Advertisement

Risk factors for bacteriological quality of bulk tank milk in Prince Edward Island dairy herds. Part 2: Bacteria count-specific risk factors

  • A.M. Elmoslemany
    Affiliations
    Department of Health Management, University of Prince Edward Island, Charlottetown, Prince Edward Island, C1A 4P3, Canada

    Faculty of Veterinary Medicine, Kafr El-Sheikh University, Egypt, PO Box 33516
    Search for articles by this author
  • G.P. Keefe
    Correspondence
    Corresponding author.
    Affiliations
    Department of Health Management, University of Prince Edward Island, Charlottetown, Prince Edward Island, C1A 4P3, Canada

    Maritime Quality Milk, University of Prince Edward Island, Charlottetown, Prince Edward Island, C1A 4P3, Canada
    Search for articles by this author
  • I.R. Dohoo
    Affiliations
    Department of Health Management, University of Prince Edward Island, Charlottetown, Prince Edward Island, C1A 4P3, Canada

    Centre for Veterinary Epidemiological Research, University of Prince Edward Island, Charlottetown, Prince Edward Island, C1A 4P3, Canada
    Search for articles by this author
  • B.M. Jayarao
    Affiliations
    Department of Veterinary and Biomedical Sciences, Pennsylvania State University, University Park 16802
    Search for articles by this author

      Abstract

      A case-control study was conducted to identify specific on-farm risk factors that influence bacteriological quality of bulk tank milk in Prince Edward Island dairy herds. Total aerobic (TAC), preliminary incubation (PIC), laboratory pasteurization (LPC), and coliform (CC) counts were used to assess the bacteriological quality of bulk tank milk. Four case-control groups were defined based on the last 6 results of each test before on farm evaluation. A herd was classified as a TAC, PIC, or CC case when the herd had at least 4 high TAC, PIC, or CC counts out of the last 6 analyses for each test, respectively. For the LPC case group, a herd was required to have at least 3 high results out of the last 6 analyses. Control groups had low counts in the last 6 analyses for each test in the corresponding case group (TAC, PIC, CC, and LPC). The results of the study showed that TAC and PIC were mainly associated with cow and stall hygiene: washing the teats with water, not using teat predip, and dirty teats were risk factors. The LPC and CC were related to equipment hygiene, with high counts being associated with low temperature of the cleaning solution, high water-hardness score, and high alkalinity of alkaline detergent wash. Based on the findings of this study it can be concluded that TAC, PIC, LPC, and CC counts are of considerable value in identifying practices that could influence milk quality.

      Key words

      Introduction

      The production of high-quality milk with low bacteriological counts begins at the farm. It involves multiple factors related to cow health and udder hygiene, hygiene of the milking environment in which the cows are housed and milked, and hygiene of the milking equipment. Microbial contamination of bulk tank milk (BTM) occurs through a variety of sources and by different types of microorganism. Therefore, using multiple bacterial tests that estimate specific groups of bacteria in BTM could provide a detailed picture on hygiene practices employed on the farm during collection, handling, and storage of BTM (
      • Murphy S.C.
      • Boor K.J.
      Trouble-shooting sources and causes of high bacteria counts in raw milk.
      ).
      In this study, 4 bacterial quality parameters were used: total aerobic count (TAC), preliminary incubation count (PIC), laboratory pasteurization count (LPC), and coliform count (CC). The TAC is an alternative to the standard plate count (SPC). It estimates the total number of aerobic bacteria in raw milk samples and is an important parameter in regulatory and quality incentive programs in many parts of the world. The TAC indicates the general hygienic conditions during milk production; therefore, it may be of less importance in identifying specific sources of contamination (
      • Chambers J.V.
      The microbiology of raw milk.
      ). The PIC quantifies psychrotrophic bacteria, which grow at inadequate refrigeration temperature. Psychrotrophs are generally found in untreated water, soil, and vegetation. They are introduced into the milk as a result of contamination of milking equipment or the exterior of the udder and teats from these sources (
      • Murphy S.C.
      Raw milk bacteria tests: Standard plate count, preliminary incubation count, lab, pasteurization count and coliform count. What do they mean for your farm?.
      ).
      The LPC estimates the number of thermoduric bacteria that survive laboratory-scale pasteurization procedures similar to batch pasteurization (62.8°C for 30 min). The main sources of contamination of milk by thermoduric bacteria are poorly cleaned and inadequately sanitized udders and equipment (
      • Murphy S.C.
      Raw milk bacteria tests: Standard plate count, preliminary incubation count, lab, pasteurization count and coliform count. What do they mean for your farm?.
      ). The CC enumerates coliform bacteria. Coliforms inhabit the intestinal tract of cows and are commonly found in manure, bedding material, soil, and contaminated water. Coliforms contaminate raw milk through the exterior of udder and teats and contaminated milking equipment.
      Previous studies found low correlations among these bacterial counts (
      • Boor K.J.
      • Brown D.P.
      • Murphy S.C.
      • Kozlowski S.M.
      • Bander D.K.
      Microbiological and chemical quality of raw milk in New York State.
      ;
      • Jayarao B.M.
      • Pillai S.R.
      • Sawant A.A.
      • Wolfgang D.R.
      • Hegde N.V.
      Guidelines for monitoring bulk tank milk somatic cell count and bacterial counts.
      ;
      • Elmoslemany A.M.
      • Keefe G.P.
      • Dohoo I.R.
      • Jayarao B.M.
      Risk factors for bacteriological quality of bulk tank milk in Prince Edward Island dairy herds. Part 1: Overall risk factors.
      ). The low correlations among these parameters indicate that they have different sources. Research-validated data on the specific on-farm risks associated with elevated bacterial counts in BTM is limited; therefore, the objective of this study was to evaluate on-farm risk factors for high bacterial counts in BTM through the use of a case-control study design. In the first part of this study (
      • Elmoslemany A.M.
      • Keefe G.P.
      • Dohoo I.R.
      • Jayarao B.M.
      Risk factors for bacteriological quality of bulk tank milk in Prince Edward Island dairy herds. Part 1: Overall risk factors.
      ), the overall risk factors for high bacterial counts in BTM as defined by combination of all 4 bacterial tests (TAC, PIC, LPC, and CC) were addressed. This paper focuses on specific risk factors for each of the 4 bacterial counts individually.

