Seizure outcomes of supratentorial brain tumor resection in pediatric patients

Seizure outcomes of supratentorial brain tumor resection in pediatric patients Abstract Background This study aims to identify the prevalence of and risk factors for seizure development after supratentorial brain tumor resection in pediatric patients. This could be used to guide the postoperative management and usage of anti-epileptic drugs (AEDs). Methods Retrospective study was conducted for patients between 0 and 21 years with supratentorial tumor resection between 2005 and 2015 at a single institution. Results Two hundred patients (114 males/86 females) were identified. Median age at resection (±SD) was 9.025 ± 5.720 years and mean follow-up was 4 ± 2 years. Resection was gross total in 82 patients (41%) and partial in 118 patients (59%); 66 patients (33%) experienced preoperative seizures, and 67 patients (34%) experienced postoperative seizures; 18 patients (27%) had early seizures, and 49 patients (73%) had late seizures. Univariate analysis identified risk factors for postoperative seizures as: preoperative seizures (P < 0.001), age less than 2 years (P = 0.003), temporal location (P < 0.001), thalamic location (P = 0.017), preoperative hyponatremia (P = 0.017), World Health Organization grade (P = 0.008), and pathology (P = 0.005). Multivariate regression identified 5 robust risk factors: temporal location (odds ratio [OR] 4.7, 95% CI: 1.7–13.3, P = 0.003), age <2 years (OR 3.9, 95% CI: 1.0–15.4; P = 0.049), preoperative hydrocephalus (OR 3.8, 95% CI: 1.5–9.4; P = 0.005), preoperative seizure (OR 2.8, 95% CI: 1.2–6.5; P = 0.016) and parietal location (OR 0.25, 95% CI: 0.06–0.99; P = 0.049). Extent of resection did not correlate with seizure development (P > 0.05). Conclusions This study reveals 5 risk factors for postoperative seizures after resection of supratentorial tumors. These factors should be considered in postoperative management of these patients. antiepileptic drugs, pediatric tumors, postoperative management, seizure outcome, supratentorial tumors At least 10% of children with intracranial central nervous system (CNS) tumors suffer from seizures. These seizures may occur preoperatively, postoperatively, or both. Postoperative seizures are classified as early (occurring within one week of surgery) or late.1 In all cases, uncontrolled seizures in young children can have severe effects on cognitive and psychosocial development. Despite the importance of this problem, seizure rates following brain tumor resections in children have not been fully evaluated. Furthermore, the use of anti-epileptic drugs (AEDs) in these patients remains unclear. Randomized clinical trials investigating the practice of prophylactic prescription of AEDs in adults have shown a decreased rate in the onset of early seizures.2 No study has shown a beneficial effect for AEDs for late-onset seizures in children or adults. It seems that a presumptive extrapolation of these results to pediatric patients has been made, notwithstanding the absence of reliable confirmatory evidence.2 Anti-epileptic drugs can have several side effects, including diplopia, dizziness, hyponatremia, nausea, sedation, gingival hyperplasia, tremor, weight gain, nephrolithiasis, and others.3,4 These drugs may provide the patient with a seizure-free recovery but also carry the potential to worsen his/her quality of life. To help guide the management of these patients, this study examines the prevalence of postoperative seizures in children who have had resection of supratentorial tumors, and identifies risk factors for postoperative seizure development in these patients. Methods Approval from the institutional review board was obtained prior to initiation of this study. The hospital’s electronic medical records were retrospectively screened using code from the International Classification of Diseases, Ninth Revision, for pediatric patients between the ages of 0 and 21 years with supratentorial tumor resection occurring at the institution between 2005 and 2015. Patients undergoing biopsy-only procedures were excluded. AEDs were not used prophylactically in seizure-free patients undergoing resection for supratentorial tumor. Predictor and Outcome Variables All inpatient and outpatient electronic medical records were reviewed. Data recorded included gender, age, tumor location, extent of resection, pathology, presence and timing of seizures, chemotherapy if given, radiation therapy if given, presence of hydrocephalus and sodium levels. Age less than 2 years old was distinguished as a categorical variable as patients in this group are considered a unique clinical entity due to immature neurocognitive status, late clinical presentation, special histological findings, and poor surgical morbidity.5 Seizures were stratified by onset (early and late) and type (focal or generalized). Extent of Resection Tumor characteristics and extent of resection were assessed by a neuroradiologist at the time of surgery based on comparison of pre- and postoperative MRI. Gross total resection (GTR) was defined as complete resection of the preoperative enhancing area with no residual enhancement on MRI. The extent of resection was recorded from postoperative and follow-up MRI reports. Tumor Location Tumors were classified into 12 intracranial locations: sellar or suprasellar, lateral ventricle, third ventricle, corpus callosum, thalamus, pineal, basal ganglia, and choroid plexus, in addition to the 4 cerebral lobes—frontal, temporal, parietal, and occipital. Each tumor location was considered an independent binary outcome, since many patients had tumors in more than one supratentorial location. Histopathology Tumors were grouped into 9 histopathological groups based on the 2016 World Health Organization (WHO) classification of CNS tumors. These were astrocytic and oligodendroglial tumors (diffuse astrocytic and oligodendroglial tumors and other astrocytic tumors), ependymal tumors, choroid plexus tumors, neuronal and mixed neuronal-glial tumors, tumors of the pineal gland, embryonal tumors, germ cell tumors, tumors of the sellar region and other gliomas. Other gliomas (N = 1) included one angiocentric glioma as per WHO classification. Classification of Postoperative Sodium Levels Sodium levels in serum before and after surgery were classified according to the range established by our institution’s department of pathology. Any value less than 135 mEq/L was considered hyponatremia and any value above 145 mEq/L was considered hypernatremia. Sodium values between 135 mEq/L and 145 mEq/L were considered normal. All recorded sodium measurements from the admission for surgical resection were included. Statistical Analysis Bivariate analysis of categorical outcome variables was performed using chi-square contingency tables or Fisher’s exact test, while Student’s t-test, ANOVA, and the Mann–Whitney test were used for continuous outcome variables. P-values ≤0.05 were considered significant. A bivariate logistic regression was conducted for significant variables (P-value ≤0.05) and borderline significant variables, and crude odds ratios (ORs) with their 95% confidence intervals (CIs) were generated. A stepwise logistic regression was then used to generate a final model starting with all significant and borderline significant variables. Borderline significance was set at a cutoff P-value of 0.2, in order to be inclusive and avoid any missed confounding variables. Adjusted ORs and their CIs were reported. Data were analyzed using the Statistical Package for the Social Sciences version 22.0. Results Demographics and Treatment Status Two hundred pediatric patients met the criteria for inclusion in this study. Demographics and treatment status are reported in Table 1. Female:male ratio was approximately 3:4. Median age (± SD) at surgery was 9.03 ± 5.72 years (range 1.2 mo—23.5 y), with a mean follow-up of 4 ± 2 years (range from initial visit). Twenty-five patients (13%) were under 2 years of age. Prior to attempted GTR, 30 (15%) patients received chemotherapy and 12 (6%) received radiation therapy. Gross total resection was confirmed in 82 (41%) cases. Table 1 Patient demographics and treatment status Variable  All Patients No. (%)*  Postoperative Seizure No. (N, %)  P-value  Total cases  200  67 (200, 34)    Sex  114 M (57)  39 M (114, 34)  0.8801    86 F (43)  28 F (86, 33)    Age in years (median ± SD)  9.03 ± 5.72  8.49 ± 6.24  0.1683  Age less than 2 years  25 (13)  15 (60)  0.0031  Preoperative radiation  12 (6)  3 (25)  0.5572  Preoperative chemotherapy  30 (15)  12 (40)  0.5301  Surgical resection  82 GTR (41)  27 GTR (82, 33)  0.8861    118 P (59)  40 P (118, 34)    Variable  All Patients No. (%)*  Postoperative Seizure No. (N, %)  P-value  Total cases  200  67 (200, 34)    Sex  114 M (57)  39 M (114, 34)  0.8801    86 F (43)  28 F (86, 33)    Age in years (median ± SD)  9.03 ± 5.72  8.49 ± 6.24  0.1683  Age less than 2 years  25 (13)  15 (60)  0.0031  Preoperative radiation  12 (6)  3 (25)  0.5572  Preoperative chemotherapy  30 (15)  12 (40)  0.5301  Surgical resection  82 GTR (41)  27 GTR (82, 33)  0.8861    118 P (59)  40 P (118, 34)    *Percentage is reported with respect to the total cohort. M: males; F: females; GTR: gross-total resection; P: partial resection; No.: number. 1 Pearson chi-square test. 2 Fisher’s exact test. 3 Student’s t-test. View Large Table 1 Patient demographics and treatment status Variable  All Patients No. (%)*  Postoperative Seizure No. (N, %)  P-value  Total cases  200  67 (200, 34)    Sex  114 M (57)  39 M (114, 34)  0.8801    86 F (43)  28 F (86, 33)    Age in years (median ± SD)  9.03 ± 5.72  8.49 ± 6.24  0.1683  Age less than 2 years  25 (13)  15 (60)  0.0031  Preoperative radiation  12 (6)  3 (25)  0.5572  Preoperative chemotherapy  30 (15)  12 (40)  0.5301  Surgical resection  82 GTR (41)  27 GTR (82, 33)  0.8861    118 P (59)  40 P (118, 34)    Variable  All Patients No. (%)*  Postoperative Seizure No. (N, %)  P-value  Total cases  200  67 (200, 34)    Sex  114 M (57)  39 M (114, 34)  0.8801    86 F (43)  28 F (86, 33)    Age in years (median ± SD)  9.03 ± 5.72  8.49 ± 6.24  0.1683  Age less than 2 years  25 (13)  15 (60)  0.0031  Preoperative radiation  12 (6)  3 (25)  0.5572  Preoperative chemotherapy  30 (15)  12 (40)  0.5301  Surgical resection  82 GTR (41)  27 GTR (82, 33)  0.8861    118 P (59)  40 P (118, 34)    *Percentage is reported with respect to the total cohort. M: males; F: females; GTR: gross-total resection; P: partial resection; No.: number. 1 Pearson chi-square test. 2 Fisher’s exact test. 3 Student’s t-test. View Large Seizure Prevalence Overall 99 patients (49.5%) experienced seizures—either before surgery, after surgery, or both. Of these ninety-nine, 32 (16% of total cohort) had strictly preoperative seizures, 33 (16.5%) had strictly postoperative seizures, while 34 patients experienced seizures both prior to and following resection. The median time to first postoperative seizure was 3.4 months (range 1 d‒9 y). Postoperative seizures were classified by seizure type and onset—44 patients (22%) experienced focal seizures, 25 (13%) experienced generalized, and 2 (1%) experienced both types. Information from the records of 1 patient was insufficient to determine seizure type. Among the 67 patients with postoperative seizures, 49 (73%) had late-onset seizures and 18 (27%) had early-onset seizures. Age The median age of patients (± SD) with postoperative seizures (8.49 ± 6.24 y) did not differ significantly from that of seizure-free patients (9.67 ± 5.42 y; P = 0.167, Mann‒Whitney U test). Stratifying patients according to seizure onset, patients with early seizures (n = 18; median age = 5.75 ± 6.67 y) were on average almost 4 years younger at surgery than those without (n = 182; median = 9.62 ± 5.52 y; P = 0.006, univariate ANOVA), while median age did not differ regarding patients with late-onset seizures (median = 8.48 ± 6.02 y; P = 0.213, univariate ANOVA; data not shown). Stratifying according to seizure type, patients with focal postoperative seizures (n = 44; median = 7.72 ± 5.47 y) were on average 2 years younger than those not developing focal postoperative seizures (n = 155; median = 9.67 ± 5.73 y; P = 0.045, univariate ANOVA; data not shown). Twenty-five of the total 200 (13%) tumor patients were under 2 years of age (Table 1). Patients in this subgroup experienced postoperative seizures at double the rate of those older than 2 (60% vs 30%, respectively; P = 0.003; Pearson chi-square test). The proportion of patients under 2 with early-onset seizures was 7-fold that of those older than 2 (36% vs 5%, respectively; P < 0.001; Fisher’s exact test). Preoperative Seizures Thirty-four of 66 patients (52%) presenting with preoperative seizures developed postoperative seizures (P < 0.001; Pearson chi-square test). Twenty-four patients (36%) with preoperative seizures had late postoperative seizures (P = 0.006; Pearson chi-square test), and 26 patients (39%) had partial seizures (P < 0.001; Pearson chi-square test), while no association was found with generalized seizures (P = 0.497; Pearson chi-square test; Table 2). Stratifying patients according to seizure type, 18 (56%) with focal preoperative seizures had postoperative seizures (P = 0.004; Pearson chi-square test), which represents nearly double the rate of those without focal preoperative seizures (29.7%). Seventeen patients (52%) with generalized preoperative seizures had postoperative seizures (P = 0.020; Pearson chi-square test), compared with those without generalized postoperative seizures (50 patients, 30.5%). Interestingly, most patients with focal preoperative seizures (n = 17, 53%) developed focal postoperative seizures (P < 0.001; Pearson chi-square test), and patients with generalized preoperative seizures (n = 8, 24%) had higher rates of generalized postoperative seizures compared with those with no prior generalized seizures (n = 17, 10%; P = 0.043; Fisher’s exact test; data not shown). Table 2 Associations of postoperative seizures, onset, and type with binary independent variables Variable  POS No. (%)**  P-value1  Early No. (%)  P-value1  Late No. (%)  P-value1  Partial No. (%)  P-value1  GEN No. (%)  P-value1  Age less than two  15 (60)  0.003*  9 (36)  <0.001*2  6 (24)  0.999  9 (36)  0.074  6 (24)  0.0982  Preoperative seizures  34 (52)  <0.001*  10 (15)  0.033*  24 (36)  0.006*  26 (39)  <0.001*  10 (15)  0.497  Preoperative hydrocephalus  27 (42)  0.095  12 (19)  0.001*  15 (23)  0.861  12 (19)  0.470  15 (23)  0.001*  Preoperative hyponatremia  5 (83)  0.017*2  0 (0)  1.0002  5 (83)  0.004*2  3 (50)  0.1252  2 (33)  0.1672  Temporal lobe location  22 (65)  <0.001*  5 (15)  0.1992  17 (50)  <0.001*  19 (56)  <0.001*  5 (15)  0.7762  Thalamic location  5 (83)  0.017*2  1 (17)  0.4362  4 (67)  0.033*2  2 (33)  0.6162  3 (50)  0.027*2  Sellar/supracellar location  12 (17)  <0.001*  5 (7)  0.610  7 (10)  <0.001*  7 (10)  0.002*  6 (9)  0.266  Parietal lobe location  5 (20)  0.126  1 (4)  0.7062  5 (20)  0.334  3 (12)  0.193  2 (8)  0.7472  Variable  POS No. (%)**  P-value1  Early No. (%)  P-value1  Late No. (%)  P-value1  Partial No. (%)  P-value1  GEN No. (%)  P-value1  Age less than two  15 (60)  0.003*  9 (36)  <0.001*2  6 (24)  0.999  9 (36)  0.074  6 (24)  0.0982  Preoperative seizures  34 (52)  <0.001*  10 (15)  0.033*  24 (36)  0.006*  26 (39)  <0.001*  10 (15)  0.497  Preoperative hydrocephalus  27 (42)  0.095  12 (19)  0.001*  15 (23)  0.861  12 (19)  0.470  15 (23)  0.001*  Preoperative hyponatremia  5 (83)  0.017*2  0 (0)  1.0002  5 (83)  0.004*2  3 (50)  0.1252  2 (33)  0.1672  Temporal lobe location  22 (65)  <0.001*  5 (15)  0.1992  17 (50)  <0.001*  19 (56)  <0.001*  5 (15)  0.7762  Thalamic location  5 (83)  0.017*2  1 (17)  0.4362  4 (67)  0.033*2  2 (33)  0.6162  3 (50)  0.027*2  Sellar/supracellar location  12 (17)  <0.001*  5 (7)  0.610  7 (10)  <0.001*  7 (10)  0.002*  6 (9)  0.266  Parietal lobe location  5 (20)  0.126  1 (4)  0.7062  5 (20)  0.334  3 (12)  0.193  2 (8)  0.7472  * Statistically significant and P < 0.050. ** Percentage is reported relative to the total count of each listed variable (row). No.: number, POS: postoperative seizures, GEN: generalized seizures. 