BACKGROUND: Morphine can inhibit inflammatory edema in experimental animals. The mechanisms and sites by which opioids exert this effect are still under debate. Since the spinal level is a site for modulation of the neurogenic component of inflammation, we investigated the effect of intrathecal (IT) administration of morphine, and the involvement of spinal nitric oxide (NO)/cyclic-guanosine monophosphate-GMP pathway in carrageenan (CG)-induced paw edema. METHODS: Male Wistar rats received IT injections of drugs (20 mu L) 30 min before paw stimulation with CG (150 mu g). Edema was measured as paw volume increase (mL), and neutrophil migration was evaluated indirectly by myeloperoxidase (MPO) assay. RESULTS: Morphine (37, 75, and 150 nmol) inhibited inflammatory edema, but had no effect on MPO activity. Coinjection with naloxone (64 nmol) reversed the effect of morphine. The corticosteroid synthesis inhibitor, aminoglutethimide (50 mg/kg, v.o.), administered 90 min before morphine injection did not modify its antiedematogenic effect. Low doses of the NO synthase inhibitor, N-omega-nitro-L-arginine (L-NNA; 10 and 30 pmol) increased, while higher doses (3 and 30 nmol) inhibited edema. The guanylate cyclase inhibitor 1H-oxadiazolo[4,3-a]quinoxalin-1-one (ODQ; 21 and 42 nmol) increased, while the phosphodiesterase type 5 inhibitor sildenafil (0.15 and 1.5 nmol) inhibited paw edema. Coadministration of a subeffective dose of L-NNA (3 pmol) or ODQ (10 nmol) with morphine prevented its antiedematogenic effect, but sildenafil (0.15 nmol) rendered a subeffective dose of morphine effective (18 nmol). ODQ also prevented the antiedernatogenic effect of the NO donor S-nitroso-N-acethyl-penicilamine. CONCLUSION: These results support the idea that morphine can act on opioid receptors at the spinal level to produce antiedematogenic, and that the NO/cGMP pathway seems to be an important mediator in this effect.