International journal of food microbiology

International Journal of Food Microbiology 124 (2008) 154–163<br /> <br /> <br /> <br /> Contents lists available at ScienceDirect<br /> <br /> <br /> International Journal of Food Microbiology<br /> j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / i j f o o d m i c r o<br /> <br /> <br /> <br /> <br /> Validation of NMKL method No. 136 — Listeria monocytogenes, detection and<br /> enumeration in foods and feed<br /> S. Loncarevic a,⁎, M. Økland a, E. Sehic a, H.S. Norli a, T. Johansson b<br /> a<br /> Section for feed and food microbiology, National Veterinary Institute, P.O.Box 750 Sentrum, 0106 Oslo, Norway<br /> b<br /> Department of Animal Diseases and Food Safety Research, Microbiology Research Unit, Finnish Food Safety Authority Evira, Mustialankatu 3, 00790 Helsinki, Finland<br /> <br /> <br /> <br /> <br /> A R T I C L E I N F O A B S T R A C T<br /> <br /> Article history: A collaborative study was organised to define the performance characteristics of the revised NMKL Method<br /> Received 20 December 2007 No.136 “Listeria monocytogenes. Detection and enumeration in foods”. Chromogenic L. monocytogenes specific<br /> Received in revised form 17 March 2008 plating medium, Agar Listeria according to Ottaviani and Agosti (ALOA) was introduced in the revised<br /> Accepted 17 March 2008<br /> method in order to improve the sensitivity and specificity of the method, and to shorten the analysis time.<br /> Efficacy of ALOA One Day from AES (ready-to-use agar in bottles), Listeria Chromogenic Agar (Agosti and<br /> Keywords:<br /> Listeria monocytogenes<br /> Ottaviani Listeria agar) from Lab M (LCA) (dehydrated powder), Chromogenic Listeria Agar Plates from Oxoid<br /> Validation (OCLA) (ready-to-use plates) and L. monocytogenes blood agar medium LMBA from Lab M (dehydrated<br /> Detection powder) were tested. Three types of food matrices (vacuum-packed hot-smoked salmon, soft cheese and<br /> Enumeration cooked ham) and one feed matrix (wheat grain) inoculated with two levels of L. monocytogenes with or<br /> without L. innocua were used in the study. A total of 24 samples were analysed both in the detection and<br /> enumeration part of the study by 18 and 17 Nordic laboratories, respectively.<br /> The sensitivities of ALOA, LCA, OCLA and LMBA in the detection of L. monocytogenes in food samples after one-<br /> step enrichment (Half-Fraser) were 94.4–96.4% and after two-step enrichment (Half-Fraser followed by Fraser)<br /> 97.7–100%. For wheat grain the respective figures were 84.7–88.9% and 90.3–93.1%, respectively. The precision<br /> characteristics were generally good for the enumeration of L. monocytogenes in the food samples with high<br /> levels of inoculation. Several poor values obtained from the food samples with low levels of inoculation<br /> probably reflect high uncertainty of measurement when less than 10 cfu/g was counted. Poor values obtained<br /> from the wheat grain samples by any of the media evaluated were due to poor precision for feed samples.<br /> According to the study, the revised NMKL Method No.136, 4th ed. showed excellent results in the detection and<br /> satisfactory results in the enumeration of L. monocytogenes in foods. The results for the detection of L.<br /> monocytogenes in wheat grain were good, but the method cannot be recommended for the enumeration of L.<br /> monocytogenes in feed-stuffs.<br /> Any one of the media evaluated can interchangeably be used as an obligatory isolation medium for the<br /> detection and enumeration of L. monocytogenes in foods, and for the detection in feed-stuffs. The L.<br /> monocytogenes specific plating media that were evaluated shorten the time of analysis and significantly reduce<br /> the work load. The detection of positive samples mostly after Half-Fraser enrichment, reduces the analysis time<br /> further, and makes it possible to skip the secondary enrichment. However, secondary enrichment cannot be<br /> totally left out, because samples with low levels of L. monocytogenes, with high levels of competing flora, and<br /> with injured L. monocytogenes, do need secondary enrichment.<br /> © 2008 Elsevier B.V. All rights reserved.<br /> <br /> <br /> <br /> <br /> 1. Introduction blood agar medium LMBA. Later on, it was decided to include LMBA in<br /> the study on a voluntary basis.<br /> Policy of the Nordic Committee on Food Analysis (NMKL) is to In the revised method, obligatory Listeria specific plating media,<br /> publish collaboratively validated methods. Thus the revised version of Oxford and PALCAM, were replaced by a chromogenic L. monocytogenes<br /> NMKL method No.136, 4th. Ed.: “Listeria monocytogenes. Detection and specific plating medium, ALOA, in order to improve the sensitivity and<br /> enumeration in foods” was validated collaboratively. It was agreed specificity of the method, and to shorten the analysis time. The other<br /> that the solid selective media included in the study would be Agar selective plating medium was made optional. Another important change<br /> Listeria according to Ottaviani and Agosti (ALOA) and L. monocytogenes in the revised method was the replacement of the LBI, primary selective<br /> enrichment broth, with Half-Fraser, and the secondary enrichment<br /> ⁎ Corresponding author. Tel.: +47 23 21 62 48; fax: +47 23 21 62 02. broth (LBII or Fraser broth) with one obligatory broth, Fraser broth. A<br /> E-mail address: semir.loncarevic@vetinst.no (S. Loncarevic). third important change was plating out the primary enrichment culture<br /> <br /> 0168-1605/$ – see front matter © 2008 Elsevier B.V. All rights reserved.