Incidence, reversibility, risk factors and the protective effect of high body mass index against sarcopenia in community-dwelling older Chinese adults

bs_bs_banner Geriatr Gerontol Int 2014; 14 (Suppl. 1): 15 28 ORIGINAL ARTICLE Incidence, reversibility, risk factors and the protective effect of high body mass index against sarcopenia in community-dwelling older Chinese adults Ruby Yu, 1 Moses Wong, 1 Jason Leung, 2 Jenny Lee, 3,4 Tung Wai Auyeung 3,5 and Jean Woo 1 1 Department of Medicine and Therapeutics, 2 Jockey Club Center for Osteoporosis Care and Control, 3 The S. H. Ho Center for Gerontology and Geriatrics, The Chinese University of Hong Kong, 4 Department of Medicine and Geriatrics, Shatin Hospital, and 5 Department of Medicine and Geriatrics, Pok Oi Hospital, New Territories, Hong Kong Aim: We examined the incidence and the reversibility of sarcopenia and their associated factors over a 4-year period using the European Working Group on Sarcopenia in Older People (EWGSOP) criteria. Methods: A total of 4000 community-dwelling older adults aged 65 years were evaluated for which detailed information regarding demographics, socioeconomic, medical history, lifestyle, and clinical factors were documented at baseline, 2 years, and 4 years later. Sarcopenia was defined according to the EWGSOP algorithm. Incident sarcopenia and its reversibility were documented at each follow-up year, and related to possible factors. Results: At baseline, of the 4000 participants, 361 (9.0%) had sarcopenia. Between baseline and 2-year follow-up, 6.0% of the participants without sarcopenia at baseline had developed sarcopenia, and 18.8% of the initially sarcopenic participants had reverted to normal. Between baseline and 4-year follow-up, the corresponding figures were 6.3% and 14.1%, respectively. The average annual incidence over 4 years was 3.1%. After multivariate adjustments, older age, female sex, presence of chronic obstructive pulmonary disease, presence of stroke, higher physical activity levels, presence of instrumental activities of daily living impairments, and lower body mass index were associated with incident sarcopenia, whereas younger age, female sex, higher body mass index and absence of instrumental activities of daily living impairments, but not physical activity, were associated with its reversibility. Protein and vitamin D intake were not significantly associated with sarcopenia incidence or its reversibility. Conclusion: Sarcopenia incidence increases with age, but is potentially reversible in a Chinese elderly population. High body mass index is protective against sarcopenia incidence and its reversibility. Increasing physical activity and maintaining a healthy weight could be beneficial in the prevention of sarcopenia. Geriatr Gerontol Int 2014; 14 (Suppl. 1): 15 28. Keywords: body mass index, Chinese, elderly, physical activity, sarcopenia. Introduction Muscle mass decreases as part of the physiological changes with age, and contributes to physical function decline. 1 However, the rate of decrease, and possible reversibility, might be amendable to intervention. The term sarcopenia was first coined to describe this condition 2,3 in order to raise awareness, and place it in a similar category as osteoporosis and osteopenia. By Accepted for publication 25 November 2013. Correspondence: Dr Ruby Yu PhD, Department of Medicine and Therapeutics, The Chinese University of Hong Kong, New Territories, Hong Kong. Email: rubyyu@cuhk.edu.hk developing a universal definition, research into prevalence, risk factors and intervention might be facilitated, as was the case for osteoporosis. Although this condition is now accepted as one of the geriatric syndromes, a universal consensus in definition is still lacking. Initially, the definition consists of the measurement of appendicular mass divided by height in meter squared. 4 However, some use body weight as the denominator. 5 Another indicator used is percentage skeletal muscle index, calculated as the total skeletal mass divided by weight 100. 6 More recently, there is a gathering consensus worldwide that the definition should include a measure of muscle power and/or physical performance measure. 7,8 At the same time, the concept of dynapenia has emerged, 9 which describes the age-related loss in 2014 Japan Geriatrics Society doi: 10.1111/ggi.12220 15

RYuet al. muscle power and seeks to differentiate muscle mass from power. Many definitions include absolute cut-off values for these measurements, and it is uncertain whether these can be translated to populations with different ethnicity with different body size and shape. For example, values of appendicular skeletal muscle mass (ASM)/height 2 related to incident physical limitation are lower in Chinese older people than Caucasians. 10 Recently, the Asian Working Group for Sarcopenia Research adopted an algorithm of sarcopenia that is similar to the European Working Group on Sarcopenia in Older People (EWGSOP), which avoided the use of absolute cut-off values by using the lowest 20th percentile of population values (unpublished data). In reality, an internationally agreed method of definition might not need to include absolute values, unless a rigorous multinational prevalence study is to be carried out. Such a study might be less meaningful than those measuring changes, such as documenting the rate of decline or reversibility, risk factors affecting the decline, the inclusion of sarcopenia in community assessments and interventional studies where sarcopenia is the primary outcome measure. From the public health perspective, it would be important to document not only the prevalence, but incidence of sarcopenia in aging populations, as it predisposes the older individual to adverse outcomes, such as falls and fractures, dependency, use of health services, and mortality. 