Studies of lead exposure and the developing central nervous system: a reply to Kaufman

Kaufman's critique of the lead-cognition hypothesis is a reiteration of well worn and weary claims raised many times in the past, primarily by spokespersons for the lead industry. They have been fully rebutted in the literature. The essence of these claims is that those studies showing an association between lead and IQ are flawed by uncontrolled confounding, multiple comparisons, and errors in measurement. Any effect of lead, Kaufman asserts, if present, is small. This response examines each of these issues and shows that they lack substance. Lead's negative impact on IQ persists in most modern studies after confounding has been controlled in many different statistical models. At least three metaanalyses have confirmed an effect of lead at low dose. Animal studies in which lead is given systematically, and the issue of confounding thereby avoided, demonstrate an unequivocal lead effect at similar doses to the human studies. The criticism of multiple comparisons similarly does not withstand examination. Measurement errors are nonsystematic and nondifferential. They are, therefore, null biasing. The actual size of the lead effect has been shown to be substantial, and to be found most prominently at the ends of the distribution. Kaufman says that lead requires study of diverse dimensions of intellect, but he restricts his scope to a sample of studies of lead and IQ, ignoring recent high quality studies that show a clear lead effect, and in those studies that he critiques he ignores data that contradict his position. His article adds nothing to the dialogue on lead neurotoxicity.