Stephen J.LurieMD, PhD, Senior EditorIndividualAuthorJody W.ZylkeMD, Contributing EditorIndividualAuthor
In Reply: Drs Watts and Mercola and Drs Kaminski and Glade acknowledge that there is variability among laboratories performing hair mineral analysis. However, they disagree with our conclusion that hair mineral testing is unreliable.
Watts and Mercola suggest that including an uncertified and illegally operating laboratory (Laboratory B) unfairly introduced greater variability into the results. Laboratory selection was based on volume of analyses and available hair as shown in Table 1 of our article. Steindel and Howanitz,1 in their accompanying Editorial, calculated a mean coefficient of variation (coefficient of variation = SD/mean) across laboratories of 103.5%. Excluding Laboratory B, and restricting the analysis to the 19 elements reported by all laboratories, the mean coefficient of variation for the 5 remaining laboratories is 68%, still above the cited normal limits of 10% for common analytes and 15% to 50% for more exotic ones. The uncertified laboratory was selected prior to learning that it was falsely advertising certification. There is no Clinical Laboratory Improvement Act certification for hair analysis, so the certification advertised by all the laboratories must be for some other application.
Seidel S. Accuracy of Hair Mineral Analysis—Reply. JAMA. 2001;285(12):1576-1578. doi:10.1001/jama.285.12.1576