Perinatal exposure to conventional synthetic disease-modifying anti-rheumatic drugs in women with rheumatic disease and neonatal outcomes: a population-based study.

ABSTRACT

OBJECTIVES: Epidemiologic studies evaluating associations between specific arthritis medications and perinatal outcomes are limited. We evaluated the association between conventional synthetic DMARD (csDMARD) use among women with rheumatic disease (RD) and neonatal outcomes. METHODS: We linked population-based data in British Columbia, Canada from 01/01/2002 to 12/31/2012 on all inpatient/outpatient visits and medications with a perinatal registry. For small-for-gestational-age (SGA) births, we assessed csDMARD exposure 90 days preconception or during pregnancy until date of delivery. For congenital anomalies, we determined csDMARD exposure 90 days preconception or during the first trimester. We used multivariable logistic regression models fitted with generalised estimating equations and calculated post-hoc power. RESULTS: There were 185 pregnancies in 175 women (31.3±5.4 years) and 6,064 pregnancies in 4,387 women (31.1±5.4 years) in the csDMARD exposed and unexposed groups, respectively. Hydroxychloroquine, azathioprine, sulfasalazine, and methotrexate exposure before or during pregnancy were not associated with SGA births. The most sufficiently powered analyses were those for hydroxychloroquine, where exposure during pregnancy resulted in an adjusted odds ratio (aOR) of 1.12 (95% confidence interval [CI], 0.65-1.94) for SGA births. Although post-hoc power calculations indicate less power to detect associations between csDMARDs and congenital anomalies, results indicate methotrexate exposure during the first trimester is associated with elevated odds for congenital anomalies (aOR 6.58, 95% CI 1.15-37.75). CONCLUSIONS: Findings are consistent with current guidelines regarding specific csDMARD use during the perinatal period for women with RD. It is important to report well-designed epidemiologic studies to facilitate future RD/csDMARD-specific meta-analyses.