Influence of Head Trauma on Outcome Following Anterior Temporal Lobectomy

ORIGINAL CONTRIBUTION Influence of Head Trauma on Outcome Following Anterior Temporal Lobectomy

Lori A. Schuh, MD; Thomas R. Henry, MD; Gail Fromes, RN; Mila Blaivas, MD; Donald A. Ross, MD; Ivo Drury, MB, BCh

Background: There is controversy in the literature re- 40 had more than 1. Frequencies were 39 febrile con- garding the importance of risk factors in developing epi- vulsions (15 complex febrile convulsions), 29 head lepsy and seizure outcome following anterior temporal trauma, 22 with lesions seen on magnetic resonance im- lobectomy. Some of the existing studies may be biased aging, 12 history of meningitis, 2 history of encephali- because of patient selection and limitations in determin- tis, 19 family history of epilepsy, and 4 perinatal insult. ing predisposition. Seventy-one (70%) were classified as Engel’s class I, with 56 patients continuously free of seizures at follow-up. Objective: To investigate the role of risk factors for epi- Those without risk factor were as likely to be rendered lepsy in determining outcome following anterior tem- free of seizures following anterior temporal lobectomy poral lobectomy. as those with a risk factor (P = .27). No risk factor alone or in combination was correlated with complete free- Patients and Methods: We identified 102 patients dom from seizures following anterior temporal lobec- in a consecutive surgery series for epilepsy from a ter- tomy, but the presence of head trauma, alone or in com- tiary center with a minimum of 1-year postoperative bination, was correlated with continued seizures following follow-up. Risk factors for epilepsy were determined anterior temporal lobectomy (P = .03; odds ratio, 2.6). Bet- prospectively on at least 3 occasions before anterior ter outcomes were not seen in those with head trauma temporal lobectomy. Risk factors investigated were a before the age of 5 years (P = .57). These findings did not history of febrile convulsions, family history of epilepsy, change if all those with lesions on magnetic resonance significant head trauma, history of meningitis, history imaging were excluded in the analysis. Those with a his- of encephalitis, or significant perinatal insult. Foreign tory of head trauma were as likely to have pathologic evi- tissue lesions on magnetic resonance imaging was also dence of mesial temporal sclerosis as others (P = .82). included if an anterior temporal lobectomy was performed for presumed dual pathologic findings (hippo- Conclusions: Patients with a history of significant head campus and lesion). Outcome was determined using trauma are less likely to become free of seizures follow- Engel’s classification. For statistical analysis we used ing anterior temporal lobectomy. No other risk factor cor- successive logistic regression analysis, ␹2 test, Fisher related with a statistically significant greater or lesser exact test, and t test. chance of freedom from seizures. This information may be used in preoperative counseling of patients. Results: Of the 102 patients, 13 had no identified risk factor for epilepsy, 49 had 1 identified risk factor, and Arch Neurol. 1998;55:1325-1328

HERE IS controversy in the tistical analysis9 or statistical bias10 in as- literature regarding the im- sessing those with more than 1 risk fac- portance of risk factors for tor in developing epilepsy. One group epilepsy in the develop- showed that age at onset of risk factor in- ment of mesial temporal duced injury rather than the nature of the lobe epilepsy and their relationship to out- risk factor itself was the important deter- T 11 From the Department of come following anterior temporal lobec- minant of outcome following ATL, but Neurology, Henry Ford tomy (ATL). Many authors1-5 found no cor- this analysis may also be biased as each in- Hospital, Detroit, Mich relation between risk factors and surgical dividual included in the study was lim- (Dr Schuh); Department of outcome, but some of these studies5 were ited to 1 or no risk factor for epilepsy, Neurology, Emory University, not limited to ATL and others1-3 did not which excludes many individuals with Atlanta, Ga (Dr Henry); and recognize currently accepted risk factors more than 1 risk factor.4,9 Departments of Nursing such as febrile convulsions (FC). Oth- We reviewed our experience with risk (Ms Fromes), Pathology 6-10 (Dr Blaivas), Surgery ers have noted an association among FC, factors for epilepsy and outcome follow- (Dr Ross), and Neurology mesial temporal lobe epilepsy, and good ing ATL without placing a priori empha- (Dr Drury), University of outcome following ATL. These studies may sis over any specific predisposition. We in- Michigan, Ann Arbor. be limited on the basis of absence of sta- vestigated whether any risk factor, alone

