WO2005010168A2 - Method and system for detecting chloramphenicol - Google Patents

Abstract

Methods and systems for detecting drugs, particularly amphenicols such as chloramphenicol, in samples such as honey and other foods, are described. An aspect includes a method and system for separating chloramphenicol from compounds that could interfere with an immunoassay for detection of chloramphenicol. In one aspect, chloramphenicol is extracted from the sample by buffering the sample and eluting the sample from a solid phase extraction column with a cyclohexyl functional group. The eluent is tested for the presence of chloramphenicol. False positive results are avoided by using mild organic solvents and a solid phase extraction column to separate interfering substances from the substances to be tested.

Description

METHOD AND SYSTEM FOR DETECTING CHLORAMPHENICOL

Reference to Prior Applications [0001] This application is based on and claims priority from U.S. Provisional Patent Application Serial Number: 60/489,737, filed July 24, 2003, which is hereby incorporated by reference.

Background of the Invention [0002] A phenicols are potent, broad-spectrum antibiotic drugs. The amphenicol family of antibiotics includes such drugs as chloramphenicol, florfenicol, and thiamphenicol. The presence of chloramphenicol in food is a particular problem. Chloramphenicol has been banned for use in food producing animals in many countries. In addition to its potential carcinogenicity, chloramphenicol can cause serious acute reactions, including aplastic anemia, in susceptible individuals and, therefore, residues of chloramphenicol are of particular concern. Recent detection of chloramphenicol in aquaculture products and various other foods, such as honey, has triggered a global response.

[0003] Chloramphenicol is accepted for use only at therapeutic doses for treatment of serious human infections. Chloramphenicol acts primarily by binding to the 50S subunit of the ribosome, preventing the binding of tRNA molecules to the aminoacyl and peptidyl binding sites of the ribosome. Consequently, peptide bonds are not formed when chloramphenicol is present in association with the bacterial ribosome. For this reason, chloramphenicol is often used in laboratories as a specific inhibitor of protein synthesis.

[0004] Due to the unpredictable effects of dose on different patient populations, federal regulations in the United States, Canada and the European Union prohibit the use of chloramphenicol in association with food producing animals and animal-feed products. Recently, the United States Food and Drug Administration (FDA) has increased the sampling of certain imported foods, such as shrimp, crawfish, honey, royal jelly, feed and milk products for the presence of chloramphenicol. FDA took this action after low levels of chloramphenicol were detected in the U.S. and other countries in imported shrimp, crayfish, honey and other foods.

[0005] For detecting chloramphenicol in honey at the desired 0.3 ppb level, FDA has published a protocol as Laboratory Information Bulletin 4281 (LIB No. 4281 Volume 18, No. 5, May 2002. For detecting chloramphenicol in shrimp FDA has published a protocol as Laboratory Information Bulletin 4290 (LIB No. 4290 Volume 18, No. 9, September 2002). Both protocols include the use of Mass Spectrometry.

