Evaluation of the BAX system for detection of Salmonella in selected foods: collaborative study

A multilaboratory study was conducted to compare the automated BAX System to the standard cultural methods for detection of Salmonella in selected foods. Five food types--frankfurters, raw ground beef, mozzarella cheese, raw frozen tilapia fish, and orange juice--at 3 inoculation levels, were analyzed by each method. A sixth food type, raw ground chicken, was tested using 3 naturally contaminated lots. A total of 16 laboratories representing government and industry participated. In this study, 1386 samples were analyzed, of which 1188 were paired samples and 198 were unpaired samples. Of the 1188 paired samples, 461 were positive by both methods and 404 were negative by both methods. Thirty-seven samples were positive by the BAX System but negative by the standard reference method, and 11 samples were positive by standard cultural method and negative by the BAX System. Of the 198 unpaired samples, 106 were positive by the BAX System and 60 were positive by the standard cultural method. A Chi square analysis of each of the 6 food types, at the 3 inoculation levels tested, was performed. For all foods, the BAX System demonstrated results comparable to those of the standard reference methods based on the Chi square results.     

Evaluation of the VIDAS Listeria (LIS) immunoassay for the detection of Listeria in foods using demi-Fraser and Fraser enrichment broths, as modification of AOAC Official Method 999.06 (AOAC Official Method 2004.06)

A multilaboratory study was conducted to compare the VIDAS LIS immunoassay with the standard cultural methods for the detection of Listeria in foods using an enrichment modification of AOAC Official Method 999.06. The modified enrichment protocol was implemented to harmonize the VIDAS LIS assay with the VIDAS LMO2 assay. Five food types--brie cheese, vanilla ice cream, frozen green beans, frozen raw tilapia fish, and cooked roast beef--at 3 inoculation levels, were analyzed by each method. A total of 15 laboratories representing government and industry participated. In this study, 1206 test portions were tested, of which 1170 were used in the statistical analysis. There were 433 positive by the VIDAS LIS assay and 396 positive by the standard culture methods. A Chi-square analysis of each of the 5 food types, at the 3 inoculation levels tested, was performed. The resulting average Chi square analysis, 0.42, indicated that, overall, there are no statistical differences between the VIDAS LIS assay and the standard methods at the 5% level of significance.     

Evaluation of VIDAS listeria monocytogenes II (LMO2) immunoassay method for the detection of Listeria monocytogenes in foods: collaborative study

A multilaboratory study was conducted to compare the VIDAS Listeria monocytogenes II (LMO2) immunoassay and the standard cultural methods for the detection of Listeria monocytogenes in foods. Five food types-vanilla ice cream, brie cheese, cooked roast beef, frozen green beans, and frozen tilapia fish-at 3 levels were analyzed by each method. A total of 26 laboratories representing government and industry participated. In this study, 1404 test portions were analyzed of which 1152 were used in the statistical analysis. There were 448 positive by the VIDAS LMO2 assay and 457 positive by the standard culture methods. A chi2 analysis of each of the 5 food types, at the 3 inoculation levels tested, was performed. The resulting chi2 value, 0.36, indicates that overall, there are no statistical differences between the VIDAS LMO2 assay and the standard methods at the 5% level of significance.     

Evaluation of VIDAS Salmonella (SLM) easy Salmonella method for the detection of Salmonella in a variety of foods: collaborative study

The VIDAS Salmonella (SLM) Easy Salmonella method is a specific enzyme-linked fluorescent immunoassay performed in the automated VIDAS instrument. The VIDAS Easy Salmonella method is a simple 2-step enrichment procedure, using pre-enrichment followed by selective enrichment in a newly formulated broth, SX2 broth. This new method was compared in a multilaboratory collaborative study to the U.S. Food and Drug Administration's Bacteriological Analytical Manual, Chapter 5 method for five food matrixes (liquid egg, vanilla ice cream, spinach, raw shrimp, and peanut butter) and the U.S. Department of Agriculture's Microbiology Laboratory Guidebook 4.04 method for deli turkey. Each food type was artificially contaminated with Salmonella at three inoculation levels. A total of 15 laboratories representing government, academia, and industry, throughout the United States, participated. In this study, 1583 samples were analyzed, of which 792 were paired replicates and 791 were unpaired replicates. Of the 792 paired replicates, 285 were positive by both the VIDAS and reference methods. Of the 791 unpaired replicates, 341 were positive by the VIDAS method and 325 were positive by the cultural reference method. A Chi-square analysis of each of the six food types was performed at the three inoculation levels tested. For all foods evaluated, the VIDAS Easy SLM method demonstrated results comparable to those of the reference methods for the detection of Salmonella.     

