Abstract
Objective
The relationship between body composition and mortality in frail older people is unclear. We used dual-x-ray absorptiometry (DXA) data to examine the association between dynamics in whole-body composition and appendicular (4 limbs) and central (trunk) compartments and all-cause mortality in frail older women.
Design
Prospective study with up to 19 years of follow up.
Setting
Community dwelling older (≥65) women.
Participants
876 frail older participants of the Women's Health Initiative Observational Study with a single measure of body composition and 581 participants with two measures.
Measurements
Frailty was determined using modified Fried's criteria. All-cause mortality hazard was modeled as a function of static (single-occasion) or dynamic changes (difference between two time points) in body composition using Cox regression.
Results
Analyses adjusted for age, ethnicity, income, smoking, cardiovascular disease, diabetes, stroke, number of frailty criteria and whole-body lean mass showed progressively decreased rates of mortality in women with higher appendicular fat mass (FM) (P for trend=0.01), higher trunk FM (P for trend=0.03) and higher whole-body FM (P for trend=0.01). The hazard rate ratio for participants with more than a 5% decline in FM between two time points was 1.91; 1.67 and 1.71 for appendicular, trunk and whole-body compartment respectively as compared to women with relatively stable adiposity (p<0.05 for all). Dynamics of more than 5% in lean mass were not associated with mortality.
Conclusion
Low body fat or a pronounced decline in adiposity is associated with increased risks of mortality in frail older women. These results indicate a need to re-evaluate healthy weight in persons with frailty.
Key words: Frailty, body composition, fat mass, lean mass, mortality
Introduction
Frailty is conceptually defined as a state of high vulnerability to adverse health outcomes including disability, morbidity, and mortality (1, 2, 3). Derangements in inflammatory, endocrine, coagulation, and metabolic systems in frail persons have been repeatedly demonstrated (4, 5). Yet, frail older adults do not universally experience adverse events, and some sustain a relatively uninterrupted lifespan. Although many factors might explain this disparity in health outcomes, we recently completed research that suggested that frail older women with a higher body mass index (BMI) had a lower mortality rate than their counterparts with normal BMI (29). In contrast to BMI, which is an imperfect measure of adiposity, especially among older adults (6, 7, 8), dual-x-ray absorptiometry (DXA) provides direct and valid quantitative measurement of fat and lean mass and allows differentiation between body compartments (9). Because lean and fat mass account for 95% of body weight, it is informative to validate our previous findings using direct measures of adiposity.
Body-composition change closely relates to aging and frailty. A previous report on adults age 65 to 102 years participating in the “Invecchiare in Chianti” study showed that frail participants have lower muscle density and muscle mass and higher fat mass than nonfrail persons (10). Extrapolating from these findings, the authors suggested that low muscle mass and high fat mass would also be linked to increased mortality in frail older men and women. This assertion, however, was refuted in a later “inChianti” report that showed a nonsignificant association of lower lean or higher fat mass area with mortality in older participants and in a subsample (n = 436) of those with frailty (11). These null findings, however, should be interpreted in light of methodological limitations. First, body-composition indexes were determined using peripheral quantitative computerized tomography (pQCT) of the right leg. Although reliable and methodologically sound, pQCT is prone to obscure information on body composition in separate body compartments. Such a partitioning into appendicular and central body-composition indexes was informative in predicting mortality in older men (12). Second, frailty was defined as a report of one or more frailty criteria, rather than using a conventional three-criteria threshold (1). Finally, change in body composition over time rather than cross-sectional measures might have provided more-nuanced information about the relationship between lean and fat mass dynamics and health outcomes. To the best of our knowledge, the only study that described longitudinal (5-year) change in muscle mass used a small subsample of 30 frail older adults participating in the Korean Longitudinal Study of Health and Aging, which showed that frailty status was independently associated with a modest decline in lean mass index of 0.46 kg/m2 (95% CI, −0.84; −0.08 (13)). Yet, the study's authors also acknowledged methodological limitations, including bioimpendance analysis to estimate body composition (bioelectrical impedance is a valid estimation of whole-body fat-free mass, but is based on an approximation using age- and sex-specific equations (14)), not accounting for mortality in a follow-up and relatively small number of frail participants.
Therefore, to address the abovementioned limitations, we investigated the association of static (single-occasion) and dynamic changes (difference between two time points) in appendicular, central, and whole-body composition measures, as assessed by DXA, with all-cause mortality in a large sample of community-dwelling frail older women participating in the Women's Health Initiative Observational Study (WHI OS). We hypothesized that frail women who have lower lean mass or who demonstrated a pronounced (change of more than 5% from a previous measure) decline in lean mass would be at an increased risk of death compared with their more-robust frail counterparts.
Methods
Study Population
The WHI OS comprised 93,676 women aged 50 to 79 years at baseline (1993–1998) from 40 U.S. clinical centers. Details of the WHI study design and baseline characteristics are reported elsewhere (15). The WHI study was approved by the institutional review boards at all 40 clinic sites, and all participants provided written informed consent at baseline.
At baseline and at the two subsequent follow-up clinic visits (3 and 6 years after baseline), OS participants recruited at the Pittsburgh, PA; Birmingham, AL; and Tucson and Phoenix, AZ, WHI clinical centers completed questionnaires on demographic medical and psychosocial characteristics, clinically provided weight and height measures, and received whole-body DXA scans.
There were 876 WHI OS women at least 65 years of age at the Year 3 clinical visit with complete data to characterize frailty using Fried's criteria, meeting the definition of frail (defined later) and who had anthropometric and DXA data. Of these, 581 also had DXA measures at the Year 6 follow-up visit (Supplementary Figure 1).
Figure 1A–1C.
