Abstract
Background
Sarcopenic obesity (SO) is associated with poorer physical performance in the elderly and will increase in relevance with population ageing and the obesity epidemic. The lack of a consensus definition for SO has resulted in variability in its reported prevalence, poor inter-definitional agreement, and disagreement on its impact on physical performance, impeding further development in the field. While sarcopenia definitions have been compared, the impact of obesity definitions in SO has been less well-studied.
Objectives
To compare 3 widely-adopted definitions of obesity in terms of SO prevalence, inter-definitional agreement, and association with muscle function.
Design
Cross-sectional.
Setting
GERILABS study, Singapore
Participants
200 community-dwelling, functionally-independent older adults.
Measurements
We utilized three commonly-used definitions of obesity: body mass index (BMI), waist circumference (WC) and DXA-derived fat mass percentage (FM%). Sarcopenia was defined using Asian Working Group for Sarcopenia criteria. For muscle function, we assessed handgrip strength, gait speed and Short Physical Performance Battery (SPPB). Subjects were classified into 4 body composition phenotypes (normal, obese, sarcopenic and SO), and outcomes were compared between groups.
Results
The prevalence rate for SO was lowest for BMI (0.5%) compared to FM% (10.0%) and WC (10.5%). Inter-definitional agreement was lowest between BMI and WC (κ=0.364), and at best moderate between FM% and WC (κ=0.583). SO performed the worst amongst body composition phenotypes in handgrip strength, gait speed and SPPB (all p<0.01) only when defined using WC. In regression analyses, SO was associated with decreased SPPB scores (β=−0.261, p=0.001) only for the WC definition.
Conclusion
There is large variation in the prevalence of SO across different obesity definitions, with low-to-moderate agreement between them. Our results corroborate recent evidence that WC, and thus central obesity, is best associated with poorer muscle function in SO. Thus, WC should be further explored in defining obesity for accurate and early characterization of SO among older adults in Asian populations.
Key words: Sarcopenic obesity, definition, agreement, waist circumference, muscle function
Introduction
Body composition changes that occur with aging can lead to sarcopenic obesity (SO), an emerging worldwide phenomenon that has been described as the confluence of two public health crises, namely the obesity epidemic and population aging (1, 2). People aged 60 and above comprise 13% of the global population, and are projected to reach 2.1 billion in 2050 (3). Commensurate with this trend of population aging, the prevalence of sarcopenia is expected to rise in tandem, along with the attendant consequences of reduced muscle mass and strength or physical function such as falls, physical disability, reduced quality of life and mortality (4). Concurrently, obesity prevalence has skyrocketed worldwide, doubling in prevalence in middle-aged and older adults since 1980 (5).
Unsurprisingly, there is increasing attention on SO, a high-risk geriatric syndrome that is predominantly observed in an ageing population and has reported prevalence rates ranging from 0.9% to 30.1% (6, 7). The development of SO is attributed to hormonal, inflammatory and myocellular mechanisms from the cross-talk between adipose and muscle tissue which promote fat deposition and loss of lean mass and strength (8). The resultant synergistic complications from both sarcopenia and obesity lead to worse outcomes in SO than either condition alone, resulting in negative health impacts such as loss of independence, disability, reduced quality of life and increased mortality (9, 10). Identification of SO is the cornerstone towards determining its prevalence, understanding its pathophysiology and clinical implications, and devising preventative and therapeutic strategies.
However, the lack of a consistent definition has been a major barrier that has resulted in marked variability in the reported prevalence of SO and conflicting data on its negative health consequences (11). For instance, the prevalence of SO in older adults was reported to differ by up to 26 times when various definitions are applied to the same study population (7). Harmonizing a definition is necessary for continued advancement in the field. While published consensus definitions for sarcopenia have emerged in recent years (4, 12, 13), there is a comparative lack of consensus regarding obesity definitions in the context of SO, leading to great variation in the methods and cut-offs being applied to define obesity (14). This is integral as different obesity definitions measure different constructs and are not interchangeable. Specifically, body mass index (BMI) provides a good indication of the disease risk of comorbidities but is unable to distinguish between lean and fat mass; waist circumference (WC) measures central obesity and is a surrogate measure of visceral adiposity; whilst fat mass percentage (FM%) measures total body fat but not its distribution (15).
This provided impetus for the current study to compare between 3 widely used definitions of obesity (BMI, WC and FM%) in SO. Specifically, we aim to determine the relative impact of the three obesity definitions in terms of SO prevalence, inter-definitional agreement, and association with muscle function.
Methods
Study population and groups
We studied 200 cognitively-intact and functionally-independent community dwelling subjects aged 50 years and above who participated in the “Longitudinal Assessment of Biomarkers for characterization of early Sarcopenia and predicting frailty and functional decline in community-dwelling Asian older adults Study” (GERI-LABS) (16). Inclusion criteria included being: 1) 50–99-years-old at study enrolment; 2) community-dwelling; and 3) independent in terms of basic activities of daily living (BADLs) and instrumental activities of daily living (IADLs). Exclusion criteria included: 1) a past medical history of dementia; 2) cognitive impairment as defined by Chinese Mini-Mental State Examination (CMMSE) ≤21 (17); 3) inability to walk 4.5m independently; and 4) residents of sheltered or nursing homes. Additional study details have been described previously (16). Ethics approval was obtained from the Domain Specific Review Board of the National Healthcare Group, and written informed consent was obtained from each participant.
