Breeding entire males is an alternative to surgical castration to improve their welfare. However, entire males may have a major quality defect called boar taint. Boar taint is partly due to the presence of androstenone in fat. In this study, we estimated the genetic parameters between androstenone and production traits to evaluate the consequences of selection against boar taint for traits of interest. We focused on growth traits, meat quality, lesions, hormone levels and computerised tomography measurements in purebred Piétrain (P) or Piétrain cross Large White (X) entire males. The number of measured animals varied from 670 P and 734 X for hormones concentrations to 553 P and 645 X for computerised tomography measurements. Skin lesions were measured on live pigs shortly after mixing, at the end of the fattening period, and on carcasses. Heritabilities of traits measured by tomography ranged from low to high: femur density (P: 0.34, X: 0.69), loin eye area (P: 0.53, X: 0.88) and loin eye density (P: 0.12, X: 0.18). The mean number of lesions at each stage was lower in purebred pigs than in crossbreds (entering the fattening stage 4.01 in P and 4.68 in X; before slaughter 3.72 in P and 4.22 in X; on carcass 4.50 in P and 4.96 in X). We also observed a decrease in the average number of lesions between the two stages in live pigs. We found high genetic correlations between stages in purebred pigs (0.74 to 0.76) but low correlations (−0.30 to 0.29) in crossbred pigs. Selection aiming to decrease fat androstenone is feasible ( h2 = 0.57 in P and h2 = 0.71 in X). It would have overall positive effects on meat production and quality traits. Selection aiming to reduce plasma oestradiol would strongly reduce the level of fat androstenone (rg = 0.89 in P and rg = 0.84 in X). Selection against oestradiol is easier and less invasive since it would only require a blood sample rather than a fat biopsy in live animals.