Consumption of unprocessed milk and its by-products is potentially hazardous and has been associated with several types of infections including brucellosis (Ramos et al., 2008), tuberculosis (Doran et al., 2009), salmonellosis (Poppe, 2011), yersinosis (Greenwood and Hooper, 1990), Escherichia coli O157:H7 (Anand and Griffiths, 2011) and staphylococcal enterotoxins poisoning (Baylis, 2009; Ostyn et al., 2010). Despite modern dairy ensuring that consumers get processed safe milk, consumption of unprocessed milk is still common in Zimbabwe especially in the rural areas with the two main types being fresh liquid milk and naturally fermented milk. Although the food regulations in Zimbabwe impose a general responsibility to ensure safety of any food brought to the market on the producer, this has not always been the case due to a number of factors including, cultural, economic and poor legislation enforcement among others.

The Zimbabwean dairy industry can be easily categorised into 2 groups, the formal and the informal sector with the latter posing greater safety concerns. The informal dairy sector mainly produces two products, raw milk termed “fresh milk” and the naturally fermented/cultured milk (Amasi/Hodzeko). The latter product taking greater preference, because it is used as a relish which provides proteins at a low cost for the lower class while it is a delicacy for the affluent that they consume as part of the main meal or dessert. Due to the faltering economy of Zimbabwe, many dairy farmers have been downsizing operations and ultimately preferring to sell their milk product directly to consumers without supplying any processor (Gadaga et al., 1999; Gran et al., 2003).

Fermentation of milk is important as it is a cheap traditional way of improving nutritional properties as well as sensory properties (Gadaga et al., 1999). It is not only effective on flavour development but has been used as a preservation method for many years; hence, fermented foods are regarded as microbiologically safe. The low pH is effective in inhibiting the growth of many microbes (Jeevartnam and Jamuna, 2004). Staphylococcus aureus among other pathogens is destroyed by pH lower than 5 (Medveďová and Valík, 2012). However, emerging research has shown that some strains of E. coli can survive the low pH (Gran et al., 2003; Bore et al., 2007). A combination of harsh conditions and prolonged exposure time has been said to result in microbes building some adaptation mechanism (Lucking et al., 2013). The ability of pathogens such as E. coli to  survive low pH emerges as a potential health hazard to consumers of the naturally fermented milk, which is generally regarded as a safe product in Zimbabwe due to its low pH. The aim of this research was to determine the incidence of pathogens (E. coli, S. aureus, and Bacillus cereus) in fresh milk, naturally fermented/cultured milk, milk handlers as well milk packaging containers used at farm level in the informal sector in Zimbabwe.

Three farms (FD 1, FD 2 and FD 3) around Harare province were selected for the research. The microbial assessment scheme (MAS) methodology developed by Jacxsens et al. (2011) was used to determine the food safety status in the informal milk processing farms. A total of 4 critical sampling locations (milk handler, milk container, fresh milk, and cultured milk) were selected. A total of 4 samples (milk handler swab, milk container swab, 50 ml fresh and cultured milk) from each dairy farm were collected during each visit and the sampling was done 4 times over a three month period. The samples were transported at 4°C to Government Analyst Laboratory for analysis within 3 h of collection. Samples were analyzed for total aerobic bacteria count (TBC), total coliform count (TCC), E. coli, Salmonella species, Pseudomonas species, S. aureus, and B. cereus. Methods of microbial analysis are summarized in Table 1. After enumeration, isolates were, purified and stored at -20°C in 30% sucrose solution awaiting polymerase chain reaction (PCR).