Based upon criteria for sarcopenia and obesity, we classified subjects into four body composition phenotypes: 1) non-obese and non-sarcopenic (“normal”); 2) non-sarcopenic and obese (“obese”); 3) non-obese and sarcopenic (“sarcopenic”); and 4) obese and sarcopenic (“SO”). Sarcopenia was defined using the Asian Working Group for Sarcopenia (AWGS) criteria as follows: 1) low muscle mass (<7.0kg/m2 in men and <5.4kg/m2 in women); and 2) low handgrip strength (<26kg in men and <18kg in women) and/or slow usual gait speed (<0.8m/s) (4). Obesity was defined using three widely used measures: BMI, WC and FM%. For BMI, we employed the cut-off of ≥27.5kg/m2 to define obesity as recommended for Asian populations by the World Health Organization (18, 19). A validation study in Singapore reported that the mortality risk increases only modestly from BMI=21.5–27.5, but was clearly higher when BMI≥27.5 (20). WC cut-offs of >90cm and >80cm were used for males and females respectively, as per the Asia Pacific Consensus by the International Diabetes Foundation Consensus Worldwide Definition of the Metabolic Syndrome (21, 22). The FM% cut-off used was ≥30% and ≥40% in males and females respectively, as per the definitions used in recent studies (23, 24).
Data collection
We collected information on demographic characteristics, comorbidities, and geriatric syndromes. Evaluating geriatric syndromes included screening for muscle symptoms using the SARC-F questionnaire (25), frailty using the FRAIL questionnaire (26), cognitive function via the locally-validated CMMSE (17), depressive symptoms using the 15-item Geriatric Depression Scale (GDS-15) (27) and nutritional status via the Mini Nutritional Assessment (MNA) (28).
Anthropometric data collected include height and weight to derive BMI, and WC which was obtained 2.5cm above the umbilicus; this anatomical landmark has been shown to be best associated with abdominal fat mass measured by dual-energy X-ray absorptiometry (DXA) (29). We also measured FM% and appendicular skeletal mass (ASM) from whole-body DXA (Discovery™ APEX 13.3; Hologic, Bedford, MA, USA). ASM was defined as the sum of fat-free lean body mass in the four limbs and standardized using height2 to derive the relative appendicular skeletal muscle mass index (RASM).
We assessed muscle function in three ways. For muscle strength, we measured hand grip strength using a hydraulic hand dynamometer (North Coast Medical, Inc, Gilroy, CA, USA), with two trials of grip strength for each hand, and the average of four trials taken. For physical performance, usual gait speed was derived through the best result of two 3m walk tests. The Short Physical Performance Battery (SPPB), a 3-component test comprising balance, gait speed and repeated chair stands, was administered as a gauge of overall physical performance (30). Functional ability was assessed in terms of BADLs, IADLs and the 15-item Frenchay Activity Index (FAI), which has previously been used to assess physical activity among older adults in the Singaporean population (31).
Statistical analysis
Statistical Package for the Social Sciences (SPSS) version 22 was used for statistical analysis. The level of significance was set at 5%. For each definition of obesity, we ascertained the prevalence of each body composition phenotype and compared demographics, comorbidities, geriatric syndromes, body composition, muscle function measures, and functional ability between them. We also compared agreement between obesity definitions on SO diagnosis.
Continuous variables were analyzed using one-way analysis of variance (ANOVA) with Bonferroni correction for post-hoc comparison and the Kruskal-Wallis test for parametric and non-parametric variables respectively. Categorical variables were analyzed using the χ2 test. Agreement between the definitions was measured using Cohen's kappa.
To further ascertain the impact of obesity definitions on physical performance, we performed multiple linear regression to determine the association between body composition phenotypes and SPPB. We adjusted for important covariates with theoretical relevance (age, gender and education level) or statistical significance across all obesity definitions on univariate analysis. The normal and obese groups were combined to form the reference group for analysis, in order to determine the deleterious impact of sarcopenia alone as well as the additional effect of concurrent obesity as in SO.
Results
Baseline characteristics
Participants were predominantly female (68.5%) and Chinese (92.0%) with a mean age of 67.9±7.9 years and median education level of 10 years. The main comorbidities were hyperlipidemia (66%), hypertension (48%), and diabetes mellitus (21.5%). The sample comprised relatively healthy older adults, as evidenced by the high scores for BADLs, IADLs and SPPB; low SARC-F and FRAIL scores; and high CMMSE scores.
Prevalence
The prevalence of phenotypes varied greatly between definitions of obesity (Table 1). BMI identified only a single individual as SO (n=1, 0.5%), which is much lower compared to WC (n=21, 10.5%) and FM% (n=20, 10.0%). Notably, the single male individual being identified as SO using the BMI definition had WC (108.0cm) and FM% (44.90%) values which were much higher than the corresponding cut-offs used to define obesity in this study. Conversely, using WC and FM% to define obesity resulted in the SO group having a much lower median BMI than 27.5kg/m2 (23.0kg/m2 and 23.6kg/m2 respectively).
Table 1.
