Introduction
Temporal lobe epilepsy (TLE) may modify the functions of respiratory, gastrointestinal/abdominal, and cardiovascular (CV) autonomic regulation during the ictal and interictal period of seizures.1–5 During the interictal period, studies have reported an impairment of parasympathetic6–9 and sympathetic7–11 CV functions. Studies have also shown that even a minimal epileptogenic discharge markedly disorganizes the autonomic nervous system, and such alterations in autonomic neural discharge are associated with arrhythmogenesis and may contribute to sudden death in patients with epilepsy.12,13
Previous studies have shown that the hemispheric center of autonomic functions differentially modulates autonomic CV function. During stressful tasks, healthy participants had increased blood flow in the right insula, which was associated with sympathetic activation. During non- tasks, left insular activation was observed, accompanied by decreases in blood pressure (BP) and heart rate (HR).14 Stimulation of the right insular cortex in patients with epilepsy revealed that the right hemisphere predominantly controlled sympathetic activity.15 Right hemispheric sympathetic control was also observed in studies where hemispheric inactivation was achieved using intracarotid amobarbital sodium.16–18 However, in many disease conditions, such as hemispheric stroke and lesion, the right hemisphere modulates parasympathetic functions.19–21
Hemispheric centers, particularly structures adjacent to the temporal lobe, play an important role in the regulation of autonomic CV functions. The adjacent interconnected structures of the temporal lobe, namely the insular cortex, amygdala, cingulate gyrus, and hippocampus, are the centers of autonomic CV modulation.22,23 Therefore, any destructive lesion such as an infarct, surgical resection, or stimulation such as epileptogenic activity in the temporal lobe may alter the autonomic modulation.
Therefore, we hypothesized that autonomic CV responses would be differentially affected in patients with left or right TLE before and after surgical resection of the seizure onset zone. Consequently, we investigated the effect of hemispheric lateralization of seizure foci on autonomic CV functions before and after TLE surgery. The effect of lateralization was assessed using a standardized battery of autonomic reactivity tests. Our study may contribute to a better understanding of the hemispheric modulation of autonomic CV function in patients with TLE before and after surgery.
Methods
Patients
The current prospective and observational cohort study was conducted in the Autonomic Function Lab at the All India Institute of Medical Sciences. Patients were referred from an intractable epilepsy clinic from January 2005 to September 2008 after the localization and lateralization of seizure foci. Only patients with refractory left TLE (LTLE) and right TLE (RTLE) who underwent surgery, were recruited. The study was approved by the local institutional ethics committee, and informed consent was obtained from each patient before testing. Patients with comorbidities that would affect autonomic functions, such as diabetes mellitus, cardiopulmonary disease, renal failure, and psychiatric illness, were excluded from the study. Chronic alcoholics and smokers were also excluded from the study.
The localization and lateralization of the seizure foci was assessed in the Department of Neurology by documenting at least 3 habitual seizures using video electroencephalography (VEEG). All patients had unilateral temporal lobe pathology with a same-sided seizure onset zone and underwent epilepsy surgery after routine pre-surgical evaluation. The pre-surgical evaluation included interictal electroencephalography (EEG) (performed in triplicate), VEEG, ictally subtracted single-photon emission tomography, positron emission tomography, and magnetic resonance imaging (MRI, epilepsy protocol at 1.5 T).24 At least 2 noninvasive evaluations, such as VEEG and MRI, had to be concordant before surgery. Patient eligibility for TLE surgery was decided after a discussion by a team of neurologists, neurosurgeons, neurophysiologists, neuroradiologists, and a neuropsychologist. Furthermore, patients underwent neuropsychological and psychiatric evaluations, visual field testing, and neurological diagnostic tests. Based on the pre-surgical evaluations and concordance for lateralization and localization, patients were divided into 2 groups: LTLE and RTLE groups. The following demographic information was collected: sex, age, type of seizures, duration of epilepsy, seizure frequency, Engel class of seizures, antiepileptic drugs (AED) taken, mean number of AEDs, EEG ictal onsets, etiology, and area resected during epilepsy surgery.
