Comparative effectiveness of ARB and ACEi for cardiovascular outcomes and risk of angioedema among different ethnic groups in England: an analysis in the UK Clinical Practice Research Datalink with emulation of a reference trial (ONTARGET)

Objective To study the comparative effectiveness of angiotensin receptor blockers (ARB) and angiotensin-converting enzyme inhibitors (ACEi) in ethnic minority groups in the UK. Design Observational cohort study using a reference trial emulation approach benchmarked against the ONTARGET trial. Setting UK Clinical Practice Research Datalink Aurum data from 01/01/2001-31/07/2019. Participants Black, South Asian, or White patients with a prescription for an ARB/ACEi who met the ONTARGET trial criteria. Main outcome measures The primary composite outcome was: cardiovascular-related death, myocardial infarction, stroke, or hospitalisation for heart failure with individual components studied as secondary outcomes. Angioedema was a safety endpoint. We assessed outcomes using a propensity-score-weighted Cox proportional hazards model for ARB vs ACEi with heterogeneity by ethnicity assessed on the relative and absolute scale. Results 17,593 Black, 30,805 South Asian, and 524,623 White patients were included. We benchmarked results against ONTARGET comparing ARB with ACEi for the primary outcome (hazard ratio [HR] 0.96, 95% CI: 0.95 to 0.98) and found no evidence of treatment effect heterogeneity( P int =0.422). Results were consistent for most secondary outcomes. However, for cardiovascular-related death, there was strong evidence of heterogeneity ( P int =0.002), with ARB associated with more events in Black individuals and with fewer events in White individuals compared to ACEi, and no differences in South Asian individuals. For angioedema, HR 0.56 (95% CI: 0.46 to 0.67) for ARB vs ACEi ( P int =0.306). Absolute risks were higher in Black individuals, for ARB vs ACEi number-needed-to-treat was 204 in Black individuals compared with 2000 in South Asian individuals and 1667 in White individuals ( P int =0.023). Conclusions These results demonstrate variation in drug effects of ACEi and ARB by ethnicity and suggest the potential for adverse consequences from current UK guideline recommendations for ARB in preference to ACEi for Black individuals.