PCR Reactions

B. cereus, S. aureus and E. coli isolates were resuscitated and DNA was extracted according to procedures described in Medina-Acosta and Cross (1993). All PCRs were carried out in 10 µl reaction tubes in a Bio-Rad T100™ Thermal Cycler (Singapore). The reaction mix consisted of 1 µl bacterial DNA template, 5 µl KAPA BIOSYSTEMS 2X KAPA Taq Ready Mix (Cape Town, South Africa), 3.4 µl PCR-grade water, and 0.6 µl primers. The primers and annealing temperatures used in the reactions are listed in Table 2. The PCR conditions for cytK are described in Swiecicka and Mahillon (2006), while conditions for rfB and sei are described in another study (Thapa et al., 2012). PCR products were separated by Gel electrophoresis using a 2% Agarose gel ULTRAPURE^TM (USA) and the image was viewed using a Bio-Rad Gel Doc™ EZ (USA).

Statistical analysis of data

aDescriptive statistics and other explanatory analysis were used as well as statistical techniques: univariate analysis (one way analysis of variance [ANOVA]) using SPSS version 16.0 and STATA 12 adjusted for multiple comparisons to detect significance difference between means. The significance test was at alpha level 0.05 or 95% confidence level.

Pseudomonas spp.

The incidence of Pseudomonas spp. was generally low in all three dairies. However, the fresh milk collected from FD 1 had high levels of Pseudomonas (4.16 log₁₀ cfu/ml) and the minimum recorded was 0.68 log₁₀ cfu/ml. No incidence of Pseudomonas spp. was recorded in FD 3 in cultured milk and milk handler from FD3.

B. cereus

The highest incidence of B. cereus in all CSLs was recorded in FD3 followed by FD1. The least incidence was recorded in FD 2. Fresh milk in all three dairies had higher levels of Bacillus species when compared with cultured milks of all the three dairies.

PCR reactions

The genomic DNA for S. aureus, B. cereus, and E. coli were subjected to PCR to investigate the presence of virulent genes, rfB, sei, and cytK with amplicon sizes of 1.0 kb, 500 bp and 320 bp, respectively. The sei gene was detected in 19% of the samples and 2.8% were found to have the cytK gene. The rfB gene could not be picked in E. coli.

TBC results were not in agreement with previous studies (Al-Tarazi et al., 2003) who reported considerably lower levels of TBC in fresh milk with mean values of 1.1 × 10⁷ cfu/ml. However, they concurred with previous studies in Zimbabwe that showed high TBC in raw milk (Gran et al., 2003). While it is reported that raw milk TBC should be less than 100 000 cfu/ml (Gosta, 1995), Zimbabwean food regulations require that raw milk should not exceed 20 000 cfu/ml (Food and Food Standard Act, 2001). High bacterial load in raw milk has been attributed to poor hygienic practices that include failure to properly clean the udder and use of contaminated water among others (Gran et al., 2003). The bacteria in milk handler are also likely to find its way into the raw milk and other subsequent stages of milk processing (Gran et al., 2003). In food safety, implementation of prerequisites and application of HACCP guarantee the control of processes (Domenech et al., 2013). The absence of documented hygienic practice followed by the processors in this study might also have contributed to high levels of contaminants. Cusato et al. (2014) also included diagnosis of prerequisites programs (PRPs),  implementation  of  good manufacturing practices (GMP), standard sanitation operating procedures (SSOP) and training of the food as the major steps in the implementation of food safety system. However, such conditions are likely to be found in formal organizations which are well established mostly with food safety management systems. Out of the fourteen PRPs (CAC, 2003; Jacxsens et al., 2010; Luning et al., 2008) informal processors in this study can concentrate on cleaning and disinfection, personal hygiene of the milk handlers, and temperature control as recommended by Kussaga et al. (2014) as a way of improving the microbiological quality of the milk. This concurs with Lee et al. (2012) who pointed out that high frequency of S. aureus isolates in dairy farms highlights the need for constant improvement of hygiene and quality assurance.

While fermentation results in increased acidity that discourages the growth of most pathogens (Gran et al., 2003), our results showed the presence of potential pathogens in both fresh milk and fermented milk. Our findings on coliforms and E. coli were in agreement with previous studies (Saudi and Moawad, 1990; Ahmed and Sallam, 1991; Sobeih et al., 2002; Al-Tarazi et al., 2003; Chye et al., 2004; Korashy and Mohammed, 2008; Altalhi and Hassan, 2009) who reported that all examined fresh milk samples were contaminated with coliforms. These results were in violation of both European and Zimbabwean standards for raw milk intended for processing and milk for consumption (Gran et al., 2003). These findings concurred with a previous study in Zimbabwe that concluded on poor hygiene and sanitary practices during milking and further handling (Gran et al., 2003).

