Shelter from the storm: Acetaminophen (paracetamol) in pregnancy, urogenital malformations, and childhood neurodevelopment

The issue of whether taking paracetamol (acetaminophen, APAP) during pregnancy is associated with an increased rate of congenital malformations and childhood neurodevelopmental disorders raises its head on a perennial basis.^1–4 In a recently high-profiled “consensus statement” published in a high-impact journal, Nature Reviews Endocrinology, these insufficiently substantiated concerns are brought forward again. ⁵ Unsurprisingly, cautionary statements made in this paper have been widely shared by news stations, newspapers and through social media platforms with very little (if any) critical appraisal of the presented evidence. It makes for good headline news, but undoubtedly will have raised doubt, fear, guilt and mistrust of the medical profession amongst pregnant women.

We challenge the scientific substantiation of the narrative in this “consensus statement”.

Whilst the authors’ overall suggestions and call for more and better research are hard to oppose, they are generic: 1) do not use APAP indiscriminately during pregnancy 2) use at the lowest dose and for the shortest duration of time possible 3) we need more and better data. Such statements can be made for the use of almost any drug during pregnancy. They also ring hollow, as they are preceded by a long and comprehensive discussion that suggest non-trivial increased risks of congenital malformations and childhood neurodevelopment following in utero exposure to APAP.

The consensus statement has been written in such a way that the authors infer a causal relationship between APAP and adverse outcome. To make causal inferences from epidemiological and observational data requires studies of high methodological quality. Such studies are not presented. The authors, themselves, acknowledge in their conclusion that more and better data quality are required. ⁶ It is also striking that the specific studies referenced only appear subject to superficial critical assessment. None of the points below are brought forward or discussed.

1. Methodological concerns regarding neurodevelopmental assessment.

   Data on neurodevelopment and ADHD/ADH are reliant on outcome measurements in the form of questionnaires, such as the Age and Stages Questionnaire (ASQ) or similar, filled by parents or teachers. These questionnaires typically use many domain-specific questions on a Likert scale. While these may serve a screening purpose to identify children at risk, they are not validated as an instrument or proxy for a quantitative assessment of neurodevelopment, nor are they a diagnostic tool for ADHD or ASD. ²

2. No adjustment for family history of neurodevelopmental disorders.

   Cited studies relating to ADHD and ASD did not adequately account for heritability of these conditions. When accounting for heritability in a quantitative bias analysis, all signals disappeared resulting in a null association. ⁷ A Norwegian analysis found the same extent of association between in utero exposure and paternal exposure prior to conception to APAP and increased risk of ADHD in the child. ⁸ These findings strongly suggest that a confounder, i.e., heritability of ADHD or familial environmental conditions, rather than exposure to APAP be the connection to an increased risk of ADHD. Few of the cited individual studies adjust for maternal BMI. In the consensus statement the Boston Birth Cohort is cited to substantiate and quantify “true” APAP exposure in the exposed liveborn children and performing an inferential analysis to the risk of ADHD or ASD in childhood. ⁹ There are several issues with this study that make its reliability highly questionable. Firstly, APAP and/or metabolites were detected in 996 of 996 umbilical cord samples. A 100% exposure rate should raise eyebrows. Secondly, the quantitative construct of “APAP burden” used to stratify levels of exposure in tertiles is not validated. Actual levels of APAP and metabolites are not presented for the respective tertiles. Thirdly, the APAP analytical method is not specified, and the potential issue of sample stability is not discussed, some samples had been stored for up to 20 years. Fourthly, there was an extremely high prevalence of neurodevelopmental disorders in the cohort. 365 out of 996 children (37%), were diagnosed with ADHD and/or ASD and just 327/996 (33%) children were NOT assigned a “developmental disability” diagnosis. These prevalences are incompatible with the disease prevalences reported in general populations. Finally, parental ADHD and ASD were insufficiently accounted for in the inferential analysis.

3. Biased representation of the literature.

   The authorś substantiation of an increased risk of cryptorchidism and anogenital distance (AGD) is unconvincing. In most individual studies referenced that adjust for maternal BMI, a known risk factor for cryptorchidism, no associations were found. An overall null association between APAP exposure and risk of cryptorchidism (adjusted OR 1.3, 95% CI 0.70–2.6) was reported in the 2011 paper cited by the authors. Yet this pivotal result is nowhere to be seen, much less discussed, in the consensus paper. ³ The original paper on AGD and APAP exposure states:” Maternal use of paracetamol was not significantly associated with AGD in boys after adjustment”. ⁴ A new study confirms this finding. ¹⁰

In summary, the evidence presented has not been systematically reviewed and presents a biased, inferred narrative that APAP is harmful. Studies included have not been critically assessed to an extent one should rightfully expect. They are inconsistent, methodologically challenged and, in their totality, do not support a clinically meaningful association between APAP exposure and increased risk of congenital malformations or childhood neurodevelopmental issues.

Acetaminophen remains a well-documented, safe analgesic during pregnancy. The alternatives, NSAIDs (which lead to premature closure of the ductus arteriosus and reduced fetal renal blood flow) and opi analgesia (which is associated with drowsiness at birth and neonatal opioid withdrawal syndrome) have far more substantiated risk. It is disappointing that such a biased and unbalanced narrative has been published in a widely respected journal. The clinical impact of the statement is likely minimal, since APAP is still first-line treatment for pain and or pyrexia for most pregnant women, only to be used when needed at the lowest effective dose; nothing has changed. However, the anxiety, fear and mistrust invoked by the ‘consensus statement’ (which, incidentally, did not include any representation from teratology information, obstetric organizations or maternity services where there is substantial expertise) will have significant, longer-lasting consequences for years to come.