Abstract Background Batroxobin has been found to have protective effect on cerebral ischemia-reperfusion, and cardiopulmonary resuscitation (CPR) is the common cause of global brain ischemia-reperfusion. Objective To observe the effect of Batroxobin on the morphological results of cerebral cortex and hippocampus in rabbit models of CPR, and the changes of serum concentration of tumor necrosis factor alpha (TNF-a) after CPR. Design A randomized controlled observation. Setting Laboratory of the Department of Burns, Changhai Hospital affiliated to the Second Military Medical University of Chinese PLA. Materials Thirty healthy New Zealand rabbits of 2.5-3.0 kg, either male or female, were used. Kits for TNF-a determination were provided by LIFEKEY BioMeditech Company (USA). Methods The experiments were carried out in the laboratory of Department of Burns, Changhai Hospital from February 2001 to January 2002. The 32 rabbits were randomly divided into sham-operated group ( n=8), conventional resuscitation group ( n=12) and Batroxobin-treated group ( n=12). The animals in the conventional resuscitation group and Batroxobin-treated group were anesthetized, then induced into modified Pittsburg's model of mechanical ventricular fibrillation. Sham-operated group was discharged on the chest wall, which did not cause ventricular fibrillation. Conventional resuscitation group and Batroxobin-treated group were exposed to 6 minutes of cardiac arrest induced by ventricular fibrillation, then the resuscitation began. A dosage of 0.3 Bu/kg of Batroxobin was administered to the rabbits in the Batroxobin-treated group at the beginning of resuscitation. Blood sample was collected at 4 and 12 hours after CPR to determine the concentration of TNF-a in serum. After the second blood collection, brain tissue was taken out immediately, and the forms of nerve cells in cerebral cortex and hippocampal CA1 region were observed under light microscope. Main outcome measures TNF-aconcentration in serum at 4 and 12 hours after CPR; Forms of nerve cells in cerebral cortex and hippocampal CA1 region at 12 hours after CPR. Results All the 31 New Zealand rabbits were involved in the analysis of results. TNF-a concentration in serum: At 4 hours after CPR, the TNF-a concentrations in serum in the conventional resuscitation group and Batroxobin-treated group [(5.947±2.366), (5.122±2.521) ng/L] were significantly higher than that in the sham-operated group [(2.604±1.623) ng/L, P < 0.05]. At 12 hours after CPR, the TNF-a concentration in serum in the conventional resuscitation group was (7.770±3.121) ng/L, it was significantly higher than that at 4 hours ( P < 0.05), also significantly higher than that in the Batroxobin-treated group [(5.425±2.280) ng/L, P < 0.05]. Forms of nerve cells: In the sham-operated group, no abnormality was found in the hippocampal CA1 region and cerebral cortex. In the conventional resuscitation group, the pyramidal cells in hippocampal CA1 region were lined up in disorders, and edema, puff, vacuolization, nucleus concentration and anachromasis were also observed appeared; Edema of nerve cells, vacuole, pyknosis appeared in cerebral cortex; microthrombosis appeared in some blood capillaries. As compared with the conventional resuscitation group, cellular edema was relieved and pyknosis of nerve cells were obviously reduced, and no microthrombosis was found in hippocampal CA1 region and cerebral cortex in the Batroxobin-treated group. Conclusion Batroxobin have neuroprotective effect on CPR rabbits, and may inhibit the excessive increase of TNF-a concentration in serum.