ARCH NEUROL / VOL 55, OCT 1998 1325

©1998 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/25/2021 Table 1. Incidence of Risk Factors for Epilepsy* PATIENTS AND METHODS Risk Factor Frequency None 13 One hundred two patients with medically intractable Febrile convulsions 19 complex partial epilepsy from a continuous series un- Head trauma 10 derwent ATL at the University of Michigan Hospital Lesion seen on MRI 10 in Ann Arbor between June 1990 and March 1996. Sei- History of meningitis 6 zure outcome was determined as of March 1997, a mini- Febrile convulsions and family history 6 mum of 1 year after surgery, using Engel’s classifica- of epilepsy tion.12 The best outcome, termed class Ia, is defined as Family history of epilepsy 3 complete freedom from seizures following ATL ex- Febrile convulsions and head trauma 3 cluding early postoperative seizures. Continued need Febrile convulsions and meningitis 2 for anticonvulsants is not included in this definition. Febrile convulsions and perinatal insult 2 Patients were chosen for ATL based on criteria pub- Head trauma and lesion seen on MRI 2 lished elsewhere.13 The presence of individual risk fac- Head trauma and family history of epilepsy 2 tors did not influence ATL candidate selection. Family history of epilepsy and meningitis 2 Risk factors for epilepsy were determined from at Head trauma, lesion seen on MRI, and 2 least 3 structured interviews with each patient, family family history of epilepsy members, and close friends before surgery. Medical Febrile convulsions, head trauma, and 2 family history of epilepsy recordswereused,whenavailable,toconfirmandsupple- Febrile convulsions, lesion seen on MRI, 2 mentinformationprovided.Allcredibleriskfactorswere and family history of epilepsy included in this analysis. These included FC, signifi- Head trauma, perinatal insult, and family 2 cant head trauma (HT), history of meningitis, history history of epilepsy of encephalitis, family history of epilepsy, and signifi- Other 14 cant perinatal insult resulting in cerebral palsy or neo- natal convulsions. These were chosen based on the work *MRI indicates magnetic resonance imaging; other, combinations of risk of Rocca et al,14 which describes risk factors for com- factors other than those individually listed. plex partial seizures from a population-based case-control study, and from a recent study by Berkovic et al.15 Those with complex FC (FC lasting Ն30 minutes, or dividual risk factors and the most common combina- with postictal focal neurologic deficits) were identified tions are presented in Table 1. and analyzed both separately and grouped with simple Outcomes were as follows: 56 patients with class Ia FC. Head trauma with any of the features described by (55%), 15 class Ib, Ic, or Id (15%), 12 class II (12%), 8 Rocca et al14 were included in the analysis: loss of con- class III (8%), 10 class IV (10%), and 1 unavailable for sciousness, posttraumatic amnesia, or evidence of skull follow-up. The percentage of frequencies of class Ia out- fracture. In their study, HT was significant even when comes for the most common risk factors are listed below. loss of consciousness was less than 30 minutes in duration. Those with a foreign tissue lesion seen on mag- Risk Factor Class Ia Outcomes, % netic resonance imaging (MRI) were included so as not None 69 to exclude those with presumed dual pathologic fea- FCs 63 tures (pathologic features in the lesion and hippocam- HT 40 pus), who presumably will have seizures of mesial tem- Lesion seen on MRI 60 poral lobe onset. The analysis was repeated excluding History of meningitis 50 those with lesions seen on MRI to ensure that the find- FCs and family history of epilepsy 67 ings of this study were not biased. Surgical pathologic Family history of epilepsy 33 FCs and HT 33 findings were reviewed by one of us (M.B.), an expe- Remainder* 50 rienced neuropathologist who was blinded to patient *Remainder indicates sole and combination risk factor frequencies other risk factors and surgical outcome. Results were catego- than those individually listed. rized for ease of analysis, as showing mesial temporal sclerosis (Ͼ50% neuron loss in CA1), a foreign tissue There was no difference in likelihood of class Ia lesion, mesial temporal sclerosis, and a lesion, or no outcomes between those with and those without an pathologic diagnosis. For statistical analysis we used identified risk factor (P = .27). Using sequential logistic the SAS statistical software and Statview (SAS Inc, Cary, regression analysis, no single risk factor for epilepsy, com- NC) to perform successive logistic regression analysis, bination, or absence of risk for epilepsy was more likely 2 ␹ test, Fisher exact test, and t test. than others to be associated with class Ia outcomes. The only correlation with outcome was HT that was negatively correlated with class Ia outcomes in this series or in combination, was positively or negatively associ- (P = .03; odds ratio, 2.6). Given these initial results we ated with the best outcomes following ATL. investigated HT further and found a positive correlation between HT and class III outcomes (P = .002). RESULTS Subset analysis of those with HT revealed no difference in outcome between those whose injury occurred be- Thirteen of the 102 patients had no identified risk fac- fore or after the age of 5 years (P = .57). Data on outcome tor for epilepsy. Of 89 with recognized risk factors, 40 by age at HT are presented in Table 2. Length of follow- had more than 1 identified risk. The frequencies of in- up was not longer for those with HT (mean ± SD, 46.7 ± 3.3

