Introduction

Elevated blood pressure (BP) has been identified as the major modifiable risk for stroke development with an estimated population-attributable risk of 48%1. Thus, strict and aggressive BP lowering is deemed the most important prevention strategy for both primary and secondary stroke prevention2,3,4,5. A significant benefit of intensive BP lowering the risk of stroke was observed in the ACCORD-BP trial (Action to Control Cardiovascular Risk in Diabetes Blood Pressure Trial)6. On the contrary, this treatment benefit was not found in the SPRINT (Systolic Blood Pressure Intervention Trial) trial with a similar study design but a larger sample size in comparison to the ACCORD-BP trial7.

Although this observed difference could be attributed to the trial population or the study power, treatment heterogeneity on all-cause mortality from the two trials should be otherwise noted. We observed a significant reduction in death from any cause in the SPRINT trial, which contrasted with the ACCORD-BP trial6,7. Clearly, it is known that the occurrence of death would preclude the observation of a stroke. As such, further analyses are needed to account for the presence of this terminal competing risk. Meanwhile, many researchers have identified that individual patients vary from one another in many ways that can affect the potential for benefit, which cannot be captured by conventional one-variable-at-a-time subgroup analyses8,9,10. However, the heterogeneity of treatment effect (HTE) of intensive BP lowering based on individual stroke risk has not been explored.

As such, the aims of this study were (1) to estimate the treatment effect of intensive BP lowering on stroke prevention using the pooled patient data of SPRINT and ACCORD-BP after considering the presence of competing risks from death; (2) to further investigate whether stroke risk score was associated with differential treatment effect.