Relationship between combined categories of change in lean and fat mass and mortality. CI = confidence interval; HR = hazard ratio
Frailty Ascertainment
Frailty was operationally defined congruently with Fried's definition (1) as the presence of three or more of the following criteria: muscle weakness, slow walking speed, exhaustion, low physical activity, and unintentional weight loss. This operationalization for frailty was adapted and validated in the WHI and has been extensively used in the WHI OS cohort (3, 16, 17). Briefly, WHI operationalization of frailty is similar to the Fried's index with a notable difference that RAND-36 physical function index substituted for slowness and weakness criteria. To align the scoring range with the Fried's frailty index, poor physical function was scored as 2 points because this scale is a proxy measure of the muscle strength and walking ability components. Similarly to the Fried's definition zero to five summary scores were calculated and a score of 3 or more indicated frailty. The participants were limited to only those meeting the definition of frail.
Dual-x-ray Absorptiometry (DXA) Measures
Using the same standard protocol at Pittsburgh, PA; Birmingham, AL; and Tucson and Phoenix, AZ clinical sites, whole-body scans (including the head) using the fan-beam mode from Hologic QDR scanners (QDR 2000, 2000+, or 4500W; Hologic, Waltham, MA) were obtained. The imaging results provided values for masses of lean and fat tissues for the whole body and specific regions. Body-compartment-specific indexes were calculated as follows: (a) appendicular lean and fat mass were calculated as the sum of respective lean or fat tissues in both arms and legs, (b) central lean and fat mass were defined as the lean and fat soft tissue of the trunk, and (c) totalbody lean and fat mass indicated the whole-body-composition indexes.
A change over time was calculated as the difference between Year 6 and Year 3 measures. Furthermore, because a 5% change in body-composition index was found to be informative in predicting mortality in other observational studies (18, 30), we constructed three categories of change: more than a 5% decline, more than a 5% increase, and a change of not more than 5% from a previous measure (stability) for each body-compartment-specific index. Finally, lean and fat mass indexes were calculated as body-compartment-specific scores in kg divided by height in meters squared.
All-Cause Mortality
Medical updates were collected annually by mail. Participants' death records were adjudicated by study physicians using hospital records, autopsy reports, and death certificates. Periodic checks of the National Death Index (NDI) for all participants, including those lost to follow-up, were performed. At the time of this analysis, the latest WHI data on mortality were available through December 1, 2013.
Covariates
Baseline data on demographic variables (race/ethnicity, family income), smoking status (Never smoker; Past Smoker; Current smoker), and history of common diseases in older age such as diabetes, stroke, and cardiovascular disease were obtained by self-report. During Year 3 clinical visits, trained and certified staff collected anthropometric measurements. Weight to the nearest 0.1 kg and height to the nearest 0.1 cm were used to compute BMI.
Statistical Approach
First we performed bivariate comparison between 581 WHI OS frail older participants who had both Year 3 and Year 6 DXA measures and 295 women who did not have Year 6 DXA data available by using a t test for variables with a Gaussian distribution and a chi-square test for categorical variables. Second, to examine the multivariate relationship between appendicular, trunk, and whole-body-mass indexes measured at Year 3 and all-cause mortality outcome, we estimated hazard ratios (HRs) and 95% confidence intervals (CIs) using separate Cox proportional hazards models. Exposure variables were modeled using quartiles according to the observed distribution. Survival models were first adjusted for age and then for ethnicity, income, smoking, cardiovascular disease, diabetes, stroke and number of frailty criteria. Furthermore, to understand whether mortality risks associated with fat mass were independent of lean mass, we included in our fully adjusted models further adjustment for a continuous measure of wholebody lean mass. The time-to-event was defined as number of years from WHI Year 3 follow-up visit to death from any cause. For women who were alive at the last known contact, censoring time was defined as number of years from WHI Year 3 follow-up visit to last known contact or the date of last available NDI search, whichever occurred later. Furthermore, because some uncertainties persist in epidemiological literature about whether the relationship between body composition and mortality might be distorted by subclinical symptoms, we conducted sensitivity analyses by censoring our sample for early mortality, defined as death within 2 years of Year 3 clinical visit. Third, we used a separate set of Cox models to evaluate the association between all-cause mortality after Year 6 clinical visit and categories of change in body-compartment-specific indexes occurring between Years 3 and 6. Similar to the abovementioned static models, dynamic models were first adjusted for age and then for ethnicity, income, smoking, cardiovascular disease, diabetes, stroke, number of frailty criteria and Year 3 BMI. Dynamics in body composition were modeled using three categories of change, with “stability within 5 %” being the reference group. Sensitivity analyses by excluding early death after Year 6 were also performed. Finally, to examine the proportionality of longitudinal dynamics in lean and fat masses, we constructed separate categories of change to indicate any gain or any loss in body-composition scores between Year 3 and Year 6. These categories were then used to model mortality hazards as a function of a joined change in lean and fat mass indexes using Cox regressions adjusted for age. Shoenfeld residuals were used to test proportionality across body-composition measures and for the models overall. All statistical analysis was completed using STATA, version 11.2 (StataCorp, College Station, TX).
Results
The mean age of the study sample at Year 3 clinical visit was 72.3 years. Most participants were White (n = 690, 79%), only 6.5% were smokers at baseline, and mean BMI was 29.3 kg/m2. There were significant differences in baseline and Year 3 characteristics by age, income, history of cardiovascular disease and number of frailty criteria between frail women who had both Year 3 and Year 6 X-ray data available versus those with non-available measure in Year 6 (Table 1).
Table 1.