1 Pearson chi-square test. 2 Fisher’s exact test. View Large Table 2 Associations of postoperative seizures, onset, and type with binary independent variables Variable  POS No. (%)**  P-value1  Early No. (%)  P-value1  Late No. (%)  P-value1  Partial No. (%)  P-value1  GEN No. (%)  P-value1  Age less than two  15 (60)  0.003*  9 (36)  <0.001*2  6 (24)  0.999  9 (36)  0.074  6 (24)  0.0982  Preoperative seizures  34 (52)  <0.001*  10 (15)  0.033*  24 (36)  0.006*  26 (39)  <0.001*  10 (15)  0.497  Preoperative hydrocephalus  27 (42)  0.095  12 (19)  0.001*  15 (23)  0.861  12 (19)  0.470  15 (23)  0.001*  Preoperative hyponatremia  5 (83)  0.017*2  0 (0)  1.0002  5 (83)  0.004*2  3 (50)  0.1252  2 (33)  0.1672  Temporal lobe location  22 (65)  <0.001*  5 (15)  0.1992  17 (50)  <0.001*  19 (56)  <0.001*  5 (15)  0.7762  Thalamic location  5 (83)  0.017*2  1 (17)  0.4362  4 (67)  0.033*2  2 (33)  0.6162  3 (50)  0.027*2  Sellar/supracellar location  12 (17)  <0.001*  5 (7)  0.610  7 (10)  <0.001*  7 (10)  0.002*  6 (9)  0.266  Parietal lobe location  5 (20)  0.126  1 (4)  0.7062  5 (20)  0.334  3 (12)  0.193  2 (8)  0.7472  Variable  POS No. (%)**  P-value1  Early No. (%)  P-value1  Late No. (%)  P-value1  Partial No. (%)  P-value1  GEN No. (%)  P-value1  Age less than two  15 (60)  0.003*  9 (36)  <0.001*2  6 (24)  0.999  9 (36)  0.074  6 (24)  0.0982  Preoperative seizures  34 (52)  <0.001*  10 (15)  0.033*  24 (36)  0.006*  26 (39)  <0.001*  10 (15)  0.497  Preoperative hydrocephalus  27 (42)  0.095  12 (19)  0.001*  15 (23)  0.861  12 (19)  0.470  15 (23)  0.001*  Preoperative hyponatremia  5 (83)  0.017*2  0 (0)  1.0002  5 (83)  0.004*2  3 (50)  0.1252  2 (33)  0.1672  Temporal lobe location  22 (65)  <0.001*  5 (15)  0.1992  17 (50)  <0.001*  19 (56)  <0.001*  5 (15)  0.7762  Thalamic location  5 (83)  0.017*2  1 (17)  0.4362  4 (67)  0.033*2  2 (33)  0.6162  3 (50)  0.027*2  Sellar/supracellar location  12 (17)  <0.001*  5 (7)  0.610  7 (10)  <0.001*  7 (10)  0.002*  6 (9)  0.266  Parietal lobe location  5 (20)  0.126  1 (4)  0.7062  5 (20)  0.334  3 (12)  0.193  2 (8)  0.7472  * Statistically significant and P < 0.050. ** Percentage is reported relative to the total count of each listed variable (row). No.: number, POS: postoperative seizures, GEN: generalized seizures. 1 Pearson chi-square test. 2 Fisher’s exact test. View Large Tumor Location Table 3 displays the breakdown of supratentorial tumors by 12 intracranial locations. Of the cerebral lobes, the frontal (18%) and temporal (17%) were most commonly involved, with the occipital least involved (6%). Fifteen (8%) patients had a lesion extending into more than one lobe, while 20% of our total cohort had tumors infiltrating 2 or more of the 8 remaining supratentorial locations. Table 3 Tumor location Tumor Location  All Patient No. (%)**  Postoperative Seizures No. (%)***  P-value1  Frontal  36 (18)  15 (42)  0.329  Temporal  34 (17)  22 (65)  <0.001*  Parietal  25 (13)  5 (20)  0.126  Occipital  12 (6)  3 (25)  0.7542  Sellar/ suprasellar  70 (35)  12 (17)  <0.001*  Pineal  11 (6)  3 (27)  0.7542  Thalamus  6 (3)  5 (83)  0.017*2  Corpus callosum  1 (0.5)  0 (0)  1.0002  Third ventricle  13 (7)  5 (39)  0.7642  Basal ganglia  1 (0.5)  0 (0)  1.0002  Lateral ventricle  21 (11)  8 (38)  0.808  Choroid plexus  13 (7)  5 (39)  0.7642  More than one location  39 (20)  14 (36)  1.000  More than one cerebral lobe  15 (8)  5 (33)  0.989  Tumor Location  All Patient No. (%)**  Postoperative Seizures No. (%)***  P-value1  Frontal  36 (18)  15 (42)  0.329  Temporal  34 (17)  22 (65)  <0.001*  Parietal  25 (13)  5 (20)  0.126  Occipital  12 (6)  3 (25)  0.7542  Sellar/ suprasellar  70 (35)  12 (17)  <0.001*  Pineal  11 (6)  3 (27)  0.7542  Thalamus  6 (3)  5 (83)  0.017*2  Corpus callosum  1 (0.5)  0 (0)  1.0002  Third ventricle  13 (7)  5 (39)  0.7642  Basal ganglia  1 (0.5)  0 (0)  1.0002  Lateral ventricle  21 (11)  8 (38)  0.808  Choroid plexus  13 (7)  5 (39)  0.7642  More than one location  39 (20)  14 (36)  1.000  More than one cerebral lobe  15 (8)  5 (33)  0.989  * Statistically significant and P < 0.050. ** Percentage is reported with respect to the total cohort (N = 200). *** Percentage is reported relative to the total count of each listed tumor location. No.: number 1 Pearson chi-square test. 2 Fisher’s exact test View Large Table 3 Tumor location Tumor Location  All Patient No. (%)**  Postoperative Seizures No. (%)***  P-value1  Frontal  36 (18)  15 (42)  0.329  Temporal  34 (17)  22 (65)  <0.001*  Parietal  25 (13)  5 (20)  0.126  Occipital  12 (6)  3 (25)  0.7542  Sellar/ suprasellar  70 (35)  12 (17)  <0.001*  Pineal  11 (6)  3 (27)  0.7542  Thalamus  6 (3)  5 (83)  0.017*2  Corpus callosum  1 (0.5)  0 (0)  1.0002  Third ventricle  13 (7)  5 (39)  0.7642  Basal ganglia  1 (0.5)  0 (0)  1.0002  Lateral ventricle  21 (11)  8 (38)  0.808  Choroid plexus  13 (7)  5 (39)  0.7642  More than one location  39 (20)  14 (36)  1.000  More than one cerebral lobe  15 (8)  5 (33)  0.989  Tumor Location  All Patient No. (%)**  Postoperative Seizures No. (%)***  P-value1  Frontal  36 (18)  15 (42)  0.329  Temporal  34 (17)  22 (65)  <0.001*  Parietal  25 (13)  5 (20)  0.126  Occipital  12 (6)  3 (25)  0.7542  Sellar/ suprasellar  70 (35)  12 (17)  <0.001*  Pineal  11 (6)  3 (27)  0.7542  Thalamus  6 (3)  5 (83)  0.017*2  Corpus callosum  1 (0.5)  0 (0)  1.0002  Third ventricle  13 (7)  5 (39)  0.7642  Basal ganglia  1 (0.5)  0 (0)  1.0002  Lateral ventricle  21 (11)  8 (38)  0.808  Choroid plexus  13 (7)  5 (39)  0.7642  More than one location  39 (20)  14 (36)  1.000  More than one cerebral lobe  15 (8)  5 (33)  0.989  * Statistically significant and P < 0.050. ** Percentage is reported with respect to the total cohort (N = 200). *** Percentage is reported relative to the total count of each listed tumor location. No.: number 1 Pearson chi-square test. 2 Fisher’s exact test View Large Patients with tumors located in the temporal lobe (n = 34) experienced postoperative seizures (n = 22; 65%) at a rate nearly double that of patients with nontemporal tumors (n = 45; 27%; P < 0.001; Pearson chi-square test). Six (3%) patients of the overall cohort had lesions located in the thalamus, of which 5 patients (83%) had postoperative seizures (P = 0.017; Fisher’s exact test; Table 2). As would be expected, patients with sellar and suprasellar tumors had low seizure rates. Patients with sellar/suprasellar tumors accounted for 35% (n = 70) of our overall cohort (Table 3). Of these, 12 patients (17%) had postoperative seizures, a rate 2½ times less than that of patients with nonsellar tumors (n = 55, 42%; P < 0.001; Pearson chi-square test). Patients with lesions in the frontal lobe (n = 36, 18% of total) experienced late-onset seizures at a higher rate than those without (n = 14, 39% vs n = 35, 21%, respectively; P = 0.027; Pearson chi-square test). No correlation was found between frontal tumors and seizure type (data not shown). The overall post-resection seizure rate did not vary according to any of the remaining 8 locations (Table 3). Seizure onset and type did feature some additional associations with tumor location. Patients with either multiple-location tumors (P = 1.000; Pearson chi-square test) or tumors extending into more than one cerebral lobe (P = 0.989; Pearson chi-square test) experienced seizures at rates similar to those of their counterparts, regardless of onset and type. Tumor Grade Table 4 displays the statistical breakdown of our cohort by 4 WHO grades. More than half (62%) of all patients had grade I tumors, 14% had grade IV tumors. The proportion of postoperative seizure increased in a stepwise fashion according to WHO grade (Table 4), with grade IV patients experiencing more than double the rate of grade I patients (54% and 25%, respectively; P = 0.008; Pearson chi-square test; Table 4). Table 4 Association of postoperative seizure rate with different WHO grades WHO Grade  All Patients  Postoperative Seizures No. (%)**  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Grade I  124 (62)  31 (25)  9 (7)  22 (18)  20 (16)  12 (10)  Grade II  21 (10.5)  8 (38)  2 (10)  6 (29)  8 (38)  1 (5)  Grade III  27 (13)  13 (48)  2 (7)  11 (41)  7 (26)  6 (22)  Grade IV  28 (14)  15 (54)  5 (18)  10 (36)  9 (32)  6 (21)  P-value    0.008*1  0.3732  0.029*1  0.0532  0.1002  WHO Grade  All Patients  Postoperative Seizures No. (%)**  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Grade I  124 (62)  31 (25)  9 (7)  22 (18)  20 (16)  12 (10)  Grade II  21 (10.5)  8 (38)  2 (10)  6 (29)  8 (38)  1 (5)  Grade III  27 (13)  13 (48)  2 (7)  11 (41)  7 (26)  6 (22)  Grade IV  28 (14)  15 (54)  5 (18)  10 (36)  9 (32)  6 (21)  P-value    0.008*1  0.3732  0.029*1  0.0532  0.1002  * Statistically significant and P < 0.050. ** Percentage is reported relative to the total count of each WHO grade. No.: number. 1 Pearson chi-square test. 2 Fisher’s exact test. View Large Table 4 Association of postoperative seizure rate with different WHO grades WHO Grade  All Patients  Postoperative Seizures No. (%)**  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Grade I  124 (62)  31 (25)  9 (7)  22 (18)  20 (16)  12 (10)  Grade II  21 (10.5)  8 (38)  2 (10)  6 (29)  8 (38)  1 (5)  Grade III  27 (13)  13 (48)  2 (7)  11 (41)  7 (26)  6 (22)  Grade IV  28 (14)  15 (54)  5 (18)  10 (36)  9 (32)  6 (21)  P-value    0.008*1  0.3732  0.029*1  0.0532  0.1002  WHO Grade  All Patients  Postoperative Seizures No. (%)**  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Grade I  124 (62)  31 (25)  9 (7)  22 (18)  20 (16)  12 (10)  Grade II  21 (10.5)  8 (38)  2 (10)  6 (29)  8 (38)  1 (5)  Grade III  27 (13)  13 (48)  2 (7)  11 (41)  7 (26)  6 (22)  Grade IV  28 (14)  15 (54)  5 (18)  10 (36)  9 (32)  6 (21)  P-value    0.008*1  0.3732  0.029*1  0.0532  0.1002  * Statistically significant and P < 0.050. ** Percentage is reported relative to the total count of each WHO grade. No.: number. 1 Pearson chi-square test. 2 Fisher’s exact test. View Large Tumor Pathology Table 5 displays the statistical breakdown of our cohort by tumor pathology. Astrocytic and oligodendroglial tumors were most commonly represented in our cohort, present in 76 (38%) patients, followed by sellar tumors (n = 41, 20.5%). Pineal (n = 3, 2.5%) and other gliomas (n = 1, 0.5%) were the least commonly represented. Table 5 Association of postoperative seizure rate with tumor pathology Tumor Pathology  All patients No. (%)  Postoperative seizures No. (%)*  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Astrocytic and oligodendroglial  76 (38)  30 (40)  5 (7)  27 (36)  23 (31)  7 (9)  Ependymal  10 (5)  4 (40)  0 (0)  4 (40)  3 (30)  1 (10)  Choroid plexus  13 (6.5)  5 (39)  4 (31)  2 (15)  3 (23)  2 (15)  Neuronal and mixed neuronal-glial  24 (12)  8 (33)  0 (0)  5 (21)  5 (21)  4 (17)  Pineal  3 (1.5)  2 (67)  0 (0)  2 (67)  1 (33)  1(33)  Embryonal  16 (8)  10 (63)  4 (25)  10 (63)  5 (31)  5 (31)  Germ cell  16 (8)  3 (19)  2 (13)  3 (19)  1 (6)  2 (13)  Sellar  41 (20.5)  5 (12)  3 (7)  4 (10)  3 (7)  3 (7)  Other gliomas  1 (0.5)  0 (0)  0 (0)  0 (0)  0 (0)  0 (0)  P-value**    0.005  0.036  0.001  0.063  0.269  Tumor Pathology  All patients No. (%)  Postoperative seizures No. (%)*  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Astrocytic and oligodendroglial  76 (38)  30 (40)  5 (7)  27 (36)  23 (31)  7 (9)  Ependymal  10 (5)  4 (40)  0 (0)  4 (40)  3 (30)  1 (10)  Choroid plexus  13 (6.5)  5 (39)  4 (31)  2 (15)  3 (23)  2 (15)  Neuronal and mixed neuronal-glial  24 (12)  8 (33)  0 (0)  5 (21)  5 (21)  4 (17)  Pineal  3 (1.5)  2 (67)  0 (0)  2 (67)  1 (33)  1(33)  Embryonal  16 (8)  10 (63)  4 (25)  10 (63)  5 (31)  5 (31)  Germ cell  16 (8)  3 (19)  2 (13)  3 (19)  1 (6)  2 (13)  Sellar  41 (20.5)  5 (12)  3 (7)  4 (10)  3 (7)  3 (7)  Other gliomas  1 (0.5)  0 (0)  0 (0)  0 (0)  0 (0)  0 (0)  P-value**    0.005  0.036  0.001  0.063  0.269  * Percentage is reported relative to the total count of each listed tumor pathology. ** Fisher’s exact test. No.: number. Astrocytic and oligodendroglial tumors: anaplastic oligoastrocytoma, anaplastic oligodendroglioma, anaplastic astrocytoma, pilocytic astrocytoma, pleomorphic xanthoastrocytoma, subependymal giant cell astrocytoma, diffuse astrocytoma, glioblastoma, pilomyxoid astrocytoma, germistocytic astrocytoma. Ependymal tumors: anaplastic ependymoma. Choroid plexus tumors: choroid plexus papilloma, choroid plexus carcinoma. Neuronal and mixed neuronal-glial tumors: anaplastic neurocytoma, ganglioglioma, anaplastic ganglioglioma, dysembryoplastic neuroepithelial tumor. Tumors of the pineal gland: pineoblastoma. Embryonal tumors: atypical teratoid/rhabdoid tumor, primitive neuroectodermal tumor. Germ cell tumors: teratoma with malignant transformation, germinoma. Tumors of the sellar region: craniopharyngioma. Other gliomas: angiocentric glioma. View Large Table 5 Association of postoperative seizure rate with tumor pathology Tumor Pathology  All patients No. (%)  Postoperative seizures No. (%)*  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Astrocytic and oligodendroglial  76 (38)  30 (40)  5 (7)  27 (36)  23 (31)  7 (9)  Ependymal  10 (5)  4 (40)  0 (0)  4 (40)  3 (30)  1 (10)  Choroid plexus  13 (6.5)  5 (39)  4 (31)  2 (15)  3 (23)  2 (15)  Neuronal and mixed neuronal-glial  24 (12)  8 (33)  0 (0)  5 (21)  5 (21)  4 (17)  Pineal  3 (1.5)  2 (67)  0 (0)  2 (67)  1 (33)  1(33)  Embryonal  16 (8)  10 (63)  4 (25)  10 (63)  5 (31)  5 (31)  Germ cell  16 (8)  3 (19)  2 (13)  3 (19)  1 (6)  2 (13)  Sellar  41 (20.5)  5 (12)  3 (7)  4 (10)  3 (7)  3 (7)  Other gliomas  1 (0.5)  0 (0)  0 (0)  0 (0)  0 (0)  0 (0)  P-value**    0.005  0.036  0.001  0.063  0.269  Tumor Pathology  All patients No. (%)  Postoperative seizures No. (%)*  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Astrocytic and oligodendroglial  76 (38)  30 (40)  5 (7)  27 (36)  23 (31)  7 (9)  Ependymal  10 (5)  4 (40)  0 (0)  4 (40)  3 (30)  1 (10)  Choroid plexus  13 (6.5)  5 (39)  4 (31)  2 (15)  3 (23)  2 (15)  Neuronal and mixed neuronal-glial  24 (12)  8 (33)  0 (0)  5 (21)  5 (21)  4 (17)  Pineal  3 (1.5)  