Comparative studies on the isolation of “sublethally injured” salmonellae in nine European laboratories*

Bull. Org. mond. Sante 1973, 48, 167-174Bull. Wld Hlth Org. Comparative studies on the isolation of "sublethally injured" salmonellae in nine European laboratories* W. EDEL & E. H. KAMPELMACHER On behalf of the members of a Working Group' Artificially or naturally contaminated minced meat samples were distributed to nine laboratories and the isolation rates obtained using standardized pre-enrichment and direct enrichment methods were compared. In all laboratories the pre-enrichment method gave higher isolation rates but the difference was less marked with the naturally contaminated samples. Extending the duration of incubation of the liquid enrichment medium to 48 h and making a second subculture improved the isolation rates. The results obtained with selective plates of the laboratories' own choice were unexpectedly poor in comparison with the standard method. In earlier studies attempts were made with the co- operation of different laboratories to develop a stan- dardized method for the isolation of Salmonella (Edel & Kampelmacher, 1968, 1969). The results were more uniform when liquid enrichment media were incubated at 43°C, when brilliant green phenol red agar (BGA) of the same composition and batch was used, and when large Petri dishes (14 cm) were used for plating. However, significant inter-laboratory dif- ferences remained in the results, especially when naturally contaminated samples, such as minced meat containing very small numbers of salmonellae, were examined. This variability may be partially accounted for by human factors, such as the degree ofexperience in reading plates and the effort to check a large num- ber of colonies, but other factors may also be in- volved-e.g., " injury " of cells caused by deep-freez- ing of the minced meat samples before shipment to the different laboratories. It is known that freezing and drying may injure salmonellae so that they are unable to multiply in selective media such as tetra- thionate bile brilliant green (TBB) incubated at 43'C * This investigation was initiated by the Common Market Scientific Veterinary Group on Salmonellosis and received financial support from the World Health Organization. Re- prints may be obtained from the Laboratory for Zoonoses and Food Microbiology, National Institute of Public Health, P.O. Box 1, Bilthoven, The Netherlands. 1 The names and addresses of the members of the Working Group are listed in Annex 1. (van Schothorst & Kampelmacher, 1968; Corry et al., 1969). Earlier experiments indicated that this type of injury factor might affect the isolation ofsalmonellae: the results from different laboratories varied when deep-frozen samples of minced meat were examined, whereas the results from samples of pig faeces were more uniform. Sublethal injury of salmonellae may occur in many food processes. It was therefore de- cided to continue the comparative studies on the methods of isolating salmonellae by investigating meat containing damaged cells of salmonellae. The standardized method (i.e., direct enrichment in TBB at 430C) described earlier was compared with a method using pre-enrichment at 37°C followed by enrichment in TBB at 43°C. MATERIALS AND METHODS As in the previous investigations, minced meat samples either artificially or naturally contaminated were used for the experiments. Preparation, packing, and dispatch were, in principal, the same as described by Edel & Kampelmacher (1968, 1969); differences are described below. The experiments included 8 weekly investigations of 15 samples, each in dupli- cate, of artificially contaminated minced meat, fol- lowed by 10 weekly investigations of another 15 sam- ples, in duplicate, of naturally contaminated minced meat. 2992 -167- 3 168 W. EDEL & E. H. KAMPELMACHER The participating laboratories were those listed in the previous publications and they are indicated by letters only. Laboratories A1, A2, and As were de- partments of the laboratory that organized the trial; the samples were prepared, packed, and dispatched by laboratory A3. Preparation of " sublethally injured " salmonellae The method used was similar to that described by van