Abstract
Objective:
We aimed (1) to evaluate the agreement between two methods (equation and bio-impedance analysis [BIA]) to estimate skeletal muscle mass (SMM), and (2) to assess if SMM was associated with all-cause mortality risk in individuals across different geographical sites in Peru.
Methods:
We used data from the CRONICAS Cohort Study (2010–2018), a population-based longitudinal study in Peru to assess cardiopulmonary risk factors from different geographical settings. SMM was computed as a function of weight, height, sex and age (Lee equation) and by BIA. All-cause mortality was retrieved from national vital records. Cox proportional-hazard models were developed and results presented as hazard ratios (HR) with 95% confidence intervals (95% CI).
Results:
At baseline, 3216 subjects, 51.5% women, mean age 55.7 years, were analyzed. The mean SMM was 23.1 kg (standard deviation [SD]: 6.0) by Lee equation, and 22.7 (SD: 5.6) by BIA. Correlation between SMM estimations was strong (Pearson’s ρ coefficient = 0.89, p<0.001); whereas Bland-Altman analysis showed a small mean difference. Mean follow-up was 7.0 (SD: 1.0) years, and there were 172 deaths. In the multivariable model, each additional kg in SMM was associated with a 19% reduction in mortality risk (HR=0.81; 95% CI: 0.75–0.88) using the Lee equation, but such an estimate was not significant when using BIA (HR=0.98; 95% CI: 0.94–1.03). Compared to the lowest tertile, subjects at the highest SMM tertile had a 56% reduction in risk of mortality using the Lee equation, but there was no such association when using BIA estimations.
Conclusion:
There is a strong correlation and agreement between SMM estimates obtained by the Lee equation and BIA. However, an association between SMM and all-cause mortality exists only when the Lee equation is used. Our findings call for appropriate use of approaches to estimate SMM, and there should be a focus on muscle mass in promoting healthier ageing.
Keywords: Skeletal muscle mass, bio-impedance analysis, all-cause mortality, cohort study
Sustainable Development Goal: Good Health and Wellbeing
INTRODUCTION
Skeletal muscle mass (SMM) is part of the lean body mass, and points to physical function, mobility and balance. Thus, the more muscle mass is present, the less prone a subject is to injury, chronic disease, and premature death [1]. After peaking during early adulthood, SMM progressively declines starting at 45 years [2, 3]. Muscle mass usually drops by about 3% to 8% per decade, and this decline is more patent after reaching 60 years of age [4]. Thus, sarcopenia, defined as the loss of SMM to below a critical threshold, may lead to functional impairment and physical disability [5], reduction in the quality of life [6] and high risk of all-cause mortality [7].
Previous studies have shown the association between low skeletal muscle mass and all-cause mortality [8], but also cardiovascular and cancer mortality [9]. Moreover, lower skeletal muscle mass at hospital admission predicts falls and mortality three months after discharge among hospitalized older patients [10]. A recent systematic review including nine studies mostly from high-income countries and one from Brazil found that muscle mass was low among those who died compared to those who survived [7]. However, many of the studies in the review were conducted among elderly individuals (60+ years), and whether such an association is present in younger age groups remains to be investigated. The distribution and amount of muscle mass are vary greatly between populations and ethnicity [11], and reports comparing methods to estimate muscle mass and its potential association with all-cause mortality are scarce.
The literature describes different methods to determine SMM, particularly in sport and clinical settings, using specialized equipment that is unsuitable for large-scale studies (i.e., dual energy X-ray absorptiometry [DXA], computed tomography [CT], and magnetic resonance imaging [MRI]) or expensive (i.e., bioelectrical impedance analysis [BIA]) [12]. To overcome problem, different equations to estimate muscle mass have been developed, although with marked variability between them [13]. The equation by Lee et al. [14] has been considered valid and appropriate for SMM in Latin American adults vs. DXA-predicted measurement [15, 16], thus being suitable for population-based studies.
The CRONICAS Cohort Study, a longitudinal study in Peru to assess cardiopulmonary risk factors among populations from different geographical settings, collected BIA information in subjects aged 35 years and over [17]; as a result, the study is unique to compare methods to determine SMM. Therefore, the present paper aims to (1) evaluate the agreement between two methods to estimate SMM, the validated Lee equation vs. BIA, and (2) to assess the association between muscle mass, estimated using an equation and BIA, and all-cause mortality among adults in different geographical settings in Peru.
