Evaluation of a creatinine-based equation to predict urine volume in nonpregnant, nonlactating Holstein cows

Accurate estimation of urine output is essential for assessing nutrient utilization, particularly for nutrients predominantly excreted through urine, such as minerals and protein-derived metabolites. An equation utilizing creatinine concentration and BW has been previously developed to estimate urinary volume (EUV) in lactating Holstein dairy cows, defined as EUV = [29 × BW (kg)]/[urinary creatinine (mg/L)]. Our objective was to evaluate whether this equation introduces bias when used for estimating urinary output in nonpregnant, nonlactating Holstein cows and to identify factors influencing its accuracy. For our study, we used 72 paired observations that included observed urinary volume (OUV) and urinary creatinine concentration. These observations were obtained from a prior research study that assessed the relative availability of various magnesium (Mg) sources using a duplicated 6 × 6 Latin square design, with cows (n = 12) grouped into squares based on lower (square 1) or higher (square 2) BW across 6 periods. Enrolled cows were of second parity (n = 8) and third or greater parity (n = 4), with BW ranging from 590 to 831 kg. To determine the agreement between EUV and OUV, we constructed 2 mixed-effects models. The first model evaluated slope bias (testing if the slope was significantly different from 1), whereas the second assessed mean bias (testing if the intercept was significantly different from 0) between EUV and mean-centered OUV. In our dataset, creatinine excretion per kilogram of BW ranged from 15.0 to 35.6 mg/kg BW with an average of 27.6 mg/kg BW. When assessing the agreement between OUV and EUV, we observed both slope and mean biases when applying the creatinine-based equation. Furthermore, there was a bias estimate across BW quartiles. Overall, 76.4% of observations fell within ±10% deviation range between EUV and OUV. These findings suggest that further research is needed to identify factors that can refine the creatinine and BW-based predictive equation specifically for nonpregnant, nonlactating Holstein dairy cows.