Corresponding Author K. K. Addo, Department of Bacteriology, Noguchi Memorial Institute for Medical Research, PO Box LG 581, Legon, Ghana. Tel.: +233 30 2501178 9; Fax: +233 30 2502182; E-mail: firstname.lastname@example.org
Objectives To investigate the microbiological quality and the presence of antibiotic residues in raw cow milk and in some indigenous milk products produced and marketed by the informal sector in the coastal savannah zone of Ghana.
Methods Milk samples were aseptically collected from 224 kraals and samples of 26 indigenous milk products were purchased from processors and retailers. Total plate counts, total coliform counts and the presence of Escherichia coli and E. coli O157:H7 were determined in all 250 samples. Milk samples were also tested for antibiotic residues.
Results Total plate counts exceeded 105 CFU/ml in 45.2% of the samples while coliforms exceeded 103 CFU/ml in 66.0% and E. coli was detected in 11.2%. E. coli was present in raw cow milk but not in the indigenous products and all E. coli isolates were negative for E. coli O157:H7. Antibiotic residues were detected in 3.1% of the raw cow milk samples.
Conclusion Bulk milk contains unacceptable levels of hygiene indicators and antibiotic residues and is a potential source of milk-borne infections. The detection of E. coli and antibiotic residues raises public health concerns about the safety of fresh unpasteurized cow milk in the coastal savannah zone of Ghana and calls for improved farm hygiene, the need for milk pasteurization and the sensible use of antibiotics in the milk industry.
Objectifs: Investiguer la qualité microbiologique et la présence de résidus d’antibiotiques dans le lait cru de vache et dans certains produits laitiers indigènes, produits et vendus via le secteur informel dans la zone de savane côtière du Ghana.
Méthodes: Des échantillons de lait ont été prélevés aseptiquement dans 224 ‘kraals’ et des échantillons de 26 produits laitiers indigènes ont été achetés chez des producteurs et des détaillants. La numération totale sur plaque, le nombre de coliformes totaux et la présence d’Escherichia coli et E. coli O157:H7 ont été déterminés dans les 250 échantillons. Des échantillons de lait ont également été testés pour les résidus d’antibiotiques.
Résultats: La numération totale sur plaque dépassait 105 UFC/ml dans 45,2% des échantillons; les coliformes dépassaient 103 UFC/ml dans 66,0% et E. coli a été détecté dans 11,2%. E. coliétait présent dans le lait cru de vache, mais pas dans les produits indigènes et tous les isolats de E. coliétaient négatifs pour E. coli O157:H7. Les résidus d’antibiotiques ont été détectés dans 3,1% des échantillons de lait cru de vache.
Conclusion: Le lait en vrac contient des taux inacceptables d’agents pathogènes et de résidus d’antibiotiques et est une source potentielle d’infections transmises par le lait. La détection d’E. coli et des résidus d’antibiotiques pose des problèmes de santé publique sur la sécurité du lait frais de vache non pasteurisé dans la zone de savane côtière du Ghana et appelle pour une hygiène agricole améliorée, à la nécessité de la pasteurisation du lait et à l’utilisation judicieuse des antibiotiques par les fermes laitières.
Objetivos: Investigar la calidad microbiológica y la presencia de residuos antibióticos en leche de vaca cruda y algunos productos lácteos indígenas producidos y vendidos por el sector informal en la zona de sabana costera de Ghana.
Métodos: Se recolectaron de forma aséptica muestras de leche en 224 kraals y se compraron muestras de 26 productos lácteos indígenas a procesadores y minoristas. Se determinaron el recuento total en placas, el recuento total de coliformes y la presencia de Escherichia coli y E. coli O157:H7 en las 250 muestras. Las muestras de leche también se examinaron en búsqueda de residuos de antibióticos.
