Urine testing for cocaine abuse: metabolic and excretion patterns following different routes of administration and methods for detection of false-negative results.

Although cocaine is typically the second-most identified drug of abuse in drug-testing programs, there is surprisingly little quantitative information on excretion patterns following different routes of administration. This report details the urinary excretion and terminal elimination kinetics for cocaine and eight metabolites [benzoylecgonine (BZE), ecgonine methylester (EME), norcocaine (NCOC), benzoylnorecgonine (BNE), m-hydroxy-BZE (m-HO-BZE), p-hydroxy-BZE (p-HO-BZE), m-hydroxy-COC (m-HO-COC), and p-hydroxy-COC (p-HO-COC)]. Six healthy males were administered approximately equipotent doses of cocaine by the intravenous (IV), smoking (SM), and inhalation (IN) routes of administration. Urine specimens were collected for a minimum of three days after drug administration, screened by immunoassay (EMIT and TDX, 300 ng/mL), and analyzed by GC-MS. Mean Cmax values were generally as follows: BZE > EME > COC > BNE approximately p-HO-BZE > m-HO-BZE > m-HO-COC > NCOC > p-HO-COC. Elimination half-lives for cocaine and metabolites were generally shorter following s.m., intermediate after i.v., and longest following i.n. administration. m-HO-BZE demonstrated the longest half-life (mean range 7.0-8.9 h), and cocaine displayed the shortest (2.4-4.0 h). Mean detection times were extended progressively by lowering cutoff concentrations. The maximum increases were approximately 55% at 50 ng/mL for the TDx assay (e.g., the detection time for the last consecutive positive changed from 32.8 h to 50.6 h for i.v. cocaine) and up to 39% for GC-MS at a cutoff concentration of 40 ng/mL (e.g., the detection time for the last consecutive positive changed from 34.8 h to 48.1 h for i.v. cocaine). Sensitivity, specificity, and predictive values for EMIT and TDx were comparable at the 300-ng/mL cutoff concentration; but at lower cutoff concentrations, predictive values of positive results for TDx were diminished indicating a higher risk of false-positive results, that is, positive results that failed to meet administrative cutoff criteria. Detection of positive results was significantly enhanced through the use of the "Zero Threshold Criteria Method", a method developed by the authors to differentiate false-negatives from true-negatives. The method was based on establishing mean immunoassay response (MIR) baselines and variance (SD) in assays of drug-free specimens. Arbitrary thresholds (MIR + 0.5 SD, MIR + 1 SD, MIR + 2 SD) were utilized to evaluate all negative specimens. Apparent true positives were identified by the presence of BZE at or above 40% GC-MS cutoff concentrations. With these criteria, up to 111 false-negative specimens were confirmed as true-positive specimens; this was in addition to the 208 true positives detected at recommended cutoff concentrations. This represents a 50% increase in positive detection rates through the use of this methodology. Such methodology is recommended for further evaluation by drug-testing programs for enhancement of positive detection rates and as an alternative to creatinine testing for dealing with dilute specimens that test negative by initial tests, but contain quantifiable concentrations of drugs of abuse.