To the Editor.
—The article by Dr Needleman and collegues1 is based on analyzing bone lead by XRF, a method of unproven clinical reliability. An XRF registry has recently been developed by the lead-testing industry in response to increasing reports of poor and highly variable data quality of XRF analyses for industrial use. Analyses for medical studies should have more exacting requirements.Needleman et al stated in their "Methods" that bone lead and psychological studies were done twice, 2 years apart. They indicated that the bone lead results of the first study were not usable (the lead x-ray levels from even the high-lead phantoms were obscured), although in the second study they were more readily "detectable." What does this mean in terms of specificity, sensitivity, and reproducibility? Further, why did the authors only compare the neuropsychological results of the first study with the XRF results of the second study?