Study Design and Study Population
This cross-sectional study was conducted in Fujian Province, China, as the preliminary phase of the project “Prospective cohort study of frailty in elderly people of Fujian Province”, which aims to explore the influence of ageing and frailty in the elderly for clinical decision making in frailty risk assessment. From July to December 2021, the eligible elderly population in Fuzhou Community Health Service Center of Fujian Province was recruited by telephone calls and posters. Inclusion criteria were men and women over the age of 60, informed consent and volunteered to participate in the study and ability to complete scale evaluation and physical examination. Exclusion criteria were life expectancy < 6 months because of critical disease or advanced tumour; long-time bedridden, completely disabled; severe visual, hearing or speech impairment. The study was in accordance with the 1975 Declaration of Helsinki and approved by the ethics committee of FuJian Provincial Hospital.
Sample
Meta-analysis suggests that the prevalence of frailty in the non-hospitalised elderly population aged 60 and above in China is 12% [7]. The sample size of the cross-sectional study was calculated by
where Zσ is the significance test statistic, α=0.05, Zσ= 1.96, p is the estimated frailty incidence rate of 12%, q=1-p; d is the allowable error, in this study 0.02; the minimum sample size calculated is 1063. Considering a projected 20% sample loss because of questionnaire quality, the minimum sample size required was 1276. In this study, among the 1508 included subjects, 50 subjects who did not provide information on all the study variables were excluded, and a total of 1458 subjects were included in the final analysis. The final number of people included meets the requirements of minimum sample size.
Measurements
PA assessment
The short from of the International Physical Activity Questionnaire (IPAQ), which has been validated in China, was used to assess physical activity (PA) level [21].
The IPAQ-SF consists of seven items and provides information on the time spent in vigorous-intensity activity (eg, jogging, swimming, running ), moderate-intensity activity (eg, dancing, riding a bike, cleaning house) and walking. The IPAQ-SF required the subjects to recall the number of days they performed each activity (frequency) and the length of time (duration) they were involved in each daily activity in the last 7 days. The formula of IPAQ was as follows: the total physical activity (MET/min/w)=the MET (metabolic equivalents) value of physical activity × the amount of time spent on physical activity per day (min/d) × the number of days of physical activity per week (d/w). MET values for vigorous-intensity activity, moderate-intensity activity, and walking were 8, 4, and 3.3, respectively. We converted the continuous variables corresponding to the total physical activity into three categorical variables, which uses cut-off values of 600 and 3000 MET min/w as follows: low total physical activity (< 600 MET/min/w), moderate total physical activity (600-3000 MET/min/w) and high total physical activity (≥ 3000 MET/min/w) [22].
Sedentary behaviour assessment
The researchers assessed ST by asking “How many h in a 24-h day do you typically spend sitting”? This includes working at a desk or computer, visiting friends, riding in a car, reading, playing cards or watching TV but does not include sleeping time. The average amount of time spent sitting per day over the past 7 days fell into four categories, 4 h/d, 4~6 h/d, 6~8 h/d and ≥8 h/d, similar to the classification used in recent studies [23].
Frailty measure
The frailty index (FI), which is based on the theory of health defects, was used to measure the degree of frailty [24]. The FI refers to the proportion of potential unhealthy measurement indicators of an individual among all measurement indicators at a certain time point. The more defects a person has, the more likely he or she is to be in a frail state. In the present study, the FI consisted of 40 variables, including multi-dimensional indicators such as medical signs, medical diagnosis, activities of daily living and performance tests (walking speed and grip strength) [25]. According to previous research, FI 0.2 was defined as the threshold for entering the frailty state, and individuals were divided into non-frailty (< 0.2) and frailty (0.2-1.0) groups [26].
Covariates
Baseline data were collected by trained researchers through face-to-face interviews using standardized questionnaires. Main contents includes general demographic information (age and gender), socioeconomic attributes (marital status, living status, education level, now or before retirement occupation, average monthly income, method for medical payments), lifestyle (smoking status, alcohol consumption, etc.) and history of disease and medication. The education level was divided into primary school and below, middle school, high school, college and master's degree and above; Marital status is classified as married, widowed, divorced or other. The average monthly income was divided into<3000 RMB, 3000-6000 RMB, 6000-10000 RMB and>10000 RMB. Occupations were classified as civil servants or professional technicians in state units, workers, commercial, service or freelance workers, manual or unemployed. Living status was divided into living with family, living with others and living alone. Smoking and drinking status were divided into current, former and never groups. Weight and height, waist circumference, blood pressure and BMI were measured and calculated using standard methods.
Statistical analysis
SAS 9.4 (Cary, NC) was used to analyse the data, and the measurement data were expressed as mean ± standard deviation, and the t-test was used to compare the two groups. The X2 test was used to compare the two groups of categorical data. If the theoretical frequency was too small, Fisher's exact probability method was used.
A multiple linear regression model was used to analyse the relationship between PA or ST and FI, expressed as β values of 95% confidence intervals (CI), with light physical activity and minimum sitting time (< 4 h/day) as reference categories, respectively. Multivariate adjusted logistic regression models were also used to assess the association between PA or ST and the prevalence of frailty, with results expressed as odds ratios (OR) with corresponding 95% CI. Two models were adopted to assess association of PA, ST and frailty. Model 1 was adjusted according to PA and ST levels; Model 2 adjusted for age, gender, education level, marital status, average monthly income, smoking status, drinking status, BMI, ST and PA.
A cross-product term was added to the logistic regression model to evaluate the statistical significance of the interaction between PA and ST on frailty. A restricted cubic spline regression was used to investigate the dose-response relationship between continuous PA-MET-h/day or ST (h/day) and frailty.
We conducted joint analysis of sitting time, physical activity and frailty, comparing groups with different amounts of sitting time and physical activity with the combined vigorous PA and lowest ST (< 4 h/day) groups serving as the reference group.
A generalized linear model was used to visualize the interaction of ST (h/day) and PA (MET-h/day) on frailty. In the interaction diagram, the effect of MET-h/day is estimated with 95% CI as a function of the increase in ST (h/day).