On distinguishing cause and consequence: do high somatic cell counts lead to lower milk yield or does high milk yield lead to lower somatic cell count?
Researchers have reported that as milk yield increases composite milk somatic cell count (SCC) is diluted in cattle with no intramammary infection (IMI) and as a consequence, estimates of SCC from high yields are lower than estimates of SCC from low yields in dairy cows without an IMI. To date, estimates of reduced milk yield associated with high SCC because of intramammary infection have not been adjusted for any dilution of SCC. Ignoring dilution is therefore likely to lead to an overestimate of reduction in yield with increasing SCC. This paper investigates scenarios of the possible impact of dilution and inflammation on the association between somatic cell count and yield. The data used to investigate this relationship come from 8373 monthly records of milk yield and composite somatic cell count, together with incidence of clinical mastitis, which were recorded on 850 cows from five dairy cattle farms in Gloucestershire, UK. Two sets of models were used to investigate dilution and inflammation using two-level hierarchical models. The first set of models was used to estimate the linear (dilution) and log10-linear (inflammation) impact of SCC on the outcome variable milk yield. Five general linear models with increasing inclusion of higher test day SCC values were run. The cumulative categories were test day SCC values of up to and inclusive of 30, 50, 100, 200 and 400x10(3)cells/ml. Linear and log linear SCC influences on milk yield were estimated. At low SCC values the linear SCC predictor was dominant, while at higher values the log linear predictor was dominant. Up to 100x10(3)cells/ml there was mostly a slightly negative linear relationship between SCC and yield, potentially indicating a dilution effect. In the second set of models, three approaches to adjust milk loss for dilution were compared with an unadjusted model. In general, dilution-adjusted SCC values fitted the data better and resulted in a slightly lower milk loss per SCC category compared with unadjusted SCC. In all models with a dilution term there was a significant reduction in yield with SCC>200x10(3)cells/ml.