Validation of the Charm 3 SL3 -Lactam Test for Screening Raw Milk in Compliance with the U.S. Pasteurized Milk Ordinance

348 SALTER ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 1, 2011 VETERINARY DRUG RESIDUES Validation of the Charm 3 SL3 -Lactam Test for Screening Raw Milk in Compliance with the U.S. Pasteurized Milk Ordinance Performance Tested Method SM 071002 Abstract The Charm 3 SL3 -Lactam Test is a 3 min receptor-based lateral-flow Rapid One-Step Assay (ROSA) that detects the six -lactam drugs of concern approved for dairy cattle in the United States. The method is a biochemical formulation change of the SL3 -Lactam Test evaluated and approved in 2007. The Charm 3 SL3 was evaluated under the AOAC Research Institute Performance Tested Method SM (PTM) program following the protocol of the U.S. Food and Drug Administration, Center for Veterinary Medicine. The method was approved as PTM 071002 on May 8, 2009. The following drugs were detected in three combined lots: penicillin G at 3.8 ppb, ampicillin at 8.0 ppb, amoxicillin at 8.4 ppb, cephapirin at 20.0 ppb, ceftiofur (total metabolites) at 79 ppb, and cloxacillin at 8.6 ppb 90% of the time with 95% confidence. These detection levels are lower than, but within 75% of, the U.S. Safe Level/Tolerances. Lot-to-lot repeatability was typically within 20% of these determined levels. The test kit was found to be suitable for testing thawed frozen samples. It was also found to respond with equal or better sensitivity to samples that contained incurred analytes, i.e., both the microbiologically active parent drug and its active metabolites. There were no interferences from somatic cells at 1.1 million/ml, bacterial cells at 300 000 CFU/mL, or 32 other non- -lactam drugs at 100 ppb. Ruggedness experiments indicated that the test procedure is robust. These results meet the fit-for-purpose approval criteria for inclusion in the National Conference for Interstate Milk Shipments milk testing program. 1 Participants METHOD AUTHORS ROBERT S. SALTER, DAVID DOUGLAS, LINDSEY MCROBBIE, JULIO QUINTANA,DAVID LEGG,JANINE SCHWARTz, DAVID CONAWAY,CARLA MCPHEE,STEVEN SAUL,andROBERT MARKOVSKY Charm Sciences Inc., 659 Andover St, Lawrence, MA 01843 Submitted for publication July 22, 2010. The method was independently tested, evaluated, and certified by the AOAC Research Institute as a Performance Tested Method SM. See http://www.aoac.org/testkits/steps.html for information on certification. Corresponding author s e-mail: bobs@charm.com SUBMITTING COMPANY Charm Sciences Inc., 659 Andover St, Lawrence, MA 01843 INDEPENDENT LABORATORY Dairy Quality Control Laboratory (DQCI), 5205 Quincy St, Mounds View, MN 55112 REVIEWERS JOE BOISON (General Referee) Canadian Food Inspection Agency, 116 Veterinary Rd, Saskatoon, SK, Canada S7N 2R3 PHIL KIJAK U.S. Food and Drug Administration, Center for Veterinary Medicine, MOD 2, 8401 Muirkirk Rd, Laurel, MD 20708 JIM AGIN Q Laboratories Inc., 1400 Harrison Ave, Cincinnati, OH 45214 2 Scope of Method 2.1 Target Drugs -Lactam drugs: penicillin G, ampicillin, amoxicillin, cephapirin, cloxacillin, and ceftiofur. 2.2 Matrix The method is applicable to raw commingled cow s milk. 2.3 Summary of Validated Performance Claims The Charm 3 SL3 -Lactam Test is a lateral-flow receptor assay designed to detect six -lactam drugs at their Safe Level/Tolerance (SL/T) levels. The method takes 3 min to detect penicillin G at 3.8 ppb, ampicillin at 8.0 ppb, amoxicillin at 8.4 ppb, cephapirin at 20.0 ppb, ceftiofur (total metabolites) at 79 ppb, and cloxacillin at 8.6 ppb with 95% confidence in raw commingled cow s milk. These drugs, which represent the six -lactam drugs approved for use in the United States, can be detected by this test method as incurred analytes in milk at levels equilavent to or lower than the levels detected in -lactam drug-fortified milk samples. The detection levels in fortified samples are lower than the SL/T values, but the method is not overly sensitive at levels below 75% of the SL/T values. It was demonstrated that 32 other animal drugs, somatic cells, and bacteria that could be encountered in the production of Grade A raw milk did not

