buernational Journal of Food Microbiology, (1991)301-308 301 © 1991 Elsevier Science Publishers B.V. 0168-1605/91/$03.50

F O O D 00424

Collaborative study of the International Office of Cocoa, Chocolate and Sugar Confectionery on Salmonella detection from cocoa and chocolate

processing environmental samples

J.M. De Smedt 1, S. Chartron 2, J.L. Cordier 3, E. Graft 4, H. Hoekstra 5, j .p. Lecoupeau 6, M. Lindblom 7, j. Milas 8, R.M.

Morgan 9, R. Nowacki 10, D. O'Donoghue 11, G. Van Gestel 12 and M. V a r m e d a l 13

l Jacobs Suchard, Herentals, Belgium; 2 Mars Alimentaire, Hagenau, France; 3 Nestec, Vetey, Switzerland," 4 Freia, Oslo, Norway," s Lindt & Sprungli, Aachen, F.R.G.; 6 Cacao Barry, Lout~,iers, France; 7 Marabou,

Sundbyberg, Sweden; 8 E.J. Brach, Chicago, Illinois, U.S.A.," 9 United Biscuits, High Wycombe, U.K.; lO Grayson & Associates, Auckland, New Zealand; it Rowntree Mackintosh, York, U.K.; 12 Mars, Veghel,

The Netherlands; and 1~ Nidar-Bergene, Trondheim, Norway

(Received 21 Januari 1991; accepted 22 May 1991)

A comparative collaborative study was performed in 13 laboratories to evaluate the use of motility enr ichment on Modified Semisolid Rappaport-Vassil iadis medium for rapid Salmonella detection from food-processing environmental samples. Artificially contaminated chocolate scrapings and naturally contaminated cocoa bean dust samples were used in the study. Pre-enrichment was performed in buffered peptone water with added casein and malachite green oxalate. Motility enr ichment was compared with a conventional cultural procedure using Rappaport-Vassil iadis broth and selenite cystine broth as selective enrichment. The productivity of motility enr ichment was 93.5% compared to a productivity of the cultural procedure of 92%. Statistical analysis showed that there was no significant difference between the two procedures. Modified Semisolid Rappaport-Vassil iadis medium is a sensitive and simple diagnostic tool for the microbiological safety evaluation of food-processing environments.

Key words: Cocoa; Chocolate; Environment; Salmonella; Motility enr ichment

Introduction

The application of the Hazard Analysis Critical Control Point concept is an important tool for the prevention and control of food-borne salmonellosis (Simon-

Correspondence address." J. De Smedt, Jacobs Suchard, Montezumalaan 1, 2200 Herentals, Belgium.

302

sen et al., 1987). In this system, prevention of recontamination of processed products from the environment is an essential step. Besides visual monitoring to assess the cleanliness of equipment, supplemental microbiological tests are per- formed to trace sources of possible contamination. If such tests show the presence of Salmonella, the source of contamination must be sought immediately and removed (Simonsen et al., 1987).

Environmental samples can contain high levels of bacteria. Competitive bacteria can seriously retard the growth of Salmonella cells during pre-enrichment (Jame- son, 1962). By studying the dynamics of Salmonella growth in the pre-enrichment of environmental samples, Van Schothorst and Renaud (1983) found that Gram- positive bacteria can cause a drastic lowering of the pH of the medium. They concluded that maintaining the pH above 5 and limiting the growth of competitive bacteria are two important factors in the Salmonella isolation procedure. To limit the growth of Gram-posit ive bacteria they introduced the addition of malachite green oxalate to the pre-enrichment medium of environmental samples.

For the microbiological safety assurance of food processing, rapid detection methods are necessary for the timely identification of contamination sources. Motility enrichment on Modified Semisolid Rappaport-Vassil iadis (MSRV) medium was introduced (De Smedt et al., 1986) as a rapid, sensitive and cost-effi- cient means for isolation of salmonellae. With this procedure a serologically confirmed Salmonella detection is obtained within 48 h of pre-enrichment. Collab- orative studies (De Smedt and Bolderdijk, 1990; De Zut ter et al., 1991) have shown that this procedure is highly reliable for food products. In the present paper the results of a collaborative study on Salmonella isolation from environmental samples by pre-enrichment in buffered peptone water with added malachite green and subsequent motility enrichment on MSRV are presented.