2 (67)  0 (0)  2 (67)  1 (33)  1(33)  Embryonal  16 (8)  10 (63)  4 (25)  10 (63)  5 (31)  5 (31)  Germ cell  16 (8)  3 (19)  2 (13)  3 (19)  1 (6)  2 (13)  Sellar  41 (20.5)  5 (12)  3 (7)  4 (10)  3 (7)  3 (7)  Other gliomas  1 (0.5)  0 (0)  0 (0)  0 (0)  0 (0)  0 (0)  P-value**    0.005  0.036  0.001  0.063  0.269  * Percentage is reported relative to the total count of each listed tumor pathology. ** Fisher’s exact test. No.: number. Astrocytic and oligodendroglial tumors: anaplastic oligoastrocytoma, anaplastic oligodendroglioma, anaplastic astrocytoma, pilocytic astrocytoma, pleomorphic xanthoastrocytoma, subependymal giant cell astrocytoma, diffuse astrocytoma, glioblastoma, pilomyxoid astrocytoma, germistocytic astrocytoma. Ependymal tumors: anaplastic ependymoma. Choroid plexus tumors: choroid plexus papilloma, choroid plexus carcinoma. Neuronal and mixed neuronal-glial tumors: anaplastic neurocytoma, ganglioglioma, anaplastic ganglioglioma, dysembryoplastic neuroepithelial tumor. Tumors of the pineal gland: pineoblastoma. Embryonal tumors: atypical teratoid/rhabdoid tumor, primitive neuroectodermal tumor. Germ cell tumors: teratoma with malignant transformation, germinoma. Tumors of the sellar region: craniopharyngioma. Other gliomas: angiocentric glioma. View Large The patterns of overall postoperative seizure, seizure onset, and seizure type varied with respect to tumor pathology (Table 5). Patients with embyronal tumors and pineal tumors experienced postoperative seizures at the highest rate (n = 10, 62.5% and n = 2, 66.7%, respectively; P = 0.005; Fisher’s exact test). In concert with our findings on fewer seizures in patients with tumors located in the sellar or suprasellar region, those with craniopharyngiomas experienced both the lowest overall and late-onset seizure rate (n = 5, 12.2% and n = 4, 9.8%, respectively; P = 0.005 and P = 0.001; Fisher’s exact test; Table 5). Hydrocephalus Seventy-five of 200 patients in our total cohort were diagnosed with hydrocephalus (Table 2). Hydrocephalus onset was preoperative in 65/75 patients and postoperative in 57. Preoperative hydrocephalus (n = 27, 42%) was not significantly associated with postoperative seizures compared with patients without preoperative hydrocephalus (n = 40, 30%; P = 0.095; Pearson chi-square test). Stratifying patients according to seizure onset, the rate of early seizure onset was 14% greater in patients with preoperative hydrocephalus than in those without (P = 0.001; Pearson chi-square test). Stratifying patients according to seizure type, 15 of 64 patients with preoperative hydrocephalus (23%; one case with insufficient data) had generalized postoperative seizures in contrast to those not developing preoperative hydrocephalus (n = 10, 7%; P = 0.001; Pearson chi-square test; Table 2). Concerning postoperative hydrocephalus, 28 patients (49%) had postoperative seizures compared with those without hydrocephalus (n = 39, 27%; P = 0.003; Pearson chi-square test). Seizures tended to occur late after resection in patients with postoperative hydrocephalus (n = 22, 39%; P = 0.046; Pearson chi-square test; data not shown). Sodium Levels Pre- and postoperative sodium measurements were available for 199 of 200 patients. The mean preoperative sodium level (± SD) was 140.30 ± 3.35 mEq/L and the mean postoperative level was 142.00 ± 10.29 mEq/L. Median pre- and postoperative sodium levels were both 140 mEq/L. Patients with postoperative seizures had lower preoperative sodium levels (139.63 ± 4.174 mEq/L vs 140.65 ± 2.804 mEq/L; P = 0.041). Patients with late postoperative seizures (139.67 ± 4.34 mEq/L) had less than one unit lower preoperative sodium levels than those without (140.56 ± 2.84 mEq/L; P = 0.088). Patients with partial/focal postoperative seizures (139.25 ± 3.78 mEq/L) did feature a mean preoperative sodium reading (still within the normal range) 2 units lower than patients without (140.58 ± 3.16 mEq/L; P = 0.02). Similarly, patients with postoperative seizures had around 3 units lower postoperative sodium levels (140.04 ± 9.16 mEq/L) compared with patients without postoperative seizures (142.98 ± 10.68 mEq/L; P = 0.056). Again, patients with partial/focal postoperative seizures (138.32 ± 7.46 mEq/L) featured a mean postoperative sodium reading 5 units lower those of their counterparts (143.03 ± 10.75 mEq/L; P = 0.007). Long-term sodium levels did not vary according to seizure onset or type (P > 0.05). Student’s t-test was used for all continuous sodium levels. Sodium levels were also assessed categorically—below, above, or within the normal range of 135–145 mEq/L. Data were available for 199 patients. Twelve patients were each noted to have sodium irregularities preoperatively: 6 (3%) below and 6 above the normal range. Sixty-one patients had sodium irregularities after surgery: 24 (12%) below and 36 above. Five patients (83%) with preoperative hyponatremia experienced postoperative seizures (P = 0.017; Fisher’s exact test), with all 5 having late-onset (P = 0.004; Fisher’s exact test). Neither seizure type nor onset was associated with hypernatremia prior to surgery (P = 0.675; Pearson chi-square test). Regarding postoperative sodium irregularities, our categorical analyses tended to concur with the quantitative: 11 patients (44%) with postoperative hyponatremia developed postoperative seizures compared with 56 patients (32%) with no hyponatremia (P = 0.234; Pearson chi-square test). Results were not statistically significant even when stratified according to type (P = 0.203; Pearson chi-square test, P = 1.000; Fisher’s exact test for focal and generalized seizures, respectively) or onset (P = 0.664, P = 0.191 for late and early seizures, respectively; Pearson chi-square test; data not shown). Long-term sodium measurements assessed categorically did not vary by seizure onset or type (P > 0.200; Pearson chi-square test). Preoperative Chemotherapy Ninety-five of 200 total patients received chemotherapy, 30 before and 65 after resection of tumor. Among patients receiving preoperative chemotherapy, 12 (40%) developed postoperative seizures (P = 0.530; Pearson chi-square test), 4 with (13%; P = 0.485; Fisher’s exact test) early onset, 8 (27%; P = 0.819; Pearson chi-square test) with late onset, 8 (27%; P = 0.633; Pearson chi-square test) with partial/focal seizures, and 5 (17%, P = 0.548; Fisher’s exact test) with generalized seizures. Additional Factors Preoperative radiation therapy, extent of resection, and sex did not have any associations with overall postoperative seizures, seizure onset, or seizure type (all P > 0.20; Pearson chi-square test). Multifactorial Analysis In order to identify robust risk factors for the development of postoperative seizure, a series of binary logistic regressions were done. Multiple permutations of each model and models of highest overall value are presented here. Factors included were: preoperative seizures, age less than 2 years, tumor pathology (all pathologies included), location (all locations included), preoperative hydrocephalus, and preoperative hyponatremia. Table 6 displays the results of this model assessing overall seizure status post-resection. Including 199 patients, the model was statistically significant (P < 0.0001) with an overall correct classification rate of 80%, fourteen points greater than that of the null. Of the 9 factors included, 5 remained independently associated with the binary outcome: temporal location (OR 4.7, 95% CI: 1.7–13.3, P < 0.01), age less than 2 years (OR 3.9, 95% CI: 1.0–15.4; P = 0.05), preoperative hydrocephalus (OR 3.8, 95% CI: 1.5–9.4; P < 0.01), preoperative seizure (OR 2.8, 95% CI: 1.2–6.5; P = 0.01), and parietal location (OR 0.25, 95% CI: 0.06–0.99; P = 0.05). Thalamic location (OR 9.5, 95% CI: 0.85–111.1; P = 0.07) and preoperative hyponatremia (OR 6.3, 95% CI: 0.59–66.7; P = 0.13) were not significantly associated with postoperative seizures. Based on the statistical test for goodness of fit, the multivariable analysis model fits the data adequately (Hosmer and Lemeshow goodness of fit test = 4.622 (df = 8), P-value = 0.797). Table 6 Multivariable model using binary logistic regression: overall postoperative seizures Model Parameters  n  Hosmer and Lemeshow Test  df  Significance    199  4.622  8  0.797  Variable    P-value  Odds Ratio  95% CI          Lower  Upper  Temporal location    0.003  4.7  1.7  13.3  Age less than 2 years    0.049  3.9  1.0  15.4  Preoperative hydrocephalus    0.005  3.8  1.5  9.4  Preoperative seizure    0.016  2.8  1.2  6.5  Parietal location    0.049  0.25  0.06  0.99  Thalamic location    0.068  9.5  0.85  111.1  Preoperative hyponatremia    0.127  6.3  0.59  66.7  Tumor pathology    0.325  --  --  --  Sellar or suprasellar location    0.311  --  --  --  Model Parameters  n  Hosmer and Lemeshow Test  df  Significance    199  4.622  8  0.797  Variable    P-value  Odds Ratio  95% CI          Lower  Upper  Temporal location    0.003  4.7  1.7  13.3  Age less than 2 years    0.049  3.9  1.0  15.4  Preoperative hydrocephalus    0.005  3.8  1.5  9.4  Preoperative seizure    0.016  2.8  1.2  6.5  Parietal location    0.049  0.25  0.06  0.99  Thalamic location    0.068  9.5  0.85  111.1  Preoperative hyponatremia    0.127  6.3  0.59  66.7  Tumor pathology    0.325  --  --  --  Sellar or suprasellar location    0.311  --  --  --  View Large Table 6 Multivariable model using binary logistic regression: overall postoperative seizures Model Parameters  n  Hosmer and Lemeshow Test  df  Significance    199  4.622  8  0.797  Variable    P-value  Odds Ratio  95% CI          Lower  Upper  Temporal location    0.003  4.7  1.7  13.3  Age less than 2 years    0.049  3.9  1.0  15.4  Preoperative hydrocephalus    0.005  3.8  1.5  9.4  Preoperative seizure    0.016  2.8  1.2  6.5  Parietal location    0.049  0.25  0.06  0.99  Thalamic location    0.068  9.5  0.85  111.1  Preoperative hyponatremia    0.127  6.3  0.59  66.7  Tumor pathology    0.325  --  --  --  Sellar or suprasellar location    0.311  --  --  --  Model Parameters  n  Hosmer and Lemeshow Test  df  Significance    199  4.622  8  0.797  Variable    P-value  Odds Ratio  95% CI          Lower  Upper  Temporal location    0.003  4.7  1.7  13.3  Age less than 2 years    0.049  3.9  1.0  15.4  Preoperative hydrocephalus    0.005  3.8  1.5  9.4  Preoperative seizure    0.016  2.8  1.2  6.5  Parietal location    0.049  0.25  0.06  0.99  Thalamic location    0.068  9.5  0.85  111.1  Preoperative hyponatremia    0.127  6.3  0.59  66.7  Tumor pathology    0.325  --  --  --  Sellar or suprasellar location    0.311  --  --  --  View Large Four of the identified risk factors remained significant in the early postoperative seizures model: temporal location (OR 52.6, 95% CI: 1.9–1000.0, P < 0.05), preoperative hydrocephalus (OR 14.7, 95% CI: 2.1–111.1; P = 0.01), age less than 2 years (OR 8.4, 95% CI: 1.0–71.4; P = 0.05), and preoperative seizure (OR 6.1, 95% CI: 1.6–24.4; P = 0.01). Concerning late postoperative seizures, only 2 risk factors were significant: preoperative hyponatremia (OR 25.0, 95% CI: 2.0–333.3; P = 0.01) and temporal location (OR 2.7, 95% CI: 1.0–7.0, P = 0.05; data not shown). Discussion This is one of the few studies that assess seizure development following supratentorial tumor resection in the pediatric population. The aim of this study was 2-fold: to provide an estimate of the prevalence of post-resection seizure development, and secondly, to identify demographic, pathological, and surgical risk factors that may predispose patients to develop such seizures. Together this information could be utilized to help manage these patients effectively. The findings of this study on seizure prevalence confirmed prior reports. The 16% seizure rate prior to surgery is consistent with Hardesty et al’s 12% at initial diagnosis,6 and the 16.5% strictly postoperative rate lies close to the 5%–15% postoperative seizure rates previously reported.6–12 A recent study by Oushy et al on adult patients showed that AEDs should be considered in patients with supratentorial tumors, intraoperative cortical stimulation, and subtotal resections.13 Our study further explores supratentorial tumors in a population of pediatric patients. The initial analysis identified several variables with statistically significant relationships with the primary outcome of interest, postoperative seizure development. Temporal lobe location, parietal lobe location, age less than 2 years, preoperative hydrocephalus, and preoperative seizure status were identified as statistically significant risk factors for the outcome of interest using binary logistic regression. As a continuous variable, age did not reach significance in bivariate testing of overall postoperative seizure development. With respect to seizure onset and type, patients experiencing early postoperative and partial/focal seizures were 4 and 2 years younger than their counterparts, respectively. Age less than 2 defined categorically did however display a significant (P = 0.003) association with overall postoperative seizure status. Hardesty et al reported that age less than 2 years increased the risk for developing postoperative seizures by 21 times compared with patients with age >2 years.6 These results pointed toward a more modest, but still important 4-fold risk ratio for this group. Brain tumors in patients younger than 2 years are considered a “unique realm of neuro-oncology”; rapid developmental changes concerning tissue growth concomitant with psychological maturation result in broad variations in tumor distribution and neurocognitive status in this population.5 Altogether, the increased susceptibility to seizure in patients under 2 years of age may be due in part to an immature inhibitory network contributing to an underregulated excitatory network. Temporal location of tumor mass increased the likelihood of postoperative seizure development 5-fold. According to Bonilha et al, patients with abnormalities within or in proximity to a temporal lobe subnetwork composed of the ipsilateral hippocampus, amygdala, lateral temporal gyri, and insula have an increased risk of postsurgical seizure development.14 Mahaley et al further suggest that the limbic and temporal areas feature some of the lowest thresholds for seizure induction in humans.15 Interestingly, thalamic involvement (P = 0.07) featured a massive upper limit in the confidence interval (111.1) and the largest odds ratio (9.5) of any factor for postoperative seizures. This might be due to the many-fold connections from the thalamus to the hippocampus via the mammillo-thalamic tract.16 Alternatively, the surgical approach for these intrinsic tumors may have played a significant role in the development of seizures. Considering late-onset seizures only (which in our series were predominantly of partial/focal type), thalamic involvement achieved significance as the preeminent risk factor, with an odds ratio of 12. While this small subcohort of 6 patients is consistent with the low prevalence of thalamic tumors in children,17,18 the low n does not preclude the robustness of the findings, particularly in the context of a multivariable regression. In summation, resecting a