Comparison of automated determination of phosphatidylethanol (PEth) in dried blood spots (DBS) with previous manual processing and testing

Abstract

Phosphatidylethanol (PEth) is a sensitive and specific biomarker of alcohol consumption in the prior 2–3 weeks. Standard, manual PEth testing using dried blood spots (DBS) is a multi-step time-consuming process. A novel, automated processing and testing method has been developed to decrease DBS processing and testing time. We conducted automated testing, using regiosimerically pure PEth reference material, on randomly selected DBS which had previously been tested via manual methods and then stored for 3–6 years at at −80°C, to compare the results (PEth 16:0/18:1 homologue). We chose samples for re-testing using categories found in the literature as follows: (1) PEth <20 ng/mL; (2) PEth 20–200 ng/mL; (3) PEth >200–1000 ng/mL; (4) PEth >1000 ng/mL. We calculated agreement between the categories using the weighted kappa statistic (n=49 DBS). We quantified agreement between continuous measures using the intraclass correlation coefficient (ICC), and further described the relationship between variables using Spearman correlation. The median PEth result was 155 ng/mL (interquartile range [IQR]: 1–1312 ng/mL) via automated methods and 98.8 ng/mL (IQR: 10.2–625.0 ng/mL) via manual methods. The weighted kappa comparing the automated to manual PEth results was 0.76 (95% Confidence Interval (CI): 0.66–0.86). The ICC was 0.69 (95% CI: 0.54–0.79), and the Spearman correlation was 0.98 (95% CI: 0.95–0.99). While the new methods yielded somewhat higher PEth values, we found good to excellent agreement between clinically relevant PEth categories. Automated DBS processing and testing using new reference standards are promising methods for PEth testing.