Drug Testing and Analysis
Research article Received: 27 May 2014
Revised: 14 July 2014
Accepted: 15 July 2014
Published online in Wiley Online Library
(www.drugtestinganalysis.com) DOI 10.1002/dta.1702
Performance characteristics of an ELISA screening assay for urinary synthetic cannabinoids Eliani Spinelli,a,b Allan J. Barnes,a Sheena Young,a Marisol S. Castaneto,a Thomas M. Martin,c Kevin L. Klettec and Marilyn A. Huestisa* Synthetic cannabinoids are marketed as legal alternatives to cannabis, as routine urine cannabinoid immunoassays do not detect synthetic cannabinoids. Laboratories are challenged to identify these new designer drugs that are widely available and represent a major public health and safety problem. Immunoassay testing offers rapid separation of presumptive positive and negative specimens, prior to more costly and time-consuming chromatographic confirmation. The Neogen SPICE ELISA kit targets JWH-018 N-pentanoic acid as a marker for urinary synthetic cannabinoids. Assay performance was evaluated by analyzing 2469 authentic urine samples with the Neogen immunoassay and liquid chromatography-tandem mass spectrometry (LC-MS/MS). Two immunoassay cut-off concentrations, 5 and 10 μg/L, classified samples as presumptive positive or negative, followed by qualitative LC-MS/MS confirmation for 29 synthetic cannabinoids markers with limits of detection of 0.5–10 μg/L to determine the assay’s sensitivity, specificity and efficacy. Challenges at ±25% of each cut-off also were investigated to determine performance around the cut-off and intra- and inter-plate imprecision. The immunoassay was linear from 1 to 250 μg/L (r2 = 0.992) with intra- and inter-plate imprecision of ≤5.3% and 98% and sensitivity >95% when evaluated against LC-MS/MS confirmation results. Our aims were to optimize assay performance and validate the Neogen SPICE ELISA kit as a screening method for the detection of synthetic cannabinoids in urine. Presumptive positive (1409) and negative (1060) urine specimens (n = 2469) were compared to the most extensive qualitative LC-MS/MS confirmation assay for 29 synthetic cannabinoid markers to determine assay performance characteristics at multiple cut-offs.
Materials and methods Reagents and materials Each Neogen SPICE ELISA kit contained antibody coated 96-well microtitre plates, drug enzyme conjugate, tetramethylbenzidine (TMB) substrate, 1 N H2SO4 stop solution, enzyme immunoassay (EIA) buffer containing bovine serum and preservative, concentrated wash buffer (10x), and qualitative positive (100 μg/L) and negative control solutions prepared in human synthetic urine. Kit components were stored at 4 °C until analysis, and the assay was performed according to manufacturer’s instructions. Samples were diluted 1:5 with EIA buffer in uncoated 96-well microtitre plates (Greiner, Monroe, NC, USA) before analysis. Blank urine was collected from drug-free volunteers and evaluated to be negative prior to use in the preparation of calibrators and quality control samples for performance challenges. Calibrators, controls and performance challenges JWH-018 N-pentanoic acid metabolite was purchased from Cayman Chemicals (Ann Arbor, MI, USA) as 1 mg/mL ampules. Stock standard solutions of cut-off calibrators and controls were prepared from different ampules by diluting contents with appropriate volumes of methanol. Cut-off calibrators (5 and 10 μg/L) were prepared by fortifying authentic drug-free urine. Working performance challenges at ±25% of the 5 and 10 μg/L cut-offs were prepared by fortifying drug-free urine with stock control solution. Specimens Authentic urine specimens (n = 20 017) were randomly collected from US Department of Defense (DoD) service members stationed
wileyonlinelibrary.com/journal/dta
worldwide to determine the prevalence of synthetic cannabinoids intake from July 2011 through June 2012. Per DoD routine drug testing, specimens screened negative for common drugs of abuse (cannabinoids, cocaine, amphetamines, phencyclidine, and opiates) prior to shipment for synthetic cannabinoid testing. Of 20,017 specimens initially analyzed with the Randox Drugs of Abuse V biochip array technology for synthetic cannabinoids, 1409 presumptive positive and 1060 presumptive negative specimens were selected for further testing. Specimens were stored at room temperature (RT) prior to DoD testing and until synthetic cannabinoid initial testing. Specimens (n = 2469) were then stored at 4–7 °C before Neogen ELISA determination and LC-MS/MS confirmation.
