Abstract
Purpose
We assessed the discrimination of lean mass estimates that have been adjusted for adiposity for physical functioning deficits and prediction of incident disability.
Methods
Included were 2,846 participants from the Health, Aging and Body Composition Study with available whole-body dual energy absorptiometry measures of appendicular lean mass index (ALMI, kg/m2) and fat mass index (FMI, kg/m2). Age-, sex-, and race-specific Z-Scores and T-Scores were determined by comparison to published reference ranges. ALMI values were adjusted for FMI (ALMIFMI) using a novel published method. Sex-stratified analyses assessed associations between lean mass estimates and the physical performance score, ability to complete a 400-meter walk, grip strength, and incident disability. Dichotomized definitions of low lean for age and sarcopenia were examined and their performance compared to the ALM-to-BMI ratio.
Results
Compared to ALMI T-Scores and Z-Scores, the ALMIFMI scores demonstrated stronger associations with physical functioning, and were similarly associated with grip strength. Greater FMI Z-Scores and T-Scores were associated with poor physical functioning and incident disability. Definitions of low lean for age and sarcopenia using ALMIFMI (compared to ALMI) better discriminated those with poor physical functioning and a greater risk of incident disability. The ALM-to-BMI ratio was modestly associated with grip strength and physical performance, but was not associated with completion of the 400-meter walk or incident disability, independent of adiposity and height.
Conclusion
Estimation of skeletal muscle mass relative to adiposity improves correlations with physical performance and prediction of incident disability suggesting it is an informative outcome for clinical studies.
Keywords: Lean mass, Fat mass, Physical function, Incident disability, Grip strength
Individuals with greater fat mass generally have greater appendicular lean mass (ALM). In addition, patients with greater fat mass demonstrate poor physical functioning despite greater ALM (1). Therefore, the extent of adiposity is likely to confound important associations between skeletal muscle deficits and physical functioning.
Methods have been proposed to adjust ALM for fat mass or body size. Compared with ALM alone, indexes scaling ALM to fat mass or body size are more strongly related to physical functioning (2–4). For example, the Foundation for the National Institutes of Health (FNIH) Sarcopenia Project identified ALM-to-Body Mass Index (BMI) ratio as a potentially useful methodology to adjust lean mass parameters for body size and demonstrated an association with incident mobility impairment (5). However, a number of limitations exist with these methods including not taking into account fat-lean associations that differ age, sex, and racial groups, and lack of adjustment for the effects of height. Oversight of these factors may result in over- or under-estimation of relative muscle loss and lead to residual confounding from demographics, body size, and adiposity.
If adiposity is truly a confounder of associations between skeletal muscle mass and physical function, an adiposity-adjusted estimate would be associated with functional measures independently of fat mass and would correlate with physical functioning better than unadjusted estimates of skeletal muscle mass. Ideally, an improved ALM index would also provide a sense of the severity of the deficits. Recently, our group developed a comprehensive method using National Health and Nutrition Examination Survey (NHANES) data to adjust appendicular lean mass index (ALMI) for fat mass index (FMI) (6). Demonstration of good construct validity of these methods is an important next step.
To date, the majority of definitions of sarcopenia have been based on cut-points for ALMI, typically based on the values in healthy young adults. This is analogous to the use of T-Scores in osteoporosis. We aimed to develop comprehensive definitions of sarcopenia that capture variations with age, sex, race/ethnicity and fat mass in order to capture both the impact of aging and the effects of chronic diseases. This approach allows for comparisons to both age-appropriate (ie, Z-score) and young adult (ie, T-score) populations.
The aims of the current study were to (a) evaluate the validity of adiposity-adjusted ALMI T-Scores (compared to young individuals of same sex and race) and Z-Scores (compared to individuals of the same age, sex, and race) by assessing relationships with disability, physical functioning, strength, and incident disability, (b) to evaluate whether associations were independent of the effects of adiposity, and (c) to compare relationships to those observed with published alternative methods of lean mass adjustments for body size.
Methods
Study Setting
Participants were enrolled in the Health, Aging and Body Composition (Health ABC) study, a prospective observational study of 3,075 well-functioning, community-dwelling older adults aged 70–79 years from 1997 to 1998. Study participants were recruited from a random sample of White and all Black Medicare beneficiaries living in Pittsburgh, Pennsylvania and Memphis, Tennessee. Participants were eligible if they self-reported no difficulty walking a quarter mile, climbing 10 steps, or performing activities of daily living. Persons who reported a life-threatening illness; who were undergoing active cancer treatment; who required an assistive walking device; or who had plans to leave the geographic area within 3 years were excluded. All individuals with available body composition data at baseline were included in this analysis (N = 2,846).
