Reexamination of a meta-analysis of the effect of antioxidant supplementation on mortality and health in randomized trials

A recent meta-analysis of selected randomized clinical trials (RCTs), in which population groups of differing ages and health status were supplemented with various doses of β-carotene, vitamin A, and/or vitamin E, found that these interventions increased all-cause mortality. However, this meta-analysis did not consider the rationale of the constituent RCTs for antioxidant supplementation, none of which included mortality as a primary outcome. As the rationale for these trials was to test the hypothesis of a potential benefit of antioxidant supplementation, an alternative approach to a systematic evaluation of these RCTs would be to evaluate this outcome relative to the putative risk of greater total mortality. Thus, we examined these data based on the primary outcome of the 66 RCTs included in the meta-analysis via a decision analysis to identify whether the results provided a positive (i.e., benefit), null or negative (i.e., harm) outcome. Our evaluation indicated that of these RCTs, 24 had a positive outcome, 39 had a null outcome, and 3 had a negative outcome. We further categorized these interventions as primary (risk reduction in healthy populations) or secondary (slowing pathogenesis or preventing recurrent events and/or cause-specific mortality) prevention or therapeutic (treatment to improve quality of life, limit complications, and/or provide rehabilitation) studies, and determined positive outcomes in 8 of 20 primary prevention studies, 10 of 34 secondary prevention studies, and 6 out of 16 therapeutic studies. Seven of the eight RCTs with a positive outcome in primary prevention included participants in a population where malnutrition is frequently described. These results suggest that analyses of potential risks from antioxidant supplementation should be placed in the context of a benefit/risk ratio.