By: 15 January 2014
JAMA Pediatrics

Paper: Association of autism with induced or augmented childbirth in North Carolina birth record
Paper authors: Gregory SG, Anthopolos R, Osgood CE, Grotegut CA, Miranda ML
Paper ref: 2013 Oct 1; 167(10):959-66

The aim of the study was to determine whether births that were induced or augmented were associated with increased odds of being diagnosed with autism. The authors obtained and then linked records from the North Carolina Detailed Birth Record and Education Record databases. Exclusions were made for incomplete data, maternal age outside the ages of 15 to 49, infants with congenital abnormalities and those born earlier than 24 weeks.

Seventy-five percent of all births were able to be linked to educational records and 92 percent of those (n = 625042) were analysed using five different statistical models to estimate the odds of autism following induction or augmentation. In this study, approximately 1.3 percent and 0.4 percent of male and female children respectively had a recorded diagnosis of autism.

In the simplest model, a child whose mother was induced and augmented during delivery had 23 percent higher odds of being diagnosed with autism than a child whose mother was neither induced nor augmented (OR 1.23, 95 percent CI 1.02-1.47). For augmentation alone the odds ratio was 1.15 and for induction alone 1.10.

The subsequent models allowed for potential confounders such as maternal age, socioeconomic status, maternal conditions harmful to pregnancy and events of labour and delivery. The inclusion of these did not substantially alter the odds ratios observed in the first model, suggesting that labour induction and augmentation continued to be independently associated with autism.

However, these models did demonstrate significant associations of other factors with autism, such as foetal distress (OR 1.25), meconium (OR 1.22), preterm delivery at ≤ 34 weeks (OR 1.25) and maternal diabetes mellitus (OR 1.23), supporting previous published data.

The final model looked at whether the association shown between autism and induction or augmentation was gender specific. This showed that male children were more sensitive to birth induction and augmentation, with statistically significant raised odds ratios for induction only (OR 1.18), augmentation only (OR 1.15) and induction plus augmentation (OR 1.35). In contrast, only augmentation was significantly associated with autism diagnosis among female children (OR 1.18).

Whilst this study only demonstrates an association, if a causative effect was established this would represent an addition diagnosis of autism for every 220 male births receiving both induction and augmentation.

In the discussion, the authors suggested one possible explanation of the association is through exposure to exogenous oxytocin, influencing social behaviour and cognitive function through as yet unidentified genetic or epigenetic factors.

The limitations of this study included the inability to control for all potential confounders due to lack of data in the NCDBR (such as paternal age and other intrapartum medication); differences in dosing and regime of induction and augmentation; and the lack of consideration of autism severity and associated conditions.

In addition, further investigation would need to focus on whether the maternal, obstetric or intrapartum conditions that drive the decision to induce or augment also underlie the association with autism.

Despite their findings, the authors did identify the benefits of labour induction and augmentation including a reduction in the number of foetal deaths and meconium aspiration syndrome as well as a lower caesarean section delivery rate.