Background Patients who present with only simple or complex partial seizures have a poorly documented prognosis. Treatment may be advocated to prevent future secondary generalised seizures, reduce the frequency of further simple or complex partial seizures or a combination of both.

Methods A full statistical analysis on 1334 patients was carried out. The outcomes measured were post-randomisation times to first seizure of any type and first tonic–clonic seizure. Methodology was adopted that accounted for individuals' underlying pre-randomisation seizure counts and allowed for the possibility that there may be a proportion of the sample that will not experience post-randomisation seizure recurrence.

Results 103 subjects randomised to the MESS (Multicentre Study of Early Epilepsy and Single Seizures) study had only partial seizures at randomisation. Only 17 of these had a tonic–clonic seizure during follow-up. The presence of an abnormal EEG at randomisation influenced this risk: an estimated 23% of those with EEG abnormality were at risk of tonic–clonic seizures during follow-up compared with 16% of those with a normal EEG. The group did, however, continue to have partial seizures during follow-up, and modelling showed that the impact of treatment on these seizures was significantly less than the effects of treatment on the frequency of tonic–clonic seizures in those patients with such pre-randomisation seizures.

Conclusion Patients presenting with a history of only partial seizures are at low risk of subsequent tonic–clonic seizures in the period of time to which therapeutic decisions are relevant. The effects of the antiepileptic drugs used in the MESS study are greater for tonic–clonic seizures than they are for partial seizures.