SALTER ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 1, 2011 349 interfere with the ability of the test kit to provide the anticipated result. 3 Introduction 3.1 Principle The Charm 3 SL3 -Lactam Test is the next generation form of the SL3 -Lactam Test that was previously an AOAC Research Institute Performance Tested Method SM (PTM) evaluated and approved by the National Conference of Interstate Milk Shipments (NCIMS) in 2007 (1). The test format uses the same equipment and the Rapid One-Step Assay (ROSA) lateral-flow receptor binding assay but with updated Charm 3 biochemistry to detect the group of -lactam drugs and all six -lactam drugs approved in the United States. The formulation change requires evaluation as a new test rather than as a modification of the existing test. The Charm 3 SL3 Test is manufactured by Charm Sciences Inc. The test uses gold-bead receptors in a lateral-flow format to detect the class of -lactam drugs. The sensitivities of the receptors are engineered to specific SL/T detection levels. Milkaddedtotheteststripina56 C incubator rehydrates and mobilizes receptors that move across a -lactam-specific test line. Unreacted receptors bind and form a visible reddish test line; drug-reacted receptors pass across the test line and bind to the control line. The test is completed after 3 min of incubation. Test lines darker than the control line are evaluated numerically and interpreted as negative by a ROSA reader, indicating the absence of detectable -lactams near the SL/T. Test lines lighter than the control line are interpreted as positive, indicating the presence of a -lactam drug near or greater than the SL/T. 3.2 General Information -Lactam drugs are the most commonly used antibiotics in the health management of dairy cattle in the United States (2). It is a public health requirement that milk tankers be tested for -lactam drugs before unloading by using methods recommended to the NCIMS by the U.S. Food and Drug Administration (FDA) as specified in the Pasteurized Milk Ordinance (PMO) Appendix N program (3). The recommendation process follows a third-party laboratory evaluation/validation protocol such as the AOAC Research Institute PTM program, followed by Center for Veterinary Medicine (CVM) review-recommendation and NCIMS final approval (4). After NCIMS approval, the FDA adds the test method to M-a-85, a list of acceptable screening methods that comply with Appendix N of the PMO (5). 3.3 Summary of Results The validation protocol consisted of eight parts: (1) Drug-fortified dose response in raw commingled cow s milk demonstrated 90% detection with 95% confidence at the following concentrations: penicillin G, 3.8 ppb; ampicillin, 8.0 ppb; amoxicillin, 8.4 ppb; cephapirin, 20.0 ppb; ceftiofur (total metabolites), 79 ppb; and cloxacillin, 8.6 ppb. (2) Ruggedness perturbations defined the limitations of the test conditions: Pipet volume of milk sample, 300 ± 30 L; incubator temperature, 56 ± 1 C; incubation time, 3 to 3 min and 30 s, time to pipet four multiple samples, up to 75 s; read result within 3 min after finishing test, milk temperature, 0 7 C; and ambient temperature 10 35 C. (3) Milk samples frozen at 20 C for up to 8 weeks, and thawed overnight in a refrigerator gave expected positive and negative results. (4) Intrafamily cross-reactivity was tested with 32 other animal drugs; none interfered at 100 ppb concentrations. Other -lactam drugs were detected at concentrations ranging from 5 to 3000 ppb. (5) Samples with incurred analytes displayed similar or more sensitive dose response, indicating the drugs were detected after prescribed administration to cows. (6) Somatic cells at 1.1 million/ml did not interfere with positive or negative results. (7) Bacterial cells at 150 000 to 300 000 CFU/mL did not interfere with positive or negative results. (8) Supplied documentation supported manufacturing quality assurance, shelf-life determination, and 3-lot manufacturing consistency. The manufacturing consistency of the lots was verified by the independent laboratory. 4 Materials and Methods 4.1 Test Kit Information (4.1.1) Kit name. Charm 3 SL3 -Lactam Test. (4.1.2) Cat. Nos. LF-SL3-100K (100 tests and five positive control tablets) and LF-SL3-500K (500 tests and 25 positive control tablets). 4.2 Additional Supplies and Reagents (4.2.1) Standard reference materials. Amoxicillin, ampicillin, sodium cephapirin, sodium cloxacillin, and potassium penicillin G reference standards were obtained from the United States Pharmacopeia (USP; Rockville, MD, http://www.usp.org). Ceftiofur hydrocloride was obtained from Pfizer Co. (Kalamazoo, MI). Stock buffers for preparing USP -lactam standards and storage conditions were as listed in the USP (6). (4.2.2) Bacterial isolates from milk. Gram positive (G+) coagulase negative (Streptococcus spp., Staphylococcus aureus), Gram negative (G ; Pseudomonas aeroginosa, Enterobacter, Escherichia coli, and Klebsiella spp.), and G+/G were isolated from mastitic cow s milk by DQCI. Freshly grown cultures of isolates in nutrient broth (equal mixes) were prepared and diluted into milk at targeted concentrations of 150 000 and 300 000 CFU/mL. (4.2.3) Matrix. Fresh raw commingled cow s milk from local dairy producer tanks/silos and tested to be -lactam free by using the Charm II -Lactam Quantitative Assay (7) was used for testing. The age of the milk obtained by cow milking was 5 days.

