With the development of L X-ray fluorescence (LXRF) to measure cortical bone lead directly, safely, rapidly, and noninvasively, the present study was undertaken to a) evaluate LXRF as a possible replacement for the CaNa2EDTA test; b) quantify lead in tibial cortical bones of mildly to moderately lead-toxic children before treatment; and c) quantify lead in tibial cortical bones of lead-toxic children sequentially following one to two courses of chelation therapy. The clinical research design was based upon a longitudinal assessment of 59 untreated lead-toxic children. At enrollment, if the blood lead (PbB) was 25 to 55 jg/dL and the erythrocyte protoporphyrin (EP) concentration was 2 35 glg/dL, LXRF measurement of tibial bone lead was carried out. One day later, each child underwent a CaNa2EDTA provocative test. If this test was positive, lead-toxic children were admitted to the hospital for 5 days of CaNa2EDTA therapy. These tests were repeated 6 weeks and 6 months after enrollment. Abatement of lead paint hazards was achieved in most apartments by the time of initial hospital discharge.The LXRF instrument consists of a low energy X-ray generator with a silver anode, a lithium-doped silicon detector, a polarizer of incident photons, and a multichannel X-ray analyzer. Partially polarized photons are directed at the subcutaneous, medial mid-tibial cortical bone. The LXRF spectrum, measured 900 from the incident beam, reveals a peak in the 10.5 KeV region, which represents the lead La line. The effective dose equivalent using tissue weighting factors according to guidelines of the National Council on Radiation Protection and Measurements (1989), was 2.5 FSv. The reproducibility of replicate LXRF measurements, including the day-to-day variation of the instrument, in 26 lead-toxic children, after repositioning the instrument within 5 cm of the first LXRF measurements, was ±9.2 (9 5% confidence limits). For an overlying tibial skin thickness of 5 mm, the minimum detection limit was 7 Ag of lead/g (wet weight) at the 95% confidence interval.Based upon a discriminant analysis, 90% of lead-toxic children were predicted correctly as being CaNa2EDTA-positive or CaNa2EDTA-negative. Using LXRF and PbB values to predict CaNa2EDTA outcomes, the specificity and sensitivity of these two predictors were 86 and 93%, respectively. In a significant fraction of CaNa2EDTA-positive and CaNa2EDTA-negative children, cortical bone lead values were similar to lead concentrations measured via bone biopsy in normal adults and lead workers in industry. By 24 weeks after enrollment, PbB, EP, and urinary lead/EDTA ratios were similar in all groups. The most dramatic decreases in net corrected photon counts by LXRF occurred in children treated twice. Mean values of cortical bone lead by LXRF at 24 weeks in all three groups of children were similar to the mean concentration in untreated CaNa2EDTA-negative children at enrollment but still three to five times greater than those measured in the tibia or whole teeth of normal European children using...