MATERIALS AND METHODS
Study design and settings
Details of the CRONICAS Cohort Study are reported elsewhere [17]. Started in 2010, it is an ongoing longitudinal study conducted to assess the cardiovascular and pulmonary profile according to urbanization and altitude. A total of 3601 subjects were enrolled in three different regions in Peru: Lima, a highly urbanized setting at sea level; Tumbes, a semiurban setting at sea level in the north of Peru; and Puno, a region in the highlands with urban and rural settings. For this analysis, information from baseline and last follow-up (2018) was used. The sample size of 3601 has >80% power to detect a reduction in the risk of mortality of at least 15% in the highest level of SMM vs. the lowest, and to control for at least 10 confounders.
Participants
Individuals in the CRONICAS Cohort Study were enrolled using sex- and age-stratified (35–44, 45–54, 55–64, 65+ years) random sampling based on the most recent census in each of the study sites [17]. Habitual residents (≥6 months) in the study settings were considered eligible. Only one participant per household was enrolled. Pregnant women, those unable to provide informed consent, those bedridden or with a physical disability preventing anthropometric measurements, and people with active tuberculosis were excluded.
Definition of variables
All-cause mortality, defined as the presence of a fatal event occurring between enrolment at baseline and any point during follow-up, was the outcome of interest. The vital status of participants (dead or alive) was retrieved from national records in 2018. Time elapsed between baseline and death or censorship was estimated in years. For this analysis, only vital status and the date of death was used; however, if the participant was alive, then the date of the search in the National Registry of Identification and Civil Status (RENIEC in Spanish) was considered as the censoring date. The RENIEC uses the National Identification Number (DNI), and the risk of misclassification of the vital status is very low [18].
Skeletal muscle mass, evaluated at baseline, was used as exposure of interest and estimated using two different approaches. In the first approach, the formula validated by Lee et al [14] was used:
where weight was in kg, height in meters, and age in years. For sex, a value of 0 is assigned for females, and 1 for males.
Skeletal muscle mass was also estimated by BIA using a TANITA TBF-300A device, using the formula described by Janssen for Caucasian populations [12]:
where height is in cm, R (electrical resistance) in Ohm, and age in years. For sex, a value of 0 is assigned to females, and 1 for males.
The TANITA TBF-300A (TANITA Corporation, Tokyo, Japan) is a professional body fat analyzer including an athlete and wrestler mode, and providing critical body composition analysis.
For analysis purposes, SMM as numerical variables (in kg) and also split into tertiles was used to assess potential dose-response of the association between muscle mass and all-cause mortality.
Sociodemographic, behavioral and cardiovascular risk factors were also considered in the analyses as potential confounders. Sociodemographic variables included were sex (male or female), age (30–39, 40–49, 50–59, 60+ years), education level (<7, 7–11, and 12+ years), socioeconomic level (based on household assets and possessions, and then split into tertiles), and study site (Lima, urban Puno, rural Puno, and Tumbes). Among behavioral variables, smoking (based on self-report history and split into never, former and current smoker), alcohol abuse (based on the Alcohol Use Disorder Identification Test and focusing on hazardous drinking [no vs. yes]), and physical activity (low vs. moderate/high) were included. Physical activity was measured using the leisure-time and transport-related domains of the International Physical Activity Questionnaire (IPAQ) using standard cutoffs as in a previous report [19]. Finally, among cardiovascular risk factors, body mass index, using traditional cutoffs to define normal (<25 kg/m2), overweight (≥25 and <30 kg/m2), and obese (≥30 kg/m2); hypertension (as systolic blood pressure ≥140 or diastolic blood pressure ≥90, or previous diagnosis by a physician, or in current anti-hypertensive medication) [20]; and type 2 diabetes (defined as fasting glucose ≥126 mg/dL, or previous diagnosis by a physician, or in current anti-diabetic therapy) [21], were also analyzed.
Statistical analysis
Initially, a description of participants’ characteristics was tabulated by study site. In addition, the mean and standard deviation (SD) of baseline SMM (in kg) estimated by each method was tabulated by each of the potential confounders, and comparisons were done using Student t test or analysis of variance, accordingly. The study population was described by tabulating demographic, behavioral and cardiovascular risk variables by SMM (in tertiles) by Lee equation and BIA.
Correlation between SMM estimates, obtained by equation and BIA, was calculated using Pearson’s coefficient (ρ). Bland-Altman plot, i.e., a plot of the difference of paired variables vs. their average, was also used to assess the agreement between the two SMM estimation approaches.
Mortality rates for the overall sample and by SMM (in tertiles) were estimated using person-time and incidence rates, and reported by 1,000 person-years of follow-up. Crude and adjusted Cox proportional hazard regression models were created to assess the association between the variables of interest. The risk of mortality was calculated by any additional increase (in kg) of SMM (by formula and BIA). SMM was also included in tertiles in different models using the lowest tertile as the reference group. For all models, hazard ratios (HR) and 95% confidence (95% CI) were reported. Proportional hazard assumptions were evaluated using Schoenfeld residuals in a post-estimation approach. Study site was the only variable violating proportional hazard assumption and, consequently, a stratified Cox procedure was used to control for this variable. The variance inflation factor (VIF) was also estimated to assess collinearity given the number of variables included in the adjusted models.