350 SALTER ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 1, 2011 Figure 1. Charm 3 SL3 procedure. 4.3 Apparatus (4.3.1) ROSA incubator. 56 1 C, manufactured by Charm Sciences Inc. (4.3.2) Fixed-volume pipets. 300 L with disposable 200 1000 L tips. (4.3.3) ROSA Pearl Reader. Manufactured by Charm Sciences Inc., or equivalent, for date, time, operator, test identity, lot number of test, line intensity reading, and interpretation. 4.4 Precautions (4.4.1) The sample should be from mixed raw commingled milk with a fat level of <6.5%. Higher fat levels may not flow properly in the test strip and cause invalid results. Clumped colostrum and abnormal individual cow s milk may cause poor flow and invalid test development. Invalid test strips should not be read in reader. (4.4.2) Keep equipment clean and test strips clean of dust and milk debris. (4.4.3) Incubator temperature should be 56 1 C with lid lowered but not latched unless test is being performed. (4.4.4) Ambient temperature range, 10 35 C. 4.5 Charm 3 SL3 -Lactam Test Procedure (Figure 1) (4.5.1) Check that incubator temperature is 56 1 C. Mix milk sample(s). Place a test strip in incubator, peel back tape of Charm 3 SL3 test strip, exposing sample compartment well and pad. (4.5.2) Slowly add 300 30 L milk into the side of the sample compartment well. Reseal tape over sample pad. (4.5.3) Incubate for 3 min, but <3 min and 30 s. Incubators have an automatic timer, beeper, and lights to indicate when incubation is completed. (4.5.4) Remove strips from incubator. Visually verify that the strip has valid development, indicated by a solid and complete C (control) line. 4.6 Interpretation of Results Reader determination. Insert the visually valid Charm 3 SL3 test strip into the ROSA Pearl Reader set at the SLBL channel 2 mode, SLBL solid/not-blinking. Negative and zero readings indicate a negative result. Positive readings indicate a positive result. 4.7 Quality Control Positive and negative controls are run daily and when initial positive samples are retested. Negative controls are visually negative, with the T-line clearly darker than the C-line, and give readings that are less than 600. Positive controls are visually positive and give readings greater than +400. Calibration strips are supplied with the ROSA Pearl Reader and must be read daily to verify results within the range printed on the strips.

SALTER ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 1, 2011 351 Table 1. Summary of independent laboratory evaluation a Concn, ppb Penicillin G Ampicillin Amoxicillin Cephapirin Cloxacillin 0 0 [0/60] 0 [0/60] 0 [0/60] 3.3 [2/60] 0 [0/60] 2.0 0 [0/30] 2.7 10 [3/30] 3.0 47 [14/30] 3.2 50 [15/30] 3.5 93 [28/30] 4.0 93 [28/30] 0 [0/30] 0 [0/30] 0 [0/30] 0 [0/30] 5.0 100 [30/30] 33 [10/30] 3 [1/30] 0 [0/30] 5.5 50 [15/30] 6.0 63 [19/30] 7 [2/30] 3 [1/30] 6.5 40 [12/30] 7.0 73 [22/30] 60 [18/30] 53 [16/30] 8.0 97 [29/30] 87 [26/30] 0 [0/30] 80 [24/30] 9.0 100 [30/30] 10.0 100 [30/30] 100 [30/30] 7 [2/30] 100 [30/30] 12.0 17 [5/30] 14.0 47 [14/30] 16.0 67 [20/30] 17.5 83 [25/30] 18.0 19.0 20.0 100 [30/30] 20.5 100 [30/30] 90% positive with 95% confidence 3.8 8.0 8.4 20.0 8.6 Safe level tolerance 5.0 10.0 10.0 20.0 10.0 a Dose response of Charm 3 SL3 expressed as percent positive and [No. positive/no. tested]. 4.8 Quality Assurance The collections of manufacturer s data and independent laboratory data were modeled after Good Laboratory Practices (8). Data were made available to FDA-CVM reviewers, who verified the authenticity of the data as supported by the printed results. 4.9 Experimental Design (4) (4.9.1) Sensitivity, selectivity, and lot-to-lot repeatability. The independent laboratory prepared drug concentrations at 1/5, 2/5, 3/5, and 4/5 of the SL/T values, at the SL/T values, and at three additional concentrations to define the 90% positive response with 95% confidence. An experiment for each drug had 30 replicates of each concentration blind-coded with 60 negative samples for a total of 300 samples ([eight concentrations per -lactam drug] [30 replicates] + 60 negatives). Each experimental set of samples was divided into 10 equivalent laboratory breakpoints (subsets) with an equal number of specific lot numbers (001, 003, or 004) performed in each subset. Data analyses were performed to determine the minimum 90% detection level with 95% confidence (9) and to establish statistical parameters for the numerical readout of the qualitative result (10). (4.9.2) Ruggedness. Manufacturer