3M Petrifilm Staph Express Count plate method for the enumeration of Staphylococcus aureus in selected types of meat, seafood, and poultry: collaborative study

The 3M Petrifilm Staph Express Count plate method was compared with AOAC Official Method 975.55 for the enumeration of Staphylococcus aureus in selected foods. Four foods--cooked, diced chicken; cured ham; smoked salmon; and pepperoni--were analyzed for S. aureus by 12 collaborating laboratories. For each food tested, the collaborators received 8 blind test samples consisting of a control sample, a low inoculation level, a medium inoculation level, and a medium inoculation level with background flora, each in duplicate. The mean log10 counts for the methods were comparable for all 4 foods. The repeatability and reproducibility variances of the 24 h Petrifilm Staph Express Count plate method were similar to those of the 72 h standard method.     

3M Petrifilm Staph Express Count plate method for the enumeration of Staphylococcus aureus in selected dairy foods: collaborative study

The 3M Petrifilm Staph Express Count plate method was compared with AOAC Official Method 975.55 for the enumeration of Staphylococcus aureus in selected foods. Five foods--ice cream, raw milk, yogurt, whey powder, and cheese--were analyzed for S. aureus by 12 collaborating laboratories. For each food tested, the collaborators received 8 blind test samples consisting of a control sample, a low inoculation level, a medium inoculation level, and a medium inoculation level with background flora, each in duplicate. The mean log10 counts for the methods were comparable for all 5 foods. The repeatability and reproducibility variances of the 24 h Petrifilm Staph Express Count plate method were similar to those of the 72 h standard method.     

Roche/BIOTECON Diagnostics LightCycler foodproof L. monocytogenes detection kit in combination with ShortPrep foodproof II Kit. Performance-Tested Method 070401

A method was developed for the detection of L. monocytogenes in food based on real-time polymerase chain reaction (PCR). This advanced PCR method was designed to reduce the time needed to achieve results from PCR reactions and to enable the user to monitor the amplification of the PCR product simultaneously, in real-time. After DNA isolation using the Roche/BIOTECON Diagnostics ShortPrep foodproof II Kit (formerly called Listeria ShortPrep Kit) designed for the rapid preparation of L. monocytogenes DNA for direct use in PCR, the real-time detection of L. monocytogenes DNA is performed by using the Roche/BIOTECON Diagnostics LightCycler foodproof L. monocytogenes Detection Kit. This kit provides primers and hybridization probes for sequence-specific detection, convenient premixed reagents, and different controls for reliable interpretation of results. For repeatability studies, 20 different foods, covering the 15 food groups recommended from the AOAC Research Institute (AOAC RI) for L. monocytogenes detection were analyzed: raw meats, fresh produce/vegetables, processed meats, seafood, egg and egg products, dairy (cultured/noncultured), spices, dry foods, fruit/juices, uncooked pasta, nuts, confectionery, pet food, food dyes and colorings, and miscellaneous. From each food 20, samples were inoculated with a low level (1-10 colony-forming units (CFU)/25 g) and 20 samples with a high level (10-50 CFU/25 g) of L. monocytogenes. Additionally, 5 uninoculated samples were prepared from each food. The food samples were examined with the test kits and in correlation with the cultural methods according to U.S. Food and Drug Administration (FDA) Bacteriological Analytical Manual (BAM) or U.S. Department of Agriculture (USDA)/Food Safety and Inspection Service (FSIS) Microbiology Laboratory Guidebook. After 48 h of incubation, the PCR method in all cases showed equal or better results than the reference cultural FDA/BAM or USDA/FSIS methods. Fifteen out of 20 tested food types gave exactly the same amount of positive samples for both methods in both inoculation levels. For 5 out of 20 foodstuffs, the PCR method resulted in more positives than the reference method after 48 h of incubation. Following AOAC RI definition, these were false positives because they were not confirmed by the reference method (false-positive rate for low inoculated foodstuffs: 5.4%; for high inoculated foodstuffs: 7.1%). Without calculating these unconfirmed positives, the PCR method showed equal sensitivity results compared to the alternative method. With the unconfirmed PCR-positives included into the calculations, the alternative PCR method showed a higher sensitivity than the microbiological methods (low inoculation level: 100 vs 98.0%; sensitivity rate: 1; high inoculation level: 99.7 vs 97.7%; sensitivity rate, 1). All in-house and independently tested uninoculated food samples were negative for L. monocytogenes. The ruggedness testing of both ShortPrep foodproof II Kit and Roche/BIOTECON LightCycler foodproof L. monocytogenes Detection Kit showed no noteworthy influences to any variation of the parameters component concentration, apparatus comparison, tester comparison, and sample volumes. In total, 102 L. monocytogenes isolates (cultures and pure DNA) were tested and detected for the inclusivity study, including all isolates claimed by the AOAC RI. The exclusivity study included 60 non-L. monocytogenes bacteria. None of the tested isolates gave a false-positive result; specificity was 100%. Three different lots were tested in the lot-to-lot study. All 3 lots gave equal results. The stability study was subdivided into 3 parts: long-term study, stress test, and freeze-defrost test. Three lots were tested in 4 time intervals within a period of 13 months. They all gave comparable results for all test intervals. For the stress test, LightCycler L. monocytogenes detection mixes were stored at different temperatures and tested at different time points during 1 month. Stable results were produced at all storage temperatures. The freeze-defrost analysis showed no noteworthy aggravation of test results. The independent validation study examined by Campden and Chorleywood Food Research Association Group (CCFRA) demonstrated again that the LightCycler L. monocytogenes detection system shows a comparable sensitivity to reference methods. With both the LightCycler PCR and BAM methods, 19 out of 20 inoculated food samples were detected. The 24 h PCR results generated by the LightCycler system corresponded directly with the FDA/BAM culture results. However, the 48 h PCR results did not relate exactly to the FDA/BAM results, as one sample found to be positive by the 48 h PCR could not be culturally confirmed and another sample which was negative by the 48 h PCR was culturally positive.     

