Background: There are differences in the efficacy and safety profiles of medications used to treat IBD that may impact a woman's perceived risk of medication exposure during pregnancy, potentially leading to medication non-adherence, poor diseasecontrol and adverse pregnancy outcomes.
Aim:To assess whether medication adherence patterns differ by drug class during pregnancy and influence birth outcomes for women with IBD.Methods: Of 143 491 women, a validated case definition was used to identify 370 women with IBD in five administrative health databases in Alberta, Canada (2012Canada ( -2015. Women who had ≥2 consecutive medications prescription for maintenance therapy for IBD in the year prior to pregnancy were included (n = 230). Prescriptionbased medication possession ratio ≥0.8 defined adherence. Chi-squared tests were used to compare adherence patterns by drug class and outcomes.
Results:Of the 159/230 women who were adherent during the year prior to pregnancy, 20 (12.6%; 95% CI: 8.2%, 18.8%) were not adherent and 21 (13.2%; 95% CI: 8.7%, 19.5%) discontinued their medications during pregnancy. Medication adherence during pregnancy differed significantly by drug class. A greater proportion of women on biologics (41.5%; 95% CI 32.9%, 50.7%) were adherent during pregnancy than women on thiopurines (22.9%; 95% CI 16.1%, 31.5%; P = 0.006); adherence was not significantly different for 5-aminosalicylates (35.6%; 95% CI 27.4%, 44.8%; P = 0.204). Women were more likely to be adherent to biologics (49.3%, 95% CI 37.3%, 61.3%) than thiopurines (20.9%; 95% CI 12.6%, 32.6%; P = 0.014) and 5-aminosalicylates (29.9%; 95% CI 19.9%, 42.1%; P = 0.030) in the first trimester. This was similar in the third trimester. In the second trimester, adherence pattern did not significantly differ by drug class (biologics vs thiopurines: P = 0.348; biologics vs 5-aminosalicylate: P = 0.999). Infants born to women with IBD (adherent: RR 1.58. 95% CI 1.02, 2.27; non-adherent: RR 1.32, 95% CI 0.97, 1.81) were more likely to be admitted into the neonatal intensive care unit than the general obstetric population, but this was not significantly different between women who were adherent or not adherent to their IBD medication (P = 0.711).Conclusion: Almost a quarter of women with IBD who were previously adherent to medical therapy were not adherent during pregnancy. Adherence pattern differed by drug class.