Intracranial hemorrhage (ICH) is a risk factor for

Transcription

1 J Neurosurg Pediatrics 13: , 2014 AANS, 2014 Levetiracetam versus (fos)phenytoin for seizure prophylaxis in pediatric patients with intracranial hemorrhage Clinical article Seema Bansal, M.D., 1 Dan Blalock, M.A., 2 Tewodros Kebede, M.D., 1 Nathan P. Dean, M.D., 3 and Jessica L. Carpenter, M.D. 1 Departments of 1 Neurology and 3 Critical Care Medicine, Children s National Medical Center, Washington, DC; and 2 Department of Psychology, George Mason University, Fairfax, Virginia Object. Seizure prophylaxis is used in a variety of conditions, including supratentorial intracranial hemorrhage (ICH). In adults, studies have demonstrated phenytoin as the drug of choice for seizure prophylaxis; in children, levetiracetam is often provided due to its favorable side effect profile and pharmacokinetics. This study evaluated the difference in efficacy between these treatment options. Methods. This retrospective review included 126 patients between 1 month and 17 years of age with acute supratentorial ICH; all received seizure prophylaxis. Demographic data and outcome assessments were compared. Results. Seizure prophylaxis was provided with (fos)phenytoin in 40 children, levetiracetam in 61 children, and both drugs in 25 patients. Baseline characteristics of the treatment groups were similar, except that more patients treated with (fos)phenytoin had seizures on presentation. Patients treated solely with (fos)phenytoin had a higher probability of early seizures (within 7 days of ICH) compared with those treated only with LVT, controlling for relevant variables including seizures on presentation (OR 24.6, p = 0.002). Patients treated with (fos)phenytoin were more likely to need additional antiepileptic drugs for seizure control (p = 0.005). There was no significant difference in the incidence of late seizures (> 7 days after ICH) (p = 0.265). Adverse events necessitating a change in therapy were uncommon. Conclusions. Levetiracetam is a reasonable alternative to (fos)phenytoin for prophylaxis of early posthemorrhagic seizures. Levetiracetam and (fos)phenytoin are well tolerated in children. Prospective studies are needed to determine superiority, optimal dosing, and impact on long-term outcomes. ( Key Words intracranial hemorrhage levetiracetam phenytoin seizure prophylaxis neurocritical care epilepsy Intracranial hemorrhage (ICH) is a risk factor for acute symptomatic seizures. Seizure prophylaxis for prevention of provoked seizures after traumatic brain injury (TBI) is standard practice in children and adults. Prophylaxis in adults is used in other conditions such as subarachnoid hemorrhage; 13 some groups use it following supratentorial neurosurgery as well, 7 although practice guidelines for this use have not yet been established. 9 Previous studies have shown that posttraumatic seizures are more common in children than in adults. 2 Up to 25% of children with TBI have electrographic seizures on continuous electroencephalogram (EEG) monitoring following injury, 11 and posttraumatic seizures may occur most commonly in children younger than 2 years. 10 Seizures within the 1st week of brain injury may complicate the acute hospital course by contributing to Abbreviations used in this paper: AED = antiepileptic drug; EEG = electroencephalogram; FOS = (fos)phenytoin; ICH = intracranial hemorrhage; ICU = intensive care unit; LVT = levetiracetam; TBI = traumatic brain injury. neuronal injury, metabolic demand, elevations in intracranial pressure, and cardiopulmonary compromise, ultimately increasing the risk of secondary brain injury. In a study examining the efficacy of various therapies for head trauma in pediatric intensive care units (ICUs), only the use of antiepileptic drugs (AEDs) was associated with a statistically significant reduction in mortality risk. 16 In adults, use of anticonvulsants to prevent posttraumatic seizures within 7 days of injury is a Level II recommendation for severe TBI. 1 Traditionally, phenytoin has been used for seizure prophylaxis due to its availability in an intravenous formulation and nonsedating properties. In adults it is well established that, compared with placebo, phenytoin is more effective in preventing posttraumatic seizures in the 1st week after injury. 1,15 In children, however, studies are limited. In a retrospective review of children with head trauma, 53% who received no AED suffered posttraumatic seizures, compared with 15% of those who received phenytoin. 5 The only randomized, double-blind trial to compare an AED to placebo for seizure prophylaxis in children with TBI 209

