Status epilepticus (SE) is a life-threatening condition defined as seizure that persists longer than 5 min (Brophy et al., 2012). Annual incidence of SE is 10–41/100,000, and SE affects 120,000–200,000 people annually in the United States. SE carries a 7–39% mortality rate (DeLorenzo et al., 1996; Coeytaux et al., 2000; Knake et al., 2001; Novy et al., 2010). Refractory SE (RSE) occurs when patients fail first- and second-line anticonvulsant therapy (Novy et al., 2010; Brophy et al., 2012). RSE occurs in approximately one third of SE patients (Mayer et al., 2002; Lambrechtsen & Buchhalter, 2008). In children, RSE is associated with development of static encephalopathy and death (Barberio et al., 2012). RSE treatment typically requires anesthetic anticonvulsant levels to achieve burst suppression on electroencephalography (EEG). Hypothermia has been used occasionally as an adjunct to anticonvulsants in RSE treatment. Existing literature describes experience with mild hypothermia in adult RSE; the largest case series comprises four patients (Corry et al., 2008). Herein, we describe five children treated with mild hypothermia (32–35°C) for RSE. This is the largest case series to date describing treatment of pediatric RSE with mild hypothermia.
- Top of page
This is the largest pediatric case series reporting use of mild hypothermia (32–35°C) to treat RSE. Hypothermia, ranging from 20 to 35°C, has been occasionally used to treat RSE in adults (Sourek & Travnicek, 1970; Karkar et al., 2002; Corry et al., 2008). Several case reports describe its use to treat RSE in children (Orlowski et al., 1984; Elting et al., 2010; Lin et al., 2012; Shein, 2012). Table 1 summarizes clinical data available for published pediatric cases. SE and RSE are neurologic emergencies, requiring aggressive treatment. SE may proceed for hours if left untreated, resulting in neuronal injury and cell death (Fountain & Lothman, 1995; Fujikawa et al., 2000; Chen et al., 2007). Animal models indicate that efficacy of anticonvulsants in terminating SE decreases as SE duration increases (Mazarati et al., 1998; Rajasekaran et al., 2010). Simply put, the longer the seizures last, the harder they are to stop and the more damage they inflict. Hence, current guidelines recommend rapid escalation of treatment using different drug classes to terminate seizures (Shorvon et al., 2008; Shearer & Riviello, 2011; Brophy et al., 2012). Notably, hypothermia in our patients terminated seizures even after hours to weeks of RSE.
Table 1. Prior reports of hypothermia in pediatric RSE
|References||No. of children||Age||Indication for cooling||Seizure duration||Other therapies attempted||Continuous infusions failed||Goal temp (°C)||Duration of HT||Outcome|
|Vastola et al. (1969)||1||16 years||RSE due to meningoencephalitis||Not reported|| ||None attempted||37||24 h||No recurrent seizures|
|Orlowski et al. (1984)||3||6 years||RSE due to Reye's syndrome||Not reported|| ||None attempted||30–31||48 h||Complete recovery, no recurrent seizures|
|11 years||RSE due to encephalitis||Not reported|| ||Paraldehyde||30–31||72 h||Good recovery with residual memory deficits|
|18 years||RSE due to unknown neurodegenerative disease||3 days|| ||Paraldehyde||30–31||5 days||Recurrent focal seizures|
|Elting et al. (2010)||1||5 months||Focal RSE due to hemimegalencephaly||48 h|| || ||35.3a||3 h||Discharged from ICU without RSE; hemispherectomy|
|Shein et al. (2012)||1||4 months||RSE due to SCN1A mutation, Dravet syndrome||Not reported|| ||Midazolam||33–34||43 h, then again for 24 h||Seizures requiring ketogenic diet; developmental delay|
|Lin et al. (2012)||2||10 years||RSE secondary to febrile-infection–related epilepsy syndrome||1 day|| ||Midazolam||33||5 days||Discharged home; mild cognitive and motor impairments|
|4 years||RSE secondary to febrile-infection–related epilepsy syndrome||10–12 h|| ||Midazolam||33||3 days||Discharged home; mild cognitive impairments and epilepsy|
Multiple animal studies indicate that hypothermia is an effective primary and/or adjunct treatment for SE. In (Vastola et al. 1969). cooled cats with experimentally induced SE below 32°C, which terminated seizures in 85% of animals. In rats with SE evoked by perforant pathway stimulation, moderate hypothermia (29–33°C) reduced severity and frequency of motor seizures, albeit it was more effective when combined with diazepam (Schmitt et al., 2006). In a similar model, deep hypothermia (20°C) suppressed SE in 40% of rats (Kowski et al., 2012). Of interest, 50% of successfully treated rats were seizure free during and after rewarming (Kowski et al., 2012). In rats with seizures induced by potassium channel antagonists (Yang & Rothman, 2001) or by fluid percussive injury (D'Ambrosio et al., 2013), focal cortical cooling significantly reduced seizure frequency. In rats injected with pilocarpine, hypothermia pretreatment reduced neuronal apoptosis and increased latency to SE onset (Maeda et al., 1999; Yu et al., 2011). Hypothermia also decreased brain edema and seizure frequency in rats with kainic acid–induced SE (Wang et al., 2011). In addition, animal studies suggest that hypothermia provides neuroprotection in SE (Wang et al., 2011; Yu et al., 2011, 2012; Zhou et al., 2012). Therefore, studies across multiple mammalian epilepsy models suggest that hypothermia may effectively abrogate SE and may provide neuroprotection during seizures.