mass extending into the temporal lobe and or thalamus may disrupt this structural network enough to elevate the risk for seizure development after surgery. Two factors—sellar/suprasellar and parietal location of intracranial mass—were noted to have inverse relationships with seizure outcome, in that a greater proportion of patients with these radiographic characteristics seemed less likely to experience seizures after resection. The decrease in postoperative seizures in sellar/suprasellar tumors could be explained by the fact that these tumors are mostly managed via a transsphenoidal approach, which has less brain manipulation compared with other techniques, decreasing the risk of seizure occurrence.19,20 However, sellar/suprasellar location did not retain significance in binary logistic regression, along with the variable tumor pathology. Parietal tumor resections were shown to decrease postoperative seizures 4-fold (P = 0.049). This finding is consistent with Hardesty et al’s study wherein none of the patients with resections of parietal tumors (n = 7) developed seizures postoperatively.6 Preoperative hydrocephalus increased risk for postoperative seizure development 4-fold. Recent studies into the physiology of hydrocephalus can provide insight into this new finding. Aquaporin 4 (AQP-4), a transmembrane water channel, is involved in the clearance of fluid in the brain. In vivo EEG characterization of seizures induced via electrical stimulation of the hippocampus showed a significantly higher threshold for seizure onset in AQP-4 null mice relative to wild-type mice, causing a decrease in seizure development.21 Histopathological examination of WHO grades I–IV gliomas from 65 patients between 14 and 71 years of age (mean 50 y) found abnormally high expression of AQP-4, with the highest concentrations sourced to the peritumor area.22 Furthermore, several studies report a linkage between AQP-4 peritumor status and hydrocephalus in humans and animals.23–33 Taken together these studies and our findings point to the following model: hyperexpression of AQP-4 around the tumor in conjunction with mass effect causing obstruction of normal CSF flow contribute to a hydrocephalic state.34 A resection attempt further destabilizes regulation of fluid gradients, altogether leading to patients’ increased susceptibility to seizure onset. Hardesty et al’s 2011 study of 223 pediatric brain tumor patients identified patients with hyponatremia with at least one of 2 criteria: a urine sodium measurement under 130 mEq/L or a urine sodium level of 130–134 mEq/L acquired after a precipitous drop of more than 10 mEq/L over the prior 24 hours. The authors concluded that postoperative hyponatremia increased the risk of seizure development by 15-fold.6 Williams et al also reported that hyponatremia was found in 12% of patients and leads to postoperative seizures in 21% of patients, highlighting its importance as one of the determinants of postoperative seizures.35 This study could not identify postoperative hyponatremia as a risk factor for postoperative seizures in a population of patients with supratentorial tumors only. This discrepancy could be attributed to the inclusion of patients with infratentorial tumors in Hardesty’s and Williams’ studies. Results could not be compared in this regard, since both studies did not include any stratification of postoperative hyponatremia according to tumor location. Very few reports discuss seizure type and onset after neurosurgical procedures. Seizures occurring early after resection are more likely due to generalized brain dysfunctions such as cerebral edema, leading to generalized seizures. Those occurring late after tumor resection are more likely due to localized brain dysfunction causing focal/partial seizures. Patients recovering from a brain tumor resection are more likely to experience late-onset seizures that are partial/focal in character and more difficult to control than early-onset seizures.36 While failing to reach significance (P = 0.21), our results generally conform to these patterns: 36 patients (81%) with partial/focal seizures experienced a late-onset (20% more than those with generalized seizures, n = 15), while 10 patients (39%) with generalized seizures had an early-onset (20% more than those with focal seizures, n = 8). The differentiated coupling of seizure type with seizure onset, in addition to our multivariable regression, may be of particular interest to neurologists and neuro-oncologists responsible for long-term monitoring of the studied population. Conclusion Clinical data linking intracranial brain tumors with the onset of postoperative seizures in children is limited.37–40 This study aimed to identify the demographic, pathological, and surgical risk factors for postoperative seizure development in pediatric patients undergoing resection of supratentorial tumors, a subpopulation not previously studied. Risk factors—tested for robustness with multivariable regression and listed in order of decreasing strength of effect—included: temporal lobe location, age less than 2 years, preoperative hydrocephalus, preoperative seizure, and parietal lobe location. While these results were not surprising with respect to tumor location, seizure status, and age, the identification of hydrocephalus as a risk factor for seizure onset is a novel finding. Altogether these comprise a concise and intelligible array of factors—easily ascertained from clinical presentation and initial MRI—that empowers the patient’s team of neurosurgeons, neurologists, and neuro-oncologists to help in adjustment of treatment strategies pre- and-post-resection as well as during clinical follow-up. Institutions that promote the prophylactic prescription of AEDs to pediatric patients recovering from resection of intracranial tumor should evaluate the necessity of this practice with respect to the risk factors identified herein. Further prospective institutional studies capable of a robust analysis of the demographic, surgical, and pathological risk factors for seizures in this setting are warranted. Funding This study has no funding sources. Conflict of interest statement All authors declare no conflict of interest. References 1. Jennett WB. Early traumatic epilepsy. Definition and identity. Lancet . 1969; 1( 7604): 1023– 1025. Google Scholar CrossRef Search ADS PubMed  2. Greene S. Caring for the pediatric neurosurgical patient. Principles and Practice of Pediatric Neurosurgery . 3rd ed. New York, NY: Thieme; 2015: 38– 50. 3. St Louis EK, Louis EK. Minimizing AED adverse effects: improving quality of life in the interictal state in epilepsy care. Curr Neuropharmacol . 2009; 7( 2): 106– 114. Google Scholar CrossRef Search ADS PubMed  4. van Breemen MS, Wilms EB, Vecht CJ. Epilepsy in patients with brain tumours: epidemiology, mechanisms, and management. Lancet Neurol . 2007; 6( 5): 421– 430. Google Scholar CrossRef Search ADS PubMed  5. Moussazadeh N, Souweidane M. Brain tumors in the first two years of life. Principles and Practice of Pediatric Neurosurgery . 3rd ed. New York, NY: Thieme; 2015: 423– 444. 6. Hardesty DA, Sanborn MR, Parker WE, Storm PB. Perioperative seizure incidence and risk factors in 223 pediatric brain tumor patients without prior seizures. J Neurosurg Pediatr . 2011; 7( 6): 609– 615. Google Scholar CrossRef Search ADS PubMed  7. De Santis A, Villani R, Sinisi M, Stocchetti N, Perucca E. Add-on phenytoin fails to prevent early seizures after surgery for supratentorial brain tumors: a randomized controlled study. Epilepsia . 2002; 43( 2): 175– 182. Google Scholar CrossRef Search ADS PubMed  8. Foy PM, Copeland GP, Shaw MD. The incidence of postoperative seizures. Acta Neurochir (Wien) . 1981; 55( 3–4): 253– 264. Google Scholar CrossRef Search ADS PubMed  9. Kern K, Schebesch KM, Schlaier Jet al.   Levetiracetam compared to phenytoin for the prevention of postoperative seizures after craniotomy for intracranial tumours in patients without epilepsy. J Clin Neurosci . 2012; 19( 1): 99– 100. Google Scholar CrossRef Search ADS PubMed  10. Matthew E, Sherwin AL, Welner SA, Odusote K, Stratford JG. Seizures following intracranial surgery: incidence in the first post-operative week. Can J Neurol Sci . 1980; 7( 4): 285– 290. Google Scholar CrossRef Search ADS PubMed  11. North JB, Penhall RK, Hanieh A, Hann CS, Challen RG, Frewin DB. Postoperative epilepsy: a double-blind trial of phenytoin after craniotomy. Lancet . 1980; 1( 8165): 384– 386. Google Scholar CrossRef Search ADS PubMed  12. Suri A, Mahapatra AK, Bithal P. Seizures following posterior fossa surgery. Br J Neurosurg . 1998; 12( 1): 41– 44. Google Scholar CrossRef Search ADS PubMed  13. Oushy S, Sillau SH, Ney DEet al.   New-onset seizure during and after brain tumor excision: a risk assessment analysis. J Neurosurg . 2017: 1– 6. doi: 10.3171/2017.2.JNS162315. 14. Bonilha L, Jensen JH, Baker Net al.   The brain connectome as a personalized biomarker of seizure outcomes after temporal lobectomy. Neurology . 2015; 84( 18): 1846– 1853. Google Scholar CrossRef Search ADS PubMed  15. Mahaley MSJr, Dudka L. The role of anticonvulsant medications in the management of patients with anaplastic gliomas. Surg Neurol . 1981; 16( 6): 399– 401. Google Scholar CrossRef Search ADS PubMed  16. Carlesimo GA, Lombardi MG, Caltagirone C. Vascular thalamic amnesia: a reappraisal. Neuropsychologia . 2011; 49( 5): 777– 789. Google Scholar CrossRef Search ADS PubMed  17. Bernstein M, Hoffman HJ, Halliday WC, Hendrick EB, Humphreys RP. Thalamic tumors in children. Long-term follow-up and treatment guidelines. J Neurosurg . 1984; 61( 4): 649– 656. Google Scholar CrossRef Search ADS PubMed  18. Puget S, Crimmins DW, Garnett MRet al.   Thalamic tumors in children: a reappraisal. J Neurosurg . 2007; 106( 5 Suppl): 354– 362. Google Scholar PubMed  19. Conte V, Carrabba G, Magni Let al.   Risk of perioperative seizures in patients undergoing craniotomy with intraoperative brain mapping. Minerva Anestesiol . 2015; 81( 4): 379– 388. Google Scholar PubMed  20. Lai L, Morgan MK, Trooboff S, Harvey RJ. A systematic review of published evidence on expanded endoscopic endonasal skull base surgery and the risk of postoperative seizure. J Clin Neurosci . 2013; 20( 2): 197– 203. Google Scholar CrossRef Search ADS PubMed  21. Verkman AS, Binder DK, Bloch O, Auguste K, Papadopoulos MC. Three distinct roles of aquaporin-4 in brain function revealed by knockout mice. Biochim Biophys Acta . 2006; 1758( 8): 1085– 1093. Google Scholar CrossRef Search ADS PubMed  22. Mou K, Chen M, Mao Qet al.   AQP-4 in peritumoral edematous tissue is correlated with the degree of glioma and with expression of VEGF and HIF-alpha. J Neurooncol . 2010; 100( 3): 375– 383. Google Scholar CrossRef Search ADS PubMed  23. Badaut J, Brunet JF, Grollimund Let al.   Aquaporin 1 and aquaporin 4 expression in human brain after subarachnoid hemorrhage and in peritumoral tissue. Acta Neurochir Suppl . 2003; 86: 495– 498. Google Scholar PubMed  24. Bloch O, Auguste KI, Manley GT, Verkman AS. Accelerated progression of kaolin-induced hydrocephalus in aquaporin-4-deficient mice. J Cereb Blood Flow Metab . 2006; 26( 12): 1527– 1537. Google Scholar CrossRef Search ADS PubMed  25. Feng X, Papadopoulos MC, Liu Jet al.   Sporadic obstructive hydrocephalus in Aqp4 null mice. J Neurosci Res . 2009; 87( 5): 1150– 1155. Google Scholar CrossRef Search ADS PubMed  26. Mao X, Enno TL, Del Bigio MR. Aquaporin 4 changes in rat brain with severe hydrocephalus. Eur J Neurosci . 2006; 23( 11): 2929– 2936. Google Scholar CrossRef Search ADS PubMed  27. McAllister JP, Miller JM. Aquaporin 4 and hydrocephalus. J Neurosurg . 2006; 105 ( 6 Suppl): 457– 458; discussion 458. Google Scholar PubMed  28. Nico B, Mangieri D, Tamma Ret al.   Aquaporin-4 contributes to the resolution of peritumoural brain oedema in human glioblastoma multiforme after combined chemotherapy and radiotherapy. Eur J Cancer . 2009; 45( 18): 3315– 3325. Google Scholar CrossRef Search ADS PubMed  29. Papadopoulos MC, Verkman AS. Aquaporin-4 and brain edema. Pediatr Nephrol . 2007; 22( 6): 778– 784. Google Scholar CrossRef Search ADS PubMed  30. Paul L, Madan M, Rammling M, Behnam B, Pattisapu JV. The altered expression of aquaporin 1 and 4 in choroid plexus of congenital hydrocephalus. Cerebrospinal Fluid Res . 2009; 6 ( Suppl 1): S7. Google Scholar CrossRef Search ADS   31. Saadoun S, Papadopoulos MC, Davies DC, Krishna S, Bell BA. Aquaporin-4 expression is increased in oedematous human brain tumours. J Neurol Neurosurg Psychiatry . 2002; 72( 2): 262– 265. Google Scholar CrossRef Search ADS PubMed  32. Shen XQ, Miyajima M, Ogino I, Arai H. Expression of the water-channel protein aquaporin 4 in the H-Tx rat: possible compensatory role in spontaneously arrested hydrocephalus. J Neurosurg . 2006; 105( 6 Suppl): 459– 464. Google Scholar PubMed  33. Tourdias T, Dragonu I, Fushimi Yet al.   Aquaporin 4 correlates with apparent diffusion coefficient and hydrocephalus severity in the rat brain: a combined MRI-histological study. Neuroimage . 2009; 47( 2): 659– 666. Google Scholar CrossRef Search ADS PubMed  34. Hao W, Wu XQ, Xu RT. The molecular mechanism of aminoguanidine-mediated reduction on the brain edema after surgical brain injury in rats. Brain Res . 2009; 1282: 156– 161. Google Scholar CrossRef Search ADS PubMed  35. Williams CN, Belzer JS, Riva-Cambrin J, Presson AP, Bratton SL. The incidence of postoperative hyponatremia and associated neurological sequelae in children with intracranial neoplasms. J Neurosurg Pediatr . 2014; 13( 3): 283– 290. Google Scholar CrossRef Search ADS PubMed  36. Shorvon S, Andermann F, Guerrini R. De novo Epilepsy After Neurosurgery. The Causes of Epilepsy: Common and Uncommon Causes in Adults and Children . Cambridge: Cambridge University Press; 2011: 407. Google Scholar CrossRef Search ADS   37. Bauchet L, Rigau V, Mathieu-Daudé Het al.  ; Société Française de Neurochirurgie Pédiatrique; Société Française de Neurochirurgie; Société Française de Neuropathologie; Association des Neuro-Oncologues d’Expression Française. Clinical epidemiology for childhood primary central nervous system tumors. J Neurooncol . 2009; 92( 1): 87– 98. Google Scholar CrossRef Search ADS PubMed  38. Khan RB, Boop FA, Onar A, Sanford RA. Seizures in children with low-grade tumors: outcome after tumor resection and risk factors for uncontrolled seizures. J Neurosurg . 2006; 104 ( 6 Suppl): 377– 382. Google Scholar PubMed  39. Khan RB, Hunt DL, Boop FAet al.   Seizures in children with primary brain tumors: incidence and long-term outcome. Epilepsy Res . 2005; 64( 3): 85– 91. Google Scholar CrossRef Search ADS PubMed  40. Sogawa Y, Kan L, Levy AS, Maytal J, Shinnar S. The use of antiepileptic drugs in pediatric brain tumor patients. Pediatr Neurol . 2009; 41( 3): 192– 194. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Neuro-Oncology Oxford University Press

Seizure outcomes of supratentorial brain tumor resection in pediatric patients

Loading next page...
 
/lp/ou_press/seizure-outcomes-of-supratentorial-brain-tumor-resection-in-pediatric-cMYF9hp7DR
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1522-8517
eISSN
1523-5866
D.O.I.