2 S. Bansal et al. did not show a significant reduction in the incidence of early posttraumatic seizures; 17 however, this study had several limitations, including failure to use continuous EEG monitoring in children treated with paralytic agents and loss of 33% of patients within 48 hours of presentation. Nevertheless, prophylaxis of early seizures with phenytoin is a Level III recommendation according to the 2012 Guidelines for Management of Severe TBI in Infants, Children and Adolescents, 4 and seizure prophylaxis after various other types of acute brain injury, including ICH, has become common practice. The use of (fos)phenytoin is complicated by the potential for multiple adverse reactions, a narrow therapeutic window requiring close monitoring of drug levels, and numerous drug interactions via its induction of the CYP450 enzyme pathway. Newer AEDs are being increasingly used to prevent early posttraumatic seizures. Levetiracetam, in particular, is commonly used as an alternative due to similar availability (intravenous formulation) and favorable side-effect profile. In addition, its pharmacokinetics are more stable and it does not require drug-level monitoring, unlike phenytoin. Several studies in adults have demonstrated similar efficacy of levetiracetam compared with phenytoin for seizure prophylaxis. In 2010, Szaflarski et al. showed that patients with severe TBI or subarachnoid hemorrhage treated with levetiracetam for 1 week had the same incidence of early seizures but had better long-term outcomes at 6 months when compared with those treated with phenytoin. 13 Milligan et al. found similar efficacy between the 2 medications when administered for seizure prophylaxis after supratentorial neurosurgery, with significantly fewer adverse events in patients treated with levetiracetam. 7 The only comparable study in children was conducted in patients undergoing hematopoietic stem cell transplantation; the 2 drugs were similarly efficacious in preventing seizures during administration of busulfan (a myeloablative drug with known CNS toxicity). 12 To our knowledge there have been no studies to compare efficacy and/or tolerability of levetiracetam versus (fos)phenytoin for seizure prophylaxis after ICH in children. In this study, our primary aim was to determine whether there was a significant difference in the occurrence of early seizures after acute supratentorial ICH in children who received levetiracetam versus (fos)phenytoin for seizure prophylaxis. Secondary aims were to compare levetiracetam and (fos)phenytoin for tolerability and effect on development of late seizures at 6 months follow-up. Methods This study was approved by the institutional review board at Children s National Medical Center in Washington, DC. It was exempt from informed consent. Patients were identified from the Pediatric NeuroICU database. Patients with acute supratentorial ICH who received seizure prophylaxis beginning on their 1st hospital day with (fos)phenytoin (FOS), levetiracetam (LVT), or both (Both) were included. Choice of AED was based on physician preference. The Both group consisted of patients who were initially treated with FOS but then switched to LVT for ongoing maintenance for various reasons (see below). Cases in which patients were switched due to a breakthrough seizure were analyzed as treatment failures of the patients original AED cohort. Cases involving patients who were already on AEDs at the time of admission, who received only a loading dose of an AED, and/or who received prophylaxis with an AED other than FOS or LVT were excluded from analysis. Electronic medical records were reviewed to supplement data on demographic characteristics, etiology of ICH, location of ICH, seizures on presentation, concurrent administration of medications with antiepileptic properties (benzodiazepines or barbiturates), duration of seizure prophylaxis provided, and FOS levels when applicable. Etiologies of ICH were grouped into traumatic and nontraumatic brain injury. Location of ICH was identified as intraaxial (parenchymal), extraaxial (intraventricular, subdural, epidural, and subarachnoid), or both. Cases in which patients received boluses or continuous infusions with benzodiazepines or barbiturates in the ICU for indications other than seizures were noted; however, we did not include benzodiazepines administered in the field or in the emergency room. The primary aim of this study was to compare the efficacy of FOS and LVT for the treatment of early posthemorrhagic seizures. We defined early posthemorrhagic seizures as those that occurred within 7 days of ICH after beginning seizure prophylaxis. This is consistent with conventions in existing literature, which distinguish hyperacute seizures (seizures on presentation) from early and late