skeletal muscle; LSM, leg skeletal muscle; ASM, arm skeletal muscle; BM, bone mineral.


The absolute and relative changes in total appendicular skeletal muscle mass in women were −0.4 ± 1.2 kg (P = 0.006) and −2.2 ± 7.1% (P = 0.001), respectively. The reduction in women of total appendicular skeletal muscle mass consisted of a significant lowering of leg skeletal muscle mass (−0.3 ± 0.8 kg, P = 0.005) and a nonsignificant reduction in arm skeletal muscle mass (−0.1 ± 0.6 kg, P = 0.20). The annual rate of total appendicular skeletal muscle mass change in women was −0.1 ± 0.4 kg/yr.

From the total sample of 54 women, 46 (85%) showed a decrease (Δ <0.5 kg) or no change in leg appendicular skeletal muscle from baseline to follow-up. The remaining 8 women (17%) showed an increase (Δ ≥0. 5 kg) in leg appendicular skeletal muscle.

There was a negligible change in fat-free body mass (+0.02 ± 1.4 kg, P = 0.93) and a significant reduction in total body water (−0.9 ± 2.1 kg, P = 0.005). Body cell mass and bone mineral mass also decreased (−0.30 ± 1.3 kg,P = 0.11; and −0.036 ± 0.106 kg,P = 0.02) over the study period in the women.

The loss of bone mineral in women, −1.6 ± 5.1%, was similar to that observed for skeletal muscle (−2.2 ± 7.1%). There was no significant change from baseline in the ratio of appendicular skeletal muscle to bone mineral (16.9 ± 4.9 to 16.8 ± 5.3 kg/kg,P = 0.61).

Body weight and fat mass in women also decreased over the study interval, although nonsignificantly by −0.8 ± 3.9 kg (P = 0.15) and −0.8 ± 3.5 kg (P= 0.12), respectively.

Figure 4 shows the observed longitudinal changes in body weight and body composition components in women. The observed 0.8-kg decrease in body weight was reflective of a loss in fat mass (−0.8 kg) with no change in measured fat-free mass (+0.02 kg). However, within the fat-free mass compartment, there was a 0.1-kg loss of arm and a 0.3-kg loss of leg skeletal muscle, and a small increase in lean soft tissue of the trunk accounted for the difference between unchanged fat-free mass and appendicular skeletal muscle loss. The observed change in bone mineral was −0.036 kg.

Fig. 4.

Changes (±SE) over study period in body weight and body composition components in women.

Bivariate correlations.

The bivariate correlations between changes in total and leg appendicular skeletal muscle mass and changes in other lean components are summarized in Table 4. The pattern for both total and leg skeletal muscle was similar, with significant or borderline significant associations observed with parallel changes in all other lean compartments, including fat-free body mass, total body water, and body cell mass.

View this table:
Table 4.

Bivariate correlations between changes in appendicular skeletal muscle mass and changes in other body compartments

Gender Differences

Neither women nor men had statistically significant reductions in body weight. However, fat mass increased significantly in men (P = 0.03) and decreased nonsignificantly (P = 0.12) in women. The gender difference in longitudinal changes in fat mass was statistically significant (P = 0.003).

The absolute change in total appendicular skeletal muscle was larger in men than in women (P = 0.08) and consisted of a nonsignificant difference in arm (P = 0.27) and a significant difference in leg skeletal muscle change (P= 0.05; Tables 2 and 3). A similar gender difference was observed in absolute fat-free mass change (P = 0.002). Multiple regression analysis of the pooled data with change in leg skeletal muscle as the dependent variable indicated two significant independent variables, baseline leg skeletal muscle mass (P = 0.002) and follow-up interval (P = 0.01). Gender entered the regression model as borderline significant (P = 0.09), and baseline age did not enter as a significant independent variable in the regression model (Table5).

View this table:
Table 5.

Multiple-regression analysis for leg skeletal muscle

Of the 78 evaluated subjects, there were 31 (11 men and 20 women) whose follow-up interval was ≥6 yr. The mean follow-up time for both men and women in this cohort was identical at 7.1 ± 1.2 yr. The absolute changes in total appendicular skeletal muscle were −1.1 ± 1.1 kg in the men (P = 0.005) and −0.6 ± 1.0 kg in the women (P = 0.012). Multiple regression analysis of the pooled data for this sample with change in leg skeletal muscle as the dependent variable indicated two significant independent variables, baseline leg skeletal muscle mass (P = 0.02) and gender (P = 0.005; total model R 2= 0.27, SE of the estimate = 0.6 kg). Gender, follow-up interval, and baseline age were not statistically significant independent variables when multiple regression models were developed with changes in body weight, fat mass, fat-free mass, and bone mineral as the dependent variables. Baseline fat-free mass entered the fat-free mass change model as a significant covariate; thus larger baseline values were associated with greater loss over time.


