Abstract
Consumption of milk and dairy products has increased significantly in Senegal in the last decade, and a large part of the local production comes from small processing units spread all over the country. We collected 85 bulk-tank milk samples from 68 smallholder dairy farms throughout the territory. Microbiological quality of milk samples was analyzed according to the official standards. Further, raw milk and pasteurized milk were screened for Mycobacterium bovis, Coxiella burnetii, and anti–Brucella abortus antibodies. Ninety-three percent of pasteurized milk samples, 92% of raw milk samples, and 81% of sour milk samples failed to meet official standards. Pathogens detected in milk were C. burnetii (6/41, 15%), which seems to be endemic in Senegal, coagulase-positive staphylococci (18/70, 26%), and Salmonella Johannesburg in one sample. Further analysis of coagulase-positive staphylococci isolated from samples containing more than 104 colony-forming units per gram showed the presence of enterotoxigenic strains in 9 of the 10 samples. These results confirm the poor microbiological quality of milk produced by small units in Senegal, especially and surprisingly of pasteurized milk. This highlights the need to implement good hygiene practices, particularly in the postpasteurization process, and an effective monitoring throughout the production and delivery chain.
Introduction
I
As collection of data on milk quality is relevant to increase awareness among national policy makers and international donor agencies about the needs of these countries, we examined the microbiological quality of milk produced by small processing units in Senegal.
Materials and Methods
Small milk-processing units and sampling techniques
Eighty-five bulk-tank milk samples (15 raw milk, 26 pasteurized milk, and 44 sour milk) were collected between November 2005 and March 2006 from 68 small milk-processing units distributed throughout Senegal. Sour milk is produced from the fermentation of milk powder or pasteurized milk. Milk samples were collected aseptically from the final products, placed directly in cool boxes, and analyzed within 12–24 h of storage at <6°C.
Bacteriological analysis
Samples were prepared and diluted according to International Organization for Standardization (ISO) standard 8261. Diluted samples were tested according to the standards of ISO or AFNOR (which is a French Association for Standardization). All samples were examined for the total coliform count (CC) according to AFNOR standard NF V08-050. Raw and pasteurized samples were subjected to standard plate counts (SPC) according to AFNOR standard NF V08-051. Sour milk was subjected to fecal CCs according to AFNOR standard NF V08-060 and to Escherichia coli counts according to AFNOR standard NF V08-053. All samples were screened for Listeria monocytogenes according to AFNOR standard NF V08-062, and for Salmonella spp. according to ISO standard 6579. Coagulase-positive staphylococci (CPS) were counted according to AFNOR standard NF V08-057-1 in raw and sour milk samples, and Campylobacter spp. in pasteurized and sour milk samples according to ISO standard 10272-1.
β-Hemolytic streptococci (BHS) were screened in raw milk as follows. Briefly, 25 g of milk was diluted in 225 g of buffered peptone water broth. After incubation, 0.1 mL of this broth was streaked on Columbia agar containing 5% sheep blood and incubated at 37°C for 48 h. Suspected BHS colonies were confirmed by Gram staining, catalase negativity, and Lancefield serogrouping. The results were expressed as the presence or absence of BHS in 25 g of milk.
CPS DNA extraction and toxin typing
CPS colonies recovered from milk samples with a CPS content higher than 104 colony-forming units per gram (cfu/g) were subjected to further study. Genomic DNA was extracted from a single colony with a standard phenol-chloroform procedure, and then submitted to polymerase chain reaction (PCR) for detection of genes encoding staphylococcal enterotoxins (se) A, B, C, D, H, K, L, M, O, P, Q, and R (sea-d, seh, sek-m, and seo-r), toxic shock syndrome toxin 1 (tst), and Panton Valentine leukocidin (luk-PV), as previously described (Tristan et al., 2003).
Detection of anti-Brucella abortus antibodies by enzyme-linked immunosorbent assay
An enzyme-linked immunosorbent assay (ELISA) method was applied to pasteurized and raw milk samples to detect specific antibodies according to the manufacturer's instructions (ELISA Brucellosis milk; Institut Pourquier, Montpellier, France). Results were expressed as a positivity index calculated by dividing the optical density of the sample by the optical density of the positive control. Samples with an index of ≥0.55 were considered positive, and values between ≥0.45 and <0.55 were considered equivocal.
Detection of Coxiella burnetii by real-time PCR
DNA extraction was performed on pasteurized and raw milk samples by using a QIAamp DNA mini kit (Qiagen, Hilden, Germany). Real-time PCR was done with the ABI 7000 sequence detection system (Perkin-Elmer–Applied Biosystems, Foster City, CA) and the TaqMan kit (Adiagene Laboratories, Saint Brieuc, France) according to the manufacturer's instructions.
