A prediction model for the risk of chronic kidney disease occurrence

Introduction. Chronic kidney disease (CKD) is a health burden and an increasing public health issue. Prevalence is estimated to be 8–16% worldwide. The goal was to construct an easily-implemented model to predict CKD occurrence. We were especially interested in models that rely solely on information available to a clinical laboratory, enabling reporting the risk of CKD occurrence with clinical results. Methods. The study cohort included patients aged>18 years in the community between 2008 and 2013 with follow-up. CKD is defined as abnormalities of kidney function, present for >3 months. Individuals developed CKD defined by the glomerular filtration rate < 60 mL/min/1.73 m². All subjects were followed-up from the date of cohort entry until they developed CKD or until the end of 2013. Data analysis was made using SAS Version 9.3. Results. There were 779 men and 873 women at baseline in the cohort. At follow-up, 25% (n=408) had developed CKD. In CKD group (N=408), the mean age was 63.9 ± 12.74 years. Age, diabetes mellitus, hypertension, medical history of gout, cardiovascular disease and genitourinary disease were significantly associated with the occurrence of CKD (P<0.05). The most fitted model for CKD occurrence included age, waist circumstance, anemia, hyperuricemia, medical history of gout, hypertension, and genitourinary disease. (AUC = 0.683). Conclusion. We evaluated predictors of CKD occurrence among individuals in the community and constructed a clinical model to predict the incidence of CKD progression. This prediction tool which may help to identify subjects at risk of CKD is routinely obtained history and laboratory examinations. Improved clinical prediction is a cornerstone of individualized medicine.