(B-215) Assessing the Risk of Hormonal Contraceptive Failure in the Presence of Drug-Drug Interactions with CYP3A4 inducers

Background: We previously developed and validated an algorithm to measure exposure to HC in claims data.

Objectives: In this study, we assessed the risk of contraceptive failures (CF) of the concomitant use of oral HCs (OHC) or long-acting reversible contraceptives (LARC) in the presence of drug-drug interactions (DDI) with CYP3A4 inducers to complement the pharmacokinetic evidence.

Methods: We used MerativeTM MarketScan® Claims Databases (2005-2021) and included women aged 18-44 years without infertility diagnoses and with a 1-year continuous enrollment before the index date. The analyses were stratified by two indication Groups: (1) epilepsy, bipolar, personality disorder and (2) migraine. Contraceptive failure is measured as pregnancy onset during the use of contraceptives. We chose carbamazepine and oxcarbazepine as CYP3A4 inducers for Group 1 and topiramate for Group 2. We selected lamotrigine and levetiracetam as references for Group 1 and amitriptyline, gabapentin, and lamotrigine for Group 2 due to their CYP3A4-neutral effect. We established concomitant treatment episodes, accounting for early refills, switches, and discontinuations. Each DDI episode started on the overlapping day of contraceptive and enzyme inducer/neutral drugs till the earliest of: overlap end, conception, disenrollment, infertility diagnosis, or 1-year follow-up from the index. We calculated Pearl index (PI, number of conceptions per 100 women during the first year of use) for each group and incidence rate ratios (IRR) to compare the risk of CF between OHC or LARC users exposed to CY3A4 inducers or neutral drugs, adjusting for age, demographics, mental health disorders, sexually transmitted diseases, and other comorbidities through propensity score matching.

Results: 105,094 OHC and 24,975 LARC episodes in Group 1 and 137,357 OHC and 43,954 LARC episodes in Group 2 were identified. In Group 1, the adjusted PI (aPI) for OHC was 2.86 (95% confidence interval, 2.27-3.61) and 0.68 (0.35-1.30) for LARC in the inducer and 1.57 (1.18-2.10) for OHC and 0.55 (0.28-1.11) for LARC in the neutral group. In Group 2, the aPI for OHC was 1.32 (1.14-1.53) and 0.28 (0.15-0.50) for LARC in the inducer and 1.36 (1.18-1.58) for OHC and 0.28 (0.16-0.51) for LARC in the neutral group. The adjusted CF risk for OHC was significantly higher in Group 1(aIRR=1.82 [1.26-2.64]) and comparable in Group 2 (aIRR=1.01 [0.81-1.28]). For LARC, the CF risk was numerically higher in Group 1(aIRR=1.22 [0.47-3.16]) and similar in Group 2 (aIRR=0.97 [0.42-2.23]).

Conclusions: Using claims data confirms the pharmacokinetic DDI between strong CYP3A4 inducers and OHCs and increased CF rate as a result. For LARCs, we may need additional data sources (e.g., Medicaid) to further assess the DDI with CYP3A4 inducers.