Study procedures
All autonomic function tests were conducted in an air-conditioned laboratory environment (22–26°C) between 2 pm and 4 pm and 2 hours after a light meal. Autonomic function tests were conducted 3 times: before surgery and at 3 and 6 months after surgery. Autonomic function tests before surgery were conducted at least 12 hours after the last seizure and prior to hospitalization for surgery to avoid anxiety due to hospitalization. Patients received their regular AEDs on the testing days, and their drug regimen was not changed after surgery. All patients underwent a standard test battery for the assessment of autonomic CV reactivity as described previously.25,26 The HR responses during the deep breathing test, Valsalva maneuver, and head-up tilt test were used as markers of parasympathetic reactivity, and BP responses to the hand grip test, cold pressor test, and head-up tilt test were used as markers of sympathetic reactivity.
Resting measurements
The resting BP, resting HR, and resting respiratory rate (RR) were measured after the patient rested in the supine position for 15 minutes. The BP was measured using a mercury sphygmomanometer, and HR was measured using electrocardiography (ECG). The RR was measured using a chest stethograph (Model Vt. 159; INCO, Chandigarh, India). Both HR and RR were recorded continuously during autonomic reactivity tests using a standard polyrite machine (Recorders and Medicare, Chandigarh, India).
Parasympathetic reactivity tests
For assessing parasympathetic function, changes in HR (ΔHR) and the expiration to inspiration (E:I) ratio during the deep breathing test (6 breaths/minute with equal cycles of inspiratory/expiratory phase, each for 5 seconds) were used. The ΔHR was calculated as the difference between the maximum and minimum HR, and the E:I ratio was calculated by dividing the average of the longest RR interval by the smallest RR interval during a 6-breath/minute interval. For the Valsalva maneuver, the Valsalva ratio (VR) was calculated as the ratio of the longest RR interval to the shortest RR interval after and during blowing into a mercury manometer at a pressure of up to 40 mmHg for 15 seconds.25,26 During head-up tilt test, the patients were tilted from horizontal to 70° over 15 seconds and kept tilted for 5 minutes. The 30:15 ratio was calculated as the ratio of the longest RR interval (around the 30th beat) to the shortest RR interval (around the 15th beat) on the ECG tracing obtained using the polyrite machine immediately after the 70° tilt in the head-up tilt test.
Sympathetic reactivity tests
For assessing sympathetic function, a hand grip test was performed by using a hand grip dynamometer (Jetter and Scheerer, Tuttlingen, Germany) at 30% of the patient’s maximum voluntary contraction for up to 4 minutes, and BP was measured on the contra-lateral arm.25,26 A maximum increase in diastolic BP (ΔDBP) from baseline was noted. The cold pressor test, in which the patient immersed his/her hand in cold water (10°C) for 1 minute, was also assessed by noting ΔDBP from baseline. For assessment using the head-up tilt test, a maximum fall in systolic BP from baseline was noted for up to 5 minutes after tilting using the intervals of 0.5 minute, 1 minute, 2.5 minutes, and 5 minutes.25,26
Using a modified Engel scale, the surgical outcome was assessed at 6 months and 12 months after surgery.27 Seizure details were documented when patients visited for autonomic function testing at 6 months and over telephonic communication at 12 months. For the current study, class I patients on the Engle scale were seizure-free or had aura only, class II patients had rare seizures (> 95% reduction), class III patients had 80–94% reduction in seizures, and class IV patients had no worthwhile improvement (< 80% reduction in seizure activity).
Statistical analysis
Continuous variables were summarized as means ± standard errors, and categorical data were summarized as frequencies. A 2-sample t-test was used to compare the mean age, duration of epilepsy, seizure frequency, and number of AEDs given between the LTLE and RTLE groups. The categorical variables of Engel class and number of male patients in the 2 groups were analyzed using the χ2 test. Normality of the autonomic variables was verified using the Kolmogorov-Smirnoff test. To evaluate differences in the groups at the 3 time points (before surgery and at 3 and 6 months after surgery), random effects models (group × time interaction for all normally distributed variables) were built using the time of the measurement as a repeated factor. This analysis was performed for resting autonomic variables (systolic BP, diastolic BP, HR, RR), parasympathetic reactivity test variables (ΔHR during deep breathing test, E:I ratio during deep breathing test, VR during Valsalva maneuver, 30:15 during head-up tilt test), and sympathetic reactivity test variables (ΔDBP during hand grip test and cold pressor test). Most of the patients in our study had mesial temporal sclerosis (MTS). Because uniformity of condition may provide more information, we performed a subgroup analysis of left MTS (LMTS) and right MTS (RMTS) patients only. Therefore, a similar analysis was performed for all the above variables to evaluate differences in the LMTS and RMTS subgroups at the 3 time points. To evaluate the difference in the fall of systolic BP during the head-up tilt test between the LTLE and RTLE groups separately at the 3 time points, a Wilcoxon rank sum test was performed. A within-group analysis was performed to compare the fall in systolic BP at the 3 time points separately in each group (LTLE and RTLE) using Friedman’s test, and a similar protocol was followed for the LMTS and RMTS subgroups. Post-hoc analyses for within-group comparisons at various time points and between groups were performed using Scheffe’s adjustment to type I error. A p value of < 0.05 was considered statistically significant. Data were analyzed utilizing SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).
