Abstract: Objective  To develop and validate a nomogram model based on a physical examination follow-up cohort to predict the risk of incident impaired fasting glucose (IFG) within 3 and 5 years among older adults with normal blood glucose at baseline. Methods  A retrospective cohort study was conducted using data from the health examination database of Ruici Medical Group (2010－2016). Older adults aged ≥60 years with normal baseline fasting plasma glucose (FPG) and complete data were included and randomly split into a training set and a validation set at a 7:3 ratio. Univariate screening and multivariable modeling were performed using the Cox proportional hazards model to identify independent predictors of IFG and construct a nomogram. Model discrimination and calibration were assessed using the concordance index (C-index), receiver operating characteristic (ROC) curves, and calibration plots. Results  A total of 21 914 participants were included (15 340 in the training set and 6 574 in the validation set). During follow-up, 1 411 incident IFG cases occurred in the training set and 602 in the validation set. Multivariable analysis showed that age, body mass index, baseline FPG, triglycerides, and alanine aminotransferase were independently associated with the risk of IFG, and a nomogram was developed accordingly. The model demonstrated good discrimination in both the training and validation sets (C-index: 0.767 and 0.763, respectively), and calibration plots indicated good agreement between predicted probabilities and observed incidence. Conclusion  The IFG risk nomogram constructed from routinely collected health examination indicators is simple and feasible, with good predictive performance, and can be used for early risk stratification and screening management in health check-up settings.

Key words: nomogram, impaired fasting glucose, risk factors, C-index, older adults, prediction model