Phthalate exposure and early thelarche.

Several years ago Colón et al. (2000) reported higher levels of phthalates, particularly di-(2-ethylhexyl) phthalate (DEHP) in serum from 41 girls experiencing premature breast development (thelarche) as compared to 35 age-matched controls. These data seem puzzling for at least two reasons. First, in light of the pharmacokinetic properties of phthalates, the reported blood levels are very high when compared to more recent exposure information, and, second, toxicologic evidence shows that phthalates do not act like estrogens under in vivo conditions, nor do they affect female sexual development and maturation in rodents. Despite these concerns, the study by Colón et al. (2000) is one of the few studies that has compared chemical exposures with sexual development and, as a consequence, it is now being cited by authors developing hypotheses for future research programs on children’s health (e.g., Chapin et al. 2003). There have been a number of recent scientific developments that relate to these matters. Thus, it seemed timely to summarize the exposure and toxicologic information that would be relevant in assessing whether low-level exposure to phthalates, which might be experienced under typical ambient conditions, could influence human female sexual development.

Colón et al. (2000) reported an average DEHP concentration of 450 ppb (nanograms per milliliter) DEHP and 3 ppb monoethylhexyl phthalate (MEHP, the first metabolite of DEHP) in serum samples from the referred cases. In comparison, they found 70 ppb DEHP in controls, whereas MEHP was below levels of detection. In general, the measurement of phthalates in biological fluids and other media can be quite problematic because of the potential for laboratory contamination due to the use of flexible vinyl in laboratory equipment and tubing. Also, the use of flexible vinyl in liners for reagent bottles is common (e.g., Kessler et al. 2001). Much of the historical data on phthalate levels is questionable, and investigators are now measuring phthalate metabolites in biological media as a way of avoiding the sample contamination problems (e.g., Blount et al. 2000a, 2000b; Kessler et al. 2001). However, taking the data at face value, the results are difficult to rationalize with information on phthalate pharmacokinetics and exposure.

Within the population at large, the principal route of exposure to phthalates is via food (e.g., Clark et al. 2003). Ingested phthalates are converted to the corresponding monoesters before absorption (e.g., Kluwe 1982), and blood levels of phthalate diesters are normally very low. To put this into perspective, a blood level of 450 ppb DEHP is approximately 1.2 μM. In a recent pharmacokinetic study of DEHP in the marmoset (Kessler et al. 2004), oral doses of DEHP at levels of 30 and 500 mg/kg resulted in peak blood levels of DEHP of 0.3 and 1.8 μM. Using a linear allometric extrapolation, an oral dose of approximately 300 mg/kg DEHP would be required to produce a blood level of 450 ppb. In contrast, the average DEHP exposure among U.S. children at ages similar to those examined by Colón et al. (2000) is 2.6 μg/kg/day (McKee et al. 2004; calculated from urinary metabolite data of Brock et al. 2003). Thus, the blood concentration data reported by Colón et al. (2000) are very unusual, and, if correct, imply extraordinary exposures by comparison to those of other children of similar ages.

Further, as indicated above, DEHP is rapidly metabolized to its corresponding monoester, MEHP [mono-(2-ethylhexyl) phthalate]. In a pharmacokinetic study in marmosets (Kessler et al. 2004), oral doses of 30 and 500 mg/kg resulted in peak MEHP concentrations of 8 and 66 μM, approximately 20 times greater than the peak DEHP levels at equivalent doses. In contrast, the average MEHP level of 3 ppb reported by Colón et al. (2000) was more than two orders of magnitude below the reported average DEHP level. The only imaginable situation that might produce such anomalous results is if the blood samples had been taken immediately after direct introduction of DEHP into the blood stream. This is theoretically possible, but the reported levels still seem unlikely. The only way in which significant levels of unmetabolized phthalates may enter the blood stream is by way of medical devices including blood bags, tubing, and other medical equipment [e.g., Food and Drug Administration (FDA) 2001]. DEHP is the primary phthalate plasticizer for such vinyl medical devices. However, as shown in an assessment by the FDA (2001), only critical-care procedures could produce blood levels that begin to approach the levels reported by Colón et al (2000). Further, in blood, DEHP is rapidly metabolized with a biological half-life in humans of < 6 hr (e.g., Peck and Albro 1982). Thus, assuming medical treatment to be the explanation for the high DEHP/MEHP ratio reported by Colón et al. (2000), the measured blood levels imply intensive care procedures performed in the few hours proceeding the blood collection. However, given the circumstances this seems unlikely.

In the study by Colón et al. (2000), the blood samples were taken from children after referral to pediatric clinics, that is, after premature thelarche had been observed. Because phthalates are rapidly metabolized and excreted, the blood level data, if not the result of laboratory contamination, could only have reflected exposures occurring shortly before the blood was drawn. For the hypothesized link between phthalate exposure and early breast development to be supported, one would have to assume that these point estimates were indicative of a pattern of exposure that had persisted for an extended period of time. But, as shown above, the conditions leading to the reported blood levels are so unusual that they cannot reflect an extended exposure profile. Analytical difficulties seem a more likely explanation to reconcile the data reported by Colón et al. (2000) with the extensive body of information on phthalate pharmacokinetics.

The second general reason why an association between phthalate exposure and the early onset of puberty in girls is puzzling and seems unlikely is that it is not supported by the toxicologic data. There is a hypothesis that premature thelarche is a consequence of unintentional exposure to estrogen or estrogen-like substances (e.g., Li et al. 2002; Tiwary 1998). It has been reported that some phthalates, although not DEHP, can bind and activate the estrogen receptor under in vitro conditions (e.g., Jobling et al. 1995; Soto et al. 1995); however, subsequent work has shown that this is an artifact of the testing conditions.

As described above, under in vivo conditions the phthalates are rapidly metabolized to the corresponding monoesters, but these metabolites are not active under in vitro conditions (Brady et al. 1998; Harris et al. 1997; Picard et al. 2001). In rats and mice, phthalates are inactive in uterotrophic assays and do not induce vaginal cornification (Kanno et al. 2003; Zacharewski et al. 1998). Similarly, in longer-term studies phthalates do not influence age at vaginal patency or estrous cyclicity and do not otherwise influence sexual development or behavior in female rats (Moore 2000). Some phthalates produce effects in male rats, apparently as a result of alterations in testosterone synthesis (Parks et al. 2000), but these alterations do not produce estrogen-like effects in female rats (Gray et al. 1999; Moore et al. 2001; Mylchreest et al. 1998, 1999). With respect to female reproductive development, the only effect that has been described in rats relates to histologic changes in the ovary that occur at very high doses (Lovekamp-Swan and Davis, 2003; Lovekamp-Swan et al. 2003).

These changes seem to be caused by inhibition of aromatase activity in the granulosa cells. Aromatase inhibition reduces conversion of testosterone to estradiol and produces what is effectively an antiestrogenic effect under these circumstances. However, this seems unlikely to have any public health implications, partly because the doses required in rodents are much higher than those to which humans are exposed and also because the aromatase inhibition is dependent on activation of the peroxisome proliferator-activated receptor α (PPARα). In other situations, there is evidence that PPARα-dependent effects exhibit profound species specificity, with rodents being very sensitive and humans being much less sensitive, if not refractory (e.g., Klaunig et al. 2003).

In summary, the purported relationship between phthalate exposure and early the-larche (Colón et al. 2000) seems highly unlikely, in part because the reported exposure levels do not seem plausible given other information on phthalate exposure, and also because phthalates do not influence the timing of female sexual development in laboratory studies.