s data included seven ruggedness perturbations following an approved multivariate experiment (11). (4.9.3) Frozen stability. Qualified negative raw milk samples, both unfortified and fortified with 4.2 ppb penicillin G, 19 ppb cephapirin, and 9 ppb cloxacillin, were prepared and frozen in blind aliquots, repeatedly thawed in the refrigerator overnight each week for 8 weeks, and tested. (4.9.4) Chemical interferences. Intrafamily animal drug cross-reaction was tested by using a series of cocktails containing animal drugs at 100 ppb (sulfadiazine, sulfanilamide, sulfathiazole, sulfamethazine, sulfapyridine, sulfadimethoxine, tetracycline, oxytetracycline, chlortetracycline, doxycycline, gentamicin, neomycin, streptomycin, ivermectin, erythromycin, pirlimycin, tilmicosin, novobiocin, enrofloxacin, florfenicol, furosemide,

352 SALTER ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 1, 2011 Figure 2. Percentage of time that penicillin G concentrations (log dose, ppb) tested positive (- -) is plotted versus probit by using various statistical models, i.e., Logit, Probit, Cloglog, and Cauchit. Figure 3. Population of responses is used to determine the level of 90% positive response with one-tailed 95% confidence. In this example the 90/95% of 3.8 ppb was less than the 5.0 ppb SL/T, but it was not more than 75% less; therefore, the Charm 3 SL3 Test met detection specifications for penicillin G. trichlormethiazide, thiabendazole, chlorothiazide, oxytocin, phenylbutazone, dexamethasone, flunixin, p-aminobenzoic acid (PABA), dipyrone, and nitrofuran metabolites AOZ and AMOZ. These drugs were added to negative raw milk, spiked with 4.2 ppb penicillin G, 18 ppb cephapirin, and 9 ppb cloxacillin. Interfamily cross-reaction was determined by using six replicates and finding the concentration that gave 5 positive of six responses with a reading intensity equivalent to that obtained for 4.2 ppb penicillin G. (4.9.5) Incurred residue studies. Incurred residue samples were obtained by treating lactating cows with intramammary infusion under the supervision of a veterinarian. Samples containing incurred penicillin G, amoxicillin, cephapirin, ampicillin, and ceftiofur were tested and commingled in 2007 and stored at 80 C (1). These concentrations were verified by HPLC analysis before use (12 16). Cloxacillin-incurred samples were collected in the spring of 2010 by using three cows that were healthy and untreated for 30 days and whose milk qualified as negative on the basis of the Charm II -Lactam Quantitative Assay. One dose of intramammary infusion Bovi-clox was administered to four udder quadrants, and milk was collected every 12 h post-administration until the milk cleared tests using the Charm II Cloxacillin Test method. Daily A.M. and P.M. milkings were collected and frozen. Samples were tested by the Charm II Cloxacillin Test method to identify samples likely to contain -lactam drugs at levels above the SL/T, but no more than four times greater than the SL/T. The frozen commingled samples were then subjected to HPLC procedures to identify and quantify the drug content in the samples (12). Specific samples were chosen for dilution with negative commingled cow s milk (milk collected and prequalified as negative before injection) to produce test samples. In one study, ceftiofur samples were prepared at 20, 40, 50, 60, 70, 80, 90, and 100 ppb total ceftiofur and ceftiofur metabolite (desfuroylceftiofur acetamide). Thirty replicates of each concentration and 60 negative samples were blind-coded (300 total samples), frozen at 80 C, and then sent to the independent laboratory for evaluation. In a second study, samples containing incurred amoxicillin, ampicillin, cephapirin, cloxacillin, and penicillin G were prepared at the SLT, at the 90% positive levels with 95% confidence levels, and at 1/2, 1/4, and 1/10 of the 90% positive levels. Ten replicates of each sample and 50 negative samples were blind-coded (300 total samples), frozen at 80 C, and sent to the independent laboratory for analysis. (4.9.6) Somatic cell interference study. The independent laboratory identified farm raw milk with a high somatic cell count (SCC), 1.1 million SCC/mL, and qualified it as negative for -lactam with the Charm II -Lactam Quantitative Assay. Milk was spiked with 4.2 ppb penicillin G and 20.0 ppb cephapirin. Negative samples were divided into 60 replicates, and the drug-fortified samples were each divided into 30 replicates. The 120 blind-coded samples were tested by technicians who were not involved in sample coding or preparation. (4.9.7) Bacterial interference study. The independent laboratory prepared blind-coded samples (negative, 4.2 ppb penicillin G, and 20.0 ppb cephapirin) by fortifying milk with freshly prepared bacterial G+, G, and G+/G cultures at 150 000 and 300 000 CFU/mL, and then dividing each of the samples into five replicates except the negative G+/G, which was split into 20 replicates. The total number of

