SEARCH

SEARCH BY CITATION

We would like to thank Dr Auer and colleagues [1] for their thoughtful comments on our manuscript [2]. The authors raise an important issue in clinical trial research: the effect of selectively excluding participants from a particular analysis post-randomization. Specifically, they wondered whether our decision to exclude 59 nicotine gum users because they had been assigned to the ‘incorrect’ dose may have influenced our findings. Before addressing this question directly, it is worth revisiting why we made this exclusion in the first place.

Briefly, the purpose of our paper was to test whether the previously observed inverse relationship between the number of pieces of gum used during a quit attempt and weight gained was seen when gum use was dichotomized into compliant versus non-compliant use (based on product guidelines). To do this, we conducted a secondary analysis on a large gum study [3]. The original study from which the data were derived did not assign gum dose according to baseline number of cigarettes per day (CPD), as the indication for nicotine gum dictates; rather, participants were first characterized as having either low or high nicotine dependence, and then gum dose (placebo, 2 mg or 4 mg) was randomized within these two dependence groups, thus deliberately cross-assigning smokers to the ‘incorrect’ dose or the ‘correct’ one. As such, some participants—by design—were assigned randomly to an active gum dose that was contrary to current guidelines. It is these participants whom we excluded from our secondary analyses; we did so because we wanted our findings to be relevant to current clinical practice. Some additional subjects were excluded because they were missing the weight data crucial to the analysis.

Turning to Dr Auer and colleagues' concern: while their calculations are slightly inaccurate, their premise is correct: more participants were excluded from the 4- mg than the 2-mg groups (43 versus 16); that is, more participants were excluded because they used 4-mg gum even though they smoked fewer than 25 CPD than those who used 2-mg gum, even though they smoked ≥25 CPD. This is to be expected, given that under-dosed participants would have been more likely to fail to achieve abstinence in the first place, and our analysis focused on abstinent subjects.

Did this sway or nullify our results? To test this, we re-ran our analysis including these subjects. With these subjects included, the relationship between gum use and weight gain was essentially identical (and actually statistically more robust than the result we originally published on the dose-appropriate sample): among active gum users, compliant users gained less weight during quitting than non-compliant users (0.6 kg versus 1.6 kg; P = 0.001).

In summary, the exclusion of the participants who had been assigned to the incorrect nicotine gum dose did not change our findings. In any case, the decision to exclude them in the first place was justified scientifically and clinically.

Declarations of interest

  1. Top of page
  2. Declarations of interest
  3. References

GlaxoSmithKline Consumer Healthcare funded this re-analysis and provided financial support to Dr Ferguson, Dr Shiffman, Dr Rohay and Mr Gitchell for the preparation of the manuscript itself. However, the sponsor did not review the manuscript or this letter and had no direct role in the drafting of either document. The original study was funded by grants DA06183 and DA10073 from NIDA and by the Department of Veterans Affairs. Through their work at Pinney Associates, Dr Ferguson, Dr Shiffman, Dr Rohay and Mr Gitchell serve as consultants to GlaxoSmithKline Consumer Healthcare on an exclusive basis on matters relating to smoking cessation. Dr Shiffman and Mr Glitchell also have an interest in a venture to develop a new nicotine replacement medication.

References

  1. Top of page
  2. Declarations of interest
  3. References