This study investigated the effect of three dietary regimes for replacement gilts on lameness and carcass traits. Diets were: a diet specifically formulated for replacement gilts (diet 1, 14.0MJ of DE/kg, 0.75% lysine), a finisher diet (diet 2, 13.5MJ of DE/kg, 1.02% lysine) and a gestation sow diet (diet 3, 13.0MJ of DE/kg, 0.69% lysine): the latter two diets are traditionally fed to replacement gilts. Thirty-six gilts were selected at day (d)0 (70.8kg±0.78 SE, aged ~130d), housed individually and allocated at random to 1) DEV (restricted access diet 1, n=12), 2) FIN (ad-libitum access diet 2, n=12) or 3) GES (initially ad-libitum access diet 2, then restricted access to diet 3 from d30, n=12). All gilts were fed ad-libitum from d71–83. Locomotory ability (0=normal to 5=severely impaired) and limb lesions (0=normal to 3=severe) were scored weekly until d82. Hind-claw lesions and toe uneveness were scored (0=normal to 3=severe) at d0, 40 and 82. Gilts were weighed at d0, 28, 70 and 82. Carcass traits were recorded at slaughter (d83) and one front leg was removed for dual-energy X-ray absorptiometry (DXA) to establish areal bone mineral density (aBMD). Joint surface lesions of the humeral condyle (HC; 1=normal to 4=severe, 5=osteochondrosis dissecans) and anconeal process (AP; 1=absent 2=present) were scored. The percentage of lame animals (locomotion score ≥2) on ≥1 occasion were DEV=0% (0/11), FIN=73% (8/11) and GES=75% (9/12) (P<0.01). Fewer DEV gilts than FIN and GES gilts had humeral condyle lesions: DEV=64% (7/11), FIN=100% (11/11) and GES=100% (12/12) (P<0.01). DEV gilts had lower scores for humeral condyle lesions (median; IQR: 2; 2) than GES (4; 1) and FIN gilts (4; 3) (P=0.05). DEV gilts weighed less (132.5kg±2.8) than FIN (142.9±2.0) gilts at d70 (P<0.05). DEV gilts had a lower average daily gain (ADG) than FIN (P<0.05) and GES (P<0.001) gilts at d0–28, lower energy intake than FIN and GES gilts from d0 to 28 (P<0.001) and higher energy intake than FIN gilts at d71–82 (P<0.05). The DEV diet formulated for replacement gilts and fed restrictively from 70kg to until ~130kg, then fed ad-libitum until ~140kg improved locomotory ability, toe eveness and humeral condyle joint lesions compared with traditional diets. It is likely these improvements were related to slower growth rates.