Jorge G. Burneo, Lorie Black, Roy Martin, Orrin Devinsky, Steve Pacia, Edward Faught, Blanca Vasquez, Robert C. Knowlton, Daniel Luciano, Werner Doyle, Sohuel Najjar, Ruben I. Kuzniecky. Race/Ethnicity, Sex, and Socioeconomic Status as Predictors of Outcome After Surgery for Temporal Lobe Epilepsy. Arch Neurol. 2006;63(8):1106–1110. doi:10.1001/archneur.63.8.1106
Several risk factors have been attributed to seizure recurrence after surgery. It is unknown whether race/ethnicity plays a role in outcome.
To evaluate whether race/ethnicity plays a role in seizure recurrence after surgery.
We evaluated data obtained from the epilepsy centers at the University of Alabama at Birmingham and New York University, New York, NY.
All patients included had a diagnosis of mesial temporal sclerosis and underwent temporal lobectomy.
Main Outcome Measures
Occurrence of seizure after surgery was registered 1 year after surgery. We used multiple logistic regression analysis to model the presence of seizure recurrence after surgery and generated odds ratios (ORs) for seizure recurrence after surgery for African American and Hispanic patients relative to white patients. An unadjusted model incorporated only race/ethnicity as the independent variable, and an adjusted model included socioeconomic status, age, duration of epilepsy, education, history of febrile seizures, sex, handedness, lateralization of epileptogenic focus, and number of antiepileptics as the independent variables.
Two hundred fifty-two patients underwent surgical treatment with pathological confirmation of mesial temporal sclerosis. No differences were found between racial/ethnic groups in terms of seizure recurrence in any models. For African American patients, the ORs were 0.9 (95% confidence interval [CI], 0.4-2.1) for the unadjusted model and 0.8 (95% CI, 0.3-2.0) for the adjusted model; for Hispanic patients, the ORs were 1.6 (95% CI, 0.8-3.2) for the unadjusted model and 1.1 (95% CI, 0.5-2.6) for the adjusted model, relative to white patients.
Our data suggest that although sex appears to play a role in the outcomes of surgery for temporal lobe epilepsy, race and socioeconomic status do not.
The success of surgery in temporal lobe epilepsy (TLE) associated with mesial temporal sclerosis (MTS) is higher than that for other types of epilepsy, reaching 58% for patients who were free of seizures impairing awareness, when compared with medical treatment, based on a randomized controlled trial.1 There are, however, marked racial/ethnic disparities in the use of this effective treatment. African American patients have lower rates of surgical intervention for treatment of MTS compared with white patients,2 whereas Hispanic patients in some areas of the United States do not have any access.2 The lack of access to surgical treatment in developing countries is an even greater challenge.3
The reasons for the disparities remain unclear. African American patients may experience a lower psychosocial impact of epilepsy4,5 and may have a preference for nonsurgical management of chronic problems. However, physicians' perceptions and expectations of patients can vary according to factors such as race, which could influence referrals for TLE surgery.6 Furthermore, the lack of patient education that results from these erroneous perceptions is a serious problem. It is critical that we have information regarding treatment outcomes in different racial/ethnic groups so that we can better inform our patients with evidence-based data.
An initial look into this problem in a small group of patients revealed that race had an effect on seizure outcome after temporal lobectomy, with African American patients having a significantly higher recurrence rate than white patients.7 However, in that study, the sample size was small, and other minorities such as Hispanic or Latino patients were not included, although they constitute the largest minority population in the United States.7,8
In the present study, we examined racial/ethnic differences in outcome among a large cohort of patients with medically intractable mesial TLE who underwent temporal lobectomy. We also examined potential modifying variables among racial groups.
Patient-specific discharge and surgical data were obtained from the videoelectroencephalographic units at the University of Alabama at Birmingham and New York University, New York, NY, from patients released with the primary diagnosis of TLE from July 1, 1994, through January 31, 2003. The data set includes information on each patient's age, ZIP code of residence, date of birth, sex, race/ethnicity, date of surgery, degree of education, duration of epilepsy (in years), handedness, lateralization of temporal epileptogenic focus, number of antiepileptics used before surgery, and history of febrile seizures.
All patients discharged with a primary diagnosis of medically intractable TLE were identified. The diagnosis was based on the confirmation of an ictal onset in either temporal region by means of videoelectroencephalographic evaluation, temporal lobe interictal epileptiform abnormalities, clinical semiology consistent with TLE, and intracranial electroencephalographic confirmation of mesial temporal onset in a subset of patients.
All patients without a magnetic resonance imaging (MRI) diagnosis of hippocampal sclerosis or with another pathology associated with hippocampal sclerosis (in addition to MTS) were excluded. The final group consisted of patients with MTS who received a temporal lobe resection.