10 13 Furthermore, the identification of risk factors could allow preventive efforts to reduce the incidence. In a study of 4000 Chinese men and women aged 65 years and older living in the community, we addressed the question of the incidence of sarcopenia over a 4-year period using the EWGSOP criteria, and examined the risk factors predisposing to the onset of sarcopenia. The reversibility of sarcopenia and its predictors were also examined. Methods Participants A total of 4000 community-dwelling Chinese men and women aged 65 years and older were recruited for a cohort study on osteoporosis and general health (Mr. Os) in Hong Kong between August 2001 and February 2003 by placing recruitment notices in community centers for older adults and housing estates. Several talks were also given at these centers explaining the purpose, procedures and investigations to be carried out. Participants were volunteers, and the aim was to recruit a stratified sample so that approximately 33% would each be aged 65 69 years, 70 74 years, and 75 years and older. Those who were unable to walk independently, had had a bilateral hip replacement or were not competent to give informed consent were excluded. Eligible participants were invited to attend a health check at the School of Public Health, The Chinese University of Hong Kong. A team of trained research assistants admonished the study questionnaire and took physical measurements for each participant on the same day. The cohort was invited to re-attend for repeat questionnaire interviews and physical measurements after 2 and 4 years. Details of the survey population have been reported elsewhere. 14 All participants gave written consent, and the study was approved by the Clinical Research Ethics Committee of the Chinese University of Hong Kong. Questionnaire The information from the questionnaire used in the present study included demographics, socioeconomic status, self-reported history of chronic diseases (chronic obstructive pulmonary disease [COPD], diabetes, hypertension, stroke and cancer), smoking, physical activity, dietary intake, cognitive function and instrumental activities of daily living (IADL). Socioeconomic status was measured on the basis of the education level, the community ladder and the Hong Kong ladder. 15 Physical activity levels were assessed using the Physical Activity Scale of the Elderly (PASE). This is a 12-item scale measuring the average number of hours per day spent in leisure, household and occupational physical activities over the previous 7 days. Activity weights for each item were determined based on the amount of energy expended, and each item score was calculated by multiplying the activity weight by activity daily frequency. A summary score of all the items reflected the daily physical activity level. 16 Dietary intake was assessed at baseline using a validated semi-quantitative food frequency questionnaire. 17 Cognitive function was assessed using the cognitive score of the Chinese version of the Community Screening Instrument of Dementia (CSI-D), the validity of which has been examined elsewhere. 18 The cut-off point for probable dementia is <28.4. 18,19 IADL impairments were assessed by noting any impairment in walking two to three blocks outside on level ground, climbing up 10 steps without resting, preparing own meals, doing heavy housework such as scrubbing floors or washing windows, and doing own shopping for groceries or clothes, a concept originally developed by Lawton and Brody. 20 Physical measurements Body weight was measured, with participants wearing a light dressing gown, by the Physician Balance Beam Scale (Health-O-Meter, Arlington Heights, IL, USA). Height was measured by the Holtain Harpenden stadiometer (Holtain, Crosswell, UK). Body mass index 16 2014 Japan Geriatrics Society

Sarcopenia incidence and its risk factors (BMI) was calculated by dividing the weight in kilogram by height in meter squared. Body composition was measured by dual energy X-ray absorptiometry (DXA) using a Hologic Delphi W4500 densitometer (Hologic Delphi, auto whole body version 12.4; Hologic, Bedford, MA, USA). Total appendicular skeletal muscle mass (ASM) was calculated as the sum of appendicular lean mass minus bone mineral content of both arms and legs. Grip strength was measured using a dynamometer (JAMAR Hand Dynamometer 5030JO; Sammons Preston, Bolingbrook, IL, USA). Two readings were taken from each side, and the average value between right and left was used for analysis. Gait speed was measured using the best time in seconds to complete a walk along a straight line 6-m long. A warm up period of <5 min was followed by two walks, and the best time recorded. Diagnosis of sarcopenia Sarcopenia was defined according to the EWGSOP algorithm, 8 in which a person who has low muscle mass, low muscle strength and/or low physical performance was categorized as having sarcopenia. With reference to the lowest quintile value of the distribution of the study population, low muscle mass was defined as ASM index (ASM/height 2 ) <6.52 kg/m 2 for males and <5.44 kg/m 2 for females; low muscle strength was defined as grip strength 28 kg for males and 18 kg for females; and low physical performance as gait speed 0.8 m/s for both males and females. Statistical analyses Characteristics of individuals at baseline were presented using means and standard deviations (SD) for continuous variables, and frequencies and percentages for categorical variables. The changes of sarcopenia categories from baseline to 2-year, 2- to 4-year, and baseline to 4-year follow-up were presented by age group and sex. Incidence proportions of sarcopenia at each follow-up year were calculated by the number of new sarcopenic cases within a specified time period divided by the size of the population initially at risk. The average annual incidence over 4 years was calculated by weighted average of the number of new sarcopenic cases per population initially at risk per 2 years from baseline to 2-year and 2- to 4-year follow-up. Risk factors for incident sarcopenia at each follow-up year were first analyzed individually using logistic regressions. Subsequently, multiple logistic regression models were constructed by stepwise and backward elimination algorithms. In these models, age, sex, education levels, socioeconomic ladders, medical history, lifestyle and nutritional factors, cognitive function, IADL impairments, and BMI were included. Analyses were repeated for the reversibility of sarcopenia at each follow-up year. As a rule of thumb for carrying out logistic regression analysis, at least 10 events per variable (EPV) are required in most instances; 21 variables with a EPV value of less than 10 were excluded in the multiple logistic regression models of the incidence and reversibility of sarcopenia. All analyses were carried out using the Windows-based SPSS Statistical Package (version 17.0; SPSS, Chicago, IL, USA), and P-values less than 0.05 were considered statistically significant. Results Baseline characteristics By February 2003, 2000 men and 2000 women aged 65 years or older with a mean age of 72.5 ± 5.2 years were