Advertisement

MILK Symposium review: Microbiological quality and safety of milk from farm to milk collection centers in Rwanda*

      ABSTRACT

      The aim of this study was to generate knowledge on the most important milk quality and safety attributes, including somatic cell count (SCC), total bacterial count (TBC), Escherichia coli, Salmonella, and Brucella spp. antibodies and antibiotic residues in milk in the chain from farm to milk collection center (MCC) in Rwanda. In addition, we investigated farm and management factors associated with high TBC, SCC, and Salmonella counts. Raw milk was sampled at the farm and MCC levels. Milk samples were taken from dairy farms linked to 2 selected MCC in each of the 4 provinces in Rwanda. In total, 406 bulk milk samples from 406 farms and 32 bulk milk samples from 8 MCC were collected and analyzed. Farm milk average SCC varied between 180 × 103 and 920 × 103 cells/mL, whereas average SCC in milk samples at MCC varied between 170 × 103 and 1,700 × 103 cells/mL. The mean milk TBC of different farms per MCC varied between 1.1 × 106 and 1.6 × 107 cfu/mL, whereas in milk samples from different MCC, the mean TBC ranged between 5.3 × 105 and 2.4 × 108 cfu/mL. The high TBC in milk from MCC suggests proliferation or recontamination of milk by bacteria during transportation. Escherichia coli was detected in 35 of 385 farm milk samples and ranged between 5 cfu/mL and 1.1 × 104 cfu/mL, whereas in milk samples from the MCC, it was detected in 20 out 32 samples varying between 5 cfu/mL and 2.9 × 103 cfu/mL. Overall farm prevalence of Salmonella in milk samples was 14%, but no milk samples from MCC were positive for Salmonella. Five out of 22 bulk milk samples from different MCC were positive for Brucella spp. antibodies, but no Brucella antibodies were detected in milk samples from farms. The prevalence of antibiotic residues as detected by the Delvotest SP NT (DSM, Delft, the Netherlands) was low: 1.3% in farm milk samples and undetected in MCC milk samples. Lack of a separate milking area was associated with high TBC, whereas offering of supplemental feeds, keeping data of past diseases, and an unhygienic milking area were associated with high SCC. Lack of teat washing before milking was the only factor associated with Salmonella contamination of milk at the farm level. This study indicated high TBC and SCC of milk samples at the farm and MCC levels, which indicates both microbial contamination of milk and poor udder health in dairy cows. Presence of E. coli, Salmonella, and Brucella antibodies in milk was common, but finding antibiotic residues in milk was uncommon.

      Key words

      INTRODUCTION

      Raw milk from dairy cows may be contaminated by microorganisms originating from the udder (mastitis associated), by zoonotic pathogens shed from infected animals, or by other microorganisms from the environment. Environmental organisms could be transferred to the milk through poor hygiene of udder and teat surfaces and from uncleaned and unsanitized milking equipment (
      • Elmoslemany A.M.
      • Keefe G.
      • Dohoo I.
      • Jayarao B.
      Risk factors for bacteriological quality of bulk tank milk in Prince Edward Island dairy herds. Part 2: Bacteria count-specific risk factors.
      ), but also from milkers or other people handling the milk. Improper cooling of milk during transport can also influence bacterial count by increasing the rate of bacterial growth before the milk reaches milk collection centers (MCC) or processors. The total bacterial count (TBC) is used to evaluate the extent to which such processes have affected milk quality or safety. However,
      • Murphy S.C.
      • Boor K.J.
      Trouble-shooting sources and causes of high bacteria counts in raw milk.
      indicated that TBC should be interpreted with caution because different types of bacteria can contaminate milk from various sources such as equipment, milk handlers, and different environmental niches. These microorganisms proliferate in milk because milk contains key nutrients and has high water activity and an ideal pH for their growth and development (
      • Hassan A.N.
      • Frank J.F.
      Microorganisms associated with milk.
      ). Numerous groups of bacteria can grow in milk, but Escherichia coli is particularly used as an indicator organism for fecal contamination of foodstuff (i.e., an indicator of hygiene) and it can be associated with foodborne outbreaks (
      • Tryland I.
      • Fiksdal L.
      Enzyme characteristics of beta-D-galactosidase-negative and beta-D-glucuronidase-positive bacteria and their interference in rapid methods for detection of waterborne coliforms and Escherichia coli..
      ). The SCC in milk may be related to the immune reaction following an IMI. Subclinical mastitis is a situation in which leukocytes increase in milk without apparent visual changes in milk appearance, whereas in clinical mastitis, there are apparent changes in milk, sometimes in combination with local signs in the udder or systemic clinical signs that can be recognized by the farmer (
      • Hillerton J.E.
      • Berry E.A.
      Treating mastitis in the cow—A tradition or an archaism.
      ). A high leukocyte level, measured as SCC, and high TBC in milk may result in the production of enzymes that degrade milk components such as fats and proteins (
      • Li N.
      • Richoux R.
      • Boutinaud M.
      • Martin P.
      • Gagnaire V.
      Role of somatic cells on dairy processes and products: A review.
      ;
      • Baur C.
      • Krewinkel M.
      • Kranz B.
      • von Neubeck M.
      • Wenning M.
      • Scherer S.
      • Stoeckel M.
      • Hinrichs J.
      • Stressler T.
      • Fischer L.
      Quantification of the proteolytic and lipolytic activity of microorganisms isolated from raw milk.
      ), thus reducing the quality of milk and milk products. This will affect the shelf life and reduces consumer acceptance of these products (
      • Elmoslemany A.M.
      • Keefe G.
      • Dohoo I.
      • Jayarao B.
      Risk factors for bacteriological quality of bulk tank milk in Prince Edward Island dairy herds. Part 2: Bacteria count-specific risk factors.
      ). Moreover, mastitis bacteria such as Staphylococcus aureus and Streptococcus agalactiae can contaminate bulk milk and be a public health concern because they are zoonotic pathogens (
      • Zadoks R.N.
      • Middleton J.
      • McDougall S.
      • Katholm J.
      • Schukken Y.
      Molecular epidemiology of mastitis pathogens of dairy cattle and comparative relevance to humans.
      ;
      • Bi Y.
      • Wang Y.J.
      • Qin Y.
      • Vallverdú R.G.
      • García J.M.
      • Sun W.
      • Cao Z.
      Prevalence of bovine mastitis pathogens in bulk tank milk in China.
      ). Several other zoonotic pathogens, including Brucella spp. and Salmonella spp., may be found in infected animals and contaminate raw milk when milking techniques, hygiene, and handling during transportation are suboptimal at the farm or in the milk chain (
      • Kamana O.
      • Ceuppens S.
      • Jacxsens L.
      • Kimonyo A.
      • Uyttendaele M.
      Microbiological quality and safety assessment of the Rwandan milk and dairy chain.
      ;
      • Habarugira G.
      • Rukelibuga J.
      • Nanyingi M.O.
      • Mushonga B.
      Bovine tuberculosis in Rwanda: Prevalence and economic impact evaluation by meat inspection at Société des Abattoirs de Nyabugogo-Nyabugogo Abattoir, Kigali.
      ;
      • Rujeni N.
      • Mbanzamihigo L.
      Prevalence of brucellosis among women presenting with abortion/stillbirth in Huye, Rwanda.
      ).
      Increasing milk quality and safety around the world is highly relevant because regulations that protect the health of consumers require adherence to key milk quality and safety guidelines such as low SCC. The maximum concentration of SCC allowed for commingled bulk milk destined for processing and for human consumption differs by region. For example, the
      • European Union
      Regulation No 853/2004 of the European Parliament and of the Council of 29. Laying down specific hygiene rules for food of animal origin.
      requires an SCC limit of bulk milk of 400 × 103 cells/mL, the United States has a limit of 750 × 103 cells/mL and Canada has a limit of 500 × 103 cells/mL (
      • Schukken Y.H.
      • Wilson D.
      • Welcome F.
      • Garrison-Tikofsky L.
      • Gonzalez R.
      Monitoring udder health and milk quality using somatic cell counts.
      ). The East African standard for SCC is 300 × 103 cells/mL (EAS 67:2006;
      • East African Community
      EAS 67:2006: Raw cow milk–Specification.
      ) although this is not generally enforced. Although payment for milk volume is widely practiced in Rwanda, there are increasing calls for differentiated milk payment according to milk quality or safety because processors and consumers are paying more attention to quality and safety of milk and milk products. The use of antibiotics in food-producing animals has resulted in practical and cost-effective ways to control disease and improve animal welfare (
      • Hillerton J.E.
      • Berry E.A.
      Treating mastitis in the cow—A tradition or an archaism.
      ). On dairy farms, antibiotics are used for therapeutic purpose; for example, to treat mastitis, metritis, respiratory disease, and foot disease, and for prophylactic purposes; for example, for blanket dry-cow therapy and medicated milk replacer for calves (
      • Redding L.E.
      • Bender J.
      • Baker L.
      Quantification of antibiotic use on dairy farms in Pennsylvania.
      ). However, overuse or misuse of antibiotics can increase the risk of antibiotic residues in milk and contributes to the rise or selection of microorganisms that are resistant to antibiotics (
      • Yan S.S.
      • Gilbert J.
      Antimicrobial drug delivery in food animals and microbial food safety concerns: An overview of in vitro and in vivo factors potentially affecting the animal gut microflora.
      ).
      In Rwanda, milk is typically produced by smallholders and is generally transported, with unreliable refrigeration, using bicycles or motorcycles to MCC; individual large-scale farmers may also supply milk directly to the MCC. There are about 100 MCC functioning in Rwanda (
      • IFAD
      Rwanda dairy development project (RDDP) detailed design report.
      ). Hand milking is widely practiced, and smallholders are characterized by low productivity, insufficient use of modern farm technologies and practices, and challenges in accessing clean water and adequate training (
      • Doyle M.M.
      • Garcia S.
      • Bahati E.
      • Karamuzi D.
      • Cullor J.S.
      • Nandi S.
      Microbiological analysis of raw milk in Rwanda.
      ;
      • IFAD
      Rwanda dairy development project (RDDP) detailed design report.
      ). Milk collection centers serve as centralized cooling and storage centers for milk from many producers before the milk is forwarded to kiosks selling fresh milk or to factories for processing (
      • Miklyaev M.
      • Shahryar A.
      • Melani S.
      Cost-benefit analysis of Rwanda's dairy value chains. Development Discussion Papers. 2017–02, JDI Executive Programs. EconPapers.
      ). Milk quality and safety testing is rare on farms in Rwanda; however, at the MCC, milk is typically tested for acidity and added water using an alcohol testing and a lactometer, respectively. Raw milk quality and safety in the chain from farm to MCC in Rwanda is important for both processors and consumers, and collecting basic data on key quality attributes is vital for problem-solving regarding farm hygiene and sanitization, mastitis control, and milk collecting hygiene. The aim of this study was to generate information on the most important milk quality attributes, including SCC, TBC, E. coli, Salmonella, and Brucella antibodies, as well as antibiotic residues in the farm-to-MCC milk chain. In addition, potential risk factors associated with TBC, SCC, and Salmonella were investigated. The knowledge generated in this project will be used to develop milk quality improvement programs for the dairy sector in Rwanda.

