To the Editor: Thompson et al. (Sept. 27 issue)1 report theresults of a study investigating the neuropsychological outcomesof early exposure to thimerosal. As a dissenting member of thepanel of external consultants for this study, I object to theauthors' conclusion that there is no causal association betweenthimerosal and children's brain function. The sample comprisedchildren who were least likely to exhibit neuropsychologicalimpairments. Specifically, children with congenital problems,those from multiple births, those of low birth weight, and thosenot living with their biological mother were excluded. The samplewas skewed toward higher socioeconomic status and maternal education— factors that are associated with lower rates of neurobehavioralproblems and higher intervention rates and that were not measured.The sampling frame included only children enrolled from birthin the health maintenance organization (HMO) and still enrolledafter 7 to 10 years, excluding children in higher-mobility families,who tend to have lower academic and behavioral function.2 Childrenwith neurobehavioral problems may have been less likely to remainwith the HMO. Only 30% of families selected for recruitmentparticipated, a low rate for scientific research. Among thefamilies selected for recruitment, 26% refused to participate.Another 28% "could not be located," which included familiesthat did not respond to multiple recruitment attempts (internaldocumentation from the study contractor, Abt Associates) —another form of refusal.
Sallie Bernard, B.A. SafeMinds Aspen, CO 81611 sbernard{at}safeminds.org
References
Thompson WW, Price C, Goodson B, et al. Early thimerosal exposure and neuropsychological outcomes at 7 to 10 years. N Engl J Med 2007;357:1281-1292. [Free Full Text]
To the Editor: Recently, I summarized several nutritional factorsthat are likely to play a large role in modulating the toxicityof the various forms of mercury.1 These factors include plasmathiol levels, zinc and selenium status, and dietary fiber intake,and they have been shown in various studies to be importantas regulating factors in both the biological transport and thedistribution of the different forms of mercury and thereforeas mediators of mercury toxicity. In addition, several geneshave been identified that are also thought to affect the toxicityof mercury. These genes are currently known to include the CPOX4polymorphism2,3 and the BDNF gene.4 The CPOX4 polymorphism isknown to lead to an atypical toxicologic response to mercuryfrom dental amalgam in up to 15% of the population.2,3 Eachof these nutritional and genetic factors represents a confoundingvariable of unpredictable magnitude that may have affected theresults of the study by Thompson et al.
James P.K. Rooney, B.Sc. Royal College of Surgeons in Ireland Dublin 2, Ireland jrooney{at}rcsi.ie
References
Rooney JPK. The role of thiols, dithiols, nutritional factors and interacting ligands in the toxicology of mercury. Toxicology 2007;234:145-156. [CrossRef][Web of Science][Medline]
Woods JS, Echeverria D, Heyer NJ, Simmonds PL, Wilkerson J, Farin FM. The association between genetic polymorphisms of coproporphyrinogen oxidase and an atypical porphyrinogenic response to mercury exposure in humans. Toxicol Appl Pharmacol 2005;206:113-120. [CrossRef][Web of Science][Medline]
Echeverria D, Woods JS, Heyer NJ, et al. The association between a genetic polymorphism of coproporphyrinogen oxidase, dental mercury exposure and neurobehavioral response in humans. Neurotoxicol Teratol 2006;28:39-48. [CrossRef][Web of Science][Medline]
Echeverria D, Woods JS, Heyer NJ, et al. Chronic low-level mercury exposure, BDNF polymorphism, and associations with cognitive and motor function. Neurotoxicol Teratol 2005;27:781-796. [CrossRef][Web of Science][Medline]
The authors reply: Bernard raises several points that we wishto address. First, children with low birth weight or seriouscongenital conditions or conditions developing in infancy thatare known to be associated with an increased probability ofneurodevelopmental problems were excluded from the study. Itwould have been difficult to distinguish the possible addedinfluence of thimerosal exposure on neuropsychological deficitsamong such children. To do so, a larger study with a differentdesign would be required.
Second, our sample was probably skewed toward higher socioeconomicstatus because participating families were members of HMOs inwhich coverage was provided by employers. In the study population,thimerosal exposure was associated with both maternal educationand maternal IQ. We therefore controlled for socioeconomic factors,maternal education, and maternal IQ in the statistical analyses.
Third, our study was less likely to include highly mobile familiesbecause the participants had to have been enrolled in the sameHMO during the first year of life and during the time of testing7 to 10 years later. These criteria ensured that we had allimmunization records during the first year of life, as wellas access to the participants' medical records during the timeof testing. This enhanced the internal validity of our studybut makes the results less generalizable to highly mobile families.
Finally, the 30% participation rate may have resulted in someunmeasured biases. Participation in the study required a substantialtime commitment from mothers and their children. Although the30% participation rate was relatively low, it was higher thanwe estimated when we planned the study. More discussion regardingparticipation and other issues can be found in the study technicalreports, available on the Web site of the Centers for DiseaseControl and Prevention.1,2
Although nutrition and genetics play an important role in neurodevelopmentaloutcomes, we believe that the factors pointed out by Mr. Rooneywere unlikely to confound the results of our study because theyare unlikely to be correlated with thimerosal exposure.
William W. Thompson, Ph.D. Centers for Disease Control and Prevention Atlanta, GA 30333 wct2{at}cdc.gov
Frank DeStefano, M.D., M.P.H. RTI International Atlanta, GA30341
References
Price C, Goodson B, Stewart G. Infant environmental exposure to thimerosal and neuropsychological outcomes at ages 7 to 10 years. Technical report. Vol. I. Bethesda, MD: Abt, 2007.
Idem. Infant environmental exposure to thimerosal and neuropsychological outcomes at ages 7 to 10 years. Technical report. Vol. II. Bethesda, MD: Abt, 2007.
Previous