Schothorst & van Leusden (1972): 1 litre of a 24-hour broth culture containing approximately 1.3 x 109 Salmonella utrecht per ml was centrifuged at 10 000 rev/min for 15 min. The sediment was sus- pended in 1 litre of condensed milk (high percentage of fat and dry matter). The milk was centrifuged at 18 000 rev/min for 15 min. The floating fat and the sediment were placed in glass Petri dishes (diameter 14 cm) and dried for several days in an incubator at 43°C. The material was then scraped from the bottom of the dishes and incubated again; this was repeated several times. The dried substance was ground to powder in a mortar. The salmonellae were assumed to be injured when they could not be isolated by direct enrichment in tetrathionate bile brilliant green broth (TBB) at 43°C, but only after pre-enrichment followed by tetrathionate. The number ofsalmonellae was determined by the " most probable number " (MPN) method using buffered peptone water as the pre-enrichment medium and TBB (at 43°C) as the enrichment medium. In order to obtain different con- tamination levels the milk powder was diluted by the addition of normal spray-dried milk powder. Artificially contaminated minced meat samples The minced meat was purchased from a small butchery (Edel & Kampelmacher, 1969). Each 10-g sample was packed in a small plastic bag and, in order to increase the amount of competitive flora, approximately 109 cells of a mixture of the genera Enterobacter, Citrobacter, Proteus, Providencia, and Pseudonmnas were added as in the earlier experi- ments. Amounts of exactly 0.1 g of the milk powder containing " sublethally injured " salmonellae, were placed in separate small plastic bags. Each sample, therefore, included a plastic bag containing 10 g of minced meat inoculated with the competitive flora and a small plastic bag containing the milk powder with injured salmonellae. All the samples were pre- pared 3 days before dispatch. The bags containing the minced meat were stored at -20°C in cardboard boxes. The bags containing the milk powder were stored in a refrigerator (+40C), since storage at -20°C resulted in a decreased number of positive isolations. In each experiment 15 samples were examined by the pre-enrichment method and 15 samples by direct enrichment. The number of samples artificially con- taminated was different in each experiment and was known only to laboratory As, where the inoculation took place. Altogether, 83 samples were artificially contaminated for each method. The number ofadded cells of S. utrecht was different in each experiment and was controlled by the MPN method. The num- ber of salmonellae per sample of 0.1 g of milk powder varied from about 4 to 240. Naturally contaminated minced meat samples The minced meat was bought weekly in portions of 250 g from different butcher's shops in one town. Each portion was divided into 10-g samples; two samples were sent to each laboratory, one for exami- nation by the direct enrichment method and the other for examination by the pre-enrichment method. All samples were stored at -20°C until dispatch. Handling of samples after arrival in the laboratory The samples were always sent to the different labo- ratories on a Monday and were investigated on the following Monday, being stored at -20°C until the day of investigation. The bags containing the milk powder with salmonellae were stored in a refrigera- tor (+4°C) until the day of investigation. Methods of bacteriological examination The methods used by all the laboratories partici- pating in this investigation were the same, except that additional selective plates could be used as de- sired by the individual laboratories. Direct enrichment method. This technique was the standardized method described by Edel & Kampel- macher (1969) except that the BGA medium, which was previously provided in different parts, was sup- plied as a single pack. Pre-enrichment method. An enrichment procedure was introduced for the reactivation of salmonellae. Samples were placed in a nonselective medium (buf- fered peptone water) 1 for 18-20 h at 37°C. 10 ml of the culture were then transferred into 100 ml of the tetrathionate liquid enrichment medium. The stan- dardized method described earlier was then applied. 