Extended follow-up after anterior temporal lobectomy demonstrates seizure recurrence 20+ years post-surgery.

Abstract

Anterior temporal lobectomy (ATL) for medication-resistant localized epilepsy results in ablation or reduction of seizures for most patients. However, some individuals who attain an initial extended period of post-surgical seizure freedom will experience a later seizure recurrence. In this study we examined the prevalence and some risk factors for late recurrence in an ATL cohort with extensive regular follow-up. Included were 449 patients who underwent ATL at Austin Health, Australia, from 1978 to 2008. Post-surgical follow-up was undertaken 2-3 yearly. Seizure recurrence was tested using Kaplan-Meier analysis, log-rank test and Cox regression. Late recurrence was qualified as a first disabling seizure >2 years post-surgery. We examined risks within the ATL cohort according to broad pathology groups and tested whether late recurrence differed for the ATL cohort compared to patients who had resections outside the temporal lobe (n=98). Median post-ATL follow-up was 22 years (range 0.1 to 38.6), 6% were lost to follow-up, 12% had died. Probabilities for remaining completely seizure free after surgery were: 51% (95%CI 53-63) at 2 post-operative years, 36% (95%CI 32-41) at 10 years, 32% (95%CI 27-36) at 20 years, 30% (95%CI 25-34) at 25 years. Recurrences were reported up to 23 years post-operatively. Late seizures occurred in all major ATL pathology groups, with increased risk in 'normal' and 'distant lesion' groups (p≤0.03). Comparison between the ATL cohort and patients who underwent extra-temporal resection demonstrated similar patterns of late recurrence (p=0.74). Some first recurrences were very late, reported decades after ATL. Late recurrences were not unique to any broad ATL pathology group and did not differ according to whether resections were ATL or extra-temporal. Reports of these events by patients with residual pathology suggest that potentially epileptogenic abnormalities outside the area of resection may be implicated as one of several possible underlying mechanisms.