Factors affecting the accuracy of blood alcohol determinations in clinical and forensic laboratories in Italy: Results of a 7-year proficiency testing study

The results of a 7-year proficiency study involving Italian laboratories measuring blood alcohol concentration (BAC) are reported. Two blood samples spiked with known amounts of ethanol were sent for analysis twice per year to approximately 50 laboratories across Italy. The blood samples were prepared according to ISO standards, which included optimized transport and storage conditions to ensure ethanol stability. Participants completed an online questionnaire detailing their analytical results along with other analytical details, including the use of GC-FID or GC-MS, storage temperature, time between sample delivery and analysis, reference materials, replicate determinations, and volume of blood aliquot. The participants also confirmed whether or not they were accredited and what the measurement uncertainty was for determination of BAC. The results from analysis of 1171 blood samples (after elimination of outliers) were evaluated by calculating z-scores using robust statistical methods. The inter-laboratory standard deviation (SD) increased with blood alcohol concentration (r = 0.94, p < 0.001). Notably, a significant reduction in robust SD was observed between 2017 and 2023 (p = 0.01). The use of matrix-matched certified reference materials gave more accurate results with mean z-scores closer to zero. The choice of analytical method, storage conditions, time delays, and laboratory accreditation did not significantly influence z-scores. This study revealed that regular participation in inter-laboratory proficiency programs did contribute to improving analytical performance. The choice of certified reference material was the principal determinant of the accuracy of the results.