Disrupting the "endocrine disruptor" hypothesis

That's the title of my latest HND piece, which speculates on how the "endocrine disruptor" mission got started. Along the way, we look at the consequences of overzealous regulators, and mention two articles (one already covered in this blog), which espouse rather different views of the situation.

The second article was written by a true believer—Stephanie Engel—who posits that "Any level higher than zero (of phthalates) is to some degree abnormal," and even objects to calling phthalates "trace chemicals." The article attempts to relate findings in an infant behavior test (BNBAS) to phthalate exposure, although the results—to be kind—are inconclusive.

I promise you that ten or fifteen years ago, this paper would not have been published.

I posed a few questions to Engel, which to her credit she did answer quickly...

Shaw Question 1.    Since sex-specific effects were hypothesized a priori, what specific hormones do you think are involved? How do these hormones relate to such factors as can be measured with BNBAS?

Engel's answer:  Research has shown that phthalates can be anti-androgenic (i.e. interfere with testosterone), anti-estrogentic, and/or estrogenic. They are known reproductive toxicants that have been shown to be related to reduced anogenital distance in both animal and human studies. Previous studies have shown that boys and girls perform slightly differently on the BNBAS overall, which may be related to differences in sex hormones/ brain development. We therefore hypothesized that phthalates may impact BNBAS differently in boys and girls.

Shaw comment:  She did not mention any specific hormones, referring only to speculations derived from earlier research. Normally, if a biological effect is proposed, some mechanism should be suggested.

Shaw Question 2.    You state that the median phthalate biomarker concentrations are within the range reported on another survey. Does this mean that most of the subjects were in a normal range?

Engel's answer:  I am not comfortable with referring to phthalate biomarker concentrations as "normal" or their range as "normal". Normal implies endogenous levels, and phthalates are exogenous environmental toxicants. Any level higher than zero is to some degree abnormal. However, the levels that we measured in these women were in the range of what has been reported in the large, population-based NHANES study. This implies that the women in our cohort were no more highly exposed than the general population is (i.e. we are not describing an unusually highly exposed population).

Shaw comment:  Normally, if one suspects an effect from a particular environmental chemical, tests are run on a normal as well as an occupationally-exposed cohort (or at least a cohort that has a higher exposure) to allow for a classic control. Since nearly everyone is exposed to phthalates, the researcher should have tried to find a cohort (possibly Amish people, who might not use modern personal care products??) with a much lower exposure.

Shaw Question 3.    Did you determine if lifestyle factors could explain the differing levels of phthalate metabolites in the women, or, again, did you simply see a normal range of titers?

Engel's answer:  We performed multivariate analyses that considered lifestyle factors that may be both associated with biomarker concentrations and BNBAS domains. In general everyone is exposed to phthalates. Women tend to have higher exposure to lower molecular weight phthalates than men do because they tend to use more personal care products that contain these chemicals. The purpose of this analysis was not to explain variation in phthalate levels (this has already been done), but to determine whether prenatal phthalate biomarker concentrations associate with neonatal behavior, after accounting for factors that might be associated with both phthalates and neonatal behavior. We found that phthalates were associated with neonatal behavior, particularly for girls, and particularly for the domains of orientation and quality of alertness.

Shaw Question 4.    Since you are trying to look at effects of a trace chemical, how wise is it to rely on a questionnaire to determine smoking, alcohol consumption, and illegal drug use--especially if CDC was running the urine tests--and presumably could have tested for such activities?

Engel's answer:I think characterizing phthalates as a "trace chemical" is inaccurate. Phthalate exposure in the general population is orders of magnitude higher than most other known environmental toxicants (PCBs, lead, methylmercury, organophosphate pesticides, bisphenol A). However, your question as to whether self-report is adequate to measure smoking, alcohol consumption and illegal drug use has also been addressed in methodological studies. Smoking is variable during pregnancy as women tend to repeatedly try to quit or cut-back, so biomarker measuring of cotonine in urine during pregnancy has almost as many problems as questionnaire based assessments (a spot cotinine level only addresses 2 week exposure). Methodological studies has found maternal self-reported smoking to be more accurate. Alcohol consumption cannot be assessed through biomarkers, and is probably under-reported. Illegal drug use can be measured through biomarkers, and is probably under-reported by questionnaire. However, these two factors would have to strongly associate with phthalate exposure to represent a substantial bias in our study. Neither do.

Shaw comment:  Sorry, not good enough. Since all the women are exposed to phthalates, how this associates with illegal drug use is irrelevant. Certainly, illegal drug use could affect the BNBAS, and to not check for it is simply ignoring a huge confounding factor. My take is that the women only agreed to the urine tests on the promise that illegal drug use was not going to be examined, although this is obviously not mentioned in the paper.

I am still waiting for an unequivocal study showing real health effects in real humans from these endocrine disruptors, but I'm not holding my breath.