Feed efficiency is a major goal for ecological and economical aquaculture durability, and selecting fish to genetically enhance this parameter could be valuable. However, no selective breeding enhancing feed efficiency exists, probably because of the difficulty to measure individual feed intake, particularly for recently domesticated species. A negative correlation between feed efficiency and weight loss at fasting demonstrated in rainbow trout and in sea bass may help solve this problem. We mated sea bass parents selected for their high or low weight loss at fasting, obtaining a FD+ line and a FD- line, respectively. As expected, FD- fish lost less weight than FD+ fish at feed deprivation. A better feed efficiency was also expected for the FD- fish, but there were not stable significant differences between the two groups for this trait, measured either with Feed Efficiency Ratio (FER), or with Residual Feed Intake (RFI). However, the FD- fish had a genetic overall better growth capacity, which could eventually be due to a better stress resistance. Moreover, FD- fish were fatter whereas FD+ fish had a better carcass yield. So, the two lines are clearly divergent for several characters, but our initial objective was not reached. A second generation of selection may be needed to observe feed efficiency differences on more divergent lines.