Results
Twenty-three LTLE patients (14 men, 9 women; age, 23.56 ± 11.31 years) and 30 RTLE patients (18 men, 12 women; age, 22.4 ± 12.29 years) completed the study. No differences were found in the mean age (p = 0.53) or in the sex distribution between the 2 groups (p = 0.90). The demographic characteristics of the patients are summarized in Table 1. No differences were found in the type of seizure, duration of epilepsy, seizure frequency, Engel class, and number of AEDs between the groups. Sixteen patients had LMTS, and 23 had RMTS. The etiology in the 2 groups and the type of TLE surgery performed is presented in Table 2.
No significant group × time interaction was found for resting autonomic CV variables: systolic BP (p = 0.82), diastolic BP (p = 0.37), HR (p = 0.09), and RR (p = 0.97). Summary data for the autonomic CV function parameters are presented in Table 3. When comparing the LMTS and RMTS subgroups, no significant group × time interaction was found: systolic BP (p = 0.87), diastolic BP (p = 0.92), HR (p = 0.23), and RR (p = 0.84).
For the variables of parasympathetic reactivity, a significant group × time interaction was found during the deep breathing test for ΔHR p < 0.001) and the E:I ratio (p = 0.03). Before surgery, ΔHR and E:I (both p < 0.001) were higher in the LTLE group than in the RTLE group (Table 4). At 3 and 6 months after surgery, ΔHR during the deep breathing test for the LTLE group significantly decreased compared to the values before surgery (both p < 0.01). The E:I ratio for the LTLE group significantly decreased after 3 months compared to the ratio before surgery (p = 0.04). For the LMTS and RMTS subgroup analysis, a significant group × time interaction was found during the deep breathing test for ΔHR (p < 0.001) and E:I (p = 0.046). Before surgery, ΔHR and E:I (both p < 0.001) were higher in the LMTS subgroup than in the RMTS subgroup. After 3 months, ΔHR for the LMTS group had significantly decreased from that before surgery (p = 0.02). No significant group × time interaction was found for the VR during the Valsalva maneuver (p = 0.09) or 30:15 ratio during the head-up tilt test (p = 0.16).
For the variables of sympathetic reactivity, no significant group × time interaction was found for ΔDBP during the hand grip test for the LTLE and RTLE groups (p = 0.46) or the LMTS and RMTS subgroups (p = 0.71). A significant group × time interaction was found during the cold pressor test for the LTLE and RTLE groups (p = 0.005). At 6 months, ΔDBP during the cold pressor test was significantly lower (p = 0.002) in the RTLE group than in the LTLE group. The ΔDBP was also significantly lower at 6 months than that before surgery in the RTLE group (p = 0.001; Table 5). When comparing the LMTS and RMTS subgroups, a significant group × time interaction was found for ΔDBP during the cold pressor test (p = 0.009). At 6 months, ΔDBP (p = 0.002) was lower in the RMTS group than in the LMTS group and was significantly decreased from that before surgery (p = 0.02) in the RMTS group. Before surgery, the fall in systolic BP during head-up tilt test was higher in the LTLE group than in the RTLE group (p = 0.002). No within-group differences at any time point were found. In the subgroup analysis, differences were found between the LMTS and RMTS groups before surgery for the head-up tilt test (p = 0.008); the fall in systolic BP was lower in the RMTS subgroup than in the LMTS subgroup. However, differences were not significant 3 or 6 months after surgery. In the within-group analysis, no difference was found in the fall in systolic BP during the head-up tilt test between the LMTS (p = 0.27) and RMTS (p = 0.53) subgroups at any of the 3 time points.
Eight patients in the LTLE group and 14 patients in the RTLE group were unable to perform the hand grip test and cold pressor test according to the protocol. During the hand grip test with these patients, BP was measured 2–4 minutes after the patient pressed the hand grip dynamometer rather than at the full 4 minutes. During the cold pressor test, these patients were unable to keep their hand in cold water (10°C) for 1 minute, and BP was thus measured approximately 25–35 seconds after the patient’s hand was placed in the cold water.
