Aims To compare the effect of nitroprusside (SNP) and nitroglycerin (NTG) on angiotensin II (ANGII), endothelin-1 (ET-1), and a 1 -adrenoceptor ( phenylephrine, PE)-mediated contraction in internal mammary artery (IMA). Methods Human IMA segments (n=120) taken from 37 patients were studied. Concentration-relaxation curves for SNP and NTG were established in IMA precontracted with these vasoconstrictors. Concentration-contraction curves were also constructed in IMA rings incubated with SNP and NTG (0.1 and 1 mm) for 10 min. Results Both SNP and NTG caused full relaxation with similar EC 50 s except NTG was four-fold more potent than SNP in PE-induced contraction (−7.92±0.06 vs −7.32±0.2 log m, mean±s.e. mean, P<0.01; 95% confidence interval for the difference of the means: 0.19, 1.01 log m). Pretreatment with SNP (0.1 and 1 mm) significantly depressed the contraction by ANGII from 56.6±7.7% (of 100 mm K + -contraction) to 18.3±8.6% and 3.9±2.1% (P=0.0001). In four rings treated with SNP, the contraction to ANGII was abolished whereas NTG did not depress ANGII-mediated contraction. Pretreatment with SNP (1 mm), but not NTG, significantly depressed the magnitude of the PE-induced contraction from 4.7±1.2 to 1.7±0.4 g ( P<0.05). Treatment with both SNP and NTG significantly increased the EC 50 (−5.09±0.17 log m, P=0.0007 for SNP and −5.40±0.06 log M, P= 0.02 for NTG). Pretreatment with SNP did not significantly change either the magnitude or the EC 50 of the ET-1-induced contraction. Conclusions SNP may be advantageous compared with NTG in preventing coronary arterial graft contraction. However, once grafts have constricted to ANGII, a 1 -adrenoceptor agonists, and ET-1, NTG may be only marginally advantageous.