This study supports the hypothesis that loss of skeletal muscle mass occurs with advancing age in elderly men and women, even in independently living healthy subjects. An additional observation, in support of earlier cross-sectional studies, is that men lost significantly more leg skeletal muscle mass than did women over the study interval. Skeletal muscle mass loss in men was masked by weight stability resulting from a corresponding increase in total body fat mass. Progression of sarcopenia, particularly in men, may therefore be clinically silent and comparable to loss of bone mineral in osteoporosis.

Aging and Skeletal Muscle Loss

Evidence ranging from tissue biochemical (27,31) to gross anatomic (17) changes in cross-sectional samples provides strong but indirect evidence of age-related skeletal muscle loss or “sarcopenia.” Interpretation of skeletal muscle changes in cross-sectional samples is, however, limited due to potential cohort effects and secular trends. Additionally, the body composition “changes” observed in most of the earlier studies lacked adequate adjustment for confounding factors such as subject weight, stature, and health status. Moreover, the applied body composition methods, such as urinary creatinine and anthropometry (22), were often limited in accuracy and had inadequate power to measure small changes in the skeletal muscle compartment.

The present investigation was designed to overcome the limitations of these earlier studies by specifically evaluating appendicular skeletal muscle and four closely related whole body components (fat-free body mass, body cell mass, total body water, and total body bone mineral mass). Our findings show, unequivocally, that independently living healthy elderly adults experience an ongoing decline in skeletal muscle mass and related lean components.

A large proportion (∼75–80%) of skeletal muscle is water (33), and over one-half of the total body potassium pool is within the cellular matrix of skeletal muscle fibers. Moreover, skeletal muscle accounts for more than one-half (∼55%) of total fat-free body mass (33). The observation of concurrent significant or borderline significant reductions in fat-free body mass, appendicular skeletal muscle, body cell mass, and total body water strengthens our study results and confirms earlier longitudinal observations. Steen et al. (35) observed a significant loss of total body water after a 5-yr follow-up of 28 Swedish ≥70-yr-old men (−3.0 kg or 2%; P < 0.05) and 37 women (−2.0 kg or 2%; P < 0.05). Body weight and body cell mass in the Swedish subjects also decreased significantly in both sexes from ages 70 to 75 yr (men, body weight and cell mass, −2.9 and −1.0 kg; women, body weight and cell mass, −1.7 and −0.6 kg; allP < 0.05), whereas body fat increased significantly in men only (+0.8 kg, P < 0.05; women, +0.6 kg,P = not significant). Flynn et al. (8) observed an age-related significant decline in total body potassium (i.e., body cell mass) in 564 longitudinally monitored men (14%) and 61 women (13%) who were followed over an 18-yr study interval. Significant declines in body mass index (−0.7%, P < 0.05) and the creatinine-to-height ratio (i.e., an index of skeletal muscle mass; −4.6%, P < 0.05) were observed in 445 men (>60 yr) who were monitored in the Baltimore Longitudinal Study of Aging (20) over a 2.6-yr mean follow-up interval. Our observations, combined with these earlier reports, indicate a dynamic remodeling of body composition as independently living healthy older adults advance in age. Fat mass remains unchanged or may even increase, whereas there are consistent losses of lean tissues, including the large fat-free body mass component and its constituent skeletal muscle mass, body cell mass, total body water, and bone mineral mass.

As suggested by results from earlier cross-sectional studies (12), men and women between the ages of 20 and 60 yr have reductions in total appendicular skeletal muscle mass of −3.2 kg (11%) and −1.6 kg (8%), respectively. From these studies, estimates of skeletal muscle loss, expressed per decade, are −0.8 kg for men and −0.4 kg for women. A recent cross-sectional study of older men (ages 49–85 yr) suggested a 1.2 kg/decade loss of appendicular skeletal muscle (34). The longitudinal observations in our elderly cohort with a mean age >70 yr indicate per decade rates of appendicular skeletal muscle loss of −1.6 kg and −0.6 kg in men and women, respectively. The possibility exists that cross-sectional studies have underestimated actual rates of skeletal muscle loss and that losses of skeletal muscle with aging are nonlinear and may accelerate in older age groups.

An important question is what mechanisms were responsible for skeletal muscle mass loss in our subjects. Disuse is a well-recognized cause of skeletal muscle atrophy, although none of our subjects reported physical disabilities. Catabolic conditions, such as rheumatoid arthritis and certain malignancies, are associated with negative protein balance and skeletal muscle atrophy. Weight loss when voluntarily carried out by subjects controlling their body weight also produces loss of lean tissue, including skeletal muscle. We excluded subjects in the present study who were nonambulatory, had physical disabilities, serious chronic illnesses, or who had experienced recent voluntary weight loss. Hence, such mediating factors are likely not the basis for the skeletal muscle atrophy observed in the present study. Our findings do suggest that future mechanistic studies need not exclude from analysis healthy, ambulatory, and weight-stable elderly subjects.