participating in the study (Table 1). Of these participants, 45.6% had primary level of education or above, 6.9% were current smokers and the mean BMI was 23.7 kg/m 2. Hypertension was the most frequent selfreported chronic disease (42.7%). There were 606 (15.2%) participants with probable dementia and 25.8% had IADL impairments. Changes in sarcopenia categories and incidence proportions Figure 1 and Table 2 show the onset of disease over the follow-up period. Of the 4000 participants at baseline, 361 (9.0%) had sarcopenia. Between baseline and 2-year follow-up, 217 (6.0%) of the participants without sarcopenia at baseline had developed sarcopenia, and 68 (18.8%) of the initially sarcopenic participants had reverted to normal. Between baseline and 4-year followup, the corresponding figures were 6.3% and 14.1%, respectively. The incidence proportions of sarcopenia from baseline to 2-year, 2- to 4-year, and baseline to 4-year follow-up were 6.9, 5.4, and 7.8%, respectively. The average annual incidence over 4 years was 3.1% (2.9% for male and 3.3% for female). Overall, the incidence of sarcopenia increased with age for both sexes, and males aged 85 years and older tended to have a substantially higher incidence than their female counterparts. Risk factors of incident sarcopenia Factors associated with incident sarcopenia from baseline to 2-year, 2- to 4-year, and baseline to 4-year follow-up are shown in Table 3. After adjustments for demographics, socioeconomic status, medical history, lifestyle and nutritional factors, cognitive function, IADL impairments, and BMI, age (adjusted OR 1.11, 95% CI 1.07 1.14), presence of stroke (adjusted OR 2.56, 95% CI 1.32 4.95), physical activity (adjusted 2014 Japan Geriatrics Society 17

RYuet al. Table 1 Baseline characteristics of study population All (n = 4,000) Demographics Age (years) 72.5 ± 5.2 Socioeconomic status Education level Have not received any education 856 (21.4) Some primary school 1324 (33.1) Primary school 683 (17.1) Secondary school / Matriculation 747 (18.7) University / College 390 (9.8) Socioeconomic status ladder Community 1 3 244 (6.4) 4 5 1030 (26.9) 6 10 2561 (66.8) Socioeconomic status ladder Hong Kong 1 3 1120 (29.8) 4 5 1715 (45.6) 6 10 925 (24.6) Medical history Chronic obstructive pulmonary disease 333 (8.3) Diabetes 579 (14.5) Hypertension 1707 (42.7) Stroke 175 (4.4) Cancer 177 (4.4) Lifestyle factors and dietary intake Current smoker 275 (6.9) Physical activity (PASE total score) 91.3 ± 43.0 Protein intake (g/day) 76.5 ± 33.5 Vitamin D intake (IU/day) 13.3 ± 21.0 Energy intake (Kcal/day) 1841.4 ± 587.5 Cognitive function CSI-D level Normal (CSI-D score 29.5) 2931 (73.3) Borderline (28.4 CSI-D score < 29.5) 463 (11.6) Probable dementia (CSI-D 606 (15.2) score < 28.4) Instrumental activities of daily living impairments No. IADL impairment 0 2967 (74.2) 1 2 875 (21.9) 3 5 154 (3.9) Body measurements Body mass index (kg/m 2 ) 23.7 ± 3.3 Figures are based on valid cases (n) observed at baseline, for socioeconomic status ladder Community (n = 3835), socioeconomic status ladder Hong Kong (n = 3760), protein intake (g/day; n = 3995), vitamin D intake (IU/day; n = 3995) and number of instrumental activities of daily living (IADL) impairments (n = 3996). Data are presented as mean ± SD or number (percentage). Percentages might not sum up to 100% due to rounding. CSI-D, Community Screening Instrument of Dementia; PASE, Physical Activity Scale of the Elderly. OR 0.995, 95% CI 0.991 0.999), IADL impairments (adjusted OR 2.12, 95% CI 1.49 3.02) and BMI (adjusted OR 0.66, 95% CI 0.62 0.70) were associated with the development of sarcopenia from baseline to 4-year follow-up. Female sex (adjusted OR 1.58; 95% CI 1.15 2.16) and presence of COPD (adjusted OR 1.84, 95% CI 1.02 3.31) were associated with incident sarcopenia from baseline to 2-year and from 2- to 4- year follow-up, respectively. Protein and vitamin D intake were not significantly associated with incident sarcopenia. Risk factors of the reversibility of sarcopenia Factors associated with the reversibility of sarcopenia from baseline to 2-year, 2- to 4-year, and baseline to 4-year follow-up are shown in Table 4. After adjustments for demographics, socioeconomic status, medical history, lifestyle factors, cognitive function, IADL impairments and BMI, age (adjusted OR 0.90, 95% CI 0.84 0.96) was the only significant predictor consistently associated with the reversibility of sarcopenia across the different study periods. BMI was associated with the reversibility of sarcopenia (adjusted OR 1.16, 95% CI 1.02 1.31) from baseline to 2-year follow-up, whereas females (adjusted OR 2.65, 95% CI 1.50 4.68) and those without IADL impairments at baseline (adjusted OR 0.35, 95% CI 0.18 0.71) were more likely to return to non-sarcopenic from 2- to 4-year followup. However, physical activity, protein or vitamin D intakes were not significantly associated with the reversibility of sarcopenia. Discussion Understanding the causes, prevention and treatment of sarcopenia is increasingly important in geriatric medicine. Although much research effort has been directed toward development of a universal definition of sarcopenia and determining its prevalence, risk factors, and consequences, there have been few studies aimed at examining the incidence and the reversibility of sarcopenia and their risk factors. To our knowledge, this is the first prospective study of community-dwelling Chinese examining the factors predisposing to the development of sarcopenia and its reversibility using updated criteria. The average annual incidence of sarcopenia over 4 years was 3.1%. The present findings also confirmed the incidence of sarcopenia increased with age. This is compatible with the extensive literature documenting the loss of muscle mass and function that occurs with aging. Annual loss of muscle mass has been reported as 1 2% at the age of 50 years onwards, 22 with the rates to be higher in men than in women. 23 However, women, on average, have a longer life expectancy than men, which implies that sarcopenia for 18 2014 Japan Geriatrics Society

Sarcopenia incidence and its risk factors (n=134) (n=207) (n=18) (n=55) (n=24) (n=361) (n=68) (n=30) (n=14) (n=11) (n=86) (n=3) (n=72) (n=4,000) (n=97) (n=217) (n=73) (n=47) (n=119) (n=3,639) (n=2,925) (n=2,486) (n=320) (n=13) Figure 1 Dynamic flow of sarcopenic participants by time of observation. *Missing observations include those who died during follow-up, default cases and returning participants who failed to complete measurement. (n=497) (n=110) (n=374) 2014 Japan Geriatrics Society 19