All patients underwent imaging on a 1.5-T MRI system using standardized protocols. We acquired sagittal and coronal T1-weighted, coronal T2-weighted spin echo, fluid-attenuated inversion recovery, and inversion recovery sequences. We diagnosed MTS on the basis of evidence of hippocampal atrophy and abnormal signal alterations according to validated diagnostic criteria.9,10
All patients who decided to undergo surgery for TLE due to hippocampal sclerosis received a resection of a portion of neocortex and a targeted resection of the amygdala and hippocampus (including the anterior two thirds or the entire hippocampus). Patients underwent regular assessment for postoperative seizure control by the treating neurologist in periodic follow-up visits in their respective epilepsy clinic. All patients had at least 1 year of follow-up. Outcome at the most recent follow-up was classified as seizure free or not seizure free. In the seizure-free group, we also included patients who still had auras (Engel I), including in the immediate postsurgical period.
Representative tissue samples of the hippocampus were routinely obtained for analysis. Specimens were fixed in formalin and embedded in paraffin. Histologic sections of the hippocampus and temporal lobe were stained with hematoxylin-eosin and glial fibrillary acidic protein. Adequate hippocampal tissue was available for the diagnosis of hippocampal sclerosis in most subjects. We diagnosed MTS according to the presence of neuronal loss in the cornu ammonis (CA) 1 and CA3-4 regions of the hippocampus by qualitative criteria. If the pathological findings were not diagnostic for MTS (inadequate sample), the patient was included on the basis of MRI identification of MTS.
Independent variables consisted of race/ethnicity, age, sex, income, duration of epilepsy, education, history of febrile seizures, handedness, lateralization of temporal epileptogenic focus, and number of antiepileptics used. Median household annual income by residence ZIP codes obtained from the 2000 US census was used as a proxy for patient income.11- 15 The following 3 specific ethnic groups constituted the race/ethnicity variable: African American, Hispanic or Latino, and white. Ethnicity was registered by the treating physician, always corroborating with the patient's last name. The following 4 specific age categories were used: younger than 25, 25 to 39, 40 to 49, and 50 years or older. Income was grouped into 4 categories, based on average annual income in US dollars: less than $35 000, $35 000 to $54 999, $55 000 to $74 999, and more than $74 999. Education was grouped as primary and secondary school, college, graduate and postgraduate training, and unknown. Finally, antiepileptics were grouped as 1, 2, and more than 2 antiepileptics.
Multivariate logistic regression was used to investigate the relationship between seizure recurrence and race/ethnicity, with potential confounders controlled. Two sets of logistic regression models were estimated to generate odds ratios (ORs) for seizure recurrence after surgery for African American and Hispanic patients relative to white patients. The first model incorporated only ethnicity as the independent variable and generated unadjusted ORs for having seizure recurrence. The second set included all independent variables described in the preceding subsection. Statistics were computed with SAS version 9.0 software (SAS Institute Inc, Cary, NC) for Windows (Microsoft Corp, Redmond, Wash).
Two hundred fifty-two patients underwent surgical treatment with pathological or MRI confirmation of MTS. Their mean (SD) age was 35.5 (11.9) years. Twenty-six patients were African American, 40 were Hispanic, and 186 were white. Men represented 49.6% of the sample; 86.5% were right-handed; 69.8% had febrile seizures; and 44.0% had left-sided TLE. Seizure recurrence was seen in 53.2% (Table 1).
All independent variables potentially associated with seizure recurrence among racial/ethnic groups were examined (Table 1). There were no differences between the groups with respect to all the variables, with the exception of sex, history of febrile seizures, and socioeconomic status.
Logistic regression results are shown in Table 2. Model 1 includes unadjusted ORs for African American and Hispanic patients, with white patients as the reference group. Hispanic patients were more likely to have seizure recurrence before adjustments were made (OR, 1.6), although the 95% confidence interval included the null value.
In the multivariate logistic regression analysis that controlled for all independent variables (model 2), ORs for African American and Hispanic patients indicated no significant difference when compared with white patients because the 95% confidence interval included the null value.
Finally, the analysis demonstrated that male sex was associated with a reduced risk of seizure recurrence (OR, 0.5; 95% confidence interval, 0.3-0.8) (Table 2).
Using prospectively collected data, we found that seizure outcomes at least 1 year after temporal lobectomy for mesial TLE were similar among white, African American, and Hispanic patients. We previously found lower rates of epilepsy surgery in African American patients compared with white patients, and although a clear explanation was not found, this was not related to access to health care in the African American group.2
Subsequently, in another publication,7 we found a trend for higher rates of seizure recurrence in African American compared with white patients. However, that study was limited by the small sample size and the retrospective design. In the present study, we did not include access to medical insurance as an independent variable in the analysis because it is more probably related to the decision to undergo surgical treatment for intractable TLE.
