A Prediction Model For Cardiovascular Events Based On A Low-sodium Salt Intervention Population

Objective: With long incubation period,CVD(cardiovascular diseases)takes a long time to develop.As a result,a proper strategy is necessary to forecast the development of cardiovascular(CV)events and give effective intervention to avert them.Therefore,in a sample of individuals who completed a 3-year low-sodium salt or conventional salt intervention and a 10-year follow-up in the hypertensive area,we constructed a CV event risk prediction model,hoping to offer a solid foundation for future customized interventions.Methods: Based on previous literature,blood pressure-associated genomic copy number variations(CNVs)were detected and analyzed,and then CNVs loci associated with CV risk were screened.The key evaluation indexes of CV events in the intervention population were selected by integrating environmental and genetic information,and a prediction model was constructed to evaluate the risk of CV events in this population.Data analysis was conducted for subjects meeting the inclusion criteria.A Cox proportional hazards model was used to build a prediction model.Both the discrimination and the calibration of the prediction model were assessed.The discrimination of the prediction model was measured using the area under curve(AUC) of receiver operator characteristic(ROC).Brier scores and calibration plots were used to assess the prediction model’s calibration.Decision curve analysis(DCA)was used for clinical applicability.The model was internally validated using the 10-fold crossvalidation method.The nomogram served as a tool for visualising the model.Results: Among the 306 total individuals,there were 100 cases and 206 control.Eighteen indexes that may affect the risk of CV events were collected and tested.In the model,there were six predictors including age,smoking,LDL-C(low-density lipoprotein cholesterol),baseline SBP(systolic blood pressure),CVD history,and CNV nsv483076.Among them,age,LDL-C,CVD history,and CNV nsv483076 independently affected the risk of CV events(P < 0.05).The fitted model has an AUC of 0.788,showing strong model discrimination,and a Brier score of 0.166,indicating that it was well-calibrated.The results of tenfold cross-validation showed that AUC=0.791 and Brier=0.166,indicating that the prediction model utilised in this study had a good level of repeatability.According to the model integrating the interaction of CNVs and baseline blood pressure,the effect of baseline SBP on CV events may be greater when nsv483076 was normal double copies than when nsv483076 was CNV.Conclusion: The efficacy of risk prediction models for CV events that include environmental and genetic components was excellent,and they may be utilised as risk assessment tools for CV events in specific groups to offer a foundation for tailored intervention strategies.