CO-05: The first risk score for sudden death prediction in primary prevention patients

Background Sudden death defined as a death within 1h from the onset of symptom, is a common cardiovascular accident, even more frequent than fatal myocardial infarction. Up to now, DUKE is the only sudden death risk predictor and it was designed specifically for patients of high coronary risk. We constructed here the first sudden death risk score for primary prevention, developed in patients with hypertension and type 2 diabetes. Methods We used the Cox model to build this risk score on R (version 3.2.2) based on 30 560 individual data from seven randomized controlled trials (six of INDANA database in hypertension & one in type 2 diabetes Diahhycab). Results There was no treatment effect and no interaction between treatment and other covariates on the risk of sudden death. This allowed us to develop the model on both treatment/placebo groups. We identified seven risk factors of sudden death : age, sex (male), smoking, cholesterolemia, systolic blood pressure, baseline of type 2 diabetes and history of myocardial infarction. The discrimination performance of the tool was fair (area under the receiver operating characteristic curve (AUC) was about 70%). Conclusions Our work provides the first risk score for sudden death prediction in primary prevention patients. This risk score, in particular in hypertension and type 2 diabetes could help stratify patients in order to optimize preventive therapeutic strategies in primary prevention. Further research on sudden death is required to better prevent this highly frequent form of death.