Prevalence of body phenotypes for different definitions of obesity
| Definition | Number of subjects, n (%) | |||
|---|---|---|---|---|
| Normal | Obese | Sarcopenic | Sarcopenic Obese | |
| Body Mass Index | 119 (59.5%) | 31 (15.5%) | 49 (24.5%) | 1 (0.5%) |
| ≥27.5kg/m2 | ||||
| Waist Circumference | 46 (23.0%) | 104 (52.0%) | 29 (14.5%) | 21 (10.5%) |
| >90cm in males | ||||
| >80cm in females | ||||
| Fat Mass Percentage | 81 (40.5%) | 69 (34.5%) | 30 (15.0%) | 20 (10.0%) |
| >30% in males | ||||
| >40% in females | ||||
Agreement
Using Cohen's kappa, inter-definitional agreement between groups was at best moderate, being the best between WC and FM% (κ=0.583) (Table 2). Amongst the 3 definitions, BMI had the poorest agreement (κ=0.364 and 0.529 with WC and FM% respectively).
Table 2.
Agreement between definitions of obesity
| 1st definition | 2nd definition | Cohen's kappa | Magnitude (35) |
|---|---|---|---|
| Body Mass Index | Waist Circumference | 0.364 | Fair |
| ≥27.5kg/m2 | >90cm in males | ||
| >80cm in females | |||
| Body Mass Index | Fat Mass Percentage | 0.529 | Moderate |
| ≥27.5kg/m2 | >30% in males | ||
| >40% in females | |||
| Waist Circumference | Fat Mass Percentage | 0.583 | Moderate |
| >90cm in males | >30% in males | ||
| >80cm in females | >40% in females | ||
Comparison of clinical characteristics between obesity definitions
Across all three obesity definitions, the SO phenotypes had the highest age, followed by sarcopenia, normal and obese phenotypes (p<0.001) (Table 3). In particular, the WC definition resulted in the SO phenotype containing a disproportionately low number of males compared with BMI and FM% definitions (19% versus 100% and 35%). SO individuals tended to have the lowest level of education. For comorbidities, the SO phenotype had a higher prevalence of strokes and transient ischemic attacks (TIAs) for all definitions of obesity (p<0.05). The prevalence of previous or current alcohol use was highest for the SO phenotype using the BMI and WC definitions, but highest for sarcopenic phenotype using the FM% definition (p<0.05). However, there was no difference in geriatric syndromes between body composition phenotypes (Supplementary Table 1).
Table 3.
Comparison of demographics, comorbidities, anthropometric data, muscle function and functional ability among body composition phenotypes for different definitions of obesity
| Variable | BMI ≥ 27.5kg/m2* | WC >90cm in males, >80cm in females | FM% >30% in males, >40% in females | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Normal (n=119) | Obese (n=31) | Sarcopenic (n=49) | SO (n=l) | p-value | Normal (n=46) | Obese (n=104) | Sarcopenic (n=29) | SO (n=21) | p-value | Normal (n=81) | Obese (n=69) | Sarcopenic (n=30) | SO (n=20) | p-value | |
| Demographics | |||||||||||||||
| Age, years | 66.6±7.5 | 66.5±6.7 | 71.9±8.1 | 80.0±0.0 | <0.001 | 66.4±7.4 | 66.7±7.3 | 70.5±7.0 | 74.1±9.1ab | <0.001 | 66.3±7.2 | 66.8±7.5 | 71.2±8.1de | 73.3±8.1ab | <0.001 |
| Male, n (%) | 38 (31.9%) | 9 (29.0%) | 15 (30.6%) | 1 (100.0%) | 0.514 | 24 (52.2%) | 23 (22.1%) | 12 (41.4%) | 4 (19.0%) | 0.001 | 28 (34.6%) | 19 (27.5%) | 9 (30.0%) | 7 (35.0%) | 0.801 |
| Education level, years | 10 (6–13) | 8 (5–11) | 8 (6–12) | 4 (4–4) | 0.024 | 12 (9–14) | 10 (6–12)f | 10 (6–12) | 6 (1–9)a | 0.001 | 10 (6–13) | 10 (6–12) | 8 (6–12) | 6 (3–12) | 0.246 |
| Race, n (%) | 0.005 | 0.294 | 0.143 | ||||||||||||
| Chinese | 111 (93.3%) | 23 (74.2%) | 49 (100.0%) | 1 (100.0%) | 44 (95.7%) | 90 (86.5%) | 29 (100.0%) | 21 (100.0%) | 74 (91.4%) | 60 (87.0%) | 30 (100.0%) | 20 (100.0%) | |||
| Comorbidities, n (%) | |||||||||||||||