Resultados: El recuento total en placa excedió el 105 UFP/ml en un 45.2% de las muestras; el recuento total de coliformes excedió el 103 UFP/ml en un 66.0% y se detectóE. coli en un 11.2%. E. coli estaba presente en la leche cruda pero no en los productos indígenas y todos los aislados de E. coli eran negativos para E. coli O157:H7. Se detectaron residuos de antibióticos en 3.1% de las muestras de leche cruda.
Conclusiones: La leche analizada contiene niveles inaceptables de patógenos y residuos de antibióticos y es una fuente potencial de infecciones transmitidas en la leche. La detección de E. coli y residuos de antibióticos plantea inquietudes con implicaciones para la sanidad pública en lo referente a la seguridad de la leche de vaca no pasteurizada proveniente de la zona de sabana costera de Ghana, y habla de la necesidad de mejoras en la higiene de las granjas y de pasteurizar la leche así como de un uso responsable de antibióticos en las granjas lecheras.
Animal protein consumption in Ghana is well below the recommended levels. The per capita consumption of milk is a mere 3 l per annum compared to the sub-Saharan average of 23 l (Delgado et al. 1999). The average annual demand of milk in Ghana is just over 70,000 metric tonnes of which more than half is imported (GOG/FAO 2002). Thus, while the potential market for the local dairy production is huge, the contribution to the total consumption is relatively low because it is perceived to be produced and marketed under unhygienic conditions and hence unsafe (Karikari et al. 1998). Raw milk is an important vehicle for the transmission of zoonotic and other pathogens to humans, as milk can easily be contaminated with cattle faeces during milking (MacDonald et al. 1988). Moreover, milk serves as an excellent growth medium for microorganisms (Soomro et al. 2002). In Ghana, microorganisms such as Yersinia, Klebsiella, Proteus, Enterobacter, Escherichia coli, Staphylococcus, Bacillus and Mycobacterium spp. have been isolated from raw milk (Donkor et al. 2007). Escherichia coli are frequently found contaminants and a reliable indicator of faecal pollution of water and food products (Diliello 1982). Also associated with the consumption of a number of contaminated foods among them meat, especially undercooked ground beef and raw milk is E. coli O157:H7 (Riley et al. 1983). It causes potentially fatal haemorrhagic enteritis or colitis leading to bloody diarrhoea and haemolytic uraemic syndrome in humans due to the production of potent verocytotoxins, associated with serious kidney damage and renal failure (Jay 1992; Besser et al. 1993).
Milk and milk products may also be contaminated with antibiotic residues such as sulphonamides, nitrofurans, beta lactams, etc. These agents are widely used at high dosages for the treatment of diseases in dairy cattle (McEvoy et al. 2000). Residues present in milk increase the number of antibiotic resistant pathogenic bacteria in people who are not allergic to the drug (Tolentino et al. 2005) while causing allergic reactions in sensitive people (Thomson & Sporns 1995). Antimicrobial agents and antibiotics in milk have been reported in both advanced and developing countries (Brady & Ketz 1998). In Ghana, there is limited information about their occurrence in raw milk. Aning et al. (2007) reported that 35% of the raw milk marketed in two major cities, Accra and Kumasi, were contaminated.
Milk produced in Ghana by the informal sector is not regulated by any agency and such milk may pose a health hazard due to contamination with pathogens and antibiotics. In this study we investigated the microbial quality of raw cow milk and dairy products produced in the coastal savannah zone of Ghana and determined the presence of antibiotic residues.
Materials and methods
Selection of milk producing kraals
Ghana is divided into 10 regions (regions are subdivided into districts) and portions of 3 of these, the Central (CR), Greater-Accra (GAR) and Volta (VR) regions form the coastal savannah zone. Within these 3 regions, 12 districts with intensive dairy production were purposively selected for the study (Figure 1). A list of all cattle owners in each district was compiled with the help of the Ministry of Food and Agriculture. A preparatory visit was paid to listed kraals where the aim of the study was explained to the producers and verbal consent sought. Based on proportional representation, 224 kraals were selected from the list of 400 for the study. Of these, 38 were located in CR, 91 in GAR and 95 in VR. The selected kraals produced about 50–500 l of milk per day, being classified as small to medium scale producers. At all the kraals, milking was manual. The milk is either sold raw or processed into wagashi for sale.