Evaluation of the assurance GDS for E. coli O157:H7 method and assurance GDS for shigatoxin genes method in selected foods: collaborative study

A multilaboratory collaborative study was conducted to compare the Assurance GDS for E. coli 0157:H7 method and the reference culture methods for the detection of E. coli 0157:H7 in orange juice, raw ground beef, and fresh lettuce. A separate companion assay, the Assurance GDS for Shigatoxin Genes method was also evaluated with the same test portions. Fifteen laboratories participated in the study. A Chi square analysis of each of the 3 food types at the high, low, and uninoculated control levels was performed. For all foods, the Assurance GDS for E. coli O157:H7 method and the Assurance GDS for Shigatoxin Genes method were equivalent to or better than the reference methods.     

Sensitivity and specificity of the BAX for screening/Listeria monocytogenes assay: internal validation and independent laboratory study

In this study to certify the BAX for Screening/Listeria monocytogenes assay (DuPont Qualicon, Wilmington, DE), an internal evaluation was conducted on 16 food types that were simultaneously analyzed with the BAX system (BAX), and the ISO method for the detection of L. monocytogenes (ISO). No statistically significant difference in performance between the BAX and ISO methods was observed. Inclusivity/exclusivity testing showed that the BAX system was able to detect 97 of 97 (100%) of L. monocytogenes strains tested. None of 56 other Listeria species or non-Listeria tested gave a reproducible positive BAX result. Ruggedness testing demonstrated that performance of the assay was not affected by reasonable variability in the operating parameters. BAX was then submitted for independent laboratory validation. In this phase, BAX was compared with standard culture methods for the detection of L. monocytogenes in chicken (USDA-FSIS), crab meat (BAM), and milk (AOAC). This study validated product claims of sensitivity and specificity >98% in accordance with AOAC Performance Tested Method requirements.     

3M Petrifilm Staph Express Count plate method for the enumeration of Staphylococcus aureus in selected types of processed and prepared foods: collaborative study

The 3M Petrifilm Staph Express Count plate method was compared with AOAC Official Method 975.55 for the enumeration of Staphylococcus aureus in selected foods. Five foods--frozen lasagna, custard, frozen mixed vegetables, frozen hashbrowns, and frozen batter-coated mushrooms--were analyzed for S. aureus by 13 collaborating laboratories. For each food tested, the collaborators received 8 blind test samples consisting of a control sample, a low inoculation level, a medium inoculation level, and a medium inoculation level with background flora, each in duplicate. The mean log10 counts for the methods were comparable for all 5 foods. The repeatability and reproducibility variances of the 24 h Petrifilm Staph Express Count plate method were similar to those of the 72 h standard method.     