<br /> doi:10.1016/j.ijfoodmicro.2008.03.032<br />  S. Loncarevic et al. / International Journal of Food Microbiology 124 (2008) 154–163 155<br /> <br /> <br /> as well as the secondary one, to improve further the sensitivity of the and homogeneity of the test materials. The samples were packed in<br /> method. In addition, the semi-quantitative method was replaced by a boxes with cooling elements and sent by courier to the participating<br /> quantitative method. A quantitative method was urgently needed, as laboratories. A vial containing water was added to each package box in<br /> quantitative microbiological criteria for L. monocytogenes became valid order to measure the temperature of the samples upon delivery. The<br /> in the EU at the beginning of the year 2006. date and time of arrival of the package, as well as the temperature of<br /> The aim of the collaborative study was to define the performance the water, was registered by each participating laboratory.<br /> characteristics of the revised method in the detection and enumera-<br /> tion of L. monocytogenes in three food matrices and one feed matrix, at 2.2. Isolation media<br /> different levels of contamination with or without L. innocua. Two<br /> forms of ALOA and a modification of ALOA, i.e. Chromogenic Listeria The participating laboratories were supplied with the solid,<br /> agar plates (OCLA), were compared: ALOA as ready-to-use agar in selective media compared:<br /> bottles with supplements and as dehydrated powder with supple-<br /> • ALOA One Day® (ALOA) — ready-to-use agar in bottles with<br /> ments (LCA), and OCLA as ready-to-use plates. In addition, most of the<br /> supplements (AES laboratoire, France)<br /> laboratories included LMBA in the study on a voluntary basis.<br /> • Listeria Chromogenic Agar (Agosti and Ottaviani Listeria agar) (LCA) —<br /> dehydrated powder with supplements (HAL010, LAB M Ltd., Lancashire,<br /> 2. Materials and methods<br /> UK)<br /> • Chromogenic Listeria Agar Plate (OCLA) — ready-to-use plates<br /> 2.1. Trial design<br /> (PO5165A, Oxoid Ltd., Basingstoke, UK)<br /> • LMBA — dehydrated powder with supplements (LAB172, LAB M Ltd.,<br /> Nineteen Nordic laboratories confirmed participation in the study.<br /> Lancashire, UK).<br /> However, 18 and 17 laboratories participated throughout the<br /> detection and enumeration part of the study, respectively. Each The laboratories were recommended to follow the manufacturers'<br /> laboratory was supplied with the description of the trial design, a instructions for the preparation of plates.<br /> copy of the revised NMKL Method No. 136, 4th ed., as well as test<br /> report sheets and test reporting tables for the registration of results 2.3. Method description<br /> for each medium used. All collected information and results were<br /> required to be sent to the leader of the trial. Detailed information The samples were analysed according to the revised NMKL Method<br /> registered in the test reports ensure that all test materials were No. 136, 4th ed., 2004 “Listeria monocytogenes. Detection and<br /> treated under the same conditions by each participating laboratory Enumeration in foods”.<br /> and indicate variations that could influence on the final results. The The detection of L. monocytogenes requires the following successive<br /> pre-trial tests were performed by the organising laboratory to define stages:<br /> suitability of the test materials and design of the trial.<br /> i) Pre-treatment procedure of samples according to a standar-<br /> Food and feed matrices used in the study were as follows: soft,<br /> dised procedure to obtain homogenised samples for detection<br /> white mould Brie cheese made from pasteurised cow's milk with 60%<br /> and/or enumeration.<br /> fat, vacuum-packed hot-smoked salmon, cooked vacuum-packed<br /> ii) Primary enrichment in an enrichment broth with reduced<br /> ham with 4% fat and wheat grain collected from different producers.<br /> selectivity (Half-Fraser broth) at 30 °C for 24 h.<br /> The matrices were checked for the absence of Listeria spp according<br /> iii) Secondary enrichment of a culture from Half-Fraser broth in an<br /> to the revised NMKL Method No. 136, 4th ed., before they were<br /> enrichment broth with full selectivity (Fraser broth) at 37 °C for<br /> included in the study. The food matrices were cut in small pieces,<br /> 48 h.<br /> thoroughly mixed and divided into portions of 10 and 25 g,<br /> iv) Plating out on ALOA and on optional solid selective medium. In<br /> respectively. The wheat grain samples were thoroughly mixed and<br /> this study the enrichment cultures were plated out on ALOA, LCA,<br /> also divided into portions of 10 and 25 g. The sample portions were<br /> OCLA and LMBA media and incubated at 37 °C for 24 and 48 h.<br /> coded with capital letters for the qualitative part of the study, and<br /> v) Sub-culturing of five (if available) presumptive L. monocytogenes<br /> small letters for the quantitative part and refrigerated at + 8 °C until<br /> colonies on blood agar (BA) plates for the detection of β-<br /> inoculation and shipment to the laboratories. The samples were<br /> haemolysis. At least one typical or suspect pure culture and four<br /> distributed to the laboratories as “blind” (without disclosing the type<br /> further pure cultures one by one, if the first one is not confirmed<br /> of sample or level of contamination).