Quantification of Medication-mediated Effects on Pregnancy Outcome in Women with Rheumatic Conditions

Background: Rheumatic disease in pregnancy is associated with an increased risk of preterm birth (PTB) and small for gestational age (SGA) offspring. Many women with rheumatic conditions in pregnancy will require the use of disease-modifying antirheumatic drugs (DMARDs) or oral corticosteroids (OCS) for disease control over the course of gestation. It is unclear whether the disease process alone increases the risk of adverse birth outcomes, or whether the excess risk is mediated through the use of these medications.

Objectives: To quantify the medication mediated effects of DMARDs and OCS on the risk of preterm birth and small for gestational age among women with rheumatic conditions.

Methods: We used data collected from MotherToBaby Pregnancy Studies between 2005-2019. Eligible women had a singleton, live birth, and either had a rheumatic condition (rheumatoid arthritis, juvenile rheumatoid arthritis, ankylosing spondylitis, or psoriatic arthritis), or had no history of any autoimmune disorder. DMARDs (biologic and non-biologic) or OCS used between LMP and 37 weeks or delivery were assessed for risk of preterm birth and small for gestational age, respectively. We performed a causal mediation analysis to identify the natural direct and natural indirect effects, and the proportion mediated by each medication according to methods described by Tchetgen Tchetgen and Phiri. Log-linear regression models for the total effect and direct effect were adjusted for maternal age, education, race/ethnicity, pre-pregnancy BMI, asthma, and tobacco use. Models for the direct effect were weighted for women with rheumatic conditions by computing the predicted probability from a log-linear regression of medication use during pregnancy on risk factors (education, planned pregnancy, tobacco use) and a disease severity measure, the Health Assessment Questionnaire.

Results: From 2,182 women, 975 had a rheumatic condition. Preterm births occurred more frequently in women with rheumatic conditions (12.9% vs 7.0%; aRR 1.8, 95% CI 1.4, 2.4). DMARDs mediated 32% of the total risk, OCS mediated 62%. Indirect effect estimates for DMARDs included the null. Small for gestational age offspring was increased in women with rheumatic conditions (9.7% vs. 5.8%, aRR 1.8, 95% CI 1.3, 2.5); 6% of the total effect was mediated through DMARDs, and 13% through OCS, although indirect effect estimates crossed the null for both agents.

Conclusions: Mediation through DMARDs and OCS were only notable for the risk of preterm birth. The strongest mediation was observed through OCS, although residual confounding by disease severity may remain.