Whole Body DXA Measures of Body Composition
Whole body dual-energy X-ray absorptiometry (DXA) was performed at both the Pittsburgh and the Memphis field centers (Hologic 4500A, version 9.03; Hologic, Inc., Waltham, MA). In addition, bone mineral–free ALM and fat mass were derived from the whole body scan. DXA quality assurance measurements were performed at both study sites to ensure scanner reliability and identical patient scan protocols were used for all participants. For soft tissue, the CVs were 1.0% and 2.1% for whole-body lean mass and fat mass, respectively.
Physical Performance and Disability Outcomes
Health ABC performance battery (7,8)
Details have been previously described (8). In brief, this battery includes five repeated chair stands, progressively more challenging tests of standing balance, a 6-meter walk to determine usual gait speed, and a narrow walk in which participants are instructed to walk between lines of colored tape 20 cm apart at their usual pace. Performance is divided by the maximum possible performance for older adults on each test to create ratio scores that are summed for the four tests to obtain a continuous scale ranging from 0 to 4, with a lower score indicating poorer function. The minimally important difference in the Health ABC Score is estimated to be 0.15 based on previous studies of the Short Physical Performance Battery (9).
Long-distance corridor walk (400-meter walk)
This is a two-stage, self-paced walking test that was designed to measure cardiorespiratory fitness longitudinally in an initially well-functioning cohort of 70-year-olds (10). The first stage consisted of a 2-minutes warm-up walk, in which distance was recorded and the first 20 meters was timed. This stage also served as a stepped-down test for persons unable to walk for a longer period. The second stage consisted of a 400-meter walk, which is about the distance an average health older adult can cover in 6 minutes. Subjects were asked to walk 400 meters as quickly as possible at a pace that they can maintain. Standard encouragement was given throughout the test and time was recorded to the nearest second. A significant proportion of subjects were unable to complete the test (24%) at baseline, and therefore we determined associations with inability to complete the 400-meter walk as a measure of mobility disability as previously defined (11).
Grip strength
Isometric grip strength was measured at baseline using a hand-held dynamometer (JAMAR Technologies, Inc., Hatfield, PA). Two trials were performed for each hand. An average of the trials performed on the strongest hand was used for analyses as previously described (12). Models assessing grip strength were adjusted for height as a measure of body size. Height was strongly associated with grip strength in these models (p < .001).
Incident disability
Adjudicated self-report data on incident physical disability was obtained from interviewer-administered questionnaires every 6 months (13,14). For physical disability, the outcome of interest was time from baseline (Visit 1) to any self-reported disability, which was defined as severe difficulty or inability to walk 1/4 mile and/or climb 10 steps, needing equipment to ambulate, or having any difficulty performing activities of daily living (ie, getting in and out of bed or chairs, bathing or showering, and dressing).
Statistical Analysis
Similar to the generation of BMI, estimates of ALM and fat mass were adjusted for nonlinear relationships with height through the generation of indices (ALMI; FMI [kg/m2]) by dividing values by height-squared. Use of indices facilitates comparison to published reference ranges from the NHANES. Estimates of ALMI and FMI from whole-body DXA were converted to T- and Z-Scores based on reference ranges generated in NHANES (15). Adiposity-adjusted estimates of ALMI (ALMIFMI) were determined using previously published methods that utilize the association of ALMI Z-Scores with FMI Z-Scores in NHANES (6). Briefly, data from NHANES were used to regress ALMI Z-Score on FMI Z-Score (and ALMI T-Score on FMI T-Score). The resulting prediction equations were used to estimate the expected value for ALMI Z-Score and T-Score relative to the FMI Z-Score or T-Score. The deviation from the expected value (observed minus predicted) was used to calculate ALMIFMIZ-Scores and T-Scores. The adjusted ALMIFMIZ-Scores therefore conceptually represent the number of standard deviations above or below the predicted value for a reference group of individuals the same age, sex, race, and FMI Z-Score. The ALMIFMIT-Scores conceptually represent the number of standard deviations above or below the mean for individuals 25 years of age of the same sex, race, and FMI T-Score.
Associations between standard ALMI T-Scores and Z-Scores and ALMIFMIT-Scores and Z-Scores with measures of physical function, and strength were determined using Pearson correlations and using linear regression. The risk of incident disability was determined using Cox Proportional Hazards models. Study site was not a confounder nor were there significant interactions by study site in the primary analyses.
