(158) Comparative Effectiveness of ARB and ACEi for Cardiovascular Outcomes and Risk of Angioedema by Ethnicity: an Analysis in the UK Clinical Practice Research Datalink with Emulation of a Reference Trial (ONTARGET)

Background: Cardiovascular disease (CVD) is a leading cause of death globally, and Black and South Asian individuals are at increased risk. Evidence for prescribing medications comes from randomised controlled trials, including the ONTARGET trial, which had limited inclusion of key higher-risk groups. Reference trial emulation can provide confidence in non-interventional data and methods which can then be applied to the analysis of trial-underrepresented groups.

Objectives: To investigate CVD treatment effect heterogeneity among ethnic minority groups in the UK.

Methods: Using data from the UK Clinical Practice Research Datalink (CPRD) Aurum linked with Hospital Episode Statistics, we applied ONTARGET trial eligibility criteria to Black, South Asian and White patients prescribed an angiotensin receptor blocker (ARB) or angiotensin-converting enzyme inhibitor (ACEi) between 1/1/2001-31/7/2019. Outcomes were compared between groups prescribed ARB vs ACEi using a propensity-score—weighted Cox proportional hazards model. Findings for the primary composite outcome of CV-related death, myocardial infarction, stroke, or hospitalisation for heart failure were benchmarked against ONTARGET results. Analyses were subsequently extended to explore treatment heterogeneity by ethnicity for the primary outcome, individual components of the primary outcome, and risk of angioedema.

Results: Among 573,021 trial-eligible patients over a maximum follow up of 5.5 years (median 5.2 years), the primary composite outcome occurred in 107,951 (18.8%) patients. We found HR 0.96 (95% CI: 0.95, 0.98) for the primary composite outcome for ARB vs ACEi, similar to ONTARGET, with no evidence of heterogeneity by ethnicity (P=0.422). However, there was strong evidence of heterogeneity by ethnicity for CV-related death on the relative and absolute scale (P=0.002 and P< 0.001). ARB (vs ACEi) use was associated with an increased risk of CV-related death in Black patients and with a lower risk in White patients. There was no difference in treatment effectiveness for South Asian patients.
Over a maximum follow up of 5.5 years, 814 (0.14%) angioedema events were recorded, HR 0.56 (95% CI: 0.46, 0.67) for ARB vs ACEi. There was evidence of heterogeneity when analysed as absolute risks (P=0.023), rather than relative risks, because of the increased baseline risk of angioedema in Black patients. Risk of angioedema was highest for users of ACEi. For ARB vs ACEi, among Black patients the number-needed-to-treat was 204 compared with 1667 among White patients. There was no difference observed for South Asian patients.

Conclusions: After emulating a landmark clinical trial and extending analysis we found evidence of differences in comparative drug effects of ARB and ACEi by ethnicity in the UK population.