Contamination flows of Bacillus cereus and spore-forming aerobic
bacteria in a cooked, pasteurized and chilled zucchini
puree processing line M.H. Guinebretiere * , H. Girardin, C. Dargaignaratz, F. Carlin, C. Nguyen-The Institut National de la Recherche Agronomique, UMR A408 Securite et Qualite des Produits dOrigine Vegetale, INRA, Domaine Saint-Paul, Site Agroparc, F-84914 Avignon Cedex 9, France Received 25 August 2001; received in revised form 10 May 2002; accepted 26 June 2002 Abstract A food processing plant producing pasteurized purees and its zucchini puree processing line were examined for contamination with aerobic and facultative anaerobic bacterial spores during a days operation. Multiplication of spores was also monitored in the product stored under different conditions. High concentrations of Bacillus cereus spores were found in the soil in which the zucchinis were grown (4.6 F0.3 log CFU/g), with a background spore population of 6.1 F0.2 log CFU/g. In the processing plant, no B. cereus or psychrotrophic bacterial spores were detected on equipment. B. cereus and psychrotrophic bacterial spores were detected after enrichment in all samples of raw zucchinis, washed zucchinis, of two ingredients (starch and milk proteins) and in processed puree at each processing step. Steam cooking of raw zucchinis and pasteurization of puree in the final package significantly reduced spore numbers to 0.5 F0.3 log CFU/g in the processed food. During storage, numbers of spore-forming bacteria increased up to 7.8 F0.1 log CFU/g in puree after 5 days at 2025 jC, 7.5 F0.3 log CFU/g after 21 days at 10 jC and 3.8 F1.1 log CFU/g after 21 days at 4 jC. B. cereus counts reached 6.4 F0.5 log CFU/g at 2025 jC, 4.6 F1.9 log CFU/g at 10 jC, and remained below the detection threshold (1.7 log CFU/g) at 4 jC. Our findings indicate that raw vegetables and texturing agents such as milk proteins and starch, in spite of their low levels of contamination with bacterial spores and the heat treatments they undergo, may significantly contribute to the final contamination of cooked chilled foods. This contamination resulted in growth of B. cereus and psychrotrophic bacterial spores during storage of vegetable puree. Ways to eliminate such contamination in the processing line are discussed. D 2002 Elsevier Science B.V. All rights reserved. Keywords: Bacillus cereus; Bacterial spores; Cooked chilled foods; Vegetables; Processing line 1. Introduction Cooked chilled foods supply a rapidly expanding market in Europe. Consumption of chilled prepared meals increased by 10% to 37% according to the product in 1997, while at the same time frozen prepared meals fell by 4% (Hauben, 1999). The main 0168-1605/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. PII: S0168- 1605( 02) 00307- 0 * Corresponding author. Tel.: +33-4-32-72-25-24; fax: +33-4- 32-72-24-92. E-mail address: guinebre@avignon.inra.fr. (M.H. Guinebretiere). www.elsevier.com/locate/ijfoodmicro International Journal of Food Microbiology 82 (2003) 223232 objective of processors of cooked chilled foods is to offer a ready-to-use product with optimized sensory qualities. A large proportion of cooked chilled foods are pasteurized in their final packaging to eliminate hazards from vegetative pathogenic bacteria. How- ever, this pasteurization does not kill bacterial spores, which can survive the process and develop during storage before consumption. The development of cooked chilled foods may thus result in an increased exposure of consumers to pathogenic spore-forming bacteria. The spore-forming bacterium Bacillus cereus has been incriminated in many cases of food poisoning (Granum, 1997). Recently, vegetable purees were the cause of a severe B. cereus outbreak in a French nursing home for elderly persons (Lund et al., 2000; De Buyser, personal communication). B. cereus is a ubiquitous bacterium found in soil and in many raw and processed foods such as rice, milk and dairy products, spices, and vegetables (Roberts et al., 1982), meat products and farinaceous foods (Kramer and Gilbert, 1989). Its ubiquitous nature and the heat resistance of its endospores, associated with psychrotrophic properties, make it a potential con- taminant of cooked chilled foods. B. cereus was isolated from 80% to 100% of samples of cooked pasteurized and chilled vegetable purees of leek, zucchini, broccoli, split pea, carrot and potato purees stored at 10 jC (Carlin et al., 2000). Among these vegetable purees, zucchini puree supported the most rapid bacterial growth and B. cereus was significantly present in products stored at 10 jC (3.6 to 5.2 log CFU per gram) at use-by date. The zucchini puree thus called for a more thorough investigation. It received a mild heat treatment (pasteurization in final