A low lean mass was defined as in previous publications (6). Low lean for age was defined as a lean mass Z-Score of less than or equal to −1 (15.9th percentile). Sarcopenia was defined as lean mass T-Score less than or equal to −2 (2nd percentile). Discrimination of disability and functional limitations for these definitions (based on ALMI or ALMIFMI) was assessed. These results were compared to the discrimination observed with the ALM-to-BMI ratio. These assessments were performed in multivariable models stratified by sex, and adjusted for age, race, height, and FMI Z-Score (in order to assess the associations of each estimate of relative lean mass deficits independent of demographics, height, and adiposity). Models were stratified by sex given previous literature suggesting sex differences in body composition relationships with physical performance (16–18). Models were adjusted for height, since height was associated with functional outcomes and strongly associated with the ALM-to-BMI ratio (R = .86). A greater ALM-to-BMI ratio was also moderately correlated with lower FMI Z-Scores (R = −.29).
Because regression models can assume linear to relationships between variables, we assessed for nonlinear relationships between body composition and physical functioning by assessing the significance of squared terms in regression models and by visualizing the data using lowess curves.
Analyses were performed on STATA 14 software (StataCorp, LP, College Station, TX).
Results
Baseline characteristics are described in Table 1. By design, the sample included older individuals (mean age 74.1) and was 41.7% African American. Body composition Z-Scores suggested that the study population had somewhat lower ALMI and lower FMI than the NHANES reference population independent of differences in age, sex, and race.
Table 1.
Baseline Characteristics of Study Population at Visit 1
| Men | Women | |
|---|---|---|
| N | 1,386 | 1,460 |
| Age | 74.2 (2.87) | 73.9 (2.9) |
| Race (% African American) | 37.1% | 46.0% |
| BMI (kg/m2) | 27.0 (3.9) | 27.5 (5.4) |
| Body composition | ||
| ALMI (kg/m2) | 7.94 (1.02) | 6.51 (1.12) |
| ALMI Z-Score | −0.061 (0.85) | −0.22 (0.89) |
| ALMI Z ≤ −1, N (%) | 176 (12.7%) | 277 (19.0%) |
| ALMI T-Score | −0.70 (0.80) | −0.61 (0.90) |
| ALMI T ≤ −2, N (%) | 77 (5.6%) | 91 (6.2%) |
| FMI (kg/m2) | 7.99 (2.32) | 11.35 (3.55) |
| FMI Z-Score | −0.28 (0.89) | −0.35 (0.90) |
| FMI T-Score | 0.48 (0.63) | 0.21 (0.72) |
| ALMIFMIZ-Score | 0.20 (0.95) | 0.046 (0.99) |
| ALMIFMIZ ≤ −1, N (%) | 127 (9.2%) | 189 (13.0%) |
| ALMIFMIT-Score | −1.29 (0.96) | −1.00 (1.05) |
| ALMIFMIT ≤ −2, N (%) | 312 (22.5%) | 230 (15.8%) |
| Low ALM-to-BMI Ratio, N (%) | 125 (9.0%) | 75 (5.1%) |
| Physical functioning | ||
| Health ABC Performance Score | 2.35 (0.51) | 2.07 (0.52) |
| Completed 400-meter Walk, N (%) | 80.7% | 72.8% |
| Grip strength (kg) | 39.9 (8.3) | 24.6 (5.8) |
| Incident disability in follow-up | 940 (68%) | 1,157 (79%) |
Note: ALMI = appendicular lean mass index; ALMIFMI = Appendicular Lean Mass Index adjusted for Fat Mass Index; BMI = body mass index; FMI = fat mass index; Health ABC = Healthy Aging and Body Composition.
As expected, there was a strong positive correlation between the ALMI and FMI Z-Scores (R: .61, p < .0001] and T-Scores (R: .59, p < .0001). In contrast, the ALMIFMI (adiposity adjusted) Z-Scores were not associated with FMI Z-Scores (R: −.022, p = .23). There was a modest association between the ALMIFMI and FMI T-Scores (R: −.046, p = .01). There was a strong correlation between the ALMIFMI and ALMI Z-Scores (R: .77, p < .0001) and T-Scores (R: .74, p < .0001).
Performance of Adiposity-Adjusted ALMI as Continuous Outcome
Results of the T-Score analysis were largely similar to analyses with Z-Scores and can be found in Supplementary Data. Greater ALMIFMIZ-Scores were positively associated with superior physical performance among men (β: 0.079 (0.050, 0.11) p < .001) (Table 2). This coefficient implies that 1 SD greater ALMIFMIZ-Score is associated with a 0.079-unit higher HABC performance score. In contrast, ALMI Z-Scores were only modestly associated with physical performance among men (β: 0.038 (0.007, 0.070) p = .02) (Table 2). Among women, higher ALMIFMIZ-Scores were not significantly associated with physical performance. However, higher ALMI Z-Scores were paradoxically associated with worse physical performance. These correlations with physical performance among men and women were statistically different for ALMIFMI compared to ALMI (all p for comparison <.0001).
