Background: Previous reports revealed poor performance in identifying drugs of abuse users through first-level workplace drug testing (WDT), based on urine samples. In a cross-sectional study, we evaluated: (i) the effect of creatinine normalization of drug values from diluted urine samples (creatinine levels ≤ 20 mg/dL) on the prevalence of drug users; (ii) the independent procedure-related predictors of positivity and dilution.
Methods: Workers' urine samples were collected at the workplace or at our certified laboratory between 2008 and 2012. All samples were analysed for drugs of abuse by immuno-enzymatic method in our laboratory, according to the Italian WDT law. Detectable drugs of abuse concentrations lower than the positive cutoff values were normalized based on mean levels of urinary creatinine. Detectable concentrations of drugs were confirmed by GC/MS. Multivariate logistic regression was used to detect independent procedure-related predictors of positive and diluted urine samples.
Results: Of the 3080 urine samples screened, 51 (1.7%) were found positive for some drugs of abuse (26 cannabinoids and 16 cocaine) and 116 (3.8%) were diluted. Seventeen out of 23 diluted urine samples with detectable concentrations of cannabinoids or cocaine were found positive after urine creatinine normalization and GC/MS confirmed both negative and positive results. This increased the percentage of positivity for cannabinoids and cocaine from 1.35% to 2.09% (+55%, p=0.0005), which is closer to the expected prevalence of drug users based on Italian self-reported surveys. Collection of samples in the laboratory was an independent predictor of positivity (OR=2.33, 95%CI 1.27-4.28) and diluted urine sample (OR=1.65, 95%CI 1.04-2.61).
Conclusions: Efficacy of first-level WDT could be improved by well-controlled pre-analytical procedures and urine creatinine normalization of detected concentrations of drugs of abuse.