Residual carbon dioxide as an index of feed efficiency in lactating dairy cows.

High feed costs make feed conversion efficiency a desirable target for genetic improvement. Residual feed intake (RFI), calculated as the difference between observed and predicted intake, is a commonly used estimate of feed efficiency. However, determination of feed efficiency in dairy herds is challenging due to difficulties in measuring feed intake of individual animals reliably. Using residual CO2 (RCO2) production as an estimate of feed efficiency would allow ranking the cows according to feed efficiency, provided that CO2 production is closely related to heat production and feed intake. The objective of this study was to evaluate the potential of RCO2 as an index of feed efficiency using data from respiration calorimetry studies (289 cow per period observations). Heat production was precisely predicted from CO2 production [root mean square error (RMSE)] adjusted for random effects was 1.5% of observed mean]. Dry matter intake (DMI) was better predicted from energy-corrected milk (ECM) yield and CO2 production than from ECM yield and body weight in the model (adjusted RSME = 0.92 vs. 1.39 kg/d). Residual CO2 production estimated as the difference between actual CO2 production and that predicted from ECM yield, metabolic body weight was closely related to RFI (adjusted RMSE = 0.42) that was calculated as the difference between actual DMI and that predicted from ECM, metabolic body weight, and energy balance (EB). When the cows were categorized in 3 groups of equal sizes on the basis of RCO2 (low, medium, and high), low RCO2 cows had lower DMI, RFI, methane production and intensity (g/kg ECM), and heat production, but higher efficiency of metabolizable energy utilization for lactation than high RCO2 cows. When RFI was predicted from RCO2, the residuals (observed - predicted) were negatively related to EB and digestibility. Predicting RFI with a 2-variable model based on RCO2 and digestibility, adjusted RMSE decreased to 0.23 kg/d, and residuals were not significantly related to EB. The cows in low RCO2 group had a higher energy digestibility than the cows in the high RCO2 group, and differences in EB were observed between the groups. Error of the model predicting residual ECM production from RCO2 was 1.41 kg/d. The residuals were positively related to ECM yield and energy digestibility. Predicting residual ECM from RCO2 and ECM yield decreased adjusted RMSE to 1.07 kg/d, and further to 0.78 kg/d when digestibility was included in the 2-variable model. It is concluded that RCO2 has a potential for ranking individual cows based on feed efficiency.