The effect of caffeine on the exercise responses of six women habituated to caffeine (greater than 600 mg/day) was examined during 1-h running at 75% VO2 max on a motorized treadmill. Each subject completed a placebo (PL) and a caffeine ingestion (CC) trial while maintaining normal caffeine intake. The subject then abstained from caffeine for 4 days and again ran after receiving caffeine (CW). Caffeine dosage for all trials was 5 mg/kg body weight. Ingestion of caffeine after withdrawal (CW) resulted in the greatest physiologic effects. Exercise oxygen uptake was significantly elevated by 0.17 l/min over the PL and CC trials (P less than 0.05). The CW trials resulted in an overall R value of 0.79 +/- 0.04 compared with 0.85 +/- 0.08 for the PL and 0.83 +/- 0.04 for the CC trials. Caffeine had its greatest effect on the resting free fatty acid levels after withdrawal: 1104 +/- 425 mu Eq/l compared with 543 +/- 288 for the PL and 839 +/- 526 for the CC. Postexercise lactates were similar for all trials. Post-exercise plasma norepinephrine and dopamine were the highest after the CW trials. The results suggest that habitually high caffeine users acquire a tolerance to caffeine which reduces its effects during prolonged exercise. Furthermore, to magnify the effect of caffeine, habitual users should withdraw from caffeine use for about 4 days.