Thirty-six primiparous sows were assigned to one of five treatments designed to mimic the lactation feed intake patterns observed in a previous study of commercial farms: high energy intake (HHH) or low energy intake during each week of a 3-wk lactation (LLL), or reduced intake during wk 1 (LHH), wk 2 (HLH) or wk 3 of lactation (HHL). The metabolizable energy intake of sows was either 16.5 (H) or 6.5 (L) Mcal/d. Diets were equal in lysine, providing 45 g/d. The HHH sows had a shorter (P < .05) weaning-to-estrus interval (9 +/- 3.2 d) than the sows in the LLL (23 +/- 3.5 d), HLH (22 +/- 3.5 d), and HHL (18 +/- 3.2 d) groups. The LH pulse frequencies of HHH (d 14: 1.5 pulses/8 h; d 21: 2.1 pulses/8 h) were greater (P < .05) than those of LLL (.2 pulses/8 h) and HLH (.5 pulses/8 h) on d 14 and those of LLL (0 pulses/8 h) and HHL (.9 pulses/8 h) on d 21 of lactation. No differences (P > .1) in mean serum concentrations of LH were observed among treatment groups either on d 21 of lactation or on d 1 postweaning. When fed a low-energy diet, sows had lower (P < .05) concentrations of serum insulin and plasma glucose than did sows fed a high-energy diet. Our results indicate that energy intake during lactation influences circulating insulin and glucose levels and LH pulse frequency and amplitude during midlactation and during the postweaning period.