[0006] Mass Spectrometry methods for detecting chloramphenicol, such as described by the above FDA bulletins, require expensive equipment and multiple organic solvents. Use of inexpensive and simpler methods, such as immunoassays, for rapid detection of chloramphenicol in various foods, including shrimp and honey, such as using ELISA methods or using radiolabeled binders such as radiolabeled antibodies, such as by the Charm II method, have recently become available. In the Charm II assay, radiolabeled analyte is added to the assay tube and competes with drug residue in the sample for binding to anti-chloramphenicol antibody or anti-amphenicol antibody. Labeled drug that is not bound to antibody is removed from the substrate prior to detection using a liquid scintillation counter, for example the Charm Sciences, Inc. 6600 or 7600 instrument. The liquid scintillation counter is used to measure bound ¹⁴C or ³H from the labeled drug. The greater the amount of drug present in the sample, the lower the counts. Another method for detection involves use of a lateral flow strip tests such as by the Charm ROSA (ROSA is a registered trademark of Charm Sciences, Inc. Lawrence, Massachusetts) lateral flow strip test for chloramphenicol in shrimp (Charm ROSA for chloramphenicol in shrimp). In some cases, however, particularly at detection at low levels, for example the 0.3 ppb level, accurate detection using such inexpensive and simpler methods can be difficult. [0007] Current rapid, inexpensive methods for detecting chloramphenicol in shrimp require a minimum of approximately 60 minutes. With the exception of the Charm II radioimmunoassay, such methods also utilize multiple organic solvents for extracting chloramphenicol. For example, the Charm ROSA test strip detects chloramphenicol in shrimp with sensitivity of about 0.2 parts per billion (ppb), in approximately 60 minutes and requires methanol, heptane and ethyl acetate. The RIDASCREEN (RIDASCREEN is a registered trademark of R-Biopharm Darmstadt, Germany) claims to detect chloramphenicol in shrimp with sensitivity of 0.4 ppb, takes 3 hours 45 minutes and requires acetonitrile, ethyl acetate and hexane. The Tecna AB628 CAP ELISA claims to detect chloramphenicol in shrimp with sensitivity of 0.0125-0.5 ppb, takes 2 hours 30 minutes and requires ethyl acetate, isooctane and trichloromethane. [0008] Development of rapid methods for detection of amphenicols, in particular chloramphenicol, in honey has been complicated by a number of factors including the wide variability among honey samples. The dry matter composition of honey generally includes approximately 95-99% sugar, organic acids such as gluconic acid, minerals such as potassium and nitrogenous compounds including enzymes from the salivary glands of bees such as invertase (saccharase), diastase (amylase) and glucose oxidase. Traces of other proteins, enzymes or amino acids, as well as water-soluble vitamins, are thought to result from pollen contamination in honey. However, the majority of substances responsible for honey color and flavor are unknown. It is likely that honey from different botanical origins contain different aromatic and other substances that contribute to the specific color and flavor of particular honeys. In addition, hydroxymemylfurfural (HMF) is known to be in higher concentrations in darker honey. Residues of phenol have also been found in honey, likely the result of use of phenol to remove bees from their hives. Such sample variation may lead to inconsistent results, particularly using an immunoassay. There is, therefore, a need for a simple, reliable, safe, rapid and inexpensive test to detect amphenicol, particularly chloramphenicol, in foods such as honey.

