Objective To investigate the effects of basic fibroblast growth factor (bFGF) on global cerebral ischemia/reperfusion (I/R) injury.Methods Twenty-four rabbits of both sexes aged 6 months weighing 2.2-3.0 kg were randomized into 3 groups ( n = 8 each) : group I sham operation; group Ⅱ I/R and group Ⅲ I/R + bFGF. The animals were anesthetized with diazepam, haloperidol and fentanyl. The right femoral artery was cannulated for BP monitoring and blood-letting. The animals were tracheotomized and mechanically ventilated (VT = 8 ml ?kg -1 , RR = 40 bpm, I: E = 1:2). PET CO2 was maintained at 40 mm Hg. Global cerebral I/R was induced by "occlusion of four vessels + hypotension" . Bilateral common carotid arteries and vertebral arteries were occluded for 30 min combined with hypotension, which was produced by blood-letting and maintained at 50%-60% of the baseline level. The four vessels were then released for reperfusion and BP was returned to baseline level by retrieval of the removed blood. In sham operation group ( Ⅰ ) the four blood vessels were exposed but not occluded. In group Ⅲ a loading dose of 30 ?g?kg-1 bFGF was given Ⅳ at the end of 30 min ischemia followed by bFGF infusion at 10 ?g? kg-1? h-1 for 6 h. In I/R group ( Ⅱ ) normal saline was given IV instead of bFGF. Blood samples were taken before I/R (T0 , baseline) and at 0, 0.5, 1, 3, 6 h (T1-5) of reperfusion for determination of serum neuron specific enolase (NSE) and S-100B concentrations. At the end of the experiment the animals were killed and brains removed for determination of brain water content and microscopic examination. Results Serum NSE and S-100B concentrations in group Ⅱ and Ⅲ were significantly increased at T4 5 as compared to the baseline (T0) and were significantly higher at T4 5 than those in sham operation group ( Ⅰ ) . Serum NSE and S-100B concentrations were significantly lower in group Ⅲ (I/R + bFGF) than in groupⅡ (I/R). Serum NSE and S-100B were positively correlated ( r = 0.736, P