Aernoud Fiolet

166 Chapter 6 TO THE EDITOR (3) In the LoDoCo2 trial, Nidorf et al. report a lower incidence of cardiovascular events with colchicine than with placebo in chronic coronary disease. A subgroup effect suggests that the benefit applies to patients in Australia but may not apply to patients in the Netherlands. Although subgroup differences may be misleading, we have ways of judging the credibility of an observed effect. 1-3 The current analysis has several strengths, including a priori specification, the use of a relative outcome measure, a clinically important effect size, and low probability of observing a difference of at least this magnitude if false, which can be inferred from the nonoverlapping 95% confidence intervals. To quantify this last criterion, it would be useful for the authors to report the P value for the heterogeneity test they performed, according to the trial protocol. Other important questions cannot be answered on the basis of the data that are provided. Were similar subgroup effects seen in secondary outcomes? Did race, ethnic group, medication adherence, or mode of coronary-disease diagnosis differ between the countries? And would regression analysis show independence of the effect? On the basis of the currently available information, additional trials are warranted before the use of colchicine in patients with chronic coronary disease is implemented worldwide. Brett G. Fischer, M.D. Weill Cornell Medicine, NewYork, NY REFERENCES 1. Wang R, Lagakos SW, Ware JH, Hunter DJ, Drazen JM. Statistics in medicine — reporting of subgroup analyses in clinical trials. N Engl J Med 2007;357:2189-2194. 2. Sun X, Briel M, Walter SD, Guyatt GH. Is a subgroup effect believable? Updating criteria to evaluate the credibility of subgroup analyses. BMJ 2010;340:c117-c117. 3. Yusuf S, Wittes J. Interpreting geographic variations in results of randomized, controlled trials. N Engl J Med 2016;375:2263-2271.