Post doctoral fellow University of Montreal, Canada
Background: Azathioprine (AZA) is an immunosuppressive drug prescribed for autoimmune diseases such as inflammatory bowel diseases (IBD). Based on small sample size, some reports suggested an increased risk of neurodevelopmental disorders (NDDs) associated with in-utero AZA exposure.
Objectives: To quantify the association between gestational use of AZA in women with IBD and the risk of NDDs in the offspring.
Methods: Using the Quebec Pregnancy Cohort, a birth cohort of singleton liveborns covered by the RAMQ prescription drug plan from 1998 to 2015, whose mothers were also covered and had an IBD diagnosis before the pregnancy was constructed. Date of entry in the cohort was date of birth, follow-up continued until index date or censuring (end of follow-up (death, 08/2015)). AZA exposure was defined as ≥1 filled prescriptions during the pregnancy. Index date was defined as having ≥1 diagnosis of attention-deficit/hyperactivity disorder (ADHD), autism spectrum disorder (ASD), specific learning disorder (SLD), combined and individually, from birth until end of follow-up, ADHD was further defined as having ≥ 1 filled ADHD medication prescriptions during the follow-up. Covariates considered were maternal sociodemographic characteristics, type of IBD, comorbidities, and other drug use during pregnancy; children characteristics at birth and calendar year of the pregnancy were also considered. A propensity score was calculated to adjust for potential confounding by indication and other risk factors. Adjusted hazard ratio (aHR) were calculated using Cox proportional hazard regression models.
Results: Of the 2,521 eligible liveborn singletons, prevalence of NDDs was 24.0% (ADHD: 17.9%, ASD: 1.9%, SLD: 11.0%). Of the 127 (5.0%) children exposed in-utero to AZA during the in-utero period, 25 (19.7%) had ≥ 1 NDD. In the crude analysis, AZA exposure was associated with an increased risk of ASD (HR 2.75, 95%CI 1.08-6.95) but not with other or any NDD. When adjusting for other covariates, through multivariate modelling or propensity score adjustment, in-utero exposure to AZA was not associated with any of the combined or specific NDDs. Male gender (aHR 1.46, 95%CI 1.07, 1.98), low birth weight (aHR 1.63, 95%CI 1.17, 2.29), preterm birth (aHR 1.46, 95%CI 1.07-1.98) and being a welfare recipient (aHR 1.30, 96%CI 1.08-1.56) were associated with an increased risk of NDDs; increasing gestational age at birth was decreasing this risk (aHR 0.96, 95%CI 0.95, 0.98).
Conclusions: Our findings suggested that in utero AZA exposure was not overall associated with an increased risk of NDD in the offspring. However, due to lack of power, we cannot exclude an increased risk of ASD. More powerful studies are needed to definitively reassure clinicians and patients.
