Feeding a high-carbohydrate (CHO) diet and administration of alkalinizing agents have both been shown to improve performance in high-intensity exercise. The effect of these treatments in combination was investigated in the present study. Six healthy male subjects exercised to exhaustion on an electrically braked cycle ergometer at a power output equivalent to 100 % of their maximum oxygen uptake (VO2,max) on four separate occasions. Each subject consumed either a diet with the same composition as his normal diet (termed the experimental normal (N) diet; 54 ± 7 % CHO, 13 ± 2 % protein, 33 ± 7 % fat) or a high-CHO diet (81 ± 2 % CHO, 13 ± 2 % protein, 6 ± 1 % fat) that had the same energy and protein content for the 3 days prior to the exercise tests. Subjects then ingested either a placebo (CaCO3) or trisodium citrate (0·3 g (kg body mass)-1) 3 h before exercise. Time to fatigue was not different between experimental conditions. Consumption of the high-CHO diet had no effect on blood acid-base status, but the ingestion of sodium citrate induced a mild metabolic alkalosis after both the N diet and the high-CHO diet. This alkalinizing effect was also evident after exercise, since blood pH, plasma bicarbonate and blood base excess were higher (P < 0·05) after the ingestion of sodium citrate than under the placebo conditions. The changes in blood lactate, pyruvate and glucose and plasma glycerol after exercise were similar for all experimental conditions. Blood lactate, glucose and pyruvate and plasma glycerol concentrations increased from resting values (P < 0·01) following exercise but this increase was similar under all experimental conditions. These data demonstrate that when the energy and protein content of the diets is the same, exercise capacity and the metabolic response to intense exercise are similar following consumption either of a high-CHO diet or a more normal diet. Acute ingestion of sodium citrate prior to exercise resulted in a reduction in post-exercise acidosis despite a blood lactate concentration that was similar to that observed after the ingestion of a placebo, but did not affect exercise performance under the conditions of this study.