packaging) that cannot be increased to levels sufficient to kill B. cereus spores and psychrotrophic spores. Therefore, a knowledge of the flows of contamination through the process- ing line should help identify critical steps in the process for B. cereus and psychrotrophic contami- nation, and so help improve safety by reducing or eliminating contamination at source. The objective of the present study was to deter- mine the possible sources of contamination for zucchini purees and the impact of processing on this contamination, with special reference to B. cereus and psychrotrophic spores. For this purpose, we monitored the flows of contamination during the production of one batch of cooked and pasteurized zucchini purees from their hypothetical sources to the final stored product. Possible primary sources of contamination were the zucchinis used for process- ing, the soil where the zucchinis were grown and the ingredients (flavouring and texturing agents) blended with the puree. Possible secondary sources of contamination were the surfaces of equipment and air in contact with the product during process- ing. B. cereus and psychrotrophic spore-forming bacteria were tested for all these possible contami- nation sources, in the product at each stage of processing, and in the puree packs stored until consumption. Only spores were enumerated or de- tected along the processing line, because pasteuriza- tion eliminates vegetative cells in the final pack- aging of the product. 2. Materials and methods 2.1. Cooked pasteurized and chilled vegetable puree Zucchini puree was manufactured in a processing plant located in Normandy, France. Raw zucchinis were harvested in a field in the same area. The raw zucchinis (300 kg) were washed, cut, steam cooked for a few minutes, ground, and mixed with 44 kg of ingredients (UHT milk cream, milk proteins, starch and salt) (Fig. 1). The resulting puree was dispensed into preformed polypropylene trays in 400-g units using an industrial liquid food dispenser, then vac- uumed and covered with a 90-Am thick polypropy- lene-polyamide film sealed on the tray. The packed zucchini puree was then pasteurized in a water bath to achieve a core temperature of 80 jC during 10 min. The equipment used to process the cooked zucchinis comprised a grinder, a blender, the distri- bution system previously described and three con- tainers for the transport of food materials between the different items of equipment. There was a phys- ical separation between the reception areas (raw vegetable reception area, ingredient storage area) and the puree production area (vegetable washing room, production room, pasteurization room, cold storage room for final products). Six hundred and sixty-eight puree packs of 400 g were produced M.H. Guinebretiere et al. / International Journal of Food Microbiology 82 (2003) 223232 224 during our study, corresponding to one batch of zucchini puree. 2.2. Sampling strategy Sampling was performed on: cultivation soil of zucchinis, raw zucchinis, washed zucchinis, processed zucchinis (cooked and ground zucchinis, cooked/ ground zucchinis mixed with ingredients, zucchini purees pasteurized in their final package), the ingre- dients (UHT cream, starch, milk proteins, salt), the tap water used to dilute ingredients, the surfaces of equip- ment in contact with the product, and production room air. Packs of puree were also analyzed after storage at the temperature recommended in France for cooked chilled foods (4 jC), and at 10 and 2025 jC, representing defective storage temperatures that may occur in consumer refrigerators and when the cold chain is broken. All these processing points were studied over 1 day of production. The production involved large volumes of foods ( f280 kg) and large equipment surfaces. We expected low levels of B. cereus spores on samples during processing from the results of Carlin et al. (2000) and Choma et al. (2000). We therefore chose to analyze a large volume of samples, covering approximately 10% of the food used for processing and 50% of the equipment surfaces. This approach reduced the accuracy of counts but allowed a better detection of B. cereus. 2.3. Sampling procedure for soil Soil samples were taken from two different field plots, on three rows for each plot. Dry soil was randomly sampled with sterile spoons (stainless steel spoons purchased from the local supermarket) on the plastic lining covering the cultivated soil and around the zucchini stems from a given row. Six samples of 500 g, one per row, were collected in sterile stomacher bags (Laboratoire Humeau, La Chapelle-sur-Erdre, France). The bags were closed hermetically with clips (Laboratoire Humeau, La Chapelle-sur-Erdre). 