The data emerging from our study are the following: the presence of an identifiable cause is important: complications like tuberous sclerosis or signs of marked cerebral damage represent an adverse risk factor for IE. The presence of epilepsy among relatives, evidence of pre- or perinatal cerebral damage, mental retardation, and early onset, long periods of uncontrolled seizures before starting an adequate therapy and frequency of seizures appear to be indicative of an adverse prognosis, since differences between the two groups of responsive or unresponsive patients are statistically significant. On the contrary, the occurrence of febrile convulsions in the past history does not seem to have an adverse prognosis. Temporal lobe epilepsy and IS bear the worst prognosis. ME, CPS, GTCS, SPS, LGS and PM have a progressively better outcome in responsiveness to AEDs. Concerning therapy in patients with IE, studies indicate the results of high dose monotherapy appear to be equal or better than with polypharmacy. Because of the gravity of the situation, trials with unconventional drugs have been performed, but it is too early to draw definite conclusions about the long-term usefulness of most of them. In conclusion, our data indicate that the appearance of an IE can be predicted utilizing the above mentioned criteria, considered either alone or in combination. The issue of IE remains undoubtedly an important one among the group of convulsive disorders. Further studies considering a greater number of patients and new therpeutic strategies are to be recommended.