Materials and Methods

Samples Cocoa bean dust was used as a naturally contaminated environmental sample.

The Most Probable Number (MPN) of Salmonella cells in this dust was deter- mined by analysing three each of aliquots of the test material, 25 g, 2.5 g and 0.25g. Artificially contaminated chocolate scrapings were also used as test samples. Artificial contamination of chocolate was achieved by the method described by Rappold et al. (1984). After contamination the MPN of salmonellae was also determined in these samples. Portions of the artificially contaminated chocolate were then mixed with chocolate scrapings of processing areas, so that the samples contained low levels of salmonellae. Before shipment to the collaborators, the MPN determination was repeated (see Table I). Chocolate scrapings not contami- nated with Salmonella were also examined. Thirteen laboratories collaborated in this study. They each received 25 samples of cocoa bean dust and 20 samples of chocolate scrapings. The Salmonella bacteria included are indicated in Table I.

303

Pre-enrichment Pre-enrichment was performed in buffered peptone water with casein and

malachite green oxalate. Buffered peptone water (Oxoid CM 509) was prepared by the participating laboratories according the manufacturer ' s instructions. Before sterilization 1% (w/v) of casein (Sigma C8654) was added. After sterilization 1 ml of malachite green solution was added to 225 ml of buffered peptone water. The malachite green solution was prepared by dissolving 2.5 g of malachite green oxalate (Merck 1398) to 100 ml of demineralised water and placed in a water bath at 37 ° C for 3 to 4 h, after which it was kept in a dark bottle at room temperature.

For the cocoa bean dust samples, 12.5 g of each sample was added to 225 ml of the pre-enrichment medium. For the chocolate scrapings 25 g of product was added to 225 ml of the pre-enrichment medium. The pre-enrichment cultures were incubated at 37 °C for 20 h.

Selective enrichment, plating and confirmation Rappaport-Vassil iadis (RV) medium was prepared as recommended by Van

Schothorst and Renaud (1983). For selenite cystine (SC), modified brilliant green agar (MBG), mannitol lysine crystal violet brilliant green agar (MLCB), triple sugar iron agar (TSI), and lysine iron agar (LIA) commercial available preparat ions of various brands were used. For the preparat ion of these media the manufacturer ' s instruction were carefully followed by the participants. From the incubated pre-en- r ichment cultures, 0.1 ml volumes were inoculated in 10 ml of RV broth and 1 ml in 10 ml of SC broth. The RV broths were incubated at 42 ° C and the SC broths at 37 ° C. After 24 h these selective enrichment cultures were plated onto MBG and MLCB. These selective agars were incubated at 37 °C for 24 h. Suspect colonies were inoculated into TSI and LIA and incubated at 37 °C for 24 h. Presumptive Salmonella cultures were confirmed serologically with commercially available poly- valent and monovalent antisera.

Motility enrichment and confirmation MSRV was prepared by the participating laboratories as described previously by

De Smedt and Bolderdijk (1987). Three drops (approx. 0.1 ml) of the incubated pre-enrichment culture were inoculated on a MSRV plate which was then incu- bated at 42 + 0.5 °C (direct motility enrichment). Three drops of the selective broths were, after 8 h of incubation of these broths, also inoculated on MSRV and incubated at 42 + 0.5 °C (indirect motility enrichment). After 24 h of incubation for the direct motility enrichment and 16 h for indirect motility enrichment, the MSRV plates were checked for migration. From the migrated cultures, bacterial suspension from the edge of the migration zone was inoculated on TSI and LIA tubes. A loopful was also inoculated in BHI broth, while another loopful was streaked on MLCB to check for purity. The MLCB, TSI and LIA were incubated at 37 ° C for 24 h. The BHI broth was incubated at 37 ° C for 4 to 6 h, after which serological slide agglutination was performed on this culture with Salmonella 0 and H antisera (various commercial brands as preferred by the participants).

