A Nomogram for Predicting the Risk of Chronic Kidney Disease among Chinese Adults

Abstract

Objective: By evaluating the associations between some factors and chronic kidney disease (CKD), a predictive nomogram is established to identify the high-risk population of CKD. Methods: A retrospective survey was conducted on the physical examination population in two hospitals in Shanghai and one hospital in Jiangsu Province from October 2022 to April 2023. Variables with P-values < 0.05 in the univariate analyses were selected as independent predictors for a multivariate logistic regression. The nomogram prediction model for the risk of CKD was established and validated using Stata15.0. Results: The prediction model in our study had the best explanatory power. In our prediction model, the influencing factors and hazard ratios were: age 1.85 (1.108–3.088), family history of hypertension 2.057 (1.248–3.392), family history of diabetes 2.675 (1.623–4.407), family history of CKD 4.142 (2.526–6.793), high-salt diet 3.814 (2.343–6.208), creatinine 1.996 (1.225–3.254), glucose 6.874 (4.129–11.443), triglyceride 4.104 (2.464–6.853), C-reactive protein 4.861 (2.817–8.387), and sodium 4.281 (2.617–7.003). The area under the curve values for the performance of the modeling and validation groups for the test nomogram were 0.9225 and 0.9466 and 0.9616, respectively. The Hosmer–Lemeshow test was: modeling set: (χ² = 2.237, P = 0.973); verification set (1): (χ² = 5.380, P = 0.716); verification set (2): (χ² = 6.752, P = 0.564). The nomogram calibration curves predicting the risk showed good consistency between the modeling and verification groups. Conclusion: A nomogram is established and verified internally and externally, which can help identify individuals with increased risk of CKD and provide a reference for the public and decision-makers to formulate primary prevention of CKD.