Aim
The present study aimed to identify specific food groups that may promote or prevent the onset of ADL disability among elderly women in Japan. We examined the baseline consumption frequency of 10 food groups (fruits, colored vegetables, soy products, milk, seaweed, fats/oil, seafood, potatoes, eggs, and meat) and studied the relationship between the baseline food consumption and incidence of ADL disability four years later.
Study Design
This four-year longitudinal cohort study was conducted with a follow-up of community-dwelling elderly women without baseline ADL disability. Baseline and follow-up measurements were conducted in autumn 2008 and 2012, respectively. Independent variables included the consumption frequencies of the 10 food groups at baseline, while the dependent variable was the incidence of ADL disability over 4 years.
Participants
The longitudinal analysis targeted 1,111 community-dwelling elderly women without ADL disability at baseline and without any missing data at follow-up. The selection process commenced in 2008, when an invitation letter was sent to all women born between October 1, 1923, and November 30, 1933, living in the southeastern area of Itabashi Ward, Tokyo (N = 10,948). Of the 1,670 respondents with the intention to participate, 1,289 attended a baseline examination from October 15 to November 3, 2008. Of the 1,289 participants, 10 had baseline ADL disability and one lacked information on ADL, resulting in 1,278 participants without baseline ADL disability. Researchers conducted the follow-up survey in October 2012. Of the 1,278 participants, 50 did not participate in the 2012 surveys, resulting in 1,228 participants who were followed up with. Of those participants, 117 lacked information regarding ADL. Thus, complete pairs of information on ADL status both at baseline and follow-up were obtained for 1,111 participants. For these participants, follow-up time ranged from 1,449 to 1,476 days, (coefficient of variation: 0.004, median: 1,461 days); we believe that the variation was small enough not to require inclusion as a covariate. The Ethics Committee of Tokyo Metropolitan Institute of Gerontology (TMIG) approved the study protocol. All participants provided written informed consent regarding academic usage of their data.
Setting
Researchers conducted the baseline data collection through on-site interviews at TMIG and partly through mail/telephone surveys at follow-up. TMIG conducts annual physical/medical examinations for elderly people in the community, the present study used data from the 2008 (baseline) and 2012 (follow-up) examinations.
Measurements
At the baseline examination, researchers determined consumption frequency of the 10 food groups and ADL status through structured interviews conducted by trained research associates with no prior knowledge about each participant’s profile. At the follow-up examination, the researchers determined ADL status through interviews for the on-site participants and through mail/telephone surveys for the others. Researchers evaluated ADL status through five of the six questions (omitting “continence”) from the Katz Index [11]; inquiries into independence regarding five activities (i.e., walking, eating, bathing, dressing, and toileting) were conducted. The initial part of the printed questionnaire on ADL consisted of a main sentence: “Concerning activities of daily living, please choose one alternative for each of the following items.” This was followed by individual questions and alternatives:
Q1. Can you walk by yourself? Q2. Can you feed yourself? Q3. Can you take a bath by yourself? Q4. Can you change clothes yourself? Q5. Can you go to the bathroom by yourself and use it?
ADL disability was defined as the inability to implement at least one of the above five activities without assistance.
A food-frequency questionnaire, which was originally developed to measure the dietary variety of Japanese people, was used to measure the consumption frequency of each of the 10 food groups (i.e., seafood, meat, eggs, milk, soy products, colored vegetables, seaweed, potatoes, fruits, and fats/oils) [12]. The initial part of the printed questionnaire consisted of a main question: “How often do you consume each of the following food groups?” and a note to the interviewer: “Ask about the situation in the past week or so,” followed by the individual names of the food groups. The individual names of the food groups were presented with a supplementary explanation in parenthesis as follows: A. Seafood (raw or processed: all fish and shellfish); B. Meat (raw or processed: all meat); C. Eggs (chicken eggs and quail eggs. Fish eggs should be classified as “A. Seafood”); D. Milk (excluding coffee/fruit-flavored milk); E. Soy products (foods using soybeans, such as tofu and natto); F. Colored vegetables (dark vegetables such as carrots, spinach, pumpkins, and tomatoes); G. Seaweed (regardless of raw or dried); H. Potatoes, I. Fruits (regardless of fresh or canned. Tomatoes should be classified as “F. Colored vegetables”); and J. Fats/oils (count dishes that use oil, such as stir-fry, butter- or margarine-coated bread). The 10 specific food groups were selected by excluding the daily foods eaten by almost all elderly Japanese: rice, miso soup, pickled vegetables, bread, and noodles, from the original 15 food groups commonly consumed in Japan [13]. The individual names of the food groups were followed by four alternatives: “1. almost every day,” “2. once every two days,” “3. once or twice a week,” and “4. almost never.” Non-fasting blood samples were also collected during the baseline assessment. The specific assays included serum albumin (bromocresol green).
Statistical Analysis
The incidence of ADL disability, depending on the baseline consumption frequency of each food group, was examined through logistic regression analyses. Univariate and multivariate logistic regression analyses were conducted with the incidence of ADL disability as the dependent variable. In the multivariate analyses, using the forced entry method, researchers adjusted the odds ratios by variables that might potentially affect both dietary habits and ADL: baseline age; binary conversion of body mass index (BMI) (1 when 23 ≤ BMI ≤ 29, 2 otherwise [14]); regular exercise and sports (yes/no); current lifestyle-related diseases including diabetes, osteoporosis, stroke, and heart disease; and the intake frequency of the other food groups. For each food group, the “almost every day” category was set as the reference. Researchers conducted all statistical analyses using SPSS Statistics 17.0 (SPSS Inc., Chicago, IL).