Summary of the Invention [0009] Inexpensive rapid methods for detecting amphenicols, particularly chloramphenicol, in honey have been complicated by false positive results, presumably by interfering compounds naturally found in honey. [0010] Numerous different phenolic compounds have been identified in honey. Darker honey tends to contain more phenolic compounds as compared to lighter honey. Phenolic compounds in honey tend to increase during storage and heating. Similarly, another possible interfering compound, HMF, is a byproduct of fructose decay formed during honey storage or during heating. Residues of phenol may result from treatment of hives. [0011] We have found that certain interfering compounds, such as phenols, and other phenolic compounds and derivatives, such as phenolic acids, may cause false positive results, for example in binding assays such as immunoassays for detecting chloramphenicol. Similarly, such compounds may increase false positive results in detecting other amphenicols and other drugs, depending on the structure of the drug to be detected and the relative affinity of the binder used. We describe herein a method and system for eliminating such interferences. [0012] In one aspect, an extraction is used, alone or in combination with, for example: i) the Charm II radioimmunoassay for amphenicol in honey; ii) the Charm II radioimmunoassay for chloramphenicol in honey; or (iii) the Charm ROSA for chloramphenicol in shrimp. The extraction method separates from the sample the above-described interferences and, thereby, allows for more accurate detection in such diverse and complicated matrices as honey or similar substances and foods. An additional aspect includes removing the interferences from a variety of foods and other sample types without using relatively non-polar and relatively toxic organic solvents, while mamtaining adequate test sensitivity. [0013] In particular, chloramphenicol can be detected in a food sample containing interfering compounds, and false positive results reduced, using an antibody based test, at a detection level of, for example, about 0.3 ppb chloramphemcol or less. The detection procedure includes extracting the chloramphenicol from a sample onto an appropriate column, eluting the chloramphenicol and detecting the chloramphenicol in the eluent using an assay containing a binding partner for chloramphemcol, such as an antibody. The presence of the chloramphenicol in the eluent indicates the presence of the chloramphenicol in the sample and the absence of the chloramphenicol in the eluent indicates the absence of the chloramphemcol in the sample. [0014] The sample can be buffered prior to application to the solid phase extraction column. For example, the sample can be buffered to a pH of about pH 6.5 to about pH 8.5, more particularly about pH 7.0 to about pH 8.0 and in a particular embodiment to about pH 7.4 to about pH 7.6. An example of useful buffering chemicals include a mixture of Trizma Base and Monobasic Potassium Phosphate, for example about 45% to about 60% Monobasic Potassium Phosphate and about 40% to about 55% Trizma Base in water. Final molar concentration of each component can be about 0.05 to about 0.5 molar. [0015] In a further aspect, the solid phase extraction column includes those with cyclohexyl as the functional group, for example a Varian BOND ELUT CH column (BOND ELUT is a registered trademark of Analyticehm International, Inc. Frampton Harbor City, California). [0016] In another aspect, a sample fraction to be tested is eluted by applying two solutions to the column. In this aspect, after the sample, for example the buffered sample, is applied to the column a variety of compounds including chloramphenicol, and possible interfering compounds such as phenolic compounds and derivatives thereof, become bound to the column. The first solution preferably elutes the compounds other than the chloramphenicol to be detected. For example, the first solution may elute phenolic compounds, such as phenolic acids, that may interfere with detection. Next, the second solution is applied to the column and the eluent is collected and tested. [0017] In an embodiment, the polarity of the first solution is greater than the polarity of the second solution. To achieve the correct polarity, various mixtures of solvents can be used. For example, either or both the first and second solutions can include a mixture of methanol and water. When methanol and water are used in both solutions, and the polarity of the first solution is greater than the polarity of the second, the first solution will have a lower concentration of methanol as compared to the second solution. For example, the first solution can include about 10% to about 35%> methanol and the second solution about 65% to about 85% methanol. In a specific example the first solution includes about 20%) methanol and about 80% water and the second solution about 75% methanol and about 25% water. It will be appreciated that a variety of combinations of solvents can be used. Although not necessarily required as a technical matter, for purposes of providing a user-friendly test, the combination of solvents preferably will include relatively non-hazardous, easily disposed solvents such as methanol and water combinations. [0018] The various aspects can be used either in an initial testing procedure or as a confirmation or validation step after an assay has reported a positive result. In either case, examples of useful detection techniques include binding assays such as the various formats for immunoassays well known in the art. Utilizing an immunoassay, an antibody is used that is either cross-reactive to, for example a variety of drugs within a family, such as a variety of amphenicols, or more specific to a particular drug such as chloramphenicol. The eluent from the column is contacted with the antibody and the binding of the drug or drugs from the sample to the antibody is measured to determine a result. [0019] An aspect includes a test system for initially screening for, and then confirming the presence of, chloramphenicol. One example of such a test system includes: an immunoassay for detecting chloramphenicol; a solid phase extraction column; and eluting solvent. If the immunoassay test result is positive the test result is confirmed by repeating the immunoassay with eluent from the column. [0020] Aspects include other methods or means to separate, extract ' or distinguish chloramphenicol from interfering compounds or substances such as phenolic compounds. For example, a variety of methods and devices may be useful to separate or distinguish interfering phenolic compounds, such as selectively derivatizing such compounds so that they no longer interfere. In an example, the result of a separation is a sample fraction and the sample fraction is tested using the detection means such as an immunoassay. [0021] Although not wishing to be constrained by theory, the usefulness of the herein described methods and systems may be related to the structural similarity between amphenicols such as chloramphenicol and a variety of interfering compounds such as phenolic compounds, and derivatives thereof, and HMF. Being structurally similar, the compounds may interfere, for example by binding to the chloramphenicol antibody in competition with chloramphenicol from the sample. Thus, it is desirable, to provide a method and system for eliminating such interferences. An aspect also, therefore, includes a method to confirm positive amphenicol immunoassay results, in particular results from the Charm II radioimmunoassay for chloramphemcol in honey or the Charm II radioimmunoassay for amphenicols in honey. Brief Description of the Drawing [0022] Figure 1 is a graphical comparison of results from honey samples after an initial screen using the Charm π assay for chloramphenicol in honey and following confirmation using a solid phase extraction column.

Figure 2 is a cross-sectional view of a syringe and column device. A sample, such as a buffered sample, is added to the syringe and pushed through the solid phase extraction column.

