I'm sorry to say that this is one medical report that I can't translate into layman's terms. Sometimes medical research studies just aren't meant to speak to the public at large.
Source: Ystrom E, Gustavson K, Brandlistuen RE, et al. Prenatal exposure to acetaminophen and risk of ADHD. Pediatrics. 2017;140(5). pii: e20163840; doi:10.1542/peds.2016-3840. See AAP Grand Rounds commentary by Dr. Dan Doherty (subscription required).
When this study first appeared it was widely reported in the lay press, and I recall being a bit stunned by the criticisms aired at the time. When it appeared in February's AAP Grand Rounds list, I knew I would need to say something about it.
Norwegian researchers used data from the longitudinal Norwegian Mother and Child Cohort Study to determine whether there was an association between use of acetaminophen during pregnancy and diagnosis of ADHD in their offspring. They looked at children born between 1999 and 2009 and collected data via the Study until the end of 2014, meaning that some children only had 5-year follow-up while some had as long as 15 years (more on that in a moment). Their final sample totaled 112,973 children, and a key finding after attempting to correct for many confounding variables was that the hazard ratio for the diagnosis of ADHD in mothers reporting 29 or more days of acetaminophen use during pregnancy was 2.20, with 95% confidence intervals of 1.50 and 3.24. It's a very well-performed study, and the authors mentioned multiple study limitations including the accuracy of diagnosis (made with administrative code data, not verified from clinical records). Unmentioned perhaps because it's obvious is that data collected by subject questionnaire is subject to the usual recall bias (e. g. all medication use information was gathered by use of questionnaires). The study's main conclusion, that "long-term maternal use of acetaminophen during pregnancy was substantially associated with ADHD even after adjusting for indications of use, familial risk of ADHD, and other potential confounders," was justified by their data.
So, I was surprised at the vitriol expressed by some experts interviewed in the lay press. Take the CNN story, for example. One expert stated that the study "...doesn't add anything to our medical knowledge." In fact, it adds a lot. Perhaps a better statement would have been that it doesn't tell us how to advise pregnant women in a definitive manner. That same expert stated that the study didn't show a dose-response curve and that the study results didn't "...make any biologic sense." In fact, the authors proposed 3 different biologic mechanisms that could explain the results, and a dose-response effect was noted (less frequent use of acetaminophen had lower hazard ratios). Those comments made me wonder if the expert had read the study completely; sometimes physicians are asked to get on the airwaves the moment a study is published online, which doesn't necessarily allow a lot of time to digest everything in the study. I'm not sure that I'd call this report "fake news," but it seemed unnecessarily and perhaps recklessly dismissive to me. (Another explanation is that she was misquoted or had her remarks taken out of context.)
I did have enough time to digest this, so I don't have an excuse for my conclusions or lack thereof. I fall short in translating the results into some useful advice to women of childbearing age. Yes, we always have problems knowing what to do with prospective cohort studies, which can only imply an association between an intervention (acetaminophen use in pregnancy) and an outcome of interest (ADHD in offspring). That is clearly different from cause and effect conclusions that are easier to believe from randomized controlled trials. However, the real problem for my translational efforts is that the children in the study had different periods of follow-up. Those born later in the cohort had only 5 years of follow-up, much less than the oldest children in the study. This will impact the rates of ADHD diagnoses in these children, some of whom will no doubt be diagnosed later in life. The researchers reported their findings in terms of hazard ratios, which is appropriate when the time to an outcome event and/or the observation period vary among the study participants. But, it makes it tough to translate numbers into a magnitude of the degree to which acetaminophen might contribute to ADHD among all the other factors involved. Or, stated differently, if we suddenly stopped all acetaminophen use by pregnant women, how many fewer cases of ADHD would we see?
The authors estimated from their cohort that the cumulative prevalence of ADHD in their cohort was about 4% by 13 years of age. In the group of 112,973 children, however, only 2246 carried this diagnosis, which is 2%. What I really wanted to do was calculate an estimated number needed to harm for prolonged acetaminophen use during pregnancy, assuming that the association is truly causal (yes, that's cheating, but still gives me an idea of the magnitude of the effect). Sadly, I couldn't pull it off. I needed more raw data to be able to perform that calculation, or perhaps use the same statistical software the authors used and try to back-calculate to the raw data by plugging in their answers.
Apologies for not being able to make more practical sense out of the results, but this study expands our understanding and should prompt further research on acetaminophen use in pregnancy. In the meantime, the findings aren't definitive enough to warn pregnant women to stay away from acetaminophen.