10.1093/neuonc/noy026
Publisher site
See Article on Publisher Site

Abstract

Abstract Background This study aims to identify the prevalence of and risk factors for seizure development after supratentorial brain tumor resection in pediatric patients. This could be used to guide the postoperative management and usage of anti-epileptic drugs (AEDs). Methods Retrospective study was conducted for patients between 0 and 21 years with supratentorial tumor resection between 2005 and 2015 at a single institution. Results Two hundred patients (114 males/86 females) were identified. Median age at resection (±SD) was 9.025 ± 5.720 years and mean follow-up was 4 ± 2 years. Resection was gross total in 82 patients (41%) and partial in 118 patients (59%); 66 patients (33%) experienced preoperative seizures, and 67 patients (34%) experienced postoperative seizures; 18 patients (27%) had early seizures, and 49 patients (73%) had late seizures. Univariate analysis identified risk factors for postoperative seizures as: preoperative seizures (P < 0.001), age less than 2 years (P = 0.003), temporal location (P < 0.001), thalamic location (P = 0.017), preoperative hyponatremia (P = 0.017), World Health Organization grade (P = 0.008), and pathology (P = 0.005). Multivariate regression identified 5 robust risk factors: temporal location (odds ratio [OR] 4.7, 95% CI: 1.7–13.3, P = 0.003), age <2 years (OR 3.9, 95% CI: 1.0–15.4; P = 0.049), preoperative hydrocephalus (OR 3.8, 95% CI: 1.5–9.4; P = 0.005), preoperative seizure (OR 2.8, 95% CI: 1.2–6.5; P = 0.016) and parietal location (OR 0.25, 95% CI: 0.06–0.99; P = 0.049). Extent of resection did not correlate with seizure development (P > 0.05). Conclusions This study reveals 5 risk factors for postoperative seizures after resection of supratentorial tumors. These factors should be considered in postoperative management of these patients. antiepileptic drugs, pediatric tumors, postoperative management, seizure outcome, supratentorial tumors At least 10% of children with intracranial central nervous system (CNS) tumors suffer from seizures. These seizures may occur preoperatively, postoperatively, or both. Postoperative seizures are classified as early (occurring within one week of surgery) or late.1 In all cases, uncontrolled seizures in young children can have severe effects on cognitive and psychosocial development. Despite the importance of this problem, seizure rates following brain tumor resections in children have not been fully evaluated. Furthermore, the use of anti-epileptic drugs (AEDs) in these patients remains unclear. Randomized clinical trials investigating the practice of prophylactic prescription of AEDs in adults have shown a decreased rate in the onset of early seizures.2 No study has shown a beneficial effect for AEDs for late-onset seizures in children or adults. It seems that a presumptive extrapolation of these results to pediatric patients has been made, notwithstanding the absence of reliable confirmatory evidence.2 Anti-epileptic drugs can have several side effects, including diplopia, dizziness, hyponatremia, nausea, sedation, gingival hyperplasia, tremor, weight gain, nephrolithiasis, and others.3,4 These drugs may provide the patient with a seizure-free recovery but also carry the potential to worsen his/her quality of life. To help guide the management of these patients, this study examines the prevalence of postoperative seizures in children who have had resection of supratentorial tumors, and identifies risk factors for postoperative seizure development in these patients. Methods Approval from the institutional review board was obtained prior to initiation of this study. The hospital’s electronic medical records were retrospectively screened using code from the International Classification of Diseases, Ninth Revision, for pediatric patients between the ages of 0 and 21 years with supratentorial tumor resection occurring at the institution between 2005 and 2015. Patients undergoing biopsy-only procedures were excluded. AEDs were not used prophylactically in seizure-free patients undergoing resection for supratentorial tumor. Predictor and Outcome Variables All inpatient and outpatient electronic medical records were reviewed. Data recorded included gender, age, tumor location, extent of resection, pathology, presence and timing of seizures, chemotherapy if given, radiation therapy if given, presence of hydrocephalus and sodium levels. Age less than 2 years old was distinguished as a categorical variable as patients in this group are considered a unique clinical entity due to immature neurocognitive status, late clinical presentation, special histological findings, and poor surgical morbidity.5 Seizures were stratified by onset (early and late) and type (focal or generalized). Extent of Resection Tumor characteristics and extent of resection were assessed by a neuroradiologist at the time of surgery based on comparison of pre- and postoperative MRI. Gross total resection (GTR) was defined as complete resection of the preoperative enhancing area with no residual enhancement on MRI. The extent of resection was recorded from postoperative and follow-up MRI reports. Tumor Location Tumors were classified into 12 intracranial locations: sellar or suprasellar, lateral ventricle, third ventricle, corpus callosum, thalamus, pineal, basal ganglia, and choroid plexus, in addition to the 4 cerebral lobes—frontal, temporal, parietal, and occipital. Each tumor location was considered an independent binary outcome, since many patients had tumors in more than one supratentorial location. Histopathology Tumors were grouped into 9 histopathological groups based on the 2016 World Health Organization (WHO) classification of CNS tumors. These were astrocytic and oligodendroglial tumors (diffuse astrocytic and oligodendroglial tumors and other astrocytic tumors), ependymal tumors, choroid plexus tumors, neuronal and mixed neuronal-glial tumors, tumors of the pineal gland, embryonal tumors, germ cell tumors, tumors of the sellar region and other gliomas. Other gliomas (N = 1) included one angiocentric glioma as per WHO classification. Classification of Postoperative Sodium Levels Sodium levels in serum before and after surgery were classified according to the range established by our institution’s department of pathology. Any value less than 135 mEq/L was considered hyponatremia and any value above 145 mEq/L was considered hypernatremia. Sodium values between 135 mEq/L and 145 mEq/L were considered normal. All recorded sodium measurements from the admission for surgical resection were included. Statistical Analysis Bivariate analysis of categorical outcome variables was performed using chi-square contingency tables or Fisher’s exact test, while Student’s t-test, ANOVA, and the Mann–Whitney test were used for continuous outcome variables. P-values ≤0.05 were considered significant. A bivariate logistic regression was conducted for significant variables (P-value ≤0.05) and borderline significant variables, and crude odds ratios (ORs) with their 95% confidence intervals (CIs) were generated. A stepwise logistic regression was then used to generate a final model starting with all significant and borderline significant variables. Borderline significance was set at a cutoff P-value of 0.2, in order to be inclusive and avoid any missed confounding variables. Adjusted ORs and their CIs were reported. Data were analyzed using the Statistical Package for the Social Sciences version 22.0. Results Demographics and Treatment Status Two hundred pediatric patients met the criteria for inclusion in this study. Demographics and treatment status are reported in Table 1. Female:male ratio was approximately 3:4. Median age (± SD) at surgery was 9.03 ± 5.72 years (range 1.2 mo—23.5 y), with a mean follow-up of 4 ± 2 years (range from initial visit). Twenty-five patients (13%) were under 2 years of age. Prior to attempted GTR, 30 (15%) patients received chemotherapy and 12 (6%) received radiation therapy. Gross total resection was confirmed in 82 (41%) cases. Table 1 Patient demographics and treatment status Variable  All Patients No. (%)*  Postoperative Seizure No. (N, %)  P-value  Total cases  200  67 (200, 34)    Sex  114 M (57)  39 M (114, 34)  0.8801    86 F (43)  28 F (86, 33)    Age in years (median ± SD)  9.03 ± 5.72  8.49 ± 6.24  0.1683  Age less than 2 years  25 (13)  15 (60)  0.0031  Preoperative radiation  12 (6)  3 (25)  0.5572  Preoperative chemotherapy  30 (15)  12 (40)  0.5301  Surgical resection  82 GTR (41)  27 GTR (82, 33)  0.8861    118 P (59)  40 P (118, 34)    Variable  All Patients No. (%)*  Postoperative Seizure No. (N, %)  P-value  Total cases  200  67 (200, 34)    Sex  114 M (57)  39 M (114, 34)  0.8801    86 F (43)  28 F (86, 33)    Age in years (median ± SD)  9.03 ± 5.72  8.49 ± 6.24  0.1683  Age less than 2 years  25 (13)  15 (60)  0.0031  Preoperative radiation  12 (6)  3 (25)  0.5572  Preoperative chemotherapy  30 (15)  12 (40)  0.5301  Surgical resection  82 GTR (41)  27 GTR (82, 33)  0.8861    118 P (59)  40 P (118, 34)    *Percentage is reported with respect to the total cohort. M: males; F: females; GTR: gross-total resection; P: partial resection; No.: number. 1 Pearson chi-square test. 2 Fisher’s exact test. 3 Student’s t-test. View Large Table 1 Patient demographics and treatment status Variable  All Patients No. (%)*  Postoperative Seizure No. (N, %)  P-value  Total cases  200  67 (200, 34)    Sex  114 M (57)  39 M (114, 34)  0.8801    86 F (43)  28 F (86, 33)    Age in years (median ± SD)  9.03 ± 5.72  8.49 ± 6.24  0.1683  Age less than 2 years  25 (13)  15 (60)  0.0031  Preoperative radiation  12 (6)  3 (25)  0.5572  Preoperative chemotherapy  30 (15)  12 (40)  0.5301  Surgical resection  82 GTR (41)  27 GTR (82, 33)  0.8861    118 P (59)  40 P (118, 34)    Variable  All Patients No. (%)*  Postoperative Seizure No. (N, %)  P-value  Total cases  200  67 (200, 34)    Sex  114 M (57)  39 M (114, 34)  0.8801    86 F (43)  28 F (86, 33)    Age in years (median ± SD)  9.03 ± 5.72  8.49 ± 6.24  0.1683  Age less than 2 years  25 (13)  15 (60)  0.0031  Preoperative radiation  12 (6)  3 (25)  0.5572  Preoperative chemotherapy  30 (15)  12 (40)  0.5301  Surgical resection  82 GTR (41)  27 GTR (82, 33)  0.8861    118 P (59)  40 P (118, 34)    *Percentage is reported with respect to the total cohort. M: males; F: females; GTR: gross-total resection; P: partial resection; No.: number. 1 Pearson chi-square test. 2 Fisher’s exact test. 3 Student’s t-test. View Large Seizure Prevalence Overall 99 patients (49.5%) experienced seizures—either before surgery, after surgery, or both. Of these ninety-nine, 32 (16% of total cohort) had strictly preoperative seizures, 33 (16.5%) had strictly postoperative seizures, while 34 patients experienced seizures both prior to and following resection. The median time to first postoperative seizure was 3.4 months (range 1 d‒9 y). Postoperative seizures were classified by seizure type and onset—44 patients (22%) experienced focal seizures, 25 (13%) experienced generalized, and 2 (1%) experienced both types. Information from the records of 1 patient was insufficient to determine seizure type. Among the 67 patients with postoperative seizures, 49 (73%) had late-onset seizures and 18 (27%) had early-onset seizures. Age The median age of patients (± SD) with postoperative seizures (8.49 ± 6.24 y) did not differ significantly from that of seizure-free patients (9.67 ± 5.42 y; P = 0.167, Mann‒Whitney U test). Stratifying patients according to seizure onset, patients with early seizures (n = 18; median age = 5.75 ± 6.67 y) were on average almost 4 years younger at surgery than those without (n = 182; median = 9.62 ± 5.52 y; P = 0.006, univariate ANOVA), while median age did not differ regarding patients with late-onset seizures (median = 8.48 ± 6.02 y; P = 0.213, univariate ANOVA; data not shown). Stratifying according to seizure type, patients with focal postoperative seizures (n = 44; median = 7.72 ± 5.47 y) were on average 2 years younger than those not developing focal postoperative seizures (n = 155; median = 9.67 ± 5.73 y; P = 0.045, univariate ANOVA; data not shown). Twenty-five of the total 200 (13%) tumor patients were under 2 years of age (Table 1). Patients in this subgroup experienced postoperative seizures at double the rate of those older than 2 (60% vs 30%, respectively; P = 0.003; Pearson chi-square test). The proportion of patients under 2 with early-onset seizures was 7-fold that of those older than 2 (36% vs 5%, respectively; P < 0.001; Fisher’s exact test). Preoperative Seizures Thirty-four of 66 patients (52%) presenting with preoperative seizures developed postoperative seizures (P < 0.001; Pearson chi-square test). Twenty-four patients (36%) with preoperative seizures had late postoperative seizures (P = 0.006; Pearson chi-square test), and 26 patients (39%) had partial seizures (P < 0.001; Pearson chi-square test), while no association was found with generalized seizures (P = 0.497; Pearson chi-square test; Table 2). Stratifying patients according to seizure type, 18 (56%) with focal preoperative seizures had postoperative seizures (P = 0.004; Pearson chi-square test), which represents nearly double the rate of those without focal preoperative seizures (29.7%). Seventeen patients (52%) with generalized preoperative seizures had postoperative seizures (P = 0.020; Pearson chi-square test), compared with those without generalized postoperative seizures (50 patients, 30.5%). Interestingly, most patients with focal preoperative seizures (n = 17, 53%) developed focal postoperative seizures (P < 0.001; Pearson chi-square test), and patients with generalized preoperative seizures (n = 8, 24%) had higher rates of generalized postoperative seizures compared with those with no prior generalized seizures (n = 17, 10%; P = 0.043; Fisher’s exact test; data not shown). Table 2 Associations of postoperative seizures, onset, and type with binary independent variables Variable  POS No. (%)**  P-value1  Early No. (%)  P-value1  Late No. (%)  P-value1  Partial No. (%)  P-value1  GEN No. (%)  P-value1  Age less than two  15 (60)  0.003*  9 (36)  <0.001*2  6 (24)  0.999  9 (36)  0.074  6 (24)  0.0982  Preoperative seizures  34 (52)  <0.001*  10 (15)  0.033*  24 (36)  0.006*  26 (39)  <0.001*  10 (15)  0.497  Preoperative hydrocephalus  27 (42)  0.095  12 (19)  0.001*  15 (23)  0.861  12 (19)  0.470  15 (23)  0.001*  Preoperative hyponatremia  5 (83)  0.017*2  0 (0)  1.0002  5 (83)  0.004*2  3 (50)  0.1252  2 (33)  0.1672  Temporal lobe location  22 (65)  <0.001*  5 (15)  0.1992  17 (50)  <0.001*  19 (56)  <0.001*  5 (15)  0.7762  Thalamic location  5 (83)  0.017*2  1 (17)  0.4362  4 (67)  0.033*2  2 (33)  0.6162  3 (50)  0.027*2  Sellar/supracellar location  12 (17)  <0.001*  5 (7)  0.610  7 (10)  <0.001*  7 (10)  0.002*  6 (9)  0.266  Parietal lobe location  5 (20)  0.126  1 (4)  0.7062  5 (20)  0.334  3 (12)  0.193  2 (8)  0.7472  Variable  POS No. (%)**  P-value1  Early No. (%)  P-value1  Late No. (%)  P-value1  Partial No. (%)  P-value1  GEN No. (%)  P-value1  Age less than two  15 (60)  0.003*  9 (36)  <0.001*2  6 (24)  0.999  9 (36)  0.074  6 (24)  0.0982  Preoperative seizures  34 (52)  <0.001*  10 (15)  0.033*  24 (36)  0.006*  26 (39)  <0.001*  10 (15)  0.497  Preoperative hydrocephalus  27 (42)  0.095  12 (19)  0.001*  15 (23)  0.861  12 (19)  0.470  15 (23)  0.001*  Preoperative hyponatremia  5 (83)  0.017*2  0 (0)  1.0002  5 (83)  0.004*2  3 (50)  0.1252  2 (33)  0.1672  Temporal lobe location  22 (65)  <0.001*  5 (15)  0.1992  17 (50)  <0.001*  19 (56)  <0.001*  5 (15)  0.7762  Thalamic location  5 (83)  0.017*2  1 (17)  0.4362  4 (67)  0.033*2  2 (33)  0.6162  3 (50)  0.027*2  Sellar/supracellar location  12 (17)  <0.001*  5 (7)  0.610  7 (10)  <0.001*  7 (10)  0.002*  6 (9)  0.266  Parietal lobe location  5 (20)  0.126  1 (4)  0.7062  5 (20)  0.334  3 (12)  0.193  2 (8)  0.7472  * Statistically significant and P < 0.050. ** Percentage is reported relative to the total count of each listed variable (row). No.: number, POS: postoperative seizures, GEN: generalized seizures. 