Details of the Invention [0023] An embodiment includes a method for extracting an amphenicol, particularly chloramphenicol, from a food sample. A food sample, for example a sample of honey or shrimp, can be applied to a solid phase extraction column, for example a silica based column with a functional group useful for amphenicol binding. The chloramphenicol is bound to the column and then eluted. In a particular embodiment, the column includes cyclohexyl as the sorbent- functional group. One column that may be used is a cyclohexyl column, for example a Varian CH BOND ELUT column/cartridge for example, containing about 500 milligrams sorbent. After binding to the column, amphenicol can be eluted with, for example, a methanol/water solution. This simple procedure can remove interferences and avoid the use of potentially hazardous, organic solvents. [0024] When using a relatively non-polar or medium polarity sorbent, such as cyclohexyl, a variety of possible solvents can be used to elute the bound chloramphenicol. Generally such solvents should be relatively polar such as, for example, solutions including methanol, isopropanol, acetonitrile, acetone or water. In a specific embodiment, 2 mL of 75% methanol can be used to elute the chloramphenicol extracted from a honey sample and bound to the column. [0025] In an embodiment, prior to elution of the chloramphenicol, the chloramphenicol is bound to the column along with certain potentially interfering substances. The potentially interfering substances, such as, for example, phenols, are then eluted using solvents such as a water/methanol solution in which the percentage of methanol as compared to the percentage of water is less than in the solution to be used to elute the amphenicol. A particular advantage to using a methanol/water solution is its relatively low toxicity as compared with generally more non-polar solvents. [0026] After elution, the eluent can be evaporated to remove the solvent, for example under nitrogen or by heating on a hot plate. The sample can then be reconstituted, for example with known amphenicol free milk standard, and tested in a chloramphenicol binding assay, for example by a Charm II assay. [0027] The binder for the binding assay is generally either a monoclonal or polyclonal antibody, such as is typically used in an immunoassay. The binder can also be a non- immunogenic binding protein such as an enzyme of other binding proteins possibly isolated from bacteria. Other possible binders include enzymes or other molecules synthesized by bacteria and involved in antimicrobial resistance, for example chloramphenicol acetyltransferase. Still other possible binders include ribosomal binding sites, for example binding sites on the 50S ribosomal subunit upon which chloramphenicol is known to act. Useful binders may also include synthetic binders such as those produced by molecular imprinting. [0028] The detection method includes, in some embodiments, a label on the binder, for example a labeled antibody specific to chloramphenicol or a labeled antibody with cross- reactivity to multiple amphenicols. Binding to the labeled antibody can be detected, for example, by capturing or isolating or otherwise separating labeled bound antibody from labeled unbound antibody. In other examples, for example in the Charm II radioimmunoassay, a label is attached to a drug to produce a radiolabeled drug. In the Charm If assay, radiolabeled chloramphemcol competes with chloramphenicol from the sample for binding to the antibody. The amount of antibody bound by chloramphenicol from the sample is determined by measuring the amount of antibody bound to the radiolabeled chloramphenicol, for example in a scintillation counter. The level of chloramphenicol in the sample is inversely related to the amount of radiolabeled chloramphenicol bound to chloramphenicol antibody. [0029] In other detection methods the label may be attached to a chloramphenicol analogue with binding characteristics, relative to the binder used, that are similar to chloramphenicol or other amphenicols. [0030] In another embodiment a detectable label, for example a gold particle, is bound to an antibody such as an antibody specific to chloramphenicol or an antibody with cross-reactivity to multiple amphenicols. The labeled antibody is applied to a test strip, for example a test strip made of nitrocellulose or other porous material such as POREX (POREX is a registered trademark of Porex Technologies Corp., Fairburn, Georgia). The sample to be tested, subsequent to extraction of amphenicol, is applied to the test strip. Chloramphenicol from the sample will bind to the labeled antibody. Labeled unbound antibody will bind to chloramphemcol, or analogue thereof with similar binding characteristics, previously attached to the strip in a test zone, for example chloramphenicol succinate attached to the test strip in a test zone, for example by a carrier protein such as bovine serum albumin (BSA), keyhole limpet hemocyanin (KLH), ovalbumin or other carrier proteins well known in the art. In embodiments including a control line or zone, labeled bound antibody will continue to flow to such a control zone for capture by, for example, an antibody binding protein such as protein A or other molecules, for example proteins that are capable of binding amphenicol bound, labeled antibody such anti-species antibodies. When a control zone is present, greater binding of labeled antibody in the control zone as compared to the test zone indicates a positive result. When a control zone is not used, binding to the test zone is compared to a known standard. [0031] Possible labels include colored particles, fluorescent labels such as quantum dots, radioisotopes, colloidal gold, other colloidal particles or any of the multitude of labels known in the art including, but not limited to, luminescent labels; chemical labels, such as electroactive agents (e.g., ferrocyanide); enzymes and radiofrequency labels. [0032] In a particular embodiment, in which the sample is honey, the honey is mixed with a buffer, the buffer chosen to adjust the pH of the sample to about pH 7 to about pH 8, prior to applying to the column. One example of such a buffer, which can be used to adjust the pH of a honey sample to between pH about 7.4 and pH about 7.6, is a combination of approximately equal concentrations of Trizma Base and Monobasic Potassium Phosphate at a molar concentration for each of about 0.05 molar to about 0.5 molar. In a specific example, about 47.2% Trizma Base and about 52.5% Monobasic Potassium Phosphate was used at a molar concentration of about 0.0185 molar Monobasic Potassium Phosphate and 0.0188 molar Trizma Base. Although this buffer is particularly useful in preparing a sample of honey, it may also be useful in adjusting the pH of other foods, prior to application to the column. After using varying combinations of, for example, methanol and water to first bind chloramphenicol to the column and then elute chloramphenicol from the column, the sample can be dried, for example under nitrogen or by heating, to evaporate the methanol or other solvents used. The sample can then be reconstituted with, for example, a known amphenicol negative milk. The reconstituted sample is then tested using, for example, the Charm II radioimmunoassay. [0033] In an example of a test to detect chloramphemcol in shrimp, other seafood, or other low moisture foods, an additional step may be required to remove or separate chloramphenicol from the substances in the food that may interfere with binding partner based tests such as immunoassays. For example, shrimp can be combined with methanol, for example 100% methanol, using a CUISINART mixer (CUISLNART is a registered trademark of Conair Corporation Stamford, Connecticut). The blended mixture can then be centrifuged and the supernatant, in the example of shrimp, poured off and applied to the extraction column. [0034] The herein described procedures can be used either as a primary screening method or as a confirmation method. An example of use as a confirmation method is to use an immunoassay, for example the Charm II radioimmunoassay, to screen a sample of honey. A positive result would be confirmed using the herein described extraction procedure. [0035] Examples of other commercially available antibody based tests for which this method may be useful include the, Charm ROSA lateral flow strip test, such as those described in U.S. Patent Nos. 5,985,675, issued November 16, 1999, 6,319,466, issued November 20, 2001 and U.S. patent Application No. 10/289,089, filed November 6, 2002, all of which are incorporated herein by this reference, and various other rapid binding partner based methods including various ELISA methods and other immunoassays including radioimmunoassay well known in the art such as described in Principles and Practice of Immunoassay, Christopher P. Price and David J. Newman 2^nd edition (1997) which is incorporated by reference or as described in U.S. Patent Nos. 4,239,745 and 4,239,852 which are also incorporated herein by this reference. [0036] In another embodiment a sample of honey is pretested for phenolic compounds by adding ferric chloride. Presence of phenol and other phenohc compounds such as derivatives of phenol, particularly oxidized derivatives of phenol, may indicate that a false positive result is more likely. [0037] Serum containing useful anti-chloramphenicol or anti-amphenicol antibodies includes serum from Lampire Biological Laboratories, Pipersville, Pennsylvania, such as anti- chloramphenicol antibody lot number 032635887. In addition, serum may be produced by techniques well-known in the art.