<br /> to be L. monocytogenes, is further confirmed by catalase test<br /> Test materials were inoculated with low or high levels of L.<br /> (optional), Gram staining (optional), rhamnose and xylose tests<br /> monocytogenes, and low or high levels of both L. monocytogenes and L.<br /> or alternatively by validated commercial identification tests,<br /> innocua. Samples inoculated only with L. innocua, were included as<br /> according to the revised method.<br /> negative controls. The levels of inoculation of the samples for the<br /> detection and enumeration part of the study are shown in Table 1. A<br /> The enumeration of L. monocytogenes requires the following<br /> total of 24 samples (25 g portions) for the qualitative part of the study,<br /> successive stages:<br /> and 24 samples (10 g portions) for the quantitative part, were analysed<br /> by each participating laboratory. Organising laboratory performed the i) Preparation of initial suspension in buffered peptone water<br /> required tests prior and during collaborative trial to ensure stability (BPW).<br /> <br /> <br /> Table 1<br /> Level (cfu/25 g or cfu/g) of inoculation of the samples with L. monocytogenes (LM) and a mixed culture of L. monocytogenes and L. innocua (LI)<br /> <br /> Listeria inoculate Detection part/mean cfu/25 g (range) Enumeration part/mean cfu/g (range)<br /> <br /> Low level High level Low level High level<br /> LM 18 (12–25) 4500 (3000–6250) 160 (152–185) 16,000 (15,200–18,500)<br /> LM 18 (12–25) 4500 (3000–6250) 160 (152–185) 16,000 (15,200–18,500)<br /> LI 52 (30–80) 13,000 (7500–20,000) 460 (400–570) 46,000 (40,000–57,000)<br /> LI-neg.contol 52 (30–80) 13,000 (7500–20,000) 460 (400–570) 46,000 (40,000–57,000)<br /> 156 S. Loncarevic et al. / International Journal of Food Microbiology 124 (2008) 154–163<br /> <br /> <br /> ii) Surface plating of the initial suspension and further decimal 25 cfu/25 g or 3000–6250 cfu/25 g of L. monocytogenes — either alone<br /> dilutions onto ALOA and in this study on LCA, OCLA and LMBA or in combination with L. innocua (Table 3). After the two-step<br /> media in addition and incubation at 37 °C for 24 and 48 h. In enrichment procedure, the sensitivities of the media varied from 97.7%<br /> this study two successive dilutions and duplicate plates were for LCA to 100% for LMBA. The sensitivities for wheat grain were lower,<br /> used, although according to the final method description and varied from 90.3% for OCLA to 93.1% for ALOA. Already after one-<br /> (Johansson and Loncarevic, 2007) it is not obligatory to use step enrichment in Half-Fraser broth, the sensitivities for the detection<br /> duplicate plates, if two successive dilutions are plated out. of L. monocytogenes in food samples were high: 94.4% for LCA as the<br /> iii) Counting of the presumptive colonies on each plate with less lowest value, and 96.4% for LMBA as the highest value. The<br /> than 100 CFU. corresponding figures for wheat grain were 84.7% for LCA and 88.9%<br /> iv) Pure-culturing for β-haemolysis detection and confirmation as for ALOA. After Fraser enrichment, more positives were detected. But,<br /> described in the detection part of the method in addition, positives already obtained after Half-Fraser enrichment,<br /> v) Calculation of the final concentration of L. monocytogenes (CFU/g were lost, especially those from ham and wheat grain samples<br /> or ml). inoculated with both L. monocytogenes and L. innocua by all the media<br /> used (Table 3). However, there was no significant difference in the<br /> 2.4. Statistical evaluation of data number of reported positive results from the samples inoculated with<br /> both L. monocytogenes and L. innocua, and those with only L.<br /> The results of the collaborative study were statistically evaluated monocytogenes. One laboratory did not isolate pure cultures and<br /> according to ISO 16140 (Anonymous, 2003). For qualitative results the confirmed the typical or suspected colonies with much background<br /> sensitivities and specificities were calculated. For quantitative results flora after Fraser enrichment from six samples already detected as<br /> the main robust estimators used for calculating repeatability and positive after Half-Fraser enrichment. It remained uncertain, whether<br /> reproducibility were: these samples would have been detected as positive by confirmation.<br /> Therefore these results were regarded as negative and thus the figures<br /> • The median<br /> for the sensitivities after Fraser enrichment were reduced. There was<br /> • The repeatability standard deviation (sr)<br /> no significant difference between the numbers of samples correctly<br /> • The repeatability limit (r)<br /> reported, whether sheep or bovine blood with sodium citrate as<br /> • The standard deviation of reproducibility (sR)<br /> anticoagulant was used for preparation of the LMBA medium.