O9 Early pregnancy exposure to hydroxychloroquine and preterm delivery in systemic lupus erythematosus

Objective To assess the risk of preterm delivery associated with hydroxychloroquine (HCQ) exposure during early pregnancy among systemic lupus erythematosus (SLE) pregnancies. Methods A population-based cohort study was conducted on singleton pregnancies of individuals with prevalent SLE in Sweden from 2007 to 2021, identified from the Swedish Medical Birth Register and National Patient Register. We required ≥2 SLE ICD-coded visits before or on the last menstrual period date (LMP), with ≥1 visit with an SLE specialist. Exposure was defined as the mother having ≥2 HCQ dispensations from LMP-93 to LMP+93 days. The outcome was preterm delivery (delivery before gestational week 37+0), further classified by timing (very or moderate-late) and onset (spontaneous or medically-indicated). Inverse probability of treatment weighting was applied to adjust for confounders (e.g., smoking, maternal comorbidities, and corticosteroids). Modified Poisson models estimated risk ratios (RR) for preterm delivery. Multinomial logistic regression models estimated odds ratios (OR) for each preterm delivery subgroup vs. term delivery. Cox models were also used to analyze time to preterm delivery in gestational days and calculate hazard ratios (HR). Results We included 894 pregnancies (351 exposed and 543 unexposed to HCQ). Exposed pregnancies vs. unexposed had similar maternal age, lower smoking prevalence, and more severe disease (i.e., more corticosteroids and comorbidities). Overall, there were 126 preterm deliveries, with 55 (16%) and 71 (13%) in the exposed and unexposed groups, respectively (table 1). HCQ use was not significantly associated with preterm delivery overall (RR 1.09 (95%CI 0.75–1.57)) nor with time to preterm delivery (HR 1.10 (95%CI 0.76–1.58)) (table 2). Multinomial logistic regression models also showed non-significant associations with preterm delivery subgroups. The RR was smaller in the parous group vs. nulliparous (0.79 (95%CI 0.46–1.34) vs. 1.06 (95%CI 0.59–1.92)). In the parous group, HCQ was associated with 52% lower odds of medically-indicated preterm vs. term delivery (OR 0.48 (95%CI 0.27–0.85)). Sensitivity analyses restricting HCQ exposure to the first 22 weeks showed similar results. Conclusions In a large population-based register linkage of SLE pregnancies, HCQ exposure during early pregnancy was not associated with preterm delivery. Our findings are reassuring regarding the safety of HCQ during pregnancy. Acknowledgements This study is supported by the NIH (NIAMS R01 AR077103).