BULK TANK SOMATIC CELL COUNT IN MILK SAMPLES FROM STATE OF PARANÁ

This research studied somatic cell count in bulk tank milk samples (BTSCC) from the state of Paraná, Brazil, at the Milk Quality Laboratory of the Dairy Herd Analysis Service of the Holstein Association of Paraná, the result of technical and scientific cooperation between UFPR and McGill University of Canada. A total of 1,950,034 bulk tank milk samples from ten regions of the state of Paraná were analyzed between January 2005 and April 2012. Fixed effects were the month and year of analysis, region and age of the sample. Means and standard deviations of BTSCC were 553,519 ± 545,532 cells/ml, respectively. All fixed effects were statistically significant (P<0.01). Highest values for BTSCC are observed in the month of February (554,000 cells/ml ± 1.45) and lowest values in September (450,000 cells/ml ± 1.47). Similarly, the highest values were found in the year 2010 (567,000 cells/ml ± 1.16), the lowest BTSCC was found in 2012 (444,000 cells/ml ± 1.57). The region effect was also significant with the highest values found in the South Center/ Guarapuava (668,000 cells/ml ± 0.87) and the lowest in the Southwest/ Francisco Beltrão (359,000 cells/ml ± 2.00). Both variables showed a reduction of their values with increasing age of the sample, from 518,000 ± 1.08 to 472,000 cells/ml ± 2.14 between the first and the seventh day, for the BTSCC. Coefficient of variation for BTSCC was 96.10%. The R2 was 0.39 for BTSCC.


INTRODUCTION
The somatic cell count (SCC) is an established evaluation parameter used by programs that monitor udder health, milk quality improvement and payment for quality in traditional producer countries.(Schukken et al., 1993, Godkin, 2000, Fonseca, 2001).
The SCC has been widely used in developed countries since electronic devices made this practice available to producers (Fonseca and Santos, 2000;Santos, 2001 andGodking, 2000) In Brazil, it  The SCC has been used by industries as payment criteria for milk quality, making it a tool for management and quality monitoring, being directly related to programs to reduce production losses and to opportunities for milk higher pay (Harmon andReneau, 1993, Rysanek andBabak, 2005).
High bulk tank somatic cell count (BTSCC) affects the composition of milk, the products shelf life, causing significant losses to dairy industry (Monardes, 1998).
According to Harmon & Reneau (1993), there is a significant association between BTSCC with the limit of 1,500,000 cells/ml.The percentage of infected udder quarters could reach up to 48% and the losses of milk yield up to 29 %.(Brazil, 2002) and 62 (Brazil, 2011) of the Ministry of Agriculture, Livestock and Supply (MAPA).This work analyzed the BTSCC samples from dairy industries under its services.
This research objective was to evaluate how environment effects (month and year of analysis, and the region and age of the sample) could influence the BTSCC.

MATERIAL AND METHODS
There were used 1,950,034 bulk tank milk samples from production sites located in 10 regions, and respective main city, in the state of Paraná (PR).They were analyzed from January 2005 to April 2012.
The bulk tank milk samples were collected monthly by personnel trained by the dairy industry, before collecting the bulk tank milk is homogenized to analyze maintained in a sterilized jar that uses the default Bronopol, according to the procedures recommended by the manuals of Field Operations (Horst, 2008), and Sample Collection (Horst, 2010) of the Milk Quality Laboratory.The samples were placed in standard vials (70 ml), using the preservative bronopol (2-bromo-2-nitro propane-1,3diol) and sent to Milk Quality Laboratory localized in Associação Paranaense de Criadores de Bovinos da Raçã Holandesa, Curitiba -PR -Brazil.
The number of days elapsed between the farm milk collection and the analysis in the laboratory, Curitiba-PR, was defined as the age of the sample.The milk samples were analyzed for BTSCC, using equipment Somacount 500 ®, by flow cytometry.
The BTSCC data was prepared from the original Milk Quality Laboratory database to achieve better accuracy: BTSCC values equal to zero or negative as well as greater than 4,525,000 cells/ml were not considered.The same was valid for age of the sample lower than 1 and greater than 7 days.
All the database preparation, data pre-sorting and statistical analysis for BTSCC values were performed in the Department of Animal Science of the SCA of the UFPR in Curitiba-PR.
All data were analyzed through computer software SAS® version 9.3.Studied traits were analyzed through the following mathematical model: Y ijkl = BTSCC values, collected in month i, year j, region k, with age of the sample l and associated random error of each observation e ijkl.µ = general average; M i = effects of month of analyzes i, being i = 1 (Jan.), 2 (Feb.),... 12 (Dec.);*In 2012, there were milk samples collected only from the first four months of the year.
The evaluated dependent variable was BTSCC, as a result of the reading by flow cytometry equipment, Somacount 500 ® (Bentley Instruments, 1995b), in thousand cells/ml.Comparisons between means were performed using the Tukey test at 1% probability.