All analyses were conducted using STATA 16 for Windows (StataCorp, College Station, TX, US) and a p<0.05 was considered as statistically significant.
Ethics
The CRONICAS Cohort Study was approved by the IRB at Universidad Peruana Cayetano Heredia (UPCH) and Asociación Benéfica PRISMA in Peru, and the Bloomberg School of Public Health at Johns Hopkins University in the United States. Follow-up assessments to assess participants’ vital status in both studies were approved by the IRB at UPCH. All research approvals conformed with the principles embodied in the Declaration of Helsinki.
RESULTS
Data from 3216 participants (89.3% from the total enrolled), mean age 55.7 years (SD: 12.7), 1657 (51.5%) females, and 686 (21.3%) with 12+ years of education, were analyzed. The characteristics of the population according to study site are available in the Online Supplement E-Table 1.
At baseline, mean SMM by equation was 23.1 (SD: 6.0) kg, whereas by BIA it was 22.7 (SD: 5.6) kg. Overall, SMM for any approach was greater among males, younger individuals, those with higher education, those from a higher socioeconomic level, and among those with high physical activity level and body mass index (Table 1).
Table 1:
Characteristics of the study population at baseline and muscle mass
| SMM by formula (in kg) | SMM by BIA (in kg) | |||
|---|---|---|---|---|
| Mean (SD) | p-value | Mean (SD) | p-value | |
| Sex | < 0.001 | < 0.001 | ||
| Male | 27.9 (3.8) | 27.4 (3.7) | ||
| Female | 18.7 (3.8) | 18.4 (3.2) | ||
| Age, in categories | < 0.001 | < 0.001 | ||
| 30 – 39 years | 25.6 (5.7) | 23.6 (5.4) | ||
| 40 – 49 years | 24.9 (5.6) | 23.5 (5.3) | ||
| 50 – 59 years | 23.6 (5.6) | 23.1 (5.7) | ||
| 60+ years | 20.7 (5.7) | 21.7 (5.7) | ||
| Education level | < 0.001 | < 0.001 | ||
| < 7 years | 20.9 (5.6) | 21.4 (5.6) | ||
| 7 – 11 years | 24.7 (5.5) | 23.8 (5.4) | ||
| 12+ years | 25.6 (5.6) | 24.0 (5.5) | ||
| Socioeconomic level | < 0.001 | < 0.001 | ||
| Low | 21.1 (5.7) | 21.4 (5.3) | ||
| Middle | 23.6 (5.8) | 23.1 (5.8) | ||
| High | 24.7 (5.9) | 23.7 (5.6) | ||
| Study setting | < 0.001 | < 0.001 | ||
| Lima | 23.1 (5.8) | 22.6 (5.5) | ||
| Urban Puno | 23.4 (6.1) | 22.1 (5.3) | ||
| Rural Puno | 21.1 (5.6) | 21.5 (5.2) | ||
| Tumbes | 24.1 (6.0) | 23.9 (6.0) | ||
| Smoking | < 0.001 | < 0.001 | ||
| Never | 21.1 (5.6) | 20.9 (5.2) | ||
| Former | 25.2 (5.4) | 24.8 (5.3) | ||
| Current | 27.4 (5.2) | 26.4 (5.0) | ||
| Alcohol abuse | < 0.001 | < 0.001 | ||
| No | 22.4 (5.8) | 22.1 (5.6) | ||
| Hazardous drinking | 27.9 (4.5) | 26.6 (4.3) | ||
| Physical activity levels | < 0.001 | < 0.001 | ||
| Low | 21.4 (5.6) | 21.3 (5.5) | ||
| Moderate/high | 24.0 (6.0) | 23.4 (5.6) | ||
| Body mass index | < 0.001 | < 0.001 | ||
| < 25 kg/m2 | 20.4 (5.6) | 21.4 (5.3) | ||
| 25 – 29.9 kg/m2 | 23.7 (5.6) | 23.1 (5.5) | ||
| 30+ kg/m2 | 25.2 (5.8) | 23.7 (6.0) | ||
| Hypertension | 0.32 | 0.91 | ||
| No | 23.2 (5.9) | 22.8 (5.6) | ||
| Yes | 23.0 (6.1) | 22.7 (5.7) | ||
| Type 2 diabetes | 0.49 | 0.32 | ||
| No | 23.2 (6.0) | 22.8 (5.6) | ||
| Yes | 23.4 (6.0) | 23.1 (6.1) | ||
BIA = Bio-impedance analysis; SD = Standard deviation; SMM = Skeletal muscle mass
Comparisons were estimated using Student t test or Analysis of variance, accordingly.
