Neurotoxic Metal Coexposures: Claus Henn et al. Respond

We thank Dorea for his comments on the importance of examining breast-feeding and ethylmercury exposure in our study of manganese–lead coexposures and neurodevelopment (Claus Henn et al. 2012). We agree that both breast-feeding and organic mercury exposure may affect neurodevelopment and have the potential to act as confounders and/or effect modifiers in analyses 
of metal effects on neurodevelopment. 
 
To be a confounder, a variable must be associated with both exposure and outcome. In our data, duration of breast-feeding was not strongly associated with exposures (measured by blood manganese and lead levels). When breast-feeding variables were forced into final models, the manganese–lead effect estimates did not change appreciably. Although breast-feeding did not appear to be an important confounder in our data, we agree with Dorea that this factor needs to be considered in studies of prenatal and early life environmental exposures and neurodevelopment. 
 
We agree with Dorea that organic mercury 
may be an important coexposure, acting potentially as a confounder and/or effect modifier of the manganese–lead association with neurodevelopment. We do not have detailed data on vaccination rates and ages. However, if the primary source of mercury exposure among these participants is via thimerosal-containing vaccines, as Dorea suggests, then in order for mercury to be a confounder, becoming vaccinated must be associated with manganese and lead exposures. We posit that any vaccination and lead–manganese exposure association would be weak at best, thereby reducing concerns that confounding by ethylmercury would explain observed associations between manganese–lead exposure and neurodevelopment. Nonetheless, we recognize that exposure to mercury and other co-pollutants may be a source of unmeasured confounding, which we stated in our article (Claus Henn et al. 2012). 
 
Three-way interactions are undoubtedly important and warrant attention. The limited sample size in our study population precludes us from looking at a manganese–lead–mercury interaction. In the future, epidemiologic studies should be designed with sufficient power to address the impact of multiple metal exposures, including three-way interactions.

http: //dx.doi.org/10.1289/ehp.1205004 Claus Henn et al. (2012) addressed a "real world scenario" of exposure to multiple neuro toxic metals in their unique and interesting study. They investigated manganese-lead coexposure and its association with neuro developmental deficiencies in Mexican children. Their rationale was that neuro developmental deficiencies of both metals together could be more severe than expected based on effects of exposure to each metal alone. Indeed, they observed a synergism between manganese and lead. Given the early age of the subjects (12 and 24 months of age), I suggest that some confounders not included in their model deserve consideration in regard to this study.
Claus Henn et al. (2012) collected informa tion on duration of breastfeeding, but it seems that in their statistical analyses, they adjusted only for sex, gestational age, hemoglobin, maternal IQ (intelligence quo tient), and maternal education. Other con founders, such as thimerosal (a compound containing ethyl mercury that is used as a preservative in some vaccines) and breast feeding, may influence neuro development outcomes. In countries such as Mexico, chil dren 12-24 months of age may be immu nized with thimerosalcontaining vaccines (TCVs) (WHO 2011). Because of opposite effects on the central nervous system, the combination of breastfeeding and ethyl mercury may influence neuro developmental outcomes. Kramer et al. (2008) showed that children who were exclusively breast fed had improved cognitive development. Indeed, Kostial et al. (1978) demonstrated that infant rats fed cow's milk diets absorbed more lead and manganese, which are asso ciated with a higher relative retention of mercury in the brain.
Blood levels of lead and manganese are indicators of ongoing exposure; however, ethylmercury has a short halflife and thus is unlikely to be concurrently measured in blood (Dórea et al. 2011). Nevertheless we can ascertain exposure from vaccination cards (Dórea et al. 2012;Marques et al. 2009). Following participants in the National Immunization Program of Mexico, the amount of ethylmercury from routine immuniza tions against hepatitis B (three doses), DTP (diphtheria, tetanus, and pertussis, three doses), and influenza can be estimated from records on vaccination cards. Additionally, during pregnancy, Mexican mothers may receive tetanus toxoid (TT) vaccines and other products, such as antiRhoD immune globulins (given to Rhnegative mothers) that may contain thimerosal (Marques et al. 2009). These sources of pre natal and post natal ethyl mercury exposure should be considered significant sources of an additional neuro toxic coexposure-organic mercury.
Claus Henn et al. (2012) realized that information on the association of neuro development and coexposure to multiple chemicals is limited; the scientific literature is even more scarce for the specific exposure to small amounts of ethyl mercury derived from TCVs (Oken and Bellinger 2008), which are largely used in non industrialized countries. However, recent work has sug gested that when studies with young children are properly adjusted for exposure to TCVs, subtle neuro developmental effects can be demonstrated (Dórea et al. 2012;Marques et al. 2009;MrozekBudzyn 2011aMrozekBudzyn , 2011b. Therefore, the potential for interaction of ethyl mercury, manganese, and lead provides an opportunity to expand our knowledge. Factors related to maternal neuro toxic exposure and neuro development (e.g., breast feeding) are significant in studies of chil dren's exposure to ethylmercury (Marques et al. 2009). The study design used by Claus Henn et al. (2012) could provide further information on this timely issue and also provide direction for future studies of con taminants and confounders that affect neuro development.
The author declares that he has no actual or potential competing financial interests. We thank Dórea for his comments on the importance of examining breastfeeding and ethyl mercury exposure in our study of manga nese-lead coexposures and neuro development (Claus Henn et al. 2012). We agree that both breastfeeding and organic mercury exposure may affect neuro development and have the potential to act as confound ers and/or effect modifiers in analyses of metal effects on neuro development.

José G. Dórea
To be a confounder, a variable must be associated with both exposure and outcome. In our data, duration of breastfeeding was not strongly associated with exposures (measured by blood manganese and lead levels). When breastfeeding variables were forced into final models, the manganese-lead effect estimates did not change appreciably. Although breastfeeding did not appear to be an important confounder in our data, we agree with Dórea that this factor needs to be considered in studies of prenatal and early life environ mental exposures and neuro development.
We agree with Dórea that organic mercury may be an important coexposure, acting potentially as a confounder and/or effect modi fier of the manganese-lead association with neuro development. We do not have detailed data on vaccination rates and ages. However, if the primary source of mer cury exposure among these participants is via thimerosalcontaining vaccines, as Dórea suggests, then in order for mercury to be a confounder, becoming vaccinated must be associated with manganese and lead expo sures. We posit that any vaccination and lead-manganese exposure association would be weak at best, thereby reducing concerns that confounding by ethyl mercury would explain observed associations between The correspondence section is a public forum and, as such, is not peer reviewed. Get it for your iPhone or iPod Touch. Get it for FREE.
To download the free app visit the iTunes App Store, or log on to www.ehponline.org/mobile Scan with your iPhone or iPod Touch QR reader to get the EHP app. manganese-lead exposure and neuro development. Nonetheless, we recognize that exposure to mercury and other copollutants may be a source of unmeasured confound ing, which we stated in our article (Claus Henn et al. 2012).
Threeway interactions are undoubtedly important and warrant attention. The limited sample size in our study population precludes us from looking at a manganeselead-mercury interaction. In the future, epidemiologic studies should be designed with sufficient power to address the impact of multiple metal exposures, including three way interactions.