December 16, 2024

expert reaction to study on the association between plastic chemical exposure and human health

A study published in PNAS looks at human health impacts of plastic chemical exposure.

Dr Stephen Burgess, Group Leader at the MRC Biostatistics Unit, University of Cambridge, said:

“This article is an exercise in statistical modelling rather than a contribution to the scientific debate on the toxicity of these chemicals. The authors do not demonstrate that these chemicals truly affect disease risk in a causal sense. That is to say, it is not clear whether reducing exposure to these chemicals would truly lower disease risk or improve health outcomes.

“The authors take correlative estimates of association of various chemicals with health outcomes, and extrapolate these estimates assuming they reflect causal quantities. Perhaps these chemicals do truly causally affect the various outcomes that they have been linked with, but perhaps they do not – if they do not, then this would be far from the first example of an observational association that does not have a causal basis.”

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

“I think it’s important to remember the inevitable limitations of this study. Impacts like these are complicated and difficult to estimate on a global scale. There are a lot of uncertainties involved, and I don’t think that the press release or the abstract (summary) of the research paper make that sufficiently clear.

“One limitation, which the researchers spell out in the Discussion section of their paper, is that their calculations are based on observational research. Observational studies measure associations, that is, correlations, between exposure to potentially harmful substances and health outcomes. They cannot show at all conclusively that the exposure is the cause of the health outcomes. These researchers do assume that the correlations observed in the studies, on which they base their work, are ones of cause and effect – in other words that it is the exposure to the plastics that causes increases in death rates and ill health. That assumption could be correct, or could be only partially correct, or even not correct at all. Maybe some other factors independently affect people’s exposure to the chemicals and their chance of suffering a stroke or the other health effects that were observed.

“The fact that it isn’t clear what’s causing what is an important source of uncertainty. For clarity, I’m not trying to say that these plastics can’t present important health hazards, but only that this research can’t clearly establish the extent to which they do cause extra ill health and higher death rates.

“Another important source of uncertainty is that there are inevitable margins of statistical error in the estimates that the researchers produced. These margins of error are generally wide. For instance, in the press release and the paper abstract, we’re told that 5.4 million cases of ischaemic heart disease (heart attacks, angina etc.) were linked to BPA exposure in 2015. Table 1 in the paper points out that the statistical margin of error around this 5.4 million estimate runs from 2.0 million to 8.9 million cases. There are similarly wide margins of statistical error around the other estimates of adverse health effects, and the same goes for the estimates of costs (given in the research paper, though not in the press release). On top of that, these estimates apply to roughly a third of the world’s population for exposure to BPA and DEHP, but rather less than that for PBDEs (because data were available for fewer countries). Estimates for the whole world would be higher, but we just can’t tell how much higher. Also the cost estimates apply only to certain types of social costs, even in the countries that are included.

“So, overall, despite the detailed analyses in the paper, all we really know is that, quite possibly, the health consequences of exposure to these plastics could be very large, but we can’t at all be sure how large. Nor can we be sure to what extent the health consequences are actually caused by the exposure to these plastics. These certainly aren’t reasons to ignore the research, but I’d strongly caution against making too much of the exact estimates.”

‘The benefits of removing toxic chemicals from plastics’ by Maureen Cropper et al. was published in PNAS at 20:00 UK Time on Monday 16th December 2024.

DOI: 10.1073/pnas.2412714121

Declared interests

Dr Stephen Burgess: I have no relevant conflict of interest to declare.

Prof Kevin McConway: Previously a Trustee of the SMC and a member of its Advisory Committee.