Serum urea concentrations were measured in Large White pigs from lines divergently selected for components of efficient lean growth rate and performance tested over three 14-d test periods starting at 30, 50, and 75 kg. Two methods of performance testing were used. Phase-fed pigs were fed to appetite isoenergetic diets differing in total lysine:energy ratio (0.58, 0.69, 0.81, 0.91, 1.01, 1.12, and 1.23 g/MJ of digestible energy), whereas diet-choice pigs were offered a choice of the 0.69 and 1.12 lysine:energy diets. Between test periods, all animals were fed one diet: 0.91 g of lysine/MJ of digestible energy. The study consisted of 230 boars and gilts with 150 pigs performance tested on phase-feeding and 80 pigs on diet-choice. The line selected for high lean food conversion had lower urea concentrations on each diet than the line selected for high lean growth rate, despite similar predicted lysine balances. Efficiency of lean growth rather than the rate of lean growth may be a better selection strategy in the context of nitrogen excretion. Urea concentrations at the end of each test period were correlated with lysine intake (0.33, 0.48 and 0.65; standard error, 0.08) and predicted lysine balance (0.39,0.44, and 0.64), but were uncorrelated with predicted lysine for protein deposition (0.01, 0.08, and 0.08) and maintenance. Urea concentration at the end of a test period was not a useful predictor of protein deposition, even after accounting for pretest variation in urea concentration and food intake during test. The expected response pattern of serum urea concentration to diets differing in total lysine:energy would be nonlinear, with the point of inflection occurring at the required dietary total lysine:energy for each genotype. However, there was no evidence of such an inflection point such that the prediction of lysine requirement from urea concentration was not possible for the selection lines in the study.