ARCH NEUROL / VOL 55, OCT 1998 1326

©1998 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/25/2021 ing ATL for medically intractable complex partial epi- Table 2. Outcome by Age at Head Trauma* lepsy, we found no correlation between individual risk factors and complete freedom from seizures following ATL. Outcome† Age at Head Those without an identified risk factor for epilepsy were Trauma, y Ia Ib, Ic, Id II III IV no less likely than any other group to be rendered free of Ͻ552122seizures after surgery for epilepsy. Only a history of HT cor- Ն564241related with outcome. Those with HT, whether alone or in combination with other risk factors, were less likely to *No difference in outcome was seen between those with head trauma become free of seizures following ATL, and more likely than before and after 5 years of age ( P = .57). any other group to have class III outcomes (worthwhile im- †Engel’s classification. provement with more than “rare” seizures, which we op- erationally defined as up to 80% reduction of seizures and months) compared with others (mean ± SD, 43.5 ± 2.2 usually meant continued monthly seizures). These poorer months) in this series. As this was a consecutive series it outcomes cannot be explained by histopathologic ab- included those with bilateral temporal lobe epilepsy if 80% sence of mesial temporal sclerosis. Those with a history of or more of ictal onsets were recorded from the same side. HT at a young age did not have better outcomes than those Eleven of 102 patients had bilateral temporal lobe epilepsy, with later HT. Patients with HT were not more likely to 2 with a history of HT. There was no increased incidence have bilateral temporal epilepsy based on preoperative ic- of HT in those with bilateral temporal lobe epilepsy (P = .72). tal recordings, require invasive ictal monitoring, or show Thirty-fourpatientsrequiredinvasivemonitoringwithdepth differences from other risk groups in age at onset of epi- and/or subdural strip electrodes before ATL, 8 had a his- lepsy, age at surgery, or length of follow-up. Of those with tory of HT. Those with HT were not more likely to require HT, preoperative MRI findings consistent with hemosid- invasive presurgical monitoring in this series (P = .44). Pre- erin deposition correlated with freedom from seizures operative MRI was reviewed in those with HT and corre- following ATL. lated with outcome. Five patients showed changes consis- These results are different from the findings of Abou- tent with mesial temporal sclerosis, 14 were normal, 2 had Khalil et al10 who found in a consecutive ATL series an as- foreign tissue lesions seen on MRI, 4 had nonspecific in- sociation between prolonged FC and good outcome after flammatory or posttraumatic changes, 1 had focal atrophy, surgery. This may be due to differences in the way risk fac- and 3 had hemosiderin deposition. Only the presence of he- tors for epilepsy were evaluated. In their study, patients were mosiderin correlated with outcome, with better seizure out- separated into those with and those without FC. Other risk comes seen in this group (P = .04), but these data should factors were determined in each of these groups, and in their be interpreted with caution because of