Correlation and agreement between SMM methods
Correlation between SMM by equation and BIA was highly significant (Pearson’s ρ coefficient = 0.89, p<0.001), whereas the Bland-Altman plot showed a small mean difference (0.39; 95% CI: 0.29 – 0.49) between measurements (Online Supplement E-Figure 1).
Males (p<0.001), young individuals (p<0.001), those with higher education level (p<0.001), and from a higher socioeconomic level (p<0.001) were in the highest tertile of SMM by both formula (Table 2) and BIA (Table 3).
Table 2:
Characteristics of the study population at baseline according to SMM by equation (in tertiles)
| Skeletal muscle mass (in tertiles) | ||||
|---|---|---|---|---|
| Low | Middle | High | p-value | |
| (n = 1073) | (n = 1071) | (n = 1072) | ||
| Sex | < 0.001 | |||
| Female | 1054 (98.2%) | 558 (52.1%) | 45 (4.2%) | |
| Male | 19 (1.8%) | 513 (47.9%) | 1027 (95.8%) | |
| Age, in categories | < 0.001 | |||
| 30 – 39 years | 75 (7.0%) | 126 (11.8%) | 182 (17.0%) | |
| 40 – 49 years | 191 (17.8%) | 268 (25.0%) | 355 (33.1%) | |
| 50 – 59 years | 279 (26.0%) | 267 (24.9%) | 323 (30.1%) | |
| 60+ years | 528 (49.2%) | 410 (38.3%) | 212 (19.8%) | |
| Education level | <0.001 | |||
| < 7 years | 695 (64.8%) | 505 (47.2%) | 283 (26.5%) | |
| 7 – 11 years | 233 (21.7%) | 375 (35.0%) | 437 (40.8%) | |
| 12+ years | 145 (13.5%) | 191 (17.8%) | 350 (32.7%) | |
| Socioeconomic level | <0.001 | |||
| Low | 461 (43.0%) | 370 (34.6%) | 202 (18.8%) | |
| Middle | 333 (31.0%) | 358 (33.4%) | 391 (36.5%) | |
| High | 279 (26.0%) | 343 (32.0%) | 479 (44.7%) | |
| Study setting | <0.001 | |||
| Lima | 349 (32.5%) | 350 (32.7%) | 347 (32.4%) | |
| Urban Puno | 185 (17.2%) | 173 (16.2%) | 204 (19.0%) | |
| Rural Puno | 253 (23.6%) | 198 (18.5%) | 126 (11.8%) | |
| Tumbes | 286 (26.7%) | 350 (32.6%) | 395 (36.9%) | |
| Smoking | <0.001 | |||
| Never | 837 (78.1%) | 627 (58.6%) | 358 (33.4%) | |
| Former | 191 (17.8%) | 359 (33.5%) | 475 (44.3%) | |
| Current | 44 (4.1%) | 85 (7.9%) | 239 (22.3%) | |
| Alcohol abuse | <0.001 | |||
| No | 1052 (98.0%) | 954 (89.1%) | 768 (71.6%) | |
| Hazardous drinking | 21 (2.0%) | 117 (10.9%) | 304 (28.4%) | |
| Physical activity levels | <0.001 | |||
| Low | 455 (42.5%) | 361 (33.7%) | 213 (19.9%) | |
| Moderate/high | 616 (57.5%) | 709 (66.3%) | 858 (80.1%) | |
| Obesity | <0.001 | |||
| < 25 kg/m2 | 430 (40.1%) | 354 (33.0%) | 164 (15.3%) | |
| 25 – 29.9 kg/m2 | 475 (44.3%) | 351 (32.8%) | 576 (53.7%) | |
| 30+ kg/m2 | 168 (15.6%) | 366 (34.2%) | 332 (31.0%) | |
| Hypertension | 0.08 | |||
| No | 786 (73.3%) | 776 (72.5%) | 820 (76.5%) | |
| Yes | 286 (26.7%) | 295 (27.5%) | 252 (23.5%) | |
| Type 2 diabetes | 0.91 | |||
| No | 940 (91.7%) | 951 (91.4%) | 961 (91.9%) | |
| Yes | 85 (8.3%) | 90 (8.6%) | 85 (8.1%) | |
Table 3:
Characteristics of the study population at baseline according to SMM by BIA (in tertiles)
| Skeletal muscle mass by BIA (in tertiles) | ||||
|---|---|---|---|---|
| Low | Middle | High | p-value | |
| (n = 1072) | (n = 1072) | (n = 1069) | ||
| Sex | <0.001 | |||
| Female | 1064 (99.3%) | 550 (51.3%) | 43 (4.0%) | |
| Male | 8 (0.7%) | 522 (48.7%) | 1029 (96.0%) | |
| Age, in categories | <0.001 | |||
