We aimed to establish a model for prediction of iCa from tCa, using multivariable regressions with diverse blood constituents. Blood was taken from 14 cows at days −2, 0, 2, 4, 7, and 14 relative to parturition. Cows were clinically healthy, and no hypocalcaemia prophylaxis and treatment were applied. Total calcium and further parameters were determined from frozen serum. Ionized calcium, blood gases, and electrolytes were determined from heparin-stabilized blood samples. Linear regression between iCa and tCa was estimated. Precision improved only slightly using a multivariable model. Best precision was achieved when estimating the iCa:tCa ratio from other blood constituents. To identify the reason behind the poorly predictive value of tCa for iCa, the relative changes of iCa and tCa around calving were calibrated to the respective values of day −2 (=100%) for each cow. An increase in the iCa:tCa ratio was observed from 0.43 at day −2 to 0.48 at day 0, followed by a gradual decrease towards 0.43 at day 7. We conclude that routine measurement of iCa should be implemented in the diagnosis of hypocalcaemia. An optimized estimate of iCa from tCa with non-esterified fatty acids (NEFA), beta-hydroxybutyric acid, cholesterol, and phosphorous as co-predictors is still poorly satisfying.