the small sample size. review 39% of the non-FC group had a history of HT, while Age at surgery did not differ between those with a history only 10% of the FC group had a similar history. The au- of HT (mean ± SD, 34.2 ± 6.3 years) and those without thors noted that HT was statistically significant in predict- (mean ± SD, 31.9 ± 9.3 years) (P = .15, t test). Neither was ing poor outcome following ATL (P = .03), however they there a difference between the 2 groups with respect to age concluded that this was less important than the presence at onset of epilepsy with a mean (± SD) of 14.9 ± 9.0 years of FC. This decision was made on the basis of absolute P for those with HT and a mean (± SD) of 12.1 ± 10.8 years values rather than a statistical analysis that compared both for those without (P = .20, t test). factors. Another difference between the 2 studies is the defi- Eighty-nine of 102 pathologic specimens were avail- nition of good outcome that included all class I outcomes able for review. Sixty-two showed mesial temporal scle- in their review. From a physician standpoint this is not un- rosis (alone), 10 dual pathologic features, 8 foreign tis- reasonable, but increasing data from quality-of-life mea- sue lesions without mesial temporal sclerosis, and 9 had sures and long-term employment status after surgery for no pathologic diagnosis. The only findings of signifi- epilepsy support a real separation between those com- cance were all those with a history of meningitis had me- pletely free of seizures and even those with continued simple sial temporal sclerosis (P = .02) and, as expected, the pres- partial seizures after ATL.16,17 ence of a lesion on MRI correlated with the presence of Our finding that those without an identified risk fac- a lesion by pathologic review (P = .001). Twenty of 29 tor for epilepsy did as well as those with a known risk is with HT had mesial temporal sclerosis by pathologic re- in agreement with that of early investigators,1,3,4,5 but con- view, similar to all other risk groups with the exception trasts with that of Mathern et al.11,18 There are differ- of those with meningitis (P = .82). ences in how outcome was defined between the 2 stud- Due to concerns about possible bias from inclu- ies and when outcome was determined, but despite this, sion of those with lesions seen on MRI, a repeated analy- fewer than 25% of their study population without a risk sis was performed excluding these individuals. Head factor for epilepsy were free of seizures in the last 12 trauma remained the only significant risk factor with re- months of follow-up. In our series nearly 70% of those gard to outcome (P = .04). without a risk for epilepsy were continuously free of seizures for the duration of follow-up, between 13 to 70 COMMENT months. These differences may be due to different patient populations, or due to differences in preoperative In this study, which examined all accepted risk factors for patient selection as all our patients had the benefit of MRI epilepsy, whether present alone or in combination, and their and functional imaging with positron emission tomog- relationship to outcome in a cohort of patients undergo- raphy before surgery, and the time frame of patient ac-

