Prediction of urinary nitrogen and urinary urea nitrogen excretion by lactating dairy cattle in northwestern Europe and North America: a meta-analysis.

A meta-analysis was conducted on the effect of dietary and animal factors on the excretion of total urinary nitrogen (UN) and urinary urea nitrogen (UUN) in lactating dairy cattle in North America (NA) and northwestern Europe (EU). Mean treatment data were used from 47 trials carried out in NA and EU. Mixed model analysis was used with experiment included as a random effect and all other factors, consisting of dietary and animal characteristics, included as fixed effects. Fixed factors were nested within continent (EU or NA). A distinction was made between urinary excretions based on either urine spot samples or calculated assuming a zero N balance, and excretions that were determined by total collection of urine only. Moreover, with the subset of data based on total collection of urine, a new data set was created by calculating urinary N excretion assuming a zero N balance. Comparison with the original subset of data allowed for examining the effect of such an assumption on the relationship established between milk urea N (MUN) concentration and UN. Of all single dietary and animal factors evaluated to predict N excretion in urine, MUN and dietary crude protein (CP) concentration were by far the best predictors. Urinary N excretion was best predicted by the combination of MUN, CP, and dry matter intake, whereas UUN was best predicted by the combination of MUN and CP. All other factors did not improve or only marginally improved the prediction of UN or UUN. The relationship between UN and MUN differed between NA and EU, with higher estimated regression coefficients for MUN for the NA data set. Precision of UN and UUN prediction improved substantially when only UN or UUN data based on total collection of urine were used. The relationship between UN and MUN for the NA data set, but not for the EU data set, was substantially altered when UN was calculated assuming a zero N balance instead of being based on the total collection of urine. According to results of the present meta-analysis, UN and UUN are best predicted by the combination of MUN and CP and that, in regard to precision and accuracy, prediction equations for UN and UUN should be derived from the total collection of urine.