The Triage Panel for Drugs of Abuse (DOA) was evaluated for detection of phencylidine (PCP), amphetamines (AMPH),

opiates (OPI), tetrahydrocannabinol (THC), the cocaine metabolite benzoylecgonine (BE), and barbiturates (BARB) in urine. This assay was compared with Syva EMIT. The comparisons were conducted on 606 positive and 325 negative samples. When there was a sufficient volume of sample available for retesting, positive and negative samples with discordant results between these two screening assays were confirmed by quantitative gas chromatography/mass spectrometry (GC/MS). The percent agreement between the two assays for positive samples ranged from 93 to 100%. For negative samples, the agreement ranged from 95 to 100%. For AMPH, 19 out of 27 discordant samples of urine that were positive for EMIT and negative for Triage DOA contained total amphetamine and methamphetamine concentrations of less than 1000 ng/mL by GC/MS. For THC, seven urine samples that were negative for Triage DOA and positive for EMIT contained 9carboxy THC concentrations at greater than 15 ng/mL by GC/MS. For BARB, three samples that were negative for Triage DOA and positive for EMIT contained barbiturates levels greater than 300 ng/mL. Two of these three samples contained phenobarbital below the concentration that produces a positive result for Triage DOA. For the majority of the urine samples studied, however, the Triage DOA produced identical results to a commercial-instrument-based immunosssay. Numerous methods are available for screening drugs of abuse in urine, including radioimmunoassay (RIA) (e.g. roche abuscreen, Nutley, nj) enzyme immunoassay (e.g. syva EMIT, palo alto, CA), and fluorescence polarization immunoassay (e.g. abbott TDx, abbott park, IL) (1-4). These assays require instruments such as a gamma-counter for RIA and analyzers such as the ETS, ADx or Cobas Mira for non isotopic analysis. Recently, newer assays have been developed that allow visual detection of the label and therefore do not require any instrumentation. Examples of these assays include the on trak assay (roche), which is based on latex agglutination and the accu-pinch (hycor biomedical, garden Grove CA), which is based on enzyme immunoassay. These assays are rapid and are designed for immediate or emergency testing of selected drug classes for clinical toxicology puposes. Only one drug class however can be analyzed per device, limiting its utility for rapid screening. The triage panel for drugs of abuse (DOA) (biosite diagnostic, san diego CA) is a multi-test screening panel that is designed to provide simultaneous and discrete visual detection of seven drug classes in approximately ten minutes (6). Five of these classes are listed by the department of health and human services (DHHS) for workplace drug testing : cocaine metabolite, benzoylecgonine (BE) opiates (OPI), amphetamine (AMPH), tetrahydrocannabiol (THC) and phencyclidine (PCP). Theother two classes, barbiturate (BARB) and benzodiasepins, are not currently listed on these guidelines. The assay was configured to meet the current DHHS screening cutoffs (300, 300, 1000, 100 and 25 ng/ml respectively). The barbiturate and bonzodiasepin cutoffs were each fixed at 300 ng/ml. Using clinical urine samples, the analytical accuracy for six of these assays on the triage DOA was evaluated against the EMIT assay. The comparison of results for benzodiazepine was conducted under different criteria and results will be reported separately. Positive samples and those with discordant results between the two screening methods were confirmed by quantitative gas chromatography/mass spectrometry (GC/MS) with ion trap and quadrupole analyzers.

Urine samples were obtained from several sources and tested for PCP and AMPH at Hermann hospital, for OPI and THC at the Nichols institute for substance abuse testing and for BE and BARB at the johns Hopkins hospital. Urine specimens from hospital patients were largely collected for clinical toxicology purposes, whereas those obtained from reference laboratories were for workplace drug testing. Using triage DOA, we assayed a total of 95 OPI samples, 100 each of PCP, THC, BE, and BARB samples, and 111 AMPH samples that were positive using EMIT. Also assayed were 100 each of OPI, THC,BE and BARB samples and 125 PCP and AMPH samples that were negative using EMIT. Samples were collected without preservatives and stored at either 40C if assayed within a few days or at -200C if long term storage was required before analysis. Urine specimens containing THC were assayed prospectively using both EMIT and triage DOA because THC samples can lose 60-100% of their THC concentrations when stored in plastic vials.