Table 2.
Cross-Sectional Associations Between ALMI (unadjusted) and ALMIFMI (adiposity-adjusted) Z-Scores (and dichotomous definitions of low lean for age) With Physical Functioning, Strength, and, Incident Disability
| Performance Score | Grip Strength (strong hand) | 400-Meter Walk Completed | Incident Disability | |
|---|---|---|---|---|
| β (95% CI) | β (95% CI) | OR (95% CI) | HR (95% CI) | |
| Men | ||||
| ALMI Z (per 1 SD) | 0.033* (0.0007, 0.065) | 2.72*** (2.21, 3.24) | 1.04 (0.89, 1.22) | 1.07 (0.99, 1.16) |
| ALMIFMIZ (per 1 SD) | 0.079*** (0.050, 0.11) | 3.03*** (2.58, 3.48) | 1.14 (0.99, 1.31) | 0.96 (0.90, 1.03) |
| Women | ||||
| ALMI Z (per 1 SD) | −0.077*** (−0.11, −0.046) | 1.21*** (0.85, 1.57) | 0.94 (0.83, 1.07) | 1.21*** (1.12, 1.30) |
| ALMIFMIZ (per 1 SD) | 0.025 (−0.0029, 0.052) | 1.20*** (0.88, 1.51) | 1.17** (1.04, 1.31) | 0.99 (0.93, 1.05) |
| Dichotomous definitions low lean for age | ||||
| β (95% CI) | β (95% CI) | OR (95% CI) | HR (95% CI) | |
| Men | ||||
| Low ALMI Z (low lean for age) | −0.13** (−0.21, −0.051) | −6.17*** (−7.51, −4.83) | 0.68* (0.47, 0.98) | 1.04 (0.86, 1.27) |
| Low ALMIFMIZ (low lean for age) | −0.26*** (−0.36 −0.17) | −6.99*** (−8.55, −5.43) | 0.50 ** (0.33, 0.75) | 1.14 (0.92, 1.41) |
| Women | ||||
| Low ALMI Z (low lean for age) | 0.064 (−0.005, 0.13) | −2.00*** (−2.81, −1.18) | 0.86 (0.64, 1.15) | 0.90 (0.77, 1.04) |
| Low ALMIFMIZ (low lean for age) | −0.12** (−0.21, −0.042) | −2.13*** (−3.09, −1.18) | 0.54*** (0.40, 0.75) | 1.30** (1.10, 1.54) |
Note: ALM= Appendicular Lean Mass; ALMIFMI = Appendicular Lean Mass Index adjusted for Fat Mass Index; CI = confidence interval; OR = odds ratio; HR = Hazard Ratio.
*p < .05; **p < .01; ***p < .001.
Greater ALMIFMIZ-Scores were also associated with greater likelihood of completing of the 400-meter walk among men (OR: 1.14 (0.99, 1.31) p = .08) and women (Table 2). In contrast, the ALMI Z-Scores were not associated (all p > .36). Among men and women, the ALMIFMIZ-Score demonstrated significantly better discrimination of completers than the ALMI Z-Score (AUC 0.54 vs. 0.50; p for comparison <.0001).
Both ALMIFMI and ALMI Z-Scores were similarly positively associated with grip strength in multivariable models (Table 2). The ALMIFMIZ-Scores were not associated with incident disability as a continuous variable in either men or women (Table 2). Higher ALMI Z-Scores were paradoxically associated with a higher risk of incident disability in men (HR: 1.07 (0.99, 1.16) p = .07) and women (HR: 1.21 (1.12, 1.30) p < .001)—however, this was fully attenuated with adjustment for FMI Z-score (Men: HR: 0.95 (0.87–1.03) p = .23; Women: HR: 0.98 (0.90, 1.08) p = .74).
Performance of ALMI and ALMIFMI Definitions of Low Lean for Age
A somewhat greater number of patients were low lean for age (Z-Score ≤ −1) by ALMI (16%) compared to ALMIFMI (11%). There was moderate agreement between ALMI and ALMIFMIZ-Score definitions of low lean for age (Kappa: 0.49, p < .0001). A greater number were defined as sarcopenia (T-Score ≤ −2) based on ALMIFMIT-Score (19.1%) compared to the ALMI T-Score (5.9%). There was moderate agreement between ALMI and ALMIFMI definitions of sarcopenia (Kappa: 0.38, p < .0001). Results of the T-Score analysis (sarcopenia, T-Score ≤ −2) were largely similar and can be found in Supplementary Data.
