Evaluation of models for somatic cell score lactation patterns in Holsteins

https://doi.org/10.1016/S0301-6226(00)00193-7Get rights and content

Abstract

Milk somatic cell score (SCS) typically reaches a minimum early in lactation and then rises. Nonlinear mixed effects models were used to describe this trajectory while accounting for between and within-cow variation. A total of 2387 SCS records from 217 Holsteins were analyzed. Four nonlinear and two linear functions were studied. Approximate maximum likelihood estimates indicated that cows free of intramammary infection had sharper SCS decreases after calving, and lower overall levels. Lactations starting between October and December had the highest fall of SCS levels at the beginning of lactation, and the smallest increases thereafter. In general, there was significant variation between cows’ individual trajectories. For some parameters and models, however, this variation was small. A four-parameter model suggested by Morant and Gnanasakthy was supported better by the data than the other five functional forms.

Introduction

Mastitis is an udder health disorder that causes substantial economic losses to the dairy industry (Shook, 1989). This inflammation of the mammary gland, usually in response to invasive agents, can be characterized by an increase in the somatic cell count (SCC) in milk. This trait or a logarithmic transform, called somatic cell score (SCS), is used as an indicator of udder health status for management and selection purposes.

After the beginning of lactation, SCS decreases to a minimum at around 60 days post-calving and increases thereafter (Wiggans and Shook, 1987). Variation in the shape and level of the SCS pattern is related to lactation number (Wiggans and Shook, 1987), to udder infection status (Sheldrake et al., 1983) and to individual cows. Heritability estimates of test-day SCS range between 0.07 and 0.44 (Ali and Shook, 1980, Kennedy et al., 1982, Emanuelson and Philipsson, 1984 Gadini et al., 1996). Monardes and Hayes (1985) reported repeatability estimates of test-day SCS of 0.36–0.42 across parities, and repeatability within lactations ranged from 0.47 to 0.59 (Emanuelson and Persson, 1984). These results suggest the need to consider both the variation within and between cows for an adequate modeling of SCS trends. Further, an important fraction of the variation seems to be related to genetic factors.

Lactation average SCS measures are used currently in some national genetic evaluation schemes, the objective being to lower the prevalence of mastitis by indirect selection for SCC. This approach does not use all the information and masks short-term variation in SCS (Shook and Schutz, 1994, Emanuelson, 1997). Test-day somatic cell records are taken on the same cow at various times, providing longitudinal information that can be related more meaningfully to episodes of infection. A statistical model for longitudinal data may provide more accurate estimates of the influence of risk factors than lactation average models. This additional information should be beneficial in the study of a trait with seemingly low heritability, such as SCS.

A reasonable model for describing SCS variation during lactation must consider the ‘typical’ lactation pattern and allow for differences associated with explanatory variables (e.g. age of cow) and with the individual cows themselves. Nonlinear mathematical functions are natural candidates for describing SCS lactation patterns, and their parameters can be modeled to account for known sources of variation. Estimates of these parameters can be used for management and breeding purposes to the extent that at least some of the variation detected is heritable.

The main objective of this study was to explore the feasibility of nonlinear mixed effects models to describe the SCS lactation curves. Another objective was to compare the fit of four nonlinear models and two linear models when applied to SCS lactation records in Holstein cows. An approximate maximum likelihood analysis was used for these purposes.

Section snippets

Data

The Ohio Agricultural Research and Development Center (Wooster) provided the SCC data used. After edits, the data consisted of 2387 test-day observations from 217 first to third lactation Holstein cows recorded between July 1982 and June 1989. The herd was maintained free of Streptococcus agalactiae, and less than 1% of the quarters were infected with Staphylococcus aureus at any one time (Todhunter et al., 1991). The bacteriological status was assessed by 0.01 ml of milk streaked onto the

Results

Mean 305-day, mature equivalent milk yield was about 7600 kg. Average SCC and standard deviation was 671,000 and 897,000 cells/ml, respectively. A few cows, with clinical mastitis symptoms, have very high levels of SCS, causing the average to be higher than the values usually associated to minor pathogens. As shown in Table 1, SCS decreased to a nadir at about 60 DIM and then increased, although not in a monotonic fashion, without regaining the initial level. The standard deviation followed an

Discussion

The overall trend of SCS during lactation is consistent with reports by Emanuelson and Philipsson (1984) and Wiggans and Shook (1987). The nonlinear (in time) SCS pattern observed could be explained by effects of stress at the onset of lactation and of dilution thereafter (Wiggans and Shook, 1987). Some of the models studied here have been used, with variable success, to describe milk yield. However, it is a challenge to find an appropriate model for SCC or SCS. Immunological responses may

Conclusion

Somatic cell score lactation patterns in dairy cows were described with nonlinear mixed effects models. Parameter estimates may be useful for selection schemes or management. In this study, four nonlinear and two linear mixed effects model were used. Within these, Morant and Gnanasakthy’s model provided the best fit. There was a significant association between intramammary infection and SCS trajectory. Cows calving between 1983 and 1985 had higher SCS levels and flatter curves, based on the

References (38)

  • E. Demidenko

    Asymptotic properties of nonlinear mixed-effects models

  • U. Emanuelson

    Clinical mastitis in the population, epidemiology and genetics

  • U. Emanuelson et al.

    Studies on somatic cell counts in milk from Swedish dairy cows. I. Non-genetic causes of variation in monthly test-day results

    Acta Agric. Scand.

    (1984)
  • U. Emanuelson et al.

    Studies on somatic cell counts in milk from Swedish dairy cows. II. Estimates of genetic parameters of monthly test-day results

    Acta Agric. Scand.

    (1984)
  • C.H. Gadini et al.

    Estimates of genetic parameters for first lactation test-day yields

    J. Dairy Sci.

    (1996)
  • D. Gianola et al.

    Prediction of breeding value in situations with nonlinear structure, categorical responses, growth functions and lactation curves

  • S.D. Kachman

    Prediction of Genetic Merit for Growth Curve Parameters in Outbred ICR Mice

    (1985)
  • B.W. Kennedy et al.

    Heritability of somatic cell count and its relationship with milk yield and composition in Holsteins

    J. Dairy Sci.

    (1982)
  • H.G. Monardes et al.

    Genetic and phenotypic statistics of lactation cell counts in different lactations of Holstein cows

    J. Dairy Sci.

    (1985)
  • Cited by (45)

    • Meta-analysis of genetic parameters for economic traits in buffaloes

      2021, Livestock Science
      Citation Excerpt :

      However, the use of these traits in a selection index to increase milk production would have a better response compared to the use of individual traits as a selection criteria (El-bramony et al., 2017). Somatic cell count in milk is an indicator of udder health and an indirect indicator of resistance to mastitis (Rodriguez-Zas et al., 2000). Cerón-Muñoz et al. (2002) reported that high somatic cell score values have negative effects on milk production and quality in buffaloes, and in the present study were estimated negative genetic correlation for MY x SCS, FY x SCS and PY x SCS.

    • Economic values for disease resistance traits in dairy goat production systems in Kenya

      2012, Small Ruminant Research
      Citation Excerpt :

      In genetic evaluation schemes however indirect selection is necessary in order to lower the prevalence of these two diseases. Mean somatic cell count (SCC, cells/μl) and faecal worm egg count (FEC, epg) can be taken as indicator traits for mastitis and helminthiosis, respectively (Baker et al., 1999; Rodriguez-Zas et al., 2000; Barillet et al., 2001; Barillet, 2007). The indicator traits SCC and FEC can be selected for in a breeding program without any detrimental effect on each other (Sechi et al., 2009).

    View all citing articles on Scopus
    View full text