QJN. CHEM. 21/4, 551-557(1975) CLINICALCHEMISTRY, Vol. 21, No.4,1975 551 Intercomparison Program of Lead, Mercury, and Cadmium Analysis in Blood, Urine, and Aqueous Solutions Robed Lauwerys,’ Jean-Pierre Buchet,’ Harry Rods,’ Alexander Berlin,2 and Jan Smeets2 Sixty-six European laboratories participated in an inter- comparison program of lead, mercury, and cadmium analysis in blood, urine, and aqueous solutions. The ex- perimental protocol was designed in such a way that the effectof precision, experience, and analytical method could be evaluated. Forallthe analyses, the scatter of the reported results is important. The major factor in- fluencing the variability of the results is the intralabora- tory variation. Analytical methods and degree of experi- ence do not seem to have a significant influence. How- ever, with the exception of mercury determination in urine, a satisfactory intralaboratory precision is not suf- ficient to make the interlaboratory variation acceptable. Itappears that systematicerrorsareresponsible forthe high interlaboratory variation observed between “pre- cise” laboratories. Additional Koyphrases inter- and intralaboratory pert or- mance #{149} toxicology #{149} effect of technique used, exper!- ence, laboratory precision #{149} analytical error . trace ele- ments Precise and accurate determination of low concen- trations of heavy metals in body fluids is of para- mount importance for the correct evaluation of envi- ronmental exposure to these metals. Intercomparison programs have been run-and in some countries are routinely undertaken-for con- trolling the accuracy of the methods used by differ- ent laboratories for the determination of lead, cadmi- um, and mercury in blood or urine. In some programs the precision of the methods has also been evaluated. In general, these studies have indicated that the reli- ability of these measurements is unsatisfactory (1-5). For example, the most important of these studies, performed in 1970 by Keppler et al. (2) in the U.S., in which 66 laboratories participated, showed that only about half the laboratories reported results of accept- able accuracy. This study gave also an indication of the precision of blood lead measurements, because the participating laboratories received two identical 1 Unite de Toxicologie Industrielle at M#{233}dicale, University of Louvain, Cbs Chapelle-aux-Champs 4,B-1200 Brussels, Belgium. 2The Commission of the European Communities, Centre Lou- vigny, 23 Avenue Monterey, (Grand Duch#{233}) Luxembourg. Received Jan. 7, 1975; accepted Jan. 20, 1975. samples coded differently. About 43% of the labora- tories obtained similar results (10 ig/100 ml) for the two identical blood samples containing approxi- mately 65 pg/100 ml. The range of the reported values, however, was considerable: from 0 to 5600 sg/100 ml for all the laboratories and, after elimina- tion of two “outliers,” was still 30 to 120 g/1O0 ml. In 1972, the Commission of the European Communi- ties ran a limited intercomparison program (26 par- ticipating laboratories) of blood lead determination (1), which also demonstrated a great scatter of the re- sults. However, the precision of the measurements performed by each laboratory could not be evaluated, because only one sample of each blood specimen was sent to the laboratories. A more extensive intercom- parison program has therefore been undertaken to evaluate accuracy and precision of lead, cadmium, and mercury determinations performed by 66 Euro- pean laboratories on blood, urine, or aqueous solu- tion. Each laboratory received three or four identical samples, coded differently. Before the program was launched most laboratories gave information about their experience and analytical methods. The effect of precision, experience, and analytical method on the accuracy of the measurements could thus be eval- uated. MaterIals and Methods Experimental Design The following samples were sent to the participat- ing laboratories: four identical samples of human urine A, three identical samples of human urine B, four identical samples of blood C (cow blood), three identical samples of blood D (cow blood), one sample of blood E (human blood), three identical samples of aqueous solution 1, and one sample of aqueous solu- tion 2 obtained by 0.1 dilution of solution 1. Preparationof Blood, Urine,and Aqueous Solution Cow blood was obtained from a veterinary school before and after oral administration to a cow of aque- ous solutions of mercury (18 mg of Hg per day during seven days, as phenylmercuric acetate, microanalyti- cal reagent, BDH, Poole, U.K.), lead (47 mg for four