seizures. Additional outcomes studied were occurrence of late posthemorrhagic seizures (defined as seizures after 7 days postinjury) and adverse reactions. We reviewed electronic medical records as well as EEG data to determine whether patients had early posttraumatic seizures. Per institution pathway, patients under sedation or those with encephalopathy were monitored with continuous EEG for at least 48 hours, while those who were awake were monitored for clinical events. Both clinical and electrographic seizures were included in our analysis. Maintenance doses of mg/kg/day were used for LVT following a 20 mg/kg bolus. For FOS, patients typically received a loading dose of 20 mg/kg followed by maintenance of 5 mg/kg/day, with additional boluses and adjustments of maintenance dosing provided to maintain target levels. Total phenytoin levels were defined as subtherapeutic if they were less than 15 ug/ml. If a patient s serum albumin was less than 3.5 gm/dl, serum phenytoin level was corrected using the Sheiner-Tozer equation: corrected Dilantin = measured level / [(0.2 albumin) + 0.1]. A series of chi-square tests were conducted to determine what variables differed significantly between treatment groups and to determine differences in the management of refractory seizures between treatment groups. To determine if the probability of early seizures differed significantly among treatment groups, a multivariate logistic regression was conducted, controlling for all other relevant variables. Nominal variables of age group and location of ICH were effects coded. The treatment group was dummy coded to allow for direct comparison between patients receiving FOS and the other 2 treatment groups. We controlled for the main effects of age, sex, diagnosis, 210

3 Seizure prophylaxis in ICH seizures on presentation, location of ICH, and use of benzodiazepines and/or barbiturates in the primary analysis. All covariates were effects coded; therefore the reference group represented the average of all groups. A 2-sided p value < 0.05 was considered to denote significance. Odds ratios were calculated with corresponding 95% confidence intervals. Because the chi-square tests for diagnosis, use of benzodiazepines and/or barbiturates, and seizures on presentation all had p values near significance, the interactions between these variables and the treatment groups were also controlled for in the multivariate analysis. The chisquare test between location of ICH and treatment group was also marginally significant; however, we were unable to include this interaction due to the overall low base rate of events in some treatment groups. A similar multivariate logistic regression was conducted to predict late seizures, controlling for relevant variables. Because of the low base rate of late seizures, we were not able to include any potential interactions between variables in the model. Results We identified 142 patients between 1 month and 17 years of age who were admitted to the pediatric ICU with acute supratentorial ICH between December 2007 and July 2012, all of whom received seizure prophylaxis beginning on the day of admission. Sixteen patients were excluded from further analysis due to the aforementioned criteria. Seizure prophylaxis was provided with LVT in 61 cases, FOS in 40 cases, and both medications in 25 cases. Seventeen patients treated with both medications received only a loading dose of FOS before being switched to LVT for maintenance therapy. The remaining 8 were switched later during hospitalization after varying durations of FOS maintenance. Patient Characteristics Treatment groups were not significantly different with respect to age, sex, or duration of therapy. Diagnosis (etiology of ICH), location of ICH, and use of benzodiazepines and/or barbiturates for sedation differed only marginally. The treatment groups differed significantly, however, with respect to the presence of seizures on presentation (Table 1). All patients received a minimum of 48 hours of seizure prophylaxis, and the majority received prophylactic treatment for at least 7 days (88.9%). Sixty-seven patients underwent EEG monitoring within 7 days of injury; duration of monitoring varied based on clinical condition. The etiology of ICH varied, but TBI was most common in all groups. Nontraumatic etiologies included coagulopathies (for example, iatrogenic and pathologic), vasculopathies (for example, arteriovenous malformation), idiopathic/cryptogenic, postsurgical, neoplasms, and miscellaneous (for example, intraparenchymal hemorrhage after thermal injury to the brain). Of note, only 3 patients in this study had ICH following elective supratentorial neurosurgery; 1 patients in the LVT group and 