<br /> • The reproducibility limit (R)<br /> There were no significant differences between the plating media in<br /> The repeatability limit (r) is defined as the absolute difference the detection of L. monocytogenes in food matrices or wheat grain after<br /> between two independent single test results obtained using the same primary (Half-Fraser) or secondary (Fraser) enrichment, or the whole<br /> method on identical test material in the same laboratory by the same procedure according to McNemar's test (used GraphpadSoftware). The<br /> operator using the same apparatus within shortest feasible time P values calculated with McNemar's test where less than 0.5. On the<br /> interval. The r value should not exceed in more than 5% of the cases. If chromogenic media, L. monocytogenes colonies with a typical green–<br /> the results exceed r, the results should be considered suspect. blue appearance could be observed already after 24 h of incubation,<br /> The reproducibility limit (R) is defined as the absolute difference but often 48 h was needed for the formation of an opaque halo around<br /> between two single test results of the two test results on the normal the colonies. On LMBA, typical colonies might be separated from<br /> scale obtained using the same method on identical test material in atypical ones already after 24 h. However, 48 h of incubation was<br /> different laboratories with different operators using different appara- preferred for the formation of typical β-haemolysis.<br /> tus. The R values should not exceed in not more than 5% of cases. The sensitivities of the media in detecting L. monocytogenes in the<br /> food samples with low levels of inoculation, ranged from 89.8% for LCA<br /> 2.5. Results excluded from further analyses<br /> <br /> Results were excluded by the following criteria: a) the test materials Table 2<br /> were exposed to temperature abuse during shipment; b) the laboratory Number of L. monocytogenes positive samples detected after confirmation by each<br /> had deviated from the specified instructions for the collaborative study, laboratory (N = 18)<br /> or c) the performance of the laboratory was questionable due to large Lab. code Half-Fraser Fraser<br /> numbers of false positives or false negatives — more than would be<br /> ALOA LCA OCLA LMBA ALOA LCA OCLA LMBA<br /> expected by chance.<br /> 1 15 15 15 15 15 15 15 15<br /> 2 16 16 16 16 11(1)⁎ 15 15 15(1)⁎<br /> 3. Results 3 16 15 15 16^ 16 16 16 16^<br /> 4 16 15 16 16 16 16 16 16<br /> 3.1. The detection part of the study 5 10 8 11 – 13 10 13 –<br /> 6 15 15 15 15 16 16 16 16<br /> 7 16 15 16 14 16 16 16 15<br /> 3.1.1. General results 8 16 16 16 16 16 16 16 16<br /> Results were received from 18 of 19 participating laboratories, 9 15 15 15 15^ 15 15 14 15^<br /> including the organising laboratory. 16 laboratories used the LMBA 10 15 15 16 15 16 15 14 15<br /> medium on voluntary basis. 13 laboratories received their samples within 11 16 16 16 16 16 16 16 16<br /> 12 13 14 15 15^ 13 13 14 15^<br /> 24 h after shipment, and five within 48 h. Two of these started the analyses<br /> 13 14 14 11(1)⁎ 10 16 16 13 16<br /> the day after the arrival of the samples. The other laboratories started the 14 14 14 14 15 14 14 15 14<br /> analyses immediately upon receipt of the samples. Two laboratories 15 15 14 14 14^ 14 14 14 14^<br /> reported parcel temperatures above 10 °C (12° and 12.1 °C). 17 15 16 16 16^ 16 15 16 16^<br /> 18 16(1)⁎ 16(1)⁎ 16(1)⁎ – 16(1)⁎ 16(1)⁎ 16(1)⁎ –<br /> The laboratories (N = 18) analysed 16 L. monocytogenes positive<br /> 19 16 16 16 16^ 16 16 16 16^<br /> samples. The results of each laboratory are shown in Table 2. Total no. detected 269 265 269 240 271 270 271 246<br /> (Total no. positives) (288) (288) (288) (256) (288) (288) (288) (256)<br /> 3.1.2. Sensitivity Results were not obtained from laboratory 16.<br /> ALOA, LCA, OCLA and LMBA were all highly sensitive in the ⁎False positive detected in a negative control sample. ^Laboratories reporting use of<br /> detection of L. monocytogenes in food samples inoculated with 12– bovine blood.<br />  S. Loncarevic et al. / International Journal of Food Microbiology 124 (2008) 154–163 157<br /> <br /> <br /> Table 3<br /> The overall sensitivities and specificities of ALOA, LCA, OCLA and LMBA in the detection of L. monocytogenes from cheese, salmon and ham samples, and from wheat grain<br /> <br /> Matrix Half-Fraser enrichment positives (N, %) Fraser enrichment more positives (+)/lost Half-Fraser + Fraser enrichment positives (N, %)<br /> positives (−)<br /> <br /> ALOA LCA OCLA LMBA ALOA LCA OCLA LMBA ALOA LCA OCLA LMBA<br /> Cheese 70 69 69 61 +1 +1 +1 +3 71 70 70 64<br /> 97.2 95.8 95.8 95.3 −1⁎ −1 0 −1 98.6 97.2 97.2 100<br /> Salmon 69 70 71 64 +1 0 +1 0 70 70 72 64<br /> 95.8 97.2 98.6 100 −2⁎ 0 0 0 97.2 97.2 100 100<br /> Ham 66 65 66 60 +5 +6 +5 +4 71 71 71 64<br /> 91.7 90.3 91.7 93.8 −2⁎ −3⁎ −5⁎ −4 98.6 98.6 98.6 100<br /> <br /> Total 205 204 206 185 212 211 213 192<br /> Sensitivity % 94.9 94.4 95.4 95.8 95.8 95.8 96.3 97.4 98.1 97.7 98.6 100<br /> FP/NC 1/108 1/108 2/108 0/96 2/108 1/108 1/108 1/96 2/108 1/108 2/108 1/96<br /> Specificity % 99.1 99.1 98.1 100 98.1 99.1 99.1 99.0 98.1 99.1 98.1 99.0<br /> Wheat grain 64 61 63 55 +3 +4 +2 +4 67 65 65 59<br /> −3⁎ −2 −2 +4<br /> Sensitivity % 88.9 84.7 87.5 85.9 93.0 90.3 90.3 92.2 93.1 90.3 90.3 92.2<br /> FP/NC 0/36 0/36 0/36 0/32 0/36 0/36 0/36 0/32 0/36 0/36 0/36 0/32<br /> Specificity % 100 100 100 100 100 100 100 100 100 100 100 100<br /> <br /> The number of cheese, salmon, ham and wheat grain samples was 72 for ALOA, LCA and OCLA, 64 for LMBA and the total number of food samples was 216 and 192, respectively.<br /> ⁎One laboratory reported typical but unconfirmed colonies with much background flora in six samples (two salmon and ham samples, one cheese sample and one wheat grain<br /> sample). They were detected as positive after Half-Fraser enrichment.<br /> FP/NC = false positive/negative control.