Among pregnant women, acetaminophen is the most commonly used pain reliever, and likely the most commonly used of any medication. Estimates suggest that anywhere from 40% to greater than 65% of pregnant women use an acetaminophen-containing product at some time in pregnancy. Recently, concerns have been raised in several studies about potential increased risks for a variety of neurobehavioral outcomes including attention-deficit/hyperactivity disorder (ADHD), autism spectrum disorders (ASD), cognitive deficits, and other behavior problems.
Aconducted in the Norwegian Mother and Child Cohort Study (pregnant women enrolled 1999-2008) evaluated 2,919 same-sex siblings at 3 years of age whose mothers were enrolled in the cohort during pregnancy. Prenatal exposure to acetaminophen was collected from maternal interview data during and shortly after pregnancy. Behavioral performance in the children was measured using validated maternal questionnaires. After adjustment for maternal febrile illness, infection, and comorbidities, use of acetaminophen for 28 days or more in pregnancy was associated with poorer gross motor and communication performance, more externalizing and internalizing behaviors, and higher activity levels in the offspring. Exposure for less than 28 days was associated with poor gross motor skills only. There were no associations found between ibuprofen use in pregnancy and any of the neurobehavioral outcomes.1
Four additional studies on this topic were published in 2016. Onefrom the Avon Longitudinal Study of Parents and Children followed 7,796 pregnant women enrolled in Bristol, England, in 1991-1992. This study’s findings were similar to the Danish cohort, with increased risks for conduct problems and hyperactivity symptoms as assessed in the Strengths and Difficulties Questionnaire at 7 years of age. In contrast, there was no association found with postnatal maternal acetaminophen use or paternal use.3
Two additional studies published in 2016 were from the Danish cohort described above. The first assessed long term follow-up for an average of 12.7 years linked to incidence of ASD or infantile autism in the Danish Psychiatric and Hospital Registries. An increased risk for ASD was found only when the diagnosis was accompanied by hyperkinetic symptoms (RR,1.51; 95% CI, 1.19-1.92), and this risk was most pronounced when acetaminophen was used in the last half of pregnancy.5
The second study examined a smaller subset of 1,491 children who were tested at 5 years of age by trained psychologists using the Wechsler Preschool and Primary Scale of Intelligence – a measure of IQ. Maternal fever and acetaminophen use in pregnancy were associated with small but significant deficits. For example, children of mothers who used acetaminophen but reported no fever had on average a 3.4 point deficit in performance IQ, compared with children of mothers who reported neither. However, mothers who reported both acetaminophen use and fever showed no differences.6
The authors of the studies described above, as well as other experts in the field, have discussed the biological plausibility of such associations being causal. The endocrine disruptive effects of acetaminophen and/or an impact on oxidative stress have been suggested as possible mechanisms, although no convincing data have supported these theories to date. Several of the studies have been criticized for reliance on maternal report of the outcomes, and all studies relied on maternal report of the exposure, which raises possibilities for bias and misclassification. However, some studies included assessments by trained psychologists or linkage to diagnoses recorded in national health registries. Control for confounding is always problematic in observational studies, especially with long-term outcomes. The sibling-pair study conducted in Norway attempted to control for genetic and environmental confounding using this design. Confounding by indication is also a concern. Adjustment for maternal fever, infections, or other comorbidities may have addressed some of this concern; however, as with acetaminophen, vague and perhaps inaccurate maternal report of these events may be problematic. Findings for another pain reliever, ibuprofen, showed no association in one study, which provides some reassurance.
At the present time, the impacts on behavior including hyperactivity and perhaps ASD are most prominent in the published reports stemming from four countries with large cohort studies. There is conflicting evidence of any impact on cognitive performance. All associations reported are quite modest with relative risks in the 1.5 or lower range. Nevertheless, acetaminophen is a medication taken by most women in pregnancy, so the broad impact, if these associations are causal, is not trivial. Clinically, an urgent concern would be that women or clinicians might ignore the current advice to take acetaminophen to reduce maternal fever. In that regard, it is reassuring that the Danish Cohort study found no impact on IQ when acetaminophen was used to treat a fever, while fever itself was associated with increased risks.
In studies which examined frequency of use, longer duration was often associated with higher risks. However, no study reported on dose. Future studies should capture dose, as well as better information on the underlying conditions being treated. The findings from these epidemiologic studies also call for additional experimental work to better understand possible mechanisms. In the meantime, as with any over-the-counter medication, judicious and appropriate use of acetaminophen should be advised.