Optimization of cloned enzyme donor immunoassay cut-offs for drugs of abuse in whole blood of drivers involved in road accidents

Introduction: Immunoassay (IA) tests should be able to detect low concentrations of illegal drugs when used for the screening of drugs in drivers. False negatives should be avoided, and false positives should be reduced as far as possible. In this study, semi-quantitative results for blood samples containing illicit drugs (cannabinoids, cocaine, amphetamines/methamphetamines, opiates and methadone) obtained with cloned enzyme donor immunoassay (CEDIA), were compared with results of confirmatory analysis performed through gas chromatography-mass spectrometry (GC\u2013MS). Screening cut-off points for each class of drugs were retrospectively optimized. Methods: Whole blood samples from drivers involved in road accidents in the period from January 2013-December 2017 were analyzed with CEDIA (4200 samples). Confirmatory analyses were performed through (GC\u2013MS) on: (i) all samples with screening concentrations above 1 ng/ml for at least one drug (positive screening results); (ii) 800 samples with screening concentration lower than 1 ng/ml (negative screening results). Recommended per se limits in relation to driving under the influence of drugs were set as fixed values. Sensitivity, specificity, positive and negative predictive values were evaluated by contingency tables and compared to ROC-analysis in order to obtain ideal screening cut-offs. Results: CEDIA results were available for 4200 blood samples and 1172 positive screening results were found. Among these, 1008 confirmation analysis were obtained through GC\u2013MS. Optimized screening cut-offs obtained through ROC analysis were as follows: 8.0 ng/ml for THC; 5.5 ng/ml for THC-COOH; 21.1 ng/ml for cocaine; 6.9 ng/ml for benzoylecgonine; 33.1 ng/ml for opiates; 61.6 ng/ml for amphetamines; 5.0 ng/ml for methadone. Using these cut-offs, sensitivity was above 97% for THC-COOH, cocaine, benzoylecgonine, amphetamines, opiates and methadone, and 92% for THC; specificity was above 90% for cocaine, benzoylecgonine, amphetamines, opiates and methadone, 80% for THC and 89% for THC-COOH; negative predictive value was above 99% for all drugs and metabolites. Conclusion: Previous studies have shown that CEDIA tests are useful for preliminary screening of serum and urine. Its implementation in whole blood is of primary importance for the assessment of impaired driving, since the per se limits of many European countries refer to whole blood, and preparation of the serum and/or the collection of urine is not always possible in the hospital emergency department, where blood samples are withdrawn. Our study shows that CEDIA tests on whole blood permit the definition of cut-off values with optimal sensitivity and negative predictive values for all analytes (near to 100%), including very good specificity