Collection of milk and samples of other dairy products
From each kraal, approximately 100 ml of milk was aseptically withdrawn from the container containing the pooled milk from that day and deposited into sterile bottles. The samples were placed in a cool box with ice packs (to maintain temperature at 4 °C) and sent to the bacteriology laboratory of the Noguchi Memorial Institute for Medical Research. Microbiological testing was done within 24 h of milking. The remaining portion of the sample was frozen at −20 °C until testing for antibiotic residues. Samples were collected between May and July 2007. The other dairy products, wagashi (n = 12 samples), burchina (n = 9 samples) and yoghurt (n = 5 samples), were obtained from processors and retailers and these were transported and stored as described for the milk samples. Wagashi is a type of cottage cheese prepared by bringing fresh milk to the boil after which a plant extract consisting of the pounded vegetative part of Calotropis procera is added. The extract serves as a coagulating agent to curdle the milk which is then drained to obtain the semi-solid cheese. Burchina is a beverage prepared by combining fresh milk with millet. Yoghurt is fermented raw milk.
Sample preparation and culture for bacteriological quality assessment
For each sample a tenfold serial dilution (10−1 to 10−7) was prepared in sterile phosphate buffered saline (PBS) pH 7.2 (Dulbecco ‘A’). For liquid samples (raw milk, burchina and yoghurt) dilutions were made by aseptically withdrawing 1 ml of each sample with a sterile disposable pipette into a sterile tube containing 9 ml of PBS. For solid samples (wagashi) 10 g taken with a sterile spatula, was placed into a stomacher bag and homogenized with 90 ml of PBS. A 1-ml portion of each homogenate was then used to prepare the tenfold dilutions. Subsequently, for total bacterial counts a 1-ml sample of each dilution was pipetted into a 90-mm diameter disposable Petri dish and mixed well with 20 ml of sterile standard plate count (SPC) agar (APHA, Oxoid). The SPC agar was prepared according to manufacturer’s instructions. Separately, the sample dilutions 10−1 to 10−4 were cultured for coliform counts. To this end a 1 ml of the dilution was pipetted into duplicate 90 mm diameter Petri dishes and mixed well with 20 ml of violet red bile (VRB) agar (Oxoid). The VRB medium was prepared according to manufacturer’s instructions. After cooling and solidification, the SPC agar plates were incubated at 32 °C for 48 h and VRB agar plates at 37 °C for 24–48 h. All samples were cultured in duplicate on each of the two media.
Quantification of colonies
SPC agar plates with countable colonies between 25 and 250 CFU/plate and VRB agar plate with countable dark red colonies (Hall et al. 1967) between 15 and 150 CFU/plate were chosen for counting with the aid of a colony counter (Gerber).
Screening for E. coli
Coliform colonies were examined for suspected E. coli. In order to increase chances of detecting E. coli and serotype 0157:H7 in particular, up to six suspected E. coli colonies per plate were purified on MacConkey agar (Oxoid) and differentiated for E. coli by plating on Eosine Methylene Blue (EMB) agar (Oxoid). On MacConkey agar only pink coloured lactose-positive colonies were selected and streaked on EMB. On EMB metallic green coloured smooth sided colonies with typical E. coli morphology were deemed as E. coli and confirmed by testing for indole, methyl red, vogues proskaeuer and citrate (IMViC) reactions.
Screening for E. coli 0157:H7
All confirmed E. coli isolates were cultured on Sorbital MacConkey for 24 h and colourless colonies indicative of non-sorbital fermenter, a characteristic of E coli 0157:H7 (March & Ratman 1986), were tested for the presence of E. coli 0157:H7 using the latex slide agglutination kit (DR062M, Oxoid). One drop of the test latex was placed close to the circle on the reaction card. A loopful of saline was then placed in the circle using a Pasteur pipette. Using a sterile loop, a portion of the colony to be tested was picked off, carefully emulsified in the saline drop and mixed to form a smooth suspension. The latex was then mixed with this smooth suspension and spread to cover the reaction area using the loop. The card was rocked in a circular motion while observing for agglutination. Colonies showing agglutination within 1-minute are identified in this assay as E. coli 0157:H7.