      Materials and Methods

      BTM Analyses, Study Design, and Data Collection

      Details on BTM analyses, study design, and data collection can be found in the companion paper of study (
      • Elmoslemany A.M.
      • Keefe G.P.
      • Dohoo I.R.
      • Jayarao B.M.
      Risk factors for bacteriological quality of bulk tank milk in Prince Edward Island dairy herds. Part 1: Overall risk factors.
      ). Additionally, descriptive statistics on different bacterial counts can be found in
      • Elmoslemany A.M.
      • Keefe G.P.
      • Dohoo I.R.
      • Jayarao B.M.
      Risk factors for bacteriological quality of bulk tank milk in Prince Edward Island dairy herds. Part 1: Overall risk factors.
      . Briefly, bulk tank raw milk quality was evaluated on all Prince Edward Island dairy herds (n = 235) over a 2-yr period (March 2005 to March 2007). Every other week, TAC, PIC, LPC, and CC were conducted using Petrifilms (3M Canada, London, Ontario, Canada).
      For the assessment of risk factors, a case-control study was conducted from January 2006 to May 2007. Four case-control groups were defined based on the last 6 results of each test before on-farm evaluation. To be classified as a TAC, PIC, or CC case, the herd was required to have at least 4 high TAC (>20,000 cfu/mL), PIC (>50,000 cfu/mL), or CC (>50 cfu/mL) counts out of the last 6 analyses for each test, respectively. For the LPC case group, a herd was required to have at least 3 high (>200 cfu/mL) LPC results out of the last 6 analyses. Thresholds for high counts were selected based on previous literature (
      • Murphy S.C.
      Raw milk bacteria tests: Standard plate count, preliminary incubation count, lab, pasteurization count and coliform count. What do they mean for your farm?.
      ). Control groups had low counts in the last 6 analyses for each test in the corresponding case group (TAC, PIC, CC, and LPC). The number of cases and controls in each group were 16 and 39, 21 and 31, 12 and 50, and 8 and 54 for TAC, PIC, CC, and LPC, respectively.
      Data collection included 3 main aspects: 1) observation and questionnaire on basic hygiene and management practices, milking procedures, and equipment cleaning and maintenance; 2) evaluation of cleaning efficiency of the milking system (thermal, chemical, and physical components of cleaning process); monitoring the presence of organic film deposits on milk contact surfaces and evaluation of efficiency of the cooling system; and 3) evaluation of environmental and cow hygiene. Complete description of on-farm data collection can be found in the companion paper (
      • Elmoslemany A.M.
      • Keefe G.P.
      • Dohoo I.R.
      • Jayarao B.M.
      Risk factors for bacteriological quality of bulk tank milk in Prince Edward Island dairy herds. Part 1: Overall risk factors.
      ).

      Statistical Analyses

      Unconditional Associations

      Unconditional associations between each of the case-control groups (TAC, PIC, LPC, and CC) and management factors were examined using simple logistic regression. For each outcome, only variables with significance level (P < 0.15) were considered for subsequent analysis. Because of the low number of cases in each of the case-control groups, multivariable models were unstable and sensitive to minor changes in the predictor variables; therefore, we could not fit multivariable models. Accordingly, the relationship between factors that were significant in the unconditional association at P < 0.05 and each of the outcomes of interest was further explored using correspondence analysis.