Table 2 demonstrates that the ALMIFMI definitions of low lean for age were associated with physical performance, grip strength, completion of the 400-meter walk, among both men and women, and with incident disability among women. These associations were either weaker or not present for ALMI definitions of low lean for age. For example, among women, a low ALMIFMIZ-Score was associated with worse physical performance (β: −0.12 (−0.21, −0.042) p = .003). A low ALMI Z-Score was not associated with physical performance (β: 0.064 (−0.005, 0.13) p = .07). Figure 1 illustrates the greater discrimination of poor physical performance using the ALMIFMI definitions among men and women.
Figure 1.
Health ABC Performance Scores among individuals classified by ALMI and ALMIFMI definitions of low lean mass for age (ALMI Z-Score ≤ −1 or ALMIFMIZ-Score ≤ −1), stratified by sex. ABC = Aging and Body Composition; ALMI = Appendicular Lean Mass Index; ALMIFMI = Appendicular Lean Mass Index adjusted for Fat Mass Index.
Independence of Adiposity-Adjusted Measures and Adiposity With Physical Function
There was evidence of a significant nonlinear association (significant squared terms) between ALMIFMIZ-Scores and physical performance and completion of the 400-meter walk among men and women, a nonlinear association with grip strength among men, and a non-linear association with incident disability among women (Supplementary Table 3). In multivariable models, adjusting for age, race, and height, there was also evidence of a non-linear association between FMI Z-Scores with physical performance, completion of the 400-meter walk, incident disability among men and women, and for grip strength among men. These nonlinear associations suggested that higher FMI Z-Scores were more closely tied to adverse outcomes among those with higher FMI Z-Scores. Conversely, lower ALMIFMIZ-Scores were more closely tied to worse outcomes among those with lower ALMIFMI. These nonlinear relationships with the Health ABC Performance Score are illustrated in Figure 2 and full models are shown in Supplementary Table 3.
Figure 2.
Lowess curves assessing the linearity of associations between ALMIFMIZ-Scores and FMI Z-Scores with HABC performance scores demonstrating that FMI Z-Scores are increasingly inversely associated among individuals with higher FMI Z-Scores while ALMIFMIZ-Scores are more strongly positively associated at lower ALMIFMIZ-Scores among both men and women. ALMI = Appendicular Lean Mass Index; ALMIFMI = Appendicular Lean Mass Index adjusted for Fat Mass Index; FMI = Fat Mass Index; HABC = Healthy Aging and Body Composition.
Table 3 demonstrates that in multivariable models adjusting for age, sex, race, FMI Z-Score, a FMI Z-Score2 term, and height, the ALMIFMIZ-Scores as well as the definitions of low lean for age were consistently associated with these outcomes. The magnitude of the association was not attenuated with these adjustments (ie, similar associations compared to the coefficients shown in Table 2).
Table 3.
Multivariable Linear, Logistic, and Cox-Proportional Hazard Models Evaluating Associations Between ALMIFMIZ-Scores, Definitions of Low Lean for Age, and Low ALM-to-BMI Ratio (in separate models) With Physical Functioning, Strength, and Incident Disability With Adjustment for FMI Z-Scores and Height
| Physical Performance Score | Grip Strength | 400-meter Walk Completed | Incident Disability | |
|---|---|---|---|---|
| β (95% CI) | β (95% CI) | OR (95% CI) | HR (95% CI) | |
| Men | ||||
| Model 1: ALMIFMIZ-Score (per SD) | 0.094 (0.067, 0.12)*** | 2.73 (2.31, 3.15)*** | 1.17 (1.01, 1.35)* | 0.96 (0.90, 1.03) |
| Model 2: Low ALMIFMIZ-Score | −0.25 (−0.34, −0.16)*** | −6.24 (−7.68, −4.80)*** | 0.51 (0.34, 0.77)** | 1.05 (0.85, 1.31) |
| Model 3: Low ALM-to-BMI | −0.094 (−0.20, 0.015) | −2.76 (−4.55, −0.98)** | 0.69 (0.40, 1.18) | 0.94 (0.72, 1.23) |
| Women | ||||
| Model 1: ALMIFMIZ-Score (per SD) | 0.012 (−0.013, 0.037) | 1.91 (1.66, 2.16)*** | 1.12 (1.00, 1.26) | 1.01 (0.95, 1.08) |
| Model 2: Low ALMIFMIZ-Score | −0.086 (−0.16, −0.012)* | −4.15 (−4.96, −3.33)*** | 0.59 (0.42, 0.82)** | 1.34 (1.13, 1.58)*** |
| Model 3: Low ALM-to-BMI | −0.16 (−0.28, −0.036)* | −2.52 (−3.68, −1.37)*** | 0.64 (0.37, 1.10) | 0.94 (0.71, 1.25) |
Notes: All models are adjusted for age, black race, height, FMI Z-Score, and FMI Z-Score2. ALM = Appendicular Lean Mass; ALMIFMI = Appendicular Lean Mass Index adjusted for Fat Mass Index; BMI = Body Mass Index; CI = confidence interval; FMI = Fat Mass Index; HABC= Healthy Aging and Body Composition; HR= Hazard Ratio; OR = odds ratio.