2.4. Sampling procedure for equipment, wash water and air Samples from equipment surfaces were collected immediately before the start of processing to esti- mate the potential transfer of contamination to the zucchini puree. Ten pieces of equipment were sampled. For each piece of equipment, one half of Fig. 1. Flow diagram of the processing line producing cooked, pasteurized and chilled zucchini puree. M.H. Guinebretiere et al. / International Journal of Food Microbiology 82 (2003) 223232 225 the surface was swabbed with a sterile cotton swab; these were ready-to-use chiffonnette ATL (40 40 cm), impregnated with a sterile buffer containing neutralizers against disinfectant solutions (Labora- toire Humeau, La Chapelle-sur-Erdre) and were used as recommended by the manufacturer. The surfaces studied were the grinder screw (2800 cm 2 ), the grinder inner surface (14,600 cm 2 ), the inner surface of the three stainless steel transport containers (18,000 cm 2 each), the blender grid (4800 cm 2 ), the blender inner surface (22,500 cm 2 ), the rotating mixing blade (8700 cm 2 ) and the outlet pipe of the distribution system. Two 70-ml samples of the water used to clean the distribution pipe were collected in sterile vials (Labo- ratoire Humeau), one before and one at the end of puree distribution. Viable airborne B. cereus spores were determined by the sedimentation method. Nineteen open Petri dishes (90 mm in diameter) containing B. cereus selective medium MYP (van Netten and Kramer, 1992) were left open near equipment for 30 min during grinding and blending operations. The dishes were then sealed with ParafilmR and incubated at 28 jC for 48 h. 2.5. Sampling procedure for zucchinis, processed zucchinis, and ingredients Six raw zucchinis per box were randomly selected to obtain a total of 120 zucchinis ( f30 kg of zucchinis). One slice ( f25 g) was cut from each zucchini with a sterile scalpel and collected in a stomacher bag. One sample unit was composed of 24 slices ( f600 g), a total of five samples was collected. The same sampling procedure was used for washed zucchinis. This sampling strategy allowed the survey of approximately 10% of the raw and washed zucchinis used for processing. For processed zucchinis, samples of cooked/ ground zucchinis ( f1 kg each) were collected directly in stomacher bags at the outlet of the grinder. Five samples were collected at regular intervals from the start to the end of the grinding operation. Five samples of processed zucchinis mixed with ingre- dients ( f1 kg each) were collected in stomacher bags with a sterile ladle after the transfer of the food into the transport container. Forty packs of processed purees (400-g unit) were randomly selected at the end of the processing line. For ingredients, a total of approximately 800 g of milk proteins, 600 g of starch, and 150 g of salt were collected with sterile spoons in stomacher bags (Table 1). Three liters of UHT cream were also collected (Table 1). Additionally, two 70-ml samples of tap water used to dilute dehydrated ingredients were collected in sterile vials. 2.6. Sample maintenance and transport In the processing plant, hydrated samples (zuc- chinis, processed zucchinis, liquid samples, surface samples) were immediately cooled and kept in melt- ing ice until the end of the production. All the samples were maintained in a cold room at 4 jC overnight, and at 2 jC during a 24-h transport step and in the laboratory until they were analyzed. During analysis, samples were maintained in melting ice. Processed zucchinis (cooked/ground zucchinis, cooked zucchinis mixed with ingredients and packs of nonstored pasteurized purees) were analyzed on arrival at the laboratory. Some packs of pasteurized puree were incubated at 4 jCF0.0, 10 jCF0.5 for 21 days and at ambient temperature (2025 jC) for 5 days. The incubation time at low temperature (4 and 10 jC) corresponded to the usual shelf life applicable to cooked chilled foods in France; the incubation time at 2025 jC was the time at which spoilage was first visible. Ten packs were stored in each storage condition. Raw and washed zucchinis, liquid samples and surface samples were analyzed within 24 h after their arrival in the laboratory. Soil samples and dehydrated ingredients were analyzed within 2 weeks after their arrival in the laboratory. 