Baseline and Year 3 Characteristics of Frail Older Participants in the Women's Health Initiative Observational Study (WHI OS) According to Availability of Year 3 and Year 6 DXA Measures
| Characteristics | Year 3 (n = 876) | Year 3 and Year 6 Measures (n = 581) | Not Available in Year 6 (n = 295) | p Value* |
|---|---|---|---|---|
| Age‡, yr (SD) | 72.30 (4.55) | 72.04 (4.44) | 72.80 (4.72) | 0.02 |
| Height‡, cm (SD) | 160.00 (6.23) | 160.05 (6.21) | 159.91 (6.28) | 0.77 |
| White‡, n (%) | 690 (78.95) | 457 (78.66) | 233 (79.52) | 0.77 |
| Income‡, n (%) | 0.01 | |||
| ≤$20000 | 310 (35.96) | 214 (37.22) | 96 (33.45) | |
| $20000-$50000 | 354 (41.07) | 246 (42.78) | 108 (37.63) | |
| >$50000 | 198 (22.97) | 115 (20.00) | 83 (28.92) | |
| Weight‡, kg (SD) | 75.26 (17.51) | 74.79 (16.89) | 76.18 (18.65) | 0.27 |
| BMI, kg‡/m2 (SD) | 29.26 (6.30) | 29.13 (6.26) | 29.51 (6.40) | 0.41 |
| BMI categories, n (%) | 0.61 | |||
| Underweight | 10 (1.15) | 6 (1.03) | 4 (1.37) | |
| Normal weight | 213 (24.43) | 149 (25.69) | 64 (21.92) | |
| Overweight | 304 (34.86) | 198 (34.14) | 106 (36.30) | |
| Obese I | 213 (24.43) | 139 (23.97) | 74 (25.34) | |
| Obese II | 77 (8.83) | 55 (9.48) | 22 (7.53) | |
| Obese III | 55 (6.31) | 33 (5.69) | 22 (7.53) | |
| Smoker‡, n (%) | 56 (6.50) | 31 (5.44) | 25 (8.56) | 0.08 |
| CVD‡, n (%) | 266 (31.86) | 163 (29.53) | 103 (36.40) | 0.04 |
| Diabetes‡, n (%) | 81 (9.28) | 50 (8.62) | 31 (10.58) | 0.35 |
| Stroke‡, n (%) | 56 (6.39) | 42 (7.23) | 14 (4.75) | 0.16 |
| Number of Frailty criteria‡, n (SD) | 3.49 (0.59) | 3.45 (0.57) | 3.58 (0.61) | 0.002 |
| Frailty criteria‡, % | ||||
| Low Physical Activity§, n (%) | 516 (58.90) | 317 (54.56) | 199 (67.46) | <0.001 |
| Weight loss*, n (%) | 137 (15.64) | 89 (15.32) | 48 (16.27) | 0.71 |
| Low Physical Function{, n (%) | 866 (98.86) | 575 (98.97) | 291 (98.64) | 0.67 |
| Fatigue#, n (%) | 676 (77.17) | 449 (77.28) | 227 (76.95) | 0.92 |
| Body-composition measures** | ||||
| ALM, kg (SD) | 14.62 (3.05) | 14.65 (2.92) | 14.57 (3.30) | 0.73 |
| AFM, kg (SD) | 17.16 (6.15) | 17.01 (6.07) | 17.45 (6.31) | 0.31 |
| TLM, kg (SD) | 19.61 (2.89) | 19.57 (2.74) | 19.68 (3.18) | 0.57 |
| TFM, kg (SD) | 16.13 (6.45) | 16.05 (6.38) | 16.29 (6.59) | 0.60 |
| WBLM, kg (SD) | 36.90 (5.67) | 36.86 (5.45) | 36.99 (6.09) | 0.74 |
| WBFM, kg (SD) | 34.60 (12.01) | 34.40 (11.87) | 35.00 (12.30) | 0.49 |
| ALM/height2, kg/m2 (SD) | 5.70 (1.12) | 5.71 (1.08) | 5.67 (1.21) | 0.77 |
| AFM/height2, kg/m2 (SD) | 6.69 (2.36) | 6.63 (2.35) | 6.81 (2.38) | 0.29 |
| TLM/height2, kg/m2 (SD) | 7.65 (1.00) | 7.63 (0.93) | 7.67 (1.12) | 0.43 |
| TFM/height2, kg/m2 (SD) | 6.29 (2.46) | 6.25 (2.44) | 6.36 (2.51) | 0.53 |
| WBLM/height2, kg/m2 (SD) | 14.40 (2.00) | 14.37 (1.91) | 14.45 (2.16) | 0.59 |
| WBFM,/height2, kg/m2 (SD) | 13.49 (4.58) | 13.41 (4.56) | 13.65 (4.62) | 0.45 |
Notes: AFM = appendicular fat mass; ALM = appendicular lean mass; DXA = dual-x-ray absorptiometry; SD = standard deviation; WBFM = whole-body fat mass; WBLM = whole-body lean mass; TFM = trunk fat mass; TLM = trunk lean mass; WHI OS = Women's Health Initiative Observational Study.
Bivariate comparison between those who had both Year 3 and Year 6 DXA measures and those with missing measure in Year 6. † Year 3 measure.
Baseline measure.
The lowest quartile on the WHI physical-activity questionnaire.
Weight loss >5% between Year 3 and baseline, and a “yes” response to the question “in the past two years, did you lose fve or more pounds not on purpose at any time?” { RAND-36 Physical Function scale score below 75. # RAND-36 Vitality scale below 55.
A total of 228 deaths from all causes were observed over a mean follow-up of 11.5 years after Year 3 clinical visit, with a maximum follow-up of 18.8 years. Main cause of death was cardiovascular disease (37%), cancer (23%), injury (3%) and other (37%). In general, crude rates of death, expressed in number of deaths per 1000 person-years, demonstrated a decrease in mortality with higher fat mass (Table 2). Hazard ratios in models adjusted for age, ethnicity, income, smoking, cardiovascular disease, diabetes, stroke, number of frailty criteria and whole-body lean mass showed progressively decreased rates of mortality in women with higher appendicular FM (P for trend=0.01), higher trunk FM (P for trend=0.03) and higher whole-body FM (P for trend=0.01). A similar set of analyses using lean mass indexes yielded insignificant results. Sensitivity analyses by excluding women with early death did not substantially change the results (Supplementary Table 1)
Table 2.