      MATERIALS AND METHODS

      Study Areas

      The study was conducted in 8 selected MCC, 2 from each of the 4 provinces in Rwanda, and with the dairy farmers associated with these MCC. The MCC were chosen to represent potential differences in agroecology conditions, milk handling practices, and cultures that are specific to each province. The MCC were located at the following sites: MCC1 and MCC2 were located in Rwamagana and Nyagatare in the eastern province, MCC3 and MCC4 in Nyankenke and Rubaya in the northern province, MCC5 and MCC6 in Mudende and Rubengera in the western province, and MCC7 and MCC8 in Rugobagoba and Muyira in the southern province. Inclusion of each MCC was based on a mean receiving capacity of at least 4,000 L of milk per day. Because we could not obtain an official list of all dairy farmers associated with each MCC, the linear snowball sampling method, as described by
      • Balinas M.E.
      A program evaluation of a Rwandan milk collection center.
      and
      • Etikan I.
      • Rukayya A.
      • Sulaiman A.
      Comparison of snowball sampling and sequential sampling technique.
      , was used. The MCC technicians and milk transporters guided the research team to enlist farmers located in all provinces (east, south, west, and north) relative to the MCC. Dairy farmers included in the study per MCC corresponded to farmers whose lactating cows were previously screened for subclinical mastitis and described in a study by
      • Ndahetuye J.B.
      • Twambazimana J.
      • Nyman A.-K.
      • Karege C.
      • Tukei M.
      • Ongol M.P.
      • Persson Y.
      • Båge R.
      A cross sectional study of prevalence and risk factors associated with subclinical mastitis and intramammary infections, in dairy herds linked to milk collection centers in Rwanda.
      . Based on these estimations, the number of dairy farmers included in the study were 50 in MCC1, 14 in MCC2, 64 in MCC3, 55 in MCC4, 58 in MCC5, 56 in MCC6, 45 in MCC7, and 64 in MCC8.

      Milk Sample Collection

      The first sampling was done at the farm level, and each farm was sampled once from May to September 2017. The second sampling was done at the level of the MCC where farmers delivered their milk. Each MCC was sampled on 4 occasions, approximately every 4 mo in total, spanning 16 mo during 2017 and 2018. Aseptic collection of milk samples at the farm and MCC levels was done according to National Mastitis Council (
      • NMC (National Mastitis Council)
      Laboratory Handbook on Bovine Mastitis.
      ) guidelines. Before milk collection, milk in the bulk tank at the MCC or in bulking containers on farms was agitated for 10 min and samples were collected from the top of the bulk tank using a clean, sanitized dipper, transferred to sterile test tubes, and then placed in an ice-cooled box for immediate transport to the microbiology laboratory at the University of Rwanda, College of Agriculture, Animal Sciences and Veterinary Medicine, Busogo Campus, Rwanda. Somatic cell count was analyzed in fresh milk within 24 h after collection. The remaining milk was stored at −20°C until further analyzed.