2.7. Microbial analyses Spores were enumerated on J-agar (Claus and Berkeley, 1986), and B. cereus spores on MYP agar (van Netten and Kramer, 1992). B. cereus spores were detected after enrichment of samples at 28 jC for 24 h in both modified MYP broth (MYP broth without egg yolk and mannitol) and Claus medium (Claus and M.H. Guinebretiere et al. / International Journal of Food Microbiology 82 (2003) 223232 226 Berkeley, 1986). Psychrotrophic bacterial spores were detected after enrichment in J-broth (Claus and Ber- keley, 1986) at 10 and 4 jC for 4 and 14 days, respectively. In a preliminary step, samples of zucchinis, soil and cotton swabs were homogenized for 2 min as described in Table 1, using, respectively, a rotating blender (Kinematica, Luzern, Switzerland), a conical flask containing a sterile magnetic rod and a Stom- acher (Seward Medical, London, UK). For soil, raw and processed zucchini(s), swabs, and water, each sample was divided up into five portions in stomacher bags; one was used for bacterial counts and the remainder for bacterial enrichments (Table 1). For ingredients, samples were distributed between direct counts and enrichments as described in Table 1. The stomacher bags were heat-sealed (Impulse sealer TISH-300, Bioblock, Illkirch, France) and sub- jected to a pasteurization. Heat treatment at 80 jC for 10 min is a common procedure used for the pasteu- rization of samples (Claus and Berkeley, 1986). In our study, most samples represented large volumes after dilution (Table 1). Therefore, stomacher bags were immersed in a water bath (Polystat 55, Bioblock) for 15 min at 80 jC, to allow heat diffusion. Small samples (cotton swabs and water samples) were heat-treated at 80 jC for 10 min. All samples were cooled in melting ice. The bags for enrichments were Table 1 Number, size and preparation of samples Samples Number Sample Direct counts a Enrichments a of samples size F2% Preliminary homogenate, containing for one sample Number of samples Sample size Sterile distilled Number of Sample size Enrichment broth Sterile phosphate Sterile distilled H 2 O added (ml) samples added (ml) buffer b (ml) H 2 O (ml) BCE c CE c Soil 6 500 g 1000 6 200 ml d 12 12 200 ml d 200 Raw courgettes 5 600 g 475 5 200 ml d 10 10 175 ml d 25 (4 ) e Washed zucchini(s) 5 600 g 475 5 200 ml d 10 10 175 ml d 25 (4 ) Cotton swabs 10 1 unit 45 10 20 ml d 20 20 7.5 ml d 2.5 (4 ) CG zucchini(s) f 5 1000 g 5 200 g 150 10 10 200 g 150 CGZ with ingredients f 5 1000 g 5 200 g 150 10 10 200 g 150 Water pipe cleaning 2 70 ml 2 10 ml 4 4 15 ml 5 (4 ) Tap water 2 70 ml 2 10 ml 4 4 15 ml 5 (4 ) UHT cream 15 200 ml 3 200 ml 200 4 8 200 ml 200 Starch 29 20 g 3 20 g 380 14 12 20 g 380 Milk proteins 7 3 g 3 3 g 20 2 2 3 g 200 Milk proteins 6 130 g 2 4 130 g 240 Salt 29 3 ou 5 g 3 3 g 100 14 12 5 g 500 Pasteurized purees 10 50 g 5 100 g 200 Stored purees (for each storage condition) 10 50 g 5 100 g 400 a For soil, raw and processed zucchini(s), swabs, and water, each sample was divided up into five portions; one was used for bacterial counts and the remainder for bacterial enrichments. For ingredients, samples were distributed between direct counts and enrichments. b Phosphate buffer, K 2 HPO 4 8.5 g/l, KH 2 PO 4 6.75 g/l, pH 6.8. c BCE, B. cereus enrichments in MYP and Claus medium; CE, cold enrichments in J-broth at 4 and at 10 jC. d Sub-sample of the preliminary homogenate. e 4 , Enrichment medium 4 concentrated. f CG, cooked and ground; CGZ, cooked and groung zucchini(s). M.H. Guinebretiere et al. / International Journal of Food Microbiology 82 (2003) 223232 227 then incubated at the desired temperature/time con- ditions. Before plating, the bags were homogenized for 1 min using a Stomacher (Seward Medical). Homoge- nates were serially diluted and 100 Al of each dilution were spread plated on J-agar and MYP agar. For plate counts, 4 to 9 100 Al and 1 500 Al of the non- diluted homogenate were also spread plated on each medium to lower the detection thresholds. MYP plates and J-agar plates were incubated at 28 jC for 24 and 48 h, respectively. For the pasteurized puree, the content of each pack was homogenized with sterile spoons. Approximately 50 g of puree per pack was collected with a sterile spoon and samples from two packs were pooled and weighed in sterile stomacher bags ( f100 g per bag). The five resulting samples were used for counts as previously described. For stored purees (puree packs stored at 4, 10 and 2025 jC), only one 100 Al aliquot of the nondiluted homogenate was spread plated. 2.8. Isolation and confirmation of B. cereus isolates For each sample, three to five colonies were randomly selected from MYP plates bearing 20 to 50 colonies. Only colonies with the typical mannitol negative and lecithinase positive characteristics of B. cereus were taken into account (van Netten and Kramer, 1992). The presumptive B. cereus cultures were grown on J-agar plates to check purity, and kept frozen in 30% (vol/vol) glycerol solution at 20 and 80 jC. Identification of isolates to the B. cereus group was confirmed by shape and position of endospore under contrast phase microscopy ( 1000) and XP V 08- 058 procedure (Anonymous, 1996). 