[0038] Example 1

This example relates to the extraction of chloramphenicol from a shrimp sample and detection using ROSA lateral flow detection.

Twenty milliliters (mL) of 100% methanol were added to a 50mL conical tube. Raw-cleaned shrimp were then added to the conical tube until the liquid level reached the 40mL mark on the tube. The contents of the conical tube were then added to a CUISLNART mixer (blade in down position) and blended for one minute before returning the mixture to the same conical tube. The tube was then heated in a large bore heater for 10 minutes at 80°c. After heating the tube was shaken vigorously and then spun down at 3800 revolutions per minute (rpm) in centrifuge for 10 minutes. The supernatant was poured off into a lOOmL beaker to which deionized water was added to the beaker until liquid level reached the 70mL mark. (If solids are observed at this time then recentrifuge at 3800rpm for an additional one minute.) Next, using a 20mL syringe setup column, chloramphenicol was extracted using a Varian CH BOND ELUT column by first pushing 5mL 100% methanol through the column to activate the column. Next, 5mL deionized water was pushed through the column to wash and then the diluted sample was pushed through the column. 5mL deionized water was then pushed through column to wash. Chloramphenicol was then eluted by pushing 4mL of 50% methanol through the column and into an aluminum dish. The extract was then evaporated on a hot plate. Next, 1 mL of dilution buffer (1.35 ml of H₃PO₃ (Phosphoric Acid); 50 gm of Bovine Serum Albumin (Protease Free); and 0.155 ml of Proclin per liter, pH 7.2, was admixed with 10.08 grams of Potassium Phosphate (KH₂PO , Monobasic) and 9.12 gm of Trizma Base) on a dish and the dish scraped to liberate the sample. 300uL (microliters) were then pipetted onto a ROSA lateral flow strip and incubated at 40°C for 8 minutes. Results were read on a ROSA reader using the "SL channel" with a control point of zero.