Hazard Ratios for All-Cause Mortality in Frail Older Participants in the Women's Health Initiative Observational Study (WHI OS) According to Year 3 Body-Composition Measures (n=876)
| Variable | Deaths/Peson-Years | Incidence Rate per 1,000 Person-Years | Mo HR (95% CI) | del 1† p Value for trend | Mo HR (95% CI) | del 2‡ p Value for trend |
|---|---|---|---|---|---|---|
| ALM/height2, kg/m2 | ||||||
| QRT1 | 61/2,403 | 25.4 | Ref | 0.22 | Ref | 0.17 |
| QRT2 | 52/2,539 | 20.5 | 0.75 (0.52–1.09) | 0.72 (0.48–1.07) | ||
| QRT3 | 64/2,513 | 25.5 | 0.93 (0.65–1.32) | 0.99 (0.68–1.44) | ||
| QRT4 | 51/2,613 | 19.5 | 0.73 (0.50–1.06) | 0.66 (0.44–1.00) | ||
| AFM/height2, kg/m2 | ||||||
| QRT1 | 70/2,481 | 28.2 | Ref | 0.01 | Ref | 0.013* |
| QRT2 | 60/2,409 | 24.9 | 0.87 (0.62–1.23) | 0.76 (0.52–1.10) | ||
| QRT3 | 50/2,571 | 19.4 | 0.71 (0.49–1.01) | 0.71 (0.48–1.05) | ||
| QRT4 | 48/2,610 | 18.4 | 0.65 (0.45–0.95) | 0.55 (0.35–0.88) | ||
| TLM/height2, kg/m2 | ||||||
| QRT1 | 54/2,493 | 21.7 | Ref | 0.95 | Ref | 0.93 |
| QRT2 | 55/2,416 | 22.8 | 1.07 (0.74-1.56) | 1.07 (0.71-1.59) | ||
| QRT3 | 64/2,580 | 24.8 | 1.16 (0.81–1.67) | 1.11 (0.75–1.64) | ||
| QRT4 | 55/2,578 | 21.3 | 1.04 (0.71–1.52) | 1.02 (0.67–1.55) | ||
| TFM/height2, kg/m2 | ||||||
| QRT1 | 69/2,514 | 27.4 | Ref | 0.11 | Ref | 0.03* |
| QRT2 | 59/2,469 | 23.9 | 0.90 (0.63–1.27) | 0.92 (0.64–1.32) | ||
| QRT3 | 50/2,473 | 20.2 | 0.78 (0.54–1.13) | 0.83 (0.55–1.23) | ||
| QRT4 | 50/2,615 | 19.1 | 0.76 (0.53–1.10) | 0.60 (0.38–0.96) | ||
| WBLM/height2, kg/m2 | ||||||
| QRT1 | 58/2,435 | 23.8 | Ref | 0.58 | Ref | 0.38 |
| QRT2 | 55/2,534 | 21.7 | 0.93 (0.64–1.34) | 0.92 (0.62–1.36) | ||
| QRT3 | 63/2,450 | 25.7 | 1.10 (0.77–1.58) | 1.14 (0.78–1.67) | ||
| QRT4 | 52/2,650 | 19.6 | 0.84 (0.58–1.24) | 0.76 (0.49–1.17) | ||
| WBFM/height2, kg/m2 | ||||||
| QRT1 | 78/2,459 | 31.7 | Ref | 0.02 | Ref | 0.01* |
| QRT2 | 53/2,500 | 21.2 | 0.70 (0.49–1.00) | 0.75 (0.52–1.09) | ||
| QRT3 | 47/2,541 | 18.5 | 0.62 (0.43–0.88) | 0.57 (0.38–0.85) | ||
| QRT4 | 50/2,566 | 19.5 | 0.66 (0.46–0.95) | 0.59 (0.37–0.94) | ||
Notes: AFM = appendicular fat mass; ALM = appendicular lean mass; CI = confdence interval; HR = hazard ratio; QRT = quartile; WBFM = whole-body fat mass; WBLM = whole-body lean mass; TFM = trunk fat mass; TLM = trunk lean mass.
Adjusted for age.
Adjusted for age, income, ethnicity, smoking, cardiovascular disease, diabetes, stroke, number of frailty criteria.
Adjusted for age, income, ethnicity, smoking, cardiovascular disease, diabetes, stroke, number of frailty criteria, and whole-body lean mass.
Descriptive data and HRs for mortality according to a 3-year difference in body-composition measures are presented in Table 3. Comparing nonsurvivors and survivors, a consistent picture emerged in that nonsurvivors had a steeper decline in bodycomposition scores. For example, nonsurvivors had -0.75% and -2.31% 3-year decline in whole-body lean and fat mass indexes respectively versus 0.51% and -0.90%, a comparable change among survivors. Contrasting categories of change in lean versus fat tissues, the results showed that a higher proportion of women maintained a change within 5% in lean mass. With regard to fat mass, frail participants were distributed fairly proportionally across the “>5% decline,” “change within 5%,” and “>5% increase” groups. In general, crude rates of death, expressed in number of deaths per person-year, demonstrated an increase in all-cause mortality in women with greater than a 5% decline in lean or fat mass. However, in models adjusted for age, ethnicity, income, smoking, cardiovascular disease, diabetes, stroke, number of frailty criteria and Year 3 BMI only a greater than 5% decline in fat mass category showed the association with mortality. For instance, frail persons who had greater than a 5% decline in appendicular fat mass had an almost twofold higher rate of death (95% CI, 1.18–3.06) than frail women with relatively stable appendicular adiposity. More than a 5% change over time in lean mass was not significantly associated with mortality. Sensitivity analyses by excluding women with early death after Year 6 were consistent with primary models (Supplementary Table 2)
Table 3.