1 Pearson chi-square test. 2 Fisher’s exact test. View Large Table 2 Associations of postoperative seizures, onset, and type with binary independent variables Variable  POS No. (%)**  P-value1  Early No. (%)  P-value1  Late No. (%)  P-value1  Partial No. (%)  P-value1  GEN No. (%)  P-value1  Age less than two  15 (60)  0.003*  9 (36)  <0.001*2  6 (24)  0.999  9 (36)  0.074  6 (24)  0.0982  Preoperative seizures  34 (52)  <0.001*  10 (15)  0.033*  24 (36)  0.006*  26 (39)  <0.001*  10 (15)  0.497  Preoperative hydrocephalus  27 (42)  0.095  12 (19)  0.001*  15 (23)  0.861  12 (19)  0.470  15 (23)  0.001*  Preoperative hyponatremia  5 (83)  0.017*2  0 (0)  1.0002  5 (83)  0.004*2  3 (50)  0.1252  2 (33)  0.1672  Temporal lobe location  22 (65)  <0.001*  5 (15)  0.1992  17 (50)  <0.001*  19 (56)  <0.001*  5 (15)  0.7762  Thalamic location  5 (83)  0.017*2  1 (17)  0.4362  4 (67)  0.033*2  2 (33)  0.6162  3 (50)  0.027*2  Sellar/supracellar location  12 (17)  <0.001*  5 (7)  0.610  7 (10)  <0.001*  7 (10)  0.002*  6 (9)  0.266  Parietal lobe location  5 (20)  0.126  1 (4)  0.7062  5 (20)  0.334  3 (12)  0.193  2 (8)  0.7472  Variable  POS No. (%)**  P-value1  Early No. (%)  P-value1  Late No. (%)  P-value1  Partial No. (%)  P-value1  GEN No. (%)  P-value1  Age less than two  15 (60)  0.003*  9 (36)  <0.001*2  6 (24)  0.999  9 (36)  0.074  6 (24)  0.0982  Preoperative seizures  34 (52)  <0.001*  10 (15)  0.033*  24 (36)  0.006*  26 (39)  <0.001*  10 (15)  0.497  Preoperative hydrocephalus  27 (42)  0.095  12 (19)  0.001*  15 (23)  0.861  12 (19)  0.470  15 (23)  0.001*  Preoperative hyponatremia  5 (83)  0.017*2  0 (0)  1.0002  5 (83)  0.004*2  3 (50)  0.1252  2 (33)  0.1672  Temporal lobe location  22 (65)  <0.001*  5 (15)  0.1992  17 (50)  <0.001*  19 (56)  <0.001*  5 (15)  0.7762  Thalamic location  5 (83)  0.017*2  1 (17)  0.4362  4 (67)  0.033*2  2 (33)  0.6162  3 (50)  0.027*2  Sellar/supracellar location  12 (17)  <0.001*  5 (7)  0.610  7 (10)  <0.001*  7 (10)  0.002*  6 (9)  0.266  Parietal lobe location  5 (20)  0.126  1 (4)  0.7062  5 (20)  0.334  3 (12)  0.193  2 (8)  0.7472  * Statistically significant and P < 0.050. ** Percentage is reported relative to the total count of each listed variable (row). No.: number, POS: postoperative seizures, GEN: generalized seizures. 1 Pearson chi-square test. 2 Fisher’s exact test. View Large Tumor Location Table 3 displays the breakdown of supratentorial tumors by 12 intracranial locations. Of the cerebral lobes, the frontal (18%) and temporal (17%) were most commonly involved, with the occipital least involved (6%). Fifteen (8%) patients had a lesion extending into more than one lobe, while 20% of our total cohort had tumors infiltrating 2 or more of the 8 remaining supratentorial locations. Table 3 Tumor location Tumor Location  All Patient No. (%)**  Postoperative Seizures No. (%)***  P-value1  Frontal  36 (18)  15 (42)  0.329  Temporal  34 (17)  22 (65)  <0.001*  Parietal  25 (13)  5 (20)  0.126  Occipital  12 (6)  3 (25)  0.7542  Sellar/ suprasellar  70 (35)  12 (17)  <0.001*  Pineal  11 (6)  3 (27)  0.7542  Thalamus  6 (3)  5 (83)  0.017*2  Corpus callosum  1 (0.5)  0 (0)  1.0002  Third ventricle  13 (7)  5 (39)  0.7642  Basal ganglia  1 (0.5)  0 (0)  1.0002  Lateral ventricle  21 (11)  8 (38)  0.808  Choroid plexus  13 (7)  5 (39)  0.7642  More than one location  39 (20)  14 (36)  1.000  More than one cerebral lobe  15 (8)  5 (33)  0.989  Tumor Location  All Patient No. (%)**  Postoperative Seizures No. (%)***  P-value1  Frontal  36 (18)  15 (42)  0.329  Temporal  34 (17)  22 (65)  <0.001*  Parietal  25 (13)  5 (20)  0.126  Occipital  12 (6)  3 (25)  0.7542  Sellar/ suprasellar  70 (35)  12 (17)  <0.001*  Pineal  11 (6)  3 (27)  0.7542  Thalamus  6 (3)  5 (83)  0.017*2  Corpus callosum  1 (0.5)  0 (0)  1.0002  Third ventricle  13 (7)  5 (39)  0.7642  Basal ganglia  1 (0.5)  0 (0)  1.0002  Lateral ventricle  21 (11)  8 (38)  0.808  Choroid plexus  13 (7)  5 (39)  0.7642  More than one location  39 (20)  14 (36)  1.000  More than one cerebral lobe  15 (8)  5 (33)  0.989  * Statistically significant and P < 0.050. ** Percentage is reported with respect to the total cohort (N = 200). *** Percentage is reported relative to the total count of each listed tumor location. No.: number 1 Pearson chi-square test. 2 Fisher’s exact test View Large Table 3 Tumor location Tumor Location  All Patient No. (%)**  Postoperative Seizures No. (%)***  P-value1  Frontal  36 (18)  15 (42)  0.329  Temporal  34 (17)  22 (65)  <0.001*  Parietal  25 (13)  5 (20)  0.126  Occipital  12 (6)  3 (25)  0.7542  Sellar/ suprasellar  70 (35)  12 (17)  <0.001*  Pineal  11 (6)  3 (27)  0.7542  Thalamus  6 (3)  5 (83)  0.017*2  Corpus callosum  1 (0.5)  0 (0)  1.0002  Third ventricle  13 (7)  5 (39)  0.7642  Basal ganglia  1 (0.5)  0 (0)  1.0002  Lateral ventricle  21 (11)  8 (38)  0.808  Choroid plexus  13 (7)  5 (39)  0.7642  More than one location  39 (20)  14 (36)  1.000  More than one cerebral lobe  15 (8)  5 (33)  0.989  Tumor Location  All Patient No. (%)**  Postoperative Seizures No. (%)***  P-value1  Frontal  36 (18)  15 (42)  0.329  Temporal  34 (17)  22 (65)  <0.001*  Parietal  25 (13)  5 (20)  0.126  Occipital  12 (6)  3 (25)  0.7542  Sellar/ suprasellar  70 (35)  12 (17)  <0.001*  Pineal  11 (6)  3 (27)  0.7542  Thalamus  6 (3)  5 (83)  0.017*2  Corpus callosum  1 (0.5)  0 (0)  1.0002  Third ventricle  13 (7)  5 (39)  0.7642  Basal ganglia  1 (0.5)  0 (0)  1.0002  Lateral ventricle  21 (11)  8 (38)  0.808  Choroid plexus  13 (7)  5 (39)  0.7642  More than one location  39 (20)  14 (36)  1.000  More than one cerebral lobe  15 (8)  5 (33)  0.989  * Statistically significant and P < 0.050. ** Percentage is reported with respect to the total cohort (N = 200). *** Percentage is reported relative to the total count of each listed tumor location. No.: number 1 Pearson chi-square test. 2 Fisher’s exact test View Large Patients with tumors located in the temporal lobe (n = 34) experienced postoperative seizures (n = 22; 65%) at a rate nearly double that of patients with nontemporal tumors (n = 45; 27%; P < 0.001; Pearson chi-square test). Six (3%) patients of the overall cohort had lesions located in the thalamus, of which 5 patients (83%) had postoperative seizures (P = 0.017; Fisher’s exact test; Table 2). As would be expected, patients with sellar and suprasellar tumors had low seizure rates. Patients with sellar/suprasellar tumors accounted for 35% (n = 70) of our overall cohort (Table 3). Of these, 12 patients (17%) had postoperative seizures, a rate 2½ times less than that of patients with nonsellar tumors (n = 55, 42%; P < 0.001; Pearson chi-square test). Patients with lesions in the frontal lobe (n = 36, 18% of total) experienced late-onset seizures at a higher rate than those without (n = 14, 39% vs n = 35, 21%, respectively; P = 0.027; Pearson chi-square test). No correlation was found between frontal tumors and seizure type (data not shown). The overall post-resection seizure rate did not vary according to any of the remaining 8 locations (Table 3). Seizure onset and type did feature some additional associations with tumor location. Patients with either multiple-location tumors (P = 1.000; Pearson chi-square test) or tumors extending into more than one cerebral lobe (P = 0.989; Pearson chi-square test) experienced seizures at rates similar to those of their counterparts, regardless of onset and type. Tumor Grade Table 4 displays the statistical breakdown of our cohort by 4 WHO grades. More than half (62%) of all patients had grade I tumors, 14% had grade IV tumors. The proportion of postoperative seizure increased in a stepwise fashion according to WHO grade (Table 4), with grade IV patients experiencing more than double the rate of grade I patients (54% and 25%, respectively; P = 0.008; Pearson chi-square test; Table 4). Table 4 Association of postoperative seizure rate with different WHO grades WHO Grade  All Patients  Postoperative Seizures No. (%)**  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Grade I  124 (62)  31 (25)  9 (7)  22 (18)  20 (16)  12 (10)  Grade II  21 (10.5)  8 (38)  2 (10)  6 (29)  8 (38)  1 (5)  Grade III  27 (13)  13 (48)  2 (7)  11 (41)  7 (26)  6 (22)  Grade IV  28 (14)  15 (54)  5 (18)  10 (36)  9 (32)  6 (21)  P-value    0.008*1  0.3732  0.029*1  0.0532  0.1002  WHO Grade  All Patients  Postoperative Seizures No. (%)**  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Grade I  124 (62)  31 (25)  9 (7)  22 (18)  20 (16)  12 (10)  Grade II  21 (10.5)  8 (38)  2 (10)  6 (29)  8 (38)  1 (5)  Grade III  27 (13)  13 (48)  2 (7)  11 (41)  7 (26)  6 (22)  Grade IV  28 (14)  15 (54)  5 (18)  10 (36)  9 (32)  6 (21)  P-value    0.008*1  0.3732  0.029*1  0.0532  0.1002  * Statistically significant and P < 0.050. ** Percentage is reported relative to the total count of each WHO grade. No.: number. 1 Pearson chi-square test. 2 Fisher’s exact test. View Large Table 4 Association of postoperative seizure rate with different WHO grades WHO Grade  All Patients  Postoperative Seizures No. (%)**  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Grade I  124 (62)  31 (25)  9 (7)  22 (18)  20 (16)  12 (10)  Grade II  21 (10.5)  8 (38)  2 (10)  6 (29)  8 (38)  1 (5)  Grade III  27 (13)  13 (48)  2 (7)  11 (41)  7 (26)  6 (22)  Grade IV  28 (14)  15 (54)  5 (18)  10 (36)  9 (32)  6 (21)  P-value    0.008*1  0.3732  0.029*1  0.0532  0.1002  WHO Grade  All Patients  Postoperative Seizures No. (%)**  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Grade I  124 (62)  31 (25)  9 (7)  22 (18)  20 (16)  12 (10)  Grade II  21 (10.5)  8 (38)  2 (10)  6 (29)  8 (38)  1 (5)  Grade III  27 (13)  13 (48)  2 (7)  11 (41)  7 (26)  6 (22)  Grade IV  28 (14)  15 (54)  5 (18)  10 (36)  9 (32)  6 (21)  P-value    0.008*1  0.3732  0.029*1  0.0532  0.1002  * Statistically significant and P < 0.050. ** Percentage is reported relative to the total count of each WHO grade. No.: number. 1 Pearson chi-square test. 2 Fisher’s exact test. View Large Tumor Pathology Table 5 displays the statistical breakdown of our cohort by tumor pathology. Astrocytic and oligodendroglial tumors were most commonly represented in our cohort, present in 76 (38%) patients, followed by sellar tumors (n = 41, 20.5%). Pineal (n = 3, 2.5%) and other gliomas (n = 1, 0.5%) were the least commonly represented. Table 5 Association of postoperative seizure rate with tumor pathology Tumor Pathology  All patients No. (%)  Postoperative seizures No. (%)*  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Astrocytic and oligodendroglial  76 (38)  30 (40)  5 (7)  27 (36)  23 (31)  7 (9)  Ependymal  10 (5)  4 (40)  0 (0)  4 (40)  3 (30)  1 (10)  Choroid plexus  13 (6.5)  5 (39)  4 (31)  2 (15)  3 (23)  2 (15)  Neuronal and mixed neuronal-glial  24 (12)  8 (33)  0 (0)  5 (21)  5 (21)  4 (17)  Pineal  3 (1.5)  2 (67)  0 (0)  2 (67)  1 (33)  1(33)  Embryonal  16 (8)  10 (63)  4 (25)  10 (63)  5 (31)  5 (31)  Germ cell  16 (8)  3 (19)  2 (13)  3 (19)  1 (6)  2 (13)  Sellar  41 (20.5)  5 (12)  3 (7)  4 (10)  3 (7)  3 (7)  Other gliomas  1 (0.5)  0 (0)  0 (0)  0 (0)  0 (0)  0 (0)  P-value**    0.005  0.036  0.001  0.063  0.269  Tumor Pathology  All patients No. (%)  Postoperative seizures No. (%)*  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Astrocytic and oligodendroglial  76 (38)  30 (40)  5 (7)  27 (36)  23 (31)  7 (9)  Ependymal  10 (5)  4 (40)  0 (0)  4 (40)  3 (30)  1 (10)  Choroid plexus  13 (6.5)  5 (39)  4 (31)  2 (15)  3 (23)  2 (15)  Neuronal and mixed neuronal-glial  24 (12)  8 (33)  0 (0)  5 (21)  5 (21)  4 (17)  Pineal  3 (1.5)  2 (67)  0 (0)  2 (67)  1 (33)  1(33)  Embryonal  16 (8)  10 (63)  4 (25)  10 (63)  5 (31)  5 (31)  Germ cell  16 (8)  3 (19)  2 (13)  3 (19)  1 (6)  2 (13)  Sellar  41 (20.5)  5 (12)  3 (7)  4 (10)  3 (7)  3 (7)  Other gliomas  1 (0.5)  0 (0)  0 (0)  0 (0)  0 (0)  0 (0)  P-value**    0.005  0.036  0.001  0.063  0.269  * Percentage is reported relative to the total count of each listed tumor pathology. ** Fisher’s exact test. No.: number. Astrocytic and oligodendroglial tumors: anaplastic oligoastrocytoma, anaplastic oligodendroglioma, anaplastic astrocytoma, pilocytic astrocytoma, pleomorphic xanthoastrocytoma, subependymal giant cell astrocytoma, diffuse astrocytoma, glioblastoma, pilomyxoid astrocytoma, germistocytic astrocytoma. Ependymal tumors: anaplastic ependymoma. Choroid plexus tumors: choroid plexus papilloma, choroid plexus carcinoma. Neuronal and mixed neuronal-glial tumors: anaplastic neurocytoma, ganglioglioma, anaplastic ganglioglioma, dysembryoplastic neuroepithelial tumor. Tumors of the pineal gland: pineoblastoma. Embryonal tumors: atypical teratoid/rhabdoid tumor, primitive neuroectodermal tumor. Germ cell tumors: teratoma with malignant transformation, germinoma. Tumors of the sellar region: craniopharyngioma. Other gliomas: angiocentric glioma. View Large Table 5 Association of postoperative seizure rate with tumor pathology Tumor Pathology  All patients No. (%)  Postoperative seizures No. (%)*  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Astrocytic and oligodendroglial  76 (38)  30 (40)  5 (7)  27 (36)  23 (31)  7 (9)  Ependymal  10 (5)  4 (40)  0 (0)  4 (40)  3 (30)  1 (10)  Choroid plexus  13 (6.5)  5 (39)  4 (31)  2 (15)  3 (23)  2 (15)  Neuronal and mixed neuronal-glial  24 (12)  8 (33)  0 (0)  5 (21)  5 (21)  4 (17)  Pineal  3 (1.5)  2 (67)  0 (0)  2 (67)  1 (33)  1(33)  Embryonal  16 (8)  10 (63)  4 (25)  10 (63)  5 (31)  5 (31)  Germ cell  16 (8)  3 (19)  2 (13)  3 (19)  1 (6)  2 (13)  Sellar  41 (20.5)  5 (12)  3 (7)  4 (10)  3 (7)  3 (7)  Other gliomas  1 (0.5)  0 (0)  0 (0)  0 (0)  0 (0)  0 (0)  P-value**    0.005  0.036  0.001  0.063  0.269  Tumor Pathology  All patients No. (%)  Postoperative seizures No. (%)*  Early No. (%)  Late No. (%)  Partial No. (%)  Generalized No. (%)  Astrocytic and oligodendroglial  76 (38)  30 (40)  5 (7)  27 (36)  23 (31)  7 (9)  Ependymal  10 (5)  4 (40)  0 (0)  4 (40)  3 (30)  1 (10)  Choroid plexus  13 (6.5)  5 (39)  4 (31)  2 (15)  3 (23)  2 (15)  Neuronal and mixed neuronal-glial  24 (12)  8 (33)  0 (0)  5 (21)  5 (21)  4 (17)  Pineal  3 (1.5)  2 (67)  0 (0)  2 (67)  1 (33)  1(33)  Embryonal  16 (8)  10 (63)  4 (25)  10 (63)  5 (31)  5 (31)  Germ cell  16 (8)  3 (19)  2 (13)  3 (19)  1 (6)  2 (13)  Sellar  41 (20.5)  5 (12)  3 (7)  4 (10)  3 (7)  3 (7)  Other gliomas  1 (0.5)  0 (0)  0 (0)  0 (0)  0 (0)  0 (0)  P-value**    0.005  0.036  0.001  0.063  0.269  * Percentage is reported relative to the total count of each listed tumor pathology. ** Fisher’s exact test. No.: number. Astrocytic and oligodendroglial tumors: anaplastic oligoastrocytoma, anaplastic oligodendroglioma, anaplastic astrocytoma, pilocytic astrocytoma, pleomorphic xanthoastrocytoma, subependymal giant cell astrocytoma, diffuse astrocytoma, glioblastoma, pilomyxoid astrocytoma, germistocytic astrocytoma. Ependymal tumors: anaplastic ependymoma. Choroid plexus tumors: choroid plexus papilloma, choroid plexus carcinoma. Neuronal and mixed neuronal-glial tumors: anaplastic neurocytoma, ganglioglioma, anaplastic ganglioglioma, dysembryoplastic neuroepithelial tumor. Tumors of the pineal gland: pineoblastoma. Embryonal tumors: atypical teratoid/rhabdoid tumor, primitive neuroectodermal tumor. Germ cell tumors: teratoma with malignant transformation, germinoma. Tumors of the sellar region: craniopharyngioma. Other gliomas: angiocentric glioma. View Large The patterns of overall postoperative seizure, seizure onset, and seizure type varied with respect to tumor pathology (Table 5). Patients with embyronal tumors and pineal tumors experienced postoperative seizures at the highest rate (n = 10, 62.5% and n = 2, 66.7%, respectively; P = 0.005; Fisher’s exact test). In concert with our findings on fewer seizures in patients with tumors located in the sellar or suprasellar region, those with craniopharyngiomas experienced both the lowest overall and late-onset seizure rate (n = 5, 12.2% and n = 4, 9.8%, respectively; P = 0.005 and P = 0.001; Fisher’s exact test; Table 5). Hydrocephalus Seventy-five of 200 patients in our total cohort were diagnosed with hydrocephalus (Table 2). Hydrocephalus onset was preoperative in 65/75 patients and postoperative in 57. Preoperative hydrocephalus (n = 27, 42%) was not significantly associated with postoperative seizures compared with patients without preoperative hydrocephalus (n = 40, 30%; P = 0.095; Pearson chi-square test). Stratifying patients according to seizure onset, the rate of early seizure onset was 14% greater in patients with preoperative hydrocephalus than in those without (P = 0.001; Pearson chi-square test). Stratifying patients according to seizure type, 15 of 64 patients with preoperative hydrocephalus (23%; one case with insufficient data) had generalized postoperative seizures in contrast to those not developing preoperative hydrocephalus (n = 10, 7%; P = 0.001; Pearson chi-square test; Table 2). Concerning postoperative hydrocephalus, 28 patients (49%) had postoperative seizures compared with those without hydrocephalus (n = 39, 27%; P = 0.003; Pearson chi-square test). Seizures tended to occur late after resection in patients with postoperative hydrocephalus (n = 22, 39%; P = 0.046; Pearson chi-square test; data not shown). Sodium Levels Pre- and postoperative sodium measurements were available for 199 of 200 patients. The mean preoperative sodium level (± SD) was 140.30 ± 3.35 mEq/L and the mean postoperative level was 142.00 ± 10.29 mEq/L. Median pre- and postoperative sodium levels were both 140 mEq/L. Patients with postoperative seizures had lower preoperative sodium levels (139.63 ± 4.174 mEq/L vs 140.65 ± 2.804 mEq/L; P = 0.041). Patients with late postoperative seizures (139.67 ± 4.34 mEq/L) had less than one unit lower preoperative sodium levels than those without (140.56 ± 2.84 mEq/L; P = 0.088). Patients with partial/focal postoperative seizures (139.25 ± 3.78 mEq/L) did feature a mean preoperative sodium reading (still within the normal range) 2 units lower than patients without (140.58 ± 3.16 mEq/L; P = 0.02). Similarly, patients with postoperative seizures had around 3 units lower postoperative sodium levels (140.04 ± 9.16 mEq/L) compared with patients without postoperative seizures (142.98 ± 10.68 mEq/L; P = 0.056). Again, patients with partial/focal postoperative seizures (138.32 ± 7.46 mEq/L) featured a mean postoperative sodium reading 5 units lower those of their counterparts (143.03 ± 10.75 mEq/L; P = 0.007). Long-term sodium levels did not vary according to seizure onset or type (P > 0.05). Student’s t-test was used for all continuous sodium levels. Sodium levels were also assessed categorically—below, above, or within the normal range of 135–145 mEq/L. Data were available for 199 patients. Twelve patients were each noted to have sodium irregularities preoperatively: 6 (3%) below and 6 above the normal range. Sixty-one patients had sodium irregularities after surgery: 24 (12%) below and 36 above. Five patients (83%) with preoperative hyponatremia experienced postoperative seizures (P = 0.017; Fisher’s exact test), with all 5 having late-onset (P = 0.004; Fisher’s exact test). Neither seizure type nor onset was associated with hypernatremia prior to surgery (P = 0.675; Pearson chi-square test). Regarding postoperative sodium irregularities, our categorical analyses tended to concur with the quantitative: 11 patients (44%) with postoperative hyponatremia developed postoperative seizures compared with 56 patients (32%) with no hyponatremia (P = 0.234; Pearson chi-square test). Results were not statistically significant even when stratified according to type (P = 0.203; Pearson chi-square test, P = 1.000; Fisher’s exact test for focal and generalized seizures, respectively) or onset (P = 0.664, P = 0.191 for late and early seizures, respectively; Pearson chi-square test; data not shown). Long-term sodium measurements assessed categorically did not vary by seizure onset or type (P > 0.200; Pearson chi-square test). Preoperative Chemotherapy Ninety-five of 200 total patients received chemotherapy, 30 before and 65 after resection of tumor. Among patients receiving preoperative chemotherapy, 12 (40%) developed postoperative seizures (P = 0.530; Pearson chi-square test), 4 with (13%; P = 0.485; Fisher’s exact test) early onset, 8 (27%; P = 0.819; Pearson chi-square test) with late onset, 8 (27%; P = 0.633; Pearson chi-square test) with partial/focal seizures, and 5 (17%, P = 0.548; Fisher’s exact test) with generalized seizures. Additional Factors Preoperative radiation therapy, extent of resection, and sex did not have any associations with overall postoperative seizures, seizure onset, or seizure type (all P > 0.20; Pearson chi-square test). Multifactorial Analysis In order to identify robust risk factors for the development of postoperative seizure, a series of binary logistic regressions were done. Multiple permutations of each model and models of highest overall value are presented here. Factors included were: preoperative seizures, age less than 2 years, tumor pathology (all pathologies included), location (all locations included), preoperative hydrocephalus, and preoperative hyponatremia. Table 6 displays the results of this model assessing overall seizure status post-resection. Including 199 patients, the model was statistically significant (P < 0.0001) with an overall correct classification rate of 80%, fourteen points greater than that of the null. Of the 9 factors included, 5 remained independently associated with the binary outcome: temporal location (OR 4.7, 95% CI: 1.7–13.3, P < 0.01), age less than 2 years (OR 3.9, 95% CI: 1.0–15.4; P = 0.05), preoperative hydrocephalus (OR 3.8, 95% CI: 1.5–9.4; P < 0.01), preoperative seizure (OR 2.8, 95% CI: 1.2–6.5; P = 0.01), and parietal location (OR 0.25, 95% CI: 0.06–0.99; P = 0.05). Thalamic location (OR 9.5, 95% CI: 0.85–111.1; P = 0.07) and preoperative hyponatremia (OR 6.3, 95% CI: 0.59–66.7; P = 0.13) were not significantly associated with postoperative seizures. Based on the statistical test for goodness of fit, the multivariable analysis model fits the data adequately (Hosmer and Lemeshow goodness of fit test = 4.622 (df = 8), P-value = 0.797). Table 6 Multivariable model using binary logistic regression: overall postoperative seizures Model Parameters  n  Hosmer and Lemeshow Test  df  Significance    199  4.622  8  0.797  Variable    P-value  Odds Ratio  95% CI          Lower  Upper  Temporal location    0.003  4.7  1.7  13.3  Age less than 2 years    0.049  3.9  1.0  15.4  Preoperative hydrocephalus    0.005  3.8  1.5  9.4  Preoperative seizure    0.016  2.8  1.2  6.5  Parietal location    0.049  0.25  0.06  0.99  Thalamic location    0.068  9.5  0.85  111.1  Preoperative hyponatremia    0.127  6.3  0.59  66.7  Tumor pathology    0.325  --  --  --  Sellar or suprasellar location    0.311  --  --  --  Model Parameters  n  Hosmer and Lemeshow Test  df  Significance    199  4.622  8  0.797  Variable    P-value  Odds Ratio  95% CI          Lower  Upper  Temporal location    0.003  4.7  1.7  13.3  Age less than 2 years    0.049  3.9  1.0  15.4  Preoperative hydrocephalus    0.005  3.8  1.5  9.4  Preoperative seizure    0.016  2.8  1.2  6.5  Parietal location    0.049  0.25  0.06  0.99  Thalamic location    0.068  9.5  0.85  111.1  Preoperative hyponatremia    0.127  6.3  0.59  66.7  Tumor pathology    0.325  --  --  --  Sellar or suprasellar location    0.311  --  --  --  View Large Table 6 Multivariable model using binary logistic regression: overall postoperative seizures Model Parameters  n  Hosmer and Lemeshow Test  df  Significance    199  4.622  8  0.797  Variable    P-value  Odds Ratio  95% CI          Lower  Upper  Temporal location    0.003  4.7  1.7  13.3  Age less than 2 years    0.049  3.9  1.0  15.4  Preoperative hydrocephalus    0.005  3.8  1.5  9.4  Preoperative seizure    0.016  2.8  1.2  6.5  Parietal location    0.049  0.25  0.06  0.99  Thalamic location    0.068  9.5  0.85  111.1  Preoperative hyponatremia    0.127  6.3  0.59  66.7  Tumor pathology    0.325  --  --  --  Sellar or suprasellar location    0.311  --  --  --  Model Parameters  n  Hosmer and Lemeshow Test  df  Significance    199  4.622  8  0.797  Variable    P-value  Odds Ratio  95% CI          Lower  Upper  Temporal location    0.003  4.7  1.7  13.3  Age less than 2 years    0.049  3.9  1.0  15.4  Preoperative hydrocephalus    0.005  3.8  1.5  9.4  Preoperative seizure    0.016  2.8  1.2  6.5  Parietal location    0.049  0.25  0.06  0.99  Thalamic location    0.068  9.5  0.85  111.1  Preoperative hyponatremia    0.127  6.3  0.59  66.7  Tumor pathology    0.325  --  --  --  Sellar or suprasellar location    0.311  --  --  --  View Large Four of the identified risk factors remained significant in the early postoperative seizures model: temporal location (OR 52.6, 95% CI: 1.9–1000.0, P < 0.05), preoperative hydrocephalus (OR 14.7, 95% CI: 2.1–111.1; P = 0.01), age less than 2 years (OR 8.4, 95% CI: 1.0–71.4; P = 0.05), and preoperative seizure (OR 6.1, 95% CI: 1.6–24.4; P = 0.01). Concerning late postoperative seizures, only 2 risk factors were significant: preoperative hyponatremia (OR 25.0, 95% CI: 2.0–333.3; P = 0.01) and temporal location (OR 2.7, 95% CI: 1.0–7.0, P = 0.05; data not shown). Discussion This is one of the few studies that assess seizure development following supratentorial tumor resection in the pediatric population. The aim of this study was 2-fold: to provide an estimate of the prevalence of post-resection seizure development, and secondly, to identify demographic, pathological, and surgical risk factors that may predispose patients to develop such seizures. Together this information could be utilized to help manage these patients effectively. The findings of this study on seizure prevalence confirmed prior reports. The 16% seizure rate prior to surgery is consistent with Hardesty et al’s 12% at initial diagnosis,6 and the 16.5% strictly postoperative rate lies close to the 5%–15% postoperative seizure rates previously reported.6–12 A recent study by Oushy et al on adult patients showed that AEDs should be considered in patients with supratentorial tumors, intraoperative cortical stimulation, and subtotal resections.13 Our study further explores supratentorial tumors in a population of pediatric patients. The initial analysis identified several variables with statistically significant relationships with the primary outcome of interest, postoperative seizure development. Temporal lobe location, parietal lobe location, age less than 2 years, preoperative hydrocephalus, and preoperative seizure status were identified as statistically significant risk factors for the outcome of interest using binary logistic regression. As a continuous variable, age did not reach significance in bivariate testing of overall postoperative seizure development. With respect to seizure onset and type, patients experiencing early postoperative and partial/focal seizures were 4 and 2 years younger than their counterparts, respectively. Age less than 2 defined categorically did however display a significant (P = 0.003) association with overall postoperative seizure status. Hardesty et al reported that age less than 2 years increased the risk for developing postoperative seizures by 21 times compared with patients with age >2 years.6 These results pointed toward a more modest, but still important 4-fold risk ratio for this group. Brain tumors in patients younger than 2 years are considered a “unique realm of neuro-oncology”; rapid developmental changes concerning tissue growth concomitant with psychological maturation result in broad variations in tumor distribution and neurocognitive status in this population.5 Altogether, the increased susceptibility to seizure in patients under 2 years of age may be due in part to an immature inhibitory network contributing to an underregulated excitatory network. Temporal location of tumor mass increased the likelihood of postoperative seizure development 5-fold. According to Bonilha et al, patients with abnormalities within or in proximity to a temporal lobe subnetwork composed of the ipsilateral hippocampus, amygdala, lateral temporal gyri, and insula have an increased risk of postsurgical seizure development.14 Mahaley et al further suggest that the limbic and temporal areas feature some of the lowest thresholds for seizure induction in humans.15 Interestingly, thalamic involvement (P = 0.07) featured a massive upper limit in the confidence interval (111.1) and the largest odds ratio (9.5) of any factor for postoperative seizures. This might be due to the many-fold connections from the thalamus to the hippocampus via the mammillo-thalamic tract.16 Alternatively, the surgical approach for these intrinsic tumors may have played a significant role in the development of seizures. Considering late-onset seizures only (which in our series were predominantly of partial/focal type), thalamic involvement achieved significance as the preeminent risk factor, with an odds ratio of 12. While this small subcohort of 6 patients is consistent with the low prevalence of thalamic tumors in children,17,18 the low n does not preclude the robustness of the findings, particularly in the context of a multivariable regression. In