1 in the Both group underwent surgery for removal of tumors and had ICH following the procedure. The most common location for ICH was extraaxial in the FOS and LVT groups, while the majority of patients in the Both group had both intraaxial and extraaxial hemorrhages. The percentage of patients treated with benzodiazepines and/or barbiturates for sedation and/or management of intracranial pressure was higher in the Both group (64%) than in the FOS group (45%) or LVT group (37.7%). As mentioned above, these differences were not statistically significant. There was a significant between-groups difference between the number of patients presenting with seizure, with proportionally more seizures on presentation in the FOS group (42.5%) and Both group (60%) than the LVT group (18%). Early Posthemorrhagic Seizures The overall incidence of early posthemorrhagic seizures, excluding seizures on presentation, was 18.3%. Seventy percent of all patients with early seizures had seizures on presentation. The location of ICH was extraaxial in 73.9% of all patients, while the etiology of ICH was TBI in 82.6%. Coagulopathy and vasculopathy were the only other ICH etiologies associated with the presence of early posthemorrhagic seizures. On average, the seizures began on the 2nd day of hospitalization. Children younger than 2 years of age had a higher probability of early seizures (OR 9.09, p = 0.003, 95% CI ) when compared with the entire cohort. Patients who received FOS had a significantly higher probability of early seizures than those who received LVT (Table 2). This was true when controlling for the effectscoded covariates of age, sex, diagnosis, location of ICH, seizures on presentation, and use of benzodiazepines and/ or barbiturates for indications other than seizure, as well as the relevant interaction terms (OR 24.6, p = 0.002, 95% CI ). Overall, the probability of early seizures was in the FOS group and in the LVT group. Patients in the Both group did not have a significantly different probability of early seizures than patients in the FOS group (p = 0.992), although the low rate of events led to an uninterpretable odds ratio. Thirty-four percent of all patients had experienced a seizure on presentation, either in the field or in the emergency room, prior to starting prophylaxis. This cohort of patients was analyzed separately. Fifty-three percent of patients treated with FOS continued to have seizures despite prophylaxis, compared with 27% in the LVT group and 27% in the Both group (Fig. 1). This difference was not statistically significant (p = 0.187). Lastly, we analyzed the subgroups of children younger than 2 years and those with TBI separately, controlling for the same variables as above. Considering only children younger than 2 years, 61.1% in the FOS group had early seizures versus 18.5% in the LVT group; although this difference was not statistically significant, there was a strong trend toward significance (OR 5.89, p = 0.057, 95% CI ). Among the children with TBI, those treated with FOS had a significantly higher probability of early seizures (34.3% vs 6.3% for those treated with LVT, OR 12.5, p = 0.006, 95% CI ). Refractory Seizures Treatment for early posthemorrhagic seizures varied. Some patients remained on monotherapy but received ad- 211

4 S. Bansal et al. TABLE 1: Demographics and clinical data stratified by cohort* Variable FOS (n = 40) LVT (n = 61) Both (n = 25) χ 2 p Value age mos 18 (45) 27 (44.3) 5 (20) 2 9 yrs 9 (22.5) 10 (32.8) 10 (40) yrs 3 (32.5) 14 (23) 10 (40) sex male 25 (62.5) 46 (75.4) 18 (72) female 15 (37.5) 15 (24.6) 7 (28) diagnosis TBI 35 (87.5) 48 (78.7) 16 (64) non-tbi 5 (12.5) 13 (21.3) 9 (36) location of ICH extraaxial 19 (47.5) 35 (57.4) 6 (24) intraaxial 4 (10) 8 (13.1) 6 (24) both 17 (42.5) 18 (29.5) 13 (52) duration of AED use <7 days 7 (17.5) 5 (8.2) 2 (8) >7 days 33 (82.5) 56 (91.8) 23 (92) Sz on presentation 17 (42.5) 11 (18) 15 (60) <0.001 use of benzodiazepines &/or barbiturates 18 (45) 23 (37.7) 16 (64) * Values represent numbers of patients (%), except where otherwise indicated. Patients in the Both cohort were initially treated with FOS and were then switched to LVT. FOS = (fos)phenytoin; LVT = levetiracetam; Sz = seizure(s). Includes boluses and/or infusions of benzodiazepines and/or barbiturates provided for sedation and/or management of intracranial pressure within the first 7 days after injury. TABLE 2: Multivariate regression analysis of predictors of early seizures Variable b (SE) p Value OR main effects FOS (intercept) LVT (1.039) Both (1856) age <2 yrs (0.751) age 2 9 yrs (0.710) sex (0.833) diagnosis (1.775) ICH, extraaxial (0.920) ICH, extra- & intraaxial (0.895) Sz on presentation (1.093) use of benzodiazepines &/or barbiturates (1.042) interactions (LVT vs FOS) vs diagnosis (2.097) (Both vs FOS) vs diagnosis (2601)* e10 (LVT vs FOS) vs Sz on presentation (1.607) (Both vs FOS) vs Sz on presentation (2648)* e6 (LVT vs FOS) vs use of benzodiazepines &/or barbiturates (1.887) (Both vs FOS) vs use of benzodiazepines &/or barbiturates (2.441) * Dummy-coded multiple logistic regression, including interaction terms, led to a highly skewed standard error for the Both group, making it difficult to interpret the parameter estimates of the FOS versus Both comparison with any confidence. 212