ELISA instrumentation Qualitative, automated synthetic cannabinoids analysis was performed on a Freedom EVO 100 configured with a microtitre plate washer and reader (Tecan Group, San Jose, CA, USA). All samples, reagents and microtitre plates were equilibrated at RT. Per manufacturer’s guidelines, a 1:5 sample dilution was performed by adding 160 μL EIA buffer to uncoated microtitre plate wells and 40 μL blank, cut-off, in-house performance controls or unknown specimens to the appropriate wells, followed by 30-s plate shaking. Positive and negative Neogen controls were processed without dilution; 20 μL diluted samples or undiluted controls were added to the coated microtitre plate; 100 μL enzyme conjugate was added and the mixture incubated at RT for 45 min. The plate was washed five times with diluted wash buffer (300 μL), manually inverted and slapped dry to remove residual liquid. TMB substrate (100 μL) was added and, after a 30-min incubation, 100 μL acid stop solution was added and absorbance (450 nm) immediately measured with a plate reader.
Method development Although analysis was performed according to manufacturer’s instructions, additional experiments were performed. We observed that the length of time reagents were in contact with antibodies during the competition phase contributed to plate drift; liquid handling parameters and plate washing instructions were optimized to ensure consistent incubation times. Although no cut-off was specified, several JWH-018 N-pentanoic acid calibrators (0, 1.0, 2.5, 4, and 6 μg/L) prepared in synthetic urine are provided, allowing laboratories to determine limits that best suit their drug testing programs. We prepared matrix-matched cut-off solutions in authentic urine at 5 and 10 μg/L that were analyzed in duplicate at the beginning, middle, and end of each plate, in columns 1, 6, and 12. Negative and positive Neogen controls were run in singulate at the beginning and end of each plate, while fortified urine performance challenges at ±25% of each cut-off were evaluated in the middle of the plate. Urine controls at 3.75 and 6.25 μg/L assessed performance of the 5 μg/L cut-off and 7.5 and 12.5 μg/L controls monitored performance of the 10 μg/L cut-off. A representative plate layout is presented in Supplemental Figure S1. Strategic placement of calibrators, controls and performance challenge specimens allowed optimal evaluation of assay performance criteria. Specimens were presumptively positive when their absorbance at 450 nm was less than or equal to the averaged cutoff of the first pair of calibrators. Similarly, a negative result was indicated when the specimen absorbance was greater than the average of duplicate cut-off calibrators in the first column.
Copyright © 2014 John Wiley & Sons, Ltd.
