Prevalence of diabetes and prediabetes among working-age adults and influencing factors of new-onset diabetes: a five-year cohort study (2018-2023).

Background: Diabetes and prediabetes in the young and middle-aged population represent a significant public health challenge in China. In recent years, the prevalence of diabetes has gradually increased within this group. This study aims to evaluate the prevalence of diabetes and prediabetes in a health check-up population of young and middle-aged individuals, and to analyze the key factors influencing the new onset of diabetes. The study provides data support for the early prevention of diabetes. Method: This study used retrospective cohort analyses to examine the data from the physical examination centers of three hospitals in Wuxi, China, for the population aged 18-59 from 2018 to 2023. Analyzing the changes in the prevalence of diabetes and prediabetes in the population. Single-factor analysis was used to examine differences in basic characteristics and laboratory indicators between individuals who developed diabetes and those who did not within five years. A multifactorial logistic regression model (MLR model), Cox proportional hazards model (Cox model), and generalized estimating equation (GEE) model were employed to analyze the factors associated with the development of diabetes. ROC curves were used to evaluate the performance of these three models. Finally, a nomogram was constructed to predict the risk of developing diabetes in the next five years. Results: From 2018 to 2023, the number of diabetes cases increased year by year, with the highest increase of 1.39% observed between 2020 and 2021. New-onset diabetes patients had poorer lifestyle and health profiles compared to those without new-onset diabetes. New-onset diabetes group also had worse metabolic and inflammatory profiles (P < 0.05), with significantly lower eGFR (P = 0.027). The AUC values for all three models were 0.64, with the GEE model performing best in Youden index (0.237), the Cox model in sensitivity (0.577), and the MLR model in specificity (0.776). The most significant factors identified were NLR, FBG, Cr, BMI, and exercise habits. The nomogram built using these five factors showed good predictive performance with AUC values of 0.705 and 0.666 in the training and test sets, respectively. Conclusion: The significant factors influencing the onset of diabetes include NLR, FPG, Cr, BMI, and exercise habits. The nomogram can effectively predict the risk of diabetes in the next five years, providing a powerful tool for early intervention. Future research could explore the interactions among these factors and validate the model's applicability in different populations.