Assays The urine specimens were initially screened with the EMIT d.a.u assay (for AMPH, OPI, THC, BE and BARB) using either the ETS (Syva) or Hitachi 717 analyzers (Boehringer Mannheim Diagnostics, Indianapolis, IN) for PCP the EMIT 700 assay was used. A cutoff of 25 ng/mL for PCP was used for both screening and confirmation. The monoclonal antibody EMIT assay was used for screening amphetamines. Cutoff of 1000 and 500 ng/mL were used for screening and confirmations, respectively. For OPI and BARB, a cutoff of 300 ng/mL was used for both screening and confirmation. For BE, the screening and confirmations cutoff were 300 and 150 ng/mL respectively. The recommendations of the manufacturers for both Triage DOA and EMIT were followed throughout. Triage DOA uses an immunochemical technique known as ASCENDTM Multimmunoassy. A schematic representation of this process is shown in Figure 1. This monoclonal antibody system simultaneously analyzes multiple analyzes in a competitive binding mode, using a solution-phase reaction followed by a solid-phase reaction. The solution-phase reaction incorporates there pre-dispensed reagent beads (one each for the antibodies drug conjugates, and buffer) that are reconstituted by the addition of a urine specimen to the reaction well. The antibody bead contains monoclonal antibodies for the seven targeted drug classes. The conjugate bead contains a representative drug of each class that is conjugated to a colloidal gold particle (Figure 1a). When a drug-free urine sample is added to the beads, all of the antibodies will bind to the drug conjugates, leaving on exposed conjugated drugs. If a urine sample contains one or more drugs at concentrations above the threshold limit, the antibodies bind to both the conjugated and the free drug (Figure 1b), leaving some drug conjugates unbound to antibodies (Figure 1c).

After a 10-min incubation, the reaction mixture is transferred to the detection area containing a nylon membrane with monoclonal antibodies immobilized in seven discrete detection zones. From the single reaction mixture, the multiple drugs are detected when they are applied to detection area. Samples that have no exposed drug conjugates (negative samples) pass through the membrane and do not bind to the monoclonal antibodies immobilized on the membrane. Samples that have exposed drug conjugates (positive samples) are captured by the second immobilized antibody, producing a distinct color bar at the zone corresponding to each positive drug (Figure 1d). After a wash solution has passed through the membrane, the presence of color bars are read visually. Only samples at or above the cutoff level produce a detectable color response; samples are considered positive as long as the built-in test valid color bar appears and a color bar is absent in the test invalid zone. A draft of the Health Care Finance Administration (HCFA) guidelines for laboratory inspectors has indicated that instrument or procedural control checks may be to used to meet the requirement for testing negative and positive controls daily. Nevertheless, external quality control materials (Hycor Biomedical, Garden Grove, CA) were also run at three levels (typically, there are 20 samples).

All positive samples were confirmed by quantitative GC/MS. The analytical methods for AMPH and PCP were described previously using the ITS40 ion trap MS (Finnigan, san Jose, CA) (9) in the full scan mode. AMPH samples were derivatized with heptafluorobutyric anhydride. For the other analytes, a quadrupole MSD analyzer (Hewlett-Packard, Palo Alto, CA) was used in the selected ion monitoring mode. OPI were derivatized with acetic anhydride, THC with methyliodide, and BE with iodopropane. PCP and BRAB were assayed underivatized.

Evaluations The precision of each test on the Triage DOA was assessed by assaying replicates of 10 blank urine samples spiked at concentrations that bracket the cutoff. Urine concentrations of PCP were varied from 20 to 30 ng/mL in 2.5-ng/mL intervals, AMPH concentrations were varied from 800 to 1200 ng/mL in 100-ng/mL intervals, THC concentrations were varied from 80 to 120 ng/mL in 10-ng/mL intervals, and OPI, BE, and BARB concentrations were varied from 240 to 360 ng/mL in 30-ng/mL intervals. A BE concentrations of 150 ng/mL was also assayed. The agreement of results between Triage DOA and EMIT was also studied on a total of 1256 positive and negative urine samples.