Drug Test. Analysis (2014)
Drug Testing and Analysis
Performance characteristics of an ELISA screening assay for SC Assay performance criteria The method was validated by determining sensitivity, linearity, imprecision (intra- and inter-assay), plate drift, cross-reactivity, specificity (interference), carryover, hook effect, matrix effect and accuracy. Limits of detection (LOD) were determined on three occasions with absorbance values of negative urine samples (n = 9) from drug-free volunteers. LOD was calculated as: mean A0 - 3SD. A representative calibration curve was constructed by plotting mean absorbance values (n = 5) from calibrators (0.05, 0.1, 0.25, 0.5, 1, 2.5, 5, 7.5, 10, 25, 50, 100, 250, 500, 750, and 1000 μg/L) prepared in pooled negative urine (n = 3). In-house controls, prepared in pooled negative urine, evaluated performance around the cut-off and intra-plate imprecision at 3.75, 5, 6.25, 7.5, 10, and 12.5 μg/L JWH-018 N-pentanoic acid. Seven replicates of each control level were assayed on a single plate and calculations performed on mean absorbance values. For inter-plate imprecision, in-house control challenges at ±25% (3.75, 6.25, 7.5, and 12.5 μg/L) of the cut-off concentration (5 or 10 μg/L) were assayed in singulate on 34 plates over 2 weeks. Plate drift was monitored across all plates (n = 34) by comparing average absorbance values of the cut-off solutions (5 or 10 μg/L) positioned in columns 1, 6, and 12. Percent drift was calculated as [(average absorbance column 12 or 6 – average absorbance column 1)/ average absorbance column 1]. Cross-reactivity was investigated by analyzing singulate negative pooled urine samples fortified with one of 73 synthetic cannabinoids at 500 μg/L. These absorbances were compared to mean absorbances (n = 2) of the calibration curve (1–250 μg/L) and blank urine (n = 2). Cross-reactivity (%) was calculated as 100 x (concentration from the calibration curve)/(target analyte concentration). Immunoassay interferences can alter antibody binding and affect concentrations by increasing or decreasing signal response. To evaluate interferences, blank urine was individually fortified (1000 μg/L) with common drugs of abuse, metabolites, coadministered drugs, over-the-counter medications and structurally similar compounds (n = 93). Absorbances of these potentially interfering compounds were compared to absorbances of blank urine samples. Carryover evaluation included comparison of blank urine absorbance (n = 4) analyzed before and after 750 μg/L carryover samples with a t-test, (p value < 0.05) comparison. Matrix effects were assessed by evaluating triplicate absorbances of cut-off calibrators (5 and 10 μg/L) prepared in urine and EIA buffer. We also evaluated the ability of each cut-off to correctly classify performance challenge concentrations as positive or negative on
each plate. Performance challenges were assayed in singulate to simulate analysis of authentic specimens. Absorbances for the 3.75, 6.25, 7.5, and 12.5 μg/L samples were assayed over 34 plates and compared with averaged duplicate absorbances for 5 and 10 μg/L cut-offs assayed in the same column. LC-MS/MS analysis Our qualitative synthetic cannabinoid LC-MS/MS method simultaneously confirms the presence of 9 synthetic cannabinoids (JWH-018, JWH-073, JWH-210, JWH-122, JWH-081, JWH-250, AM2201, MAM2201, and RCS-4) and 19 metabolites in urine by comparison with an in-house spectral library.[12] Briefly authentic urine (100 μL) was fortified with deuterated internal standards and ammonium acetate buffer to adjust pH before enzymatic hydrolysis with beta-glucuronidase. Samples were extracted/ precipitated with acetonitrile, vortexed and centrifuged for 10 min at 15000 g and 4 °C to produce a supernatant suitable for LC-MS/MS analysis. The method was fully validated with good analytical recovery (53–95%), low matrix effect (95–122%) and LODs between 0.5 and 10 μg/L for all analytes. Diagnostic efficiency A true positive (TP) had a positive immunoassay and a positive LC-MS/MS result. If both results were negative, the sample was a true negative (TN). When samples were positive by immunoassay but negative by LC-MS/MS, the result was a false positive (FP) and a negative immunoassay result that confirmed positive by LC-MS/MS was a false negative (FN). We calculated sensitivity at each cut-off as TP/(TP + FN) x 100 and specificity as TN/(TN + FP) x 100. Efficiency was calculated as (TP + TN)/total number × 100.