304

Statistical analysis of data The results were statistically analysed according to the McNemar test (Siegel,

1956). A pair-wise comparison of the rate of false-negative reactions by each method was made using the relationship:

X 2= ( l a - b l - 1)2/(a +b)

where a is the number of samples positive by motility enrichment and negative by the cultural procedure, and b is the number of samples negative by motility enrichment and positive by the cultural procedure. A X 2 value > 3.84 indicates significance at the 0.05 level.

Results

Raw cocoa beans are frequently contaminated with Salmonella. The first process that precedes the manufacture of cocoa or chocolate is the cleaning of the beans. The dust from the cleaning operation has a high microbial load (more than 10 7 microorganisms per gram) and frequently also contains low levels of salmonel- lae, i.e. between 1 and 10 per 25 g (unpublished data). In the present study cocoa

TABLE I

Most Probable Number (MPN) of Salmonella cells per 25 g in chocolate scrapings

Sample number Serotype ~ MPN

1 S. eastbourne 0.9 2 S. tennessee 20

3 S. branderup 4 4 Not contaminated 5 S. colorado 4 6 S. circhow 0.9 7 S. anatum 2 8 Not contaminated 9 S. anatum 4

10 Not contaminated 11 S. colorado 9 12 S. virchow 0.9 13 S. eastbourne 0.4 14 Not contaminated 15 S. branderup 20 16 S. tennessee 0.4

17 S. colorado 4 18 S. anatum 0.4

19 S. eastbourne 0.4 20 Not contaminated

The serotypes included were isolated from cocoa dust except for S, eastbourne and S, tennessee which were isolated from milk chocolate.

305

TABLE I1

Tests with chocolate scrapings: number of Salmonella isolations with motility enrichment versus cultural procedure

Labor- Motility enrichment Cultural procedure Total a

atory Direct Indirect Total RV SC Total

A. 12 12 12 12 12 12 12 B. 13 13 13 13 13 13 13 C. 7 7 7 7 7 7 7 D. 11 11 12 11 11 11 12 E. 9 9 9 9 9 9 9 F. 6 6 6 6 6 6 6 G. 13 10 13 12 12 12 13 H. 6 6 6 6 6 6 6 I. 2 2 2 0 2 2 2 J. 8 8 8 9 9 9 9 K. 2 2 2 2 2 2 2 L. NT b NT NT NT NT NT NT M. 4 4 4 4 4 4 4

Total 93 90 94 91 93 93 95

a Motility enrichment and cultural procedure combined. b not tested.

bean dust was used as a naturally contaminated environmental sample. Addition of cocoa dust to liquid gave a rather viscous solution, so that a 5% (w/v) solution was used instead of the usual 10% solution used for food samples. In this way 12.5 g of dust was added to 225 ml of pre-enrichment medium. The MPN determination showed that there were two salmonellae per 12.5 g. Besides the naturally contami- nated samples, artificially contaminated chocolate scrapings were also used as samples. For these samples the usual amount of 25 g was added to 225 ml of pre-enrichment medium. The level of contamination of these artificially contami- nated samples is given in Table I.

The results of the comparative tests with chocolate scrapings are presented in Table II. Besides the five uncontaminated samples, there were also four samples for which only one out of three 25-g units was positive (MPN 0.4 cells per 25 g). In four other samples the level was also very low (MPN 0.9-2 cells per 25 g). There was also a decrease in the number of viable cells as a function of time. 4 months after the shipment and analysis of the samples, the tests were repeated in the organizing laboratory: the number of positive results dropped from 13 in the first series of tests to only three after 4 months. In some of the collaborating laborato- ries the tests were per formed 2-3 months after the receipt of the samples. The combination of a low level of contamination and a decrease in numbers with time explains the low number of positive results in these laboratories. The results of the comparative tests with cocoa bean dust are presented in Table III . There was also a decrease of positive results in these samples with time: the number of positives in

