Belarusian psychiatrist questions study results on the influence of popular drugs on the risk of autism in children

Vladimir Pikirenya. Photo: Facebook

A study conducted by scientists at the University of Nebraska Medical Center found an unexpected link between fifteen popular drugs (including statins, a number of antidepressants, antipsychotics, and beta-blockers) and an increased risk of autism spectrum disorders (ASD) in children.

The results were published in the journal Molecular Psychiatry and quickly spread through the media. Psychiatrist Vladimir Pikirenya, in his Telegram channel, expressed the opinion that the authoritative scientific journal, in pursuit of a sensation, published a low-quality study.

What science already knows about the causes of autism

To substantiate his view, Pikirenya presents the following arguments. First of all, the doctor suggests recalling what science has long and precisely known about the causes of autism. For example, the genetic factor plays a colossal role: if one of monozygotic twins has ASD, the risk for the other increases by 150 times.

It has also been scientifically proven that parental age (every 10 years adds 20% probability), male sex of the child (increases risk by 4 times), and even the use of specific epilepsy medications (valproic acid) increase the risk by 5 times.

Any mental disorder in either parent itself doubles the risk. In addition, premature birth, low birth weight, complications during childbirth, gestational diabetes, preeclampsia, autoimmune diseases, asthma, and obesity increase the risk by 10-30%.

The main problem is the differing health status of mothers

Regarding the study itself, the psychiatrist's main criticism concerns how the scientists formed comparison groups. It was a so-called observational study, where conclusions are drawn based on medical databases and patient records.

However, as Pikirenya notes, women who took medications during pregnancy and those who did not initially differed significantly in their health status.

If you look at the study tables, factors already associated with an increased risk of ASD were significantly more common in the group of children with autism. For example, premature births occurred in 15% of cases versus 11% in the control group. Maternal obesity was 35% versus 21%, diabetes 19% versus 15%, bipolar disorder 10% versus 2%.

An especially large difference was observed in mental health: in the group where children later developed autism, depression was present in 50% of mothers versus 9.8% in the control group, and anxiety disorders in 68% versus 18%.

Pikirenya notes that it is precisely these women who are more often prescribed statins, antidepressants, and antipsychotics. Therefore, the researchers might have confused the effect of the drugs themselves with the effect of the diseases for which these medications were prescribed.

Correction of results

The psychiatrist also draws attention to the results after statistical correction. When the authors tried to mathematically account for mental disorders and other maternal factors, the risk estimates from drug intake significantly decreased. For example, for aripiprazole, the risk after adjustments fell from 28% to 13%; for sertraline, from 54% to 10%; for haloperidol, from 87% to 13%.

At the same time, as Pikirenya notes, the researchers did not make the code for their calculations publicly available, although they referred to it in the article (the link to the GitHub repository turned out to be empty). This means that it is simply impossible to independently verify the correctness of their complex statistics.

Comparison with other drugs

He sees another problem in how control drugs were selected. For a correct analysis, it was necessary to choose drugs with the same indications, but without the presumed mechanism of influence on the development of autism (disruption of cholesterol synthesis).

Instead, according to him, the researchers took random drugs, among which there were no antidepressants or antipsychotics. And it turned out that even these, according to their scheme, showed an increased risk of autism.

Pikirenya believes that this may indicate not the harm of specific drugs, but that women with serious health problems themselves are already in a high-risk group.

Harmful habits

The psychiatrist also questions the quality of the initial data itself. According to the article, less than 1% of pregnant women in the sample smoked or consumed alcohol. This contradicts official American statistics, which show that about 3% of pregnant women in the USA smoke, and approximately 8.5% consume alcohol (with about half of them in high doses).

The authors of the study did not explain these anomalous figures, which casts doubt on the reliability of the entire database on which the conclusions were built.

"The published article once again showed how, in pursuit of sensational results, researchers—consciously or unconsciously—can distort data, and scientific journals, in pursuit of ratings and attention, readily publish such works. And this once again reminds us that even publications in the most authoritative scientific journals should be treated with great caution," the expert concludes.

• 10
• 0
• 1
• 5
• 0
• 0
• 0
• 1