Differential Associations of Healthy Lifestyle Latent Classes with Glycemic Transitions in Prediabetes: An Observational Study Based on CHARLS

Abstract

Highlights What are the main findings? Over a four-year follow-up, among older adults with prediabetes in China, the patterns of glycemic transition were as follows: 1644 individuals (65%) remained in the prediabetic range, 322 (13%) progressed to diabetes, and 563 (22%) returned to normal glucose levels. Three distinct latent classes of healthy lifestyles were identified among older adults with prediabetes—smoking and drinking (27.6%), high-BMI (28.2%), and healthy lifestyle group (44.3%). These latent classes provide a reference for classifying health-risk profiles. Compared with the healthy lifestyle group, the odds ratios for reverting to normoglycemia were 0.736 and 0.628 for the smoking and drinking group and the high-BMI group, respectively, and the odds ratio (OR) for progression to diabetes in the high-BMI group was 2.17. What are the implications of the main findings? In the prediabetes population, healthy behaviour clustering patterns were identified, and three latent classes were distinguished. This finding suggests that health behaviours among individuals with prediabetes do not exist in isolation; they exhibit intrinsic clustering, and traditional single-dimension classifications or scoring methods may mask the population’s true phenotypic heterogeneity. The clustering approach used in this study provides directions for future research. Its stability and reproducibility should be validated in prospective cohorts; the drivers underlying the latent subtypes and their associations with long-term glycemic outcomes should be explored, thereby offering a methodological basis for stratified management research. Abstract Introduction: Current clinical guidelines recommend comprehensive management of multiple healthy lifestyles as fundamental to prediabetes care; however, the evidence base regarding the status and clustering of these risk factors remains insufficient. We identified distinct latent classes of healthy lifestyle in individuals with prediabetes and examined their impact on glycemic progression. Methods: This study included 2529 baseline prediabetic participants from the China Health and Retirement Longitudinal Study (CHARLS). We performed latent class analysis (LCA) to stratify participants according to healthy lifestyle factors: physical activity, smoking, alcohol consumption, sleep duration, body mass index (BMI), and blood pressure. Multiple imputation was used for missing covariate data. Multinomial logistic regression was used to examine associations with glycemic transition. Results: Three healthy lifestyle classes were identified: Class 1 (27.6%, the smoking and drinking group), Class 2 (28.2%, the high-BMI group), and Class 3 (44.3%, the healthy lifestyle group). Compared with Class 3, Class 1 (OR, 0.736; 95% CI, 0.554 to 0.978) and Class 2 (OR, 0.628; 95% CI, 0.489 to 0.806) had significantly decreased odds of returning to normoglycemia. Conversely, Class 2 (OR, 2.167; 95% CI, 1.626 to 2.887) had higher odds of progressing to diabetes than Class 3. Conclusions: Prediabetic individuals fall into three distinct healthy lifestyle latent classes. The distinct associations between the three lifestyle latent classes and glycemic progression to diabetes and regression to normoglycemia support the potential value of lifestyle-based risk stratification in prediabetes management, pending validation in prospective cohorts.