n=4 Strips/concentration (we tested four different extractions of shrimp, two of which were negative, and two of which were spiked at 0.1 ppb Chloramphemcol) n=2 Strips/ sample (we ran two lateral flow strips per extraction)

Negative Shrimp (two results from each extraction): -781, -652, -824, -801

Shrimp Spiked at O.lppb (two results from each extraction): +1550, +1100, +958, +1438

[0039] Example 2

This example relates to removal of interferences from honey and detection using the Charm II radioimmunoassay.

In this example extraction buffer containing 47.2% Trizma Base and 52.5% Potassium Phosphate Monobasic was prepared and 4.8 grams of the mixture was added to 1000 mL's deionized water (the pH should be 7.5 +/- .10 and the final molar concentration of about 0.0185 molar Monobasic Potassium Phosphate and about 0.0188 molar Trizma Base). 10 grams of honey was the combined with 30 mL of the extraction buffer and the pH was adjusted to 7.5 (using either 0.1 normal sodium hydroxide or 0.1 normal hydrochloric acid, depending on the pH of the particular sample). Next, 5mL 100% methanol followed by 5.0 mL distilled water was pushed through a 20mL syringe setup on a a Varian CH BOND ELUT column. 5 ml of the sample, previously diluted in the extraction buffer, was then pushed through the activated column and the liquid that flowed through column was discarded. The column was then washed with 5 ml of 20% methanol and again the liquid that flowed through the column was discarded. Chloramphenicol was then eluted by slowly pushing through 2 ml of 75% methanol. The eluent was collected and dried under nitrogen (a hot plate can also be used to dry the sample on metal planchette). The dried sample was reconstituted with 1.0 ml Zero Control Standard (milk pretested to be antibiotic free) and tested for the presence of chloramphenicol using the Charm II radioimmunoassay for chloramphenicol in honey. The results result shown in Table 1 were obtained using the above procedure. Results under the column titled CHARM II ASSAY represent test results prior to column extraction. The antibody used in that Charm II Assay was a polyclonal antibody purified from anti-chloramphenicol serum purchased from Lampire Biological Laboratories, Pipersville, Pennsylvania. Results under column titled CHARM II CONFIRMATION ASSAY represent results subsequent to column extraction as described herein. Results under column title HPLC RESULTS represent results from same sample tested on HPLC with a limit of detection of 0.075 ppb.

TABLE 1 Kit Used ATBL-0C4 Control used: Honey Negative Control HNC-002 HPLC CHARM II CONFIRMATION RESULTS SAMPLE CHARM II ASSAY ASSAY LOD 0.075 PPB Honey Negative Control (HNC-002) 1594 1185 1669 1203 1505 1088 1606 1135 1415 1129 1450 1219 AVERAGE 1540 1160 CONTROL POINT (NEG. AVERAGE - 1232 928 20%) lative Control spiked at 903 707 3.2 ppb CAP 963 718 1041 688 876 692 AVERAGE 946 701

SAMPLE 1 1184 1020 1152 950 1127 1030 NOT AVERAGE 1154 1000 DETECTABLE SAMPLE 2 1163 1030 1070 1048 1147 997 NOT 1127 1025 DETECTAVERAGE ABLE SAMPLE 3 1152 1003 1088 963 1058 1033 NOT AVERAGE 1109 10Q0 DETECTABLE SAMPLE 4 991 913 1026 931 AVERAGE 1008 922 > 0.3 PPB SAMPLE 5 1153 1066 992 1141 NOT AVERAGE 1072 1104 DETECTABLE SAMPLE 6 1082 1108 1214 1103 NOT AVERAGE 1148 1105 DETECTABLE SAMPLE 7 1081 960 1126 966 AVERAGE 1104 963 0.15 PPB

In the above results, the control point is the cutoff number between a negative and "retest for positive" result. Test results greater than the control point indicate a negative sample, while results less than or equal to the control point indicate the sample is presumptive positive and needs to be retested using the confirmation method described herein.

