Study participants
This study analyzed data from the SONIC study (Septuagenarians, Octogenarians, Nonagenarians, and Investigation with Centenaries), a longitudinal cohort survey of community-dwelling older people in Japan [13]. The study began in 2010 with a three-year follow-up survey of community-dwelling older people in four locations in Japan's Kansai and Kanto regions. The study recruited 2144 randomly selected participants in the baseline years of 2011, 2012, and 2013, involving 900 people aged 70-73, 972 people aged 80-81, and 272 people aged 90-91, respectively. Of these, 1341 were participants in the 3-year follow-up survey: 657 people aged 73-76, 610 people aged 83-84, and 74 people aged 92-94, respectively. This study excluded those receiving dietary guidance, those with missing weight measurements, and those with missing BDHQ (brief-type self-administered diet history questionnaire) [14]. Figure 1 shows a flow chart of the study participants. The SONIC study protocol was approved by the institutional review boards of Osaka University Graduate School of Medicine, Dentistry, and Human Sciences, and the Tokyo Metropolitan Institute of Gerontology (approval numbers: 266, H22-E9, 22 018, and 38, respectively). Informed consent was obtained from all participants.
Weight assessment
In this study, those who lost 5% of their bodyweight from the baseline weight to the 3-year follow-up weight were defined as weight losers, using anthropometric measurements at the survey. In a meta-analysis of weight loss and life expectancy in community-dwelling older people, most studies evaluated 5% weight loss and there was a significant correlation with death in subjects with 5% weight loss over several years [10]. In addition, weight loss of 5% or more in 6 months without dietary restriction or excessive exercise is defined as pathological weight loss [15]. Therefore, 5% weight loss was defined as such in this study. Weight was classified as 5% weight loss or maintenance.
Health status
This survey was conducted by a physician or nurse using a questionnaire that included physical factors, medical history, and prescribed medications. Blood pressure measurements, body measurements, and blood draws were done by a doctor or nurse [11,12]. BMI was calculated from weight and height measurements. Serum albumin, total protein, blood glucose, and HbA1C were from blood data. Since previous studies showed that serum albumin 3.8 g/dL below significantly increases the risk of death, serum albumin of 3.8 g/dL or lower was used in this study [16]. Hypertension was defined as a systolic blood pressure of 140 mmHg or higher, a diastolic blood pressure of 90 mmHg or higher, and the use of antihypertensive medication, according to the Japanese Society of Hypertension guidelines 2019 [17]. DM was defined by the Japanese Diabetes Society as fasting blood glucose of 126 mg/dL or higher, blood glucose of 200 mg/dL or higher at any time, hemoglobin A1C of 6.5 % or higher, and the use of diabetic medication [18]. Smoking and drinking histories were categorized into two: current smoking and no smoking or current drinking and no drinking. The grip strength was measured using a Smedley grip strength meter (Model YD-100; Yagami Ltd.., Tokyo, Japan), and the average of two measurements was used.
Dietary assessment
Dietary intake was assessed by the brief-type self-administered diet history questionnaire (BDHQ) to evaluate caloric intake and the energy ratio (% energy) of carbohydrate, protein, animal protein, vegetable protein, and fat [14]. We also classified it using the energy ratios in the Dietary Intake Standards for Japanese (2020 version) : carbohydrates were classified as less than 50%, 50-65%, and 65% or more, protein as less than 15%, 15-20%, and 20% or more, and lipids as less than 20%, 20-30%, and 30%or more [19].
Other factors
The survey included social factors such as household, economic status, and years of education. For household, the options of single, a married couple, and other were used. The economic status was based on household income, with the options of no financial comfort, normal, and financial comfort. Years of education were defined as 9 years or fewer, 10-12 years, and 13 years or more.
Cognitive function was assessed using Moca-J (The Japanese version of the Montreal Cognitive Assessment), which was developed as a screening test to detect mild cognitive impairment (MCI) [20]. MOCA-J was developed as a screening test to detect mild cognitive impairment (MCI). MCI is suspected if the score is below 25 points.
Statistical analysis
Descriptive statistics are summarized as the mean ± SD or median (% tiling) for continuous variables and percentages for categorical variables. The Kolmogorov-Smirnov test was used to assess the normality of continuous variables. We used the chi-square test for categorical variables, the t-test for continuous variables, and the Mann-Whitney U test for comparison between the two groups. Cochran-Armitage trend tests were conducted. Odds ratios and 95% confidence intervals for factors associated with 5% weight loss after 3 years were calculated using logistic regression analysis with gender, sex, BMI, household, cognitive function, grip strength, serum albumin level, and the carbohydrate energy ratio as explanatory variables. All data were analyzed using the statistical software SPSS Ver. 25 (IBM Japan, Tokyo, Japan). The significance level was set at less than 5%.