| Stroke/transient ischemic attack | 1 (0.8%) | 0 (0.0%) | 3 (6.1%) | 1 (100.0%) | <0.001 | 1 (2.2%) | 0 (0.0%) | 2 (6.9%) | 2 (9.5%) | 0.026 | 1 (1.2%) | 0 (0.0%) | 2 (6.7%) | 2 (10.0%) | 0.029 |
| Previous/current smoker | 12 (10.0%) | 2 (6.5%) | 6 (12.3%) | 1 (100.0%) | 0.011 | 5 (10.9%) | 9 (8.7%) | 4 (13.8%) | 3 (14.3%) | 0.476 | 10 (12.3%) | 4 (5.8%) | 4 (13.3%) | 3 (15.0%) | 0.458 |
| Previous/current alcohol use | 6 (5.1%) | 0 (0.0%) | 4 (8.2%) | 1 (100.0%) | <0.001 | 4 (8.7%) | 2 (1.9%) | 2 (6.9%) | 3 (14.3%) | 0.026 | 4 (4.9%) | 2 (2.9%) | 4 (13.3%) | 1 (5.0%) | 0.048 |
| Anthropometric Data | |||||||||||||||
| Body mass index, kg/m2 | 23.5 (21.1–24.9) | 29.7 (29.0–31.0) | 21.9 (19.9–23.1) | 29.8 (29.8–29.8) | <0.001 | 21.1 (19.7–23.2) | 25.6 (23.9–28.3)f | 20.6 (18.9–22.4)e | 23.0 (22.0–24.1)bc | <0.001 | 22.2 (20.6–24.0) | 26.5 (25.0–29.4)f | 20.6 (19.0–21.9)de | 23.6 (22.4–24.3)bc | <0.001 |
| Waist circumference, cm | 84.9±7.4 | 98.7±5.7 | 82.2±7.3 | 108.0±0.0 | <0.001 | 79.3±7.0 | 91.5±7.1f | 78.3±6.4e | 88.7±6.2ac | <0.001 | 82.7±7.4 | 93.7±6.9f | 79.7±7.0e | 87.2±7.8bc | <0.001 |
| Fat mass percentage, % | 35.36±6.53 | 42.48±6.34 | 35.63±6.41 | 44.90±0.00 | <0.001 | 30.83±5.73 | 39.49±5.92f | 32.88±5.68e | 39.88±5.28ac | <0.001 | 32.96±5.70 | 41.38±5.74f | 32.65±4.67e | 40.57±5.96ac | <0.001 |
| RASM, kg/m2 | 5.89 (5.55–6.55) | 7.15 (6.27–7.97) | 5.21 (4.88–5.94) | 6.25 (6.25–6.25) | <0.001 | 5.96 (5.31–6.92) | 6.13 (5.72–7.13) | 5.23 (4.74–6.12)de | 5.16 (4.96–5.37)ab | <0.001 | 5.94 (5.59–6.65) | 6.28 (5.78–7.30) | 5.15 (4.95–5.96)de | 5.24 (4.82–6.17)ab | <0.001 |
| Muscle Function | |||||||||||||||
| Hand grip strength, kg | 21.5 (18.8–27.8) | 21.3 (18.3–27.3) | 16.0 (13.8–18.0) | 22.9 (22.9–22.9) | <0.001 | 26.1 (20.0–30.9) | 20.9 (18.5–25.4) | 16.5 (15.4–21.1)de | 15.3 (13.0–17.0)ab | <0.001 | 23.0 (19.0–29.4) | 20.8 (18.3–26.9) | 16.0 (13.7–20.6)de | 16.8 (14.0–17.7)ab | <0.001 |
| Gait speed (3m), m/s | 1.19±0.19 | 1.10±0.24 | 1.05±0.22 | 0.91±0.00 | <0.001 | 1.21±0.21 | 1.15±0.20 | 1.08±0.20 | 0.99±0.23ab | <0.001 | 1.20±0.20 | 1.13±0.20 | 1.04±0.24d | 1.05±0.18a | 0.001 |
| Short Physical Performance Battery | 12 (11–12) | 11 (11–12) | 12 (10–12) | 12 (12–12) | 0.008 | 12 (12–12) | 12 (11–12) | 12 (11–12) | 11 (10–12)a | 0.009 | 12 (11–12) | 12(11–12) | 12 (10–12) | 12 (11–12) | 0.019 |
| Functional Ability | |||||||||||||||
| Barthel Index of ADLs | 100 (100–100) | 100 (100–100) | 100 (100–100) | 100 (100–100) | 0.736 | 100 (100–100) | 100 (100–100) | 100 (100–100) | 100 (98–100) | 0.581 | 100 (100–100) | 100 (100–100) | 100 (95–100) | 100 (100–100) | 0.140 |
| Instrumental ADLs | 23 (23–23) | 23 (23–23) | 23 (23–23) | 23 (23–23) | 0.306 | 23 (23–23) | 23 (23–23) | 23 (23–23) | 23 (22–23)abc | 0.003 | 23 (23–23) | 23 (23–23) | 23 (23–23) | 23 (23–23) | 0.170 |
| Frenchay Activity Index | 34 (30–36) | 33 (31–36) | 31 (28–34) | 19 (19–19) | 0.004 | 34 (30–36) | 33 (30–36) | 30 (27–34)de | 31 (29–35) | 0.004 | 34 (30–36) | 33 (30–36) | 31 (28–33)d | 31 (27–34) | 0.007 |
ADLs = Activities of Daily Living; BMI = Body Mass Index; FM% = Fat Mass Percentage; RASM = Relative Appendicular Skeletal Muscle Mass Index; WC = Waist Circumference; * Post-hoc analyses were not performed due to the small sample size of the SO group; a. Significant post-hoc Bonferroni test between normal and SO (P<0.05); b. Significant post-hoc Bonferroni test between obese and SO (P<0.05); c. Significant post-hoc Bonferroni test between sarcopenic and SO (P<0.05); d Significant post-hoc Bonferroni test between normal and sarcopenic (P<0.05); e. Significant post-hoc Bonferroni test between obese and sarcopenic (P<0.05); f. Significant post-hoc Bonferroni test between normal and obese (P<0.05).