1 Draft ISO recommendation on the isolation of salmo- nellae from meat and meat products. ISOLATION OF SUBLETHALLY INJURED SALMONELLAE :S. I . I . . B C D E F G 1+ 1 L a b o ra tory *w0ro Fig. 1. Comparison of pre-enrichment and enrichment methods for the isolation of salmonellae from artificially contaminated minced meat. Shaded columns, pre-enrichment method; stippled columns, direct enrichment method. The formulae of the media and the methods used are given in Annex 2. This description is the same as that sent to the laboratories conducting the experiment. Selective plates. All laboratories made subcultures on BGA medium (Oxoid CM 329) prepared from the same batch (designated " a " in Tables 2 and 3) and each laboratory used at least one extra plate, with a medium of its own choice (designated "b "). The media chosen by the laboratories were as fol- lows: A1, A2, and A3: BGA (home-made); B: deoxy- cholate citrate agar (home-made), and BGA (Oxoid CM 263); C: xylose lysine decarboxylate agar and deoxycholate citrate agar; D: pril-mannitol agar and BGA (home-made); E: deoxycholate citrate agar; F: BGA (Difco); G: BGA (Difco); H: two sugars novobiocin agar; 1: BGA (home-made). RESULTS Salmonella isolation Experiments with artificially contaminated minced meat. By means of the pre-enrichment method, 95.0% (86.7-100 %) ofthe contaminated samples were found positive, while with the direct enrichment me- thod only 51.7% (30.1-73.5%) of the contaminated samples were detected. The results of these experi- ments are summarized in Fig. 1. Experiments with naturally contaminated minced meat. Of the 150 duplicate samples used in the 10 ex- periments, 36 (24.0%) were found to be positive. A sample was called positive if at least one laboratory isolated salmonellae by either the pre-enrichment or the direct enrichment method. When only the pre- enrichment method was used 33 samples (22.0%) were found positive (91.7% of the total number of positive samples). The different laboratories found from 1.3% to 10.0% of the samples positive. When only the direct enrichment method was used 25 sam- ples (16.7 %) were found positive (69.4% of the total). The different laboratories found from 2.7% to 8% of the samples positive. The results are summarized in Fig. 2. The data concerning the 17 different sero- types are given in Table 1. Influence of the enrichment time and the selective plates used, on the isolation of Salmonella Incubation time. Of the total number of artificially contaminated samples found positive (plates " a " and/or " b ") after incubation for 48 h, 94.7% (84.7-100%) were positive with the pre-enrichment method after 24 h, the remaining 5.3% (0.0-15.3 %) 80 [ 70 _ 60 F 50k_ 401- 30 _ 201- nI Number of samples known to be contaminated ? T TW______ _-l_____g_ 10 _ uS.- Al A2 A3 169 Oh^ W. EDEL & E. H. KAMPELMACHER Fig. 2. Comparison of pre-enrichment and enrichment methods for the isolation of salmonellae from naturally contaminated minced meat. Shaded columns, pre-enrichment method; stippled columns, direct enrichment method. Table 1. Salmonella serotypes isolated from naturally contaminated minced meat Serotype No. of isolations S. typhimurium 21 2 25 (40.3%) S. typhimurium var. copenhagen 4J S. livingstone 6 (10.0 %) S. brandenburg 4 (6.5%) S. infantis 4 S. panama 4 S. ohio 3 S. derby 2 S. dublin 2 S. eimsbuettel 2 S. give 2 S. senftenberg 2 S. anatum 1 S. mendoza 1 S. montevideo 1 S. thompson 1 S. victoria 1 S. westhampton 1 total 62 a a Of 36 positive samples, 16 contained I serotype, 16 contained 2 serotypes, 3 contained 3 serotypes, and 1 contained 5 serotypes. being detected at the second subculture. With the direct enrichment method, only 52.5% (22.2-76.7 %) of the positive samples were detected at 24 h, the remaining 47.5% (23.3-77.8%.) being found positive at 48 h. Of the total number of naturally contaminated samples found positive, 88.7% (70.0-100%) were found positive with the pre-enrichment method after 24 h, the remaining 11.3% (0-30.0%Y.) being detected at 48 h. With the direct enrichment method 70.8% (50.0-83.3 %) were detected at 24 h and the remaining 29.2% (16.7-50.0%) at 48 h. These results are summarized in Table 2 for the artificially contaminated minced