Letters Section Editor: Robert M. Golub,
MD, Senior Editor.
In Reply: Dr Barbui asks what the results of
our study would have been if we had chosen amitriptyline rather than dothiepin
as the comparator. Correcting for multiple comparisons and using amitriptyline
as the reference exposure, the odds ratios (confidence intervals) for nonfatal
suicidal behaviors among users of the other 3 antidepressants we studied are
dothiepin, 1.21 (0.80-1.83); fluoxetine, 1.40 (0.92-2.13); and paroxetine,
1.55 (0.99-2.43). These are estimated 99.17% confidence intervals, which we
believe is most appropriate in this post hoc analysis to maintain overall
type I error less than 5% while allowing for all 6 pairwise comparisons among
the 4 exposure groups. These results and their interpretation are consistent
with those reported in our study.
Jick H, Kaye JA, Jick SS. Antidepressants and the Risk of Suicidal Behaviors—Reply. JAMA. 2004;292(23):2833-2834. doi:10.1001/jama.292.23.2833-c