Results We present performance criteria from a fully validated immunoassay method for screening synthetic cannabinoids in urine targeting JWH-018 N-pentanoic acid. Daily LODs of 0.58, 0.66, and 0.33 μg/L were obtained yielding a mean 0.52 μg/L LOD. The non-linear regression model employed an exponential function (Supplemental Figure S2A), represented by the equation: Y0 = 0.0515 + 1.5352*e(-0.4459x) + 0.6372*e(-0.04864x). The goodness of fit was expressed by the coefficient of determination (r2) = 0.996. Natural logarithms of mean absorbance vs. concentration exhibited linearity from 1 to 250 μg/L and r2 = 0.992 (Supplemental Figure S2B). Intra- (n = 7) and inter-plate imprecision (n = 34) evaluated around each cut-off was expressed as percent coefficient of
Table 1. Intra- and inter-plate imprecision of the Neogen SPICE ELISA Kit and discrimination power for performance challenges at ± 25% of 5 and 10 μg/L cut-offs Absorbance Intra-plate (N = 7)
Absorbance Inter-plate (N = 34)
μg/L
Mean (SD)
% CV
Mean (SD)
% CV
3.75 5 6.25 7.5 10 12.5
0.564 (0.014) 0.475 (0.019) 0.385 (0.020) 0.354 (0.015) 0.283 (0.011) 0.243 (0.010)
2.4 4.1 5.3 4.2 3.9 4.1
0.789 (0.047) 0.601 (0.035) 0.537 (0.041) 0.445 (0.040) 0.350 (0.025) 0.309 (0.027)
6.0 5.9 7.6 8.9 7.1 8.6
Drug Test. Analysis (2014)
Copyright © 2014 John Wiley & Sons, Ltd.
Difference to cutoff (N = 7) Significant — Significant Significant — Significant
p value >0.0001 — >0.0001 >0.0001 — >0.0001
wileyonlinelibrary.com/journal/dta
Drug Testing and Analysis
E. Spinelli et al.
variation (%CV) (Table 1). Acceptable intra- (≤5.3%) and inter-plate (≤8.9%) imprecision were observed over 3.75–12.5 μg/L. Plate drift for 5 and 10 μg/L cut-off calibrators in columns 6 and 12 ranged from -9 to 16% (Median 9%) and from -8 to 17% (Median 4%), respectively. For replicates (n = 2) in consecutive wells, absorbance differences for 5 and 10 μg/L cut-off calibrators (n = 102, three pairs of each concentration per plate) was 0–12% (Median 3%) and 0–14% (Median 4%), respectively. Of 73 synthetic cannabinoids evaluated at 500 μg/L, 19 samples exhibited absorbances equal to or greater than the blank and were considered non-cross-reactive (Table 2a). Twenty-two sample absorbances were between the blank and 5 μg/L urine calibrator displaying low cross reactivity (
Research article Received: 27 May 2014
Revised: 14 July 2014
Accepted: 15 July 2014
Published online in Wiley Online Library
(www.drugtestinganalysis.com) DOI 10.1002/dta.1702
Performance characteristics of an ELISA screening assay for urinary synthetic cannabinoids Eliani Spinelli,a,b Allan J. Barnes,a Sheena Young,a Marisol S. Castaneto,a Thomas M. Martin,c Kevin L. Klettec and Marilyn A. Huestisa* Synthetic cannabinoids are marketed as legal alternatives to cannabis, as routine urine cannabinoid immunoassays do not detect synthetic cannabinoids. Laboratories are challenged to identify these new designer drugs that are widely available and represent a major public health and safety problem. Immunoassay testing offers rapid separation of presumptive positive and negative specimens, prior to more costly and time-consuming chromatographic confirmation. The Neogen SPICE ELISA kit targets JWH-018 N-pentanoic acid as a marker for urinary synthetic cannabinoids. Assay performance was evaluated by analyzing 2469 authentic urine samples with the Neogen immunoassay and liquid chromatography-tandem mass spectrometry (LC-MS/MS). Two immunoassay cut-off concentrations, 5 and 10 μg/L, classified samples as presumptive positive or negative, followed by qualitative LC-MS/MS confirmation for 29 synthetic cannabinoids markers with limits of detection of 0.5–10 μg/L to determine the assay’s sensitivity, specificity and efficacy. Challenges at ±25% of each cut-off also were investigated to determine performance around the cut-off and intra- and inter-plate imprecision. The immunoassay was linear from 1 to 250 μg/L (r2 = 0.992) with intra- and inter-plate imprecision of ≤5.3% and 98% and sensitivity >95% when evaluated against LC-MS/MS confirmation results. Our aims were to optimize assay performance and validate the Neogen SPICE ELISA kit as a screening method for the detection of synthetic cannabinoids in urine. Presumptive positive (1409) and negative (1060) urine specimens (n = 2469) were compared to the most extensive qualitative LC-MS/MS confirmation assay for 29 synthetic cannabinoid markers to determine assay performance characteristics at multiple cut-offs.