      Somatic Cell Count

      The SCC was determined in milk samples from farms (n = 393) and from MCC (n = 32) using an electronic portable somatic cell counter (DeLaval Cell Counter, DCC, DeLaval, Sweden). A cut-off level of 300 × 103 cells/mL was used to compare levels of SCC, representing the standard used in East African region (
      • East African Community
      EAS 67:2006: Raw cow milk–Specification.
      ; EAS 67:2006).

      Total Bacterial Count

      To determine TBC in farm milk samples (n = 386) and MCC milk samples (n = 32), 1 mL of the milk sample was mixed with 9 mL of diluent (sterilized peptone physiological saline solution) and the mixture vortexed thoroughly. Then, serial dilutions (10−1 to 10−9) were prepared. From each dilution starting from the highest, 0.1 mL of test sample was inoculated onto plate count agar (Titan Biotech Ltd., Rajasthan, India) plates in duplicate. The sample was spread evenly on the surface of the plate using a sterile spreading glass rod. Samples were incubated at 37°C for 24 h. At the end of the incubation period, plates with between 30 and 300 colonies were counted. The number of colony-forming units was then converted, considering the dilution factor and the plated sample volume, into colony-forming units per milliliter of raw milk.

      Escherichia coli

      Enumeration of β-glucuronidase-positive E. coli in bulk milk samples from farm (n = 385) and samples from MCC (n = 32) was performed according to
      • ISO (International Organization for Standardization)
      ISO 16649-1:2001: Microbiology of food and animal feeding stuffs—Horizontal method for the enumeration of beta-glucuronidase-positive Escherichia coli—Part 1: Colony-count technique at 44 degrees C using membranes and 5-bromo-4-chloro-3-indolyl beta-D-glucuronide.
      ; 16649-1:2001). The milk sample (100 µL) was inoculated directly onto tryptone bile x-glucuronide (TBX) medium (BioMérieux, Marcy l'Etoile, France) plates in duplicate and spread evenly. Plates were incubated at 44°C for 24 h. At the end of incubation period, plates with between 30 and 300 colonies were counted.

      Salmonella

      The ISO 6579:2002-A1 2007 method (
      • ISO (International Organization for Standardization)
      SO 6579:2002/AMD 1:2007 Microbiology of food and animal feeding stuffs—Horizontal method for the detection of Salmonella spp.—Amendment 1: Annex D: Detection of Salmonella spp. in animal faeces and in environmental samples from the primary production stage.
      ) was followed to detect Salmonella in milk samples from the farm (n = 313) and MCC samples (n = 22). For each sample, 4.1 mL of each milk sample was added to 9 mL of peptone water (BiolaZrt, Budapest, Hungary) and the mixture was incubated at 37°C for 24 h for pre-enrichment. Then, 0.1 mL of suspension was added to 10 mL of modified semisolid Rappaport-Vassiliadis agar (Oxoid, Basingstoke, UK) and the mixture was incubated at 41.5°C for 48 h. Suspected Salmonella colonies were subcultured on xylose lysine deoxycholate (BiolaZrt). Final verification of Salmonella was done using the Oxoid Salmonella Latex Test (Oxoid) following the manufacturer's instructions.

      Brucella spp. Antibodies

      Antibodies to Brucella abortus and Brucella melitensis were analyzed by ELISA (Svanovir Brucella-Ab, Boehringer Ingelheim, Uppsala, Sweden) in milk samples from farms (n = 313) and samples from MCC (n = 22). Test kit specificity for milk samples was reported by the manufacturer to be 99 to 100%. Relative test kit sensitivity for the Rose Bengal test is 89.6% and that for the complement fixation test is 100% (). Milk samples were thawed at room temperature, and ELISA was performed according to the manufacturer's protocol for milk samples. On each ELISA plate, positive and negative control sera were included to ensure accuracy of the test, and all samples and controls were run in duplicate. Skanlit Software for Thermo Scientific Multiskan FC (Thermo Scientific, Ratastie, Finland) was used to read the ELISA plates and to calculate sample optical density (OD) values. Percent positivity (PP) was calculated as (OD of sample or negative control/OD of positive control) × 100. A milk sample with PP ≥10% was considered positive according to the manufacturer's instructions.

      Antibiotic Residues

      The prevalence of antibiotics, as detected by Delvotest SP NT kit (DSM, Heerlen, the Netherlands), was evaluated in milk by incubating 100 μL of homogenized milk sample for 2 to 3 h at 64°C and observing a color change of the lower two-thirds of the test panel to yellow (negative test) or completely purple for positive. According to the manufacturer, this test can detect more than 40 antibiotics. The kit has been previously validated and its sensitivity were found to be 1.5 ng/g for penicillin G, 2.5 ng/g for amoxicillin, 3.0 ng/g for ampicillin, and 5.8 ng/g for cephapirin (
      • Hennart S.L.A.
      • Faragher J.
      Validation of the Delvotest ® SP NT.
      ). In total, 372 and 32 milk samples from farms and MCC, respectively, were tested for antibiotic residues.

      Questionnaire

      Data collection and observations on dairy husbandry practices at the farms were done by the research team using a semi-structured questionnaire. Milking practices, housing, and hygiene routines were recorded. Variables included in the questionnaire are presented in Table 1. The hygiene concepts referred to in Table 1 (e.g., good/poor, slightly dirty/very dirty) were taken from mastitis studies such as
      • Schreiner D.A.
      • Ruegg P.
      Relationship between udder and leg hygiene scores and subclinical mastitis.
      or
      • Abrahmsén M.
      • Persson Y.
      • Kanyima B.M.
      • Båge R.
      Prevalence of subclinical mastitis in dairy farms in urban and peri-urban areas of Kampala, Uganda.
      and modified for our study. To classify farm environment as having good or poor hygiene or to describe milking as slightly or very dirty was based on whether these environment were visually completely free of dirt (i.e., good hygiene or clean), partially loaded with dirt (i.e., slightly dirty), or full of dirt (i.e., poor hygiene or very dirty). Data collectors were trained to ensure consistent scoring of hygiene. The interviews were conducted after milking.
      Table 1Factors analyzed at the farm level (n = 406) in 4 regions in Rwanda
      VariableCategory
      Type of cattle kraalIndividual, grouped, or no kraal
      Type of floor of cow housingConcrete, earthen, or raised wood
      Type of bedding materialsSawdust, grass, or none
      Wet beddingYes or no
      Frequency of bedding material replacementOnce a week or twice a week
      Grazing typeZero grazing, semi-grazing, or free grazing
      Separate calving area; separate milking areaYes or no
      Farm hygieneGood or poor
      Milking area hygieneClean, slightly dirty, or very dirty
      Frequency of cleaning milking areaBefore every milking; once per day; once, twice, or thrice per week; other
      Technique of milkingStripping or full hand
      Milking frequencyOnce or twice daily
      Who milks the cowOwner, worker, or child
      Hand washing before milkingWith water only, with water and soap, or no hand washing
      Teat and udder washing before milking; teat and udder drying; use of clean towel for dryingYes or no
      Premilking teat dipping; postmilking teat dippingYes or no
      Foremilk stripping; performing California Mastitis Test regularly; milking mastitic cows last; culling chronically infected cowsYes or no
      Feed cows after milkingYes or no
      Feeds sometimes concentratesYes or no
      Knowledge of clinical/subclinical mastitisYes or no
      Dry-cow therapyYes or no
      Availability of veterinary service; fly control; data record of past diseasesYes or no