ARCH NEUROL / VOL 55, OCT 1998 1327

©1998 American Medical Association. All rights reserved. Downloaded From: https://jamanetwork.com/ on 09/25/2021 quisition in the study by Mathern et al11,18 precluded such plete freedom from seizures following ATL. Those under- an evaluation for all their patients. going ATL without an identified risk factor for epilepsy are The absence of a positive correlation between indi- not less likely to be rendered free of seizures following vidual risk factors and good outcome following ATL is, surgery. Those with a history of HT, whether the sole risk we believe, due to the overall high success rate of ATL, factor for epilepsy or combined with other risk factors, are especially in light of current surgical selection criteria that significantly less likely to be rendered free of seizures fol- now include MRI and functional imaging. With a proce- lowing ATL, regardless of the age of injury. This informa- dure for which the chances of long-term complete suc- tion may be used in preoperative counseling of patients. cess are higher than 50%, it is likely that studies of large numbers of patients will be necessary to show any statis- Accepted for publication January 29, 1998. tically significant difference between those patients who Corresponding author: Lori Schuh, MD, Department do well and those who are continuously free of seizures. of Neurology, Henry Ford Hospital, 2799 W Grand Blvd, Fifty-five percent in this series had class Ia outcomes. As Detroit, MI 48202-2689. we reported, continuous outcome since ATL, rather than outcome in the most recent 12 months of follow-up or at some specific point of time after surgery, our results do REFERENCES not differ significantly from long-term outcome reported 19,20 1. Bengzon ARA, Rasmussen T, Gloor P, Dussault J, Stephens M. Prognostic fac- by others. We note, however, that the sample size of tors in the surgical treatment of temporal lobe epileptics. Neurology. 1968;18: some etiologic groups was small (ie, meningitis, encepha- 717-731. litis, or perinatal insult) and conclusions regarding out- 2. Stepien L, Bidzinski J, Mazurowski W. The results of surgical treatment of temporal lobe epilepsy. Pol Med J. 1969;8:1184-1190. come among these should be viewed with caution. 3. Rasmussen T. Surgical treatment of patients with complex partial seizures. Adv Animal models of HT using fluid percussion injury Neurol. 1975;11:415-442. have shown that even mild unilateral HT results in bi- 4. Jensen I. Temporal lobe epilepsy: etiological factors and surgical results. Acta Neurol Scand. 1976;53:103-118. lateral (ipsilateral greater than contralateral) neuron loss 5. Dodrill CB, Wilkus RJ, Ojemann GA, et al. Multidisciplinary prediction of seizure in the hilus and CA3 region of hippocampus, and en- relief from cortical resection surgery. Ann Neurol. 1986;20:2-12. 21-23 6. Falconer MA, Serafetinides EA, Corsellis JA. Etiology and pathogenesis of tem- hanced limbic epileptogenesis. It is possible that bi- poral lobe epilepsy. Arch Neurol. 1964;10:233-248. lateral mesial temporal injury may account for the poorer 7. Falconer MA. Mesial temporal (Ammon’s horn) sclerosis as a common cause of outcomes seen in those with HT; however, this does not epilepsy: aetiology, treatment, and prevention. Lancet. 1974;2:767-770. 8. Falconer MA. Genetic and related aetiological factors in temporal lobe epilepsy. appear to be the case in this population based on preop- Epilepsia. 1971;12:13-31. erative assessment, and would not be unique to HT as 9. French JA, Williamson PD, Thadani, et al. Characteristics of medial temporal lobe epi- most mesial temporal lobe epilepsy appears to be bilat- lepsy, I: results of history and physical examination. Ann Neurol. 1993;34:774-780. 10. Abou-Khalil B, Andermann E, Andermann F, Olivier A, Quesney LF. Temporal lobe eral with asymmetric severity irrespective of the cause epilepsy after prolonged febrile convulsions: excellent outcome after surgical treat- of epilepsy based on autopsy series.24 The nature of the ment. Epilepsia. 1993;34:878-883. 