| 30 – 39 years | 102 (9.5%) | 130 (12.1%) | 151 (14.1%) | |
| 40 – 49 years | 223 (20.8%) | 294 (27.4%) | 297 (27.7%) | |
| 50 – 59 years | 282 (26.3%) | 279 (26.0%) | 308 (28.7%) | |
| 60+ years | 465 (43.4%) | 369 (34.5%) | 316 (29.5%) | |
| Education level | <0.001 | |||
| < 7 years | 637 (59.4%) | 485 (45.2%) | 361 (33.7%) | |
| 7 – 11 years | 263 (24.5%) | 378 (35.3%) | 404 (37.8%) | |
| 12+ years | 172 (16.0%) | 209 (19.5%) | 305 (28.5%) | |
| Socioeconomic level | <0.001 | |||
| Low | 430 (40.1%) | 366 (34.1%) | 237 (22.1%) | |
| Middle | 337 (31.4%) | 358 (33.4%) | 387 (36.1%) | |
| High | 305 (28.5%) | 348 (32.5%) | 448 (41.8%) | |
| Study setting | <0.001 | |||
| Lima | 354 (33.0%) | 350 (32.7%) | 342 (31.9%) | |
| Urban Puno | 204 (19.0%) | 184 (17.1%) | 174 (16.2%) | |
| Rural Puno | 233 (21.8%) | 206 (19.2%) | 138 (12.9%) | |
| Tumbes | 281 (26.2%) | 332 (31.0%) | 418 (39.0%) | |
| Daily smoking | <0.001 | |||
| Never | 846 (79.0%) | 608 (56.7%) | 368 (34.3%) | |
| Former | 186 (17.4%) | 351 (32.7%) | 483 (45.5%) | |
| Current | 39 (3.6%) | 113 (10.5%) | 216 (20.2%) | |
| Alcohol abuse | <0.001 | |||
| No | 1047 (97.7%) | 938 (87.5%) | 789 (73.6%) | |
| Hazardous drinking | 25 (2.3%) | 134 (12.5%) | 283 (26.4%) | |
| Physical activity levels | <0.001 | |||
| Low | 450 (42.1%) | 348 (32.5%) | 231 (21.6%) | |
| Moderate/high | 620 (57.9%) | 722 (67.5%) | 841 (78.4%) | |
| Obesity | <0.001 | |||
| < 25 kg/m2 | 386 (36.0%) | 330 (30.8%) | 232 (21.6%) | |
| 25 – 29.9 kg/m2 | 462 (43.1%) | 404 (37.7%) | 536 (50.0%) | |
| 30+ kg/m2 | 224 (20.9%) | 338 (31.5%) | 304 (28.4%) | |
| Hypertension | 0.76 | |||
| No | 793 (74.0%) | 787 (73.4%) | 802 (74.8%) | |
| Yes | 278 (26.0%) | 285 (26.6%) | 270 (25.2%) | |
| Type 2 diabetes | 0.31 | |||
| No | 953 (92.4%) | 938 (90.6%) | 962 (91.9%) | |
| Yes | 78 (7.6%) | 97 (9.4%) | 85 (8.1%) | |
Muscle mass and all-cause mortality
A total of 172 (5.8%) fatal events occurred during 7.0 (SD: 1.0) years of follow-up, accruing 20,894.5 person-years and an overall all-cause mortality of 8.2 (95% CI: 7.1 – 9.6) per 1,000 person-years of follow-up. The mortality rate was higher among those in the lowest tertile of SMM by any definition (Table 4).
Table 4:
Skeletal muscle mass and all-cause mortality: Crude and adjusted Cox models
| Mortality rate | All-cause mortality | ||
|---|---|---|---|
| (per 1,000 person-year) | Crude model | Adjusted model* | |
| Estimate (95%IC) | HR (95% CI) | HR (95% CI) | |
| SMM by Lee equation | |||
| Per each kg increase | -- | 0.91 (0.89 – 0.94) | 0.81 (0.75 – 0.88) |
| In tertiles | |||
| Low | 11.4 (9.1 – 14.2) | 1 (Reference) | 1 (Reference) |
| Middle | 9.6 (7.6 – 12.2) | 0.85 (0.61 – 1.18) | 0.63 (0.36 – 1.09) |
| High | 4.0 (2.8 – 5.8) | 0.35 (0.23 – 0.54) | 0.44 (0.20 – 0.96) |
| SMM by BIA | |||
| Per each kg increase | -- | 0.97 (0.94 – 0.99) | 0.98 (0.94 – 1.03) |
| In tertiles | |||
| Low | 9.3 (7.2 – 11.9) | 1 (Reference) | 1 (Reference) |
| Middle | 8.9 (7.0 – 11.5) | 0.96 (0.68 – 1.37) | 1.04 (0.63 – 1.73) |
| High | 6.6 (4.9 – 8.7) | 0.70 (0.48 – 1.03) | 1.01 (0.53 – 1.94) |
BIA = Bioimpedance analysis; Kg = kilogram; SMM = Skeletal muscle mass.