      Correspondence Analyses

      Correspondence analysis is an exploratory technique designed to analyze the relationship among a set of categorical variables. This method can be used to produce a graphical display of the relationship among categories of different variables in 2 or more dimensions. The technique defines a measure of distance between any 2 points, where points are the values (categories) of the discrete variables. When interpreting the graph, the proximity of the points to each other indicates the strength of association. The closer together the points are, the more closely they are associated. The value of the outcome variable can also be presented on the same graph to determine which clusters of predictor variable values are associated with the outcome (
      • Greenacre M.
      Correspondence Analysis in Practice.
      ). As only categorical variables can be used in this technique, continuous variables such as cow hygiene score, teat end cleanliness score, and detergent wash temperature and alkalinity were first categorized based on the percentiles of the distribution of the variables or, when available, based on recommended thresholds.

      Results

      The results of the unconditional association between each of the outcomes of interest and different management factors are shown in Tables 1 and 2. These variables were classified into 4 different groups as indicated in the companion paper (
      • Elmoslemany A.M.
      • Keefe G.P.
      • Dohoo I.R.
      • Jayarao B.M.
      Risk factors for bacteriological quality of bulk tank milk in Prince Edward Island dairy herds. Part 1: Overall risk factors.
      ).
      Table 1Unconditional association (P < 0.15) between environment- and cow hygiene-related factors and each of total aerobic (TAC), preliminary incubation (PIC), laboratory pasteurization (LPC), and coliform (CC) counts
      ParameterOdds ratio
      TACPICLPCCC
      Milking cow stall hygiene score3.98
      P<0.01;
      4.00
      P<0.05;
      Dry cow stall hygiene score7.42
      P<0.01;
      4.72
      P<0.01;
      Cow hygiene score
      Average of udder, leg, and flank scores.
      7.00
      P<0.01;
      Milking cow udder hygiene score7.71
      P<0.01;
      Milking cow leg hygiene score4.73
      P<0.01;
      3.22
      P<0.05;
      Milking cow flank hygiene score3.24
      P<0.05;
      Dry cow udder hygiene score4.00
      P<0.15.
      Teat cleanliness score4.60
      P<0.01;
      5.52
      P<0.05;
      3.21
      P<0.05;
      Udder hair clipping0.26
      P<0.05;
      Teat wash7.62
      P<0.01;
      3.77
      P<0.05;
      Pre-dip0.26
      P<0.05;
      0.31
      P<0.05;
      1 Average of udder, leg, and flank scores.
      *** P < 0.01;
      ** P < 0.05;
      * P < 0.15.
      Table 2Unconditional association (P < 0.15) between equipment hygiene related factors and each of total aerobic (TAC), preliminary incubation (PIC), laboratory pasteurization (LPC), and coliform (CC) counts
      ParameterOdds ratio
      TACPICLPCCC
      Temperature (°C)
       Prewash fill0.93
      P<0.01;
       Prewash drain0.92
      P<0.05;
      0.89
      P<0.01;
       Detergent wash temperature score
      Average of alkaline wash start and end temperatures of bulk tank and pipeline.
      0.89
      P<0.01;
      0.89
      P<0.01;
       Bulk tank alkaline wash fill0.93
      P<0.05;
      0.95
      P<0.15.
      0.90
      P<0.01;
       Bulk tank alkaline wash drain0.89
      P<0.01;
      0.93
      P<0.05;
       Pipeline alkaline wash start0.91
      P<0.05;
      0.92
      P<0.05;
       Pipeline alkaline wash end0.92
      P<0.15.
      0.86
      P<0.05;
      Chemistry
       Bulk tank alkaline wash alkalinity (ppm)1.001
      P<0.05;
       Bulk tank alkaline wash chlorine > 100 ppm0.24
      P<0.05;
       Bulk tank alkaline wash pH4.3
      P<0.15.
       Bulk tank acid rinse (yes vs. no)0.22
      P<0.15.
       Water hardness score (grain per gallon)1.33
      P<0.15.
      1.25
      P<0.05;
       Pipeline alkaline wash alkalinity (ppm)1.003
      P<0.05;
      1.002
      P<0.05;
      1.002
      P<0.15.
       Pipeline alkaline wash chlorine > 100 ppm0.23
      P<0.05;
       Pipeline alkaline wash pH2.43
      P<0.15.
      3.23
      P<0.15.
      Other equipment
       Slug score0.84
      P<0.15.
      0.79
      P<0.15.
       Bulk tank automatic cleaning0.12
      P<0.01;
       Milk house water bacterial check: yearly vs. longer2.85
      P<0.15.
      2.85
      P<0.15.
      8.75
      P<0.05;
      3.52
      P<0.15.
       Milk house water hardness check: yearly vs. longer2.42
      P<0.15.
      7.00
      P<0.05;
      3.24
      P<0.15.
       Bulk tank outlet bioluminescence score > 02.62
      P<0.15.
      2.43
      P<0.15.
      3.91
      P<0.15.
       Water softener (yes vs. no)0.32
      P<0.15.
      0.08
      P<0.01;
      1 Average of alkaline wash start and end temperatures of bulk tank and pipeline.
      *** P < 0.01;
      ** P < 0.05;
      * P < 0.15.
      Group 1 (Table 1) includes factors related to environment and cow hygiene. The TAC was associated with a large number of hygiene-related factors. High hygiene scores (i.e., dirty) of stall, cow, and teat end were associated with elevated TAC in BTM. Additionally, using water to wash the teats for premilking udder preparation was also a risk for high TAC. On the other hand, udder hair clipping and the use of teat predip were protective. A similar level of association was also evident between hygiene-related factors and PIC except for udder, leg, and flank hygiene scores, which were not associated with PIC. The CC was mainly associated with leg hygiene and teat end cleanliness scores, with dirty teats or legs being risk factors.
