Composite patient profiles predict blood pressure control more accurately than individual factors (p<0.001)Patient profiles predict blood pressure control better than single factors

Background Randomized outcome trials have shown the benefit of antihypertensive treatment, but subgroup analyses exploring treatment responses in relation to patient characteristics are usually based on single factors rather than comprehensive patient profiles. We hypothesized that a combination of individual predictors better identifies patients who are more or less likely to achieve blood pressure (BP) control. Methods To test our hypothesis, we reanalysed the data from the double-blinded VALUE-outcome trial (n=15,313). The patient population was divided into an exploratory cohort (2/3) and a validation cohort (1/3). We constructed composite patient profiles by combining four proven predictors of BP control: age, severity of hypertension, comorbidity and previous treatment status. Logistic regression and Cox proportional hazard models were used to test whether BP control and cardiovascular event rates differed among these profiles, and whether profiling conferred additional predictive information beyond the individual factors alone. Results BP control rates differed significantly between patient profiles, regardless of treatment (likelihood ratio test, p<0.001). Results from the exploratory cohort were reproducible in the validation cohort. Profiling predicted BP control significantly better than any individual factor (p<0.001 for all comparisons). In addition, there were significant associations between profiles and the incidence of cardiovascular events. Adverse events, however, were not related to profiles. Conclusions We conclude that composite patient profiles may predict BP responses to antihypertensive treatment more accurately than single factors. These findings may help to better individualize antihypertensive therapy.