Discussion
The current study showed that hemispheric sidedness of the seizure onset zone had a differential effect on autonomic CV functions in patients with TLE. We found that before surgery, patients having Seizure foci in the left temporal lobe had higher parasympathetic reactivity than that of patients with seizure foci in the right temporal lobe. After surgery, parasympathetic reactivity was lower in the LTLE group and LMTS subgroup than that before surgery. Moreover, at 3 and 6 months after surgery, the LTLE and RTLE groups and the LMTS and RMTS subgroups had comparable parasympathetic reactivity. Before surgery, patients with seizure foci in the right temporal lobe had a higher sympathetic reactivity than patients with seizure foci in the left temporal lobe. On the basis of these findings, we suggest that LTLE increases the parasympathetic response to HR, and TLE surgery reduces this heightened response. Therefore, LTLE surgery may help in stabilizing parasympathetic functions. We also suggest that epilepsy in the right temporal lobe increases the sympathetic response to BP. Thus, RTLE surgery may reduce this heightened response, although additional studies are required to substantiate this finding.
In general, the HR and the E:I ratio are robust markers of parasympathetic reactivity.28,29 In the current study, higher values of these variables in the LTLE group and LMTS subgroup indicated a higher parasympathetic response to HR than that of the RTLE group and RMTS subgroup, respectively. However, these variables decreased after surgery and became comparable among the groups. These findings suggested that patients with LTLE have heightened parasympathetic control of HR, which may cause bradycardia, hypotension, and asystole during the ictal period of seizures, as reported in various studies.30,31 Furthermore, our results suggested that TLE surgery in the LTLE group and LMTS subgroup may reduce these cardiac risks and improve autonomic function in patients with TLE.
A fall in systolic BP during the head-up tilt test is a marker of sympathetic reactivity. A lower fall of systolic BP suggests higher sympathetic activation during the head-up tilt test.32,33 In the current study, the smaller fall in systolic BP during the head-up tilt test in the RTLE group suggested higher sympathetic function than that in the LTLE group. However, both groups showed similar sympathetic reactivity after surgery. Furthermore, the fall in systolic BP in the RTLE group showed a tendency to increase after surgery. This finding suggests that RTLE surgery mitigates the heightened sympathetic function. Similarly, ΔDBP during the cold pressor test is a marker of sympathetic function, wherein a higher ΔDBP indicates increased sympathetic function.34 In the current study, the LTLE and RTLE groups had comparable ΔDBP during the cold pressor test before surgery, but ΔDBP significantly decreased at 6 months in the RTLE group and RMTS subgroup compared to that before surgery. Additionally, ΔDBP was significantly lower at 6 months after surgery in the RTLE groups than in the LTLE group and LMTS subgroup. This finding also indicated that sympathetic reactivity was reduced after RTLE surgery.
Consistent with our present findings, a previous study35 examining patients with hippocampal sclerosis associated with TLE found that before surgery, patients with hippocampal sclerosis and LTLE had elevated parasympathetic response, while those with RTLE had elevated sympathetic response, which mediated excessive bradycardia and tachycardia counter-regulation. To understand the mechanism of autonomic modulation in patients with LTLE and RTLE, we included 2 time points after surgery and found that surgical removal of the affected lesion or area influenced autonomic CV modulation differentially. The LTLE group and LMTS subgroup showed a significant decrease in parasympathetic response after surgery, while the RTLE group and RMTS subgroup tended to show a decrease in heightened sympathetic responses, which may consequently decrease the risk of cardiac emergencies in TLE.
Changes in autonomic CV modulation at different time points after surgery have been reported in previous studies.36–40 One study37 reported sympathetic over-activity 1 month after a temporal lobectomy, and other studies36,40 demonstrated a decrease in sympathetic CV modulation in patients with TLE at 3–4 months after surgery. Previous findings have also suggested that patients with good outcomes 12 months after surgery exhibited a decrease in sympathetic CV modulation at 6 months, while patients with poor outcomes showed no change.38,39 In the current study, all patients were seizure-free at 12 months after surgery. We found a change in parasympathetic modulation as early as 3 months after surgery, and a change in sympathetic modulation was also observed at 6 months after surgery. This variation in the changes in autonomic function at different time points may have been affected by various factors like seizure severity,5 epileptic drugs,7,11 type of surgery, the actual surgery, number of patients, and the methods used to assess the autonomic functions. However, other unknown and uncontrolled factors may have also influenced these results. Future studies should be conducted with a longer follow-up period to understand the effect of time on this variation.