Gender Differences

An important observation in the present study was that the relative rate of skeletal muscle loss in men substantially exceeded the relative rate of bone mineral loss. In women, the relative losses of skeletal muscle and bone proceeded at similar rates. Aging men and women would therefore be anticipated to develop very different musculoskeletal relationships in old age. Falls, weakness, frailty, and ultimately fractures can potentially arise from changes in bone composition and quality, a loss of supporting and protective skeletal muscle, or a combination of the two. Our observations suggest that in elderly men there is a predominant skeletal muscle loss, whereas in women proportional reductions occur in both skeletal muscle and bone. These findings point toward a possible gender difference in the pathophysiology of osteoporotic fractures and also suggest potentially different prevention strategies in men and women.

The absolute loss of leg skeletal muscle mass was greater in men than in women (P < 0.05) over the study period, confirming the earlier cross-sectional findings of Gallagher et al. (12) that were also based on DEXA measurements. Forbes and Reina (10) reviewed available literature on urinary creatinine excretion, a measure of fat-free and skeletal muscle mass, in subjects ranging in age from 1 to 80 yr. Creatinine excretion was lower in older subjects, particularly men, compared with young adults. This observation was supported by a similar analysis of available literature on total body potassium (9). In the 18-yr longitudinal study by Flynn et al. (8), there was a greater total body potassium loss in men (9%) compared with women (2%) between the ages 41 and 60 yr. After age 60 yr, the investigators observed a similar rate of total body potassium loss in men (7%) and women (9%). Although these observations suggest that the magnitude of skeletal muscle loss over time is greater in men than in women, this difference is partially accounted for by the larger initial muscle mass in men as shown in our regression analysis. The remaining gender effect was only borderline significant in the whole cohort (follow-up range 1–10 yr) but became statistically significant in those subjects followed for ≥6 yr.


The current study demonstrates that weight stability in older individuals does not imply body composition stability. The finding of significant muscle loss in what likely presents as a healthier than normal group of elderly adults, while maintaining weight stability, exposes sarcopenia as a silent, progressive phenomenon, similar perhaps to osteoporosis. Although the current study cannot directly attest to the consequences of observed skeletal muscle changes, the loss of leg skeletal muscle mass and reductions in muscle strength may have implications for overall mobility and physical function. Hip fracture risk in older subjects (15, 16) is higher for persons who lose more weight starting at age 50 in men and 65 in women. Conversely, a weight increase of 10% or more reduces the risk (15). An argument could therefore be made that, even in older weight-stable men and women where known leg muscle atrophy is occurring, the risk of hip fracture is likely to be elevated.

It is well established that loss of muscle mass and strength can be reversed with resistance training exercises, even into the seventh decade in older men (11) and frail elderly women (7). Moreover, evidence is available that demonstrates the hypertrophic effects of exercise on muscle, including at older ages (7, 11). It may be necessary to advise elderly subjects to increase specific activities, focusing on resistance training for the maintenance of muscle mass and function.

Study Limitations

The current study population is limited in size and does not represent a random sample of older individuals but rather a convenience sample of healthy survivors. Moreover, we excluded subjects with large weight fluctuations secondary to identifiable mechanisms. As such, our findings cannot be considered representative of the aging population in general. However, the unequivocal finding of an appendicular skeletal muscle mass loss over time in a healthy cohort suggests that the extent of muscle loss in a randomly selected sample would be greater than that observed in our study.

It is possible that older adults who slightly decrease physical activity with age show small decreases in skeletal muscle mass and increases in body fat. A limitation of our study was the lack of information on daily physical activity at both measurement points, thereby preventing us from investigating if a portion of the individual variation in skeletal muscle loss was accounted for by variations in activity.

Although the applied body composition methods were state of the art, our ability to accurately detect small compartmental changes over time was limited. Some of the observed changes in secondary variables (e.g., body cell mass) were only modestly greater than method measurement errors. Although our combined methods in a relatively large sample provide a clear and consistent pattern of group change in appendicular skeletal muscle and associated water and potassium over time (Table 4), results for individuals and sometimes specific methods were variable (e.g., fat-free mass was unchanged in women even though there was a significant decrease in total body water). Improving body composition assessment method accuracy and reproducibility, particularly in light of the potential clinical applicability of these methods, is an important future goal.

Based on our analysis of body composition changes in an elderly cohort living in New York City, using technologically advanced and multiple independent body composition methodologies, we present strong evidence of progressive skeletal muscle loss over a 5-yr observation period. These findings extend earlier cross-sectional and limited longitudinal studies and lend support to the hypothesis that dynamic remodeling of soft tissues occurs even in healthy, ambulatory, weight-stable elderly subjects.


This work was supported in part by National Institutes of Health Grants F32-AG-05679, R29-AG-14715, RO1-AG-13021, and P01-DK-42618.


  • Address for reprint requests and other correspondence: D. Gallagher, Obesity Research Center, 1090 Amsterdam Ave., New York, New York 10025.

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.


View Abstract