2.9. Statistical analysis and counts Counts were expressed as log CFU per gram or per milliliter of food samples, or CFU per piece of equip- ment. Variation among replicate samples was expressed by standard deviation. To evaluate the effect of washing, cooking, addition of ingredients and pasteurization on spore levels in processed zuc- chinis, means before and after each operation were compared using Students t-test (Systat version 9, SPSS Chicago). 3. Results and discussion 3.1. Confirmation of B. cereus isolates A total of 196 presumptive B. cereus isolates were obtained from the different samples. Subse- quent confirmation showed that 100% of the isolates had the typical spore morphology of B. cereus. Among the 196 isolates, 20 had atypical characters (hemolysis negative or weakly positive), but on MYP plates and J-agar plates, they exhibited char- acteristic rhizoid outgrowths of B. mycoides that belongs to the B. cereus group. Thus, 100% of the isolates was regarded as belonging to the B. cereus group. 3.2. Contamination levels in soil Levels of B. cereus spores in the cultivation soil of the zucchinis were quite high: 4.6 F0.3 log CFU/ g. Spore counts were 6.1 F0.2 log CFU/g. It is generally accepted that the primary habitat of most Bacillus species is the soil where they are considered as forming the greater part of the metabolically active bacterial flora (Felske et al., 1998). The levels and ratio of spores to total cell numbers of Bacillus spp. vary according to irrigation period, fertilizer practice, season and climate (Von Stetten et al., 1999; Watanabe and Hayano, 1995). In soil, spore number is usually high; Ueda and Kuwabara (1986) found levels of 4.3 F0.4 to 5.1 F0.5 log CFU/g in rice field soils, and Watanabe and Hayano (1995) found 10 5 to 10 7 log CFU/g in rice and wheat rotation soils, consistent with our results. In temper- ate soils, the abundance of B. cereus spores ranged from 10 5 to 10 6 CFU/g according to Von Stetten et al. (1999) and 10 2 to 10 5 CFU/g according to Te Giffel et al. (1995). Zucchinis are the aerial part of the plant. However, many fruits are in contact with soil or aerosols of soil particles. The high levels of spore-forming bacteria measured in soil samples where the zucchinis were grown suggest that the cultivation soil constituted a significant source of contamination for zucchinis with B. cereus and other Bacillus spp. Soil is also consid- ered as a source of B. cereus contamination for other foods such as milk (Andersson et al., 1995; Te Giffel et al., 1995). M.H. Guinebretiere et al. / International Journal of Food Microbiology 82 (2003) 223232 228 3.3. Contamination from equipment surfaces and air No bacterial spores were detected in the piped water used to clean the puree distribution system. The total number of spores collected on half the total surface of equipment before the start of production ranged from 160 to 19,000 spores according to the piece of equipment. The highest contamination occurred on rotating equipment such as the screw and mixing blade (Table 2). As surfaces of equip- ment came into contact with 280 kg of zucchini puree, we can consider that contamination potentially transferred by equipment surfaces to the final prod- uct was low. No B. cereus spores or psychrotrophic bacteria spores were detected in samples from equipment surfaces. Current procedures for cleaning the distribution system included three successive steps of washing with hot water. Equipment surfaces were cleaned with hot water under high pressure before and after each production step. At the end of each processing day, surfaces were cleaned with disinfectant solutions (Penngar neige, Penngar, Le Mans, France; Anios W4, Laboratoire Anios, Lilles, France) containing a fungicide, bactericides and notably one B. cereus sporicide (Norm NF T 72-300). Wirtanen et al. (1996) showed that the high mechanical forces of the cleaning flow, heat treatment and chemical treat- ment were efficient in removing Bacillus biofilms on equipment surfaces. Given the low levels of contam- ination measured on equipment in our study, we can conclude that the cleaning procedures employed were effective and satisfactorily controlled B. cereus on equipment surfaces. On MYP plates exposed to the air of the process- ing room for 30 min, only one viable B. cereus particle was detected among 19 plates. Counts of these bacteria in air are poorly documented. Ueda et al. (1988) found averages of airborne bacterial counts ranging from 0.2 to 10.7 CFU per liter of air in mill plants, composed of 26% to 33% of viable B. cereus particles. In our study, the product was in contact with air mainly during its transport in the three containers. Contamination transferred by air to the final product was thus presumably very low. 3.4. Contamination levels of ingredients In the ingredients, spore