5 Seizure prophylaxis in ICH TABLE 3: Percentage of patients with refractory seizures Variable FOS (%) LVT (%) Both (%) p Value failure of monotherapy failure of 2 AEDs requirement for continuous infusion Fig. 1. Percentage of patients in each treatment group with early posthemorrhagic seizures in cohort of patients with seizures on presentation. ditional boluses of the same AED, while others were treated with additional medications. Patients in whom the first AED was considered to have failed were treated with a second drug. After failure of 2 AEDs, some patients were treated with a continuous infusion (either midazolam or pentobarbital) while others received a third AED. This decision was based on physician preference. Monotherapy failed in 30% of patients treated with FOS, compared with 6.6% of those treated with LVT and 12.5% of those treated with both (p = 0.005) (Table 3). Failure of 2 AEDs was seen in a higher percentage of patients on FOS (17.5%) than in the LVT group (3.3%) or the Both group (8%) (p = 0.046). Similarly, more patients on FOS required a continuous infusion (15% vs 3.3% in the LVT Group and 4% in the Both group; p = 0.065). All patients who required a continuous infusion for seizure control were younger than 2 years of age, and all but 1 had TBI. Subtherapeutic (fos)phenytoin Levels Of all 40 patients treated with FOS, 18 had subtherapeutic phenytoin levels at some point during treatment. Controlling for relevant variables, patients with subtherapeutic levels had a higher probability of early posthemorrhagic seizures (OR 6.64, p = 0.032, 95% CI ). Fourteen patients had early seizures while being treated with FOS monotherapy; 7 had therapeutic levels and 7 had subtherapeutic levels on the day that early seizures occurred. The frequency of monitoring of drug levels was similar among these patients. Patients in the group with therapeutic levels underwent monitoring on an average of 3.35 hospital days within the 1st week of stay, while those in the group with subtherapeutic levels underwent monitoring on an average of 3.28 hospital days. Late Posthemorrhagic Seizures Seventy-three patients were followed for at least 6 months after ICH, and in 9 of these patients, seizures were observed at least 7 days after ICH. In the FOS group, 15% of the patients had late seizures versus 10.3% in the LVT group and 14.3% in the Both group (no significant difference, p = 0.265). All of these patients had been treated for a minimum of 7 days with seizure prophylaxis. Only 2 of the patients who experienced late seizures had also experienced early seizures, and both were in the FOS group. Both of these patients required continuous infusions to control their early seizures. Both had been discharged home on 2 AEDs and began to have seizures after being weaned from 1 medication. Five patients were not receiving an AED at the time of late seizures. In 2 of these 5 cases, the seizures occurred with provocation: 1 patient had an active Varicella zoster virus infection, and the other had failure of his ventriculoperitoneal shunt. Lastly, in 2 additional cases that were analyzed in this group, the patients had late seizures provoked by recurrent hemorrhages within 3 weeks of initial ICH; both were receiving prophylactic medication at the time of late seizures. Adverse Events Adverse drug reactions resulting in a change or discontinuation of treatment occurred in 4 patients. One patient developed a rash immediately after receiving a loading dose of FOS; her medication was changed to LVT, and the case was analyzed in the Both group. Three other patients developed an adverse reaction outside the initial 7 days for treatment. One patient developed a rash while being treated with oral phenytoin after discharge and selfdiscontinued the drug; similarly, a patient discharged on oral LVT self-discontinued the drug after discharge because of rash. Both had received at least 7 days of therapy. The fourth patient tolerated intravenous FOS; however, on hospital Day 11 he was transitioned to oral phenytoin and developed a rash. Discussion In this retrospective study, the overall incidence of early posthemorrhagic seizures after supratentorial ICH in pediatric patients treated with seizure prophylaxis was 18%. The risk of early seizures in patients with ICH due to TBI was 19.2%. This is consistent with previous estimates of rates of early seizures after TBI, which vary from 10% to 35% when trauma is severe. 