      Groups 2, 3, and 4 comprised factors related to equipment hygiene (wash solution temperature, chemistry, and physical cleaning). There was an association between temperature-related factors and LPC and CC. In all cases, high temperature was always protective (Table 2).
      Chemistry-related factors were mainly associated with LPC and CC. High alkalinity of detergent wash was a risk factor for all counts. High chlorine concentration of the detergent wash was protective for LPC and CC, whereas high hardness score was associated with elevated LPC and CC (Table 2).
      The last group comprised physical cleaning and other equipment-related factors (Table 2). High slug score (sufficient physical cleaning) of the pipeline was protective for both TAC and LPC. Less-frequent evaluation of milk house water for bacteria and hardness was a risk for all bacterial counts. High bioluminescence at the bulk tank outlet was associated with elevated TAC, PIC, and CC. Finally, the use of a water softener was protective for TAC and PIC.
      The categories used for correspondence analysis are shown in Table 3. Figures 1, 2, 3, and 4 summarize the results of correspondence analysis. Figure 1 illustrates the risk factors for TAC. Low TAC (control) was closely associated with using teat predip, clean cow (udder, leg, and flank), clean teat end, and udder hair clipping. On the other hand, high TAC (case) was associated with not using teat predip and high cow hygiene score (i.e., dirty cow). The graph also highlights the close association between udder hair clipping and hygienic condition of the teat end and the cow.
      Table 3Categories of the continuous predictors that were used for correspondence analysis
      VariableRange%Merging categories
      Low and medium categories were merged if they were originally located close to each other in correspondence analysis graph.
      Cow hygiene score (udder, leg, and flank)1–3.25
       Low <1.325Low and medium (TAC)
      TAC = total aerobic count.
       Medium 1.31–1.9950
       High ≥225
      Teat cleanliness1–3.12
       Low <1.125Low and medium (TAC)
       Medium 1.11–1.6950
       High ≥1.725
      Pipeline alkaline wash alkalinity (ppm)
      Thresholds were selected according to IDF Bulletin (Reinemann et al., 2003).
      0–1500
       Low <2509Low and medium (TAC)
       Medium 250–50053
       High ≥50038
      Bulk tank alkaline wash alkalinity (ppm)
      Thresholds were selected according to IDF Bulletin (Reinemann et al., 2003).
      0–4000
       Low <40032
       Medium 400–80043
       High ≥80025
      Bulk tank alkaline wash fill temperature (°C)
      Thresholds were selected according to IDF Bulletin (Reinemann et al., 2003).
      8.3–80
       Adequate ≥7125
       Inadequate <7175
      Detergent wash temperature score (°C)
      Average of alkaline wash start and end temperatures of bulk tank and pipeline.
      23.5–70.5
       Low <5025
       Medium 50–5750
       High >5725
      1 Low and medium categories were merged if they were originally located close to each other in correspondence analysis graph.
      2 TAC = total aerobic count.
      3 Thresholds were selected according to IDF Bulletin (
      • Reinemann D.J.
      • Wolters G.M.V.H.
      • Billon P.
      • Lind O.
      • Rasmussen M.D.
      Review of Practices for Cleaning and Sanitation of Milking Machines.
      ).
      4 Average of alkaline wash start and end temperatures of bulk tank and pipeline.
      Figure thumbnail gr1
      Figure 1Multiple correspondence analysis of risk factors for high total aerobic count (TAC) in bulk tank raw milk in Prince Edward Island dairy herds.
      Figure thumbnail gr2
      Figure 2Multiple correspondence analysis of risk factors for high preliminary incubation count (PIC) in bulk tank raw milk in Prince Edward Island dairy herds.
      Figure thumbnail gr3
      Figure 3Multiple correspondence analysis of risk factors for high laboratory pasteurization count (LPC) in bulk tank raw milk in Prince Edward Island dairy herds.
      Figure thumbnail gr4
      Figure 4Multiple correspondence analysis of risk factors for high coliform count (CC) in bulk tank raw milk in Prince Edward Island dairy herds.
      Low PIC (control) was closely related to using teat predip and high bulk tank alkaline wash temperature, whereas high PIC (case) was associated with dirty teat end, and not using teat predip (Figure 2).
      Low LPC (control) was associated with high temperature of the detergent wash, low water hardness, and medium level of bulk tank alkaline wash alkalinity, whereas high LPC (case) was mainly related to the use of low temperature of the detergent wash and high level of bulk tank alkaline wash alkalinity (Figure 3).
      Figure 4 shows that low CC (control) was associated with clean teat end, automatic cleaning of bulk tank milk, low water hardness, and high or medium detergent wash temperature. On the other hand, high CC (case) was mainly related to dirty teat ends and low temperature of detergent wash.