Results The results of the precision testing on the Triage DOA are shown in Table I. The PCP, AMPH, THC, BE and BARB tests produced a positive results 100% of the time when tested at the cutoff concentrations. OPI produced a positive result 70% of the time when tested at the cutoff concentration. The PCP,

AMPH, OPI and BARB tested produced a negative result when tested using control specimens containing drugs at 80% of the cutoff concentration, while the THC and BE test produced a positive result 80% and 100% of the time, repectively.

Table II summarizes the agreement between triage DOA an EMIT. For PCP and BE, complete agreement for all positive and negative samples was observed. The GC/MS results for positive PCP samples ranged from 30 to 1487 ng/mL, while positive BE samples ranged 310 to 170,000 ng/mL. None of the negative samples were tested by GC/MS Considerable differences were seen between the two screening assays for AMPH, as summarized in table II. A total of 31 out of 236 positive and negative samples tested for AMPH exhibited differences between triage DOA and EMIT. There was insufficient urine for GC/MS confirmation in four of these samples. In 24 of the remaining 27, combined amphetamine and methamphetamine concentration by GC/MS were less than the 1000 ng/ml cutoff value used for screening (range 0-924 ng/mL). 19 out off these 24 specimens (group 1, table III) were positive using EMIT and negative using triage DOA, two (group 2) were negative by EMIT and positive by triage DOA and three (group 3) were positive using both assays. Other structurally related drugs such as sympathomimetic amines may have been present in these urines, however, the GC/MS analysis was not targeted to detect for the presence of these drugs.

The remaining three urines from the 31 discordant samples (group4) had total amphetamine results by GC/MS that were greater than the 100 ng/ml screening cutoff (range 1026-2133 ng/ml). each of these were positive using EMIT and negative using triage DOA. GC/MS was performed on sample 3 of this group following chiral derivatization (10). This sample contained high L-methamphetamine concentrations. These L-isomers are present in some commercially available antihistamines (e.g. vicks inhaler) (11) and are not listed on the drug enforcement agencys controlled substances list. Thus, for this sample, triage DOA correctly indicated that the result for the D-isomers of amphetamine was below the 100 ng/ml cutoff, while EMIT and GC/MS (using conventional derivatization) produced results that suggested that these isomers were greater than this cutoff. This observation is further supported by the manufactures claim that L-amphetamine and L-methamphetamine do not react in the test at 10000 and 32000 ng/ml respectively.

The comparison between EMIT and triage DOA for OPI showed good agreement between the two methods. All positive samples by EMIT , ranging from 349 to 104940 ng/ml, were also positive using triage DOA. In three of these samples, the GC/MS result for codeine and morphine was below the 300 ng/ml screening cutoff. GC/MS analysis revealed the presence of hydrocodone in another (table IV). For THC , the comparison of EMIT and triage DOA produced 100% agreement among negative samples and 93% agreement among positive sample ranging from 18-1486 ng/ml. for the seven samples

positive by EMIT and negative by triage DOA, all had a GC/MS concentration that was above the 15 ng/ml cutoff for 9-carboxy THC metabolite by GC/MS although, only one was greater than the triage DOA cutoff 100 ng/ml for total THC metabolites. The results between the screening assays may have variable reactivity to the numerous THC metabolites, whereas the confirmation assay is directed at the 11-nor-A9-THC-9-carboxylic acid metabolite. These seven sample were also tested on the abbott ADx (north Chicago,iL) and were also found to be negative at a cutoff concentration of 100 ng/ml (data not presented)

The comparison of EMIT and triage DOA for the barbiturates also showed 100% agreement among negative samples and 97% agreement among positive samples ranging from 330 to greater than 1200 ng/ml. as shown in table IV, two of the discordant results were urine samples confirmed by GC/MS to contain 370 and 400 ng/ml of Phenobarbital. The cutoff concentration of 300 ng/ml for the triage DOA barbiturate assay was adjusted to be sensitive to secobarbital. It was experimentally determined that Phenobarbital will only begin to produce positive result when the concentration exceeds 700 ng/ml. the third sample was retested using the TDx and contained barbiturate reactivity greater than 1200 ng/ml, wich was not detected by triage DOA. Insufficient urine specimen remained for resolution of this discordant result.