      Data Analysis

      Prevalence of E. coli, Salmonella, or Brucella spp. antibodies was calculated as the number of positive samples against the total number of samples analyzed at the farm and MCC levels, respectively. Mean and median of TBC and SCC of data from farms within MCC and in different MCC were calculated and tabulated accordingly. The TBC and SCC were transformed on a log10 basis to achieve a normal distribution before analysis. Thereafter, associations between TBC or SCC and potential risk factors were analyzed by linear regression analysis as follows. To evaluate on-farm risk factors associated with TBC or SCC, unconditional associations between each independent variable and the dependent variable, first with TBC, and subsequently in a separate analysis with SCC, were investigated using univariable linear or univariable mixed-effect linear regression analysis, including MCC as random factor. Statistical significance in this step was assessed at P < 0.20. Factors that were significant in the univariable analyses were then investigated using Spearman rank correlation to assess collinearity; if 2 variables showed high collinearity (r ≥ 0.70), the one with the lowest P-value was then offered to the multivariable regression models. If MCC as a random factor was not significant (P ≥ 0.05), an ordinary linear regression model was used. The multivariable models were reduced using a manual, stepwise backward variable selection procedure where the initial model included all independent variables (with P-value <0.20 in the univariable analysis) as main effects. Variables with a significant association (P ≤ 0.05) with the dependent variable were kept in their respective final models. In each model, all variables with P ≤ 0.20 for TBC and SCC in the univariable analyses were then retested one at a time in their respective final model and kept in the model if they were significantly associated with the dependent variable. In parallel, confounding was checked if removal of a variable in final multivariable models changed the regression coefficients of the remaining variables (>25%). All plausible 2-way interactions between the significant main effects were tested in all final models. Model fit was assessed by determination of multiple correlation coefficient (R) and coefficient of determination (R2). Risk factors associated with Salmonella in bulk milk were analyzed in similar manner but using logistic regression models. The statistical analyses were performed using Stata 15 (Stata Corp LLC, College Station, TX).

      RESULTS

      Somatic Cell Counts

      The average milk SCC of farms varied between 180 × 103 and 920 × 103 cells/mL, whereas average milk SCC in all MCC varied between 170 × 103 and 1,700 × 103 cells/mL. The median SCC of milk at the farm level varied between 85 × 103 and 760 × 103 cells/mL, whereas that at the MCC level varied between 105 × 103 and 1,091 × 103 cells/mL (Table 2). The results of the final multivariable mixed-effect linear regression analysis showed that feeding concentrates, keeping records of past diseases, and unhygienic milking area were associated with a high SCC in milk at the farm level (Table 3).
      Table 2Somatic cell counts (×103 cells/mL) of bulk milk from farms (n = 406) and milk collection centers (MCC; n = 8) in 4 provinces in Rwanda
      MCCSCC at farm levelSCC at MCC level
      MeanSDMedianQ1
      Q1, Q3 = first and third quartiles, respectively.
      Q3
      Q1, Q3 = first and third quartiles, respectively.
      MeanSDMedianQ1Q3
      134047017062.5503480190485305652
      2440350270213604450210406285670
      34305301901075471,7001,8001,0915403,692
      49201,1007602201,274450360437124802
      536041019068.55296803606183721,057
      6180270853118217019010536.5373
      735038017052.5619450140433326604
      836050015047327350140304260497
      1 Q1, Q3 = first and third quartiles, respectively.
      Table 3On-farm factors associated with SCC in bulk milk from farms (n = 406) in 4 regions in Rwanda
      FactorRegression coefficientSEP-value95% CI
      Sometimes feeds concentrates
       NoReferent
       Yes0.220.080.0070.38–0.59
      Data record of past diseases
       NoReferent
       Yes0.320.140.020.58–0.05
      Milking area hygiene0.001
       CleanReferent
       Slightly clean0.260.080.0010.11–0.42
       Very dirty0.300.090.0010.11–0.48
      Intercept5.540.15<0.0015.24–5.83

      Total Bacterial Count

      Results of the TBC analyses are found in Table 4. Average TBC in farm milk varied between 1.1 × 106 and 1.6 × 107 cfu/mL, whereas average TBC of milk at the MCC varied between 5.3 × 105 and 2.4 × 108 cfu/mL. The farm milk median TBC varied between 7 × 103 and 1.1 × 106 cfu/mL, whereas that of milk at MCC varied between 2.5 × 105 and 1.4 × 108 cfu/mL (Table 4). The variable “lack of separate milking area” was significantly (P < 0.05) associated with higher TBC levels at farm level. The TBC was 0.49 cfu/mL higher (95% CI = 0.15–0.88, P = 0.005) in milk samples from farms without a separate milking area than in those from farms with a separate milking area.
      Table 4Total bacterial count (TBC; ×104 cfu/mL) of milk from farms (n = 406) and milk collection centers (MCC; n = 8) in 4 provinces in Rwanda
      MCCTBC at farm levelTBC at MCC level
      MeanSDMedianQ1
      Q1, Q3 = first and third quartiles, respectively.
      Q3
      Q1, Q3 = first and third quartiles, respectively.
      MeanSDMedianQ1Q3
      13206501102644325030017417579
      21405500.70.094.35320.32510136
      31,6004,900867.58887,20013,0004163421,190
      41604603521432,0003,800157495,879
      5110260101.8425,8006,1006,07826511,183
      61504607.82.3472,8005,200341488191
      72008409.41.83124,0003,50014,23023462,921
      8450180102.788.46,10011,0005345317,863
      1 Q1, Q3 = first and third quartiles, respectively.

      Escherichia coli and Salmonella

      Escherichia coli was detected in 8.5% of farm milk samples (range: 5.0 cfu/mL to 1.2 × 104 cfu/mL) and in 63% (20/32 samples) from MCC milk samples (range: 5.0 cfu/mL to 2.9 × 103 cfu/mL). Overall, Salmonella prevalence in farm milk samples was 14.0%. No Salmonella were detected in milk samples from MCC. The only on-farm factor remaining after the multivariable mixed-effect linear regression analysis was “lack of teat washing before milking.” Farms that did not wash cows' teats before milking had a significantly higher odds of also having a higher level of Salmonella in milk samples (odds ratio = 2.22, 95% CI = 1.13–4.36, P = 0.02).

      Brucella Antibodies in Milk

      No Brucella antibodies were detected in farm bulk milk samples. Five of 22 bulk milk samples from different MCC were positive for Brucella spp. antibodies. The positive samples came from the 2 MCC in the eastern province: MCC2 was positive on 2 occasions and MCC3 was positive on 3 occasions.