      Discussion

      To achieve high raw milk quality, producers should be aware of the factors that influence contamination of raw milk and how they can be controlled. In this study, variables that had significant unconditional associations with milk quality parameters were mainly related to cow and equipment hygiene and were classified into 4 groups for ease of discussion.
      Group one included variables measuring environmental and cow hygiene. It was observed that high cow (udder, leg, and flank) hygiene and teat end cleanliness scores (dirty) were associated with increased risk of elevated bacterial count in BTM (
      • Elmoslemany A.M.
      • Keefe G.P.
      • Dohoo I.R.
      • Jayarao B.M.
      Risk factors for bacteriological quality of bulk tank milk in Prince Edward Island dairy herds. Part 1: Overall risk factors.
      ). The current study showed that cow hygiene score was mainly related to TAC, whereas teat end cleanliness score was associated with TAC, PIC, and CC. Additionally, there was a significant (P < 0.001) association between milking stall hygiene and cow and teat end cleanliness scores, with cleaner stalls being associated with cleaner cows and teats. A clean and dry cow environment is important in preventing environmental mastitis (
      • Bartlett P.C.
      • Miller G.Y.
      • Lance S.E.
      • Heider L.E.
      Managerial determinants of intramammary coliform and environmental streptococci infections in Ohio dairy herds.
      ). Previous studies indicated that bacterial counts in bedding are positively correlated with stall cleanliness (
      • Zdanowicz M.
      • Shelford J.A.
      • Tucker C.B.
      • Weary D.M.
      • von Keyserlingk M.A.G.
      Bacterial populations on teat ends of dairy cows housed in free stalls and bedded with either sand or sawdust.
      ) and as the percentage of dirty stalls increased, the risk of clinical mastitis also increased (
      • Schukken Y.H.
      • Grommers F.J.
      • Van de Geer D.
      • Erb H.N.
      • Brand A.
      Risk factors for clinical mastitis in herds with a low bulk milk somatic cell count. 1. Data and Risk factors for all cases.
      ). Additionally,
      • Barkema H.W.
      • Schukken Y.H.
      • Lam T.J.G.M.
      • Beoboer M.L.
      • Benedictus G.
      • Brand A.
      Management practices associated with low, medium and high somatic cell counts in bulk milk.
      reported that herds with SCC >250,000 cells/mL had more manure in stalls, had less-frequently cleaned stalls, and used less bedding in stalls. In our study, dirty stall was associated with high TAC, which could be caused by many factors including mastitis microorganisms (
      • Murphy S.C.
      Raw milk bacteria tests: Standard plate count, preliminary incubation count, lab, pasteurization count and coliform count. What do they mean for your farm?.
      ). Furthermore, dirty stall was associated with elevated PIC, which could be attributed to contamination from the exterior of the udder and teats by bacteria from the cows’ environment (
      • Jayarao B.M.
      • Pillai S.R.
      • Sawant A.A.
      • Wolfgang D.R.
      • Hegde N.V.
      Guidelines for monitoring bulk tank milk somatic cell count and bacterial counts.
      ). Although hygiene score of the stall was associated with both TAC and PIC, we did not find any associations between bacterial counts in bedding and any of the milk quality parameters (data not shown). This may indicate that hygiene score of the stall is a better predictor of milk quality than bacterial counts in bedding.
      Dirty teats and udders are considered one of the main sources of environmental bacteria in milk. Between milking, the teats and udder often become soiled with manure and bedding materials. If the teats were not thoroughly cleaned and dried before milking, this dirt with the associated microorganisms will be transferred into the milk (
      • Chambers J.V.
      The microbiology of raw milk.
      ). Contamination from the exterior of the udder and teats can contribute microorganisms from the cow environment such as streptococci, staphylococci, spore-formers, coliforms, and other gram-negative bacteria, which in turn can elevate TAC, PIC, LPC, and CC (
      • Murphy S.C.
      Raw milk bacteria tests: Standard plate count, preliminary incubation count, lab, pasteurization count and coliform count. What do they mean for your farm?.
      ). Our results indicate that dirty teats are risk factor for elevated TAC, PIC, and CC counts. A high number of environmental bacteria in BTM indicates a problem related to environmental and milking hygiene (
