A new study regarding phthalates has garnered media attention this month, but readers should recognize the study’s limitations. Some media coverage of this study blurs the important distinction between “association” and “causation.”
What Are Phthalates?
Phthalates, sometimes called plasticizers, are a group of chemicals generally used to make plastics more durable, or to dissolve other materials. Phthalates may be found in products such as vinyl flooring, food wraps, intravenous tubing, lubricating oils, and some personal care products such as shampoos, soaps, and hairsprays.
The Recent Study
The August 20, 2021 edition of Environmental Pollution features a study entitled “Phthalates and Attributable Mortality: A Population-Based Longitudinal Cohort Study and Cost Analysis.” The study evaluated a relatively small sample size of 5,303 adults ages 20 years and older (median age 56.5 years old) who participated in the US National Health and Nutrition Examination Study 2001-2010 and who provided urine samples for phthalate metabolite measurements. The authors conclude that the study shows a “potential increase [in] the probability that phthalates contribute to cardiovascular mortality,” particularly in adult men, among other things. But, the authors themselves clarify that this study shows only an association between certain phthalates and certain mortalities – not a causal relationship.
The Recent Study vs. Recent Media Coverage
The difference between association and causation is an important one. A statistical association between two variables only implies that knowledge of the value of one variable provides information regarding the value of another. It does not, however, imply that one variable causes the other.
There is a dissonance between some of the recent media coverage of this study and what the study itself actually shows. If you only look to the media stories regarding this study, you will have missed:
- The study relies on a survey technique with questionable validity – a single spot urinary sampling – which even the authors acknowledge is a weakness in their methodology, further exacerbated by extrapolating broadly based upon a relatively small sample;
- The study showed no statistically significant associations with low molecular weight phthalates, and only very weak associations in high molecular weight phthalates;
- The study showed only weak associations, in the highest exposure group, for DEHP (one of the most common members of the class of phthalates); and
- The study showed no association with cancer mortality.
The methodology, conclusions, and inferences in this study may face some serious criticisms if relied upon to support individual or population-based opinions or action. The lead author, when interviewed by the press, admitted, “I’m never going to tell you this is a definitive study. . . . It is a snapshot in time and can only show an association.”
The study also includes discussion of a correlating annual economic burden, which is estimated at between $40-$47 billion. However, the authors themselves acknowledge that “some will argue that economic analyses are premature in the case of phthalate-induced cardiovascular mortality[.]” It is worth noting that this study’s estimate is nearly quadruple that of previous estimates. The authors justify making an early estimate of this kind here by merely stating “it is important to consider the” Bradford Hill criteria “for evaluating whether sufficient evidence exists to proscribe public health action.” To date – perhaps for any of the reasons explored here – no such action has occurred.