Materials and methods Reagents and materials Each Neogen SPICE ELISA kit contained antibody coated 96-well microtitre plates, drug enzyme conjugate, tetramethylbenzidine (TMB) substrate, 1 N H2SO4 stop solution, enzyme immunoassay (EIA) buffer containing bovine serum and preservative, concentrated wash buffer (10x), and qualitative positive (100 μg/L) and negative control solutions prepared in human synthetic urine. Kit components were stored at 4 °C until analysis, and the assay was performed according to manufacturer’s instructions. Samples were diluted 1:5 with EIA buffer in uncoated 96-well microtitre plates (Greiner, Monroe, NC, USA) before analysis. Blank urine was collected from drug-free volunteers and evaluated to be negative prior to use in the preparation of calibrators and quality control samples for performance challenges. Calibrators, controls and performance challenges JWH-018 N-pentanoic acid metabolite was purchased from Cayman Chemicals (Ann Arbor, MI, USA) as 1 mg/mL ampules. Stock standard solutions of cut-off calibrators and controls were prepared from different ampules by diluting contents with appropriate volumes of methanol. Cut-off calibrators (5 and 10 μg/L) were prepared by fortifying authentic drug-free urine. Working performance challenges at ±25% of the 5 and 10 μg/L cut-offs were prepared by fortifying drug-free urine with stock control solution. Specimens Authentic urine specimens (n = 20 017) were randomly collected from US Department of Defense (DoD) service members stationed
wileyonlinelibrary.com/journal/dta
worldwide to determine the prevalence of synthetic cannabinoids intake from July 2011 through June 2012. Per DoD routine drug testing, specimens screened negative for common drugs of abuse (cannabinoids, cocaine, amphetamines, phencyclidine, and opiates) prior to shipment for synthetic cannabinoid testing. Of 20,017 specimens initially analyzed with the Randox Drugs of Abuse V biochip array technology for synthetic cannabinoids, 1409 presumptive positive and 1060 presumptive negative specimens were selected for further testing. Specimens were stored at room temperature (RT) prior to DoD testing and until synthetic cannabinoid initial testing. Specimens (n = 2469) were then stored at 4–7 °C before Neogen ELISA determination and LC-MS/MS confirmation.
ELISA instrumentation Qualitative, automated synthetic cannabinoids analysis was performed on a Freedom EVO 100 configured with a microtitre plate washer and reader (Tecan Group, San Jose, CA, USA). All samples, reagents and microtitre plates were equilibrated at RT. Per manufacturer’s guidelines, a 1:5 sample dilution was performed by adding 160 μL EIA buffer to uncoated microtitre plate wells and 40 μL blank, cut-off, in-house performance controls or unknown specimens to the appropriate wells, followed by 30-s plate shaking. Positive and negative Neogen controls were processed without dilution; 20 μL diluted samples or undiluted controls were added to the coated microtitre plate; 100 μL enzyme conjugate was added and the mixture incubated at RT for 45 min. The plate was washed five times with diluted wash buffer (300 μL), manually inverted and slapped dry to remove residual liquid. TMB substrate (100 μL) was added and, after a 30-min incubation, 100 μL acid stop solution was added and absorbance (450 nm) immediately measured with a plate reader.