Descriptive Characteristics and Hazard Ratios for All-Cause Mortality in Frail Older Participants in the Women's Health Initiative Observational Study (WHI OS) According to Year 6 to Year 3 Changes in Body-Composition Measures (n=581)
| Model 1† |
Model 2‡ |
|||||||
|---|---|---|---|---|---|---|---|---|
| Variable | Survivors (n = 461) | Non Survivors (n = 120) | Death/Person-Years | Incidence Rate per 1,000 Person-Years | HR (95% CI) | p Value | HR (95% CI) | p Value |
| ALM | ||||||||
| Change in ALM, % (SD) | 2.18 (10.25) | 0.16 (8.82) | ||||||
| Categories of change, n (%) | ||||||||
| >5% decline | 101 (21.91) | 31 (25.83) | 31/1,733 | 17.9 | 1.09 (0.70-1.68) | 0.71 | 1.16 (0.72–1.85) | 0.54 |
| Change within 5% | 228 (49.46) | 64 (53.33) | 64/3,787 | 16.9 | Ref | Ref | ||
| >5% increase | 132 (28.63) | 25 (20.83) | 25/1,868 | 13.4 | 1.00 (0.62-1.59) | 0.99 | 0.98 (0.59–1.62) | 0.93 |
| AFM | ||||||||
| Change in AFM, % (SD) | −1.20 (12.52) | −1.62 (14.23) | ||||||
| Categories of change, n (%) | ||||||||
| >5% decline | 150 (32.54) | 48 (40.00) | 48/2,352 | 20.4 | 2.08 (1.34-3.21) | 0.001 | 1.91 (1.18–3.06) | 0.04 |
| Change within 5% | 179 (38.83) | 37 (30.83) | 37/2,873 | 12.9 | Ref | Ref | ||
| >5% increase | 132 (28.63) | 35 (29.17) | 35/2,163 | 16.2 | 1.29 (0.81-2.06) | 0.29 | 1.21 (0.75–2.09) | 0.39 |
| TLM | ||||||||
| Change in TLM, % (SD) | 1.53 (6.97) | −0.07 (6.34) | ||||||
| Categories of change, n (%) | ||||||||
| >5% decline | 69 (14.97) | 28 (23.33) | 28/1,314 | 21.3 | 1.35 (0.87–2.09) | 0.19 | 1.14 (0.71–1.85) | 0.58 |
| Change within 5% | 262 (56.83) | 67 (55.83) | 67/4,232 | 15.8 | Ref | Ref | ||
| >5% increase | 130 (28.20) | 25 (20.83) | 25/1,842 | 13.6 | 1.07 (0.67–1.70) | 0.78 | 1.08 (0.64–1.80) | 0.78 |
| TFM | ||||||||
| Change in TFM, % (SD) | −2.53 (18.59) | −4.36 (18.67) | ||||||
| Categories of change, n (%) | ||||||||
| >5% decline | 195 (42.30) | 54 (45.00) | 54/2,992 | 18.0 | 1.31 (0.85–2.01) | 0.22 | 1.67 (1.05–2.66) | 0.03 |
| Change within 5% | 136 (29.50) | 35 (29.17) | 35/2,218 | 15.8 | Ref | Ref | ||
| >5% increase | 130 (28.20) | 31 (25.83) | 31/2,178 | 14.2 | 0.83 (0.51–1.34) | 0.44 | 0.85 (0.50–1.46) | 0.56 |
| WBLM | ||||||||
| Change in TBLM, % (SD) | 0.51 (5.60) | −0.75 (4.96) | ||||||
| Categories of change, n (%) | ||||||||
| >5% decline | 63 (13.67) | 22 (18.33) | 22/1,107 | 19.9 | 1.35 (0.83–2.19) | 0.22 | 0.97 (0.55–1.67) | 0.90 |
| Change within 5% | 316 (68.55) | 84 (70.00) | 84/5,178 | 16.2 | Ref | Ref | ||
| >5% increase | 82 (17.79) | 14 (11.67) | 14/1,103 | 12.7 | 1.01 (0.57–1.79) | 0.97 | 0.81 (0.43–1.55) | 0.53 |
| WBFM | ||||||||
| Change in TBFM, % (SD) | −0.90 (13.20) | −2.31 (14.69) | ||||||
| Categories of change, % | ||||||||
| >5% decline | 163 (35.36) | 45 (37.50) | 45/2,472 | 18.2 | 1.36 (0.89–2.06) | 0.15 | 1.71 (1.09–2.69) | 0.02 |
| Change within 5% | 165 (35.79) | 45 (37.50) | 45/2,270 | 16.2 | Ref | Ref | ||
| >5% increase | 133 (28.85) | 30 (25.00) | 30/2,146 | 14.0 | 0.86 (0.54–1.37) | 0.54 | 1.00 (0.60–1.69) | 0.98 |
Notes: AFM = appendicular fat mass; BMI = body mass index; ALM = appendicular lean mass; SD = standard deviation WBFM = whole-body fat mass; WBLM = whole-body lean mass; TFM = trunk fat mass; TLM = trunk lean mass
Adjusted for age
Adjusted for age, income, ethnicity, smoking, cardiovascular disease, diabetes, stroke, number of frailty criteria, and BMI in Year 3.
Finally, Figures 1A–1C illustrate HRs for mortality as a function of joint changes in lean and fat mass measures. The results showed that combined categories of change were not associated with mortality, with the notable exception of an increase in both trunk lean and trunk fat mass group that was associated with lower mortality (HR 0.41; 95% CI, 0.21–0.78) compared with a decline in the trunk lean and trunk fat mass reference category. Of note, the correlation between changes in lean mass and changes in fat mass was modest and ranged from −0.16 for appendicular compartment to −0.25 for trunk measures (results not shown).
