Effects of nitrogen supplementation associated with different starch proportions on nutritional performance and metabolic characteristics of cattle fed low-quality tropical forage (Brachiaria decumbens hay, 7.4% CP) were evaluated using ruminal and abomasal cannulated steers. Five European × Zebu young bulls (186 kg BW) were distributed according to a 5 × 5 Latin square. The following treatments were evaluated: control (only forage), supplementation with 300 g CP/d (0:1), supplementation with 300 g starch/d and 300 g CP/d (1:1), supplementation with 600 g starch/d and 300 g CP/d (2:1), and supplementation with 900 g starch/d and 300 g CP/d (3:1). Supplements increased (P < 0.05) DM intake, but did not affect (P > 0.05) forage intake. There was a cubic effect (P < 0.05) of starch, which was attributed to the highest forage intake (g/kg BW) when using the 2:1 starch:CP ratio. Supplements increased (P < 0.05) OM digestibility, but did not affect (P > 0.05) NDF digestibility. There was a positive linear effect (P < 0.05) of the amount of starch supplemented on OM digestibility. Total NDF digestibility was not affected (P > 0.05) by the amount of starch supplemented. Ruminal ammonia nitrogen concentrations were higher (P < 0.05) in supplemented animals, however, a negative linear effect (P < 0.05) of amount of starch was observed. Supplements increased (P < 0.05) the nitrogen balance (NB) and efficiency of nitrogen utilization. These effects were attributed to increased body anabolism, supported by higher (P < 0.05) serum concentration of IGF1. Increasing the amount of starch tended to linearly increase the NB (P < 0.06) and serum concentration of IGF1 (P < 0.05). However, comparison of NB means between treatments with supplementation showed a higher value for the 2:1 starch:CP ratio. Nitrogen supplementation in cattle fed low-quality tropical forage increases nitrogen retention in the animal body. Additional supply of starch increases nitrogen retention by increasing energy availability for both rumen and animal metabolism.