Method development Although analysis was performed according to manufacturer’s instructions, additional experiments were performed. We observed that the length of time reagents were in contact with antibodies during the competition phase contributed to plate drift; liquid handling parameters and plate washing instructions were optimized to ensure consistent incubation times. Although no cut-off was specified, several JWH-018 N-pentanoic acid calibrators (0, 1.0, 2.5, 4, and 6 μg/L) prepared in synthetic urine are provided, allowing laboratories to determine limits that best suit their drug testing programs. We prepared matrix-matched cut-off solutions in authentic urine at 5 and 10 μg/L that were analyzed in duplicate at the beginning, middle, and end of each plate, in columns 1, 6, and 12. Negative and positive Neogen controls were run in singulate at the beginning and end of each plate, while fortified urine performance challenges at ±25% of each cut-off were evaluated in the middle of the plate. Urine controls at 3.75 and 6.25 μg/L assessed performance of the 5 μg/L cut-off and 7.5 and 12.5 μg/L controls monitored performance of the 10 μg/L cut-off. A representative plate layout is presented in Supplemental Figure S1. Strategic placement of calibrators, controls and performance challenge specimens allowed optimal evaluation of assay performance criteria. Specimens were presumptively positive when their absorbance at 450 nm was less than or equal to the averaged cutoff of the first pair of calibrators. Similarly, a negative result was indicated when the specimen absorbance was greater than the average of duplicate cut-off calibrators in the first column.
Copyright © 2014 John Wiley & Sons, Ltd.
Drug Test. Analysis (2014)
Drug Testing and Analysis
Performance characteristics of an ELISA screening assay for SC Assay performance criteria The method was validated by determining sensitivity, linearity, imprecision (intra- and inter-assay), plate drift, cross-reactivity, specificity (interference), carryover, hook effect, matrix effect and accuracy. Limits of detection (LOD) were determined on three occasions with absorbance values of negative urine samples (n = 9) from drug-free volunteers. LOD was calculated as: mean A0 - 3SD. A representative calibration curve was constructed by plotting mean absorbance values (n = 5) from calibrators (0.05, 0.1, 0.25, 0.5, 1, 2.5, 5, 7.5, 10, 25, 50, 100, 250, 500, 750, and 1000 μg/L) prepared in pooled negative urine (n = 3). In-house controls, prepared in pooled negative urine, evaluated performance around the cut-off and intra-plate imprecision at 3.75, 5, 6.25, 7.5, 10, and 12.5 μg/L JWH-018 N-pentanoic acid. Seven replicates of each control level were assayed on a single plate and calculations performed on mean absorbance values. For inter-plate imprecision, in-house control challenges at ±25% (3.75, 6.25, 7.5, and 12.5 μg/L) of the cut-off concentration (5 or 10 μg/L) were assayed in singulate on 34 plates over 2 weeks. Plate drift was monitored across all plates (n = 34) by comparing average absorbance values of the cut-off solutions (5 or 10 μg/L) positioned in columns 1, 6, and 12. Percent drift was calculated as [(average absorbance column 12 or 6 – average absorbance column 1)/ average absorbance column 1]. Cross-reactivity was investigated by analyzing singulate negative pooled urine samples fortified with one of 73 synthetic cannabinoids at 500 μg/L. These absorbances were compared to mean absorbances (n = 2) of the calibration curve (1–250 μg/L) and blank urine (n = 2). Cross-reactivity (%) was calculated as 100 x (concentration from the calibration curve)/(target analyte concentration). Immunoassay interferences can alter antibody binding and affect concentrations by increasing or decreasing signal response. To evaluate interferences, blank urine was individually fortified (1000 μg/L) with common drugs of abuse, metabolites, coadministered drugs, over-the-counter medications and structurally similar compounds (n = 93). Absorbances of these potentially interfering compounds were compared to absorbances of blank urine samples. Carryover evaluation included comparison of blank urine absorbance (n = 4) analyzed before and after 750 μg/L carryover samples with a t-test, (p value < 0.05) comparison. Matrix effects were assessed by evaluating triplicate absorbances of cut-off calibrators (5 and 10 μg/L) prepared in urine and EIA buffer. We also evaluated the ability of each cut-off to correctly classify performance challenge concentrations as positive or negative on
each plate. Performance challenges were assayed in singulate to simulate analysis of authentic specimens. Absorbances for the 3.75, 6.25, 7.5, and 12.5 μg/L samples were assayed over 34 plates and compared with averaged duplicate absorbances for 5 and 10 μg/L cut-offs assayed in the same column. LC-MS/MS analysis Our qualitative synthetic cannabinoid LC-MS/MS method simultaneously confirms the presence of 9 synthetic cannabinoids (JWH-018, JWH-073, JWH-210, JWH-122, JWH-081, JWH-250, AM2201, MAM2201, and RCS-4) and 19 metabolites in urine by comparison with an in-house spectral library.[12] Briefly authentic urine (100 μL) was fortified with deuterated internal standards and ammonium acetate buffer to adjust pH before enzymatic hydrolysis with beta-glucuronidase. Samples were extracted/ precipitated with acetonitrile, vortexed and centrifuged for 10 min at 15000 g and 4 °C to produce a supernatant suitable for LC-MS/MS analysis. The method was fully validated with good analytical recovery (53–95%), low matrix effect (95–122%) and LODs between 0.5 and 10 μg/L for all analytes. Diagnostic efficiency A true positive (TP) had a positive immunoassay and a positive LC-MS/MS result. If both results were negative, the sample was a true negative (TN). When samples were positive by immunoassay but negative by LC-MS/MS, the result was a false positive (FP) and a negative immunoassay result that confirmed positive by LC-MS/MS was a false negative (FN). We calculated sensitivity at each cut-off as TP/(TP + FN) x 100 and specificity as TN/(TN + FP) x 100. Efficiency was calculated as (TP + TN)/total number × 100.
