A non-fatal MI increases the risk of death but an analysis of trials suggests that it is not a valid surrogate for all-cause and CV mortality.
A non-fatal myocardial infarction (MI) does not appear to represent a valid surrogate marker for all-cause and cardiovascular (CV) mortality according to an analysis by a researchers from the Department of Medicine, Washington University, US. Surrogate markers and their endpoints are commonly used in clinical trials but a potential problem is that there needs to be some certainty over whether changes in the surrogate are indicative of changes in the final outcome of interest. There are essentially three criteria which need to be met for a surrogate to be deemed valid: a biologically plausible relationship between the surrogate and the final outcome; a consistent association between surrogate and final outcome and that any treatment effect on the surrogate corresponds to a treatment effect on the final outcome. Although there is an abundance of evidence of the association between a non-fatal MI (NFMI) and both all-cause and CV mortality, there is little objective evidence to support the third criteria.
In trying to establish whether a non-fatal MI is a valid surrogate for both types of mortality in trials, the Washington team searched for all randomised studies that considered the treatment or prevention of coronary artery disease and reported a non-fatal MI as an outcome. The inclusion criteria was set as only randomised controlled trials with at least 1,000 patients and 24 months of follow-up. Trial level correlation between NFMI and all-cause and CV mortality was assessed for surrogacy using the coefficient of determination, R-squared, by plotting the individual trial outcome (in this case mortality) against NFMI. The R-squared value corresponds to the explained variation (or association) between these two measures and varies from 0 (no surrogacy) to 1 (perfect surrogacy). The threshold for validating NFMI as a surrogate for all-cause and CV mortality was set at 0.80.
Findings
A total of 144 randomised trials with 5,726,395 patients were included in analysis. In the pooled analysis of these trials, the value of R-squared for all-cause mortality was 0.02 (95% CI 0.0 – 0.08) and 0.11 (95% CI 0.02 – 0.27) for CV mortality. This lack of surrogacy was evident in trials of both primary, secondary and mixed primary/secondary prevention studies. Furthermore, the duration of follow-up had no impact on the association. For example, R-squared was not a surrogate for trials lasting from 2 to 3.9 years (R-squared = 0.004) or even trials with a follow-up of 6 or more years (R-squared = 0.06).
Commenting on their findings, the authors noted that while it is generally assumed that an intervention to reduce NFMI would also reduce all-cause and CV mortality, the results implied that any treatment or intervention designed to reduce a NFMI cannot be assumed to also reduce either all-cause or CV mortality. They also noted that there was a high degree of heterogeneity for the included trials and whilst this is generally considered to be a disadvantage, it actually becomes an advantage for the present analysis as it allows for subgroup analyses to determine if surrogacy exists.
They concluded that while NFMI is not a validated surrogate endpoint for mortality in trials, it’s inclusion may be justified due to the relationship between impaired quality of life and health-care related costs.
Citation
O’Fee K et al. Assessment of Nonfatal Myocardial Infarction as a Surrogate for All-Cause and Cardiovascular Mortality in Treatment or Prevention of Coronary Artery Disease. A Meta-analysis of Randomized Clinical Trials. JAMA Int Med 2021
