Study design and participants
Nanjing municipality, a typical megacity in eastern China, comprised 12 administrative districts and had approximately 9.5 million registered residents at the end of 2022, with 19.5% aged 60 years and above9. A large-scale cross-sectional survey—Healthy Aging, Healthy Elders 2023 (HAHE-2023 study)—was conducted in 2023 among urban and rural residents aged 60 years and above in Nanjing municipality. As a periodical surveillance program, the HAHE-study was developed to: (1) investigate the prevalence of selected chronic conditions (primarily including diabetes, HTN, EBW, abnormal lipid profiles, elevated blood glucose); (2) collect information on lifestyle and behavior (including dietary consumption, physical activity [PA], smoking, and drinking); and (3) assess health-related quality of life among local registered residents aged 60 years and above in Nanjing. The first HAHE study was conducted in mid-201810, and the second, HAHE-2023 study, was implemented in mid-2023.
Individuals were eligible to participate in the HAHE study if they met the following criteria: (1) registered as a resident in Nanjing; (2) aged 60 years or older; (3) had no cognitive, psychiatric, or literacy-related problems. The sample size for the HAHE-2023 study was primarily based on the estimation used in the HAHE-2018 study, with additional consideration of the expected response rate. In China, the emergency containment measures against the COVID-19 epidemic were terminated on January 8 of 2023, after which regular prevention strategies were implemented11. As the HAHE-2023 study was conducted in the first year following the end of COVID-19 emergency measures in Nanjing, a conservative response rate was assumed. Additionally, the information collected in the HAHE-2023 study was intended to serve as reference data for future surveillance of chronic diseases and lifestyle/behavioral patterns among older residents in the post-COVID-19 context. Therefore, the overall sample size for the HAHE-2023 study was projected to exceed that of HAHE-2018 (N = 21000) to ensure adequate statistical power and population representativeness. Ultimately, the sample size for the HAHE-2023 study was determined to be approximately 30,000.
Participants were randomly selected from across the 12 districts in Nanjing municipality using a multi-stage sampling strategy10. First, the sub-sample size was calculated for each district based on its proportion of older residents (aged 60 years and above) relative to the municipality. Second, the number of participating households was estimated for each district, assuming that two eligible participants would be available per household. Third, 15 administrative communities or villages were randomly chosen from each district. Then, eligible households were randomly selected from each participating community based on the calculated number of households for each district. Finally, all eligible older residents within each selected household were invited to participate in the HAHE-2023 study. Consequently, 35,071 participants were recruited, and 32,735 (93.3%) successfully completed the survey.
It is well-documented that individuals often seek clinical treatment and/or modify their lifestyle and behaviors to manage blood glucose levels following a diabetes diagnosis12,13. To ensure that blood glucose levels were not influenced by such diabetes-related clinical or behavioral interventions, it was scientifically necessary and appropriate to limit the analysis to individuals without a diagnosed diabetes condition. Therefore, all participants with diagnosed diabetes in the HAHE-2023 survey were excluded from the present study. This resulted in a final analytical sample of 26,769 non-diabetic individuals used to examine the combined associations of EBW, HTN, and HTG with hyperglycemia. Figure 1 presents the selection flowchart of participants included in the analysis.
The selection flowchart of participants in the study.
Written informed consent was obtained from all participants prior to the survey. The Ethics Committee of Nanjing Medical University Affiliated Nanjing Municipal Center for Disease Prevention and Control approved the original data collection protocol. All methods used in this study adhered to the principles outlined in the Declaration of Helsinki. For the present analysis, only de-identified secondary data were used. Therefore, the Ethics Committee of Geriatric Hospital of Nanjing Medical University granted a waiver of ethical approval.
Data collection
For all participants, socio-demographic characteristics, lifestyle and behaviors, history of selected chronic diseases, and family history of diabetes were self-reported through a face-to-face questionnaire survey, while body weight, height, BP, fasting plasma glucose (FPG), and TG were objectively measured10,14. Information on socio-demographic characteristics included age, gender, residence, educational level, marital status, and health insurance. Lifestyle and behavioral factors in this study included PA, smoking, drinking, and meat and fruit consumption. Personal history of diabetes and HTN, as well as family history of diabetes, were also self-reported by each participant.
Body weight was measured to the nearest 0.1 kg and height to the nearest 0.01 m for each participant. Each measurement was taken twice, and the average was used to calculate body mass index (BMI), defined as body weight (kg) divided by height squared (m2)14. BP was recorded using a calibrated sphygmomanometer based on Korotkoff sounds, following the recommended procedure for measuring BP among Chinese adults15. For each participant, BP was measured at least twice and the average reading was used in the analysis15. A 5-ml fasting venous blood sample was collected from each participant at the survey site and sent to a designated laboratory for analysis. FPG and TG levels were assessed using a HITACHI 7180 analyzer (Hitachi Co., Japan), with detection kits provided by Shanghai Fosun Long March Medical Science Co., China.
Study variables
Outcome variable
The outcome measure was FPG status. In this study, a cutoff value of 6.1 mmol/L was used to categorize participants into either “normoglycemia (FPG < 6.1 mmol/L)” or “hyperglycemia (FPG ≥ 6.1 mmol/L)” for the analysis1,16. This FPG cutoff was recommended by the WHO for defining hyperglycemia1. Moreover, it has been officially adopted to identify hyperglycemia among older adults aged 60 years and above in China16.