      • Jayarao B.M.
      • Wolfgang D.R.
      Bulk-tank milk analysis. A useful tool for improving milk quality and herd udder health.
      ).
      The influence of dirty cows on bacterial counts depends on the extent of soiling of the teat surface and on premilking udder preparation practices. The purpose of premilking udder hygiene is to reduce bacterial contamination on the teats to minimal numbers before the attachment of the milking unit (
      • Pankey J.W.
      Premilking udder hygiene.
      ).
      • Galton D.M.
      • Petersson L.G.
      • Merrill W.G.
      Effects of premilking udder preparation practices on bacterial counts in milk and on teats.
      reported that SPC fell from approximately 13,500 cfu/mL with no udder preparation to between 2,701 and 4,354 cfu/mL after application of an iodophor predip. Similarly, coliform counts were reduced from >18,000 cfu/mL to <6,000 cfu/mL using the same treatment. Recently,
      • Jayarao B.M.
      • Pillai S.R.
      • Sawant A.A.
      • Wolfgang D.R.
      • Hegde N.V.
      Guidelines for monitoring bulk tank milk somatic cell count and bacterial counts.
      reported a reduction in environmental mastitis pathogens and psychrotrophic and thermoduric bacteria with the use of teat predipping. On the other hand, washing the teats with water before milking resulted in elevated SPC (
      • Galton D.M.
      • Adkinson R.W.
      • Thomas C.V.
      • Smith T.W.
      Effects of premilking udder preparation on environmental bacterial contamination of milk.
      ) and increased prevalence of coliform mastitis (
      • Bartlett P.C.
      • Miller G.Y.
      • Lance S.E.
      • Heider L.E.
      Managerial determinants of intramammary coliform and environmental streptococci infections in Ohio dairy herds.
      ).
      Our results corroborate earlier findings with respect to SPC and PIC; however, our findings differed with respect to CC as our study revealed that there was no association between CC and methods of premilking udder preparation. In our study, the lack of association between CC and premilking udder preparation procedures could be attributed to the limited number of CC cases or may indicate that coliforms had contaminated BTM through routes other than contaminated teats, such as milking equipment or contaminated water source (
      • Kagkli D.M.
      • Vancanneyt M.
      • Vandamme P.
      • Hill C.
      • Cogan T.M.
      Contamination of milk by enterococci and coliforms from bovine faeces.
      ). A lack of correlation between premilking teat cleaning regimens and environmental bacteria was also reported by
      • Feldmann M.
      • Zimmermann A.
      • Hoedemaker M.
      Influence of milking technique, milking hygiene and environmental hygiene parameters on the microbial contamination of milking machines.
      and
      • Gibson H.
      • Sinclair L.A.
      • Brizuela C.M.
      • Worton H.L.
      • Protheroe R.G.
      Effectiveness of selected premilking teat-cleaning regimes in reducing teat microbial load on commercial dairy farms.
      , which suggests that teat end contamination may not be the primary source of coliforms in BTM.
      The last 3 groups of predictors are all related to cleaning and sanitation of the milking equipment, which illustrates the importance of milking equipment as a source of bacterial contamination of raw milk. Adequate cleaning and disinfection of milking equipment is necessary to remove residues and microorganisms from equipment surfaces (
      • Reinemann D.J.
      • Wolters G.M.V.H.
      • Billon P.
      • Lind O.
      • Rasmussen M.D.
      Review of Practices for Cleaning and Sanitation of Milking Machines.
      ).
      In this study, factors measuring the temperature of the wash solution were mainly associated with LPC and CC, with higher temperatures being protective. Previous reports indicated that the temperature of cleaning solution could affect the type of microorganisms on milking equipment surfaces. The predominance of thermoduric over other microflora from the milking equipment may be related to the use of high temperatures during cleaning of the equipment, whereas the dominance of thermolabile species such as Pseudomonas spp. and coliforms, indicate the use of low (<42°C) cleaning temperature (
      • Murphy S.C.
      • Boor K.J.
      Trouble-shooting sources and causes of high bacteria counts in raw milk.
      ;
      • Feldmann M.