Comparison of anthropometric data, muscle function and functional ability measures between obesity definitions
Consistent with the operational definition of body composition phenotypes, significant differences were found in terms of anthropometric data such as BMI, WC, FM% and RASM across all 3 definitions of obesity (p<0.05) (Table 3). In particular, SO groups as defined by WC and FM% had much lower median BMIs (23.0kg/m2 and 23.6kg/m2 respectively) than the BMI cut-off for obesity of ≥27.5kg/m2. In addition, the WC definition resulted a divergent trend of RASM being lower in the SO compared to sarcopenia group (5.16kg/m2 vs 5.23kg/m2, post-hoc p=1.000), in contrast to the FM% definition where the converse was observed (5.24kg/m2 vs 5.15kg/m2, post-hoc p=1.000) (Table 3).
There was a statistically significant difference between phenotypes across all definitions of obesity in term of muscle strength and physical performance measures (p<0.05) (Table 3). The sarcopenic and SO phenotypes persistently performed poorer than the normal and obese groups, with the exception of the BMI definition where only a single individual was classified as SO. Notably, using the WC definition, SO consistently performed worse than the sarcopenic phenotype in all three measures of muscle function, in contrast to the inconsistent results between SO and sarcopenic phenotypes for the FM% definition. In terms of functional ability, the SO phenotype was found to have significantly lower IADL scores than normal, obese and sarcopenic phenotypes (post-hoc comparison with Bonferroni correction, p<0.05) only when defined using WC. There was no other significant difference between phenotypes in BADL for all definitions of obesity, whereas FAI scores were lowest for SO and sarcopenic phenotypes for all 3 definitions of obesity.
Multiple linear regression for SPPB
We performed multiple linear regression to assess the impact of sarcopenia and SO phenotypes on SPPB adjusted for significant covariates (Table 4). We adjusted for age, gender, education, stroke/transient ischemic attack, previous/current alcohol use and RASM. Small SO sample size (n=1) precluded multivariate analysis using the BMI definition. Using WC and FM% to define obesity, the total variance explained by the regression models was 15.1% [F(7, 192)=5.415, p<0.001] and 15.5% [F(7, 192)=4.388, p<0.001] respectively. Neither model violated assumptions of normality, linearity, multicollinearity and homoscedasticity.
Table 4.
Multiple linear regression analysis for Short Physical Performance Battery
| Variables | WC >90cm in males, >80cm in females | FM% >30% in males, >40% in females | ||||||
|---|---|---|---|---|---|---|---|---|
| Adjusted R2 | Standardized coefficient (β) | B ± Std. Error | p-value | Adjusted R2 | Standardized coefficient (β) | B ± Std. Error | p-value | |
| Age | 0.151 | −0.132 | −0.023 ± 0.013 | 0.086 | 0.155 | −0.170 | −0.029 ± 0.013 | 0.030 |
| Gender | −0.123 | −0.359 ± 0.289 | 0.216 | −0.169 | −0.494 ± 0.290 | 0.091 | ||
| Education (years) | 0.156 | 0.044 ± 0.020 | 0.034 | 0.178 | 0.050 ± 0.021 | 0.017 | ||
| Stroke/TIA | 0.040 | 0.350 ± 0.607 | 0.565 | 0.022 | 0.192 ± 0.620 | 0.758 | ||
| Alcohol consumption | −0.138 | −0.478 ± 0.246 | 0.053 | −0.144 | −0.501 ± 0.251 | 0.048 | ||
| RASM | −0.103 | −0.137 ± 0.128 | 0.286 | −0.127 | −0.169 ± 0.130 | 0.196 | ||
| Sarcopenic | 0.002 | 0.008 ± 0.287 | 0.977 | −0.184 | −0.698 ± 0.291 | 0.017 | ||
| Sarcopenic Obese | −0.261 | −1.154 ± 0.333 | 0.001 | −0.018 | −0.080 ± 0.338 | 0.813 | ||
FM% = Fat Mass Percentage; RASM = Relative Appendicular Skeletal Muscle Mass Index; TIA: Transient Ischemic Attack; WC = Waist Circumference
For the WC model, SO, but not sarcopenia, was significantly associated with a decreased SPPB score (β=−0.261, p=0.001). In contrast, the FM% model showed that sarcopenia, but not SO, was associated with decreased SPPB scores (β=−0.184, p=0.017). Other significant covariates included the association of increasing age and current or previous alcohol use with decreased SPPB (β=−0.170, p=0.030 and β=−0.144, p=0.048 respectively) in the FM% model, and the protective effect of more years of formal education in both WC and FM% models (β=0.156, p=0.034 and β=0.178, p=0.017 respectively).
Discussion
This study is, to our knowledge, the first to elucidate the impact of different obesity definitions on the prevalence, inter-definitional agreement, and muscle function of SO. These vary greatly between obesity definitions, supporting our initial position that different obesity definitions measure different constructs and are not interchangeable, and reiterating the importance of finding a consensus definition of SO. Of the three definitions studied, WC had the highest case detection rate for SO, and consistently identified individuals with the worst muscle function and IADL outcomes. It is noteworthy that we chose a method of WC measurement which is best associated with abdominal adiposity, as different anatomical locations for measuring WC can yield different results (29). Considering how WC is currently not widely adopted in SO studies despite its ease of measurement (14), our results support the case to further explore the consistent use of WC to define obesity in SO for research and clinical purposes.