Evaluation of the GeneQuence DNA hybridization method in conjunction with 24-hour enrichment protocols for detection of Salmonella spp. in select foods: collaborative study

A multilaboratory study was conducted to compare performance of the GeneQuence DNA hybridization (DNAH) method incorporating new 24 h enrichment protocols and reference culture procedures for detection of Salmonella spp. in select foods. Six food types (raw ground turkey, raw ground beef, dried whole egg, milk chocolate, walnuts, and dry pet food) were tested by the DNAH method and by the culture methods of either the U.S. Department of Agriculture-Food Safety and Inspection Service (USDA-FSIS) or the U.S. Food and Drug Administration's Bacteriological Analytical Manual (FDA/BAM). Fifteen laboratories participated in the study. Four of the foods tested (raw ground turkey, dried whole egg, milk chocolate, and dry pet food), showed no statistically significant differences in performance between the DNAH method and the reference procedure as determined by Chi square analysis. Sensitivity rates for the DNAH method ranged from 92 to 100%. The DNAH method, with the specific enrichment protocol evaluated, was found to be ineffective for detection of Salmonella spp. in walnuts. For raw ground beef, results from one trial showed a statistically significant difference in performance, with more positives obtained by the reference method. However, evidence suggests that the difference in the number of positives was likely due to lack of homogeneity of the test samples rather than to DNAH method performance.     

Evaluation of VIDAS Listeria species Xpress (LSX) immunoassay method for the detection of Listeria species in foods: collaborative study

In a multilaboratory study, the effectiveness of an alternative method for rapid screening of Listeria species compared to traditional reference methods was demonstrated in a variety of food products. A collaborative study was conducted to compare the VIDAS Listeria species Xpress (LSX) method and the standard cultural methods for the detection of Listeria species in foods. Six food types were tested: vanilla ice cream, cheddar cheese, raw ground beef, frozen green beans, deli turkey, and cooked shrimp. Each food, inoculated with a different Listeria strain at two levels and uninoculated test portions, was analyzed by each method. A total of 15 laboratories representing government and industry participated. In this study 1134 tests were analyzed in the statistical analysis. There were 490 positives by the VIDAS LSX method using the sample boiling step, 483 positives by the VIDAS LSX method using the Heat and Go system, and 439 positives by the standard culture methods. Overall, the Chi-square result for the VIDAS LSX method with boiling for all foods was 7.25, indicating a significant statistical difference between the VIDAS method and the standard methods at the 5% confidence. For the VIDAS LSX method with the Heat and Go system, the Chi-square result for all foods was 5.37, indicating a significant statistical difference between the VIDAS LSX assay with the Heat and Go system and the standard methods at the 5% level of significance. In both cases, the VIDAS method was more sensitive than the standard methods. The LSX method detects Listeria species in foods with negative or presumptive positive results in a minimum of 30 h compared to at least 5 days for the cultural methods. Based on the results of this collaborative study, it is recommended that the VIDAS LSX method be adopted as an AOAC Official Method for the detection of Listeria species in dairy products, vegetables, seafood, raw meats and poultry, and processed meats and poultry.     

Dry rehydratable film method for rapid enumeration of coliforms in foods (3M Petrifilm Rapid Coliform Count plate): collaborative study

A rehydratable dry-film plating method for coliforms in foods, the 3M Petrifilm Rapid Coliform Count plate method, was compared with the U.S. Food and Drug Administration's Bacteriological Analytical Manual method for nondairy foods and the American Public Health Association's Standard Methods for the Examination of Dairy Products (SMEDP) method for dairy foods. Six food types, vanilla ice cream, cheddar cheese, fresh refrigerated uncooked pasta, wheat flour, prepared frozen macaroni and cheese, and frozen hash browns, were analyzed for coliforms by 11 collaborating laboratories. For each food product tested, the collaborators received 8 blind samples consisting of a control sample and 3 levels of inoculated sample, each in duplicate. The mean log counts for the methods were comparable. The repeatability and reproducibility variances of the Petrifilm Rapid Coliform Count method at 14 and 24 h were not significantly different from those of the standard methods.     