summation, resecting a mass extending into the temporal lobe and or thalamus may disrupt this structural network enough to elevate the risk for seizure development after surgery. Two factors—sellar/suprasellar and parietal location of intracranial mass—were noted to have inverse relationships with seizure outcome, in that a greater proportion of patients with these radiographic characteristics seemed less likely to experience seizures after resection. The decrease in postoperative seizures in sellar/suprasellar tumors could be explained by the fact that these tumors are mostly managed via a transsphenoidal approach, which has less brain manipulation compared with other techniques, decreasing the risk of seizure occurrence.19,20 However, sellar/suprasellar location did not retain significance in binary logistic regression, along with the variable tumor pathology. Parietal tumor resections were shown to decrease postoperative seizures 4-fold (P = 0.049). This finding is consistent with Hardesty et al’s study wherein none of the patients with resections of parietal tumors (n = 7) developed seizures postoperatively.6 Preoperative hydrocephalus increased risk for postoperative seizure development 4-fold. Recent studies into the physiology of hydrocephalus can provide insight into this new finding. Aquaporin 4 (AQP-4), a transmembrane water channel, is involved in the clearance of fluid in the brain. In vivo EEG characterization of seizures induced via electrical stimulation of the hippocampus showed a significantly higher threshold for seizure onset in AQP-4 null mice relative to wild-type mice, causing a decrease in seizure development.21 Histopathological examination of WHO grades I–IV gliomas from 65 patients between 14 and 71 years of age (mean 50 y) found abnormally high expression of AQP-4, with the highest concentrations sourced to the peritumor area.22 Furthermore, several studies report a linkage between AQP-4 peritumor status and hydrocephalus in humans and animals.23–33 Taken together these studies and our findings point to the following model: hyperexpression of AQP-4 around the tumor in conjunction with mass effect causing obstruction of normal CSF flow contribute to a hydrocephalic state.34 A resection attempt further destabilizes regulation of fluid gradients, altogether leading to patients’ increased susceptibility to seizure onset. Hardesty et al’s 2011 study of 223 pediatric brain tumor patients identified patients with hyponatremia with at least one of 2 criteria: a urine sodium measurement under 130 mEq/L or a urine sodium level of 130–134 mEq/L acquired after a precipitous drop of more than 10 mEq/L over the prior 24 hours. The authors concluded that postoperative hyponatremia increased the risk of seizure development by 15-fold.6 Williams et al also reported that hyponatremia was found in 12% of patients and leads to postoperative seizures in 21% of patients, highlighting its importance as one of the determinants of postoperative seizures.35 This study could not identify postoperative hyponatremia as a risk factor for postoperative seizures in a population of patients with supratentorial tumors only. This discrepancy could be attributed to the inclusion of patients with infratentorial tumors in Hardesty’s and Williams’ studies. Results could not be compared in this regard, since both studies did not include any stratification of postoperative hyponatremia according to tumor location. Very few reports discuss seizure type and onset after neurosurgical procedures. Seizures occurring early after resection are more likely due to generalized brain dysfunctions such as cerebral edema, leading to generalized seizures. Those occurring late after tumor resection are more likely due to localized brain dysfunction causing focal/partial seizures. Patients recovering from a brain tumor resection are more likely to experience late-onset seizures that are partial/focal in character and more difficult to control than early-onset seizures.36 While failing to reach significance (P = 0.21), our results generally conform to these patterns: 36 patients (81%) with partial/focal seizures experienced a late-onset (20% more than those with generalized seizures, n = 15), while 10 patients (39%) with generalized seizures had an early-onset (20% more than those with focal seizures, n = 8). The differentiated coupling of seizure type with seizure onset, in addition to our multivariable regression, may be of particular interest to neurologists and neuro-oncologists responsible for long-term monitoring of the studied population. Conclusion Clinical data linking intracranial brain tumors with the onset of postoperative seizures in children is limited.37–40 This study aimed to identify the demographic, pathological, and surgical risk factors for postoperative seizure development in pediatric patients undergoing resection of supratentorial tumors, a subpopulation not previously studied. Risk factors—tested for robustness with multivariable regression and listed in order of decreasing strength of effect—included: temporal lobe location, age less than 2 years, preoperative hydrocephalus, preoperative seizure, and parietal lobe location. While these results were not surprising with respect to tumor location, seizure status, and age, the identification of hydrocephalus as a risk factor for seizure onset is a novel finding. Altogether these comprise a concise and intelligible array of factors—easily ascertained from clinical presentation and initial MRI—that empowers the patient’s team of neurosurgeons, neurologists, and neuro-oncologists to help in adjustment of treatment strategies pre- and-post-resection as well as during clinical follow-up. Institutions that promote the prophylactic prescription of AEDs to pediatric patients recovering from resection of intracranial tumor should evaluate the necessity of this practice with respect to the risk factors identified herein. Further prospective institutional studies capable of a robust analysis of the demographic, surgical, and pathological risk factors for seizures in this setting are warranted. Funding This study has no funding sources. Conflict of interest statement All authors declare no conflict of interest. References 1. Jennett WB. Early traumatic epilepsy. Definition and identity. Lancet . 1969; 1( 7604): 1023– 1025. Google Scholar CrossRef Search ADS PubMed  2. Greene S. Caring for the pediatric neurosurgical patient. Principles and Practice of Pediatric Neurosurgery . 3rd ed. New York, NY: Thieme; 2015: 38– 50. 3. St Louis EK, Louis EK. Minimizing AED adverse effects: improving quality of life in the interictal state in epilepsy care. Curr Neuropharmacol . 2009; 7( 2): 106– 114. Google Scholar CrossRef Search ADS PubMed  4. van Breemen MS, Wilms EB, Vecht CJ. Epilepsy in patients with brain tumours: epidemiology, mechanisms, and management. Lancet Neurol . 2007; 6( 5): 421– 430. Google Scholar CrossRef Search ADS PubMed  5. Moussazadeh N, Souweidane M. Brain tumors in the first two years of life. Principles and Practice of Pediatric Neurosurgery . 3rd ed. New York, NY: Thieme; 2015: 423– 444. 6. Hardesty DA, Sanborn MR, Parker WE, Storm PB. Perioperative seizure incidence and risk factors in 223 pediatric brain tumor patients without prior seizures. J Neurosurg Pediatr . 2011; 7( 6): 609– 615. Google Scholar CrossRef Search ADS PubMed  7. De Santis A, Villani R, Sinisi M, Stocchetti N, Perucca E. Add-on phenytoin fails to prevent early seizures after surgery for supratentorial brain tumors: a randomized controlled study. Epilepsia . 2002; 43( 2): 175– 182. Google Scholar CrossRef Search ADS PubMed  8. Foy PM, Copeland GP, Shaw MD. The incidence of postoperative seizures. Acta Neurochir (Wien) . 1981; 55( 3–4): 253– 264. Google Scholar CrossRef Search ADS PubMed  9. Kern K, Schebesch KM, Schlaier Jet al.   Levetiracetam compared to phenytoin for the prevention of postoperative seizures after craniotomy for intracranial tumours in patients without epilepsy. J Clin Neurosci . 2012; 19( 1): 99– 100. Google Scholar CrossRef Search ADS PubMed  10. Matthew E, Sherwin AL, Welner SA, Odusote K, Stratford JG. Seizures following intracranial surgery: incidence in the first post-operative week. Can J Neurol Sci . 1980; 7( 4): 285– 290. Google Scholar CrossRef Search ADS PubMed  11. North JB, Penhall RK, Hanieh A, Hann CS, Challen RG, Frewin DB. Postoperative epilepsy: a double-blind trial of phenytoin after craniotomy. Lancet . 1980; 1( 8165): 384– 386. Google Scholar CrossRef Search ADS PubMed  12. Suri A, Mahapatra AK, Bithal P. Seizures following posterior fossa surgery. Br J Neurosurg . 1998; 12( 1): 41– 44. Google Scholar CrossRef Search ADS PubMed  13. Oushy S, Sillau SH, Ney DEet al.   New-onset seizure during and after brain tumor excision: a risk assessment analysis. J Neurosurg . 2017: 1– 6. doi: 10.3171/2017.2.JNS162315. 14. Bonilha L, Jensen JH, Baker Net al.   The brain connectome as a personalized biomarker of seizure outcomes after temporal lobectomy. Neurology . 2015; 84( 18): 1846– 1853. Google Scholar CrossRef Search ADS PubMed  15. Mahaley MSJr, Dudka L. The role of anticonvulsant medications in the management of patients with anaplastic gliomas. Surg Neurol . 1981; 16( 6): 399– 401. Google Scholar CrossRef Search ADS PubMed  16. Carlesimo GA, Lombardi MG, Caltagirone C. Vascular thalamic amnesia: a reappraisal. Neuropsychologia . 2011; 49( 5): 777– 789. Google Scholar CrossRef Search ADS PubMed  17. Bernstein M, Hoffman HJ, Halliday WC, Hendrick EB, Humphreys RP. Thalamic tumors in children. Long-term follow-up and treatment guidelines. J Neurosurg . 1984; 61( 4): 649– 656. Google Scholar CrossRef Search ADS PubMed  18. Puget S, Crimmins DW, Garnett MRet al.   Thalamic tumors in children: a reappraisal. J Neurosurg . 2007; 106( 5 Suppl): 354– 362. Google Scholar PubMed  19. Conte V, Carrabba G, Magni Let al.   Risk of perioperative seizures in patients undergoing craniotomy with intraoperative brain mapping. Minerva Anestesiol . 2015; 81( 4): 379– 388. Google Scholar PubMed  20. Lai L, Morgan MK, Trooboff S, Harvey RJ. A systematic review of published evidence on expanded endoscopic endonasal skull base surgery and the risk of postoperative seizure. J Clin Neurosci . 2013; 20( 2): 197– 203. Google Scholar CrossRef Search ADS PubMed  21. Verkman AS, Binder DK, Bloch O, Auguste K, Papadopoulos MC. Three distinct roles of aquaporin-4 in brain function revealed by knockout mice. Biochim Biophys Acta . 2006; 1758( 8): 1085– 1093. Google Scholar CrossRef Search ADS PubMed  22. Mou K, Chen M, Mao Qet al.   AQP-4 in peritumoral edematous tissue is correlated with the degree of glioma and with expression of VEGF and HIF-alpha. J Neurooncol . 2010; 100( 3): 375– 383. Google Scholar CrossRef Search ADS PubMed  23. Badaut J, Brunet JF, Grollimund Let al.   Aquaporin 1 and aquaporin 4 expression in human brain after subarachnoid hemorrhage and in peritumoral tissue. Acta Neurochir Suppl . 2003; 86: 495– 498. Google Scholar PubMed  24. Bloch O, Auguste KI, Manley GT, Verkman AS. Accelerated progression of kaolin-induced hydrocephalus in aquaporin-4-deficient mice. J Cereb Blood Flow Metab . 2006; 26( 12): 1527– 1537. Google Scholar CrossRef Search ADS PubMed  25. Feng X, Papadopoulos MC, Liu Jet al.   Sporadic obstructive hydrocephalus in Aqp4 null mice. J Neurosci Res . 2009; 87( 5): 1150– 1155. Google Scholar CrossRef Search ADS PubMed  26. Mao X, Enno TL, Del Bigio MR. Aquaporin 4 changes in rat brain with severe hydrocephalus. Eur J Neurosci . 2006; 23( 11): 2929– 2936. Google Scholar CrossRef Search ADS PubMed  27. McAllister JP, Miller JM. Aquaporin 4 and hydrocephalus. J Neurosurg . 2006; 105 ( 6 Suppl): 457– 458; discussion 458. Google Scholar PubMed  28. Nico B, Mangieri D, Tamma Ret al.   Aquaporin-4 contributes to the resolution of peritumoural brain oedema in human glioblastoma multiforme after combined chemotherapy and radiotherapy. Eur J Cancer . 2009; 45( 18): 3315– 3325. Google Scholar CrossRef Search ADS PubMed  29. Papadopoulos MC, Verkman AS. Aquaporin-4 and brain edema. Pediatr Nephrol . 2007; 22( 6): 778– 784. Google Scholar CrossRef Search ADS PubMed  30. Paul L, Madan M, Rammling M, Behnam B, Pattisapu JV. The altered expression of aquaporin 1 and 4 in choroid plexus of congenital hydrocephalus. Cerebrospinal Fluid Res . 2009; 6 ( Suppl 1): S7. Google Scholar CrossRef Search ADS   31. Saadoun S, Papadopoulos MC, Davies DC, Krishna S, Bell BA. Aquaporin-4 expression is increased in oedematous human brain tumours. J Neurol Neurosurg Psychiatry . 2002; 72( 2): 262– 265. Google Scholar CrossRef Search ADS PubMed  32. Shen XQ, Miyajima M, Ogino I, Arai H. Expression of the water-channel protein aquaporin 4 in the H-Tx rat: possible compensatory role in spontaneously arrested hydrocephalus. J Neurosurg . 2006; 105( 6 Suppl): 459– 464. Google Scholar PubMed  33. Tourdias T, Dragonu I, Fushimi Yet al.   Aquaporin 4 correlates with apparent diffusion coefficient and hydrocephalus severity in the rat brain: a combined MRI-histological study. Neuroimage . 2009; 47( 2): 659– 666. Google Scholar CrossRef Search ADS PubMed  34. Hao W, Wu XQ, Xu RT. The molecular mechanism of aminoguanidine-mediated reduction on the brain edema after surgical brain injury in rats. Brain Res . 2009; 1282: 156– 161. Google Scholar CrossRef Search ADS PubMed  35. Williams CN, Belzer JS, Riva-Cambrin J, Presson AP, Bratton SL. The incidence of postoperative hyponatremia and associated neurological sequelae in children with intracranial neoplasms. J Neurosurg Pediatr . 2014; 13( 3): 283– 290. Google Scholar CrossRef Search ADS PubMed  36. Shorvon S, Andermann F, Guerrini R. De novo Epilepsy After Neurosurgery. The Causes of Epilepsy: Common and Uncommon Causes in Adults and Children . Cambridge: Cambridge University Press; 2011: 407. Google Scholar CrossRef Search ADS   37. Bauchet L, Rigau V, Mathieu-Daudé Het al.  ; Société Française de Neurochirurgie Pédiatrique; Société Française de Neurochirurgie; Société Française de Neuropathologie; Association des Neuro-Oncologues d’Expression Française. Clinical epidemiology for childhood primary central nervous system tumors. J Neurooncol . 2009; 92( 1): 87– 98. Google Scholar CrossRef Search ADS PubMed  38. Khan RB, Boop FA, Onar A, Sanford RA. Seizures in children with low-grade tumors: outcome after tumor resection and risk factors for uncontrolled seizures. J Neurosurg . 2006; 104 ( 6 Suppl): 377– 382. Google Scholar PubMed  39. Khan RB, Hunt DL, Boop FAet al.   Seizures in children with primary brain tumors: incidence and long-term outcome. Epilepsy Res . 2005; 64( 3): 85– 91. Google Scholar CrossRef Search ADS PubMed  40. Sogawa Y, Kan L, Levy AS, Maytal J, Shinnar S. The use of antiepileptic drugs in pediatric brain tumor patients. Pediatr Neurol . 2009; 41( 3): 192– 194. Google Scholar CrossRef Search ADS PubMed  © The Author(s) 2018. Published by Oxford University Press on behalf of the Society for Neuro-Oncology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com

Journal

Neuro-OncologyOxford University Press

Published: Mar 20, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off