Determine whether the outcome of immunochemical screening for synthetic cannabinoids is affected by the presence of drugs frequently abused by prisoners.

Current surveillance of drug and drug abuse in prisons does not routinely use synthetic cannabinoid (SC) screening, so their possible occurrence may go undetected. Since the thresholds for the detection of SC must be low (ng/mL and less) due to their high efficiency, theoretically other substances present in the urine at orders of magnitude higher concentrations may influence the results of the SC screeqning.

Urine samples of 450 prisoners from detention prisons in Prague were tested with immunochemical tests for the presence of buprenorphine, benzodiazepines, fentanyl, methamphetamine-amphetamine-ecstasy, morphine, cannabinoids, and tramadol. The results were in binary format, 1 or 0 (Dataset 1). These samples were further tested for the presence of SC [K2 SC–1 (JWH-210); K2-II SC–2 (UR-144); K2-III SC–3 (AB-PINACA); all Thermo Scientific] on the Indiko Plus (Thermo Scientific) instrument. The format of their results are numeric values (Dataset 2). The statistical data evaluation and analysis were done with Statistica software.

The dataset 1 was analyzed for distribution of zero, single, double, triple etc. occurrence of positive and negative screening results. SC-1, SC-2, and SC-3 values of dataset 2 were analyzed for the possible correlation, and the principal component analysis (PCA) was applied to evaluate the data structure, where the SC-1 and SC-2 data were found more intercorrelated in contrast to SC-3 data which do not correlate with the previous SC-1 and SC-2 values.

Furthermore, the possibilities of dependencies between the two datasets 1 and 2 were studied. First, the multiple linear regression analysis (MLR) was used, where the dependent variables were the values SC-1, SC-2 and SC-3 and the binary values of dataset 1 were used as independent variables. A statistically significant (but close to critical value) relationship was found between SC-1, SC-2 and SC-3 and positive results for the test from dataset 1. No dependence was found between the occurrence of other substances and the SC test values.

As an additional statistical method for determining possible relationships, the logistic regression analysis (LogReg) was used. Binary values corresponding to the presence or absence (1 or 0) of the test substance (dataset 1) were used as a dependent variables and dataset values 2 (SC-1, SC-2, SC-3) were used as independent ones. As with MLR, a statistically significant dependence (also close to the critical value) was found between the presence or absence of methamphetamine-amphetamine-ecstasy (dataset 1) and the SC values of dataset 2.

According to the PCA, a correlation was found between the values of the SC-1 and SC-2 test results.

Although correlations were found among the methamphetamine-amphetamine-ecstasy incidence and the SC-1, SC-2 and SC-3 test result values, the statistical significance parameters for both the MLR and the LogReg analyses were very close to their critical values. In addition, the results of the analyses were rather sensitive to the choice of the regression dataset.

Based on these findings, it is not possible to unequivocally answer the question whether the presence of commonly abused substances affects the results of immunochemical urine SC screening. To make a clear decision, it is necessary to statistically evaluate a larger set of data.