RYuet al. Table 2 Sarcopenia status and age- and sex-specific incidence proportion by follow-up periods Both sexes Male Female All ages 65 74 years 75 84 years 85 years All ages 65 74 years 75 84 years 85 years All ages n % n % n % n % n % n % n % n % n % Baseline to 2-year Remained normal 2925 85.6 1128 91.3 364 76.2 13 46.4 1505 86.4 1027 88.8 367 75.5 26 78.8 1420 84.8 Normal to sarcopenic 217 6.4 46 3.7 41 8.6 4 14.3 91 5.2 73 6.3 50 10.3 3 9.1 126 7.5 Sarcopenic to normal 68 2.0 21 1.7 11 2.3 2 7.1 34 2.0 18 1.6 16 3.3 0 0.0 34 2.0 Remained sarcopenic 207 6.1 41 3.3 62 13.0 9 32.1 112 6.4 38 3.3 53 10.9 4 12.1 95 5.7 Total 3417 100.0 1236 100.0 478 100.0 28 100.0 1742 100.0 1156 100.0 486 100.0 33 100.0 1675 100.0 Incidence proportion (%) 217 6.9 46 3.9 41 10.1 4 23.5 91 5.7 73 6.6 50 12.0 3 10.3 126 8.2 2- to 4-year Remained normal 2516 84.4 1009 89.9 270 72.8 9 50.0 1288 85.2 913 81.4 299 73.8 16 72.7 1228 83.5 Normal to sarcopenic 143 4.8 45 4.0 31 8.4 3 16.7 79 5.2 36 3.2 27 6.7 1 4.5 64 4.4 Sarcopenic to normal 91 3.1 20 1.8 6 1.6 0 0.0 26 1.7 41 3.7 24 5.9 0 0.0 65 4.4 Remained sarcopenic 231 7.7 48 4.3 64 17.3 6 33.3 118 7.8 53 4.7 55 13.6 5 22.7 113 7.7 Total 2981 100.0 1122 100.0 371 100.0 18 100.0 1511 100.0 1043 100.0 405 100.0 22 100.0 1470 100.0 Incidence proportion (%) 143 5.4 45 4.3 31 10.3 3 25.0 79 5.8 36 3.8 27 8.3 1 5.9 64 5.0 Baseline to 4-year Remained normal 2559 85.8 1014 90.4 272 73.3 9 50.0 1295 85.7 938 89.9 310 76.5 16 72.7 1264 86.0 Normal to sarcopenic 216 7.2 56 5.0 48 12.9 5 27.8 109 7.2 60 5.8 44 10.9 3 13.6 107 7.3 Sarcopenic to normal 48 1.6 15 1.3 4 1.1 0 0.0 19 1.3 16 1.5 13 3.2 0 0.0 29 2.0 Remained sarcopenic 158 5.3 37 3.3 47 12.7 4 22.2 88 5.8 29 2.8 38 9.4 3 13.6 70 4.8 Total 2981 100.0 1122 100.0 371 100.0 18 100.0 1511 100.0 1043 100.0 405 100.0 22 100.0 1470 100.0 Incidence proportion (%) 216 7.8 56 5.2 48 15.0 5 35.7 109 7.8 60 6.0 44 12.4 3 15.8 107 7.8 Average annual incidence 3.1 2.0 5.1 12.1 2.9 2.7 5.2 4.3 3.3 over 4 years Individual cells might not sum up to total due to rounding. A total of 583 missing cases at 2-year follow-up were excluded among 4000 participants from baseline. Disease-free observations at the beginning of each follow-up period were used as denominator. A total of 436 missing cases at 4-year follow-up were excluded among 3417 participants who attended 2-year follow-up. A total of 583 missing cases at 2-year follow-up were excluded among 4000 participants from baseline for error adjustment, regardless of the subsequent disease status at 4-year follow-up. Another 436 missing cases at 4-year follow-up were excluded among 3417 participants who attended 2-year follow-up. The average annual incidence over 4 years was calculated by weighted average of the number of new sarcopenic cases per population initially at risk per 2 years from baseline to 2-year and 2- to 4-year follow-up. For example, average annual incidence over 4 years for all ages, both sexes = (3142 6.9%/2+2605 5.4% / 2) / (3142 + 2605) = 3.1%. 20 2014 Japan Geriatrics Society

Sarcopenia incidence and its risk factors Table 3 Binary logistic regression on incidence of sarcopenia during different periods of observation Baseline to 2-year 2- to 4-year Baseline to 4-year Total Incident cases OR (95% CI) Total Incident OR (95% CI) Total Incident OR (95% CI) Univariate Multivariate cases Univariate Multivariate cases Univariate Multivariate Demographics Age (years) 3142 217 1.097 (1.069 1.126) 1.095 (1.064 1.127) 2659 143 1.097 (1.062 1.134) 1.091 (1.053 1.130) 2775 216 1.109 (1.080 1.140) Sex Male 1596 91 Referent Referent 1367 79 Referent 1404 109 Referent Female 1546 126 1.467 (1.110 1.940) 1.577 (1.152 2.157) 1292 64 0.850 (0.606 1.192) 1371 107 1.006 (0.762 1.328) Socioeconomic status Education level Primary and below 2201 166 Referent 1843 96 Referent 1934 153 Referent Secondary and above 941 51 0.702 (0.508 0.971) 816 47 1.112 (0.777 1.593) 841 63 0.943 (0.695 1.279) SES ladder-community 1 3 176 8 Referent 143 8 Referent 145 10 Referent 4 5 814 47 1.287 (0.597 2.774) 6 10 2047 154 1.708 (0.825 3.538) 698 41 1.053 (0.483 2.297) 1737 87 0.890 (0.422 1.874) 717 51 1.034 (0.512 2.087) 1829 148 1.189 (0.612 2.309) SES ladder-hong Kong 1 3 855 63 Referent 715 35 Referent 747 57 Referent 4 5 1383 90 0.875 (0.627 1.222) 6 10 736 47 0.858 (0.580 1.268) 1183 62 1.075 (0.702 1.644) 632 38 1.243 (0.775 1.993) 1234 93 0.987 (0.700 1.390) 658 57 1.148 (0.783 1.684) Medical history COPD No 2906 193 Referent 2479 125 Referent Referent 2583 193 Referent Yes 236 24 1.591 (1.018 2.487) 180 18 2.092 (1.245 3.516) 1.838 (1.020 3.309) 192 23 1.685 (1.064 2.669) Diabetes No 2693 185 Referent 2292 123 Referent 2386 181 Referent Yes 449 32 1.040 (0.705 1.535) 367 20 1.016 (0.625 1.652) 389 35 1.204 (0.825 1.759) Hypertension No 1798 134 Referent 1511 79 Referent 1589 124 Referent Yes 1344 83 0.817 (0.616 1.085) 1148 64 1.070 (0.763 1.502) 1186 92 0.994 (0.750 1.316) Stroke No 3026 209 Referent 2563 131 Referent Referent 2676 202 Referent Referent Yes 116 8 0.998 (0.480 2.075) 96 12 2.652 (1.413 4.979) 3.142 (1.555 6.350) 99 14 2.017 (1.126 3.615) Cancer No 3013 204 Referent 2555 140 Referent 2663 209 Referent Yes 129 13 1.543 (0.855 2.785) 104 3 0.512 (0.160 1.636) 112 7 0.783 (0.360 1.704) 1.106 (1.073 1.139) 2.557 (1.321 4.952) 2014 Japan Geriatrics Society 21

RYuet al. Table 3 Continued Baseline to 2-year 2- to 4-year Baseline to 4-year Total Incident cases OR (95% CI) Total Incident OR (95% CI) Total Incident OR (95% CI) Univariate Multivariate cases Univariate Multivariate cases Univariate Multivariate Lifestyle factors Current smoker No 2940 197 Referent 2501 129 Referent 2608 195 Referent Yes 202 20 1.530 (0.943 2.482) PASE total score 3142 217 0.992 (0.989 0.996) Dietary intake Protein (g/day) 3138 217 0.996 (0.992 1.001) Vitamin D (IU/day) 3138 217 1.002 (0.997 1.008) Energy (1000 Kcal/day) 3138 217 0.784 (0.614 1.002) Cognitive function CSI-D level Normal to borderline (CSI-D score 28.4) Probable dementia (CSI-D score <28.4) 0.995 (0.991 0.999) 158 14 1.788 (1.004 3.182) 2659 143 0.992 (0.988 0.997) 2656 143 0.999 (0.994 1.004) 2656 143 1.003 (0.997 1.009) 2656 143 0.938 (0.702 1.252) 0.995 (0.990 0.999) 167 21 1.780 (1.101 2.877) 2775 216 0.992 (0.989 0.996) 2772 216 1.000 (0.996 1.004) 2772 216 1.003 (0.998 1.008) 2772 216 0.933 (0.735 1.185) 2720 175 Referent 2330 127 