SALTER ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 1, 2011 353 Table 2. Total ceftiofur, ppb a Results of the incurred ceftiofur study Number positive Number negative samples was 135. Samples were prepared, blind-coded, and tested by the independent laboratory using technicians who were not involved in sample preparation or coding. (4.9.8) Quality assurance manufacture. Information about manufacturing quality assurance and quality control, data for three manufactured batches, and real-time 6 month shelf-life data were supplied to support consistency in manufacturing and shelf-life claims. Validation Results and Discussion 5.1 Sensitivity, Selectivity, and Lot-to-Lot Repeatability Study Positive Negative 0 60 0 20 0 30 0 40 0 30 0 55 8 24 27 60 15 15 50 65 18 12 60 70 22 8 73 80 30 0 100 100 30 0 100 a Parented and related metabolites. The independent laboratory-determined drug sensitivity data obtained by using fortified samples are shown in Table 1. Figures 2 and 3 are graphical depictions of the results of the statistical analysis, using penicillin G data, and the determination of the 90% detection level with 95% confidence (90/95%). The 90/95% levels determined from the fortified-drug data were less than or equal to the SL/T values and within 25% of the SL/T values that met sensitivity specifications for penicillin G, ampicillin, amoxicillin, cephapirin, and cloxacillin. The selectivity of the Charm 3 SL3 Test was measured with 60 negative samples in each of the five dose response experiments. In four experiments there were zero positives in 60 samples. In the fifth experiment there were two positives in 60 samples. There was no accountable laboratory error to explain the two positives in that experiment. All five experiments met FDA-CVM specifications for 90% selectivity with 95% confidence. The results of the incurred ceftiofur study are shown in Table 2. Ceftiofur sensitivity was determined on the basis of incurred residues, because the variety of ceftiofur-related metabolites on which the tolerance is based cannot be duplicated in fortified samples. The determined 90/95% level was 79 ppb, which met SL/T sensitivity claim specifications. There were zero positives in the 60 negative samples, meeting the FDA 90% selectivity with 95% confidence specifications. The test strips used in the independent laboratory evaluations were an even mixture of three manufactured lots. Therefore, dose response data for each lot (n = 10 replicates) were determined and compared with the manufacturer s submitted data (n = 30). Table 3 compares the 90% positive with 95% confidence values of the three lots. In a few cases, e.g., cephapirin and lot 001 ampicillin, the independent sensitivity calculated was greater than the SL/T value, reflecting the larger uncertainty of the 95% confidence values when only n = 10 replicates are used. The levels determined for the three lots are consistent to within 20% of the SL/T values at both the independent and the manufacturer s laboratories. The results of the fortified -lactams and incurred ceftiofur sensitivity experiments were that all six U.S.-approved -lactam drugs were detected at or below the SL/T values and were not more sensitive than 75% of SL/T values. The principle of the Charm 3 SL3 -Lactam Test uses bacterial receptors in a lateral-flow design similar to other ROSA assays manufactured by Charm Sciences Inc. The Table 3. Results of the lot-to-lot repeatability study comparing the 90/95% values of the three lots Lot 001, ppb Lot 003, ppb Lot 004, ppb Drug Mfr. a Ind. lab b Mfr. a Ind. lab b Mfr. a Ind. lab b Penicillin G 4.1 4.1 4.0 4.4 3.3 3.9 Ampicilllin 8.3 10.4 6.2 7.5 6.5 7.6 Amoxicillin 8.8 8.4 7.8 9.2 8.2 8.4 Cephapirin 17.6 20.1 16.1 22.4 15.4 22.8 Ceftiofur 70 80 c 68.7 70 80 c 81.2 70 80 c 83.3 Cloxacillin 8.4 9.3 7.2 8.9 6.9 8.7 a b c Value submitted by the manufacturer. Value determined by the independent laboratory. Because the data were not sufficient for precise determination, the range is reported.