Singlepath Salmonella. Performance-Tested Method 060401

Singlepath Salmonella is an immunochromatographic (lateral flow) assay for the presumptive qualitative detection of Salmonella spp. in food. The AOAC Performance-Tested Method study evaluated Singlepath Salmonella as an effective method for the detection of Salmonella spp. in the following selected foods: dried skimmed milk, black pepper, dried pet food, desiccated coconut, cooked peeled frozen prawns, raw ground beef, and raw ground turkey. When the foods were inoculated with Salmonella spp. at levels ranging from low [0.23-1.08 colony forming units (CFU)/25 g] to high (2.3-6.0 CFU/25 g), a Chi-square value of 0.9 indicated that there was no significant difference between Singlepath Salmonella and the ISO 6579:2002 reference method. Singlepath Salmonella gave a false-positive rate of 7.3% and a false-negative rate of 2.5%. For the inclusivity study, all 105 Salmonella serovars reacted with Singlepath Salmonella. For the exclusivity study, 58 non-Salmonella spp. were tested. There were no cross-reactions with Singlepath Salmonella from these strains.     

Modification of enrichment protocols for TECRA Listeria Visual Immunoassay method 995.22: collaborative study

A collaborative study was conducted to validate new enrichment methods for the TECRA Listeria Visual Immunoassay (TLVIA). These new methods incorporate a newly formulated medium, TECRA Listeria Enrichrment Broth, which does not contain the highly toxic antifungal agent, cycloheximide. The new procedures will provide an alternative to the enrichment procedures described in AOAC Method 995.22. Three food types (raw ground beef, lettuce, and ice cream) were analyzed in the United States, and 2 food types (cooked turkey and cooked fish fillets) were analyzed in Australasia. Thirty collaborators participated in the study, 16 in Australasia and 14 in the United States. With the exception of one batch of ground beef, comparison of the proportion of positive test portions (p > or = 0.05) showed no significant difference between the TLVIA and the reference method for the 5 foods at 3 inoculation levels. For the one batch of naturally contaminated raw ground beef, the TLVIA gave significantly more confirmed positive results than the reference method.     

TECRA Unique test for rapid detection of Salmonella in food: collaborative study

The TECRA Unique Salmonella test uses the principle of immunoenrichment to allow rapid detection of Salmonellae in food. A collaborative study was conducted to compare the TECRA Salmonella Unique test with the reference culture method given in the U.S. Food and Drug Administration's Bacteriological Analytical Manual. Three food types (milk powder, pepper, and soy flour) were analyzed in Australia and 2 food types (milk chocolate and dried egg) were analyzed in the United States. Forty-one collaborators participated in the study. For each of the 5 foods at each of the 3 levels, a comparison showed no significant differences (p > or = 0.05) in the proportion of positive test samples for Unique and that for the reference method using the Chi-square test for independence with continuity correction.     

Comparison of assurance gold salmonella EIA, BAX for screening/Salmonella, and GENE-TRAK Salmonella DLP rapid assays for detection of Salmonella in alfalfa sprouts and sprout irrigation water

The Assurance Gold Salmonella EIA, BAX for Screening/Salmonella, and GENE-TRAK Salmonella DLP rapid assays were compared with official cultural methods described in the Bacteriological Analytical Manual (BAM) for analysis of alfalfa sprouts and sprout irrigation water for the presence of Salmonella. The lower limits of detection of 4 serovars of Salmonella cells (S. tennessee, S. muenchen, S. mbandanka, and S. cubana) in pure culture were determined as approximately log10 2, 5, and 6 for the BAX, GENE-TRAK, and Gold EIA, respectively. Despite its low detection limit, the BAX did not perform as well as the other assays in analyzing contaminated sprouts and sprout irrigation water. For 4 different lots of sprouts and sprout irrigation water samples inoculated with the 4 serovars at low [1-2 colony forming units (CFU/g)] and high (68-180 CFU/g) levels, the BAX detected Salmonella in 58/64 (90.6%) of the samples, compared with 64/64 (100%) by the GENE-TRAK, Gold EIA, and BAM methods. Assay performance was also compared for analysis of naturally contaminated sprouts and sprout irrigation water with 3 lots of alfalfa sprouted seeds associated with different salmonellosis outbreaks. Positive assay results for the naturally contaminated samples were Gold EIA 41, GENE-TRAK 36, BAM 33, and BAX 13.     