meat and in Table 3 for the naturally contaminated meat. Selective plates. A comparison of the results be- tween the selective plates " a " (standard) and " b " (medium of the laboratory's choice) showed that with the pre-enrichment method the isolation rates were 99.0% (89.9-100 %) with " a " and 83.3 % (5.4-100%) with " b ". With the direct enrichment method the figures were 97.2% (86.7-100%) and 80.1% (1.9- 100%) for " a " and " b ", respectively. For naturally contaminated minced meat samples the isolation rates 170 ISOLATION OF SUBLETHALLY INJURED SALMONELLAE Table 2. Number of positive isolations from artificially contaminated meat samples by the pre-enrichment and direct enrichment methods Pre-enrichment Direct enrichment No. of to a b Total " a Totalartificially a *abb a- - Totalr 'b a b" TndotalbLaboratory contami- nated Total Total Total Total Total Total samples 24 h after 24 h after 24 h after 24 h after 24 h after 24 h after 48 h 48 h 48 h 48 h 48 h 48th A 1 83 69 79 71 79 72 79 14 48 17 47 19 49 A2 83 75 80 76 80 77 80 11 41 12 43 12 45 A3 83 73 81 77 81 77 81 32 57 32 57 32 57 B c 83 74 77 71 74 74 77 36 61 31 52 36 61 Cd 83 79 83 10 11 79 83 8 36 1 9 8 36 D 83 77 80 76 80 80 80 17 25 13 24 17 25 E 83 69 74 3 4 69 74 35 52 1 1 35 52 F 83 72 81 73 81 74 81 10 26 21 28 23 30 G 83 67 71 79 79 79 79 27 53 40 54 40 54 H 83 73 80 79 81 79 81 12 29 12 32 14 32 1 83 39 72 61 72 61 72 12 31 12 31 12 31 total 913 767 858 676 722 821 867 214 459 192 378 248 472 percentage 100 84.0 94.0 74.0 79.1 89.9 95.0 23.4 50.3 21.0 41.4 27.2 51.7 range of percentage in (47.0- (85.5- (3.6- (4.8- (73.5- (86.7- (9.6- (30.1- (1.2- (1.2- (9.6- (30.1-different laboratories 95.2) 100) 95.2) 97.6) 96.4) 100) 43.4) 73.5) 48.2) 68.7) 48.2) 73.5) a BGA plate-Oxoid CM 329, the same batch number for all laboratories. b Plate of the laboratory's own choice. c Results on deoxycholate citrate agar are shown in column' b": results on BGA (Oxoid 263) were lower than those in column " a'. d The results in the table were obtained on xylose decarboxylase agar. Results on deoxycholate agar: 24 h, 12; 48 h, 2-a total of 14. with the pre-enrichment method were 88.7% (50.0- The results from the naturally contaminated 100%) and 97.4% (0.0-100%) for " a " and " b " and minced meat samples were also better with the pre- for direct enrichment 93.3% (50.0-100%) and 94.4% enrichment method than with the direct enrichment (0.0-100%) for " a " and " b ". These results are method. However, the differences were less marked summarized in Tables 2 and 3. in some of the laboratories. As was also observed by Edel & Kampelmacher (1969) better results were DISCUSSION obtained when the duration of incubation of theliquid enrichment medium was extended to 48 h and The results from the experiments with artificially a second subculture was made. However, with pre- contaminated minced meat show that in all labora- enrichment the percentage of positive isolations was tories the pre-enrichment method gave higher iso- already high after the first subculture (24 h). It may lation rates than the direct enrichment method. There be concluded that pre-enrichment favours the repro- was almost 100% recovery when all laboratories used ducibility of results for the isolation of salmonellae the pre-enrichment method; with the direct enrich- and thus this method is to be recommended. ment method there were marked differences between With the selective agar media the somewhat poor the results from different laboratories, although it is results with the media of the laboratories' own choice assumed that the detailed instructions (see Annex 2) were unexpected. The inter-laboratory differences, the were strictly followed in all laboratories. reasons for which are not understood, appear to be 171 W. EDEL & E. H. KAMPELMACHER Table 3. Number of positive isolations from naturally contaminated minced meat by the pre-enrichment and enrichment methods Pre-enrichment Direct enrichment a a "b"b Toala# Total .Total no. Laboratory a' and/or" b "a " b'"a and/or' b of positive Total Total Total Total Total Total samples 24 h after 24 h after 24 h after 24 h after 24 h after 24 h after 48 h 48 h 48 h 48 h 48 h 48 h Ai 7 10 7 10 7 10 6 10 4 10 6 10 14 A2 5 