<br /> <br /> <br /> <br /> to 93.8% for LMBA after primary enrichment, and from 95.4% for LCA to of L. monocytogenes and L. innocua, both after Half-Fraser and Fraser<br /> 100% for LMBA after the whole procedure (Table 4). The corresponding enrichment, excluding ALOA after both the enrichment steps (Table 6).<br /> figures for wheat grain ranged from 75.0% for LCA to 83.3% for ALOA LMBA was the only medium on which also the samples inoculated<br /> and OCLA after primary enrichment and from 83.3% for LCA to 88.9% with low levels of L. monocytogenes and L. innocua were detected as<br /> for ALOA and OCLA after the whole procedure (Table 4). positive, both after Half-Fraser and Fraser enrichment, whereas both<br /> For food samples with high levels of inoculation, the respective figures ALOA and LCA each gave two false negatives after Fraser enrichment in<br /> ranged from 98.1% for ALOA to 99.0% for LMBA after primary enrichment two laboratories.<br /> and from 99.0% for ALOA and OCLA to 100% for LMBA after the whole L. monocytogenes was easiest to detect in salmon samples<br /> procedure. The corresponding figures for wheat grain ranged from 91.7% compared to the other three sample matrices (Table 3). The<br /> for OCLA to 94.4% for ALOA and LCA after primary enrichment and from sensitivities of LMBA, OCLA and LCA after enrichment in Half-Fraser<br /> 91.7% for OCLA to 100% for LMBA after the whole procedure (Table 4). and Fraser broth, were higher for salmon than for the other matrices.<br /> However, ALOA gave one more positive result for cheese and ham<br /> 3.1.3. Specificity samples than for salmon samples, and so did LCA for ham samples<br /> Each laboratory received eight negative samples (inoculated with L. after the two-step enrichment.<br /> innocua). Nine false positive results were reported. Six of them were One false positive result was reported for a negative control sample<br /> reported by one laboratory on each chromogenic medium; LMBA was for salmon with a high level of L. innocua after enrichment in Half-<br /> not used (Lab 18). The specificity varied from 98.1% to 100% (Table 3). Fraser broth and plating on OCLA (Table 6).<br /> For LMBA, only one false positive result was obtained, while ALOA gave<br /> three, and LCA and OCLA two false positives, respectively.<br /> <br /> Table 4<br /> 3.1.4. Results for each matrix<br /> The overall sensitivities of ALOA, LCA, OCLA and LMBA in the detection of L.<br /> monocytogenes in food samples and in wheat grain per level of inoculum<br /> 3.1.4.1. Cheese. There were only slight differences between ALOA, LCA,<br /> Inoculum level Half-Fraser enrichment Half-Fraser + Freser<br /> OCLA and LMBA in the detection of L. monocytogenes in positive<br /> positives enrichmentpositives<br /> cheese samples of different inoculation levels, with or without L.<br /> innocua (Table 5). After Half-Fraser enrichment, the media detected ALOA LCA OCLA LMBA ALOA LCA OCLA LMBA<br /> <br /> more or an equivalent number of positives in the samples inoculated Food samples<br /> Low level<br /> with L. monocytogenes and L. innocua, compared to the respective<br /> Positives (N) 99 97 99 90 105 103 106 96<br /> samples inoculated with only L. monocytogenes, excluding one result Sensitivity (%) 91.7 89.8 91.7 93.8 97.2 95.4 98.1 100<br /> for OCLA. However, after Fraser enrichment, ALOA, LCA and LMBA lost High level<br /> one positive result from those already obtained after Half-Fraser Positives (N) 106 107 107 95 107 108 107 96<br /> enrichment. A combination of LCA and LMBA resulted in a sensitivity Sensitivity (%) 98.1 99.0 99.0 99.0 99.0 100 99.0 100<br /> <br /> of 100% already after Half-Fraser enrichment. Wheat grain<br /> One false positive result was reported for negative control samples Low level<br /> for cheese with a high level of L. innocua after enrichment in Fraser broth Positives (N) 30 27 30 25 34 34 33 30<br /> and plating on ALOA and LMBA (Table 5). Furthermore, one laboratory Sensitivity (%) 83.3 75.0 85.3 78.1 94.4 94.4 91.7 93.8<br /> High level<br /> (Lab 18) consistently reported false positive results on all the media used<br /> Positives (N) 32 30 32 27 35 35 33 32<br /> for that sample, both after Half-Fraser and Fraser enrichment (Table 5). Sensitivity (%) 88.9 83.3 88.9 84.4 97.2 97.2 91.7 100<br /> <br /> The number of food samples per level was 108 for ALOA, LCA and OCLA, and 96 for<br /> 3.1.4.2. Salmon. All the salmon samples inoculated with only L. LMBA.<br /> monocytogenes were detected as positive on ALOA, LCA, OCLA and The number of wheat grain samples per level was 36 for ALOA, LCA and OCLA, and 32 for<br /> LMBA. The same applies to the samples inoculated with high levels LMBA.<br /> 158 S. Loncarevic et al. / International Journal of Food Microbiology 124 (2008) 154–163<br /> <br /> <br /> Table 5<br /> Detection of L. monocytogenes in cheese samples with low or high levels of L. monocytogenes and low or high levels of both L. monocytogenes and L. innocua after enrichment in Half-<br /> Fraser followed by Fraser broth<br /> <br /> Inoculum level Half-Fraser enrichment positives Fraser enrichment more positives (+)/ Half-Fraser + Fraser enrichment positives<br /> lost positives (−)<br /> <br /> ALOA LCA OCLA LMBA ALOA LCA OCLA LMBA ALOA LCA OCLA LMBA<br /> LM H 17 17 18 15 +1 +1 0 +1 18 18 18 16<br /> LM L 17 16 15 17 0 0 +1 +1 17 17 17 16<br /> LM H + LI H 18 17 16 18 −1 0 0 0 18 18 17 16<br /> LM L + LI L 17 18 15 18 0 −1 0 +1/−1 18 17 18 16<br /> Total 70 69 69 61 0 0 +1 +2 71 70 70 64<br /> Sensitivity (%) 97.2 95.8 95.8 95.3 97.2 95.8 97.2 98.4 98.6 97.2 97.2 100<br /> FP/NC (N) 1/36 1/36 1/36 0/32 2/36 1/36 1/36 1/32 2/36 1/36 1/36 1/32<br /> Specificity (%) 97.2 97.2 97.2 100 94.4 97.2 97.2 96.9 94.4 97.2 97.2 96.9<br /> <br /> The number of L. monocytogenes positive cheese samples at each inoculum level was 18 for ALOA, LCA and OCLA, and 16 for LMBA. The total number of L. monocytogenes positive<br /> samples was 216 and 192, respectively.