Model adjusted by sex, age, education level, socioeconomic level, daily smoking, alcohol use, physical activity levels, and body mass index. Study site was included as a stratified variable in the Cox regression model.
In a multivariable model and using the Lee equation, each additional kg in SMM was associated with a 19% reduction of mortality risk (HR = 0.81; 95% CI: 0.75 – 0.88). Moreover, those in the highest tertile of SMM had a 56% lower mortality risk (HR = 0.44; 95% CI: 0.20 – 0.96). However, there was no such association with reduced mortality risk (HR = 0.98; 95% CI: 0.94 – 1.03) in the multivariable model when using the BIA estimation. Similarly, there was no association when SMM by BIA was evaluated in tertiles (Table 4).
DISCUSSION
Main findings
Using data of a Peruvian cohort of adults aged ≥35 years from different geographical settings, we found a strong correlation, assessed by Pearson coefficient, and good agreement, evaluated by Bland-Altman analysis, between estimates of SMM obtained by the Lee equation and bioimpedance analysis. Nevertheless, an association between SMM and all-cause mortality was found only when the Lee equation was used, but not BIA.
Comparison with previous studies
Previous reports have shown a good performance of the Lee equation to estimate SMM. Thus, a previous work demonstrated that such equation had good prediction qualities, with an R2 of 0.86 compared to magnetic resonance imaging [14]. Similarly, another study reported R2 between 0.80 and 0.83 for males and females compared to DXA [16]. Thus, our results are in line with those studies but using BIA.
Regarding all-cause mortality, a recent meta-analysis found a significant pooled difference in appendicular skeletal muscle mass (i.e., the sum of the muscle mass of the 4 limbs) between older adults who died and who survived across different cohort studies of older subjects [7]. This review also reported that not only skeletal muscle quality matters, but also quantity. Based on our results, a standardization process between different approaches to estimate SMM is needed. Although BIA devices are used to estimate SMM in different settings, usually equations are based on Caucasian populations which cannot be directly applied to other ethnic groups.
Four longitudinal studies conducted in Asia (two in Japan, and one each in South Korea and Hong Kong) reported that the difference in muscle mass between elderly who died and those who survived was smaller [22–25], suggesting that muscle mass has less impact on mortality in this population, and therefore highlighting the need to better explore the impact of muscle mass in mortality risk. Our study expands on this by showing that different approaches to estimate SMM may have different associations with all-cause mortality.
Chuang et al. reported a threshold relationship between lean body mass and mortality, and as a result, the lowest quartile had the highest mortality risk [8]. This study used a skeletal muscle mass index by a formula built on Caucasian populations using bioimpedance analysis [26]. The MINOS (Montceau les MINes OSteoporosis) Cohort Study also demonstrated that appendicular skeletal muscle mass loss was associated with higher mortality among elderly subjects followed for 10 years [27]. However, this study also showed that a single measure of muscle (or lean) mass may not necessarily reflect health status.
Finally, different outcomes have been associated with low skeletal muscle mass. For example, using a machine learning approach, 5-year mortality can be predicted with good accuracy among elderly men by sarcopenia characteristics (i.e., based on lean mass estimates) [28]. Similarly, skeletal muscle mass presented an inverse association with cardiovascular disease incidence in a large community-based cohort study [9]: the highest SMM tertile had the lowest 10-year CVD risk among subjects aged 45+ years. Thus, SMM appears to be related to different health outcomes, and a further increase in mortality risk, albeit with little evidence from low and middle-income countries.
Public health relevance
A low level of skeletal muscle mass has been related to unfavorable physical function and an unhealthy inflammatory status [8]. In addition, reduction of muscle mass has been associated to cardiovascular disease, tuberculosis, type 2 diabetes, falls and mortality [10], highlighting that loss of muscle mass can be a relevant public health problem as populations age globally. In view of this, our findings confirm the need to standardize, validate and recalibrate existing approaches to obtain adequate SMM estimates, and the necessity of strategies to preserve skeletal muscle mass, as well as reduce or reverse their loss.