RESULTS AND DISCUSSION
The estimated mean, its standard deviation and the variation coefficient of BTSCC from the 1,950,034 analyzed samples from 2005 to 2012, were respectively 553,519 ± 545,532 cells/ml and 96.10%.
The large variability (CV 96.10%) in the BTSCC values is characterized by the standard deviation similar to the mean.Machado et al. (2000) and Paula et al. (2004) also found the same.
The average BTSCC found in this study is higher than those reported by Franks (2001) and Godkin (1999), considering averages of BTSCC in different countries.Franks (2001) observed the lowest average in Switzerland (112,000 cells/ml) and the highest in Israel (382,000 cells/ml), and Godkin (1999), in the Province of Ontario, Canada, observed an average of 250,000 cells/ml.
In Brazil, similar results were obtained by Machado et al. (2000), working with a smaller number of data (4785 samples).They found mean and standard deviation of 505,000 ± 593,000 cells/ml, and Paula et al. (2004) analyzed 257,540 bulk tank somatic cell count in the states of Santa Catarina, Paraná and São Paulo, from 1999 to 2001, and found an average of 486,812 ± 401,547 cells/ml with a variation coefficient of 62.35%.
According to Harmon & Reneau (1993), the limit of 500,000 cells/ml found in this study would be associated to 16% of infected udder quarters in herds and to 6% loss in milk production.
These results are considered high when compared with those countries with well-developed dairy industry, reflecting carelessness by producers regarding the health of the mammary gland.That seems to be the result of the lack of incentives by milk industries to establish programs to pay for milk quality based on SCC.
In Brazil, since the publication of the MAPA IN 51 (Brazil, 2002), the SCC criteria has been used to promote the improvement of milk quality.According to Fonseca (2001), a survey conducted in 93 dairy plants in the 1998 offseason, only 7% of the producers used the SCC test.
It is expected that, after the publication of IN 62 (Brazil, 2011), establishing to 2015 new legal limit of 400,000 cells/ml, industries and producers should give greater attention to SCC and that would decrease over the years.According Monardes (1998), the establishment of legal limits on economic blocks and countries like European Union (400,000 cells/ml), Canada (500,000 cells/ml) and New Zealand (400,000 cells/ml) respectively in 1992, 1994 and 1995 were decisive in enabling public policies for food security.
The results of variance analysis of BTSCC in Paraná are shown in Table 1.All the variables in the mathematical model were significant (P <0.01).
The month of analysis significantly influenced the BTSCC (P <0.01), as can be seen in the summary of the variance analysis (Table 1).
In Table 2, it can be identified higher estimates of adjusted means for BTSCC in summer, which corresponds to the period from January to March (525,000 to 532,000 cells/ml).Moreover, lower estimates were found in late winter and early spring from August to October (477,000 to 463,000 cells/ml).Those results resemble findings from other authors (Harmon & Reneau, 1993;Harmon, 1998b;Ott et al. 1999;Pritchard et al. 2001 andPaula et al. 2004).
According to Harmon & Reneau (1993), Harmon (1998b) and Paula et al. (2004), the stress of high temperatures and humidity, increasing the susceptibility to infection and the number of pathogens to which cows are exposed, make the summer the period with the highest incidence of clinical mastitis, mainly of environmental origin.
Table 2 shows that the months of March and April, June and July, June and November are similar to each other (P <0.01), why in the state of Paraná temperatures in these months are milder and therefore they have no differences significant.