<br /> LM: L. monocytogenes; LI: L. innocua; H: high; L: low; FP = false positive; NC = negative control.<br /> <br /> <br /> <br /> <br /> 3.1.4.3. Ham. All the ham samples inoculated with a high level of L. No false positive results were reported for wheat grain control<br /> monocytogenes were detected as positive on ALOA, LCA, OCLA and samples.<br /> LMBA, both after Half-Fraser and Fraser enrichment. The same applies<br /> to the samples inoculated with high or low levels of L. monocytogenes 3.1.5. Confirmation<br /> and L. innocua after Half-Fraser enrichment (Table 7). However, after In the detection part of the study, morphological and traditional<br /> Fraser enrichment, 1–3 positives already obtained were lost for each biochemical tests for the confirmation of L. monocytogenes were<br /> medium at each inoculation level because of overgrowth by L. innocua. performed according to the draft method by 11 laboratories. Seven<br /> A low level of L. monocytogenes was difficult to detect after Half-Fraser laboratories used commercial identification tests. In the detection part<br /> enrichment by all the media, and 4–7 positives were lost per medium. of the study confirmation was performed using traditional biochem-<br /> Enrichment in Fraser broth improved the detection and resulted in a ical tests by nine laboratories (catalase test — optional, 7 labs; gram<br /> sensitivity of 100% for LMBA, while each chromogenic medium failed staining — optional, 3 labs; xylose and rhamnose test, 9 labs), and<br /> to detect one of the positive samples (result of one laboratory). using validated commercial biochemical identification tests by six<br /> Excluding ALOA, the sensitivities of the media in detection of L. laboratories (Accu-Probe, 3 labs and API Listeria, 3 labs).<br /> monocytogenes in ham samples were lower than in cheese and salmon<br /> samples after Half-Fraser and Fraser enrichment and after both combined. 3.1.6. Previous experience and use of media<br /> No false positive results were reported for ham control samples. 13 and 17 laboratories had previous experience with Half-Fraser<br /> and Fraser broth, respectively. Previous experience with plating media<br /> 3.1.4.4. Wheat grain. Samples inoculated with a high level of L. used in the collaborative study was rather limited. Seven laboratories<br /> monocytogenes were detected as positive already after Half-Fraser had experience with ALOA media, four with LMBA, two with OCLA and<br /> enrichment on ALOA and LCA, and after Fraser enrichment on all the only one with LCA.<br /> media excluding OCLA for one sample (Table 8). Samples with a low 10 and 16 out of 18 laboratories used Half-Fraser and/or Fraser<br /> level of L. monocytogenes were detected as positive after Fraser enrichment, respectively, for routine analysis of L. monocytogenes.<br /> enrichment, and the same applies to those 1–3 samples reported as Nine laboratories used mostly LB I, and five used LB II. 16 laboratories<br /> false negatives for each medium after Half-Fraser enrichment. The used PALCAM, and 14 used Oxford media for detection of L.<br /> samples inoculated with both L. monocytogenes and L. innocua were monocytogenes in food and feed samples. Only four laboratories<br /> difficult to detect even after the whole enrichment procedure, and 1–3 used ALOA and LMBA routinely, and no one reported routine usage of<br /> positive results already obtained by Half-Fraser enrichment, were lost LCA or OCLA.<br /> for each medium after Fraser enrichment. LMBA medium contains sheep blood for haemolysis detection. The<br /> The sensitivities of the media in the detection of L. monocytogenes laboratories acquired the blood (with sodium citrate used as antic-<br /> from wheat grain samples were lower than for food samples. oagulant) used for the LMBA themselves. For the qualitative part of the<br /> <br /> <br /> <br /> Table 6<br /> Detection of L. monocytogenes in salmon samples with low or high levels of L. monocytogenes and low or high levels of both L. monocytogenes and L. innocua after enrichment in Half-<br /> Fraser followed by Fraser broth<br /> <br /> Inoculum level Half-Fraser enrichment positives Fraser enrichment more positives (+)/ Half-Fraser + Fraser enrichment positives<br /> lost positives (−)<br /> <br /> ALOA LCA OCLA LMBA ALOA LCA OCLA LMBA ALOA LCA OCLA LMBA<br /> LM H 18 18 18 16 0 0 0 0 18 18 18 16<br /> LM L 18 18 18 16 0 0 0 0 18 18 18 16<br /> LM H + LI H 17 18 18 16 −1 0 0 0 17 18 18 16<br /> LM L + LI L 16 16 17 16 +1/−1 0 +1 0 17 16 18 16<br /> Total 69 70 71 64 −1 0 +1 0 70 70 72 64<br /> Sensitivity (%) 95.8 97.2 98.6 100 94.4 97.2 100 100 97.2 97.2 100 100<br /> FP/NC (N) 0/36 0/36 1/36 0/32 0/36 0/36 0/36 0/32 0/36 0/36 1/36 0/32<br /> Specificity (%) 100 100 97.2 100 100 100 100 100 100 100 97.2 100<br /> <br /> The number of L. monocytogenes positive salmon samples at each inoculum level was 18 for ALOA, LCA and OCLA, and 16 for LMBA. The total number of L. monocytogenes positive<br /> samples was 216 and 192, respectively.<br /> LM: L. monocytogenes; LI: L. innocua; H: high; L: low; FP = false positive; NC = negative control.