Stephen J.LurieMD, PhD, Senior Editor
To the Editor: Dr Glassman and colleagues1 found no statistically significant difference between sertraline and placebo for the trial's primary outcome, the number of patients showing a greater than 5% decrease in LVEF. However, no confidence intervals (CIs) are provided, and this prevents the reader from assessing the uncertainty involved in the results. The authors do report an absolute risk increase of 0.4%, for which I calculate a 95% CI of −4.4% to 5.3%. As a clinician, I would consider an absolute risk increase of 2% or more clinically important enough to consider, and this possibility is well within the CI.
Steinberg DI. Sertraline for Treatment of Depression in Acute Coronary Syndromes. JAMA. 2002;288(19):2403. doi:10.1001/jama.288.19.2403-JLT1120-1-2