Increased Risk of Intrahepatic Cholestasis of Pregnancy in Women with Systemic Lupus Erthematosus Exposed to Azathioprine

Objectives Intrahepatic cholestasis of pregnancy (ICP) is linked to adverse maternal and fetal outcomes. Emerging data from IBD cohorts and a 2024 FDA safety report suggest a strong association between thiopurines and ICP.[1] This is concerning as azathioprine (AZA), a thiopurine, is the immunosuppressive of choice in SLE pregnancies. However, evidence in SLE is limited to a 2025 administrative study reporting a 3-fold increased ICP risk with AZA, without considering disease activity or thiopurine metabolites.[2] Thus, we performed a multicenter prospective cohort study to evaluate the risk of ICP in AZA-exposed vs unexposed SLE pregnancies. Methods The Lupus in prEGnAnCY (LEGACY) cohort is conducted at SLICC centers in Canada, South Korea, Peru, and Mexico. Pregnant women with SLE are enrolled before 17 weeks and followed in the second (20–24 weeks), third (30–34 weeks) trimesters, and postpartum (8-12 weeks). Since ICP occurs after 20 weeks, only pregnancies with a second-trimester visit were included. Follow-up began at that visit and continued until delivery. AZA exposure was modeled as time varying. The primary outcome was delivery for ICP and/or early-onset ICP (<28 weeks). Multivariable Cox proportional hazards models with frailties adjusted for maternal demographics, co-morbidities, disease activity, and glucocorticoid use. At the Montreal site, thiopurine metabolites and shunting were assessed, using established cut-offs.[3] Results Of 127 SLE pregnancies, 46 were AZA-exposed and 81 unexposed (Table 1). Ten ICP cases occurred (each in a distinct woman): 8 among AZA-exposed (17.4%, 95% CI 9.1-30.7) and 2 among unexposed (2.5%, 95% CI 0.7-8.6). AZA was continued until and beyond delivery in most (6/8) exposed ICP cases. All ICP cases required delivery except one AZA-exposed case (who stopped AZA at ICP diagnosis). AZA exposure was associated with a substantially increased risk of ICP (unadjusted HR 9.1, 95% CI 1.9-44.5; adjusted HR 11.9, 95% CI 2.2-65.1). Of note, all ICP cases with metabolite data (4/4) exhibited second-trimester shunting. Among all pregnancies with second-trimester metabolite data (n=22), 36.4% (95% CI 19.7-57.0) were shunting, and 50.0% (95% CI 21.5-78.5) of these developed ICP. No ICP occurred with tacrolimus alone (n=14). All ICP pregnancies resulted in live births, although AZA-exposed cases tended to have higher bile acid levels, earlier delivery, and lower birth weight for gestational age vs unexposed cases. Table 1. Characteristics of SLE pregnancies at the second trimester visit according to AZA exposure (n=127) Conclusion We observed that AZA exposure was strongly associated with ICP in SLE pregnancies. Second-trimester thiopurine shunting may identify women at highest risk, supporting the value of metabolite monitoring. References [1.] Joudaki S, Aliment Pharmacol Ther 2025;61:1430-6. [2.] Nguyen NV. Am J Gastroenterol 2025;121:1183-91. [3.] Lambert-Fliszar F, Lupus Sci Med 2021;8:e000519. Supported by a CIORA grant. Best Abstract by a Post-Graduate Research Trainee Award