354 SALTER ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 1, 2011 Table 4. Summary P values for assay perturbations (n = 12 determinations) Assay parameter Perturbation Raw milk PenG, Cephapirin, 4.2 ppb a 18 ppb Cloxacillin, 9 ppb Milk 1 Milk temperature 0 C 0.367 0.355 0.633 0.783 Control, 2 C 7 C 0.221 0.160 0.433 0.833 Milk volume 270 L 0.559 0.837 0.886 0.641 Control, 300 L 330 L 0.741 0.760 0.892 0.223 Incubation time 3.5 min 0.071 0.616 0.529 0.185 Control, 3 min 3.5 min 0.762 0.628 0.191 0.069 Reading after completion 3 min 0.140 0.083 0.111 0.021 b Control, <1 min 3 min 0.202 0.013 b 0.036 b 0.497 Ambient temperature 10 C 0.327 0.083 0.701 0.218 Control, 22 C 35 C 0.285 0.935 0.036 b 0.221 Incubation temperature 55 C 0.363 0.299 0.692 0.156 Control, 56 C 57 C 0.743 0.896 0.387 0.112 Assay set up time 75 s 0.116 0.391 0.109 0.345 Control, <30 s 75 s 0.502 0.946 0.072 0.118 Milk 2 Milk temperature 0 C 0.146 0.344 0.840 0.356 Control, 2 C 7 C 0.681 0.762 0.524 0.483 Milk volume 270 L 0.202 0.196 0.991 0.387 Control, 300 L 330 L 0.107 0.498 0.878 0.188 Incubation time 3.5 min 0.316 0.029 b 0.292 0.127 Control, 3 min 3.5 min 0.585 0.494 0.128 0.197 Reading after completion 3 min 0.209 0.060 0.198 0.021 b Control, <1 min 3 min 0.408 0.033 b 0.099 0.058 Ambient temperature 10 C 0.611 0.849 0.225 0.018 b Control, 22 C 35 C 0.158 0.034 b 0.450 0.558 Incubation temperature 55 C 0.125 0.729 0.596 0.322 Control, 56 C 57 C 0.691 0.102 0.563 0.951 Assay set up time 75 s 0.352 0.726 0.630 0.443 Control, <30 s 75 s 0.004 c 0.397 0.207 0.660 a PenG = Penicillin G. b Minor effect, 0.05 P 0.01. c Major effect, P 0.01. Charm 3 SL3 receptors have affinity for all -lactam drugs, and they are manufactured by a proprietary technique to achieve detection near regulatory levels without being hypersensitive (17). 5.2 Ruggedness Ruggedness parameters of the Charm 3 SL3 Test were determined by multivariate analysis using 12 replicates in two separate experiments and examining their resulting averages and SD values in a t-test analysis. Differences were expressed as probability of difference (P) shown in Table 4. Perturbations were considered minor if P <0.05andmore significant if P < 0.01. Insignificant perturbation levels, P > 0.05, were observed with pipetting 300 L 10%; incubator temperature 1 C, and milk temperature 0 to 7 C. Minor perturbation effects, P < 0.05 but generally P > 0.01, occurred with positive spiked samples with the following perturbations: ambient temperature, 10 to 35 C; time of reading, 3 min after completion of a test; and time of incubation, 3 to 3.5 min. The only minor perturbation that

SALTER ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 1, 2011 355 Table 5. Concentration response of incurred drug residues a Penicillin G Ampicillin Amoxicillin Cephapirin Cloxacillin ppb Positive, % ppb Positive, % ppb Positive, % ppb Positive, % ppb Positive, % 5 100 10 100 10 100 20 100 11 100 3.8 100 8 100 8.4 100 19 100 9.5 100 1.9 0 4 100 4.2 100 9.5 80 4.75 10 0.95 0 2 100 2.1 0 4.75 0 2.38 10 0.38 0 0.8 100 0.84 0 1.9 0 0.95 0 0 0 0 0 0 0 0 0 0 0 a Expressed as percent positive (positive, %); n = 10. repeated with the two milk samples was 3.5 min with 9.0 ppb cloxacillin testing more positive. Multiple tests could be performed at one time if samples were set up in the incubator, and the incubator timer was activated within 75 s of the first sample addition. Generally, exceeding these perturbation parameters will trend to more positive readings of samples containing drug(s) and more negative readings of samples containing no drug. An increase in positive readings would potentially cause a more positive dose response, but it does not result in false positives. This more positive trend for positive spiked samples was observed with a 3 min 30 s incubation time, 3 min to read results, and >75 s to set up and pipet multiple samples. Parameters that affect the composition of a milk sample, such as poor sampling, inadequate mixing, and fat content of >6.5%, are also considered critical and can cause invalid or erroneous results. 5.3 Incurred Residue Studies No false positive results were observed in the incurred residue study (Table 5). The positive response of samples containing incurred penicillin G and cloxacillin was very similar to the response of spiked samples. The sensitivity data obtained for the amoxicillin and cephapirin samples were slightly more sensitive than the sensitivity data obtained for the fortified samples. Samples with incurred ampicillin showed much greater sensitivity, compared with fortified samples, with a 1/10 dilution from the fortified-determined sensitivity still producing 10 of 10 positive responses. It was not until the incurred ampicillin sample was diluted 1/20 from the SL/T value that a negative response was observed. This increase in sensitivity with these ampicillin incurred samples may be explained by metabolite that is not quantitated by HPLC but is cumulatively detected along with the parent compound by the screening method. Metabolites in these samples were observed previously by HPLC Receptorgram analysis (1, 16). Similar results have been observed