October 30, 2018

expert reaction to study looking at prenatal phthalate exposure and language development in early childhood

Research published in JAMA Paediatrics demonstrates that prenatal exposure to to phthalates was associated with language delay in children.  

A Before the Headlines accompanied this Roundup

Prof Dorothy Bishop FRS FMedSci, Professor of Developmental Neuropsychology, University of Oxford, said:

“I am commenting on this paper as someone with expertise in children’s language disorders.  I am not an expert on the impact of prenatal chemicals on children’s development.

“I felt the authors’ claims for an effect of prenatal phthalates on language delay were not particularly convincing.

“Positive features of the study were that they included two independent population samples (from Sweden and the USA), and also controlled for the effects of possible confounding factors (e.g. mother’s educational level, child’s sex).  However, six of the eight chemicals they investigated showed no effect, and the effects they observed in the other two were small and did not clearly replicate across samples.  They looked for a dose-response effect but did not demonstrate this clearly.

“Furthermore, the two phthalates showing association in this study were different from those showing an effect on language in a previous Danish study.  And, remarkably, levels of these chemicals were between 5 and 10 times as high in the Swedish sample compared to the US sample, but rates of language delay were very similar in the two groups.

“It would be helpful if future studies on this topic would pre-register their hypotheses and analysis plans, as otherwise it is hard to know whether the authors have just focused on a subset of results that look strongest.  Future studies on this topic should also control for family history of language problems, which is one of the stronger predictors of language delay.”

Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“We shouldn’t read too much into this study, and it’s important to note that all the researchers are proposing, in the light of their results, is that “the association of phthalates with language delay may warrant further examination.”  I wouldn’t disagree with that, but it’s premature to conclude any more.

“An important difficulty in any study like this is that the mothers with high levels of phthalates will probably differ, on average, from those with low levels in many ways, and those differences will generally not have been caused by the phthalates.  So any differences between their children might be caused by these other factors and not by the phthalates themselves.  The researchers adjusted their results to allow for some such differences, namely the sex of the baby, whether the birth was premature, the mother’s educational level smoking status, weight, and (in the American part of the study) the mother’s ethnicity, but these are not the only factors that might affect the children’s chances of delayed language development.  Particularly in the US data, these adjustments made quite large differences to the measures of association between phthalates and language delay, and that does make me question whether further adjustments of that kind might have made the observed associations disappear anyway.

“In any case, the size of the association is arguably not so large.  According to the researchers, “a doubling of prenatal exposure of [certain phthalate metabolites] increased the odds ratio for language delay by approximately 25% to 40%”, but this seems to ignore the facts that the ratio for only one of these metabolites was around 40%, only in the US data, another of the US measurements showed no increase at all in the chances of language delay, and the other ratios were nearer 25% than 40%.  Also, because of the smaller number of children in the US study, none of the results there was statistically significant, so that we can’t be confident that they aren’t just random statistical blips.

“In this study overall, out of every 100 children, 10 showed language delay.  If the level of these phthalates in their mother’s blood doubled for all of them, surely quite a big increase, then according to these findings, about another two would show language delay.  If we could be confident that the phthalates were a cause of language delay, perhaps that would be concerning, but we can’t be confident.  Furthermore, the findings from this study differ from those in other studies, for example in Denmark and in Mexico.  So this new study has moved things on to some extent, but there isn’t yet a smoking phthalate gun.”

Dr Wayne Carter, Associate Professor, School of Medicine, University of Nottingham, said:

“This study provides an insight into the possible link between prenatal exposure to phthalate and language development in early childhood.

“Although the results of the study suggest a statistical link between metabolites of phthalate and a delay in language development, this was not observed for all metabolites and nor was a clear dose response established.

“Furthermore, the sampling time for prenatal mothers was not uniform and the size of the cohorts considered relatively small.

“Hence the study is useful and of interest but studies with more expansive numbers and with a consistent study design are required to further inform the general public.”

‘Association of prenatal phthalate exposure with language development in early childhood’ by Carl-Gustaf Bornehag et al. was published in JAMA Pediatrics at 15:00 UK time on Monday 29 October 2018.

Declared interests

Prof Dorothy Bishop: “I confirm I have no conflict of interest.”

Prof Kevin McConway: “I am a member of the Advisory Committee of the SMC.”

Dr Wayne Carter: “I can confirm no conflicts of interest associated with my review of the paper.”