Is myocardial infarction a surrogate endpoint for all-cause mortality? A trial-level meta-analysis of 144 randomized controlled trials enrolling 1.2 million patients from 1972–2020. (14th October 2021)

• Record Type: Journal Article
• Authors: O'Fee, K; Deych, E; Ciani, O; Brown, D L
• Is Part Of: European heart journal. Volume 42(2021)Supplement 1
• Publication Date: 2021-10-14
• Subjects: Epidemiology, Prognosis, Outcome; Cardiology -- Periodicals; Heart -- Diseases -- Periodicals; 616.12005
• DOI: 10.1093/eurheartj/ehab724.1324
• Languages: English
• ISSNs: 0195-668X

Abstract

Abstract: Background: Nonfatal myocardial infarction (MI) is commonly included as an endpoint in studies of treatment or prevention of coronary artery disease driven by the assumption that MI is a surrogate for downstream mortality and that preventing MI will reduce mortality. Though biologically plausible and frequently associated in epidemiological studies, the validity of MI as a surrogate marker for all-cause mortality in randomized controlled trials (RCTs) has not been previously demonstrated with trial-level meta-analytic evidence. Purpose: To assess MI as a surrogate endpoint for all-cause mortality in RCTs Methods: In December 2020, PubMed was searched for all RCTs reporting all-cause mortality and MI published in the New England Journal of Medicine, the Lancet, or the Journal of the American Medical Association. RCTs with a minimum sample size of 1000 patients and 24 months of follow up were included. Trial-level correlation between MI and all-cause mortality was then assessed for surrogacy using the coefficient of determination (R2) between the natural logarithm of the odds ratios for MI and mortality using a weighted linear regression where each study was weighted by the number of observations. Criteria for surrogacy was set at 0.8. Prespecified subgroup analyses based on era of trial enrollment (before 2000, 2000–2009, 2010+), duration of follow up (2–3.9, 4–5.9, or 6+ years), and study subject (revascularization, primary prevention, secondary prevention, mixed primary/secondary prevention) were also assessed. Results: 1025 RCTs were retrieved and reviewed with 144 articles representing 1, 211, 897 patients ultimately meeting criteria for inclusion in the meta-analysis. Overall, MI had no correlation with all-cause mortality (R2=0.02, 95% CI: 0.00–0.08) (figure 1). In terms of era of trial enrollment, before year 2000 MI had low correlation with all-cause mortality (R2=0.22, 95% CI: 0.08–0.36) and had no correlation for the periods 2000–2009 (R2=0.02, 95% CI: 0.00–0.17) and 2010 and beyond (R2=0.01, 95% CI: 0.00–0.09). By follow-up period, MI had low correlation with all-cause mortality at 6+ years (R2=0.30, 95% CI: 0.01–0.55) and had no correlation with mortality at 2–3.9 years (R2=0.004, 95% CI: 0.00–0.08) or 4–5.9 years (R2=0.06, 95% CI: 0.001–0.16). MI had low correlation with all-cause mortality in revascularization trials (R2=0.21, 95% CI: 0.002–0.50) and no correlation with mortality in primary (R2=0.10, 95% CI: 0.001–0.26), secondary (R2=0.03, 95% CI: 0.00–0.20), and mixed primary/secondary prevention trials (R2=0.001, 95% CI: 0.00–0.08). Conclusions: MI cannot be validated as a surrogate endpoint for all-cause mortality in RCTs of treatments for or to prevent CAD. Thus, treatments that reduce MI cannot be assumed to reduce mortality. Inclusion of MI as an endpoint in RCTs may still be justified based upon its association with impaired quality of life and increased utilization of health care resources but not based on its surrogacy for mortality. FUNDunding Acknowledgement: Type of funding sources: None.