      Antibiotic Residues in Milk

      Antibiotic residues were found in 5 of 372 screened farm bulk milk samples as detected by Delvotest SP NT, yielding a prevalence of 1.3%. No antibiotic residues were detected in MCC milk samples.

      DISCUSSION

      The milk chain from farm to MCC is the cornerstone of the formal dairy market in Rwanda. This study showed milk to have high SCC and TBC and to be contaminated with E. coli.

      SCC at the Farm and MCC Levels

      The MCC included in the study did not regularly screen milk for SCC and were therefore unable to enforce the SCC standard for threshold limits, whether it concerned requirements for acceptance or rejection or payment incentives (e.g., premium payment for a high-quality product or penalty for low-quality product). Milk samples from farms and MCC (7/8 MCC) had average SCC >300 × 103 cells/mL, which is the limit for raw milk set by the
      • East African Community
      EAS 67:2006: Raw cow milk–Specification.
      ; EAS 67:2006). This standard is stricter than those in the European Union and the United States, likely because it was adopted directly from ISO 13366 (
      • ISO (International Organization for Standardization)
      ISO 13366-2:2006 [IDF 148-2:2006] Milk — Enumeration of somatic cells — Part 2: Guidance on the operation of fluoro-opto-electronic counters.
      ) without consideration of local conditions. The high SCC levels in milk indicate udder health problems in the cows. We observed considerable variation between the lowest and the highest recorded SCC in milk from farms: the lowest recorded SCC was 2 × 103 cells/mL and the highest was 7,900 × 103 cells/mL. This considerable variation demonstrates the difficulty in setting and complying with a relevant threshold for milk acceptance or rejection or for quality compensation. Our results showed that 36% of the farms had a bulk milk SCC >300 × 103 cells/mL, which is lower than that found in a study of smallholder farms in Lusaka, Zambia, where 61.4% of the milk samples had SCC above the recommended limit of 300 × 103 cells/mL (
      • Kunda B.
      • Pandey G.S.
      • Muma J.B.
      Somatic cell count and antibiotic residues in raw milk produced by smallholder dairy farmers in Lusaka province of Zambia.
      ). To give good advice on how to lower the bulk milk SCC at the farm level, an understanding of the factors that affect bulk milk SCC is needed. Our results showed that the variables “sometimes feeding concentrates,” “keeping records of diseases,” and “unhygienic milking area” were associated with high bulk milk SCC. Improvement of these factors could result in lowering bulk milk SCC. It is not clear why feeding concentrate was associated with high bulk milk SCC. It could be that it is more common to feed concentrates to high-yielding cows, and these cows are more commonly Holsteins, a breed shown to have a higher risk of mastitis in Rwanda (
      • Ndahetuye J.B.
      • Persson Y.
      • Nyman A.
      • Tukei M.
      • Ongol M.
      • Båge R.
      Aetiology and prevalence of subclinical mastitis in dairy herds in peri-urban areas of Kigali in Rwanda.
      ). It is not known whether feed manufacturers in Rwanda add selenium and vitamin A and E to feeds during feed formulation; these supplements are known to minimize mastitis incidence in dairy cows (
      • Sandholm M.
      • Honkanen-Buzalski T.
      • Kaartinen L.
      • Pyörälä S.
      The Bovine Udder and Mastitis.
      ). Similarly, it is not clear why keeping records was associated with higher SCC. It is possible that farmers who keep records are those who have recently experienced mastitis in their farms and therefore want to keep records on the cases. An explanation to why farms with cleaner milking areas in this study had lower bulk milk SCC is that good hygienic conditions prevent and reduce transmission of mastitis bacteria from one cow to another (
      • Philpot W.N.
      Control of mastitis by hygiene and therapy.
      ). The latter author stated that if transmission of mastitis pathogens is prevented by good hygiene, a parallel decrease in incidence of IMI will occur. By applying best practices, several of these issues can be mitigated or overcome. Bearing in mind that cattle owners, compared with farmers rearing other animal species, are more likely to adopt innovations, management technologies, and practices and that cattle are prioritized before other species in preventive health care and veterinary treatments (
      • Amadou H.
      • Dossa L.H.
      • Lompo D.J.P.
      • Abdulkadir A.
      • Schlecht E.
      A comparison between urban livestock production strategies in Burkina Faso, Mali and Nigeria in West Africa.
      ). Thus, the potential exists to increase and improve milk production and quality in Rwanda by inexpensive and simple means, such as application of the 10-point mastitis control plan (
      • Middleton J.R.
      • Saeman A.
      • Fox L.
      • Lombard J.
      • Hogan J.
      • Smith K.
      The National Mastitis Council: A global organization for mastitis control and milk quality, 50 years and beyond.
      ) and other best practices.

      TBC at the Farm and MCC Levels

      Except for 2 MCC, a higher TBC was detected in MCC milk samples than in farm milk samples. This suggests proliferation of bacteria in milk during transportation in unrefrigerated equipment. This agrees with
      • Doyle M.M.
      • Garcia S.
      • Bahati E.
      • Karamuzi D.
      • Cullor J.S.
      • Nandi S.
      Microbiological analysis of raw milk in Rwanda.
      , who detected an increase in total microbial load in the chain from farm through milk transporters to MCC and finally consumers in Rwanda. The same trend was reported in Uganda, where a 150-fold proliferation of bacteria occurred in milk from the farm level through transportation to consumers (
      • Grimaud P.
      • Sserunjogi M.
      • Grillet N.
      An evaluation of milk quality in Uganda: Value chain assessment and recommendations.
      ). In our study, TBC recorded in milk samples from farms were very high, suggesting that mixing such milk with milk of better quality at the MCC would increase the overall TBC of the milk at the MCC. Therefore, there is a need for infrastructure and equipment to separate out low-quality milk as early as possible in the milk chain, or to introduce economic incentives for farmers to produce and deliver milk with very low TBC. We speculate that the reason why 2 MCC did not experience an increase in TBC from farm to MCC was because the farmers were located close to the MCC and milk delivery took less time, allowing less opportunity for proliferation of microorganisms in the milk. The highest recorded TBC at the farm level (1.6 × 107 cfu/mL) was comparable to that reported in Zimbabwe (6.7 ± 5.8 log10 cfu/mL) in raw milk samples (
      • Mhone T.A.
      • Matope G.
      • Saidi P.
      Aerobic bacterial, coliform, Escherichia coli and Staphylococcus aureus counts of raw and processed milk from selected smallholder dairy farms of Zimbabwe.
      ), and comparable to the 7.08 log10 cfu/mL reported in milk samples from chilling centers in Sri Lanka (
      • De Silva S.
      • Kanugala K.
      • Weerakkody N.
      Microbiological quality of raw milk and effect on quality by implementing good management practices.
      ). The lowest median TBC (7 × 103 cfu/mL) was recorded in milk from MCC2, in Nyagatare, where farmers are known to have received more training on dairy husbandry and milk handling (). In this study, we found that the lack of a separate milking area (such that farmers milk in the same place where the cow is housed) was significantly associated with increased risk of contamination of milk with environmental microorganisms, reflected by high TBC. Hence, a recommendation that farmers do not milk cows in the same place where cows are housed would likely improve the hygienic quality of milk.