In the summary of the variance analysis (Table 1), it is shown the year of the analysis significantly influenced the variable BTSCC (P <0.01).
Several authors have also studied the effect of the years in BTSCC analysis (Schukken et al. 1990;Schukken et al. 1992a;Schukken et al. 1993;Sargeant et al. 1998;Godkin, 1999, andPaula et al ., 2004), showing a reduction in BTSCC over the years, due to the implementation of legal limits, mammary gland health monitoring programs and payment for quality.
In Table 4, it is shown that the region affected significantly (P <0.01) the characteristic studied (BTSCC).
The adjusted means and standard errors for the ten regions and their main cities are described in Table 4.The lowest BTSCC estimates were 359,000 cells/ml for the Southwest region, Francisco Beltrão -PR and highest 668,000 cells/ml for the Center South, Guarapuava -PR.
In Brazil, Noro (2004) and Paula et al. (2004) found similar results to this study, highlighting the differences found in herd size, production system, level of technology adopted, implementation of mammary gland health control programs and milk quality payment, stimulated by the dairy industries of those regions.
In the variance analysis summary (Table 1), it is shown that the age of the sample significantly influenced the BTSCC (P <0.01).These results contrast with those obtained by Monardes et al. (1996), who found no significant effect of the sample age on the SCC.
Table 5 presents the adjusted means for BTSCC that show similar values for two and three days, as four, five and six days are similar to each other (P<0.01).
However, as a tendency, when the age of the sample increased from 1 day to 7 days the BTSCC decreased from 518 to 472 (x 1000 cells/ml).Paula et al. (2004), found that from the first to the seventh day of age the samples had a reduction in the BTSCC values, contrasting with Kennedy et al. (1982), cited by Ostrensky (1999), who observed that the SCC values in the first three days remained virtually unchanged, and with Ostrensky (1999), who found an increase in the SCC values.Kennedy et al. (1982), cited by Ostrensky (1999), as in this study, also reported that the SCC was reduced as the age of the sample increased, showing a reduction of 28% (Ostrensky, 1999).
Table 5 shows that 68.07% of the samples were analyzed by the third day, while 96.42% until the sixth day.Monardes et al. (1996) reported that in Canada, 93% of the samples were analyzed by the third day and 99% by the seventh day after collection.

CONCLUSION
The averages for BTSCC in the State of Paraná are high compared to countries with well-developed dairy industry, reflecting a high level of mastitis in the studied herds associated with a significant amount of infected udder quarters that would lead to significant milk losses.Based on the results found in this study, it is suggested a demand for producers to reassess the management of mastitis control programs that would focus on the health of the mammary gland.The dairy industries must also reevaluate its payment programs for milk quality, correlating BTSCC with the reduction on production losses and improving opportunities for larger returns on milk production.The new legal limit to be implemented, as an instrument of public policy in favor of food security, should was introduced in 1991 by the Milk Quality Laboratory of the Dairy Herd Program Analysis of Paraná, from the Holstein Association of Paraná (APCBRH), Curitiba-PR.It was the result of technical and scientific cooperation between Universidade Federal do Paraná (UFPR) and McGill University in Canada.
A j = effects of year of analyzes j, being j = 2005, 2006, ... 2012*; R k = effects of the region k, being k = 1, 2, .., 10; I l = effects of the age of the sample in days l, being l = 1, 2, ... 7; e ijkl = associated random error of each observation Y ijkl.