Characterization of hemorrhagic complications after surgery for temporal lobe epilepsy

Objectives: To assess the significance of symptomatic hemorrhagic complications occurring after different temporal resections for temporal lobe epilepsy (TLE) and to compare this data to findings with postoperative hematomas after temporal surgery for mostly glial or metastatic tumors. Patients and Methods: Patient charts were retrospectively reviewed for 442 patients who underwent surgery for refractory TLE between 1995 and 2000. Procedures were 247 transsylvian amygdalohippocampectomies (AH), 40 transcortical AH, 57 anterior temporal lobectomies (ATL), 23 lesionectomies plus AH, and 75 lesionectomies without AH. All patients with delayed awakening or new neurological deficits due to hemorrhages were included in the study. An identical procedure was performed to detect symptomatic hemorrhages after 208 procedures for temporal tumor resection during the same time period. Results: Symptomatic postoperative hemorrhages were found in 17 patients (3.8 %) undergoing epilepsy surgery, while the incidence was 3.0 x% in a group with space-occupying temporal tumors (six patients). Hemorrhages showed a characteristic distribution after epilepsy surgery: in eight patients they were located remote from the site of surgery in the upper cerebellar vermis and foliae. Five typical hemorrhages associated with dysphasia were found in the left frontal operculum, only three patients had hematomas in the resection cavity, and one was located epidurally. Two patients had more than one location of hemorrhage. Transsylvian AH and ATL had a similar risk for postoperative hemorrhage, whereas none was found after lateral lesionectomies or transcortical AH. Intraoperative manipulations were associated with opercular hemorrhages; the only predisposing factor for resection site hematomas was older age, whereas cerebellar hemorrhages were associated with cerebrospinal fluid (CSF) loss during AH and ATL. There was no mortality in the TLE group, and 0.75 % permanent mild deficits. Seizure outcome did not differ from the rest of the group (82.5 % satisfactory seizure control). In contrast, all intraaxial hematomas after tumor surgery (N = 4, incidence 1.9 %) were located in or adjacent to the resection cavity. Prognosis was much worse with parenchymal hemorrhages after tumor surgery: three of four patients died, one survived with a severe hemiparesis, only two patients with extraaxial hematomas (incidence 1 %) had a complete recovery. The 3 % incidence of symptomatic hemorrhages was only insignificantly lower compared to the TLE group, patients with tumor surgery were older than TLE patients (49 versus 33 years), and in five of six patients only incomplete tumor resection was achieved. Conclusion: Although associated with a low permanent morbidity, features of postoperative hemorrhages after TLE surgery are characteristically different to complications after surgery for other indications, which has to be kept in mind for patient counseling and obtaining informed consent.