BIOTECON diagnostics foodproof Listeria monocytogenes Detection Kit, 5' nuclease in combination with the foodproof ShortPrep II Kit

A method was developed for the detection of Listeria monocytogenes in food. The method is based on real-time PCR using hydrolysis probes (5' Nuclease). This advanced PCR method was designed to reduce the time necessary to achieve results from PCR reactions and to enable the user to monitor the amplification of the PCR product simultaneously, in real-time. After DNA isolation using the BIOTECON foodproof ShortPrep II Kit designed for the rapid preparation of L. monocytogenes DNA for direct use in PCR, the real-time detection of L. monocytogenes DNA is carried out using the foodproof Listeria monocytogenes Detection Kit. The kit provides primers and hydrolysis probes for sequence-specific detection, convenient premixed reagents, and controls for reliable interpretation of results. For the internal comparison study, three different foods (soft cheese, coalfish, and smoked ham) were analyzed, chosen from the 15 food groups recommended by the AOAC Research Institute for detection of L. monocytogenes. From each food, 20 samples were inoculated with a low level (1-10 CFU/25 g) and 20 samples with a high level (10-50 CFU/25 g) of L. monocytogenes. Additionally, five nonspiked samples were prepared from each food. Depending on the matrix, the food samples were examined with the test kits and compared with the cultural methods according to the U.S. Food and Drug Administration's Bacteriological Analytical Manual or the U.S. Department of Agriculture/Food Safety and Inspection Service Microbiology Laboratory Guidebook.     

Modification of the BAX System PCR assay for detecting Salmonella in beef, produce, and soy protein isolate. Performance Tested Method 100201

The BAX System PCR assay for Salmonella detection in foods was previously validated as AOAC Research Institute (RI) Performance Tested Method (PTM) 100201. New studies were conducted on beef and produce using the same media and protocol currently approved for the BAX System PCR assay for E. coli O157:H7 multiplex (MP). Additionally, soy protein isolate was tested for matrix extension using the U.S. Food and Drug Administration-Bacteriological Analytical Manual (FDA-BAM) enrichment protocols. The studies compared the BAX System method to the U.S. Department of Agriculture culture method for detecting Salmonella in beef and the FDA-BAM culture method for detecting Salmonella in produce and soy protein isolate. Method comparison studies on low-level inoculates showed that the BAX System assay for Salmonella performed as well as or better than the reference method for detecting Salmonella in beef and produce in 8-24 h enrichment when the BAX System E. coli O157:H7 MP media was used, and soy protein isolate in 20 h enrichment with lactose broth followed by 3 h regrowth in brain heart infusion broth. An inclusivity panel of 104 Salmonella strains with diverse serotypes was tested by the BAX System using the proprietary BAX System media and returned all positive results. Ruggedness factors involved in the enrichment phase were also evaluated by testing outside the specified parameters, and none of the factors examined affected the performance of the assay.     

Single-laboratory validation of a method for the determination of furan in foods by using static headspace sampling and gas chromatography/mass spectrometry

A headspace gas chromatography/mass spectrometry method was developed and validated in-house for the determination of furan in foods. The method of standard additions with d4-furan as the internal standard was used to quantitate furan. The limit of detection and limit of quantitation (LOQ) values ranged from 0.2 and 0.6 nglg, respectively, in apple juice to 0.9 and 2.9 ng/g, respectively, in peanut butter. Recoveries were obtained at 0.5, 1, 2, and 3 times the LOQ. At 1, 2, and 3 times the LOQ, the recoveries ranged from 89.4 to 108%, and the relative standard deviations ranged from 3.3 to 17.3% for all the matrixes. For apple juice, chicken broth, and infant formula, the averaged coefficients of determination from the linear regression analyses were >0.99 with each food fortified at 0.5, 1, 2, and 3 times the LOQ. The coefficients of determination were >0.99 for green beans and 0.96 for peanut butter with the foods fortified at 1, 2, and 3 times the LOQ. Within-laboratory precision was determined by comparing the amounts of furan found in 18 samples by 2 analysts on different days with different instruments. For most of the foods, the difference between the amounts found by each analyst was <18%. The method was used to conduct a survey of >300 foods. The furan levels found ranged from none detected to 174 ng/g.