Detection of antibiotic residues
The Charm Blue-Yellow antibiotic residue test kit (Charm Sciences Incorporated, MA, USA) was used. Each raw milk sample was shaken thoroughly and 50 μl pipetted into the purple agar portion of the test well. Samples were run in duplicates and each run included positive and negative controls. The wells were covered with clear sealing tape to seal the rims and incubated at 64 ±1 °C for 2 h and 45 min. After incubation, the wells were cooled to room temperature for 5 min and the colour read under white light within 15 min. Colours were compared to reference colours; yellow or yellow to green colours indicate the absence of antibiotic residues and blue to purple colours indicate the presumptive presence of antibiotic residues. To avoid false positive results all samples showing a blue or purple colour were validated by retesting the sample after boiling 500 μl of the sample for 2 min. The sample was cooled, mixed thoroughly and tested in duplicate as outlined above. Only samples still positive after heat treatment were considered as ‘Blue Yellow screening test positive’. If the sample was negative after heat treatment, it was considered to probably contain a non-antibiotic, heat-sensitive component.
Data analysis was performed using Microsoft Office Excel 2003 (Microsoft Excel, Palisade Corp., Newfield, NY, USA). Relative proportions were compared using the chi-squared test and Fisher’s exact test (http://home.clara.net/sisa/twoby2.htm) and a probability value of less than 0.05 was defined statistically significant. Comparisons of means were achieved using two-tailed Student’s t-tests using Microsoft Excel.
Ethical approval for the study was granted by the Ghana Health Service’s ethical review board.
In 45.2% of the samples, the total plate counts were above 105 CFU/ml (Table 1). The counts were lowest (with 68.4% of samples recording counts below 105) for CR and highest in GAR (with only 40% of samples with counts below 105) but these differences were not significant (P > 0.05). Similar values were observed for both fresh milk samples and dairy products. The number of coliforms was higher than 103 CFU/ml in 66% of the samples and was higher for fresh milk than for other dairy products (Table 2). There was no significant difference (P > 0.05) between the regions in terms of either level of total bacteria or coliform count, indicating that hygienic conditions in all three regions were comparable. E. coli was detected in 11.2% of the raw milk samples as the other diary products were all negative for E. coli (Table 3). E. coli serotype O157:H7 was not detected. Antibiotic residues were found in 3.1% of the raw milk samples.
Table 1. Total plate counts of raw milk and milk product samples
Sample type by region (number)
Number (%) of samples in the following range*
Total raw milk (224)
Milk products (26)
Total dairy (250)
Table 2. Total coliform counts of raw milk and milk product samples
Sample type by region (number)
Number (%) of samples in the following range*
Total raw milk (224)
Milk Products (26)
Total dairy (250)
Table 3. Prevalence of E. coli, E. coli O157:H7 and antibiotic residues in raw milk samples produced within the coastal savannah zone regions of Central, Volta and Greater-Accra
Sample type by region (number)
Number (%) of samples positive for the following:
E. coli O157:H7
nt = not tested.