counts were close to or below the limit of detection, except for milk proteins and starch, which carried 2.8 F0.6 and 1.6 F0.1 log CFU per gram (Table 3), respectively. B. cereus spore Table 2 Counts of bacterial spores on equipment surfaces before the beginning of production a Total (CFU) b B. cereus (CFU) b Grinder screw 5.3 10 3 V1.6 10 2 Grinder inside surface 2.3 10 3 V1.6 10 2 Blender grid 1.7 10 3 V1.6 10 2 Rotating mixing blade 1.9 10 4 V1.6 10 2 Blender inside surface 1.7 10 2 V1.6 10 2 Container inside surface no. 1 3.3 10 2 V1.6 10 2 Container inside surface no. 2 V1.6 10 2 V1.6 10 2 Container inside surface no. 3 1.7 10 2 V1.6 10 2 Distributor tip 1.3 10 3 V1.6 10 2 Distributor conveyor belt 1.7 10 2 V1.6 10 2 Pipe cleaning water A c V2 V2 Pipe cleaning water B c V2 V2 a Enrichments for B. cereus spores and for psychrotrophic bacterial spores were all negative. b Numbers are the number of spores present on half of the surface of each piece of equipment and the number of spores per milliliter for water samples. c A, before the beginning of production; B, after the end of the production. Table 3 Counts and detection of bacterial spores in ingredients added to the puree Spore counts a Enrichments b Total B. cereus B. cereus Psychrotrophs (log CFU/g) (log CFU/g) 10 jC 4 jC UHT cream < 0.6 < 0.6 0/4 0/4 0/4 Starch 1.6 F0.1 < 1.6 4/4 2/2 2/2 Milk proteins 2.8 F0.6 < 1.4 4/4 3/3 3/3 Salt < 1.2 < 1.2 0/4 0/2 0/2 Tap water < 0.3 < 0.3 0/4 0/2 0/2 a Means Fstandard deviation (n =3). b Number of positive samples relative to the total number of enriched samples; for starch and salt, the 26 enriched samples were pooled during analyses so as to obtain four analyses for B. cereus enrichment and four analyses for cold enrichments; samples sizes were 200 ml for UHT cream, 15 ml for tap water, 60 to 80 g for starch (pooled samples), 15 to 20 g for salt (pooled samples), and 3 to 130 g for milk proteins. M.H. Guinebretiere et al. / International Journal of Food Microbiology 82 (2003) 223232 229 counts were below the level of detection in all the ingredients. After enrichment, B. cereus and psychro- trophic bacterial spores were detected in all the samples of starch and milk proteins. Of the ingredients, milk proteins and starch con- stituted the principal sources of spore contamination. B. cereus and psychrotrophic spore-forming bacteria have often been considered as contaminants of farina- ceous foods, dried foods, dehydrated milk and milk products (Blakey and Priest, 1980; Kramer and Gil- bert, 1989; Roberts et al., 1982; Wong et al., 1988). Levels of B. cereus generally measured on milk powder and dried products ranged from 10 to 10 5 CFU/g (Te Giffel, 1997). 3.5. Contamination of zucchinis, processed zucchinis and stored product On raw zucchinis, spore counts were 1.5 F0.2 log CFU/g and washing operations failed to reduce the contamination on washed zucchinis (Table 4). Contamination level significantly decreased after cooking according to Students t-test ( P= 0.02) and increased after the addition of ingredients ( P= 0.001). Finally, pasteurization resulted in a signifi- cant decrease in spore counts ( P = 0.001) to 0.5 F0.3 log CFU/g. Purees were pasteurized in their final package. The marked effect of the pasteu- rization may be explained by the germination of a proportion of spores during the distribution and packaging of the puree. The resulting contamination level in unstored pasteurized product represented 640 to 2,500 spores per puree pack of 400 g. Spore- forming bacteria were able to multiply in the product stored at cold temperatures, reaching, after 21 days, 3.8 F1.1 log CFU/g at 4 jC and 7.5 F0.3 log CFU/g at 10 jC (Table 5). B. cereus counts were less than or equal to 0.4 log CFU/g on raw and processed zucchinis (Table 4). However, B. cereus and psychrotrophic bacterial spores were detected after enrichment in most samples at all stages of processing. The number of positive samples did not increase after the addition of ingredients. Thus, levels of B. cereus contamination transferred by starch and milk pro- teins were probably of the same order of magnitude as that transferred by zucchinis after cooking. On packaged purees, B. cereus counts were less than or equal to 0.2 log CFU/g (Table 4). This level was nevertheless sufficient to initiate growth of B. cereus in products stored at 10 and 2025 jC. Counts of B. cereus reached 4.6 F1.9 log CFU/g in purees stored at 10 jC for 21 days (Table 5), higher than the upper limit of 4.0 log CFU/g recommended in France for some processed vege- tables (Jouve, 1996). In purees maintained at 4 jC, no significant growth of B. cereus was measured. As 4 jC is the recommended storage temperature for these products in France (Anonymous, 1991), B. cereus should not