Reply to Montastruc et al.

Abstract

To the Editor—We thank Drs Montastruc and de Canecaude for their comments [1]. Immortal time bias implicates a study design wherein participants cannot experience the outcome during some period of follow-up time. Montastruc et al posit that immortal time bias impacts our study and limits the significance of our results by generating an illusion of protective effect. The concern stems from a belief that both our exposure and control cohorts were composed of persons diagnosed with acne, seborrheic keratosis, or melanocytic nevi (A/Sk/Mn). If this were the study design, participants allocated to the exposure group would include an “immortal” time period between A/Sk/Mn diagnosis and the exposure (T&C administration), increasing the observation window for the unexposed group (which now includes the exposure group prior to exposure), and decreasing the cardiovascular event rate in the unexposed. This could lead to an observation of a higher cardiovascular event rate, and thus immortal time bias could be implicated in our results.