Gastrointestinal bleeding and increased intestinal permeability have been observed in marathon runners. We sought to determine if L-arginine would be useful for prevention of these complications. Twenty-three runners were randomized to receive L-arginine (A) or glycine (placebo) (G), 10 grams 3 times daily for 14 days prior to the 1997 Houston-Methodist Marathon. Serum, stool hemoccults and lactulose:mannitol permeabilities were obtained at baseline, immediately after completion of the marathon and approximately 48 hours later. Runners rated their symptoms of nausea and vomiting, belching and indigestion, abdominal pain and bloating, diarrhea, and extremity pain on a 1-5 scale of increasing severity. The L:M was unchanged in either group during the three collections. Occult bleeding occurred in 8%/20% in A and G groups, respectively, p = NS) immediately post-marathon. No runners had occult bleeding 48 hours post-race. Gastrointestinal symptom scores were minimal to nonexistent. Extremity pain scores were similar for groups A and G (2.1+/-1.4 and 2.8+/-1.6, respectively, (p = NS). Fluid intake was similar between both groups (1875+/-1547 vs. 1506+/-970 ml, p = NS). Serum amylase was normal at baseline and remained virtually unchanged. Serum lipase was normal at baseline and immediately post-race in both groups, but increased at 48 hours post-race (82.2+/-34.3 to 121.5+/-53.3 mg/dl [A], p = 0.02 and 114.3+/-55.7 to 181.9+/-162.2 mg/dl [G], p = 0.09). CPK increased significantly and similarly in both groups immediately post-race, and even more dramatically 48 hours post-race (130.3+/-130.8 to 738.8+/-902.9, p = 0.007 to 1966.5+/-3.166.0 mg/dl [A] and 140.9+/-77.9 to 863.0+/-772.3, p = 0.003 to 5619+/-10636.8mg/dl [G]). Modest post-race decreases were seen in most serum amino acids in both groups. Finish times were longer than predicted (23+/-21 and 9+/-7 min for A and G groups, respectively, p = 0.049). Our study failed to show a clear benefit of arginine supplementation for the prevention of intestinal ischemia/reperfusion injury associated with endurance running, but either a detrimental affect on performance with arginine, or enhanced performance with glycine. Skeletal muscle injury was unaffected by arginine or glycine supplementation. The delayed increase in serum lipase suggests mild pancreatic injury, affected by either arginine or glycine supplementation.