Results of a multilaboratory checksample analysis program for carbon monoxide and cyanide

A blind check sample analysis program, designed to evaluate global performance in the determination of fire-related analytes (carbon monoxide and cyanide), was implemented and processed using 10 field forensic toxicology laboratories. Each participant received, on a quarterly basis, a set of four calibrators and a set of four unknown check samples. Each participant was asked to determine the presence and quantity of carbon monoxide (carboxyhemoglobin in percent saturation of hemoglobin units) and cyanide (in mg/L units), using the calibrators for reference. The results were to be reported in an anonymous manner. Prior to and during the analysis period all batches of calibrators and check samples were validated for analyte stability using gas chromatography, CO- oximetry, and ion chromatography. The performance of the laboratories in the analysis of carboxyhemoglobin was evaluated over a concentration range of 3 to 40% saturation yielding a correlation with validation concentrations of 0.986. In some cases, false negatives and false positives were reported. The performance of laboratories in the analysis of cyanide was evaluated over a concentration range of 0.1 to 2.5 mg/L. Laboratory results yielded a correlation of 0.443 with validated concentrations. False negatives and false positives were also reported. Based upon the correlation estimates, it was concluded that forensic toxicology laboratories are likely to generate reliable carboxyhemoglobin data.