2,10 Fifteen percent of patients with diagnoses other than TBI had early seizures. This suggests that the risk of early seizures in multiple types of acute supratentorial ICH is similar. Seizures were most likely to occur approximately 48 hours after ICH, coinciding with the period of maximum edema and risk for secondary injury. Furthermore, we found that children younger than 2 years of age had a much higher probability (OR 9.09) of early seizures than the entire patient cohort, regardless of confounding factors. This confirms that this is a high-risk subgroup. This was in contrast to earlier studies that did not show any difference. 2 Patients treated with levetiracetam had a lower probability of early seizures than those treated with (fos)phe- 213

6 S. Bansal et al. nytoin. This remained true within the subgroups of patients with TBI and children less than 2 years of age; the difference reached significance in the former and nearsignificance in the latter. A major potential confounder is seizure on presentation, which may suggest an increased susceptibility to early seizures, possibly due to more severe underlying pathology. There were significantly more patients with seizure on presentation in the FOS and Both groups. To exclude this as a confounder, we corrected for it in our multivariate analysis; thus the apparent advantage of levetiracetam is not explained by the potentially increased susceptibility to seizures in the other groups. Furthermore, we analyzed the cohort of patients with seizures on presentation separately. Presumably, these patients are similarly predisposed to early posthemorrhagic seizures. Patients in the FOS and the Both groups received an initial loading dose of (fos)phenytoin in the emergency room due to seizure on presentation; however, the Both group represents patients who were then switched to levetiracetam following admission to the ICU for ongoing maintenance. One patient had a vein of Galen malformation and was switched from (fos)phenytoin because of concern for its potential to induce arrhythmia. A second patient was transitioned to levetiracetam because of an adverse event (cutaneous hypersensitivity). The remaining patients were switched because of physician preference; as more studies established the safety and efficacy of levetiracetam, its use has increased over time in our ICUs. Half of the patients in the FOS group had early seizures, while only 27% of patients in the LVT and Both groups did. These results were not statistically significant, suggesting that these regimens are similar in efficacy. One could argue that the increased probability of early seizures in patients treated with FOS may be due to subtherapeutic drug levels; however, 50% of patients on FOS with early seizures had therapeutic levels. Furthermore, we found that regardless of whether a therapeutic level is attained, the frequency of monitoring of drug levels was similar, indicating that some children are simply unable to maintain therapeutic levels. Thus, diligent monitoring of levels alone is not sufficient to prevent early seizures in patients treated with FOS. This difficulty in maintaining levels makes FOS a more problematic agent to use in children, underscoring this advantage of LVT. Early seizures are often refractory to treatment, and children may require multiple AEDs to achieve seizure control. Compared with those treated with levetiracetam, patients treated with (fos)phenytoin were significantly more likely to experience failure of monotherapy. Furthermore, more patients treated with (fos)phenytoin subsequently required continuous infusions, which can lead to additional complications including respiratory depression, hypotension, and longer ICU course. Although it did not reach statistical significance, there was a trend toward an increased need for continuous infusions in the FOS group. A larger sample size may reveal significant differences. Early posttraumatic seizures in adults are associated with an increased risk of subsequently developing epilepsy, 3,6 but to date no study has shown that the treatment of early seizures reduces the risk of developing epilepsy later. Approximately 12% of our patients with supratentorial ICH who received seizure prophylaxis experienced late posthemorrhagic seizures by 6 months post-ich, which is consistent with previous studies (all had received at least 7 days of