Surrogacy of composite endpoints for mortality in trials of atherosclerotic cardiovascular disease.

BACKGROUND

The surrogacy of modern composite endpoints for cardiovascular and all-cause mortality is unknown.

METHODS

We reviewed all primary publications of RCTs for non-acute atherosclerotic cardiovascular disease utilizing composite endpoints published 2019-2023 in first tier journals, and included all trials utilizing composite outcomes of ≥3 components. Trial-level surrogacy of outcomes was assessed using the R2 coefficient of determination, with a threshold for validating surrogacy at R2 = 0.8.

RESULTS

Forty-five (n = 45) RCTs met the inclusion criteria, of which 19 (42%) tested secondary prevention strategies. The median sample size was 6033 (Q1-Q3, 3684-11,287), median patient age was 64 (Q1-Q3, 62-67), and most participants were males (69%, Q1-Q3 55%-74%). Half (n = 22, 49%) of the RCTs were funded by industry sponsors and the remainder by government agencies. None of the components of the published composite endpoints reached the pre-specified threshold of 0.8 for surrogacy for all-cause mortality. The correlation coefficient for cardiovascular death was 0.54 (95% CI: 0.36-0.72); other correlation coefficients were notably lower. Similar lack of correlation was observed between cardiovascular death and MI, stroke, unstable angina, revascularization and HF.

CONCLUSIONS

Large ASCVD RCTs published in high-impact journals use composite endpoints components that do not correlate with patient prognosis. Journal editors, regulators and healthcare professionals should demand clinical trials provide evidence that is better aligned with patient outcomes to improve informed decision making for cardiovascular patients.