be a hazard provided refrigeration is properly main- tained throughout the cold chain (cold storage room Table 4 Counts and detection of bacterial spores in raw zucchini(s) and processed zucchinis Spore counts a Enrichments b Total B. cereus B. cereus Psychrotrophs (log CFU/g) (log CFU/g) 10 jC 4 jC Raw zucchini(s) 1.5 F0.2 < 0.4 10/10 5/5 5/5 Washed zucchini(s) 1.7 F0.6 < 0.4 10/10 5/5 5/5 Cooked/ground zucchini(s) 0.9 F0.4 < 0.3 2/10 4/5 2/5 Cooked/ground zucchini(s) mixed with ingredients 3.2 F0.8 < 0.4 2/10 4/5 2/5 Pasteurized purees 0.5 F0.3 < 0.2 nt nt nt a Means Fstandard deviation (n = 5). b Number of positive samples relative to the total number of enriched samples; sample weighs were 100 to 200 g. M.H. Guinebretiere et al. / International Journal of Food Microbiology 82 (2003) 223232 230 of the processing plant, transport, distribution, and consumer refrigerator). Fluctuations of storage before consumption of the product, most often occurring during distribution of the product or in the refriger- ator of consumers, may induce growth of B. cereus, as some strains from pasteurized vegetable purees were able to grow at temperatures below 10 jC (Choma et al., 2000). 4. Conclusion B. cereus and psychrotrophic bacterial spores were detected at many points along the processing line of zucchini puree and in particular in all the expected primary sources of contamination (soil, zucchini(s), ingredients). The efficient cleaning procedure used for equipment surfaces and especially the use of B. cereus sporicide prevented the installation of B. cereus on equipment surfaces. Also, salt and UHT cream added to the puree did not carry B. cereus and psychrotro- phic spores. Raw material (zucchinis) and texturing agents (milk proteins and starch) used to process the puree were, therefore, the main sources contributing to the contamination with B. cereus and psychrotro- phic bacterial spores. They both transferred a low level of B. cereus spores, but that was sufficient to induce bacterial growth in the stored product. Pre- vious work (Carlin et al., 2000; Choma et al., 2000) showed that the low levels of contamination recovered in unstored vegetable purees did not vary significantly according to season. Our work shows that secondary sources of contam- ination (equipment surfaces) can be effectively avoided. Research and efforts from processors should therefore be focused on means to significantly reduce primary sources of contamination. The final pasteuri- zation is applied to a complex mixture of vegetables, dairy cream and texturing agents, and any increase in temperature over 80 jC, or any increase in treatment time, would result in unacceptable adverse changes in the product. In addition, heat transfer through the packs of puree is slow and this further reduces any possibility for intensifying the pasteurization treat- ment. In contrast, there are more possibilities for strengthening heat treatment applied during the cook- ing of zucchinis. They can be cut into small pieces to permit rapid heat transfer. Improved steam cooking technologies should enhance the heat treatment with- out causing unwanted alterations (Varoquaux and Nguyen-The, 1994). This would permit a further reduction in the number of spores carried by zucchinis. However, such reduction would be worthwhile only if milk proteins and starch free of B. cereus spores could be obtained, as no step except the final pasteurization could reduce this source of contamination. This work represents the first analysis of contam- ination flows occurring during processing of cooked chilled foods. It shows that although texturing agents carry a very low number of B. cereus spores, they represent a critical step for contamination and pre- sumably limit reduction of the incidence of B. cereus in cooked chilled foods. It emphasizes the importance of hygienic processing of dehydrated ingredients for the general safety of the cooked chilled foods sector. Acknowledgements This work was supported by research project FAIR CT 97-3159 (Commission of the European Commun- ities). The authors thank the food company involved for active co-operation, and Dr Cindy Morris (Station de Pathologie Vegetale, INRA, Avignon, France) for the help in designing the sampling strategy. References Andersson, A., Ronner, U., Granum, P.E., 1995. What problems does the food industry have with the spore-forming pathogens Bacillus cereus and Clostridium perfringens? International Jour- nal of Food Microbiology 28, 145155. Anonymous, 1991. Guide de bonnes pratiques hygieniques des plats Table 5 Counts of spore-forming bacteria in pasteurized puree according to storage conditions a log CFU/g B. cereus Total (log CFU/g) Purees stored at 4 jC (21 days) 3.8 F1.1 < 1.7 Purees stored at 10 jC (21 