<br />  S. Loncarevic et al. / International Journal of Food Microbiology 124 (2008) 154–163 159<br /> <br /> <br /> Table 7<br /> Detection of L. monocytogenes in ham samples with low or high levels of L. monocytogenes and low or high levels of both L. monocytogenes and L. innocua after enrichment in Half-<br /> Fraser followed by Fraser broth<br /> <br /> Inoculum level Half-Fraser enrichment positives Fraser enrichment more positives (+)/lost Half-Fraser + Fraser enrichment positives<br /> positives (−)<br /> <br /> ALOA LCA OCLA LMBA ALOA LCA OCLA LMBA ALOA LCA OCLA LMBA<br /> LM H 18 18 18 16 0 0 0 0 18 18 18 16<br /> LM L 12 11 12 12 +5 +6 +5 +4 17 17 17 16<br /> LM H + LI H 18 18 18 16 −1 −2 −2 −1 18 18 18 16<br /> LM L + LI L 18 18 18 16 −1 −1 −3 −3 18 18 18 16<br /> Total 66 65 66 60 +3 +3 0 0 71 71 71 64<br /> Sensitivity (%) 91.7 90.3 91.7 93.8 95.8 94.4 91.7 93.8 98.6 98.6 98.6 100<br /> FP/NC (N) 0/36 0/36 0/36 0/32 0/36 0/36 0/36 0/32 0/36 0/36 0/36 0/32<br /> Specificity (%) 100 100 100 100 100 100 100 100 100 100 100 100<br /> <br /> The number of L. monocytogenes positive ham samples at each inoculum level was 18 for ALOA, LCA and OCLA, 16 for LMBA. The total number of L. monocytogenes positive samples<br /> was 216 and 192, respectively.<br /> LM: L. monocytogenes; LI: L. innocua; H: high; L: low; FP = false positive; NC = negative control.<br /> <br /> <br /> <br /> <br /> collaborative study, 10 and six laboratories (labs 3, 9, 12, 15, 17, and 19) enumerated from samples inoculated with both L. monocytogenes<br /> reported use of sheep and bovine blood, respectively For the and L. innocua, and with L. monocytogenes. Nor were there any<br /> quantitative part of the study, 10 laboratories (labs 3, 9, 12, 15, 17, significant statistical difference in the results obtained on the different<br /> and 19) used sheep blood and six bovine blood. One laboratory did not culture media.<br /> report the type of blood used (lab 13). The standard deviation of the repeatability (sr) in the enumeration of<br /> L. monocytogenes from the food matrices varied from 0.08–0.45 log cfu/g<br /> 3.2. The enumeration part of the study (Table 9). Correspondingly, the repeatability limit (r) ranged from 0.22–<br /> 1.26 log cfu/g. This means that the absolute difference between two<br /> 3.2.1. General results independent single test results obtained under repeatability conditions,<br /> Results were received from 17 of 19 participating laboratories, i.e. same test method, identical test material in the same laboratory by the<br /> including the organising laboratory. All 17 laboratories included the same operator using the same apparatus within a short time interval,<br /> LMBA medium in the collaborative study on a voluntary basis. 15 should not exceed 1.26 log cfu/g.<br /> laboratories received their samples within 24 h after shipment, and Best precision (lowest values of sr and r) was obtained by LCA for<br /> two within 48 h. Three laboratories (labs 11, 12 and 15) started the cheese samples inoculated with a high level of L. monocytogenes. The<br /> analyses the day after receipt, due to late arrival of the samples. The poorest precision (highest values of sr and r) was obtained by OCLA for<br /> other laboratories started the analyses immediately upon receipt of salmon samples with a low inoculum level of L. monocytogenes.<br /> the samples. One laboratory (lab 5) started the analysis five days after The standard deviation of the reproducibility (sR) in the enumer-<br /> receiving the samples. All except one laboratory reported sample ation of L. monocytogenes from the food samples, ranged from 0.15–<br /> temperatures of less than 10 °C (between 0 and 9 °C). One laboratory 0.44 log cfu/g. Correspondingly, the reproducibility limit (R) varied<br /> (lab 13) reported 15 °C, and was excluded from the statistical analysis from 0.41–1.23 log cfu/g. As for the repeatability values, the best<br /> because of temperature abuse during shipment. Laboratories 5, 13 and precision among laboratories was obtained for the analyses of cheese<br /> 14 did not use duplicate plates, and therefore their results were samples inoculated with a high level of L. monocytogenes on LCA, and<br /> excluded from further calculations. the poorest precision for the cheese samples with a low inoculum<br /> The precision data for cheese, salmon, ham and wheat grain are level of L. monocytogenes on LMBA (sR = 1.23 log cfu/g). As the limit of<br /> shown in Tables 9 and 10. The results from the dilutions 10− 1 and 10− 3 repeatability is above the limit of reproducibility, the limit of<br /> of low and high levels of inoculation, respectively, were used for the repeatability should also be used for the interpretation of the<br /> analysis. All the median values (log cfu/g) were lower than the reproducibility. That means that two results within reproducible<br /> theoretical concentrations calculated by the organising laboratory, conditions should not exceed 1.26 log cfu/g.