Total raw milk (224)
Milk Products (26)
Total dairy (250)
Bacterial counts in milk reflect the level of hygiene practised during milking and milk collection, the storage temperature and the time elapsed since milking (Soler et al. 1995). While total bacterial counts mainly reflect the time elapsed since milking or the processing at ambient temperature, coliform bacteria generally reflect faecal contamination due to poor hygiene. According to international regulations milk should be delivered and refrigerated within 2 h (EU 1994) or 3 h (IDF 1990) after milking. The majority of the raw milk samples cultured had acceptable levels (≤105 CFU/ml) of total bacteria count but only about one third had acceptable levels (<103 CFU/ml) of coliforms. These findings compare with similar studies done in Tanzania and Ghana (Omore et al. 2009) and indicate that poor farm hygiene and practices during milking are a major concern for the informal milk market in the coastal savannah zone of Ghana. Faecal contaminants in the milk may originate from the udder of the animal or from the farm environment (Oliver et al. 2005). A typical kraal in our study consists of an enclosure of varying size surrounded by wooden or bamboo sticks with a bare ground covered with thick layer of fresh and dried cow dung. Teats are often not washed prior to milking because of the belief that allowing the heifers to drink milk before manual milking cleans the teats. Furthermore, herdsmen tie the hind legs of a cow with a rope and start drawing milk without first cleaning their hands. Also water sources at kraals may form a rich source of faecal contamination as these often consists of dug out wells or stagnant water bodies contained by dams that during rainfall collects the water from the kraal and its direct environment. Use of this water for cleaning milking cans prior to milking subsequently contaminates the milk.
Escherichia coli was isolated from raw milk samples but not from the dairy products. A possible explanation is that the heat treatment given these products may be sufficient to inactivate heat-labile organisms including E. coli or as these products were sampled soon after production (early morning), contamination from excessive handling and the environment were minimal. All E. coli isolates tested negative for E. coli 0157:H7 which is consistent with the fact that no outbreak related to food poisoning caused by E coli 0157:H7 has been reported in Ghana.
In 3.1% of the raw milk samples antibiotic residues above the European Union Maximum Residue Limit (EU MRL) were found. Antimicrobials detected by this assay include beta-lactams, sulphonamides, amino-glycosides, tetracycline and macrolides.
In Kenya 9.4% of milk samples at the consumer households and 5.7% from market agents had antibiotic residues above the EU MRL (Aboge et al. 2000). A much higher level of 35.5% was previously reported for milk samples from the peri-urban areas of Accra and Kumasi, Ghana using the Charm-AIM screening test kit (Aning et al. 2007), and of marketed milk samples from Mwanza and Dar es Salaam in Tanzania, 36% tested positive with the Charm-AIM screening test kit (Kurwijila et al. 2006) which has now been replaced by the Charm Blue-Yellow assay used in this study. The presence of antibiotic residues in milk is of concern because of the risk of developing drug-resistance of common human pathogens (Nijsten et al. 1996). Besides, antibiotics may cause allergic reactions and may interfere with the growth of starter cultures used in making yoghurt and cottage or other cheeses resulting in economic losses to raw milk processors.
In developed countries where manual milking is not common anymore, standards of hygiene and laws on milk pasteurization are enforced. In many developing countries in spite of the existence of regulations that require milk pasteurization, a high percentage of the milk is sold raw through informal channels where hygienic measures during milking and distribution are not implemented (Omore et al. 1999). The informal milk markets are believed to be thriving in some of these countries because they provide social and economic benefits to smallholder producers, small market agents and consumers in terms of farm-gate prices, creation of jobs and competitive consumer prices. Cow milk and other dairy products may harbour a variety of microorganisms and can be important sources of human infections.
The quality of milk sold through the informal market in the coastal savannah region of Ghana indicates that farm hygiene and standards of milk production need to be improved. The unacceptable levels of contamination with microorganisms and antibiotic residues poses important human health risk. Consumer practices, such as boiling, to reduce or eliminate potential infection by milk-borne zoonoses should be encouraged.
This research was funded by the Ghanaian–Dutch collaboration for health research and development (Project ID number 2005/OD/08). The authors would like to thank the following institutions; The Animal Research Institute (ARI), the Veterinary Services Directorate (VSD) and the Ministry of Food and Agriculture (MOFA) of the study districts for the support of their staff during the sample collection. Alhaji Maama of the ARI is acknowledged for his immense support during the sample collection. Finally gratitude is extended to all farmers and stakeholders in the small scale diary business whose cooperation contributed to the success of this work.