נועם ברדה

Background: Cardiovascular risk prediction models use single blood pressure measurements. Cumulative blood pressure exposure over time may better predict cardiovascular events. Methods: We used data from a large community database to examine the incremental value of 10-year cumulative blood pressure exposure over standard risk based on a single value using the Pooled Cohort Equations (PCE). The index date was set to January 1, 2018, with a 5-year prediction horizon and 10 years of prior exposure. Individuals included were 45 years of age or older, had no known cardiovascular disease and had at least 3 prior blood pressure values. The primary exposure was cumulative blood pressure over the previous 10 years, expressed as the area under the systolic blood pressure curve over 140 mmHg. We assessed the model's ability to improve prediction of "Hard cardiovascular disease" (cardiac mortality, myocardial infarction or stroke), stroke and all cause mortality beyond that achieved by the PCE. Results: Our cohort of 614,084 individuals was divided into a derivation set of 428,826 and a validation set of 184,222 individuals. The C statistic improved by 0.8% (95% CI: 0, 0.8%), 1.3% (1.0, 1.7%), and 0% (0, 0.1%), and continuous net reclassification improvement was 0.492, 0.656 and 0.539 for the "Hard cardiovascular disease", stroke, and all-cause mortality outcomes, respectively. Conclusions: Cumulative blood pressure exposure over 10 years significantly improves the prediction of incident cardiovascular events and mortality above that derived from a single value. Introduction of cumulative blood pressure exposure into electronic health records might facilitate personalized risk prediction.