*p < .05; **p < .01; ***p < .001.
Performance of Novel Cutoffs Compared to a Low ALM-to-BMI Ratio
There was only fair agreement between the low ALM-to-BMI ratio and the ALMIFMI definitions of low lean for age (Kappa: 0.20, p < .0001) and sarcopenia (Kappa: 0.19, p < .0001). More patients were identified as low lean for age (ALMIFMIZ-Score ≤ −1) compared to the number identified as having a low ALM-to-BMI ratio among men (9.2% vs. 9.0%) and women (13% vs. 5.1%). There were substantial differences in body composition among those with a low ALMIFMIZ-Score compared to individuals with a low ALM-to-BMI ratio. Those with a low ALM-to-BMI ratio had high FMI Z-Scores (mean (SD): 0.70 (0.77)) compared to the overall population and normal ALMI Z-Scores, on average (mean (SD): −0.13 (0.91); Figure 3). In contrast, those who had a low ALMIFMI (Z-Score ≤ −1) had very low ALMI Z-Scores (p < .0001) compared to the overall population (mean (SD): −1.26 (0.80)). As expected, those with low ALMIFMIZ-Scores had greater FMI Z-Scores compared to those with low ALMI Z-Scores (mean (SD): −0.26 (1.00) vs. −1.16 (0.74), p < .0001).
Figure 3.
Summary of ALMI and FMI Z-Scores among those meeting thresholds for low lean for age and low ALM-to-BMI ratio and compared to the overall population. Those with a low ALM-to-BMI ratio have high FMI Z-Scores compared to the overall population. In contrast, those with low ALMI Z-Scores and low ALMIFMIZ-Scores have very low ALMI Z-Scores.
A low ALM-to-BMI ratio was modestly associated with grip strength and the HABC performance score after adjustment for FMI Z-Score, height, and demographics. However, a low ALM-to-BMI ratio was not associated with the odds of completing the 400-meter walk, or with an increased risk of incident disability after adjustment (Table 3). In contrast, as noted above, the ALMIFMI definition of low lean for age was similarly and significantly associated with physical performance, completion of the 400-meter walk, grip strength, and incident disability before and after adjustment for these variables.
Discussion
This study demonstrates the construct validity of measures of skeletal muscle mass that have been adjusted for fat mass in addition to height. The comprehensive method of adjustment applied to these data accounts for age-, sex-, and race-specific associations between ALMI Z-Scores and FMI Z-Scores. Thus these definitions allow us to estimate for a given individual, how does their lean mass compare to the expected range for an individual of the same age, sex, race, and adiposity. Adjustment for adiposity using this method provides a variable that is more clearly associated with physical performance and incident disability while remaining independently associated with these outcomes after adjustment for the impact of fat mass and height.
There are a number of important distinctions between these novel methods and existing methods. Methods described by Newman and colleagues utilized a similar approach using a residual method to adjust lean mass estimates for adiposity (2,4). The current method builds on this work and makes several specific advances. Firstly, these methods allow for a comprehensive adjustment of adiposity that considers that the relationship between lean and fat varies with age, sex, race, and adiposity. Secondly, this method, which is based on national published reference data, does not assume that relationships between lean and fat that are similar in diseased populations (19). Finally, the current method allows for a quantification of deficits in a way that has intrinsic meaning. For example, a patient with an ALMIFMIZ-Score of −1 can be said to have a lean mass at approximately the 16th percentile compared to the NHANES population of similar age, sex, race, and adiposity. Overall, this method is widely applicable to existing data and could be incorporated in software to supply this information to investigators or clinicians utilizing whole-body DXA. These adjusted estimates of lean mass correlate with important measures of physical functioning independent of adiposity, conserving the face-validity of the construct.