Referent 2423 189 Referent 422 42 1.607 (1.129 2.289) 329 16 0.887 (0.520 1.511) 352 27 0.982 (0.645 1.494) IADL impairments No. IADL impairments 0 2420 154 Referent 2085 94 Referent Referent 2174 145 Referent Referent 1 5 719 63 1.413 (1.041 1.918) Body measurements BMI (kg/m 2 ) 3142 217 0.669 (0.632 0.709) 0.646 (0.606 0.688) 571 49 1.988 (1.389 2.846) 2659 143 0.698 (0.652 0.747) 2.045 (1.351 3.097) 0.680 (0.631 0.733) 598 71 1.885 (1.397 2.544) 2775 216 0.691 (0.653 0.731) 0.995 (0.991 0.999) 2.122 (1.491 3.019) 0.656 (0.615 0.699) Extra number of missing observations (n) being removed from the regression model socioeconomic status ladder Community (n = 105), socioeconomic status ladder Hong Kong (n = 168), protein intake (g/day) (n = 4), vitamin D intake (IU/day; n = 4), energy intake (Kcal/day; n = 4), and number of instrumental activities of daily living (IADL) impairment (n = 3) in baseline to 2-year model; socioeconomic status ladder Community (n = 81), socioeconomic status ladder Hong Kong (n = 129), protein intake (g/day; n = 3), vitamin D intake (IU/day; n = 3), energy intake (Kcal/day; n = 3) and number of IADL impairments (n = 3) in 2- to 4-year model; and socioeconomic ladder Community (n = 84), socioeconomic ladder Hong Kong (n = 136), protein (g/day; n = 3), vitamin D (IU/day; n = 3), energy intake (Kcal/day; n = 3) and number of IADL impairments (n = 3) in baseline to 4-year model. The logistic regression model for baseline to 2-year observation was based on disease-free observations at baseline. Participants lost to 2-year follow-up were excluded. The 2- to 4-year model was based on disease-free observations at 2-year follow-up. Participants lost to 4-year follow-up were excluded. Baseline to 4-year model was based on disease-free observations at baseline. Participants lost to 2-year and 4-year follow-up were excluded. The backward stepwise (Wald) procedure was used in multivariate binary logistic regression with all variables entered at the beginning, except those variables with events per variable value less than 10. Socioeconomic status ladder Community, and disease status of stroke were excluded in baseline to 2-year model; socioeconomic status ladder Community, and disease status of cancer were excluded in 2- to 4-year model; and disease status of cancer was excluded in baseline to 4-year model accordingly. Odds ratios and confidence intervals were presented only for significant variables with P < 0.05. BMI, body mass index; COPD, Chronic Obstructive Pulmonary Disease; CSI-D, Community Screening Instrument of Dementia; PASE, Physical Activity Scale of the Elderly; SES, socioeconomic status. 22 2014 Japan Geriatrics Society

Sarcopenia incidence and its risk factors Table 4 Binary logistic regression on participants who had returned to non-sarcopenic during different periods of observation Baseline to 2-year 2- to 4-year Baseline to 4-year Total Revert cases OR (95% CI) Total Revert OR (95% CI) Total Revert OR (95% CI) Univariate Multivariate cases Univariate Multivariate cases Univariate Multivariate Demographics Age (years) 275 68 0.938 (0.892 0.986) 0.935 (0.886 0.986) 322 91 0.905 (0.861 0.951) 0.923 (0.875 0.975) 206 48 0.906 (0.851 0.966) Sex Male 146 34 Referent 144 26 Referent Referent 107 19 Referent Female 129 34 1.179 (0.681 2.040) 178 65 2.611 (1.548 4.404) 2.651 (1.501 4.681) 99 29 1.919 (0.994 3.706) Socioeconomic status Education level Primary and below 199 52 Referent 240 70 Referent 149 36 Referent Secondary and above 76 16 0.754 (0.399 1.424) 82 21 0.836 (0.473 1.476) 57 12 0.837 (0.400 1.753) SES ladder-community 1 3 28 4 Referent 20 3 Referent 18 3 Referent 4 5 77 15 1.452 (0.437 4.816) 6 10 159 44 2.296 (0.753 6.994) 77 20 1.988 (0.526 7.509) 215 63 2.349 (0.665 8.297) 58 11 1.170 (0.288 4.758) 123 32 1.758 (0.478 6.473) SES ladder-hong Kong 1 3 101 25 Referent 106 30 Referent 74 20 Referent 4 5 96 25 1.070 (0.563 2.034) 6 10 62 13 0.807 (0.377 1.725) 124 36 1.036 (0.584 1.839) 74 17 0.756 (0.380 1.502) 73 16 0.758 (0.356 1.613) 48 10 0.711 (0.299 1.688) Medical history COPD No 240 64 Referent 282 84 Referent 178 48 Referent Yes 35 4 0.355 (0.121 1.045) 40 7 0.500 (0.213 1.175) 28 0 Diabetes No 245 62 Referent 279 76 Referent 185 40 Referent Yes 30 6 0.738 (0.288 1.889) 43 15 1.431 (0.725 2.825) 21 8 2.231 (0.865 5.755) Hypertension No 176 40 Referent 208 61 Referent 130 28 Referent Yes 99 28 1.341 (0.765 2.352) 114 30 0.861 (0.515 1.437) 76 20 1.301 (0.672 2.517) Stroke No 257 64 Referent 306 87 Referent 193 45 Referent Yes 18 4 0.862 (0.274 2.712) 16 4 0.839 (0.263 2.673) 13 3 0.987 (0.260 3.741) Cancer No 258 67 Referent 302 86 Referent 194 47 Referent Yes 17 1 0.178 (0.023 1.369) 20 5 0.837 (0.295 2.375) 12 1 0.284 (0.036 2.261) 0.897 (0.839 0.959) 2014 Japan Geriatrics Society 23

RYuet al. Table 4 Continued Baseline to 2-year 2- to 4-year Baseline to 4-year Total Revert cases OR (95% CI) Total Revert OR (95% CI) Total Revert OR (95% CI) Univariate Multivariate cases Univariate Multivariate cases Univariate Multivariate Lifestyle factors Current smoker No 251 63 Referent 296 85 Referent 189 44 Referent Yes 24 5 0.785 (0.282 2.190) PASE total score 275 68 1.002 (0.994 1.009) Dietary intake Protein (g/day) 274 68 1.000 (0.992 1.008) Vitamin D (IU/day) 274 68 1.002 (0.986 1.019) Energy (1000Kcal/day) 274 68 0.963 (0.594 1.560) Cognitive function CSI-D level Normal to borderline (CSI-D score 28.4) Probable dementia (CSI-D score <28.4) 26 6 0.745 (0.289 1.919) 322 91 1.003 (0.996 1.010) 322 91 0.995 (0.988 1.002) 322 91 0.997 (0.986 1.008) 322 91 0.732 (0.470 1.140) 17 4 1.014 (0.315 3.268) 206 48 1.005 (0.996 1.014) 206 48 0.998 (0.988 1.007) 206 48 0.995 (0.976 1.014) 206 48 0.776 (0.430 1.400) 227 55 Referent 265 71 Referent 172 40 Referent 48 13 1.162 (0.574 2.351) 57 20 1.477 (0.804 2.713) 34 8 1.015 (0.426 2.418) IADL impairments No. IADL impairments 0 178 47 Referent 228 76 Referent Referent 139 38 Referent 1 5 97 21 0.770 (0.428 1.385) Body measurements BMI (kg/m 2 ) 275 68 1.145 (1.023 1.281) 1.155 (1.021 1.307) 94 15 0.380 (0.205 0.704) 322 91 1.105 (1.002 1.219) 0.354 (0.177 0.707) 67 10 0.466 (0.216 1.006) 206 48 1.119 (0.976 1.283) Extra number of missing observations (n) being removed from the regression model socioeconomic status ladder Community (n = 11), socioeconomic status ladder Hong Kong (n = 16), protein intake (g/day); n = 1), vitamin D intake (IU/day; n = 1), and energy intake (Kcal/day; n = 1) in baseline to 2-year model; socioeconomic