(206) Patient characteristics associated with Topical Calcineurin Inhibitor Discontinuation: A retrospective cohort study

Background: In 2005, the FDA and Health Canada released safety advisories related to Topical Calcineurin Inhibitors (TCIs) and the hypothetical risk of malignancy. Research that followed showed no clear evidence of the association. In 2018, the Canadian Dermatology Association released a positioning statement communicating the lack of evidence associated with the initial TCI warnings. These announcements and changes in evidence were associated with changes in treatment selection.

Objectives: To examine patient characteristics associated with TCI discontinuation following the safety announcements.

Methods: We conducted a retrospective cohort study of patients who were continuous users of TCI treatment within 1 year of each announcement (April 26th, 2004 – April 25th, 2005 and May 15th, 2017 – May 14th, 2019). The primary outcome was TCI discontinuation. We used multinomial logistic regression to estimate the association between patient characteristics (age, sex, geographic location, history of cancer, physician specialty, income quintile, number of unique prescriptions, physician visits, dermatologist visits and ED visits) and medication discontinuation (before announcement, after announcement, or no discontinuation).

Results: Among 1,423 TCI users between April 2004 to April 2006, a total of 811 users (57%) discontinued TCI treatment. Similarly, within a cohort of 6,228 TCI users between May 2017 and May 2019, 3,610 users (58%) discontinued treatment. Patients with fewer ED visits had significantly lower odds of TCI discontinuation following the 2005 safety announcement (OR: 0.68; 95% CI: 0.47, 0.97) than patients who visited the ED more frequently. Patients with fewer ED visits had significantly lower odds of TCI discontinuation before the 2018 safety announcement (OR: 0.82; 95% CI: 0.69, 0.99), however, this was not observed after the announcement (OR: 0.85, 95% CI:0.71-1.02). Of the remaining predictor variables, few patient characteristics were significantly associated with discontinuation before or after a respective safety announcement.

Conclusions: Having lower ED visits was associated with a lower likelihood of TCI discontinuation. Most patient characteristics were not meaningfully associated with the discontinuation of TCIs before or after the announcements. These findings suggest the discontinuation of topical medications should be further studied, specifically in relation to safety announcements or risk communications.