6 7 8 7 8 2 5 3 6 3 6 10 A3 11 11 12 12 12 12 10 12 10 12 10 12 17 B c 12 15 13 15 13 15 7 9 7 9 7 9 19 C d 11 12 10 11 11 12 5 8 5 8 5 8 16 D 9 13 11 15 12 15 5 7 4 7 5 7 16 E 2 2 0 0 2 2 4 5 0 0 4 5 6 F 11 13 12 14 12 14 4 6 5 7 5 7 16 G 4 5 10 10 10 10 5 9 7 11 7 11 15 H 10 11 10 11 11 11 8 10 7 10 8 10 15 1 4 4 5 6 5 6 2 2 3 4 3 4 8 total no. of positive subsamples 86 102 97 112 102 115 58 83 55 84 63 89 - no. of positive samples 28 31 30 33 30 33 20 24 19 25 20 25 36 percentage 77.8 86.1 83.3 91.7 83.3 91.7 55.6 66.7 52.8 69.4 55.6 69.4 100 range of percentage in (5.6 (5.5 (0.0 (0.0 (5.5 (55. (5.5 (5.5 (0.0 (0.0 (8.3 (11.1 - different laboratories -33.3) -41.7) -36.1) -41.7) -36.1) -41.7) -27.8) -33.3) -27.8) -33.3) -27.8) -33.3) a BGA plate-Oxoid CM 329, the same batch number for all laboratories. b Plate of laboratory's own choice. c Results on deoxycholate citrate agar are shown in column" b": results on BGA (Oxoid 263) were the same as those in column " a". d Results on xylose lysine decarboxylase agar are shown in column" b". Results on deoxycholate citrate agar: 24 h, 9; 48 h, 1-a total of 10. more marked than those arising from the use of different methods within a given laboratory. As sug- gested by Kampelmacher (1967) the human factor is significant even when a standardized method is used. Better training and education and exchange of labo- ratory workers might improve efficiency. R1 SUM f tTUDES COMPARATIVES SUR L'ISOLEMENT DE SALMONELLA AYANT SUBI DES DOMMAGES SUBL1TAUX, DANS NEUF LABORATOIRES D'EUROPE Les auteurs rapportent les r6sultats d'etudes compara- tives menees dans neuf laboratoires d'Europe sur les m6thodes d'isolement - pr6enrichissement ou enrichis- sement direct - de Salmonella, pr6alablement lesees par une serie de manipulations, a partir de viande hachee. Les divers laboratoires ont utilis6 les memes mnthodes et leurs recherches ont port6 sur un total de 120 &han- tillons de 10 g de viande hachee artificiellement conta- min6e et de 150 6chantillons de 10 g de viande naturelle- ment contaminee fournis par le laboratoire organisateur. La contamination artificielle a ete realis&e en ajoutant 'a la viande un nombre variable (4 A 240 par 0,1 g d'ino- culat) de cellules de Salmonella utrecht et environ 109 cel- lules d'un m6lange des genres Citrobacter, Enterobacter, Proteus, Providencia et Pseudomonas. Les echantillons, en double exemplaire, ont ete examin6s A raison de 15 par semaine. La m6thode de preenrichissement a donn6 des resultats 172 ISOLATION OF SUBLETHALLY INJURED SALMONELLAE 173 positifs avec 86,7 a 100% (selon les laboratoires) des echantillons artificiellement contamines, alors que par la methode de l'enrichissement direct la proportion des 6chantillons positifs n'a ete que de 30,1 a 73,5 %. Au total, 36 (24,0%Y.) des 150 6chantillons naturellement conta- mines ont e trouv6s positifs (un specimen letant consi- der6 comme positif lorsqu'un laboratoire au moins avait isole des Salmonella apres preenrichissement ou enrichis- sement direct). La methode du pr6enrichissement, a elle seule, a decel 91,7% des 6chantillons positifs, et la methode de 1'enrichissement direct 69,4%. On a obtenu de meilleurs r6sultats en portant la dur6e d'incubation en milieu liquide d'enrichissement a48 heures et en proc6dant i une deuxieme sous-culture. L'emploi par les laboratoires de milieux s6lectifs de leur choix a donn6 des r6sultats relativement moins satisfaisants. REFERENCES Corry, J. E. L. et al. (1969) J. appl. Bact., 32, 415 Edel, W. & Kampelmacher, E. H. (1968) Bull. Wld Hlth Org., 39, 487 Edel, W. & Kampelmacher, E. H. (1969) Bull. Wld Hith Org., 41, 297 Kampelmacher, E. H. (1967) Zbl. Bakt., L Abt. Orig., 204, 100 van Schothorst, M. & Kampelmacher, E. (1968) In: Proceedings of the Sixth International Symposium on Food Microbiology, Bilthoven, Netherlands, June 1968, pp. 193-205 van Schothorst, M. & van Leusden, F. M. (1972) Zbl. Bakt., I. Abt. Orig., A221, 19-29 Annex I MEMBERS OF THE WORKING GROUP E. Bulling, Director and Professor, Federal Institute of Public Health, Robert von Ostertag Institute for Veterinary Medicine, Berlin W. Edel, Bacteriologist, Laboratory for Zoonoses and Food Microbiology, National Institute ofPublic Health, Bilthoven, The