The purpose of this study was to compare the efficacy of diazepam and the pro-diazepam avizafone in preventing the severity of soman-induced pathology in guinea pig. Survival, respiration and seizures of experimental animals were investigated with on-line monitoring of respiratory and EEG parameters. Guinea pigs were pretreated with pyridostigmine (0.1mg/kg i.m.) and 30 min later challenged with 1 or 2 LD50 soman. One minute after intoxication they were treated with atropine (3 or 33.8 mg/kg), pralidoxime chloride (32 mg/kg) and either diazepam (2 mg/kg), avizafone (3.5 mg/kg) or saline solution. The highest dose of atropine (33.8 mg/kg) gave a protective effect in groups treated without anticonvulsants by reducing the severity of clinical signs and death within 24 h but also by decreasing seizure occurrence and brain injuries. When injected at the similar molar dose of 7 micromoles/kg, the protection of anticonvulsants against soman neurotoxicity was higher with the atropine/pralidoxime/avizafone combination than with atropine/pralidoxime/diazepam. Indeed, when atropine was used at the lowest dose, avizafone was found to prevent early mortality and seizures occurrence with better efficacy than diazepam. On the other hand, when added to the therapy, the both anticonvulsants did not prevent the moderate EEG depression (reduction of amplitude by 30-52%) observed under 2 LD50 soman. Moreover, the number of animals suffering from respiratory distress (defined as a decrease of minute ventilation of more than 20% from the baseline value) was enhanced when diazepam or avizafone were used in the therapy. This effect was dependent on the atropine dose and the nature of the anticonvulsant. The beneficial effects of the different therapeutics tested were assessed and compared to the previous data obtained with the same therapies against sarin and from the pharmacokinetics properties of the atropine/diazepam mixture.