It is not known whether warm-up protocols typically employed by athletes are beneficial to performance. The aim of this study was to investigate the effect of warm-up on a perimaximal run to exhaustion, VO2 kinetics, energy metabolism and running economy. Nine male distance runners ran to exhaustion at a speed corresponding to 105% maximal oxygen uptake (VO2max) after each of three different warm-up protocols: no warm-up (NW), jog warm-up (WM), or jog with strides (WH). Warm-up did not affect pre-run blood lactate concentration (BLC), BLC-increase (DeltaBLC), net oxygen consumption, or running economy. WH increased the amplitude of the primary VO2 response (mean (confidence intervals of difference); 4083 ml min(-1) vs. 3763 ml min(-1) (-638, -2)), with no change in the time constant; reduced the rate of BLC-increase (DeltaBLC-rate) (0.02 mmol l(-1) s(-1) vs. 0.03 mmol l(-1) s(-1) (0.003, 0.01)); reduced anaerobic lactic power (109 W vs. 141 W (13, 51)); reduced the relative anaerobic lactic energy contribution (7.0% vs. 9.1% (0.8, 3.4)) compared to NW. The reduction in anaerobic power associated with DeltaBLC-rate between NW and WH was significantly correlated with the increase in aerobic power associated with the primary amplitude (r=0.674, p<0.05). Despite these theoretically beneficial metabolic effects, WH did not significantly increase time to exhaustion vs. NW (290 s vs.316 s (-77, 43)) but might be considered in events where the winning margin is often small.