Netherlands A. C. Ghosh, Food Hygiene Laboratory, Central Public Health Laboratory, London, England G. Gottal, Director, National Veterinary Medical Labora- tory, Luxembourg B. C. Hobbs, Head, Food Hygiene Laboratory, Central Public Health Laboratory, London, England E. H. Kampelmacher, Head, Laboratory for Zoonoses and Food Microbiology, National Institute of Public Health, Bilthoven, The Netherlands; Extraordinary Professor, Agricultural University, Wageningen, The Netherlands G. Lott, Head, Meat Hygiene Laboratory, Veterinary Bacteriological Institute, University of Zurich, Switzer- land A. Linton, Senior Lecturer in Veterinary Bacteriology, Medical School, University of Bristol, England J. Pantaleon, Scientific Director, Central Laboratory of Veterinary Research, Alfort, France M. van Schothorst, Bacteriologist, Laboratory for Zoono- ses and Food Microbiology, National Institute of Public Health, Bilthoven, The Netherlands G. Tiecco, Institute of Public Health, Rome, Italy J. Thomas, Head, Bacteriological Department, National Institute of Veterinary Research, Brussels, Belgium Annex 2 DIRECTIONS FOR THE SALMONELLA ISOLATION METHODS Thawing ofsamples Remove cardboard box from the deep-freezer on Mon- day morning (day of investigation) at 10.30 h. Open box and spread samples in a refrigerator at +40C for 4 h to thaw. Inoculation ofsamples Start at 14.30 h after thawing pre-enrichment: sample of 10 g into 100 ml of buffered peptone water enrichment: sample of 10 g into 100 ml of tetra- thionate Put the whole sample (minced meat and milk powder 1 bearing the same sample number), including plastic bags, into the pre-enrichment or enrichment medium, respec- tively, by cutting them in small pieces. For each sample sterile instruments should be used. Preparation of the pre-enrichment medium-buffered pep- tone water peptone (Difco B118) 10 g sodium chloride 5 g disodium hydrogen orthophosphate (12 H2O) 9 g potassium dihydrogen orthophosphate 1.5 g distilled water 1 000 ml 1 Only for samples of artificially contaminated meat. W. EDEL & E. H. KAMPELMACHER Dissolve the ingredients in the water by heating. Adjust the pH to 7.2±0.1. Dispense into jars (ready for use) in 100-ml quantities. Sterilize for 15 min at 121°C. Prepare the medium the week before it is to be used. (All the ingredients were sent by laboratory A to the participating laboratories.) Preparation of the enrichment medium-tetrathionate bile brilliant green broth As described by Edel & Kampelmacher (1969) Duration of incubation of the liquid media pre-enrichment medium: 18 h enrichment medium: 48 h Incubation temperature of the liquid media pre-enrichment medium: 37C enrichment medium: 43°C (see Edel & Kampel- macher, 1969) Enrichment after pre-enrichment After incubation of the pre-enrichment medium at 37°C for 18 h, with a pipette transfer 10 ml into 100 ml of the tetrathionate medium and then follow the same procedure as regards incubation time and temperature as mentioned above. Subcultures from the enrichment medium only First subcultures should be made after 18-24-h; second subcultures after 48 h. Method ofstreaking As described by Edel & Kampelmacher (1969) Selective plate-preparation of brilliant green phenol red agar (BGA) The BGA medium used was similar to the one described by Edel & Kampelmacher (1969) but it is now available as a single pack (Oxoid CM 329). Hence: BGA (Oxoid CM 329) 52 g Distilled water 1 000 ml To prepare the medium use a pot, without a lid, that can be placed on a gas flame (e.g., an enamelled saucepan). Put the water in the pot and place the ingredients on top of the water and let them soak for 15 min. Stir with a spoon and bring just to the boil (until one or two bubbles appear) with occasional agitation. Do not auto- clave! Mix well, cool and pour the plates. Store the plates in the dark at room temperature. When the plates are used, they must be sufficiently dry, otherwise Proteus sp. may spread. The preparation of the plates should be performed during the week before they are to be used. If necessary the plates may be dried in an incubator at <500C. Incubation of the selective plates 18-24 h at 370C Identification of suspicious colonies As described by Edel & Kampelmacher (1969). 174