      • Zimmermann A.
      • Hoedemaker M.
      Influence of milking technique, milking hygiene and environmental hygiene parameters on the microbial contamination of milking machines.
      ).
      In our study, CC controls were associated with using either medium or high cleaning temperatures, whereas CC cases were associated with the use of low cleaning temperature. However, low LPC was associated with high water temperature, which indicates that the dominance of specific types of microorganisms from milking equipment is influenced by additional factors such as chemical cleaning.
      The majority of cleaning chemical-related factors were associated with LPC and CC. High water hardness score was associated with elevated LPC and CC, whereas high alkalinity of the detergent wash was a risk for all bacterial counts. Very hard water or use of highly alkaline cleaners enhances the development of milk-stone. Additionally, high alkalinity may also lead to corrosion of surfaces and increase deterioration of rubber parts and gaskets, resulting in conditions for bacterial adherence and formation of biofilms (
      • Austin J.W.
      • Bergeron G.
      Development of bacterial biofilms in dairy processing lines.
      ). Microorganisms such as micrococci, enterococci, coliforms, aerobic spore-formers, certain lactobacilli, and other gram-negative bacteria become embedded in the biofilm, multiply in it, and are protected from the effects of detergent and disinfectant solutions (
      • Braunig J.
      • Hall P.
      Milk and dairy products.
      ). Based on the composition of the microflora embedded in the biofilm, it could be associated with elevated LPC, CC, PIC, or TAC.
      The associations between LPC, CC, and the majority of equipment hygiene-related factors support previous reports by
      • Thomas S.B.
      • Druce R.G.
      • King K.P.
      The microflora of poorly cleansed farm dairy equipment.
      and
      • Reinemann D.J.
      • Wolters G.M.V.H.
      • Billon P.
      • Lind O.
      • Rasmussen M.D.
      Review of Practices for Cleaning and Sanitation of Milking Machines.
      , who indicated that improperly cleansed dairy equipment and bacterial incubation on milk contact surfaces are the main source of thermoduric and gram-negative rods. These results also support the conclusion by
      • Villar A.
      • García J.A.
      • Iglesias L.
      • García M.L.
      • Otero A.
      Application of principal component analysis to the study of microbial populations in refrigerated raw milk from farms.
      that thermoduric and coliform counts characterize hygienic condition of dairy equipment.
      Less frequent evaluation of water source was a risk for high counts in all quality parameters. Untreated water supply could be a source of contamination with coliforms, pseudomonas, and other gram-negative bacteria (
      • Chambers J.V.
      The microbiology of raw milk.
      ), which could incubate on milking equipment and elevate CC and PIC. Furthermore, water hardness minerals can react with cleaning agents and reduce their cleaning efficiency (
      • Reinemann D.J.
      • Wolters G.M.V.H.
      • Billon P.
      • Lind O.
      • Rasmussen M.D.
      Review of Practices for Cleaning and Sanitation of Milking Machines.
      ). Regular checking of water supply could also indicate a positive attitude of the farmers toward hygiene in general.
      High bulk-tank-outlet bioluminescence score was associated with elevated TAC, PIC, and CC. The outlet valve is considered one of the major sources for contamination of raw milk stored in bulk tank. This is because it is difficult to clean and may allow accumulation of milk residues that provide a good environment for bacterial growth (
      • Chambers J.V.
      The microbiology of raw milk.
      ).
      Finally, although correspondence analysis does not assess the statistical significance of the relationship between independent variables and the outcome, it summarizes the complex relationships that exist among these variables. Our correspondence analysis results showed that preventive categories were tightly clustered around controls, whereas risk categories were more diffuse around cases (less strongly associated with cases). These results suggest that managers in control herds were doing most practices correctly. The results also imply that managers in case herds did not necessarily fail to follow correct practices; however, failing to follow a comprehensive control system may cause the herd to be a case.

