(048) Antidepressant Adherence Trajectories Before Tamoxifen Initiation and Risk of Subsequent Antidepressant Discontinuation Among US Women with Breast Cancer

Publication Number: 711

Background: For women with breast cancer taking antidepressants before initiating tamoxifen, concern remains about drug-drug interactions that may limit tamoxifen’s efficacy. Clinical guidelines recommend detailed assessment of women’s past history of antidepressant use before considering antidepressant discontinuation.

Objectives: Our goal was to explore the association between new tamoxifen users’ antidepressant adherence trajectories 12 months before tamoxifen initiation and the risk of antidepressant discontinuation in the 12 months after initiating tamoxifen.

Methods: Retrospective cohort of female, new tamoxifen users (index date) aged 18-64 years with incident breast cancer and antidepressant use in the 12 months before index date, IQVIA PharMetrics® Plus for Academics claims, 2006-2020. In a prior study, we stratified the cohort into four groups with distinct longitudinal patterns of antidepressant adherence (trajectories) pre-index date, using group-based trajectory modelling. The first trajectory group comprised 47% of the cohort with consistently high adherence while 21%, 13% and 19% exhibited steady increase, declining and recent start adherence, respectively. Discontinuation after tamoxifen initiation was defined as a 60-day gap in antidepressant days supply. A Cox proportional hazards regression model estimated the risk of antidepressant discontinuation by adherence trajectory group, adjusted for demographics, antidepressant class, prescriber specialty and mental health comorbidities. Women were censored at earliest of antidepressant discontinuation, loss of enrollment or 1-year post index date (i.e., end of study).

Results: There were 557 new tamoxifen users with prior antidepressant use; mean age 50 (SD 7) years. During follow up, 41% of both the consistently high and steady increase groups but 66% of the declining and 54% of the recent start groups discontinued antidepressants. Median times to discontinuation were shorter in the declining (5 months) and recent start groups (8 months), relative to the consistently high and steady increase groups (both 12 months),p <.01. Hazard ratios (95% CI) of antidepressant discontinuation were higher in the declining [2.2 (1.6 – 3.1)] and recent start [1.7 [1.2 – 2.4] groups, relative to the consistently high group.

Conclusions: Breast cancer survivors with past trajectories of consistently high antidepressant adherence experience lower risk of antidepressant discontinuation after initiating tamoxifen. However, relatively higher discontinuation among recent antidepressant initiators may suggest clinical practice differences in the management of breast cancer survivors with chronic, relative to new mood disorders (i.e., arising after cancer diagnosis).