previously for other incurred residue studies (17). HPLC Receptorgram analysis of the incurred ampicillin sample isolated a parent compound peak at 6.0 7.0 min and a secondary unidentified peak at 14.5 16.5 min. This secondary peak was 4 times more sensitive with the Charm 3 SL3 Test, as compared with the original SL3 Test approved in 2007, explaining the results observed in this study. This peak elution time was not consistent with the infused hetacillin; however, the elution time was consistent with ampicillin polymer synthesized as prescribed by Aki et al. (18) and that is known to form in aqueous solution. This synthesized polymer also had the characteristic of being more sensitive when the Charm 3 SL3 Test was used, compared with the SL3 Test evaluated in 2007. Ampicillin does not have a high incidence of use in U.S. dairy practices; however, the Charm 3 SL3 Test could detect an incurred level of this drug in milk lower than the SL/T as determined by HPLC (12). It is possible that the HPLC method underestimates ampicillin residue and ampicillin biologic activity because it does not detect polymers formed in the parenteral drug solution (19). The Charm 3 SL3 -Lactam Test detected antibiotics at levels similar to or lower than the levels in fortified samples and demonstrates that the test reliably detects drugs under the prescribed conditions of administration to lactating cows. 5.4 Interference and Frozen Milk Studies There were no effects on the test performance from somatic cells, bacteria, or 32 non- -lactam animal drugs. Frozen milk thawed repeatedly at 1, 2, 3, 4, and 8 weeks had no effect on results. Table 6 summarizes the results from the four studies. There were no false positive or false negative results from the frozen, somatic, or microbial interference experiments, indicating no effect on the test. There were no false positive results with negative raw milk or false negative results with penicillin G or cephapirin samples in other drug interference studies; however, there was one false negative in the 21 positive cloxacillin (9 ppb) samples that was not repeatable. All the results support no interference from other non- -lactam drugs. Table 7 summarizes interfamily experiments for cross-reactivity with other -lactam drugs that are not used

356 SALTER ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 1, 2011 Table 6. Results for the interference studies a Interference challenge Spiking material Negative milk Penicillin G at 4.2 ppb Cephapirin at 20.0 ppb b Cloxacillin at 9.0 ppb c Bacteria Gram+/ mix, 110 150 k/ml 0/20 5/5 5/5 Gram+/ mix, 250 340 k/ml 0/20 5/5 5/5 Gram+, 120 160 k/ml 0/5 5/5 5/5 Gram+, 280 320 k/ml 0/5 5/5 5/5 Gram, 110 150 k/ml 0/5 5/5 5/5 Gram, 220 250 k/ml 0/5 5/5 5/5 Control milk 64 78 k/ml 0/5 5/5 5/5 Bacteria summary 0/65 35/35 35/35 Somatic cells 1100000 SCC/mL 0/60 30/30 30/30 Interfering chemicals 6 Sulfa drugs 0/3 3/3 3/3 3/3 4 Tetracyclines 0/3 3/3 3/3 3/3 3 Aminoglycosides 0/3 3/3 3/3 3/3 4 Macrolides and ivermectin 0/3 3/3 3/3 2 d /3 Dewormers and pesticides 0/3 3/3 3/3 3/3 Steroids and anti-inflamatory 0/3 3/3 3/3 3/3 Enrofloxacin, PABA, nitrofuran 0/3 3/3 3/3 3/3 metabolites, and florfenicol Chemical interference summary 0/21 21/21 21/21 20/21 Frozen milk Thawed weekly and tested at 1, 2, 3, 4, and 8 weeks 0/35 35/35 35/35 35/35 a b c d Results are expressed as # positive results/# tested. 19.0 ppb in freeze-thaw study and 18.0 ppb in interfering-chemical study. Cloxacillin was not tested in the microbial or somatic interference study. Sample was retested 3 more times and was positive 3 of 3 times. Table 7. Cross-reactivity of other -lactam drugs not used in U.S. dairy management practices at the estimated 90% positive concentration with 95% confidence Drug Concn, ppb Cefacetrile 50 Cefalexin 2500 Cefalonium 7 Cefazolin 30 Cefoperazone 1 Cefquinome 50 Cefuroxime 200 Dicloxacillin 5 10 Nafcillin 150 Oxacillin 15 Ticarcillin 50 Cefadroxil 3000 in U.S. dairy management practices. The listed cross-reactivity levels are based on at least five positives in six replicates and on the reader value at that concentration giving a value equivalent to 4.2 ppb penicillin G. The Charm 3 SL3 formulation, in comparison with the original SL3 -Lactam Test submitted in 2007, was more sensitive to cloxacillin, oxacillin, and dicloxacillin -lactam drugs and less sensitive to cefalexin and cefadroxil -lactam drugs. 5.5 Quality Control and Quality Assurance In Table 3 the dose responses of the three manufactured lots tested by the independent laboratory in the sensitivity studies are compared with data submitted by the manufacturer. Results indicate that the individual lot 90/95% levels were duplicated within 20% of the SL/T values of the combined results (Table 1) and are consistent with the manufacturer-submitted data. In the manufacturer-submitted data, not shown, the 90/95% levels of the submitted lots were duplicated within 10% by tests after 6 months of refrigerated storage. These results are within 20% of the sensitivity levels determined by the independent laboratory. The results meet the protocol acceptance criteria and justify expiration dating of test kits (4).