While BMI has been used extensively in defining SO, our results indicate it was significantly less sensitive in identifying SO compared to the other two definitions despite using Asian-appropriate cut-offs and has the poorest agreement with the other obesity definitions. This suggests that BMI is unsuitable as a measure of obesity for the definition of SO in older adults in our local population. WHO cut-offs for BMI were designed to detect adverse cardiovascular and metabolic outcomes (19), and appear unsuitable for identifying poor muscle function outcomes in SO. BMI fails to account for the higher body fat composition and loss of lean body mass with age (11), and does not assess body fat distribution. This is pertinent as it is central obesity rather than peripheral fat deposition which is associated with higher morbidity (18).
Our results suggest that WC and FM% intrinsically measure different constructs and are not interchangeable. Despite having similar SO case detection rates (10.5% and 10.0% respectively), the agreement between them was only moderate (κ=0.583), as has been highlighted in the existing literature (32). Only the WC definition consistently resulted in the SO phenotype having the worst scores amongst the body composition phenotypes for muscle function performance measures and IADLs. In particular, for the composite physical performance measure of SPPB, SO performed significantly worse in multivariate analysis only for the WC definition. This is pertinent as lower SPPB scores are associated with a higher risk of hospitalization, institutionalization, morbidity and mortality (30).
The observed differences in association with muscle function measures and IADLs can be attributed to WC being a more specific measure of abdominal obesity as opposed to FM%, which fails to distinguish body fat distribution. This corroborates the putative role of central obesity, rather than generalized obesity per se, in the pathogenesis of adverse physical performance and functional outcomes arising from SO. In support of this, the English Longitudinal Study of Aging (ELSA) recently reported that abdominal obesity is associated with a decline in muscle strength (33). We posit that WC, as a specific measure of central obesity, provides an indication of the risk of ectopic fat deposition within muscle tissue and thus increased intramuscular adipose tissue (IMAT) (8). IMAT has been linked with adipose tissue inflammation due to the secretion of cytokines such as monocyte chemoattractant protein-1, and the resultant proinflammatory milieu and accelerated muscle catabolism can then lead to decreased muscle mass and impaired strength (34, 35). Thus, using WC to define obesity among older adults may result in more accurate characterization of SO and potentially earlier detection of the downstream consequences of reduced physical performance.
This study was made possible through a comprehensive evaluation which permitted accurate characterization of body composition phenotypes, and thorough assessment of clinically relevant outcomes comprising various measures of muscle function and functional ability. Limitations include the cross-sectional study design, such that reverse causality cannot be excluded; study results thus represent point-prevalence and associations rather than definitive conclusions about causality. The results may not be generalizable to other ethnic groups or other Asian populations due to the predominance of Chinese individuals. They also may not apply to less robust older adults. Further studies in non-Chinese populations and in more heterogeneous populations of older adults, as well as longitudinal studies to determine the trajectory of SO relative to other body composition phenotypes, are necessary to validate our findings. Lastly, the small sample size may result in inadvertent type II error and precluded subgroup analysis for important covariates such as gender.
Conclusion
To our knowledge, this is the first study that compares the impact of obesity definitions on SO, reiterating the importance of finding a consensus definition of obesity for SO. Our results demonstrate that obesity definitions significantly affect the prevalence of SO, agreement between definitions is at best moderate, and muscle function and functional ability differ depending on the definition being employed.
Our findings suggest that WC should be further explored as a means of defining obesity in SO. It is associated with poorer physical and functional performance outcomes, has an established pathophysiological association with impaired muscle strength and mass, and is easily measured clinically, thus enhancing its utility and relevance for defining sarcopenic obesity. In contrast, BMI appears unsuitable for use in the context of SO. It has a low case detection rate, poor agreement with other obesity definitions, and is intrinsically unable to distinguish between the components of body composition.
Disclosure: The authors report no conflicts of interest in this work.
Acknowledgments
This study was supported by Lee Foundation grant 2013. We extend our appreciation to the Senior Activity Centers and the study participants who have graciously consented to participate in the study.
Footnotes
Electronic Supplementary Material
Supplementary material is available for this article at https://doi.org/10.14283/jfa.2019.28 and is accessible for authorized users.
Electronic supplementary material
Supplementary Table 1: Additional comparisons of demographics, comorbidities and geriatric syndromes among body composition phenotypes for different definitions of obesity
References
- 1.Frontiers in Medicine. 2018;5(168) [Google Scholar]
- 2.Roubenoff R. Sarcopenic obesity: the confluence of two epidemics. Obes Res. 2004;12(6):887–888. doi: 10.1038/oby.2004.107. 10.1038/oby.2004.107 [DOI] [PubMed] [Google Scholar]
- 3.Organisation WH. Global Health and Aging. 2011.