[0040] Example 3 Using the Charm II radioimmunoassay for chloramphenicol, we found an increase in false positive results by spiking honey, known not to contain chloramphenicol, with phenol at varying concentrations. As the concentration of phenol was increased, the sample became progressively less negative and finally positive. In this example we spiked chloramphenicol negative honey with varying concentrations of phenol as follows: 100 parts per million (ppm), 250ppm, 500ppm, lOOOppm, 5000ppm. Results using the Charm II radioimmunoassay for amphenicols are shown in Table 2.

TABLE 2

PPM Concentrations o 100 250 500 1000 500C 1285 1326 1244 1165 1064 720 1315 1239 1184 1069 1137 700 1381 1302 1249 1113 1058 671 1294 1270 1308 1053 ni l 810 average 1319 1284 1246 1100 1092 725 B/Bo 1.0 0.97 0.94 0.83 0.83 0.55

Control Point from honey = 1055

Control Point from freeze dried honey ~ 1200

Control Point is the cutoff between positive and negative. Results below the control point are positive at the particular detection level, in this case 0.3ppb chloramphenicol. [0041] Example 4 In non-spiked honey we observed an increase in false positive rates in darker honey. False positives were consistent with phenol positives and confirmed negative by HPLC. Phenol as a potential interference (for example, false positive results for the Charm LT radioimmunoassay) in a sample was determined by addition of 5 drops of Ferric Chloride solution (9g/ 100ml) to 2.0 ml of honey sample. Observing a black/violet precipitate at the solution honey interface indicates phenol is present.

Detailed Description of the Drawings [0042] Figure 1 is a graph comparing test results before and after separating chloramphenicol from interfering compounds using the procedure described in Example 2. The negative sample (as labeled along the X axis) has a B/Bo (as labeled along the Y axis) of 1, indicating no reduction, or inhibition, of binding of radiolabeled chloramphenicol to anti- chloramphenicol antibody. The positive sample has a B/Bo of 0.6 indicating a 40% reduction in binding of radiolabeled chloramphenicol to anti-chloramphenicol antibody. The control point - the B/Bo distinguishing a positive from a negative result - is B/Bo 0.8 (20% reduction in binding). Samples 1, 2, 3, 4, 5, 6, and 7 all showed an increase in B/Bo (move toward negative) after using the confirmation procedure described in Example 2. With the exception of sample 4, the change in B/Bo resulted in an initial positive sample failing to be confirmed positive (chloramphenicol not present above 0.3ppb). These results - positive on initial screen by the Charm II radioimmunoassay and negative (except sample 4) after confirmation - were consistent with results from HPLC. [0043] Figure 2 is a cross-section of a typical syringe/column setup. After activation and washing of the column the buffered sample is added to the syringe 1 and pushed through the bivalve 2 into the column 4 containing sorbent 5 such as a silica based sorbent with cyclohexyl functional group. The syringe 1 and bivalve 2 is attached to the column by an adapter 3. Interfering compounds such as phenols will be bound to the sorbent 5 along with amphenicols such as chloramphenicol . Next a solvent is pushed through the column 4, via the syringe 1. A first solvent can elute the interfering compounds, such as phenols, which are discarded. If amphenicol is in the sample, it can be eluted in the second solvent, which can then be evaporated and tested for the presence of an amphenicol.

Claims

Claims

1. A method for detecting an amphenicol in a sample, the method comprising: a) applying the sample to a solid phase extraction column; b) eluting the amphenicol from the solid phase extraction column; and c) detecting the amphenicol in the eluent, wherein the presence of the amphenicol in the eluent indicates the presence of the amphenicol in the sample and wherein the absence of the amphenicol in the eluent indicates the absence of the amphenicol in the sample.

2. The method of claim 1 wherein the amphenicol detected comprises chloramphenicol.

3. The method of claim 2 wherein the sample comprises honey.

4. The method of claim 2 wherein the sample comprises a mixture of solid and liquid portions and wherein the solid portion of the sample comprises greater than about 50% sugar.

5. The method of claim 3 further comprising buffering the sample to a pH of about pH 6.5 to about pH 8.5, wherein the buffered sample is applied to the column.

6. The method of claim 3 further comprising buffering the sample to a pH of about pH 7.4 to about pH 7.6, wherein the buffered sample is applied to the column.

7. The method of claim 3 further comprising buffering the sample, wherein the buffering comprises combining the sample with a mixture of Trizma Base and Monobasic Potassium Phosphate, wherein the buffered sample is applied to the column.

8. The method of claim 3 wherein the solid phase extraction column comprises cyclohexyl as a functional group.

9. The method of claim 3 wherein said eluting comprises applying a first and second solution to the column, and wherein after the second solution is applied to the column the eluent is collected and tested for the presence of chloramphenicol.

