We read with great interest the study by Romero-Gómez et al.1 demonstrating that the combination of metformin, peginterferon alfa-2, and ribavirin improved insulin resistance in >50% of patients and increased sustained virological response (SVR) rate in 10% of patients with hepatitis C genotype 1 and homeostasis model assessment (HOMA) >2. Intriguingly, in female participants, the addition of metformin to the standard of care for chronic HCV infection doubled the SVR rate.1
Since 1994, the U.S. National Institutes of Health requires that at least half of all clinical trial participants enrolled are females,2 and increasing interest in women's health and sex-specific outcomes have led to the increase in subgroup analyses stratified by sex. However, improperly conducted sex-based subgroup analysis in clinical trials can yield incorrect conclusions that may result in adverse effects on women's health. It has been therefore suggested that: (1) sex-based subgroup analysis should be planned a priori to the study commencement; (2) hypothesis or rationale for the analysis should be provided; (3) a statistical tests for interaction with sex should be performed when analyzing the outcomes; and (4) the overall treatment results should be emphasized more than the findings of the sex-based subgroup analysis.3 The study by Romero-Gómez et al.1 clearly had not planned the sex-based subgroup analysis a priori because the authors did not mention any adjustment by sex in the Methods section of the article. In addition, there is no clear hypothesis or rationale for a sex-based subgroup analysis in the Introduction or Methods section. The Methods and Results sections were searched for information regarding a statistical test for interaction between sex and the SVR, but no mention was found. In contrast, the authors properly placed equal emphasis on sex-based subgroup results as they did with the overall trial results. Given the frequency with which subgroup analyses by sex are now being performed, it is paramount that investigators should caution the scientific community in their interpretation. Discussion of the sex-specific effects of metformin in chronic HCV infection should be considered proper only when sex-specific analysis are viewed as hypothesis-generating, and further research to confirm these observations is recommended.