Discussion
This study demonstrated that having more central or appendicular body fat mass was associated with lower mortality in older frail women. However, low lean body mass was not a significant determinant of mortality. Analyses of 3-year change in body composition has also highlighted that a pronounced decline in fat mass, but not in lean mass, had a significant association with higher mortality. These results further validate our previous findings on lower mortality rates in frail older women with overweight (25.0–29.9 kg/m2) and class I obesity (30.0–34.9 kg/m2) BMI categories as compared with frail older women of normal weight (18.5–24.9 kg/m2) (29).
To the best of our knowledge, this is the first study to evaluate the association of direct measures of body composition and their dynamics over two time points with rates of allcause mortality in a sample of only frail, older adults. Thus, comparisons with other studies are limited to nonfrail or mixed populations. A negative linear association between fat mass and risk of death was demonstrated in French women 75 years and older so that a 10% increase in fat mass was associated with a 12% reduction in mortality risk (22). Among aged women participating in a Chinese project aimed at examining bone health, neither truncal nor total-body fat mass was associated with mortality (20). Similarly, calf fat mass, measured by pQCT, was not a risk factor for mortality in an Italian study of community-dwelling older adults (11). Finally, bodycomposition measurement by bioelectrical impedance yielded a null association between body composition and mortality in a Swiss study of older women (21). Taken together, this and our findings, the results indicate that higher fat mass may have a positive impact on survival in older women, and this association might be pertinent even to a greater extent in those with frailty.
Scarce literature exists on the association between changes over time in body composition and mortality in older women. Consistent with our findings, among men age 65 to 93 who underwent repeated DXA measurements, those who lost trunk fat mass had the highest risk of mortality compared with men who remained stable (18), However, in contrast to our results, which showed no association between fat mass gain and mortality, Lee and colleagues (18) demonstrated that men with total fat mass gain exhibited a trend toward a greater mortality risk than those who remained stable. This U-shaped relationship resonated in another study that highlighted detrimental role of changes in appendicular and central fact mass indexes in predicting mortality in age-adjusted models (12). Thus, it is plausible to conclude that although decline in fat mass might be detrimental regardless of one's functional status, increase in adiposity needs to be viewed in the context of older adult frailty status.
Surprisingly, we found no association between lean mass and mortality in older women with physical frailty. These null findings, however, were consistent with those of previous reports that highlighted the insignificant relationship of muscle area and mass with mortality. Using Health ABC data, Newman and colleagues (19) indicated that strength, but not muscle mass, was associated with mortality in an initially not disabled older population. Similarly, the “inChianti” report showed that walking speed and not skeletal muscle density/area was a significant risk factor for mortality in community-dwelling older adults and in a small subsample of frail individuals (11). Finally, an extensive recent study of 4,574 French women 75 years and older also demonstrated an insignificant association between lean mass/height2 index and mortality in, on average, 17.7 years of follow-up (22). In summary, muscle function might be a more powerful predictor of death than muscle mass in nonfrail individuals. In frail older adults, a combination of low muscle strength and reduced muscle mass is especially pronounced, rendering the frail population even more sensitive to changes in functional performance and less attuned to dynamics in lean body mass. This assertion, however, deserves further confirmation in future studies.
Before considering causality, we must consider several possible methodological explanations. The first is the “survival effect.” Individuals susceptible to the complications of excessive adiposity may have died at a younger age, and those with higher fat mass who survived to old age may have characteristics that protect them from the adverse effects of adiposity (7). Another important methodological explanation is reverse causation. Unrecognized systemic illness or subclinical symptoms might lead to biased estimates of mortality risks in those with low fat mass, thereby making individuals with higher adiposity appear to be protected (7). To address this potential concern, we excluded those with early mortality and the results were consistent with those in primary models, thus, reverse causation is less likely to explain our findings. Finally, mortality rates might be distored because frailty definition includes a body-weight-change component, however our previous sub group analyses demonstrated that excluding those who experienced unintentional weight loss did not substantially altered HR estimates (29).
A causal association between higher fat and lower mortality has been suggested (22, 23, 24). Specifically, higher adiposity may benefit older adults by exerting a favorable effect on cognitive function (23), protecting against osteoporotic fractures (24), and serving as an energy reserve to be mobilized in the event of acute illness (22). The last assertion might be particularly pertinent to frail older adults because of their impaired ability to withstand stress. Finally, low adiposity may represent indicators of reduced metabolic reserves and undernutrition, which also were implicated in the pathophysiology of frailty (1). Thus, pending further research, a cautious interpretation of these results is that salutary role of higher fat mass in the context of overall mortality in older frail female populations might relate to availability of energy reserves.
The strengths of this study include its large sample size, the use of well-validated frailty criteria, and, most important, direct measures of body composition. Furthermore, we distinguished between static and dynamic changes in body composition, with the latter being an important, and often underinvestigated, concomitant of health outcomes in old age (25). We also acknowledge some limitations of our study. As with any observational study, there exists the possibility of residual and unmeasured confounding. Given the preponderance of White participants, our results may not be widely generalizable. Furthermore, although the DXA provides a reliable estimate of trunk fat, it cannot distinguish between subcutaneous and visceral fat, which may exert different effects on survival (26) due to their distinct metabolic implications (27, 28).
In conclusion, we demonstrated that low body fat or a pronounced decline in adipose tissue is associated with increased rates of all-cause mortality in frail older women. These results confirm our previous findings indicating that being overweight and even slightly obese might be beneficial in terms of long-term survival in frail older women. From a clinical point of view, our results might indicate a need to re-evaluate the definition of healthy weight in persons with physical frailty.
Acknowledgments: The authors thank the WHI investigators and staff for their dedication and the study participants for making the program possible. A listing of WHI investigators can be found at: https://cleo.whi.org/researchers/Documents%20%20 Write%20a%20Paper/WHI%20Investigator%20Short%20List.pdf
Funding sources: The WHI program is funded by the National Heart, Lung and Blood Institute, National Institute of Health, U.S. Department of Health, and Human Services through Contracts N01WH22110, 24152, 32100-2, 32105-6, 32108-9, 32111-13, 32115, 32118-32119, 32122, 42107-26, 42129-32, and 44221.