therapy). Neither the choice of prophylaxis nor the presence of early posthemorrhagic seizures impacted the risk for late seizures, although our analysis may have been limited by the low base rate of events. Overall we found a very low rate of adverse events. Cutaneous hypersensitivity was the only adverse reaction observed, and it occurred with oral levetiracetam, intravenous (fos)phenytoin, and oral phenytoin. One patient with congenital heart disease was switched to levetiracetam from (fos)phenytoin due to its potential to induce arrhythmia; while this was not included as an adverse event in our study, it is worth noting as a potentially significant side effect. Limitations of this study include those inherent to the retrospective design. We recognize that the upper limits of many of the calculated confidence intervals were very high; however, these limits should be interpreted with great caution. In conducting our multivariate regressions, extremely conservative models were used, which controlled for various factors, including both those that differed significantly and those that showed nonsignificant differences between treatment groups, as well as their potential interactions. Narrower confidence intervals may have been achieved with more liberal analysis; however for the purposes of a retrospective study we felt that it was best to err on the side of caution. Similarly, the large standard errors that were observed were likely a result of the paucity of residual variance remaining when all other variables were accounted for in the model. Although we are confident these treatment groups led to different probabilities of early seizures, further studies may find different parameter estimates of the overall probabilities of early seizures. Only pediatric patients with supratentorial hemorrhage were included in this study. These patients may have a different underlying tendency to seizures compared with adults with similar pathology; therefore interpretation of our results should be limited to this population. Furthermore, the incidence of acute symptomatic seizures in children is not well established. Although we have compared different therapies, we cannot be certain that this reduces the overall incidence of early seizures. Patients who did not receive prophylaxis were excluded from this study, but future studies should be undertaken to compare outcomes for patients who do and do not receive prophylaxis. As mentioned, only patients who have acute encephalopathy and/or are being treated with pharmacological sedation receive EEG monitoring per institution protocol. Not all patients with ICH met these criteria. Although there is a possibility that subclinical seizures were missed in some cases, it is neither practical nor a standard of care to monitor all patients with continuous electroencephalography for 7 days. Differences in general ICU care are unlikely to account for these results. Although use of (fos)phenytoin has recently declined in favor of levetiracetam as the initial drug choice for seizure prophylaxis, no major changes in standards of care for common ICU pathologies has occurred. For example, at our institution, the last major modification of the severe TBI protocol was 2007, prior to admission of the first patient included in the current study. 214

7 Seizure prophylaxis in ICH Cognitive outcomes were not assessed but may play a substantial role in the preferred choice of seizure prophylaxis. Phenytoin has been linked to worse cognitive outcomes when administered to adults with subarachnoid hemorrhage, 8 and Taylor et al. demonstrated that levetiracetam is associated with greater retention of cognitive function while providing better seizure prophylaxis when compared with phenytoin in adults with acute ICH. 14 Future studies are needed to assess long-term cognitive outcomes after seizure prophylaxis with various AEDs in children. There is considerable variation in practice based on provider preference, with different thresholds of when to treat seizures and when to escalate AED treatment. Because we could not reliably quantify doses of benzodiazepines given prior to arrival in our institution, these data were not accounted for. In the case of longer-acting medications, such as lorazepam, this may have led to bias within the first 12 hours; however, as previously mentioned, we found that early posthemorrhagic seizures typically occurred on hospital Day 2, by which time medications given on presentation would have already been metabolized. Thus it is unlikely this influenced our outcome. For children who experience early seizures, there is great variation in the duration of treatment, and this may have had an impact on long-term