In the current study, only ΔHR and E:I ratio during the deep breathing test showed significant interaction for the parasympathetic response. The VR assessed using the Valsalva maneuver and 30:15 ratio during the head-up tilt test, which are markers of parasympathetic function, showed no significant interaction of group and time. Because the VR indicates the interaction of sympathetic, parasympathetic, and baroreflex response,41 this versatility may explain our results for this variable. We suggest that future studies use different methods for assessing autonomic functions, such as HR variability and baroreflex sensitivity.
Our findings on the sympathetic response were based on a single autonomic function test, namely, the head-up tilt test. The other autonomic function tests for sympathetic reactivity we used were not successful because patients had difficulty performing the hand grip test, which requires some effort, and the cold pressor test, which requires the ability to tolerate cold. These challenges during the hand grip and cold pressor tests limited the value of our results with reference to non-baroreflex-based autonomic responses. Therefore, future studies should consider using a robust sympathetic activity marker, such as muscle sympathetic nerve activity, to assess sympathetic function in conjunction with simultaneous EEG recordings.
Previous studies have suggested that AEDs like carbamazepine may suppress parasympathetic and sympathetic functions in patients with epilepsy.7,11 However, another study found that carbamazepine had no association with decreased autonomic functions in patients with epilepsy.42 In the current study, 14 patients in the LTLE group and 30 in the RTLE group were administered carbamazepine, which may have influenced our results.
Because a majority of the patients with epilepsy in the current study had MTS, we performed a subgroup analysis (LMTS and RMTS) to determine if these subgroups had distinct results from the main TLE groups (LTLE and RTLE) that included all types of insults in the temporal lobe. We found that the parasympathetic and sympathetic modulations were similar in both groups (TLE groups and MTS subgroups) before surgery. After surgery, there were changes in the significance within the MTS subgroups although the mean values were still comparable with those of the main group (TLE). We believe this change in significance in the MTS subgroups was due to the fewer number of patients in those subgroups, which contributed to the high variation in the mean values. In the temporal lobe, all parts of the structure are interconnected and associated with the modulation of the autonomic nervous system;3,22 any insult in one region may affect the function of other neighboring structures. This interconnectedness could be one of the reasons for the observed comparable results between the TLE groups and MTS subgroups. A larger study with a larger sample size is required to confirm if these changes are maintained in the MTS subgroups.
The epileptogenic discharges that arise from the temporal lobe during the ictal or interictal period of seizures may interfere with autonomic CV modulation because the interconnected structures of the temporal lobe, such as the insular cortex, amygdala, cingulate gyrus, and hippocampus, and their pathways play an important role in the regulation of autonomic CV modulation.3,22 Similarly, surgical removal of the seizure foci may alter autonomic CV functions. Studies demonstrated an increase in cerebral glucose metabolism after TLE surgery in areas of the autonomic nervous system.43,44 However, the mechanism of epileptogenic activity in TLE before and after surgery is unclear. The pathophysiology and clinical manifestations associated with TLE are very complex and may represent the combined effect of hyper-excitation observed with seizures and hypo-function from loss of neuronal tissue in the affected brain regions.
The observed heightened parasympathetic and sympathetic responses before epilepsy surgery in the current study may be attributed to the presence of subclinical seizures during autonomic function tests. It has been reported that even minimal epileptogenic activity can alter cardiac neural discharge and arrhythmia.12 Therefore, any subclinical epileptogenic activity may alter the functions of different areas of the brain, such as the cerebral cortex or amygdala, and may result in changes in the cardiac autonomic function. We could not use EEG in the present study because it was not feasible during testing; thus, we conducted the tests at least 12 hours after the last seizure (according to the patient’s verbal report) to obtain information about subclinical epileptogenic activities.
In summary, the comparatively higher parasympathetic response in the LTLE group and the higher sympathetic response in the RTLE group suggested lateralization of the hemispheric center of autonomic CV controls in patients with epilepsy. Furthermore, TLE surgery seemed to stabilize autonomic CV regulation by reducing the parasympathetic CV responses in the LTLE group and LMTS subgroup; the results also indicated that surgery may reduce sympathetic CV responses in the RTLE group and RMTS subgroup. Given these results, we believe that lower parasympathetic outflow in LTLE and sympathetic outflow in RTLE should be accompanied by a reduced risk of CV emergencies after surgery in patients with TLE.