The literature describes that resistance exercise may be particularly beneficial as this can attenuate the decrease of muscle mass associated with aging [5]. Interventions should be initiated at the start of the fifth decade of life and highlight the need of increasing public awareness of the consequences of loss muscle mass. But the onset of loss of skeletal muscle mass can begin at younger ages if individuals develop life-threatening chronic conditions (i.e., cancer, heart disease, etc.) whereby reductions in SMM would be expected as part of the disease process. As a result, skeletal muscle is a strong predictor of mortality, as observed in the large reported effect sizes.
Strengths and limitations
This study benefits from a population-based longitudinal study spanning multiple and heterogeneous study sites. We used a well-known and validated equation to estimate the association of interest as well as a BIA assessment. Nevertheless, our study has some limitations. First, it is an observational study and causal conclusions cannot be inferred. Second, an important proportion of subjects had no follow-up information; however, the attrition rate is comparable to other cohort studies in different settings. Third, only baseline estimations, and not changes over time of our SMM exposure, were used for the analysis. Fourth, although validity of death assessment was not confirmed, the the risk of misclassification should be low because RENIEC uses the national identification number to track vital status. Finally, physical activity levels were evaluated using only two domains of the International Physical Activity Questionnaire (IPAQ). Physical activity is considered a relevant variable to adjust our models, and some residual confounding could be present.
Conclusions
There is a strong correlation and good agreement between estimates of SMM obtained by the Lee equation and by bioimpedance analysis. But an association between SMM and all-cause mortality was found only when the Lee equation was used, not BIA. Our findings call for appropriate use of approaches to estimate SMM. There should be a focus on muscle mass in promoting healthier ageing.
Supplementary Material
Funding:
The CRONICAS Cohort Study was supported by the National Heart, Lung, and Blood Institute Global Health Initiative under the contract Global Health Activities in Developing Countries to Combat Non-Communicable Chronic Diseases (Project Number 268200900033C-1-0-1). RMC-L is supported by a Wellcome Trust International Training Fellowship (214185/Z/18/Z).
Data availability:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to the use of different information pooled for these analyses.
REFERENCES
- 1.Frontera WR, Ochala J. Skeletal muscle: a brief review of structure and function. Calcif Tissue Int 2015;96(3):183–95. [DOI] [PubMed] [Google Scholar]
- 2.Janssen I, Heymsfield SB, Wang ZM, Ross R. Skeletal muscle mass and distribution in 468 men and women aged 18–88 yr. J Appl Physiol (1985) 2000;89(1):81–8. [DOI] [PubMed] [Google Scholar]
- 3.Melton LJ 3rd, Khosla S, Riggs BL. Epidemiology of sarcopenia. Mayo Clin Proc 2000;75 Suppl:S10–2; discussion S2–3. [PubMed] [Google Scholar]
- 4.Holloszy JO. The biology of aging. Mayo Clin Proc 2000;75 Suppl:S3–8; discussion S-9. [PubMed] [Google Scholar]
- 5.Janssen I, Heymsfield SB, Ross R. Low relative skeletal muscle mass (sarcopenia) in older persons is associated with functional impairment and physical disability. J Am Geriatr Soc 2002;50(5):889–96. [DOI] [PubMed] [Google Scholar]
- 6.Woo T, Yu S, Visvanathan R. Systematic Literature Review on the Relationship Between Biomarkers of Sarcopenia and Quality of Life in Older People. J Frailty Aging 2016;5(2):88–99. [DOI] [PubMed] [Google Scholar]
- 7.de Santana FM, Premaor MO, Tanigava NY, Pereira RMR. Low muscle mass in older adults and mortality: A systematic review and meta-analysis. Exp Gerontol 2021;152:111461. [DOI] [PubMed] [Google Scholar]
- 8.Chuang SY, Chang HY, Lee MS, Chia-Yu Chen R, Pan WH. Skeletal muscle mass and risk of death in an elderly population. Nutr Metab Cardiovasc Dis 2014;24(7):784–91. [DOI] [PubMed] [Google Scholar]
- 9.Tyrovolas S, Panagiotakos D, Georgousopoulou E, Chrysohoou C, Tousoulis D, Haro JM, et al. Skeletal muscle mass in relation to 10 year cardiovascular disease incidence among middle aged and older adults: the ATTICA study. J Epidemiol Community Health 2020;74(1):26–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Reijnierse EM, Verlaan S, Pham VK, Lim WK, Meskers CGM, Maier AB. Lower Skeletal Muscle Mass at Admission Independently Predicts Falls and Mortality 3 Months Post-discharge in Hospitalized Older Patients. J Gerontol A Biol Sci Med Sci 2019;74(10):1650–6. [DOI] [PubMed] [Google Scholar]