Conflict of Interest Disclosure: OZ, ERS, WL, SG, MD, LS and SZS declared no conflict of interest
Ethical standards: The authors declare that the study complies with the current ethical standards for investigation involving human participants in the US.
References
- 1.Fried LP, Tangen CM, Walston J, et al. Frailty in older adults: evidence for a phenotype. J Gerontol A Biol Sci Med Sci. 2001;56:M146–M156. doi: 10.1093/gerona/56.3.m146. 10.1093/gerona/56.3.M146 PubMed PMID: 11253156. [DOI] [PubMed] [Google Scholar]
- 2.Ensrud KE, Ewing SK, Taylor BC, et al. Frailty and risk of falls, fracture, and mortality in older women: the study of osteoporotic fractures. J Gerontol A Biol Sci Med Sci. 2007;62:744–751. doi: 10.1093/gerona/62.7.744. 10.1093/gerona/62.7.744 PubMed PMID: 17634322. [DOI] [PubMed] [Google Scholar]
- 3.Woods NF, LaCroix AZ, Gray SL, et al. Frailty: Emergence and consequences in women aged 65 and older in the Women's Health Initiative Observational Study. J Am Geriatr Soc. 2005;53:1321–1330. doi: 10.1111/j.1532-5415.2005.53405.x. 10.1111/j.1532-5415.2005.53405.x PubMed PMID: 16078957. [DOI] [PubMed] [Google Scholar]
- 4.Walston J, McBurnie MA, Newman A, et al. Frailty and activation of the inflammation and coagulation systems with and without clinical comorbidities: results from the Cardiovascular Health Study. Arch Intern Med. 2002;162:2333–2341. doi: 10.1001/archinte.162.20.2333. 10.1001/archinte.162.20.2333 PubMed PMID: 12418947. [DOI] [PubMed] [Google Scholar]
- 5.Xue QL. The frailty syndrome: Definition and natural history. Clin Geriatr Med. 2011;27:1–15. doi: 10.1016/j.cger.2010.08.009. 10.1016/j.cger.2010.08.009 PubMed PMID: 21093718; PMCID 3028599. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Baumgartner RN, Heymsfield SB, Roche AF. Human body composition and the epidemiology of chronic disease. Obes Res. 1995;3:73–95. doi: 10.1002/j.1550-8528.1995.tb00124.x. 10.1002/j.1550-8528.1995.tb00124.x PubMed PMID: 7712363. [DOI] [PubMed] [Google Scholar]
- 7.Oreopoulos A, Kalantar-Zadeh K, Sharma AM, Fonarow GC. The obesity paradox in the elderly: potential mechanisms and clinical implications. Clin Geriatr Med. 2009;25:643–659. doi: 10.1016/j.cger.2009.07.005. 10.1016/j.cger.2009.07.005 PubMed PMID: 19944265. [DOI] [PubMed] [Google Scholar]
- 8.Villareal DT, Banks M, Siener C, Sinacore DR, Klein S. Physical frailty and body composition in obese elderly men and women. Obes Res. 2004;12:913–920. doi: 10.1038/oby.2004.111. 10.1038/oby.2004.111 PubMed PMID: 15229329. [DOI] [PubMed] [Google Scholar]
- 9.Chen Z, Wang Z, Lohman T, Heymsfield SB, Outwater E, Nicholas JS. Dual-energy X-ray absorptiometry is a valid tool for assessing skeletal muscle mass in older women. J Nutr. 2007;137:2775–2780. doi: 10.1093/jn/137.12.2775. PubMed PMID: 18029498. [DOI] [PubMed] [Google Scholar]
- 10.Cesari M, Leeuwenburgh C, Lauretani F, Onder G, Bandinelli S, Maraldi C. Frailty syndrome and skeletal muscle: results from the Invecchiare in Chianti study. Am J Clin Nutr. 2006;83:1142–1148. doi: 10.1093/ajcn/83.5.1142. PubMed PMID: 16685058; PMCID 2668161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Cesari M, Pahor M, Lauretani F, Zamboni V, Bandinelli S, Bernabei R. Skeletal muscle and mortality results from the InCHIANTI Study. J Gerontol A Biol Sci Med Sci. 2009;64:377–384. doi: 10.1093/gerona/gln031. 10.1093/gerona/gln031 PubMed PMID: 19181709. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Szulc P, Munoz F, Marchand F, Chapurlat R, Delmas PD. Rapid loss of appendicular skeletal muscle mass is associated with higher all-cause mortality in older men: the prospective MINOS study. Am J Clin Nutr. 2010;91:1227–1236. doi: 10.3945/ajcn.2009.28256. 10.3945/ajcn.2009.28256 PubMed PMID: 20237137. [DOI] [PubMed] [Google Scholar]
- 13.Jung HW, Kim SW, Lim JY, Kim KW, Jang HC, Kim CH. Frailty status can predict further lean body mass decline in older adults. J Am Geriatr Soc. 2014;62:2110–2117. doi: 10.1111/jgs.13107. 10.1111/jgs.13107 PubMed PMID: 25370293. [DOI] [PubMed] [Google Scholar]
- 14.Deurenberg P, van der Kooij K, Evers P, Hulshof T. Assessment of body composition by bioelectrical impedance in a population aged greater than 60 y. Am J Clin Nutr. 1990;51:3–6. doi: 10.1093/ajcn/51.1.3. PubMed PMID: 2296928. [DOI] [PubMed] [Google Scholar]
- 15.Anderson GL, Manson J, Wallace R, Lund B, Hall D, Davis S. Implementation of the Women's Health Initiative study design. Ann Epidemiol. 2003;13:S5–S17. doi: 10.1016/s1047-2797(03)00043-7. 10.1016/S1047-2797(03)00043-7 PubMed PMID: 14575938. [DOI] [PubMed] [Google Scholar]