      Escherichia coli at the Farm and MCC Levels

      Detection of E. coli was less frequent at the farm level than at the MCC level, suggesting contamination during handling at MCC or proliferation of bacteria during transport to MCC. Potential routes of contamination at the MCC level include personnel, equipment, and tools, whereas contamination at the farm level may be due to animal feces or poor hygienic level of animal husbandry practices (
      • Kateřina B.
      • Marcela V.
      • Vladimír B.
      • Libor K.
      • Ivana K.
      • Renáta K.
      Microbiological quality of raw milk in the Czech Republic.
      ). Our results agree with those of
      • Grimaud P.
      • Sserunjogi M.
      • Grillet N.
      An evaluation of milk quality in Uganda: Value chain assessment and recommendations.
      , who reported high E. coli counts (2 × 106 cfu/mL) in raw milk samples at the farm level in Uganda.

      Prevalence of Salmonella at the Farm and MCC Levels

      The prevalence of Salmonella in milk from farms in this study was 14% but no MCC samples tested positive. It is possible that due to the dilution effect, Salmonella concentrations in MCC milk samples were below the detection limit of the method used. This prevalence at the farm level is higher than results from Rwanda reported by
      • Kamana O.
      • Ceuppens S.
      • Jacxsens L.
      • Kimonyo A.
      • Uyttendaele M.
      Microbiological quality and safety assessment of the Rwandan milk and dairy chain.
      , who found a prevalence of Salmonella of 5.2% in raw milk samples from dairy farms, MCC, and milk shops. Our results are in the range of those reported in a study in Tanzania, where a prevalence of 10.1% was found in raw milk (
      • Schoder D.
      • Maichin A.
      • Lema B.
      • Laffa J.
      Microbiological quality of milk in Tanzania: From Maasai stable to African consumer table.
      ). Farm environment is likely where reservoirs and vehicles for the Salmonella can be found (
      • Quintana Á.R.
      • Seseña S.
      • Garzón A.
      • Arias R.
      Factors affecting levels of airborne bacteria in dairy farms: A review.
      ). It is possible that the Salmonella found in milk originated from milkers' hands, which may have touched reservoirs of Salmonella such as infected calves, shedding cows, or contaminated water (
      • Marth E.H.
      Salmonellae and salmonellosis associated with milk and milk products. A review.
      ). Our study revealed that a lack of teat washing before milking was associated with Salmonella contamination of bulk milk. Because shedding of Salmonella is common in cattle (
      • Wells S.J.
      • Fedorka-Cray P.J.
      • Dargatz D.A.
      • Ferris K.
      • Green A.
      Fecal shedding of Salmonella spp. by dairy cows on farm and at cull cow markets.
      ), poor hygiene through lack of teat washing will facilitate the transmission of the pathogen from the cow to the milk.

      Prevalence of Brucella Antibodies at the Farm and MCC Levels

      No sample tested positive for Brucella antibodies among farm bulk milk samples, but antibodies were detected in 3 samples in MCC bulk milk. These antibodies may have come from farm milk that was not sampled because we did not visit all farmers associated with the MCC. It may also imply that some cows could be harboring the Brucella pathogen and zoonotically infecting themselves and humans in the region around the MCC. This type of transmission of brucellosis between animals and humans has been suggested in Uganda, where the presence of Brucella antibodies in humans was associated with Brucella antibodies in milk samples from cattle (
      • Miller R.
      • Nakavuma J.L.
      • Ssajjakambwe P.
      • Vudriko P.
      • Musisi N.
      • Kaneene J.B.
      The prevalence of brucellosis in cattle, goats and humans in rural Uganda: A comparative study.
      ). The level of detection of Brucella antibodies in milk at the MCC level in this study (22.72%) was markedly higher than the level (11%) reported in Gulu in Uganda (
      • Rock K.T.
      • Mugizi D.
      • Ståhl K.
      • Magnusson U.
      • Boqvist S.
      The milk delivery chain and presence of Brucella spp. antibodies in bulk milk in Uganda.
      ).

      Antibiotic Residues at the Farm and MCC Levels

      Detecting antibiotic residues in bulk milk was not common in the present study. Antibiotic residues were detected only in milk samples from farm delivering milk that had high SCC levels, suggesting that treating mastitis with antibiotics without observing the withholding period could explain the presence of antibiotic residues in the milk samples. The consequences of antibiotic residues in milk are severe; for example, antibiotic residues can prevent optimum growth of starter cultures during processing of dairy products, and β-lactam antibiotics, if present, can cause allergic reactions in some individuals (
      • Dewdney J.M.
      • Maes L.
      • Raynaud J.P.
      • Blanc F.
      • Scheid J.P.
      • Jackson T.
      • Lens S.
      • Verschueren C.
      Risk assessment of antibiotic residues of β-lactams and macrolides in food products with regard to their immuno-allergic potential.
      ;
      • Griffiths M.W.
      Improving the safety and quality of milk. Volume 1: Milk production and processing.
      ). Our results showed a markedly lower prevalence of antibiotic residues in farm bulk milk than has been reported by others: 44.5% reported in Kenya (
      • Teresiah W.
      • Patrick S.
      • Mary O.
      • Gerard O.
      • Anton J.
      Quality control of raw milk in the smallholder collection and bulking enterprises in Nakuru and Nyandarua counties, Kenya.
      ) in a study that used a kit similar to the one used here; 30% in Zambia (
      • Kunda B.
      • Pandey G.S.
      • Muma J.B.
      Somatic cell count and antibiotic residues in raw milk produced by smallholder dairy farmers in Lusaka province of Zambia.
      ), using the Copan milk test (Copan Italia spa, Brescia, Italy); and 36% reported in Tanzania (
      • Kurwijila L.R.
      • Omore A.
      • Staal S.
      • Mdoe N.
      Investigation of the risk of exposure to antimicrobial residues present in marketed milk in Tanzania.
      ), using the Charm AIM-96 (Charm Sciences Inc., Lawrence, MA) antimicrobial inhibition assay.

      CONCLUSIONS

      Milk delivered to MCC in Rwanda by farmers or intermediaries had high microbial contamination and SCC, which contributes to high TBC and SCC of milk at the MCC. Improved testing and separating low- from high-quality milk, followed by rejection of milk with high TBC and SCC upon receipt at the MCC is recommended. Overall, the increase in TBC from farm to MCC suggests bacterial proliferation during transport, emphasizing the need for refrigeration and proper handling during transport. Contamination of milk with E. coli seemed to be more frequent at the MCC level, suggesting that conditions were less hygienic at milk bulk collection sites. The 14% prevalence of Salmonella on dairy farms suggests that it is a key pathogen, and prevention and control measures are required to safeguard public health from the risk associated with consumption of Salmonella-contaminated milk. Antibiotic residues were rarely detected but Brucella spp. antibodies were common in milk samples from MCC.

