A Novel Nomogram Based on a Competing Risk Model Predicting Cardiovascular Death Risk in Patients With Chronic Kidney Disease

ObjectiveChronic kidney disease (CKD) patients are more likely to die from cardiovascular disease (CVD) than develop renal failure. This study aimed to develop a new nomogram for predicting the risk of cardiovascular death in CKD patients. MethodsThis study enrolled 1656 CKD patients from NHANES 2003 to 2006 survey. Data sets from 2005 to 2006 survey population were used to build a nomogram for predicting the risk of cardiovascular death, and the nomogram was validated using data from 2003 to 2004 survey population. To identify the main determinants of cardiovascular death, we performed univariate analysis and backward-stepwise regression to select the key factors. The probability of cardiovascular death for each patient in 5, 7, and 9 years was calculated using a nomogram based on the predictors. To assess the nomogram's performance, the area under receiver operating characteristic curve (AUC) and the calibration curve with 1,000 bootstraps resamples were utilized. The prediction model's discrimination was examined using cumulative incidence function (CIF). ResultsAge, homocysteine, potassium levels, CKD stage, and anemia were included in the nomogram after screening risk factors using univariate analysis and backward-stepwise regression. Internal validation revealed that this nomogram possesses high discrimination and calibration (AUC values of 5-, 7-, and 9-years were 0.79, 0.81, and 0.81, respectively). External validation confirmed the same findings (AUC values of 5-, 7- and 9-years were 0.76, 0.73, and 0.73, respectively). According to CIF, the established nomogram effectively differentiates patients at a high risk of cardiovascular death from those at low risk. ConclusionThis work develops a novel nomogram that integrates age, homocysteine, potassium levels, CKD stage, and anemia and can be used to more easily predict cardiovascular death in CKD patients, highlighting its potential value in clinical application.