Maternal antidepressant use in pregnancy may be associated with a small increased risk of autism

The prevalence of autistic spectrum disorders has risen dramatically in the last thirty years, with recent estimates suggesting that 1 in 88 children in the United States have an autistic spectrum disorder. This increasing prevalence could simply be due to changes in the diagnosis of such disorders and so be of little interest. However, there is the possibility that increases in prevalence are due to changing environmental factors, such as medication use during pregnancy. These issues have been blogged about before on the Mental Elf – for example, foetal exposure to sodium valproate is linked with autism risk.

A recent population-based study from Sweden investigated the possibility of a relationship between antidepressant use in pregnancy, especially selective serotonin reuptake inhibitors (SSRIs), and the subsequent development of autistic spectrum disorders (ASDs). A possible genetic relationship between autism and depression was also investigated by looking at parental depression rates.

Methods

From a cohort of over 500,000 0-17 year olds in Stockholm County in Sweden, 4,429 individuals with autism were identified, of which 2,601 were without intellectual disability. Cases of autism were identified from various registers recording autism diagnosis and care. Individuals with autism were matched with 10 living controls without autism, matched on month of birth and gender.

Parental depression was identified using psychiatric registers, with only cases of depression before the birth of the child being included. Maternal antidepressant use was determined from drug use data collected at interview at around 10 weeks gestation – this data was available for 1,679 ASD cases and 16,845 controls. Other data, with possible links with autism, were also collected, including parental age, educational attainment of parents and maternal region of birth.

Results

Around 1% of mothers of children with autism had a diagnosis of depression, compared with 0.6% of mothers of controls

Investigating a genetic relationship, the authors found that:

Paternal rates of depression did not differ between autism and control groups (0.4%)
Around 1% of mothers of individuals with autism had a diagnosis of depression, compared with 0.6% of mothers of controls

From the data available on maternal antidepressant use, odds ratios of children developing autistic spectrum disorders were calculated compared with the children of mothers who were not depressed or taking antidepressants. Odds ratios were adjusted using data collected on possible confounders, such a parental age.

From the sample with maternal antidepressant data use, maternal history of depression was only associated with an increased risk of ASD when antidepressant use was reported during pregnancy – with odds ratios of 1.06 (95% CI [0.68 to 1.66]) for depression with no drug use, compared with 3.34 (95% CI [1.50 to 7.47]) for depression with drug use
The above relationship between antidepressant use and ASD was largely driven by increased risk of developing autism without intellectual disability (ID) – odds ratios of 1.30 (95% CI [0.72 to 2.34]) for ASD without ID and 1.04 (95% CI [0.50 to 2.13])
Antidepressant use with or without depression was associated with increased risks of ASD – odds ratio 1.90 (95% CI [1.15 to 3.14])
Looking at specific drugs, both SSRIS and non-selective monoamine reuptake inhibitors were associated with increased risks of ASD, with greater odds of developing autism without intellectual disability
If there is a real, causal relationship between ASD and antidepressant use in pregnancy, 0.6% of cases of autism may be preventable

Conclusions

The authors conclude that:

In utero exposure to both SSRIs and non-selective monoamine reuptake inhibitors was associated with an increased risk of autistic spectrum disorders.

This study shows us that the effect that antidepressants have on autism prevalence is very small

Although this conclusion is statistically true, there are some important issues to take into account. Confidence intervals in the study are wide, and sample sizes for drug data are quite small – results from the study may need to be supported before being completely accepted. More importantly, the present study design is unable to distinguish the effects of depression from those of antidepressant use. It is possible, as the authors point out, that antidepressant use is a marker for more severe cases of depression. Alterations in foetal exposure to both serotonin from antidepressants or glucocorticoids from depression may lead to alterations in development, and potentially to ASDs, so it is important that future research separates these two possible effects.

Overall, the effect on ASD prevalence that could be attributed to antidepressant use is very small. Although some cases of ASD may be preventable by discouraging antidepressant use during pregnancy, this raises the potential of harm due to the lack of treatment of depression. Because of this issue, further evidence on the potential relationship between ASD and antidepressants is needed before firm conclusions can be drawn and so clinical advice given.