Gastrointestinal (GI) complaints are frequently experienced during running. Sports drinks to prevent dehydration and hypoglycemia during exercise are generally used. The aim was to investigate the effect of 3 different drinks on GI complaints and performance during competitive running in a controlled field study. Ninety-eight well-trained subjects (90 M, 8 F, age 41 +/- 8 y) performed a competitive 18-km run three times within 8 days. The study was a controlled, standardized field experiment following a randomized, crossover design. Three different drinks were compared: water, a sports drink (CES), and a sports drink with added 150 mg/l caffeine (CAF). The incidence of GI complaints and the effect of the drinks on performance was studied. Each subject consumed 4 times 150 ml as follows: at the start, after 4.5 km, 9 km, and 13.5 km. Fluid intake was controlled. Incidence and intensity of GI complaints during the run were determined using a 10 points scale questionnaire. There were no significant differences in performance between the 3 drinks. Run time (18 km, mean +/- SD): WAT 1 : 18 : 03 +/- 08 : 30, CES 1 : 18 : 23 +/- 08 : 47, CAF 1 : 18 : 03 +/- 08 : 42. The use of carbohydrate-containing sports drinks led to higher incidences of all types of GI complaints compared to water. Significant differences (p < 0.05) were reached for flatulence; incidence: WAT 17.9 %, CES 28.6 %, CAF 30.6 %, and reflux; incidence: WAT 55.7 %, CES 78.6 %, CAF 72.5 %. There were no significant differences in intensity of the GI complaints. Addition of caffeine to CES had no effect on GI complaints, compared to CES alone. We conclude that sports drinks used during an 18-km run in cool environmental conditions do not support the performance better than mineral water. The use of sports drinks during an 18-km run leads to a higher incidence of both upper and lower GI complaints compared to water. Addition of caffeine to the sports drink has no effect on either running performance or GI complaints.