      Conclusions

      This study highlights the importance of different bacterial counts (TAC, PIC, LPC, and CC) as indicators of on-farm hygienic conditions during milk production. The TAC and PIC were mainly associated with environment and cow hygiene with poor hygiene score associated with elevated bacterial count. The LPC and CC were mainly related to equipment hygiene, with high temperature of the wash solution being protective, whereas high water hardness score and high alkalinity of the detergent wash were risk factors.

      Acknowledgments

      The authors acknowledge the technical support of Ron Sampson (Dairy Farmers of Prince Edward Island, Charlottetown, PE), Ricky Milton, Theresa Andrews, and Lloyd Dalziel (Atlantic Veterinary College, Charlottetown, PE). This research was funded by Dairy Farmers of Prince Edward Island, Agricultural Research Investment Fund (PEI Department of Agriculture, Charlottetown, PE), Purity Dairy (Charlottetown, PE), and Amalgamated Dairies Limited (Summerside, PE). Personal funding for A. M. Elmoslemany was provided by Mission Office, Ministry of Higher Education, Egypt.

      References

        • Austin J.W.
        • Bergeron G.
        Development of bacterial biofilms in dairy processing lines.
        J. Dairy Res. 1995; 62: 509-519
        • Barkema H.W.
        • Schukken Y.H.
        • Lam T.J.G.M.
        • Beoboer M.L.
        • Benedictus G.
        • Brand A.
        Management practices associated with low, medium and high somatic cell counts in bulk milk.
        J. Dairy Sci. 1998; 81: 1917-1927
        • Bartlett P.C.
        • Miller G.Y.
        • Lance S.E.
        • Heider L.E.
        Managerial determinants of intramammary coliform and environmental streptococci infections in Ohio dairy herds.
        J. Dairy Sci. 1992; 75: 1241-1252
        • Boor K.J.
        • Brown D.P.
        • Murphy S.C.
        • Kozlowski S.M.
        • Bander D.K.
        Microbiological and chemical quality of raw milk in New York State.
        J. Dairy Sci. 1998; 81: 1743-1748
        • Braunig J.
        • Hall P.
        Milk and dairy products.
        Microorganisms in Foods 6: Microbial Ecology of Foods Commodities. Kluwer Academic/Plenum Publisher, New York, NY2005 (Pages 643–715)
        • Chambers J.V.
        The microbiology of raw milk.
        in: Robinson R.K. Dairy Microbiology Handbook. 3rd ed. John Wiley & Sons Inc., New York, NY2002: 39-90
        • Elmoslemany A.M.
        • Keefe G.P.
        • Dohoo I.R.
        • Jayarao B.M.
        Risk factors for bacteriological quality of bulk tank milk in Prince Edward Island dairy herds. Part 1: Overall risk factors.
        J. Dairy Sci. 2009; 92: 2634-2643
        • Feldmann M.
        • Zimmermann A.
        • Hoedemaker M.
        Influence of milking technique, milking hygiene and environmental hygiene parameters on the microbial contamination of milking machines.
        Dtsch. Tierarztl. Wochenschr. 2006; 113: 274-281
        • Galton D.M.
        • Adkinson R.W.
        • Thomas C.V.
        • Smith T.W.
        Effects of premilking udder preparation on environmental bacterial contamination of milk.
        J. Dairy Sci. 1982; 65: 1540-1543
        • Galton D.M.
        • Petersson L.G.
        • Merrill W.G.
        Effects of premilking udder preparation practices on bacterial counts in milk and on teats.
        J. Dairy Sci. 1986; 69: 260-266
        • Gibson H.
        • Sinclair L.A.
        • Brizuela C.M.
        • Worton H.L.
        • Protheroe R.G.
        Effectiveness of selected premilking teat-cleaning regimes in reducing teat microbial load on commercial dairy farms.
        Lett. Appl. Microbiol. 2008; 46: 295-300
        • Greenacre M.
        Correspondence Analysis in Practice.
        2nd ed. Chapman and Hall/CRC Press, New York, NY2007
        • Jayarao B.M.
        • Pillai S.R.
        • Sawant A.A.
        • Wolfgang D.R.
        • Hegde N.V.
        Guidelines for monitoring bulk tank milk somatic cell count and bacterial counts.
        J. Dairy Sci. 2004; 87: 3561-3573
        • Jayarao B.M.
        • Wolfgang D.R.
        Bulk-tank milk analysis. A useful tool for improving milk quality and herd udder health.
        Vet. Clin. North Am. Food Anim. Pract. 2003; 19: 75-92
        • Kagkli D.M.
        • Vancanneyt M.
        • Vandamme P.
        • Hill C.
        • Cogan T.M.
        Contamination of milk by enterococci and coliforms from bovine faeces.
        J. Appl. Microbiol. 2007; 103: 1393-1405
        • Murphy S.C.
        Raw milk bacteria tests: Standard plate count, preliminary incubation count, lab, pasteurization count and coliform count. What do they mean for your farm?.
        in: Natl. Mastitis Council. Reg. Mtg. Proc., Syracuse, New York. Natl., Mastitis Council Inc., Madison, WI.1997: 34-41
        • 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
        • Pankey J.W.
        Premilking udder hygiene.
        J. Dairy Sci. 1989; 72: 1308-1312
        • Reinemann D.J.
        • Wolters G.M.V.H.
        • Billon P.
        • Lind O.
        • Rasmussen M.D.
        Review of Practices for Cleaning and Sanitation of Milking Machines.
        in: Bull. 381. Int. Dairy Fed., Brussels, Belgium2003
        • Schukken Y.H.
        • Grommers F.J.
        • Van de Geer D.
        • Erb H.N.
        • Brand A.
        Risk factors for clinical mastitis in herds with a low bulk milk somatic cell count. 1. Data and Risk factors for all cases.
        J. Dairy Sci. 1990; 73: 3463-3471
        • Thomas S.B.
        • Druce R.G.
        • King K.P.
        The microflora of poorly cleansed farm dairy equipment.
        J. Appl. Bacteriol. 1966; 29: 409-422
        • Villar A.
        • García J.A.
        • Iglesias L.
        • García M.L.
        • Otero A.
        Application of principal component analysis to the study of microbial populations in refrigerated raw milk from farms.
        Int. Dairy J. 1996; 6: 937-945
        • Zdanowicz M.
        • Shelford J.A.
        • Tucker C.B.
        • Weary D.M.
        • von Keyserlingk M.A.G.
        Bacterial populations on teat ends of dairy cows housed in free stalls and bedded with either sand or sawdust.
        J. Dairy Sci. 2004; 87: 1694-1701