days) 7.5 F0.3 4.6 F1.9 Purees stored at 2025 jC (5 days) 7.8 F0.1 6.4 F0.5 a Means Fstandard deviation (n = 5). M.H. Guinebretiere et al. / International Journal of Food Microbiology 82 (2003) 223232 231 prepares refrigeres, elabores par le Syndicat National des Fabri- cants de plats prepares. Bulletin Officiel CCRF 6, 261269 (Septembre). Anonymous, 1996. Microbiologie des alimentsDenombrement de Bacillus cereus par comptage des colonies a` 30 jC. In: AFNOR (Ed.), Analyse microbiologiqueMethodes horizontales, Tome 1. AFNOR, Tour Europe, Paris la Defense, pp. 445457. Blakey, L.J., Priest, F.G., 1980. The occurrence of Bacillus cereus in some dried foods including pulses and cereals. Journal of Applied Bacteriology 48, 297302. Carlin, F., et al., 2000. Spore-forming bacteria in commercial cooked, pasteurized and chilled vegetable purees. Food Micro- biology 17, 153165. Choma, C., et al., 2000. Prevalence, characterization and growth of Bacillus cereus in commercial cooked chilled foods containing vegetables. Journal of Applied Microbiology 88, 617625. Claus, D., Berkeley, R.C.W., 1986. Genus Bacillus Cohn 1872, 174 AL . In: Sneath, P.H.A., Mair, N.S., Sharpe, M.E., Holt, J.G. (Eds.), Bergeys Manual of Systematic Bacteriology, vol. 2. Williams and Wilkins, Baltimore, pp. 11051139. Felske, A., Wolterink, A., Van Lis, R., Akkermans, A.D.L., 1998. Phylogeny of the main bacterial 16S rRNA sequences in Dren- tse A grassland soils (The Netherlands). Applied and Environ- mental Microbiology 64, 871879. Granum, P.E., 1997. Bacillus cereus. In: Doyle, M.P., Beuchat, M.P., Montville, T.J. (Eds.), Food Microbiology: Fundamentals and Frontiers. ASM Press, Washington, D.C., pp. 327336. Hauben, K., 1999. Sous vide cooking: state of the art. Proceedings of the Third European Symposium on Sous Vide, Leuven, Bel- gium, March 2526, 1999. ALMA, Leuven, pp. 1227. Jouve, J.L., 1996. Produits vegetaux et dorigine vegetale. In: Poly- technica (Ed.), La qualite microbiologique des aliments, Ma t- rise et crite`res. Europe Media duplication, Lassay-les-Chateaux, France, pp. 67198. Kramer, J.M., Gilbert, R.J., 1989. Bacillus cereus and other Bacillus species. In: Doyle, M.P. (Ed.), Foodborne Bacterial Pathogens. Marcel Dekker, New York, pp. 2170. Lund, T., DeBuyser, M.L., Granum, P.E., 2000. A new cytotoxin from Bacillus cereus that may cause necrotic enteritis. Molec- ular Microbiology 38, 254261. Roberts, D., Watson, G.N., Gilbert, R.J., 1982. Contamination of food plants and plant products with bacteria of public health significance. In: Rhodes-Roberts, M.E., Sckinner, F.A. (Eds.), Bacteria and Plants. Academic Press, London, pp. 169195. Te Giffel, M., 1997. Isolation, identification and characterization of Bacillus cereus from the dairy environment. Thesis. Landbou- wuniversiteit Wageningen, The Netherlands, 160 pp. Te Giffel, M.C., Beumer, R.R., Slaghuis, B.A., Rombouts, F.M., 1995. Occurrence and characterization of (psychrotrophic) Ba- cillus cereus on farms in the Netherlands. Netherlands Milk and Dairy Journal 49, 125138. Ueda, S., Kuwabara, Y., 1986. The ecological study of Bacillus cereus in rice crop processing (1). On the distribution of B. cereus in rice field soil. Journal of Antibacterial and Antifungal Agents 14, 377382. Ueda, S., Yamazaki, M., Machida, A., Akao, A., Kuwabara, Y., 1988. The ecological study of Bacillus cereus in rice crop pro- cessing (3). Airborne B. cereus in rice mills. Journal of Anti- bacterial and Antifungal Agents 16, 459464. van Netten, P., Kramer, J.M., 1992. Media for the detection and enumeration of Bacillus cereus in foods: a review. International Journal of Food Microbiology 17, 8599. Varoquaux, P., Nguyen-The, C., 1994. Vacuum processing: a new concept for pre-cooking fruits and vegetables. Food Science and Technology Today 8, 4249. Von Stetten, F., Mayr, R., Scherer, S., 1999. Climatic influence on mesophilic Bacillus cereus and psychrotolerant Bacillus weihen- stephanensis populations in tropical, temperate and alpine soil. Environmental Microbiology 1, 503515. Watanabe, K., Hayano, K., 1995. Seasonal variation of soil pro- tease activities and their relation to proteolytic bacteria and Bacillus spp. in paddy field soil. Soil Biology & Biochemistry 27, 197203. Wirtanen, G., Husmark, U., Mattila-Sandholm, T., 1996. Microbial evaluation of the biotransfer potential from surfaces with Bacil- lus biofilms after rinsing and cleaning procedures in closed food- processing systems. Journal of Food Protection 59, 727733. Wong, H.C., Chang, M.H., Fan, J.Y., 1988. Incidence and charac- terization of Bacillus cereus isolates contaminating dairy prod- ucts. Applied and Environmental Microbiology 54, 699702. M.H. Guinebretiere et al. / International Journal of Food Microbiology 82 (2003) 223232 232