Response to Gorelick
To the Editor: Dr. Gorelick raises a reasonable question as to whether the sex-by-treatment interaction reported in our article on combined varenicline/bupropion sustained-release treatment could have been due in part to a lower efficacy of varenicline alone in men compared with women. Although we, along with the editorialist, highlight this possibility, two factors argue against it being a major contributor to the interaction effect. First, the previous literature has not reported sex differences in varenicline treatment. Second, the statistical argument advanced in Dr. Gorelick's comment is imprecise. Error-bar overlap is not a reliable criterion for assessing statistical significance. Using the information presented in Table 2 of our article, a chi-square calculation yields a p value of 0.27 for the difference between varenicline plus placebo treatment in men compared with women. Thus, although the possibility remains that there is an effect that contributed to the overall interaction effect, there is no compelling data in support of that interpretation at the present time.