Detection of Mycobacterium bovis by touch-down PCR
DNA isolation from pasteurized and raw milk samples and touch-down amplification were carried out as previously described (Zumarraga et al., 2005). The detection limit of touch-down PCR was determined by testing 10-fold dilutions of M. bovis in sterile DNA-free water. The detection limit was 10 cfu (equivalent to about 50 fg of DNA).
Data analysis
SPC and CC values were compared between pasteurized and raw milk samples by using the Kruskal–Wallis test. Differences were considered significant at p < 0.05. Statistical analysis was performed using Stata software version 9. Microbiological results were interpreted according to official Senegalese microbiological criteria (
Results and Discussion
Ninety-three percent of pasteurized milk samples, 92% of raw milk samples, and 81% of sour milk samples failed to meet official standards, indicating the poor microbiological quality of the milk tested. Milk quality according to each microbiological parameter is displayed in Table 1. The high levels of contamination observed indicate unsanitary conditions or practices during the pre- and postpasteurization processes and/or a deficiency in pasteurization itself (Chambers, 2002). Surprisingly, the average levels of contamination in raw and pasteurized milk (4.5 × 107 cfu/g and 3.5 × 108 cfu/g, respectively, for SPC, and 8.5 × 104 cfu/g and 3.2.107 cfu/g, respectively, for CC) were not significantly different (p > 0.05). These data, combined with that from a previous study conducted in pasteurization centers in The Gambia, Senegal, and Guinea (Hempen et al., 2004), and also in Brazil (Silva et al., 2009), suggest that pasteurization is not the only critical step for improving the microbiological quality of milk products. The unsatisfactory quality of pasteurized milk is the consequence of the poor quality of raw milk used and/or a high level of recontamination after pasteurization. Postpasteurization contamination was also observed in restaurants and markets in developed and emerging countries like Spain and Brazil (Silva et al., 2009; Sospedra et al., 2009). Our findings highlight the fact that pasteurized milk of such poor microbiological quality, found as early as the farm level, is obviously of serious concern particularly in a developing country.
NS, unsatisfactory quality; SPC, standard plate count; cfu/g, colony-forming units per gram; BHS, β-hemolytic Streptococcus; CPS, coagulase-positive Staphylococcus.
None of the samples tested contained BHS, Campylobacter spp., L. monocytogenes, or M. bovis. B. abortus antibodies and Salmonella enterica serotype Johannesburg were detected in only one raw milk sample. The results in this regard were by and large similar to those of previous surveys of bulk-tank milk in Senegal, which also showed the absence or a low frequency of all these pathogens (Hempen et al., 2004; Unger and Münstermann, 2004). Nevertheless, C. burnetii was present in 6 (15%) of the 41 milk samples tested (6 farms, 4 of which were in the Dakar region). All were raw milk (6/15, 40.0%) as pasteurization process normally inactivates this microorganism. To the best of our knowledge, no other data are available on C. burnetii in Senegal and only two studies based on serological tests have previously been conducted in Africa (Kelly et al., 1993; Nakoune et al., 2004). C. burnetii seems to be endemic to Senegalese cattle, particularly in the Dakar region. Given the possibility of abortion in dairy cattle and fatal complications in patients at risk (Maurin and Raoult, 1999), further investigations are needed to evaluate its prevalence in humans, as well as the economic implications for farmers. CPS, a common cause of mastitis in dairy cattle, was detected in 18 (26%) of the 70 samples tested. Ten (14%) of these samples (one raw milk and nine sour milk) contained more than 104 cfu/g. Nine of the 10 CPS strains contained enterotoxin genes, inferring that their consumption could have caused illness. Two other major virulence genes (tst and luk-PV) were also detected, each in one isolate. The presence of CPS with potent superantigenic activity (tst gene) (Kikuchi et al., 2003) and/or the capacity to cause invasive disease (luk-PV gene) (Prevost et al., 1995) and/or food poisoning (enterotoxins) is a major public health concern.
Conclusion
Overall, these findings call for (1) implementation of good hygiene practices throughout the milk chain by the training of the professionals involved in milk collection and processing, including pasteurization, transport, and delivery, to ensure the safety and quality of milk, (2) adequate inspection of production facilities with microbiological controls of milk, and (3) funds to supply farmers with adequate equipment and facilities.
Footnotes
Acknowledgments
The authors thank Babacar Gning for his technical assistance, and Ikem Eronini and David Young for editorial assistance. This study was supported by a grant from the French Ministry for Foreign Affairs.
Disclosure Statement
No competing financial interests exist.