Independent variables
There were three independent variables in the study: body weight, BP, and TG status. Body weight status was classified as “non-EBW (BMI < 24kg/m2),” “overweight (BMI: 24–27 kg/m2),” or “obesity (BMI ≥ 28 kg/m2)” with EBW referring to either overweight or obesity, according to BMI cutoffs recommended for Chinese adults14. Participants were categorized as having “HTN”, if they had previously been diagnosed with HTN by a physician or if their systolic/diastolic BP was ≥ 140/90 mmHg assessed during the survey15. Otherwise, participants were classified as having “normotension”. TG status was classified as “normal TG (< 1.7 mmol/L)” or “HTG (≥ 1.7 mmol/L)” according to the cutoff recommended for identifying elevated TG in Chinese populations17.
In addition to analyzing the separate associations of EBW, HTN, and HTG with hyperglycemia, participants were categorized into sub-groups based on two- and three-factor combinations to examine the joint associations with hyperglycemia. For the two-factor combinations: (1) Participants were grouped as non-EBW + normal BP (reference), non-EBW + HTN, overweight + normal BP, overweight + HTN, obesity + normal BP, or obesity + HTN (highest risk) to assess the combined association of EBW and HTN with hyperglycemia; (2) Participants were categorized as non-EBW + normal TG (reference), non-EBW + HTG, overweight + normal TG, overweight + HTG, obesity + normal TG, or obesity + HTG (highest risk) to evaluate the combined association of EBW and HTG with hyperglycemia; and (3) Participants were classified as normal BP + normal TG (reference), normal BP + HTG, HTN + normal TG, or HTN + HTG (highest risk) to assess the combined association of HTN and HTG with hyperglycemia.
Furthermore, for the three-factor combinations, participants were grouped as follows: non-EBW + normal BP + normal TG (reference), non-EBW + normal BP + HTG, non-EBW + HTN + normal TG, non-EBW + HTN + HTG, overweight + normal BP + normal TG, overweight + normal BP + HTG, overweight + HTN + normal TG, overweight + HTN + HTG, obesity + normal BP + normal TG, obesity + normal BP + HTG, obesity + HTN + normal TG, or obesity + HTN + HTG (highest risk) to investigate the three-factor combined associations with hyperglycemia.
Covariates
Potential influencing factors associated with hyperglycemia or diabetes—including socio-demographic attributes, lifestyle and behaviors, and family history of diabetes—were treated as covariates in the multivariate regression analysis. The socio-demographic factors adjusted for in analysis included age (60–69, 70–79, or 80 + years), sex (men or women), area of residence (urban or rural), educational level (≤ 6, 7–12, or 13 + years of schooling), marital status (single or with a spouse/partner), and type of health insurance (resident basic medical insurance [RBMI] or employee basic medical insurance [EBMI]).
Definitions of family history of diabetes, smoking, and drinking were adopted from those used in the official survey of Chinese chronic non-communicable disease and risk factor surveillance14. Participants were classified as having a positive family history of diabetes (“Yes”) if one of their parents had been diagnosed with diabetes. Otherwise, the participants were categorized as having a negative family history of diabetes (“No”)14. In addition, participants were classified as either “smoker” or “non-smoker”, and “drinker” or “non-drinker” in the analysis14.
PA level was measured using the validated Chinese version of the International Physical Activity Questionnaire (IPAQ-CHN)18,19. Participants were categorized as having “sufficient PA (≥ 150 min/week)” or “insufficient PA (< 150 min/week)” based on the total time spent in moderate PA plus twice the time spent in vigorous PA during the past 7 days20. Meat and fruit intake were assessed using a validated Chinese version of the Food Frequency Questionnaire (FFQ)21, which collected data on consumption frequency in the previous week. Based on weekly intake frequency recommendations for older Chinese adults from the Chinese Nutrition Society22, participants were classified as either “reached recommendation (Yes)” or “did not reach recommendation (No)” for meat and fruit intake, respectively. Participants with HTN or HTG were asked to report whether they were taking medications to manage these conditions. Accordingly, they were categorized as either “taking medications” or “not taking medications” in the analysis.
Data analysis
First, differences in selected participant characteristics (%) by sex and blood glucose status were examined separately using the Chi-square test. Then, mixed-effects logistic regression models were applied to calculate odds ratios (ORs) and 95% confidence intervals (CIs) for evaluating the separate and combined associations of EBW, HTN, and HTG with hyperglycemia among non-diabetic older adults. Model 1 was a univariate analysis, where body weight, BP, TG, or their combined measures served as the single independent variable, and the survey community was included as a random effect. Model 2 was a multivariate analysis, adjusting for age, sex (overall participants only), urban/rural residence, educational attainment, marital status, health insurance, PA, smoking, drinking, meat and fruit consumption, family history of diabetes, medication use for managing HTN or HTG, and—where applicable—body weight status, BP status, and TG status, in addition to the variable(s) considered in Model 1. The significance level was set at p < 0.05 (two-sided). EpiData version 3.1 (The EpiData Association, 2008, Odense, Denmark) and SPSS version 25.0 for Windows (SPSS Inc., Chicago, IL, USA) were used for data entry and analysis, respectively.