ladder Community (n = 10), socioeconomic ladder Hong Kong (n = 18) in 2- to 4-year model; and socioeconomic status ladder Community (n = 7) and socioeconomic status ladder Hong Kong (n = 11) in baseline to 4-year model. Logistic regression model for baseline to 2-year observation was based on disease observations at baseline. Participants lost to 2-year follow-up were excluded. The 2- to 4-year model was based on disease observations at 2 years. Participants lost to 4-year follow-up were excluded. The baseline to 4-year model was based on disease observations at baseline. Participants lost to 2-year and 4-year follow-up were excluded. The backward stepwise (Wald) procedure was used in multivariate binary logistic regression with all variables entered at the beginning, except those variables with events per variable value less than 10. Socioeconomic status ladder Community, disease status of chronic obstructive pulmonary disease, diabetes, stroke, cancer, and smoking status were excluded in baseline to 2-year model; socioeconomic status ladder Community, disease status of chronic obstructive pulmonary disease, stroke, cancer, and smoking status were excluded in 2- to 4-year model; socioeconomic status ladder Community, disease status of chronic obstructive pulmonary disease, diabetes, stroke, cancer, smoking status and Community Screening Instrument of Dementia (CSI-D) level were excluded in baseline to 4-year model accordingly. Odds ratios and confidence intervals are presented only for significant variables with P < 0.05. BMI, body mass index; COPD, Chronic Obstructive Pulmonary Disease; IADL, instrumental activities of daily living; PASE, Physical Activity Scale of the Elderly, SES socioeconomic status. 24 2014 Japan Geriatrics Society

Sarcopenia incidence and its risk factors women is a greater public health concern. Nevertheless, we did not find a female preponderance after the 2-year follow-up in the present study. It is possibly a result of the natural bias, as those who are more sarcopenic or frail might default from follow-up, resulting in an underestimation of incidence. Higher physical activity levels, as measured with the PASE, were associated with lower sarcopenic risk, as noted in our previous study and others. 24 26 However, these studies were cross-sectional, and could not establish a causal relationship between physical activity and sarcopenia. Being longitudinal, our results of the protective effect of physical activity indicate a need for intervention. Past studies have shown that resistance strength training, such as weight lifting, has particularly strong beneficial effects on increasing muscle protein synthesis, muscle mass and strength in the elderly, including the oldest old, 27 possibly by evoking muscle hypertrophy along with neuromuscular adaptations. 28 However, strength training was usually intensive, which might not be practical to many untrained or sedentary older adults with various stages of functional decline. Aerobic exercise training could be an alternative in maintaining or increasing lean muscle mass. Such regimens have been shown to result in stimulating muscle protein synthesis, 29 and improving muscle fibers size and function. 30 Previously, we have also shown that heavy housework is associated with reduced mortality and cancer deaths over a 9-year period in the same study population; 31 further studies on the role of non-leisure time physical activity on sarcopenia, especially housework participation, are warranted. Although obesity is believed to be a risk factor for many adverse outcomes, in elderly populations, being slightly overweight might be beneficial. Previously, we have shown that older men were resistive to hazards of overweight and adiposity; and mild-grade overweight, obesity, and even central obesity could favor survival. 32 Similarly, a study in hospitalized elderly individuals found that fat mass was associated with a lower risk of death or complications. 33 The findings from the present study give further support to the inverse relationship between BMI and sarcopenia. Furthermore, BMI was positively associated with ASM (aged-adjusted Pearson s correlation coefficient [r] = 0.578, P < 0.001) and grip strength (age-adjusted r = 0.033, P < 0.05; data not shown), components of sarcopenia. This is in line with a previous study of Caucasian women, which found that participants who were overweight had a significantly reduced risk in developing sarcopenia when compared with their normal weight counterparts. 34 Despite possible favorable effects of BMI on muscle mass and strength, those with BMI <18.5 kg/m 2 or 25 kg/m 2 had slower walking speeds compared with their counterparts (walking speeds of 0.97, 1.03, 1.03, 0.99 m/s for BMI groups of <18.5, 18.5 to <23.0, 23.0 to <25.0 and 25 kg/m 2, respectively, data not shown), as has been reported in our previous study 35 as well as one other. 36 Other than muscle mass, BMI was also positively associated with fat mass (age-adjusted r = 0.843, P < 0.001, data not shown), which is thought to be an energy reserve in older adults that helps the individual survive illnesses and chronic conditions. 37 It has been pointed out that fat mass can have several age-rated effects on lean mass, whereas individuals with higher fat mass might have a higher protein intake, which is a protective factor against sarcopenia. 38 Given this, we postulated that high BMI might serve as a protective buffer in countering losses in muscle performance in the elderly population. Therefore, maintaining a healthy weight is important for older adults in order to preserve muscle mass and strength. A number of studies have shown that protein intake is a key factor for optimal muscle and bone health in older adults. 39,40 However, of greater practical importance, is the determination of the optimal quantity and quality of protein intake to preserve muscle mass and maintain physical functions in older adults. Although the current recommended dietary allowance of protein intake for male aged 50 years and older is 56 g protein/day, and for their female counterparts is 46 g/day, 41 a higher protein intake might be required for optimizing muscle health, particularly in older adults. 42 Insufficient or ineffectual protein intake in elderly individuals might facilitate the loss of muscle by blunting muscle protein synthesis and thus promoting net muscle protein catabolism. 