      ACKNOWLEDGMENTS

      The authors acknowledge funding from the Swedish International Development Agency (SIDA; Stockholm, Sweden), within the University of Rwanda-Sweden program for research, higher education and institutional advancement, subprogram agricultural sciences, project no. 20290000. This work was also funded in whole or part by the United States Agency for International Development (USAID; Washington, DC) Bureau for Food Security under Agreement #AID-OAA-L-15-00003 as part of Feed the Future Innovation Lab for Livestock Systems (University of Florida, Gainesville). Travel to the meeting was funded by the American Dairy Science Association. Any opinions, findings, conclusions, or recommendations expressed here are those of the authors alone. The authors have stated that they have no conflicts of interest.

      REFERENCES

        • Abrahmsén M.
        • Persson Y.
        • Kanyima B.M.
        • Båge R.
        Prevalence of subclinical mastitis in dairy farms in urban and peri-urban areas of Kampala, Uganda.
        Trop. Anim. Health Prod. 2014; 46 (23955012): 99-105
        • Amadou H.
        • Dossa L.H.
        • Lompo D.J.P.
        • Abdulkadir A.
        • Schlecht E.
        A comparison between urban livestock production strategies in Burkina Faso, Mali and Nigeria in West Africa.
        Trop. Anim. Health Prod. 2012; 44 (22430479): 1631-1642
        • Balinas M.E.
        A program evaluation of a Rwandan milk collection center.
        (MS thesis.) Texas A&M University, College Station, TX2014
        • Baur C.
        • Krewinkel M.
        • Kranz B.
        • von Neubeck M.
        • Wenning M.
        • Scherer S.
        • Stoeckel M.
        • Hinrichs J.
        • Stressler T.
        • Fischer L.
        Quantification of the proteolytic and lipolytic activity of microorganisms isolated from raw milk.
        Int. Dairy J. 2015; 49: 23-29
        • Bi Y.
        • Wang Y.J.
        • Qin Y.
        • Vallverdú R.G.
        • García J.M.
        • Sun W.
        • Cao Z.
        Prevalence of bovine mastitis pathogens in bulk tank milk in China.
        PLoS One. 2016; 11 (27187065)e0155621
        • De Silva S.
        • Kanugala K.
        • Weerakkody N.
        Microbiological quality of raw milk and effect on quality by implementing good management practices.
        Procedia Food Sci. 2016; 6: 92-96
        • Dewdney J.M.
        • Maes L.
        • Raynaud J.P.
        • Blanc F.
        • Scheid J.P.
        • Jackson T.
        • Lens S.
        • Verschueren C.
        Risk assessment of antibiotic residues of β-lactams and macrolides in food products with regard to their immuno-allergic potential.
        Food Chem. Toxicol. 1991; 29 (1894215): 477-483
        • Doyle M.M.
        • Garcia S.
        • Bahati E.
        • Karamuzi D.
        • Cullor J.S.
        • Nandi S.
        Microbiological analysis of raw milk in Rwanda.
        Afr. J. Food Sci. 2015; 06: 141-143
        • East African Community
        EAS 67:2006: Raw cow milk–Specification.
        https://law.resource.org/pub/eac/ibr/eas.67.2006.html
        Date: 2006
        Date accessed: July 1, 2019
        • Elmoslemany A.M.
        • Keefe G.
        • Dohoo I.
        • Jayarao B.
        Risk factors for bacteriological quality of bulk tank milk in Prince Edward Island dairy herds. Part 2: Bacteria count-specific risk factors.
        J. Dairy Sci. 2009; 92 (19447997): 2644-2652
        • Etikan I.
        • Rukayya A.
        • Sulaiman A.
        Comparison of snowball sampling and sequential sampling technique.
        Biometric. Biostat. Int. J. 2016; 3: 1-2
        • European Union
        Regulation No 853/2004 of the European Parliament and of the Council of 29. Laying down specific hygiene rules for food of animal origin.
        Off. J. Eur. Union L. 2004; 226: 22-82
        • Griffiths M.W.
        Improving the safety and quality of milk. Volume 1: Milk production and processing.
        CRC Press, Boca Raton, FL2019
        • Grimaud P.
        • Sserunjogi M.
        • Grillet N.
        An evaluation of milk quality in Uganda: Value chain assessment and recommendations.
        Afr. J. Food Agric. Nutr. Dev. 2007; 7nd07045
        • Habarugira G.
        • Rukelibuga J.
        • Nanyingi M.O.
        • Mushonga B.
        Bovine tuberculosis in Rwanda: Prevalence and economic impact evaluation by meat inspection at Société des Abattoirs de Nyabugogo-Nyabugogo Abattoir, Kigali.
        J. S. Afr. Vet. Assoc. 2014; 85 (25685979)1062
        • Hassan A.N.
        • Frank J.F.
        Microorganisms associated with milk.
        in: Fuquay J.W. Fox P.F. McSweeney P.L.H. Encyclopedia of Dairy Sciences. 2nd ed. Academic Press, San Diego, CA2011: 447-457
        • Hennart S.L.A.
        • Faragher J.
        Validation of the Delvotest ® SP NT.
        J. AOAC Int. 2012; 95 (23451401): 1807-1815
        • Hillerton J.E.
        • Berry E.A.
        Treating mastitis in the cow—A tradition or an archaism.
        J. Appl. Microbiol. 2005; 98 (15916638): 1250-1255
        • IFAD
        Rwanda dairy development project (RDDP) detailed design report.
        • ISO (International Organization for Standardization)
        ISO 13366-2:2006 [IDF 148-2:2006] Milk — Enumeration of somatic cells — Part 2: Guidance on the operation of fluoro-opto-electronic counters.
        https://www.iso.org/standard/40260.html
        Date: 2006
        Date accessed: June 7, 2019
        • ISO (International Organization for Standardization)
        SO 6579:2002/AMD 1:2007 Microbiology of food and animal feeding stuffs—Horizontal method for the detection of Salmonella spp.—Amendment 1: Annex D: Detection of Salmonella spp. in animal faeces and in environmental samples from the primary production stage.
        https://www.iso.org/standard/42109.html
        Date: 2007
        Date accessed: June 7, 2019
        • ISO (International Organization for Standardization)
        ISO 16649-1:2001: Microbiology of food and animal feeding stuffs—Horizontal method for the enumeration of beta-glucuronidase-positive Escherichia coli—Part 1: Colony-count technique at 44 degrees C using membranes and 5-bromo-4-chloro-3-indolyl beta-D-glucuronide.
        https://www.iso.org/standard/29823.html
        Date: 2001
        Date accessed: June 7, 2019
        • Kamana O.
        • Ceuppens S.
        • Jacxsens L.
        • Kimonyo A.
        • Uyttendaele M.
        Microbiological quality and safety assessment of the Rwandan milk and dairy chain.
        J. Food Prot. 2014; 77 (24490925): 299-307
        • Kateřina B.
        • Marcela V.
