Abstract Animals with superior feed efficiency (FE) may also have an improved nitrogen use efficiency (NUE), which would be beneficial to economic profitability while reducing environmental impacts. When genetically selecting animals on FE, it is preferable to use residual traits [e.g. residual feed intake (RFI) or residual body gain (RG)] rather than ratios because of their predictable genetic outcomes. We studied the relationship of RFI and RG with NUE, estimated from the validated 15N abundance in plasma, across two contrasted diets based on corn or grass silages. We evaluated FE of 588 (half by diet) Charolais bulls (545 ± 57 kg BW) from 12 experimental cohorts (different farms and periods) over 200 days. Before the end of the FE test, plasma was sampled and analyzed for δ 15N. NUE was related to FE through simple-linear models with variables previously corrected for the cohort and diet effects. The models’ slopes were standardized according to FE deviation in order to compare the response of NUE to FE between indices. Higher NUE was related to higher FE (P < 0.001), showing positive correlations with RG (r=-0.40) and negative with RFI (r=0.29). However, the standardized slope of NUE to RG was significantly higher (+28%; P < 0.05) than that of NUE to RFI. This stronger NUE relation to RG compared to RFI could reflect a higher potential of RG animals to deposit N as compared to a more conservational N metabolism in RFI individuals. Regarding diets, and despite the correction of NUE and FE for this effect, the slopes of NUE to FE were numerically (P > 0.05) higher (-16% and +36%; for RG and RFI) in corn-based diets, which agrees with superior NUE observed in corn-vs-grass diets. Results suggested that superior RG animals may present proportionally higher NUE than superior RFI animals, with even better results in corn-vs-grass diets.