Results We present performance criteria from a fully validated immunoassay method for screening synthetic cannabinoids in urine targeting JWH-018 N-pentanoic acid. Daily LODs of 0.58, 0.66, and 0.33 μg/L were obtained yielding a mean 0.52 μg/L LOD. The non-linear regression model employed an exponential function (Supplemental Figure S2A), represented by the equation: Y0 = 0.0515 + 1.5352*e(-0.4459x) + 0.6372*e(-0.04864x). The goodness of fit was expressed by the coefficient of determination (r2) = 0.996. Natural logarithms of mean absorbance vs. concentration exhibited linearity from 1 to 250 μg/L and r2 = 0.992 (Supplemental Figure S2B). Intra- (n = 7) and inter-plate imprecision (n = 34) evaluated around each cut-off was expressed as percent coefficient of
Table 1. Intra- and inter-plate imprecision of the Neogen SPICE ELISA Kit and discrimination power for performance challenges at ± 25% of 5 and 10 μg/L cut-offs Absorbance Intra-plate (N = 7)
Absorbance Inter-plate (N = 34)
μg/L
Mean (SD)
% CV
Mean (SD)
% CV
3.75 5 6.25 7.5 10 12.5
0.564 (0.014) 0.475 (0.019) 0.385 (0.020) 0.354 (0.015) 0.283 (0.011) 0.243 (0.010)
2.4 4.1 5.3 4.2 3.9 4.1
0.789 (0.047) 0.601 (0.035) 0.537 (0.041) 0.445 (0.040) 0.350 (0.025) 0.309 (0.027)
6.0 5.9 7.6 8.9 7.1 8.6
Drug Test. Analysis (2014)
Copyright © 2014 John Wiley & Sons, Ltd.
Difference to cutoff (N = 7) Significant — Significant Significant — Significant
p value >0.0001 — >0.0001 >0.0001 — >0.0001
wileyonlinelibrary.com/journal/dta
Drug Testing and Analysis
E. Spinelli et al.
variation (%CV) (Table 1). Acceptable intra- (≤5.3%) and inter-plate (≤8.9%) imprecision were observed over 3.75–12.5 μg/L. Plate drift for 5 and 10 μg/L cut-off calibrators in columns 6 and 12 ranged from -9 to 16% (Median 9%) and from -8 to 17% (Median 4%), respectively. For replicates (n = 2) in consecutive wells, absorbance differences for 5 and 10 μg/L cut-off calibrators (n = 102, three pairs of each concentration per plate) was 0–12% (Median 3%) and 0–14% (Median 4%), respectively. Of 73 synthetic cannabinoids evaluated at 500 μg/L, 19 samples exhibited absorbances equal to or greater than the blank and were considered non-cross-reactive (Table 2a). Twenty-two sample absorbances were between the blank and 5 μg/L urine calibrator displaying low cross reactivity (