306

TABLE III

Tests with cocoa bean dust: number of Salmonella isolations with motility enrichment versus cultural procedure

Labo- Motility enrichment Cultural procedure Total a

ratory Direct Indirect Total RV SC Total

A. 19 18 20 18 2 18 20 B. 22 20 23 23 10 23 23 C. 16 11 16 15 13 20 24 D. 19 20 20 22 22 22 22 E. 18 14 18 14 7 15 15 F. 16 14 16 14 14 15 16 G. 25 25 25 25 25 25 25 H. 14 15 15 15 15 15 15 I. 10 2 10 2 4 5 11 J. 21 20 21 20 9 20 21 K. 16 10 17 16 11 18 21 L. 18 10 19 21 5 21 21 M. 15 15 16 15 14 16 24

Total 229 196 236 221 151 233 258

Motility enrichment and cultural procedure combined.

the organizing laboratory dropped from 23 in the first series of tests to 12 after 4 months.

Discussion

For cocoa products the presence of casein in the pre-enrichment medium is necessary to inhibit the bactericidal substances present in these products (Zapatka et al., 1977). Non-fat milk with brilliant green (AOAC, 1984) is widely used as pre-enrichment medium for cocoa and chocolate. For environmental samples non-fat milk is less suitable because high levels of competitive bacteria can cause a rapid decrease of the pH during incubation (Van Schothorst and Renaud, 1985; De Zutter et al., 1991). This drop in pH adversely affects the success of subsequent detection of Salmonella. To reduce the growth of competitive bacteria Van Schothorst and Renaud (1985) introduced the addition of malachite green oxalate to buffered peptone water for Salmonella detection from environmental samples. In the present study a combination of buffered peptone water with casein and malachite green was used as pre-enrichment medium for environmental samples from cocoa and chocolate factories.

Motility enrichment on MSRV in petri dishes is a simple and rapid tool for Salmonella detection (De Smedt et al., 1986). For food samples the method is at least as sensitive as the standard cultural method (De Smedt and Bolderdijk, 1990; De Zutter et al., 1991). The medium can be used after pre-enrichment (direct

307

motility enrichment) or after short enrichment in selective broth (indirect motility enrichment). The present study shows that motility enrichment on MSRV is also applicable to environmental samples.

With chocolate scrapings a total of 240 samples were examined. Of these, 60 samples were not contaminated, while 96 samples were contaminated with only 0.4-2 ceils per 25 g. These low levels were chosen because it is necessary in the microbiological safety assurance of food processing that such low levels are detected. The rest of the samples (84) showed a contamination level of four ceils or more per 25 g. Of all these samples 95 were found positive of which 99% were detected with MSRV and 98% with the cultural procedures.

Two-hundred-and-fifty-eight samples of naturally contaminated cocoa bean dust were found positive on a total of 325 samples, examined by the participating laboratories. MSRV detected 93% of these positive samples and the cultural procedures 92%. Combination of the results with chocolate scrapings and cocoa bean dust gives a total of 353 positive samples. MSRV gave 330 positive results (productivity 93.5%) and 23 false-negative results, while the cultural procedures gave 326 positive (productivity 92%) and 27 false-negative results. Statistical calculation using the McNemar test (Siegel, 1956) gives a X 2 value of 0.18, which means there is no statistical significant difference between the two procedures.

With the chocolate scrapings and the cocoa dust samples together, the cultural procedure produced 326 positive samples. With RV broth, 96% of these tests were positive and with SC 75%. The superiority of RV agrees with the results of studies with food samples (Vassiliadis, 1983; De Smedt and Bolderdijk, 1990). With the combination of direct motility enrichment and indirect motility enrichment, 330 tests were found positive for Salmonella. With direct motility enrichment, 98% of these tests were positive and with indirect motility enrichment 87%. These findings agree with the findings of a collaborative study on Salmonella detection from cocoa and chocolate products (De Smedt and Bolderdijk, 1990). Recently a collaborative study was performed with a broad range of food samples in which direct motility enrichment on MSRV was compared with a cultural procedure in which RV was used as selective broth (De Zut ter et al., 1991). In this study the productivity for naturally contaminated foods was 96% for MSRV and 90% for RV. In the present study direct motility enrichment on MSRV and the cultural procedure with RV produced together 349 positive samples. The productivity of direct motility enrichment was 322 positives or 92% and the productivity of RV 312 positives or 89%. Statistical calculation according the McNemar test (Siegel, 1956) shows a X 2 value of 1.27 which is also not significant at the 0.05 level of significance.