Furthermore, because the number of African American patients was very small, we could not rule out the possibility of ascertainment bias for those undergoing surgery. The present study has a larger sample with patients from 2 different medical centers and different geographic locations. Also, it includes a third group, Hispanic or Latino patients, the size of which has tripled in certain states such as Alabama, and which has a large representation in the New York area and has become the largest minority population in the United States.12 To our knowledge, most studies examining racial differences in neurological diseases have focused on the African American population, and ours is the first large series (to our knowledge) to examine the outcome of epilepsy surgery in Hispanic patients.
Potential limitations of this study include differences in the patient groups, limited geographic sampling, a small sample size for 1 ethnic group, and the methodology for ethnic classification. First, there were some significant differences in the baseline characteristics of the patients, such as socioeconomic status, but this may be an effect of the sample size. Second, although the patient population consisted of white, African American, and Hispanic patients and the information came from 2 geographically different epilepsy centers, it is not representative of the general population because other racial/ethnic groups such as Asians were not included. Third, the relatively small number of African American patients limits the power to detect differences involving subgroups within that group. Fourth, the potential misclassification of race/ethnicity remains an issue, particularly with Hispanic patients; although the name origin can help with this, reports of race/ethnicity can be subjective and are infrequently validated.16 This important consideration is not unique to our analysis.17 Nonetheless, it is unlikely that systematic underreporting or overreporting of a particular group is correlated with the propensity to perform invasive treatments like surgery. Fifth, the study was restricted to patients with MTS and intractable TLE. However, although racial differences for other types of epilepsy may not be inferred from this study, the homogeneous population (MTS only) is a major strength. Finally, detailed data on neurosurgeon characteristics (eg, experience with the procedure, individual surgeon volume, and technique) may have influenced the rate found here in a particular direction. However, any effect would be minimal unless large intraindividual racial differences were present among the surgeons.
As with other medical conditions, patients of racial/ethnic minorities have concerns about surgical outcomes and adverse events after surgical procedures, and these concerns can influence patients' willingness to consider the treatment.18,19 This may partly account for the disparities observed in epilepsy.
We need to examine race/ethnicity subgroup data when possible (especially when previous work suggests significant differences by race or ethnicity), as recommended by the National Institutes of Health.20 Current evidence-based knowledge does not allow us to know the potential relationship between race/ethnicity and trial outcomes, and this lack of understanding constitutes a barrier to using results of clinical research to address racial/ethnic disparities in health.21
Our findings of a sex difference in outcome (worse outcome in women compared with men with TLE) do not have a clear explanation. Studies found that male animals were more prone to seizure activity than were female animals.22,23 In humans, men may be more susceptible to temporal lobe–like seizures because of high levels of testosterone24 and more vulnerable to seizure-associated brain damage25 and may have a slightly higher overall incidence of epilepsy when compared with women,26,27 although this may not be the case in the localization-related epilepsies.28 On the other hand, greater plasticity has been found in women when compared with men, in terms of memory, in response to the onset of epilepsy.29 Sex differences in seizure recurrence after surgical treatment of medically refractory epilepsy have not been shown to be a predictive factor, but most of the previous analyses7,30- 32 have included smaller samples than our study did, or simply sex was not considered as a variable under study.
In conclusion, race/ethnicity and socioeconomic status do not appear to play an important role in TLE surgery outcome, and being male appears to have a protective effect in terms of seizure recurrence.
Correspondence: Ruben I. Kuzniecky, MD, Comprehensive Epilepsy Program, New York University, 403 E 34th St, New York, NY 10016 (email@example.com).
Accepted for Publication: February 17, 2006.
Author Contributions:Study concept and design: Burneo, Najjar, and Kuzniecky. Acquisition of data: Burneo, Black, Pacia, Faught, Vasquez, Knowlton, Doyle, and Kuzniecky. Analysis and interpretation of data: Burneo, Martin, Devinsky, Luciano, and Kuzniecky. Drafting of the manuscript: Burneo, Black, Devinsky, Doyle, and Najjar. Critical revision of the manuscript for important intellectual content: Burneo, Martin, Devinsky, Pacia, Faught, Vasquez, Knowlton, Luciano, and Kuzniecky. Statistical analysis: Burneo and Najjar. Administrative, technical, and material support: Burneo, Black, Devinsky, Faught, Knowlton, and Kuzniecky. Study supervision: Vasquez, Luciano, and Doyle.