The strengths of these novel measures as a construct are emphasized in the comparison of their performance to the ALM-to-BMI ratio. The ALM-to-BMI ratio was correlated with height and fat mass and was not as consistently associated with poor physical functioning and incident disability after adjusting for these variables. By design, the adiposity-adjusted estimates presented here are independent of fat mass and height and correlate with physical functioning independently of these covariables.
This study demonstrated a significant nonlinear association between fat mass and physical functioning and incident disability. While we are not aware of similar previous findings in the literature, previous studies have suggested that the most severely obese individuals are at greatest risk for disability (20,21). This observation may be intuitive: that more obese individuals are impacted to a greater degree by the extent of their excess adiposity. The implications of this observation may also be intuitive, namely that obese individuals are likely to derive the most benefit from weight loss (and loss of fat mass) in terms of physical functioning.
Similarly, this study observed a similar nonlinear association between ALMIFMIZ-Scores and physical functioning. This nonlinear association suggests that relationships between low muscle mass and physical functioning are most notable among those with lower muscle mass. This may also suggest that those with the greatest deficits in these measures potentially have the most to gain from interventions to reverse them.
Stratified analysis by sex suggested differential effects of the ALMIFMIZ-Scores and physical functioning outcomes. For example, the relationship between ALMIFMIZ-Scores and physical performance was stronger among men, while the relationship with incident disability was stronger among women. Differential effects by sex were not a focus of the current validation study. However, differential effects of body composition measures on physical functioning have been previously described and these differences deserve further study (16,18).
There are several limitations to the current study. It was out of the scope of the current analysis to study the impact of comorbid conditions, medication usage, or other factors that might be impacted by body composition. Similarly, whether associations identified in this study are directly causal, or represent a marker of illness and comorbidity is not clear. Further study in this area may help clarify the nature of these associations and the impact of interventions to ameliorate them. There are a number of important strengths that are also worth noting. This is a large study in an at-risk population assessing several distinct and important measures of physical functioning with direct relevance to patients.
In conclusion, adjustment for the confounding effects of fat mass significantly improves the correlation of estimates of lean mass with physical performance and incident disability. The validation of the construct of low lean mass relative to fat mass has implications for clinical research and potentially in clinical practice. The comprehensive approach presented here can be applied to existing data and used in prospective studies as an alternative outcome that may have greater relevance to patients, at least with regard to disability.
Supplementary Material
Supplementary data is available at The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences online.
Funding
This research was supported by National Institute on Aging Contracts N01-AG-6-2101; N01-AG-6-2103; N01-AG-6-2106; NIA grant R01-AG028050, and National Institute of Nursing Research grant R01-NR012459. J.F.B. is supported by a Veterans Affairs Clinical Science Research & Development Career Development Award (IK2 CX000955). D.R.W. was supported by National Institute of Health grant K12HD068373.
Conflicts of Interest
The authors have no conflicts to disclose.
Supplementary Material
Acknowledgments
J.F.B. would like to acknowledge the support of a Veterans Affairs Clinical Science Research & Development Career Development Award (IK2 CX000955). The contents of this work do not represent the views of the Department of the Veterans Affairs or the United States Government. This research was supported in part by the Intramural Research Program of the NIH, National Institute on Aging.