In the application of the Hazard Analysis Critical Control Point concept for the prevention of food-borne salmonellosis (Simonsen et al., 1987) it is important to keep the food-processing environment under control.

Microbiological tests for the presence of Salmonella in environmental samples are essential for the timely identification of possible contamination sources. The present study showed that motility enrichment on MSRV is a sensitive method for such tests. The great advantages of this procedure are its simplicity, cost-efficiency

308

(De Zutter et al., 1991) and rapidity. Serologically confirmed test results are obtained within 48 h after the beginning of pre-enrichment. A larger number of tests can easily be performed, which gives a better safety assurance for keeping the environment under control.

References

Association of Official Analytical Chemists (AOAC) (1984) Bacteriological Analytical Manual (6th edn.). AOAC, Arlington, VA.

De Smedt, J.M. and Bolderdijk, R. (1987) Dynamics of Salmonella isolation with modified semisolid Rappaport-Vassiliadis medium. J. Food Protect. 50, 658-661.

De Smedt, J.M. and Bolderdiik, R. (1990) Collaborative study of the International Office of Cocoa, Chocolate and Sugar Confectionery on the use of motility enrichment for Salmonella detection in cocoa and chocolate. J. Food Protect. 53, 659-664.

De Smedt, J.M., Bolderdijk, R., Rappold, H. and Lautenschlaeger, D. (1986) Rapid Salmonella detection in foods by motility enrichment on modified semisolid Rappaport-Vassiliadis medium. J. Food Protect. 49, 510-514.

De Zutter, L., De Smedt, J.M., Abrams, R., Beckers, H., Catteau, M., de Borchgrave, J., Debevere, J., Hoekstra, J., Jonkers, F., Lenges, J., Notermans, S., Van Damme, L., Vandermeersch, R., Ver- braeken, R. and Waes, G. (1991) Collaborative study on the use of motility enrichment on Modified Semisolid Rappaport-Vassiliadis medium for the detection of Salmonella from foods. Int. J. Food Microbiol. 13, 11-20.

Jameson, J.E. (1962) A discussion of the dynamics of Salmonella enrichment. J. Hyg. (Cambridge) 60, 193-207.

Rappold, H., Bolderdijk, R. and De Smedt, J.M. (1984) Rapid cultural method to detect Salmonella in Foods. J. Food Protect. 47, 46-48.

Siegel, S. (1956) Nonparametric Statistics for the Behavioral Sciences. McGraw Hill, New York, NY. Simonsen, B., Bryan, F.L., Christian, J.H.B., Roberts, T.A., Tompkin, R.B. and Silliker, J.H. (1987)

Prevention and control of food-borne salmonellosis through application of Hazard Analysis Critical Control Point (HACCP). Int. J. Food Microbiol. 4, 227-247.

Van Schothorst, M. and Renaud, A.M. (1983) Dynamics of Salmonella isolation with modified Rappaport's medium (R10). J. Appl. Bacteriol. 54, 209-215.

Van Schothorst, M. and Renaud, A.M. (1985) Malachite green pre-enrichment medium for improved salmonellae isolation from heavily contaminated samples. J. Appl. Bacteriol. 59, 223-230.

Vassiliadis, P. (1983) The Rappaport-Vassiliadis (RV) enrichment medium for the isolation of salmonellas: an overview. J. Appl. Bacteriol. 54, 69-76.

Zapatka, F.A., Varney, G.W. and Sinskey, A.J. (1977) Neutralization of the bactericidal effect of cocoa powder on salmonellae by casein. J. Appl. Bacteriol. 42, 21-25.