- 16.Beasley JM, LaCroix AZ, Neuhouser ML, Huang Y, Tinker L, Woods N. Protein intake and incident frailty in the Women's Health Initiative observational study. J Am Geriatr Soc. 2010;58:1063–1071. doi: 10.1111/j.1532-5415.2010.02866.x. 10.1111/j.1532-5415.2010.02866.x PubMed PMID: 20487071; PMCID 2924946. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.LaCroix AZ, Gray SL, Aragaki A, et al. Statin use and incident frailty in women aged 65 years or older: prospective findings from the Women's Health Initiative Observational Study. J Gerontol A Biol Sci Med Sci. 2008;63:369–375. doi: 10.1093/gerona/63.4.369. 10.1093/gerona/63.4.369 PubMed PMID: 18426960. [DOI] [PubMed] [Google Scholar]
- 18.Lee CG, Boyko EJ, Nielson CM, et al. Mortality risk in older men associated with changes in weight, lean mass, and fat mass. J Am Geriatr Soc. 2011;59:233–240. doi: 10.1111/j.1532-5415.2010.03245.x. 10.1111/j.1532-5415.2010.03245.x PubMed PMID: 21288234; PMCID 3403719. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Newman AB, Kupelian V, Visser M, Simonsick EM, Goodpaster BH, Kritchevsky SB. Strength, but not muscle mass, is associated with mortality in the health, aging and body composition study cohort. J Gerontol A Biol Sci Med Sci. 2006;61:72–77. doi: 10.1093/gerona/61.1.72. 10.1093/gerona/61.1.72 PubMed PMID: 16456196. [DOI] [PubMed] [Google Scholar]
- 20.Auyeung TW, Lee JS, Leung J, Kwok T, Leung PC, Woo J. Survival in older men may benefit from being slightly overweight and centrally obese—a 5-year follow-up study in 4,000 older adults using^DXA. J Gerontol A Biol Sci Med Sci. 2010;65:99–104. doi: 10.1093/gerona/glp099. 10.1093/gerona/glp099 PubMed PMID: 19628635. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Graf CE, Karsegard VL, Spoerri A, Makhlouf AM, Ho S, Herrmann FR. Body composition and all-cause mortality in subjects older than 65 y. Am J Clin Nutr. 2015;101:760–767. doi: 10.3945/ajcn.114.102566. 10.3945/ajcn.114.102566 PubMed PMID: 25833973. [DOI] [PubMed] [Google Scholar]
- 22.Rolland Y, Gallini A, Cristini C, Schott AM, Blain H, Beauchet O. Body-composition predictors of mortality in women aged >= 75 y: data from a large population-based cohort study with a 17-y follow-up. Am J Clin Nutr. 2014;100:1352–1360. doi: 10.3945/ajcn.114.086728. 10.3945/ajcn.114.086728 PubMed PMID: 25332333. [DOI] [PubMed] [Google Scholar]
- 23.Han C, Jo SA, Seo JA, Kim BG, Kim NH, Jo I. Adiposity parameters and cognitive function in the elderly: application of «Jolly Fat» hypothesis to cognition. Arch Gerontol Geriatr. 2009;49:133. doi: 10.1016/j.archger.2008.11.005. 10.1016/j.archger.2008.11.005 [DOI] [PubMed] [Google Scholar]
- 24.Johansson H, Kanis JA, Oden A, McCloskey E, Chapurlat RD, Christiansen C. A meta-analysis of the association of fracture risk and body mass index in women. J Bone Miner Res. 2014;29:223–233. doi: 10.1002/jbmr.2017. 10.1002/jbmr.2017 PubMed PMID: 23775829. [DOI] [PubMed] [Google Scholar]
- 25.Arnold AM, Newman AB, Cushman M, Ding J, Kritchevsky S. Body weight dynamics and their association with physical function and mortality in older adults: the Cardiovascular Health Study. J Gerontol A Biol Sci Med Sci. 2010;65:63–70. doi: 10.1093/gerona/glp050. 10.1093/gerona/glp050 PubMed PMID: 19386574. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Kuk JL, Katzmarzyk PT, Nichaman MZ, Church TS, Blair SN, Ross R. Visceral fat is an independent predictor of all-cause mortality in men. Obesity (Silver Spring, Md). 2006;14:336–341. doi: 10.1038/oby.2006.43. 10.1038/oby.2006.43 [DOI] [PubMed] [Google Scholar]
- 27.Goodpaster BH, Krishnaswami S, Harris TB, et al. Obesity, regional body fat distribution, and the metabolic syndrome in older men and women. Arch Intern Med. 2005;165:777–783. doi: 10.1001/archinte.165.7.777. 10.1001/archinte.165.7.777 PubMed PMID: 15824297. [DOI] [PubMed] [Google Scholar]
- 28.Ibrahim MM. Subcutaneous and visceral adipose tissue: structural and functional differences. Obes Rev. 2010;11:11–18. doi: 10.1111/j.1467-789X.2009.00623.x. 10.1111/j.1467-789X.2009.00623.x PubMed PMID: 19656312. [DOI] [PubMed] [Google Scholar]
- 29.Zaslavsky O, Rillamas-Sun E, LaCroix A, et al. The association of anthropometric measures with long-term survival among frail older women: observations from the Women's Health Initiative Observational Study (WHI OS) Journal of American Geriatric Society. 2016;64(2):277–284. doi: 10.1111/jgs.13930. 10.1111/jgs.13930 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Newman AB, Yanez D, Harris T, et al. Weight change in old age and its association with mortality. Journal of the American Geriatrics Society. 2001;49(10):1309–1318. doi: 10.1046/j.1532-5415.2001.49258.x. 10.1046/j.1532-5415.2001.49258.x PubMed PMID: 11890489. [DOI] [PubMed] [Google Scholar]