Citations 0
July 25, 2007

Antioxidant Supplements and Mortality—Reply

JAMA. 2007;298(4):400-403. doi:10.1001/jama.298.4.402

In Reply: Dr Albanes and Dr Huang and colleagues stress the potential importance of trials with 0 events. In our primary analyses, we calculated RRs and therefore excluded trials with 0 events (405 trials with 40 000 participants). Sensitivity analyses including 1 hypothetical trial with 1 death in each study arm and 40 000 participants reached the same conclusion as the primary analysis (RR, 1.02; 95% CI; 0.98-1.06). Trials with 0 events are likely to have specific characteristics (eg, short duration, low dose). We did not use imputation techniques because they are based on the assumption that any trial can be replaced by a new randomly chosen trial in the same sample.1

First Page Preview View Large
First page PDF preview
First page PDF preview