In this study, fresh milk with 4.16 log₁₀ cfu/ml of Pseudomonas spp. was found to be the most prevalent. The presence of Pseudomonas spp. in both fresh and fermented milk indicates that Pseudomonas spp. can survive in fermented foods. This is in violation of food regulations (Dairy Regulations, 1977; Food and Food Standards Act, 2001; Public Health Act, 2001; CAC, 2003) where the food is not supposed to have either pathogenic or spoilage organisms.

In spite of generally low E. coli counts, its presence is suggestive of poor hygienic indicators, faecal contamination and implies a risk that other enteric pathogens may be present in the product (Gran et al., 2003). The findings concur with previous studies which indicated that poor hygiene was evidenced by the presence of E. coli (Ghafir et al., 2008; Jacxsens et al., 2009; Sampers et al., 2010). The presence of E. coli violates regulations which prohibit the presence of pathogenic and spoilage microorganisms in any food item intended for human consumption (Dairy Regulations, 1977; Food and Food Standards Act, 2001; Public Health Act, 2001; CAC, 2003). The results in this study indicated that the E. coli survives in fermented milk. This concurs with another study which reported that acid adapted  cells showed a marked increase in levels of resistance to lactic acid even though the level of resistance varied among strains (Gregory et al., 1995). Our results showed that none of our strains belonged to O157:H7. This contradicted with a previous study that suggested the presence of the pathogenic E. coli in a number of food items (Prasad et al., 2012).

The presence of S. aureus at all the CSLs in all the dairies contravening various food regulations was in agreement with a previous study (Gran et al., 2003). The isolation rate observed in this study was higher than was reported by Abdel-Hammed and El Malt (2009) who reported that 24% were contaminated with S. aureus. Dissemination of S. aureus from humans to food can occur by direct contact, indirectly by skin fragments, or through respiratory tract droplet nuclei (Jablonski and Bohach, 1997). S. aureus is also commonly found in mastitis udder (Wellenberg et al., 2002), hence common in milk from animals suffering from mastitis. The minimum amount of S. areus required to produce intoxication in human is estimated to be about 5 log₁₀ cfu/g (Rørvik and Granum, 1999). To produce sufficient enterotoxin, the pH should be higher than 4.6 and the temperature should be above 15°C for more than 3 to 4 h (Rørvik and Granum, 1999). If S. aureus gains access to the milk before fermentation, the pH would have been higher than 4.6 for longer than 6 h and therefore pose a definite risk of toxin production during the early part of the fermentation. Our findings on the presence of the toxin producing sei gene in S. aureus was in line with a previous study (Nazari et al., 2014), although we could not distinguish between the most toxic sei 1 and sei 2. Similar to previous studies (Larsen and Jogernsen, 1997), B. cereus was isolated in milk and other sampling locations. The prevalence of the cytK in B. cereus from our study was in line with previous studies that concluded that its distribution is low in B. cereus.

Fresh and fermented milk from informal processors in Harare pose a potential threat to consumers and this is in violation of both the Zimbabwean and international food regulations. Poor hygiene practices by the milk handlers maybe the greatest link of milk contamination. The presence of virulent genes (19% S. aureus and 2.8% B. cereus) confirms that consumers of fermented milk are at safety risk.

The authors have not declared any conflict of interests.

The  authors  greatly  appreciate  the  assistance  of   Mr. Musiyambiri (Director Government Analyst) and his team for the support during laboratory work; Dr. Mutambu (Director Institute of Health Research), for the PCR machine; and also Dr. Mharakurwa for his advice during PCR work. Thanks are also given to the farmers and the milk handlers for their cooperation.