SALTER ET AL.: JOURNAL OF AOAC INTERNATIONAL VOL. 94, NO. 1, 2011 357 Additional real-time shelf-life studies are on-going for refrigerated 1-year storage. 6 Summary The Charm 3 SL3 -Lactam Test meets the conditions required for approval under the FDA-CVM performance test verification process administered by the AOAC Research Institute. The Charm 3 SL3 -Lactam Test is identical to and uses the same equipment as Charm SL3 -Lactam Test previously approved in 2007. The Charm 3 SL3 used a biochemical formulation change to attain cloxacillin sensitivity, and therefore, was evaluated as a new test. The data submitted by the manufacturer were satisfactorily duplicated by DQCI, the independent laboratory, which verified the detection of -lactam drugs at the following 90/95% detection levels: amoxicillin, 8.4 ppb; ampicillin, 8.0 ppb; cephapirin, 20.0 ppb; ceftiofur (total metabolites), 79 ppb; cloxacillin, 8.6 ppb; and penicillin G, 3.8 ppb. The Charm 3 SL3 -Lactam Test method met the other approval criteria for ruggedness, lack of interferences, frozen stability, selectivity, samples with incurred residus, lot repeatability, and quality assurance. Acknowledgments We thank Vicki Lancaster, FDA-CVM, Rockville, MD, for the statistical calculations depicted in Figures 2 and 3 and performed by R software. 7 References (1) Salter, R.S., Conaway, D., & Markovsky, R. (2008) J. AOAC Int. 91, 675 683 (2) Maturin, L. (2008) Milk Residues: What, When, Where and HowtoDoDrugResidueTests, National Mastitis Council Proceedings, New Orleans, LA, pp 28 33 (3) U.S. Department of Health and Human Services (2007) Pasteurized Milk Ordinance, Appendix N, U.S.Foodand Drug Administration, Public Health Service, Publication 229, Washington, DC (4) U.S. Food and Drug Administration, Center for Veterinary Medicine (2009) Data Requirements for Milk Screening Tests Labeled for Testing Milk at the Bulk Tank/Tanker Truck for Drug Residues-Tests with Instrument Readers/Printers Only-Antibiotic Class: -Lactams, Laurel, MD (5) U.S. Food and Drug Administration (Jan. 2010) Coded Memorandum to NCIMS M-a-85 Revision 13 Approved -Lactam Tests and Their Detection Levels, IMS List HFS 626, Vol. 07-4, Washington, DC (6) U.S. Pharmacopeia (2009) 32nd Ed., U.S. Pharmacopeial Convention Inc., Rockville MD, pp 86 90 (7) U.S. Food and Drug Administration (July 31, 1996) Coded Memorandum to NCIMS M-a-85 Rev. 5, Charm II -Lactam Quantitative Assay Screening and Detection Levels, IMS List, HFS 626, Vol. 97-2, Washington, DC (8) Code of Federal Regulations (1994) Title 21, Part 58, U.S. Government Printing Office, Washington, DC (9) XLStat Software, http://www.xlstat.com/en/ho_ Hlt191261487_Hlt191261487_Hlt191261488me (accessed February 2008) (10) ISO 5725-2 (1994) Accuracy (Trueness and Precision) of Measurement Methods and Results, International Organization for Standardization, Geneva, Switzerland, pp 6 9 (11) Youden, W.J., & Steiner, E.H. (1975) Statistical Manual of the AOAC, AOAC INTERNATIONAL, Gaithersburg, MD, pp 33 36 (12) Chu, P., Von Bredow, J., & Carson, M. (2000) Multiresidue Detection of 5 -Lactam Residues in Milk by HPLC, U.S. Food and Drug Administration, Center for Veterinary Medicine, Laurel, MD (13) Jaglan, P., Yein, F., Hornish, R., Cox, B., Arnold, T., Roof, R., & Gilbertson, T. (1992) J. Dairy Sci. 75, 1870 1876 (14) Moats, W., & Raida, H. (1995) J. AOAC Int. 79, 49 54 (15) Schermerhorn, P., Ngoh, M.A., Chu, P., & Von Bredow, J. (Nov. 22, 2000) Procedure for Determination of Cephapirin and Ceftiofur Residues in Bovine Milk Using Liquid Chromatography and Ultraviolet Light Detection, U.S. Food and Drug Administration, Center for Veterinary Medicine, Laurel, MD (16) Zomer, E., Quintana, J., Saul, S., & Charm, S. (1995) J. AOAC Int. 78, 1165 1172 (17) Salter,R.S.,Legg,D.,Ossana,N.,Boyer,C.,Skeemaker,J., Markovsky, R., & Saul, S. (1999) J. AOAC Int. 84, 29 36 (18) Aki, H., Sawawai, N., Yamamoto, K., & Yamamoto, M. (1991) Pharm. Res. 8, 119 122 (19) Van Der Bijl, P., Seifart, H.I., Parkin, D.P., & Mattheyse, F.J. (1988) S. Afr. Medical J. 73, 453 455