- 4.Chen LK, Liu LK, Woo J, Assantachai P, Auyeung TW, Bahyah KS, et al. Sarcopenia in Asia: consensus report of the Asian Working Group for Sarcopenia. J Am Med Dir Assoc. 2014;15(2):95–101. doi: 10.1016/j.jamda.2013.11.025. 10.1016/j.jamda.2013.11.025 [DOI] [PubMed] [Google Scholar]
- 5.Wannamethee SG, Atkins JL. Muscle loss and obesity: the health implications of sarcopenia and sarcopenic obesity. The Proceedings of the Nutrition Society. 2015;74(4):405–412. doi: 10.1017/S002966511500169X. 10.1017/S002966511500169X [DOI] [PubMed] [Google Scholar]
- 6.Kemmler W, von Stengel S, Engelke K, Sieber C, Freiberger E. Prevalence of sarcopenic obesity in Germany using established definitions: Baseline data of the FORMOsA study. Osteoporos Int. 2016;27(1):275–281. doi: 10.1007/s00198-015-3303-y. 10.1007/s00198-015-3303-y [DOI] [PubMed] [Google Scholar]
- 7.Batsis JA, Barre LK, Mackenzie TA, Pratt SI, Lopez-Jimenez F, Bartels SJ. Variation in the prevalence of sarcopenia and sarcopenic obesity in older adults associated with different research definitions: dual-energy X-ray absorptiometry data from the National Health and Nutrition Examination Survey 1999–2004. Journal of the American Geriatrics Society. 2013;61(6):974–980. doi: 10.1111/jgs.12260. 10.1111/jgs.12260 [DOI] [PubMed] [Google Scholar]
- 8.Kalinkovich A, Livshits G. Sarcopenic obesity or obese sarcopenia: A cross talk between age-associated adipose tissue and skeletal muscle inflammation as a main mechanism of the pathogenesis. Ageing research reviews. 2017;35:200–221. doi: 10.1016/j.arr.2016.09.008. 10.1016/j.arr.2016.09.008 [DOI] [PubMed] [Google Scholar]
- 9.Baumgartner RN. Body composition in healthy aging. Annals of the New York Academy of Sciences. 2000;904(1):437–448. doi: 10.1111/j.1749-6632.2000.tb06498.x. 10.1111/j.1749-6632.2000.tb06498.x [DOI] [PubMed] [Google Scholar]
- 10.Baumgartner RN, Wayne SJ, Waters DL, Janssen I, Gallagher D, Morley JE. Sarcopenic obesity predicts instrumental activities of daily living disability in the elderly. Obes Res. 2004;12(12):1995–2004. doi: 10.1038/oby.2004.250. 10.1038/oby.2004.250 [DOI] [PubMed] [Google Scholar]
- 11.Batsis JA, Villareal DT. Sarcopenic obesity in older adults: aetiology, epidemiology and treatment strategies. Nature Reviews Endocrinology. 2018;14(9):513–537. doi: 10.1038/s41574-018-0062-9. 10.1038/s41574-018-0062-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Studenski SA, Peters KW, Alley DE, Cawthon PM, McLean RR, Harris TB, et al. The FNIH sarcopenia project: rationale, study description, conference recommendations, and final estimates. The journals of gerontology Series A, Biological sciences and medical sciences. 2014;69(5):547–558. doi: 10.1093/gerona/glu010. 10.1093/gerona/glu010 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Fielding RA, Vellas B, Evans WJ, Bhasin S, Morley JE, Newman AB, et al. Sarcopenia: an undiagnosed condition in older adults. Current consensus definition: prevalence, etiology, and consequences. International working group on sarcopenia. J Am Med Dir Assoc. 2011;12(4):249–256. doi: 10.1016/j.jamda.2011.01.003. 10.1016/j.jamda.2011.01.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Akishita M, Kozaki K, Iijima K, Tanaka T, Shibasaki K, Ogawa S, et al. Chapter 1 Definitions and diagnosis of sarcopenia. Geriatr Gerontol Int. 2018;18(Suppl1):7–12. doi: 10.1111/ggi.13311. 10.1111/ggi.13311 [DOI] [PubMed] [Google Scholar]
- 15.Garvey WT, Mechanick JI, Brett EM, Garber AJ, Hurley DL, Jastreboff AM, et al. American association of clinical endocrinologists and American college of endocrinology comprehensive clinical practice guidelines for medical care of patients with obesity. Endocrine practice: official journal of the American College of Endocrinology and the American Association of Clinical Endocrinologists. 2016;22(Suppl3):1–203. doi: 10.4158/EP161365.GL. 10.4158/EP161365.GL [DOI] [PubMed] [Google Scholar]
- 16.Tay L, Ding YY, Leung BP, Ismail NH, Yeo A, Yew S, et al. Sex-specific differences in risk factors for sarcopenia amongst community-dwelling older adults. Age (Dordrecht, Netherlands) 2015;37(6):121. doi: 10.1007/s11357-015-9860-3. 10.1007/s11357-015-9860-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Sahadevan S, Lim PP, Tan NJ, Chan SP. Diagnostic performance of two mental status tests in the older chinese: influence of education and age on cut-off values. International journal of geriatric psychiatry. 2000;15(3):234–241. doi: 10.1002/(sici)1099-1166(200003)15:3<234::aid-gps99>3.0.co;2-g. 10.1002/(SICI)1099-1166(200003)15:3<234::AID-GPS99>3.0.CO;2-G [DOI] [PubMed] [Google Scholar]