outcomes. Additionally, only 6-month follow-up data were available in this study. Longer follow-up and further clinical assessments are needed to evaluate the effect of treatment of early posthemorrhagic seizures on the development of epilepsy years later. Finally, optimal dosing for levetiracetam has not been established, so it is possible that our patients received suboptimal doses of levetiracetam. Conclusions To date, there have been no studies comparing (fos) phenytoin and levetiracetam for seizure prophylaxis in children with ICH. Although this study is limited by its retrospective design, our results suggest that levetiracetam is a reasonable alternative to (fos)phenytoin for prevention of early posthemorrhagic seizures in children. Adverse events are rare in both groups, and treatment choice does not impact the occurrence of late seizures. Disclosure The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. Author contributions to the study and manuscript preparation include the following. Conception and design: Carpenter. Acquisition of data: Bansal, Kebede, Dean, Carpenter. Analysis and interpretation of data: all authors. Drafting the article: Bansal, Blalock. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Bansal. Statistical analysis: Blalock. Administrative/technical/material support: Kebede. References 1. Bratton SL, Chestnut RM, Ghajar J, McConnell Hammond FF, Harris OA, Hartl R, et al: Guidelines for the management of severe traumatic brain injury. XIII. Antiseizure prophylaxis. J Neurotrauma 24 (Suppl 1):S83 S86, 2007 (Erratum in J Neu rotrauma 25: , 2008) 2. Hahn YS, Fuchs S, Flannery AM, Barthel MJ, McLone DG: Factors influencing posttraumatic seizures in children. Neurosurgery 22: , Herman ST: Epilepsy after brain insult: targeting epileptogenesis. Neurology 59 (9 Suppl 5):S21 S26, Kochanek PM, Carney N, Adelson PD: Guidelines for the early medical management of severe traumatic brain injury in infants, children, and adolescents second edition. Chapter 17. Antiseizure prophylaxis. Pediatr Crit Care Med 13 (1 Suppl): S72 S82, Lewis RJ, Yee L, Inkelis SH, Gilmore D: Clinical predictors of post-traumatic seizures in children with head trauma. Ann Emerg Med 22: , Lowenstein DH: Epilepsy after head injury: an overview. Epilepsia 50 (Suppl 2):4 9, Milligan TA, Hurwitz S, Bromfield EB: Efficacy and tolerability of levetiracetam versus phenytoin after supratentorial neurosurgery. Neurology 71: , Naidech AM, Kreiter KT, Janjua N, Ostapkovich N, Parra A, Commichau C, et al: Phenytoin exposure is associated with functional and cognitive disability after subarachnoid hemorrhage. Stroke 36: , Pulman J, Greenhalgh J, Marson AG: Antiepileptic drugs as pro phylaxis for post-craniotomy seizures. Cochrane Database Syst Rev 2:CD007286, Ratan SK, Kulshreshtha R, Pandey RM: Predictors of posttraumatic convulsions in head-injured children. Pediatr Neurosurg 30: , Schreiber JM, Zelleke T, Gaillard WD, Kaulas H, Dean NP, Carpenter JL: Continuous video EEG for patients with acute en cephalopathy in a pediatric intensive care unit. Neurocrit Care 17:31 38, Soni S, Skeens M, Termuhlen AM, Bajwa RP, Gross TG, Pai V: Levetiracetam for busulfan-induced seizure prophylaxis in children undergoing hematopoietic stem cell transplantation. Pediatr Blood Cancer 59: , Szaflarski JP, Sangha KS, Lindsell CJ, Shutter LA: Prospective, randomized, single-blinded comparative trial of intravenous levetiracetam versus phenytoin for seizure prophylaxis. Neurocrit Care 12: , Taylor S, Heinrichs RJ, Janzen JM, Ehtisham A: Levetiracetam is associated with improved cognitive outcome for patients with intracranial hemorrhage. Neurocrit Care 15:80 84, Temkin NR, Dikmen SS, Wilensky AJ, Keihm J, Chabal S, Winn HR: A randomized, double-blind study of phenytoin for the prevention of post-traumatic seizures. N Engl J Med 323: , Tilford JM, Simpson PM, Yeh TS, Lensing S, Aitken ME, Green JW, et al: Variation in therapy and outcome for pediatric head trauma patients. Crit Care Med 29: , Young KD, Okada PJ, Sokolove PE, Palchak MJ, Panacek EA, Baren JM, et al: A randomized, double-blinded, placebo-controlled trial of phenytoin for the prevention of early posttraumatic seizures in children with moderate to severe blunt head injury. Ann Emerg Med 43: , 2004 Manuscript May 22, Submitted October 15, Portions of this work were presented in poster form at the Child Neurology Society Annual Meeting in Huntington Beach, California, in October Please include this information when citing this paper: published online November 29, 2013; DOI: / PEDS Address correspondence to: Seema Bansal, M.D., Department of Neurology, Ste. W4-800, 111 Michigan Ave. NW, Washington, DC sbansal@luriechildrens.org. 215