L-arginine supplementation in sepsis is controversial. Septic shock has been alternatively viewed as an L-arginine-deficient state or as a syndrome caused by excess nitric oxide, an end-product of L-arginine metabolism. Randomized, placebo-controlled, and double-blinded (investigators, veterinarians, and pharmacists). Laboratory. Purpose-bred, 1- to 2-yr-old, 10- to 12-kg beagles. The effects of parenteral L-arginine alone or in combination with N-acetylcysteine were compared with vehicle alone in a well-characterized canine model of Escherichia coli peritonitis. Two doses were studied that delivered approximately 1.5-fold (10 mg x kg(-1) x hr(-1)) and 15-fold (100 mg x kg(-1) x hr(-1)) the L-arginine dose typically administered with standard total parenteral nutrition. Animals in the low- and high-dose L-arginine arms were further randomized to receive vehicle alone or N-acetylcysteine (20 mg x kg(-1) x hr(-1)) as an antioxidant to prevent peroxynitrite formation. The main measurements were hemodynamics, plasma arginine and ornithine, serum nitrate/nitrite, laboratory studies for organ injury, and survival. Both doses of L-arginine similarly increased mortality (p = .02), and worsened shock (p = .001 for reduced mean arterial pressure). These effects were associated with significant increases in plasma arginine (p = .0013) and ornithine (p = .0021). In addition, serum nitrate/nitrite (p = .02), liver enzymes (p = .08), and blood urea nitrogen/creatinine ratios (p = .001) rose, whereas arterial pH (p = .001) and bicarbonate levels (p = .001) fell. N-acetylcysteine did not significantly decrease any of the harmful effects of L-arginine. Thus, parenteral L-arginine monotherapy was markedly harmful in animals with septic shock. These findings suggest that supplemental parenteral L-arginine, at doses above standard dietary practices, should be avoided in critically ill patients with septic shock.