        • Vladimír B.
        • Libor K.
        • Ivana K.
        • Renáta K.
        Microbiological quality of raw milk in the Czech Republic.
        Czech J. Food Sci. 2016; 34: 189-196
        • Kunda B.
        • Pandey G.S.
        • Muma J.B.
        Somatic cell count and antibiotic residues in raw milk produced by smallholder dairy farmers in Lusaka province of Zambia.
        Asian J. Multidisciplin. Stud. 2016; 4: 36-40
        • Kurwijila L.R.
        • Omore A.
        • Staal S.
        • Mdoe N.
        Investigation of the risk of exposure to antimicrobial residues present in marketed milk in Tanzania.
        J. Food Prot. 2006; 69 (17066932): 2487-2492
        • Li N.
        • Richoux R.
        • Boutinaud M.
        • Martin P.
        • Gagnaire V.
        Role of somatic cells on dairy processes and products: A review.
        Dairy Sci. Technol. 2014; 94 (25309683): 517-538
        • Marth E.H.
        Salmonellae and salmonellosis associated with milk and milk products. A review.
        J. Dairy Sci. 1969; 52 (4885967): 283-315
        • Mhone T.A.
        • Matope G.
        • Saidi P.
        Aerobic bacterial, coliform, Escherichia coli and Staphylococcus aureus counts of raw and processed milk from selected smallholder dairy farms of Zimbabwe.
        Int. J. Food Microbiol. 2011; 151 (21944662): 223-228
        • Middleton J.R.
        • Saeman A.
        • Fox L.
        • Lombard J.
        • Hogan J.
        • Smith K.
        The National Mastitis Council: A global organization for mastitis control and milk quality, 50 years and beyond.
        J. Mammary Gland Biol. Neoplasia. 2014; 19 (25524293): 241-251
        • Miklyaev M.
        • Shahryar A.
        • Melani S.
        Cost-benefit analysis of Rwanda's dairy value chains. Development Discussion Papers. 2017–02, JDI Executive Programs. EconPapers.
        Örebro University, School of Business, Örebro, Sweden2017
        • Miller R.
        • Nakavuma J.L.
        • Ssajjakambwe P.
        • Vudriko P.
        • Musisi N.
        • Kaneene J.B.
        The prevalence of brucellosis in cattle, goats and humans in rural Uganda: A comparative study.
        Transbound. Emerg. Dis. 2016; 63 (25660343): e197-e210
        • Murphy S.C.
        • Boor K.J.
        Trouble-shooting sources and causes of high bacteria counts in raw milk.
        Dairy Food Environ. Sanit. 2000; 20: 606-611
        • Ndahetuye J.B.
        • Persson Y.
        • Nyman A.
        • Tukei M.
        • Ongol M.
        • Båge R.
        Aetiology and prevalence of subclinical mastitis in dairy herds in peri-urban areas of Kigali in Rwanda.
        Trop. Anim. Health Prod. 2019; 51 (31030333): 2037-2044
        • Ndahetuye J.B.
        • Twambazimana J.
        • Nyman A.-K.
        • Karege C.
        • Tukei M.
        • Ongol M.P.
        • Persson Y.
        • Båge R.
        A cross sectional study of prevalence and risk factors associated with subclinical mastitis and intramammary infections, in dairy herds linked to milk collection centers in Rwanda.
        Prev. Vet. Med. 2020; 179 (32380364)105007
        • NMC (National Mastitis Council)
        Laboratory Handbook on Bovine Mastitis.
        Rev. ed. National Mastitis Council Inc., New Prague, MN2017
        • Philpot W.N.
        Control of mastitis by hygiene and therapy.
        J. Dairy Sci. 1979; 62 (379060): 168-176
        • Quintana Á.R.
        • Seseña S.
        • Garzón A.
        • Arias R.
        Factors affecting levels of airborne bacteria in dairy farms: A review.
        Animals (Basel). 2020; 10 (32245161): 526
        • Redding L.E.
        • Bender J.
        • Baker L.
        Quantification of antibiotic use on dairy farms in Pennsylvania.
        J. Dairy Sci. 2019; 102 (30594359): 1494-1507
        • Rock K.T.
        • Mugizi D.
        • Ståhl K.
        • Magnusson U.
        • Boqvist S.
        The milk delivery chain and presence of Brucella spp. antibodies in bulk milk in Uganda.
        Trop. Anim. Health Prod. 2016; 48 (27026231): 985-994
        • Rujeni N.
        • Mbanzamihigo L.
        Prevalence of brucellosis among women presenting with abortion/stillbirth in Huye, Rwanda.
        J. Trop. Med. 2014; 2014 (25101131)740479
        • Sandholm M.
        • Honkanen-Buzalski T.
        • Kaartinen L.
        • Pyörälä S.
        The Bovine Udder and Mastitis.
        Gummerus Kirjapaino Oy, Jyvaskyla, Finland1995
        • Schoder D.
        • Maichin A.
        • Lema B.
        • Laffa J.
        Microbiological quality of milk in Tanzania: From Maasai stable to African consumer table.
        J. Food Prot. 2013; 76 (24215695): 1908-1915
        • Schreiner D.A.
        • Ruegg P.
        Relationship between udder and leg hygiene scores and subclinical mastitis.
        J. Dairy Sci. 2003; 86 (14672175): 3460-3465
        • Schukken Y.H.
        • Wilson D.
        • Welcome F.
        • Garrison-Tikofsky L.
        • Gonzalez R.
        Monitoring udder health and milk quality using somatic cell counts.
        Vet. Res. 2003; 34 (14556696): 579-596
        • Svanova
        SVANOVIR Brucella-Ab I-ELISA—Summary of evaluation data.
        • TechnoServe Rwanda
        The dairy value chain in Rwanda.
        • Teresiah W.
        • Patrick S.
        • Mary O.
        • Gerard O.
        • Anton J.
        Quality control of raw milk in the smallholder collection and bulking enterprises in Nakuru and Nyandarua counties, Kenya.
        J. Food Sci. 2016; 10: 70-78
        • Tryland I.
        • Fiksdal L.
        Enzyme characteristics of beta-D-galactosidase-negative and beta-D-glucuronidase-positive bacteria and their interference in rapid methods for detection of waterborne coliforms and Escherichia coli..
        Appl. Environ. Microbiol. 1998; 64 (9501441): 1018-1023
        • Wells S.J.
        • Fedorka-Cray P.J.
        • Dargatz D.A.
        • Ferris K.
        • Green A.
        Fecal shedding of Salmonella spp. by dairy cows on farm and at cull cow markets.
        J. Food Prot. 2001; 64 (11198437): 3-11
        • Yan S.S.
        • Gilbert J.
        Antimicrobial drug delivery in food animals and microbial food safety concerns: An overview of in vitro and in vivo factors potentially affecting the animal gut microflora.
        Adv. Drug Deliv. Rev. 2004; 56 (15191795): 1497-1521
        • Zadoks R.N.
        • Middleton J.
        • McDougall S.
        • Katholm J.
        • Schukken Y.
        Molecular epidemiology of mastitis pathogens of dairy cattle and comparative relevance to humans.
        J. Mammary Gland Biol. Neoplasia. 2011; 16 (21968538): 357-372