11. Mathern GW, Pretorius JK, Babb TL. Influence of the type of initial precipitating injury in HT may result in 1 or more of the following: injury and at what age it occurs on course and outcome in patients with tempo- multiple ictal onset zones where only the most active zone ral lobe seizures. J Neurosurg. 1995;82:220-227. was resected, activation of a previously latent ictal onset 12. Engel J Jr. Outcome with respect to epileptic seizures. In: Engel J Jr, ed. Surgi- cal Treatment of the Epilepsies. New York, NY: Raven Press; 1987:553-571. zone occurs postoperatively, the presence of neocorti- 13. Henry TR, Ross DA. Summary of epilepsy surgery protocols. In: Luders HO, ed. cal temporal lobe epilepsy simulating mesial temporal lobe Epilepsy Surgery. New York, NY: Raven Press; 1991:781. epilepsy, or that HT injures subcortical structures that 14. Rocca WA, Sharbrough FW, Hauser WA, Annegers JF, Schoenberg BS. Risk factors for complex partial seizures: a population-based case-control study. Ann do not in themselves act as ictal onset zones, but which Neurol. 1987;21:22-31. modulate or regulate cortical epileptic excitability, such 15. Berkovic SF, McIntosh A, Howell RA, Mitchell A, Sheffield LJ, Hopper JL. Famil- ial temporal lobe epilepsy: a common disorder identified in twins. Ann Neurol. that new ictal onset zones develop after surgery for epi- 1996;40:227-235. lepsy because these subcortical structures remain. As 16. Vickrey BG, Hays RD, Engel J, et al. Outcome assessment for epilepsy surgery: the shown in animal models, it may be differences in the his- impact of measuring health-related quality of life. Ann Neurol. 1995;37:158-166. 17. Sperling MR, O’Connor MJ, Saykin AJ, Plummer C. Temporal lobectomy for re- topathologic features of mesial temporal structures in fractory epilepsy. JAMA. 1996;276:470-475. those with HT, with maximal neuron loss in CA3 and 18. Mathern GW, Babb TL, Vickrey BG, Melendez M, Pretorius JK. The clinical- absence of mossy fiber sprouting, which account for the pathogenic mechanisms of hippocampal neuron loss and surgical outcomes in temporal lobe epilepsy. Brain. 1995;118:105-118. poorer outcomes seen in this etiologic group. 19. Rougier A, Dartigues JF, Commenges D, Clovine B, Loiseau P, Cohadan F. A lon- Marks et al25 reported a series of patients with a history gitudinal assessment of seizure outcome and overall benefit from 100 corticec- tomies for epilepsy. J Neurol Neurosurg Psychiatry. 1992;55:762-767. of HT and outcome following surgery for epilepsy, not lim- 20. Berkovic SF, McIntosh AM, Kalnins RM, et al. Preoperative MRI predicts outcome ited to ATL, and reported good outcome only in those whose of temporal lobectomy: an actuarial analysis. Neurology. 1995;45:1358-1363. trauma occurred before 5 years of age. Their overall impres- 21. Lowenstein DH, Thomas MJ, Smith DH, McIntosh TK. Selective vulnerability of dentate hilar neurons following traumatic brain injury: a potential mechanistic sion, however, was that those with HT had epilepsy that was link between head trauma and disorders of hippocampus. J Neurosci. 1992;12: difficult to localize, and in general deterred surgical inter- 4846-4853. vention in this group. We did not find any difference in out- 22. Shumate M, Rafiq A, Lyeth B, Gong QZ, Coulter D. Traumatic brain injury pro- duces loss of neurons in hilus and CA3 but no significant mossy fiber sprouting come comparing groups with HT before and after the age in hippocampus [abstract]. Epilepsia. 1995;36:118. of 5 years. Once again study design may account for some 23. Rafiq A, Shumate M, Lyeth B, Gong QZ, Coulter D. Enhanced epileptogenesis in hippocampal/entorhinal cortical slices following traumatic brain injury [ab- of these differences. The Yale series excluded patients with stract]. Epilepsia. 1995;36:118. any other possible risk factor for epilepsy. 24. Mourtizen-Dam A. Hippocampal neuron loss in epilepsy and after experimental In summary, in a cohort undergoing ATL, we found seizures. Acta Neurol Scand. 1982;66:601-642. 25. Marks DA, Kim J, Spencer DD, Spencer SS. Seizure localization and pathology no single risk factor for epilepsy, whether the sole risk or following head injury in patients with uncontrolled epilepsy. Neurology. 1995; combined with other risk factors, was predictive of com- 45:2051-2057.

ARCH NEUROL / VOL 55, OCT 1998 1328