43,44 However, the present findings showed no association between protein intake and incident sarcopenia. It is possible that associations are only apparent with a wide variation in protein intake in the study population, in that our participants could have a fairly high or adequate mean protein intake (76.5 g/day), 45 compared with other population-based studies. 46,47 Furthermore, 82% of our male participants and 76% of our female participants had protein intake at or above the recommended dietary allowance levels. Alternatively, the protein intake per mealtime of our participants might not be high enough to achieve significant protective effects; whereas an intake of 25 30 g at each mealtime could be beneficial in increasing protein synthesis and conserving muscle mass in older adults. 48 Nevertheless, data regarding protein intake per mealtime are not available in the present study, its association with sarcopenia remains to be explored. The role of vitamin D for sarcopenia in aging populations remains controversial. Although a number of studies have shown an independent association between low serum 25-hydroxyvitamin D (25OHD) and muscle mass or physical function; 49,50 others have found no association. 51 Previously, we have shown no association of serum 25OHD levels with baseline or 4-year change in muscle mass and physical performance measures 2014 Japan Geriatrics Society 25

RYuet al. in the same study population of older Chinese men. 52 Similar results were obtained in the present study by using dietary vitamin D intake, suggesting that vitamin D does not appear to be important in this elderly cohort. As noted, the prevalence of vitamin D deficiency was lower than in other published studies, which could explain the absence of such an association. 52 The present study also showed that stroke was an independent risk factor of incident sarcopenia. This observation is not unexpected, given the close association of stroke-associated disability with muscle atrophy and neuromuscular changes. 53 COPD, associated with inflammation and muscle wasting, 54 was also associated with a higher risk. IADL impairment, denoting a critical physical limitation and a level of dependency, also seems to play an important role in the development of sarcopenia. Several prospective studies have shown the relationship between sarcopenia and functional decline in older adults. 55,56 However, no significant associations were found between cognitive impairment and incident sarcopenia; although IADL impairments have been associated with cognitive decline. 57 Perhaps IADL impairments imply how poor the functional health really is, and therefore it was more predictive of sarcopenia than cognitive function. Despite cognitive function not being a significant predictor in the present study, those who were cognitively impaired were associated with a higher prevalence of IADL impairments (IADL score 3; 8.9%) compared with their counterparts (2.9%), χ 2 = 49.3, P < 0.001 (data not shown); therefore, it might be possible that the role of IADL impairments in the prediction of sarcopenia could be partly attributed to cognitive impairments, thus suggesting that improvement in cognitive function might improve muscle mass and delay progression to sarcopenia in the elderly population. Nevertheless, stroke, COPD, IADL impairments and cognitive function might not be easy to modify. Efforts should be made to increase participation in physical exercise for preserving muscle mass and prevention of sarcopenia. The present study also showed the reversibility of sarcopenia during follow-up, and that age was consistently the independent predictor across the different study periods. Perhaps older age denoted a higher risk of persistent sarcopenia. Those with IADL impairments at baseline were also less likely to return to nonsarcopenic during follow-up. On the contrary, sarcopenia might be partly reversible with increasing body weight. We also found that females were also more likely to return to non-sarcopenic compared with their male counterparts. The mechanism underlying such sex-related differences with aging remains to be elucidated. However, physical activity could play a role, and that women might be more health conscious than men. Among the 322 sarcopenic participants at 2-year followup, there was a significant increase in physical activity levels (baseline 80.02 to 4-year 92.94, P < 0.001) for women, whereas for men the levels tended to decline (baseline 89.59 to 4-year 85.29, P = 0.359; data not shown). Nevertheless, physical activity was not significantly associated with the reversibility of sarcopenia, although it was significantly associated with a lower risk of incident sarcopenia. This observation is not unexpected, given the small number of reverted cases during the follow-up years, which could decrease the power to detect an association. In addition, the difference between the characteristics of participants included in the analyses for the development of sarcopenia and its reversibility exists. For example, participants with sarcopenia were older, had a higher prevalence of COPD and stroke, and had lower mean values on their MMSE score, PASE score, and BMI than those without sarcopenia; therefore factors associated with incident sarcopenia might be different from its reversibility. The present study had some limitations. Our cohort was more educated and more physically active than the general elderly population in Hong Kong; therefore, findings should not be generalized to those who are institutionalized or frailer, or with lower education levels. Nutrient quantitation might not be exact. The use of a food frequency questionnaire rather than 24-h recall might have overestimated the intake. Serum 25OHD levels were accurate; however, data were only available in a subsample of men. Furthermore, those defaulting from follow-up were older, had more disabling diseases including sarcopenia and could have deteriorated more, which probably will result in an underestimation of incidence, particularly with longer duration of follow-up. In the present study, the proportion of missing cases differed between those with and without sarcopenia (2-year follow-up, 23.8% vs 13.7% and 4-year follow-up, 39.1% vs 20.4%); which could underestimate the true rate of muscle and functional loss in older adults over time if the rate was different between the two groups. For accurate data, we need to visit all participants at home if they do not come back. Finally, the number of participants who had returned to non-sarcopenic during follow-up was small (n = 68 at 2-year, and n = 48 at 4-year follow-up, respectively), thus findings warrant confirmation in larger studies with longer follow-up. In conclusion, the present study confirmed that sarcopenia incidence increased with age, but is potentially reversible, with several modifiable lifestyle-related factors as predictors. High BMI was protective against incident sarcopenia and its reversibility. Increasing physical activity and maintaining a healthy weight could be beneficial in the prevention of sarcopenia. Further studies with longer duration of follow-up are required to confirm these associations, and to examine other potential lifestyle behaviors that might contribute to sarcopenia and its reversibility. 26 2014 Japan Geriatrics Society