10. The method of claim 9 wherein the polarity of the first solution is greater than the polarity of the second solution.

11. The method of claim 10 wherein the first solution comprises a mixture of methanol and water.

12. The method of claim 10 wherein the second solution comprises a mixture of methanol and water.

13. The method of claim 11 wherein the solid phase extraction column comprises cyclohexyl as a functional group.

14. The method of claim 3 wherein said detecting comprises: a) contacting the eluent with an antibody to chloramphenicol; and b) measuring the binding of chloramphemcol to the antibody.

15. The method of claim 3 wherein the sample applied to the column has been previously tested positive, and wherein the method is used to confirm the previous positive result.

16. The method of claim 3 wherein chloramphenicol in honey is detected at about 0.3 ppb or less.

17. A method for extracting chloramphenicol from a sample of honey, the method comprising: a) buffering the honey; b) applying the honey to a solid phase extraction column; c) eluting the chloramphemcol, said eluting comprising adding a first and second solution, wherein the second solution is less polar than the first solution and wherein eluent resulting from applying the first solution to the column is discarded and wherein eluent from applying the second solution contains chloramphenicol.

18. The method of claim 17 wherein the first solution comprises a mixture of methanol and water.

19. The method of claim 17 wherein the second solution comprises a mixture of methanol and water.

20. The method of claim 17 wherein the column comprises a cyclohexyl functional group.

21. A method for extracting chloramphenicol from honey, said honey including compounds that could interfere with an immunoassay for detecting chloramphenicol, comprising: a) buffering the sample to a pH of about 7.0 to about 8.0; b) applying the sample to a solid phase extraction column comprising cyclohexyl as a functional group; c) eluting the chloramphenicol, said eluting comprising adding a first and second solution, wherein the first and second solutions comprise a combination of methanol and water, and wherein said second solution is less polar than the first solution, and wherein eluent resulting from applying the first solution to the column is discarded and wherein eluent from applying the second solution contains the chloramphenicol.

22. The method of claim 21 further comprising detecting chloramphenicol in the eluent.

23. A test system for confirming the presence of an amphenicol in honey, the test system comprising: a) an immunoassay for detecting the amphenicol; b) a solid phase extraction column; and c) a solvent for eluting amphenicol, wherein a positive immunoassay test result is confirmed by applying the honey sample to the column and repeating the immunoassay with eluent from the column.

24. The test system of claim 23 wherein the amphenicol detected comprises chloramphenicol.

25. The test system of claim 24 wherein the column comprises a cyclohexyl functional group.

26. The test system of claim 25 wherein the eluting solvent comprises a first and second solution and wherein eluent resulting from applying the first solution to the column is discarded and wherein eluent from applying the second solution is tested.

27. The test system of claim 26 wherein the polarity of the first solution is greater than the polarity of the second solution.

28. The test system of claim 26 wherein the first solution comprises a mixture of methanol and water.

29. The test system of claim 26 wherein the second solution comprises a mixture of methanol and water.

30. The test system of claim 26 wherein the first solution and the second solution comprise a mixture of methanol and water and wherein the first solution has a lower concentration of methanol as compared to the second solution

31. The test system of claim 25 further comprising buffering the honey to a pH of about pH 7.0 to about pH 8.0 prior to applying to the column.

32. The test system of claim 25 further comprising buffering the honey to a pH of about pH 7.4 to about 7.6 prior to applying to the column.

33. The test system of claim 32 wherein the buffering comprises combining the honey with a mixture of Trizma Base and Monobasic Potassium Phosphate.

34. The test system of claim 24 wherein chloramphenicol in honey is detected at about 0.3 ppb or less.

35. A test system for detecting an amphenicol in a sample containing phenolic compounds, the test system comprising: a) means to separate the amphenicol from the interfering compounds; and b) means to detect the amphenicol in the sample, whereby the results of said separation means comprises a sample fraction and whereby the sample fraction is tested using the detection means.

36. The test system of claim 35 wherein the amphenicol comprises chloramphenicol and the sample comprises honey.

37. The test system of claim 36 wherein the means to detect the chloramphenicol comprises an immunoassay.

38. The test system of claim 36 further characterized in that the separation is used to confirm the presence of chloramphenicol in a previously positive sample and wherein a separated fraction is used to test for chloramphenicol.

39. The test system of claim 36 wherein the chloramphenicol is detected at about 0.3 ppb or less.