- 11.Woo J, Arai H, Ng TP, Sayer AA, Wong M, Syddall H, et al. Ethnic and geographic variations in muscle mass, muscle strength and physical performance measures. European Geriatric Medicine 2014;5(3):155–64. [Google Scholar]
- 12.Janssen I, Heymsfield SB, Baumgartner RN, Ross R. Estimation of skeletal muscle mass by bioelectrical impedance analysis. J Appl Physiol (1985) 2000;89(2):465–71. [DOI] [PubMed] [Google Scholar]
- 13.Fernandez Vieitez JA, Aguilera RR. Muscle mass estimation by different anthropometrics equations in high performance weightlifters. Archivos de Medicina del Deporte 2001;XVIII(86):585–91. [Google Scholar]
- 14.Lee RC, Wang Z, Heo M, Ross R, Janssen I, Heymsfield SB. Total-body skeletal muscle mass: development and cross-validation of anthropometric prediction models. Am J Clin Nutr 2000;72(3):796–803. [DOI] [PubMed] [Google Scholar]
- 15.Daskalopoulou C, Wu YT, Pan W, Giné Vázquez I, Prince M, Prina M, et al. Factors related with sarcopenia and sarcopenic obesity among low- and middle-income settings: the 10/66 DRG study. Sci Rep 2020;10(1):20453. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Rech CR, Dellagrana RA, Nunes Marucci MF, Petroski EL. Validity of anthropometric equations for the estimation of muscle mass in the elderly. Rev Bras Cineantropom Desempenho Hum 2012;14(1):23–31. [Google Scholar]
- 17.Miranda JJ, Bernabe-Ortiz A, Smeeth L, Gilman RH, Checkley W. Addressing geographical variation in the progression of non-communicable diseases in Peru: the CRONICAS cohort study protocol. BMJ Open 2012;2(1):e000610. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Vargas-Herrera J, Ruiz KP, Nuñez GG, Ohno JM, Pérez-Lu JE, Huarcaya WV, et al. [Preliminary results of the strengthening of the national death registry information system]. Rev Peru Med Exp Salud Publica 2018;35(3):505–14. [DOI] [PubMed] [Google Scholar]
- 19.Miranda JJ, Carrillo-Larco RM, Gilman RH, Avilez JL, Smeeth L, Checkley W, et al. 0PLoS One 2016;11(11):e0166344. [Google Scholar]
- 20.Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL Jr., et al. The seventh report of the joint national committee on prevention, detection, evaluation, and treatment of high blood pressure: The JNC 7 report. Jama 2003;289(19):2560–72. [DOI] [PubMed] [Google Scholar]
- 21.American Diabetes Association Professional Practice Committee. 2. Classification and diagnosis of diabetes: Standards of medical care in diabetes-2022. Diabetes Care 2022;45(Supplement_1):S17–s38. [DOI] [PubMed] [Google Scholar]
- 22.Moon JH, Kim KM, Kim JH, Moon JH, Choi SH, Lim S, et al. Predictive values of the new sarcopenia index by the foundation for the national institutes of health sarcopenia project for mortality among older Korean adults. PLoS One 2016;11(11):e0166344. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Nakamura K, Yoshida D, Honda T, Hata J, Shibata M, Hirakawa Y, et al. Prevalence and Mortality of Sarcopenia in a Community-dwelling Older Japanese Population: The Hisayama Study. J Epidemiol 2021;31(5):320–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Seino S, Kitamura A, Abe T, Taniguchi Y, Yokoyama Y, Amano H, et al. Dose-Response Relationships Between Body Composition Indices and All-Cause Mortality in Older Japanese Adults. J Am Med Dir Assoc 2020;21(6):726–33.e4. [DOI] [PubMed] [Google Scholar]
- 25.Woo J, Leung J. Sarcopenic Obesity Revisited: Insights From the Mr and Ms Os Cohort. J Am Med Dir Assoc 2018;19(8):679–84.e2. [DOI] [PubMed] [Google Scholar]
- 26.Segal KR, Van Loan M, Fitzgerald PI, Hodgdon JA, Van Itallie TB. Lean body mass estimation by bioelectrical impedance analysis: a four-site cross-validation study. Am J Clin Nutr 1988;47(1):7–14. [DOI] [PubMed] [Google Scholar]
- 27.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(5):1227–36. [DOI] [PubMed] [Google Scholar]
- 28.Kruse C, Goemaere S, De Buyser S, Lapauw B, Eiken P, Vestergaard P. Predicting mortality and incident immobility in older Belgian men by characteristics related to sarcopenia and frailty. Osteoporos Int 2018;29(6):1437–45. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to the use of different information pooled for these analyses.