Sarcopenia incidence and its risk factors Acknowledgements We thank Dr Edith Lau who set up the elderly cohort, and the support from the SH Ho Centre for Gerontology and Geriatrics, Faculty of Medicine, The Chinese University of Hong Kong. Disclosure statement The authors declare no conflict of interest. References 1 Janssen I, Heymsfield SB, Ross R. Low relative skeletal muscle mass (sarcopenia) in older persons is associated with functional impairment and physical disability. JAm Geriatr Soc 2002; 50: 889 896. 2 Evans WJ, Campbell WW. Sarcopenia and age-related changes in body composition and functional capacity. J Nutr 1993; 123 (2 Suppl): 465 468. 3 Rosenberg IH. Sarcopenia: origins and clinical relevance. J Nutr 1997; 127 (5 Suppl): 990S 991S. 4 Baumgartner RN, Koehler KM, Gallagher D, Romero L. Epidemiology of sarcopenia among the elderly in New Mexico. Am J Epidemiol 1998; 147: 755 763. 5 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 2000; 89: 81 88. 6 Kim J, Wang Z, Heymsfield SB, Baumgartner RN, Gallagher D. Total-body skeletal muscle mass: estimation by a new dual-energy X-ray absorptiometry method. Am J Clin Nutr 2002; 76: 378 383. 7 Fielding RA, Vellas B, Evans WJ 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: 249 256. 8 Cruz-Jentoft AJ, Baeyens JP, Bauer JM et al. Sarcopenia: European consensus on definition and diagnosis: report of the European Working Group on Sarcopenia in Older People. Age Ageing 2010; 39: 412 423. 9 Manini TM, Clark BC. Dynapenia and aging: an update. J Gerontol A Biol Sci Med Sci 2012; 67 (1): 28 40. 10 Woo J, Leung J, Sham A, Kwok T. Defining sarcopenia in terms of risk of physical limitations: a 5-year follow-up study of 3,153 chinese men and women. J Am Geriatr Soc 2009; 57: 2224 2231. 11 Landi F, Liperoti R, Russo A et al. Sarcopenia as a risk factor for falls in elderly individuals: results from the ilsirente study. Clin Nutr 2012; 31: 652 658. 12 Gariballa S, Alessa A. Sarcopenia: prevalence and prognostic significance in hospitalized patients. Clin Nutr 2013; 32: 772 776. 13 Landi F, Cruz-Jentoft AJ, Liperoti R et al. Sarcopenia and mortality risk in frail older persons aged 80 years and older: results from ilsirente study. Age Ageing 2013; 42: 203 209. 14 Wong SY, Kwok T, Woo J et al. Bone mineral density and the risk of peripheral arterial disease in men and women: results from Mr. and Ms Os, Hong Kong. Osteoporos Int 2005; 16: 1933 1938. 15 Woo J, Lynn H, Leung J, Wong SY. Self-perceived social status and health in older Hong Kong Chinese women compared with men. Women Health 2008; 48: 209 234. 16 Washburn RA, Smith KW, Jette AM, Janney CA. The Physical Activity Scale for the Elderly (PASE): development and evaluation. J Clin Epidemiol 1993; 46: 153 162. 17 Woo J, Leung S, Ho S, Lam T, Janus ED. A food frequency questionnaire for use in the Chinese population in Hong Kong: description and examination of validity. Nurs Res 1997; 17: 1633 1641. 18 Prince M, Acosta D, Chiu H, Scazufca M, Varghese M. Dementia diagnosis in developing countries: a crosscultural validation study. Lancet 2003; 361: 909 917. 19 Chan TS, Lam LC, Chiu HF, Prince M. Validity and applicability of the Chinese version of community screening instrument for dementia. Dement Geriatr Cogn Disord 2003; 15: 10 18. 20 Lawton MP, Brody EM. Assessment of older people: selfmaintaining and instrumental activities of daily living. Gerontologist 1969; 9: 179 186. 21 Peduzzi P, Concato J, Kemper E, Holford TR, Feinstein AR. A simulation study of the number of events per variable in logistic regression analysis. J Clin Epidemiol 1996; 49: 1373 1379. 22 Hughes VA, Frontera WR, Roubenoff R, Evans WJ, Singh MA. Longitudinal changes in body composition in older men and women: role of body weight change and physical activity. Am J Clin Nutr 2002; 76: 473 481. 23 Gallagher D, Visser M, De Meersman RE et al. Appendicular skeletal muscle mass: effects of age, gender, and ethnicity. J Appl Physiol 1997; 83: 229 239. 24 Lee JS, Auyeung TW, Kwok T, Lau EM, Leung PC, Woo J. Associated factors and health impact of sarcopenia in older chinese men and women: a cross-sectional study. Gerontology 2007; 53: 404 410. 25 Kim SH, Kim TH, Hwang HJ. The relationship of physical activity (PA) and walking with sarcopenia in Korean males aged 60 years and older using the Fourth Korean National Health and Nutrition Examination Survey (KNHANES IV-2, 3), 2008 2009. Arch Gerontol Geriatr 2013; 56: 472 477. 26 Ryu M, Jo J, Lee Y, Chung YS, Kim KM, Baek WC. Association of physical activity with sarcopenia and sarcopenic obesity in community-dwelling older adults: the Fourth Korea National Health and Nutrition Examination Survey. Age Ageing 2013; 42: 734 740. 27 Liu CJ, Latham NK. Progressive resistance strength training for improving physical function in older adults. Cochrane Database Syst Rev 2009; (3)CD002759. 28 Aagaard P, Suetta C, Caserotti P, Magnusson SP, Kjaer M. Role of the nervous system in sarcopenia and muscle atrophy with aging: strength training as a countermeasure. Scand J Med Sci Sports 2010; 20: 49 64. 29 Sheffield-Moore M, Yeckel CW, Volpi E et al. Postexercise protein metabolism in older and younger men following moderate-intensity aerobic exercise. Am J Physiol Endocrinol Metab 2004; 287 (3): E513 E522. 30 Harber MP, Konopka AR, Douglass MD et al. Aerobic exercise training improves whole muscle and single myofiber size and function in older women. Am J Physiol Regul Integr Comp Physiol 2009; 297 (5): R1452 R1459. 31 Yu R, Leung J, Woo J. Housework reduces all-cause and cancer mortality in Chinese men. PLoS ONE 2013; 8 (5): e61529. doi: 10.1371/journal.pone.0061529 32 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 2014 Japan Geriatrics Society 27