- 18.Lee YS, Biddle S, Chan MF, Cheng A, Cheong M, Chong YS, et al. Health Promotion Board-Ministry of Health Clinical Practice Guidelines: Obesity. Singapore Medical Journal. 2016;57(6):292–300. doi: 10.11622/smedj.2016103. 10.11622/smedj.2016103 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Who EC. Appropriate body-mass index for Asian populations and its implications for policy and intervention strategies. Lancet (London, England) 2004;363(9403):157. doi: 10.1016/S0140-6736(03)15268-3. 10.1016/S0140-6736(03)15268-3 [DOI] [PubMed] [Google Scholar]
- 20.Odegaard AO, Pereira MA, Koh WP, Gross MD, Duval S, Yu MC, et al. BMI, all-cause and cause-specific mortality in Chinese Singaporean men and women: the Singapore Chinese health study. PloS one. 2010;5(11):e14000. doi: 10.1371/journal.pone.0014000. 10.1371/journal.pone.0014000 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Organization WH. Health Communications Australia; Sydney: 2000. The Asia-Pacific perspective: redefining obesity and its treatment. [Google Scholar]
- 22.Alberti KG, Zimmet P, Shaw J. Metabolic syndrome—a new world-wide definition. A Consensus Statement from the International Diabetes Federation. Diabetic medicine: a journal of the British Diabetic Association. 2006;23(5):469–480. doi: 10.1111/j.1464-5491.2006.01858.x. 10.1111/j.1464-5491.2006.01858.x [DOI] [PubMed] [Google Scholar]
- 23.Scott D, Daly RM, Sanders KM, Ebeling PR. Fall and fracture risk in sarcopenia and dynapenia with and without obesity: the role of lifestyle interventions. Current osteoporosis reports. 2015;13(4):235–244. doi: 10.1007/s11914-015-0274-z. 10.1007/s11914-015-0274-z [DOI] [PubMed] [Google Scholar]
- 24.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. The journals of gerontology Series A, Biological sciences and medical sciences. 2013;68(2):168–174. doi: 10.1093/gerona/gls109. 10.1093/gerona/gls109 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Malmstrom TK, Morley JE. SARC-F: a simple questionnaire to rapidly diagnose sarcopenia. J Am Med Dir Assoc. 2013;14(8):531–532. doi: 10.1016/j.jamda.2013.05.018. 10.1016/j.jamda.2013.05.018 [DOI] [PubMed] [Google Scholar]
- 26.Morley JE, Malmstrom TK, Miller DK. A simple frailty questionnaire (FRAIL) predicts outcomes in middle aged African Americans. J Nutr Health Aging. 2012;16(7):601–608. doi: 10.1007/s12603-012-0084-2. 10.1007/s12603-012-0084-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Sheikh JI, Yesavage JA. Geriatric Depression Scale (GDS): Recent evidence and development of a shorter version. Clinical Gerontologist: The Journal of Aging and Mental Health. 1986;5(1–2):165–173. [Google Scholar]
- 28.Guigoz Y. The Mini Nutritional Assessment (MNA) review of the literature—What does it tell us? J Nutr Health Aging. 2006;10(6):466–485. PubMed PMID: 17183419, discussion 85-7. [PubMed] [Google Scholar]
- 29.Guerra RS, Amaral TF, Marques EA, Mota J, Restivo MT. Anatomical location for waist circumference measurement in older adults: a preliminary study. Nutricion hospitalaria. 2012;27(5):1554–1561. doi: 10.3305/nh.2012.27.5.5922. PubMed PMID: 23478705. [DOI] [PubMed] [Google Scholar]
- 30.Puthoff ML. Outcome measures in cardiopulmonary physical therapy: short physical performance battery. Cardiopulmonary physical therapy journal. 2008;19(1):17. 10.1097/01823246-200819010-00005 [PMC free article] [PubMed] [Google Scholar]
- 31.Yeo A, Chong M, Tay L, Yew SP, Tan C, Ismail N, et al. TRENDS AND PREDICTORS OF 2-YEAR ACTIVITY PARTICIPATION IN COMMUNITY DWELLING OLDER ADULTS. Innovation in Aging. 2017;1(suppl_1):905–906. 10.1093/geroni/igx004.3245 [Google Scholar]
- 32.Kemmler W, Teschler M, Weisenfels A, Sieber C, Freiberger E, von Stengel S. Prevalence of sarcopenia and sarcopenic obesity in older German men using recognized definitions: high accordance but low overlap! Osteoporos Int. 2017;28(6):1881–1891. doi: 10.1007/s00198-017-3964-9. 10.1007/s00198-017-3964-9 [DOI] [PubMed] [Google Scholar]
- 33.de Carvalho DHT, Scholes S, Santos JLF, de Oliveira C, Alexandre TdS. Does abdominal obesity accelerate muscle strength decline in older adults? Evidence from the English Longitudinal Study of Ageing. The Journals of Gerontology: Series A. 2018 doi: 10.1093/gerona/gly178. gly178-gly. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Lim JP, Leung BP, Ding YY, Tay L, Ismail NH, Yeo A, et al. Monocyte chemoattractant protein-1: a proinflammatory cytokine elevated in sarcopenic obesity. Clinical Interventions in Aging. 2015;10:605–609. doi: 10.2147/CIA.S78901. PubMed PMID: 25848236, PMCID 4378871. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.McHugh ML. Interrater reliability: the kappa statistic. Biochemia Medica. 2012;22(3):276–282. 10.11613/BM.2012.031 [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Table 1: Additional comparisons of demographics, comorbidities and geriatric syndromes among body composition phenotypes for different definitions of obesity