<br /> and especially low for wheat grain samples. Statistically, there were no For wheat grain, the standard deviation of repeatability was<br /> significant differences in the results (cfu/g of L. monocytogenes) generally slightly higher than for food matrices, the range being<br /> <br /> <br /> <br /> Table 8<br /> Detection of L. monocytogenes in wheat grain samples with low or high levels of L. monocytogenes and low or high levels of both L. monocytogenes and L. innocua after enrichment in<br /> Half-Fraser followed by Fraser broth<br /> <br /> Inoculum level Half-Fraser enrichment positives Fraser enrichment more positives (+)/lost Half-Fraser + Fraser enrichment positives<br /> positives (−)<br /> <br /> ALOA LCA OCLA LMBA ALOA LCA OCLA LMBA ALOA LCA OCLA LMBA<br /> LM H 18 18 17 15 0 0 0 +1 18 18 17 16<br /> LM L 17 15 17 14 +1 +3 +1 +2 18 18 18 16<br /> LM H + LI H 16 16 16 15 +1/− 3 +1/−1 −1 0 17 17 16 16<br /> LM L + LI L 13 12 13 11 +1 −1 0 0 14 12 14 11<br /> Total 64 61 63 55 0 +2 0 +3 67 65 65 59<br /> Sensitivity (%) 88.9 84.7 87.5 85.9 88.9 87.5 87.5 90.6 93.0 90.2 90.2 92.2<br /> FP/NC (N) 0/36 0/36 0/36 0/32 0/36 0/36 0/36 0/32 0/36 0/36 0/36 0/32<br /> Specificity (%) 100 100 100 100 100 100 100 100 100 100 100 100<br /> <br /> The number of L. monocytogenes positive wheat grain samples at each inoculum level was 18 for ALOA, LCA and OCLA, 16 for LMBA. The total number of L. monocytogenes positive<br /> samples was 216 and 192, respectively.<br /> LM: L. monocytogenes; LI: L. innocua; H: high; L: low; FP = false positive; NC = negative control.<br /> 160 S. Loncarevic et al. / International Journal of Food Microbiology 124 (2008) 154–163<br /> <br /> <br /> 0.08–1.15 log cfu/g. The other precision values for wheat grain were not collaborative study according to EN/ISO 11290 Part 1:1997 (Anon-<br /> satisfactory. ymous, 1999 one laboratory used ALOA and one used LMBA for the<br /> The precision data derived from this collaborative study may not analyses on a voluntary basis. Both these media detected as positive 13<br /> be applicable to concentration ranges and matrices other than those out of 15 samples (cheese, meat and ham) containing 5 cfu/25 g of L.<br /> included in the study. Factors such as Listeria strains, the competitive monocytogenes in the presence of L. innocua. The high sensitivity of<br /> flora and the physiological status of target and competitors, have an LMBA in the detection of low levels of L. monocytogenes, was also<br /> influence on the data obtained. shown in the study by Johansson et al. (2000), where all the naturally<br /> contaminated Gorgonzola cheese samples (N = 8) containing 1–10 cfu<br /> 3.2.2. Confirmation L. monocytogenes in 25 g of cheese, were detected as positive.<br /> Morphological and traditional biochemical tests for the confirma- The advantages of L. monocytogenes specific plating media were<br /> tion of L. monocytogenes in the enumeration part of the study were shown by facilitated detection in samples containing both L.<br /> performed according to the draft method by 15 laboratories. 13 monocytogenes and L. innocua, especially after Fraser enrichment,<br /> laboratories used traditional identification tests. The laboratories that although the level of L. innocua was almost three times higher than<br /> used traditional confirmation, reported that they performed the that of L. monocytogenes. Some positives already obtained by Half-<br /> following tests: catalase test (12 labs), gram staining (optional) (7 Fraser enrichment could be lost because of overgrowth by L. innocua.<br /> labs), xylose and rhamnose test (13 labs). The use of commercial Plating out from Half-Fraser is important, because the Fraser broths<br /> biochemical identification tests (Accu-Probe) was reported by two support better the growth of L. innocua than L. monocytogenes (Petran<br /> labs. and Swanson, 1993). If plated out only after Fraser enrichment, L.<br /> innocua heavily overgrows L. monocytogenes. On the chromogenic<br /> 4. Discussion media compared, the Listeria colonies were very large, especially after<br /> the incubation of 48 h, and the identification of L. monocytogenes<br /> 4.1. Detection colonies with opaque halos among those of L. innocua may cause<br /> difficulties. On LMBA, the identification of L. monocytogenes in the<br /> According to the present study, there were only slight differences presence of L. innocua may often be easier compared to ALOA and<br /> (p N 0.05) between ALOA, LCA, OCLA and LMBA in the detection of L. OCLA, due to small colony size and β-haemolysis of L. monocytogenes.<br /> monocytogenes in cheese, salmon and ham samples, and thus these Sometimes L. monocytogenes strains with very narrow haemolysis<br /> media can be used as alternatives in the analyses of food. The zones may occur in foods. It is recommended to read LMBA against a<br /> sensitivities of the media were high, varying from 97.7% to 100% after light source. Combining a chromogenic medium and LMBA would<br /> two-step enrichment, thus showing excellent performances for the further improve the detection.<br /> matrices and the levels of L. monocytogenes evaluated. Most of the Feed-stuffs have not been included in the collaborative studies of<br /> positives (94.9%–96.4%) were detected already after Half-Fraser the methods for L. monocytogenes previously, and no comparisons of<br /> enrichment from all the matrices, which shortens the analysis time the media have been published, either. In this study, the sensitivities of<br /> for the majority of the samples. This result supports the results of the the media in the detection of L. monocytogenes were substantially<br /> previous studies (Vlaemynck et al., 2000; Johansson et al., 2000; lower for wheat grain than for the food matrices, but still moderately<br /> Normark, 2002). However, in case of low levels of L. monocytogenes, good — 90.3%–93.1% after the two-step enrichment. This inferiority is<br /> injured cells and/or heavy competitive flora, further enrichment in obviously caused by numerous competitors in wheat grain. However,<br /> Fraser broth could be needed (Johansson et al., 2000; Anonymous, the category of ready-to-eat fo