References
- 1. Batsis JA, Zbehlik AJ, Barre LK, Bynum J, Pidgeon D, Bartels SJ. Impact of obesity on disability, function, and physical activity: data from the Osteoarthritis Initiative. Scand J Rheumatol. 2015;44:495–502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Delmonico MJ, Harris TB, Lee JS et al. ; Health, Aging and Body Composition Study Alternative definitions of sarcopenia, lower extremity performance, and functional impairment with aging in older men and women. J Am Geriatr Soc. 2007;55:769–774. doi:10.1111/j.1532-5415.2007.01140.x. [DOI] [PubMed] [Google Scholar]
- 3. Dufour AB, Hannan MT, Murabito JM, Kiel DP, McLean RR. Sarcopenia definitions considering body size and fat mass are associated with mobility limitations: the Framingham Study. J Gerontol A Biol Sci Med Sci. 2013;68:168–174. doi: 10.1093/gerona/gls109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Newman AB, Kupelian V, Visser M et al. ; Health ABC Study Investigators Sarcopenia: alternative definitions and associations with lower extremity function. J Am Geriatr Soc. 2003;51:1602–1609. [DOI] [PubMed] [Google Scholar]
- 5. McLean RR, Shardell MD, Alley DE et al. Criteria for clinically relevant weakness and low lean mass and their longitudinal association with incident mobility impairment and mortality: the foundation for the National Institutes of Health (FNIH) sarcopenia project. J Gerontol A Biol Sci Med Sci. 2014;69:576–583. doi:10.1093/gerona/glu012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Weber D, Long J, Leonard MB, Zemel B, Baker JF. Development of novel methods to define deficits in appendicular lean mass relative to fat mass. PLoS One. 2016;11:e0164385. doi:10.1371/journal.pone.0164385. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Brach JS, Simonsick EM, Kritchevsky S, Yaffe K, Newman AB; Health, Aging and Body Composition Study Research Group The association between physical function and lifestyle activity and exercise in the health, aging and body composition study. J Am Geriatr Soc. 2004;52:502–509. doi:10.1111/j.1532-5415.2004.52154.x. [DOI] [PubMed] [Google Scholar]
- 8. Simonsick EM, Newman AB, Nevitt MC et al. ; Health ABC Study Group Measuring higher level physical function in well-functioning older adults: expanding familiar approaches in the Health ABC study. J Gerontol A Biol Sci Med Sci. 2001;56:M644–M649. [DOI] [PubMed] [Google Scholar]
- 9. Kwon S, Perera S, Pahor M et al. What is a meaningful change in physical performance? Findings from a clinical trial in older adults (the LIFE-P study). J Nutr Health Aging. 2009;13:538–544. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Simonsick EM, Montgomery PS, Newman AB, Bauer DC, Harris T. Measuring fitness in healthy older adults: the Health ABC Long Distance Corridor Walk. J Am Geriatr Soc. 2001;49:1544–1548. [DOI] [PubMed] [Google Scholar]
- 11. Vestergaard S, Patel KV, Walkup MP et al. Stopping to rest during a 400-meter walk and incident mobility disability in older persons with functional limitations. J Am Geriatr Soc. 2009;57:260–265. doi:10.1111/j.1532-5415.2008.02097.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Klepin HD, Geiger AM, Tooze JA et al. ; Health, Aging and Body Composition Study Physical performance and subsequent disability and survival in older adults with malignancy: results from the health, aging and body composition study. J Am Geriatr Soc. 2010;58:76–82. doi:10.1111/j.1532-5415.2009.02620.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Chen DS, Betz J, Yaffe K et al. ; Health ABC study Association of hearing impairment with declines in physical functioning and the risk of disability in older adults. J Gerontol A Biol Sci Med Sci. 2015;70:654–661. doi:10.1093/gerona/glu207. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Sanders JL, Boudreau RM, Penninx BW et al. ; Health ABC Study Association of a Modified Physiologic Index with mortality and incident disability: the Health, Aging, and Body Composition study. J Gerontol A Biol Sci Med Sci. 2012;67:1439–1446. doi:10.1093/gerona/gls123. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Kelly TL, Wilson KE, Heymsfield SB. Dual energy X-Ray absorptiometry body composition reference values from NHANES. PLoS One. 2009;4:e7038. doi:10.1371/journal.pone.0007038. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Tseng LA, Delmonico MJ, Visser M et al. Body composition explains sex differential in physical performance among older adults. J Gerontol A Biol Sci Med Sci. 2014;69:93–100. doi:10.1093/gerona/glt027. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Lisko I, Stenholm S, Raitanen J et al. Association of body mass index and waist circumference with physical functioning: the vitality 90+ study. J Gerontol A Biol Sci Med Sci. 2015;70:885–891. doi:10.1093/gerona/glu202. [DOI] [PubMed] [Google Scholar]
- 18. Bouchard DR, Héroux M, Janssen I. Association between muscle mass, leg strength, and fat mass with physical function in older adults: influence of age and sex. J Aging Health. 2011;23:313–328. doi:10.1177/0898264310388562. [DOI] [PubMed] [Google Scholar]
- 19. Baker JF, Von Feldt J, Mostoufi-Moab S et al. Deficits in muscle mass, muscle density, and modified associations with fat in rheumatoid arthritis. Arthritis Care Res (Hoboken). 2014;66:1612–1618. doi:10.1002/acr.22328. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Sharkey JR, Ory MG, Branch LG. Severe elder obesity and 1-year diminished lower extremity physical performance in homebound older adults. J Am Geriatr Soc. 2006;54:1407–1413. doi:10.1111/j.1532-5415.2006.00842.x. [DOI] [PubMed] [Google Scholar]
- 21. De Stefano F, Zambon S, Giacometti L et al. Obesity, muscular strength, muscle composition and physical performance in an elderly population. J Nutr Health Aging. 2015;19:785–791. doi:10.1007/s12603-015-0482-3. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.