Background: Among women with breast cancer treated with antidepressants, discontinuation is a common therapeutic option to prevent drug-drug interaction with tamoxifen. Guidelines recommend evaluating historical patterns of antidepressant use before considering discontinuation. Few studies have characterized antidepressant treatment trajectories before tamoxifen initiation, yet such knowledge would be clinically helpful in determining discontinuation strategies.
Objectives: Our goal was to identify antidepressant adherence trajectories before tamoxifen initiation among women with breast cancer and to assess characteristics predictive of trajectory group membership.
Methods: A retrospective cohort study included women with incident breast cancer who newly initiated tamoxifen, were 18-64 years-old, and had ≥2 antidepressant fills in the 12-month (baseline) before tamoxifen initiation (index date). We used IQVIA PharMetrics® Plus for Academics claims data from 2006-2020. Antidepressant adherence in the 12-month baseline was measured as monthly proportion of days covered (PDC) before tamoxifen initiation. Group-based trajectory modeling (GBTM) was conducted using polynomial functions of monthly PDC and compared 2-6-group finite mixture models to determine the model that best fit the data, as determined by the Bayesian Information Criterion, expert-adjudged parsimony, and ≥10% of the sample in each identified group. Multinomial logistic regression explored characteristics predictive of group membership.
Results: We observed 557 (20%) of 2,741 new tamoxifen users with prior antidepressant use; mean age 50 (SD 7) years. The final GBTM estimated 4 antidepressant adherence trajectory groups: consistently high [47% of cohort; mean PDC 0.81 (SD 0.1)], steady increase [21% of cohort; mean PDC 0.52 (SD 0.1)], declining [13% of cohort; mean PDC 0.33 (SD 0.1)] and recent start [19% of cohort; mean PDC 0.18 (SD 0.1)]. The recent start trajectory group had a lower mean age 47 (SD 7) years relative to the consistently high trajectory group, 51 (SD 7) years, p<.01. Prior venlafaxine use was associated with significantly higher likelihood (OR 2.4; 95% CI: 1.3-4.4] of being a recent starter but lower likelihood (OR 0.3; 95% CI: 0.2-0.8) of being in the declining group. Baseline depression was 3.5 (95% CI: 1.1-10.9) times more likely in the consistently high than the recent start group.
Conclusions: Women with breast cancer initiating tamoxifen have distinct and diverse baseline antidepressant use trajectories. Understanding these distinct histories is important as they may reflect survivors at various levels of medical and mental healthcare need during tamoxifen treatment.