In addition to noting that hypothermia effectively abrogated RSE in our patients, we can make two additional clinical observations. First, none of our RSE patients treated with hypothermia relapsed into SE after rewarming. In contrast, when treated with pentobarbital at normothermia, 80% of our patients (4/5) experienced SE relapse on pentobarbital discontinuation (Table 2). The one patient that did not relapse (EZ) was treated with pentobarbital and hypothermia simultaneously. SE relapse rate after pentobarbital coma in our series is consistent with that reported previously (Barberio et al., 2012). Second, degree of hemodynamic instability and, consequently, use of inotropic agents appeared to decrease during hypothermia in our series. In three of five patients, inotrope requirements decreased during hypothermia. Hemodynamic improvement likely reflects decreased pentobarbital dose required to achieve burst suppression during hypothermia versus normothermia. In a meta-analysis, 77% of RSE patients treated with pentobarbital developed hypotension and required inotropic support (Claassen et al., 2002). Pentobarbital depresses cardiac function (Jiang et al., 2011). Hence, it is not surprising that inotrope requirements decreased as pentobarbital dose decreased.
Table 2. Patient and clinical course characteristics
|Age at onset||10 years old||5 months old||11 months old||10 years old||15 years old|
|SE etiology||Unknown||Hydrocephalus, perinatal injury||POLG-1 mutation||Epilepsy||Anti-NMDAR encephalitis|
|Preexisting comorbidities||None||Prematurity||Chiari type I RAD||Epilepsy DD, ADHD||None|
|Medications attempted and failed (alphabetical order)|| || || || || |
|Initial burst suppression achieved with||Pentobarbital 3 mg/kg/h||Pentobarbital 10 mg/kg/h||Pentobarbital 3.5 mg/kg/h + hypothermia||Pentobarbital 5 mg/kg/h||Pentobarbital 2.5 mg/kg/h|
|Did SE recur after initial burst suppression?||Yes||Yes||No||Yes||N/A|
|Target temperature (°C)||34||32–34||34||33||33–35|
|Time to target temp (h)||2||2||3||3||13|
|Time at target temp (days)||1||3||2||5||5|
|Rewarming rate||0.5°C/24 h||1°C/6 h||0.5°C/12 h||0.5°C/12 h||2°C/8 h|
|Acid-base abnormalities||None||Lactic acidosis with NEC and septic shock||None||Lactic acidosis with pentobarbital||None|
|Electrolyte abnormalities||None|| ||Hypokalemia|| ||None|
|BP changes during cooling or rewarming||None||None||None||Hypertension||None|
|Infection during TH||None||Clinical sepsis||None||None||UTI|
|Coagulopathy (INR > 1.5)||None||Yes||None||None||None|
|Did SE recur after rewarming?||No||No||No||No||No|
|Seizures postrewarming||None||Subclinical||Two brief clinical seizures||None||None|
|Discharge medications(alphabetical order)||Levetiracetam||N/A|| || || |
|Discharge disposition||Home||Death||Home (after first admission)||Home||Long-term care facility|
In two patients (EZ and DB), however, inotrope requirement escalated with initiation of hypothermia. In EZ, hypothermia and pentobarbital were started simultaneously. Therefore, it is impossible to distinguish which of these treatments, hypothermia and/or pentobarbital, contributed most to development of hemodynamic instability. In DB, escalation of inotropes coincided with development of necrotizing enterocolitis (NEC) and septic shock 24 h after initiation of hypothermia. It is unclear whether hypothermia exacerbated these complications.
In neonates treated with hypothermia for hypoxic–ischemic encephalopathy (HIE), risk of NEC appears unaffected. Multiple clinical trials of therapeutic hypothermia in neonatal HIE demonstrated similar rates of NEC in infants treated with hypothermia compared to infants treated with usual care (normothermia) (Gluckman et al., 2005; Shankaran et al., 2005; Compagnoni et al., 2008; Shankaran et al., 2008; Azzopardi et al., 2009; Jacobs et al., 2011). None of these trials, however, was powered to detect a difference in NEC incidence between hypothermic and normothermic groups. More recently, a prospective nonrandomized trial evaluated safety of mild hypothermia in infants with advanced NEC and multiorgan dysfunction (Hall et al., 2010). In this trial, rates of death, laparotomy, and intestinal perforation were similar between 15 cooled and 10 control infants. Therefore, existing data suggest that mild hypothermia does not increase NEC risk in critically ill infants.
Body cooling poses several well-documented risks, although few are clinically significant with mild hypothermia (32–35°C). Cardiac arrhythmias are unlikely at core temperature >30°C (Welton et al., 1978; Piktel et al., 2011). Coagulopathy appears mild above 33°C (Rohrer & Natale, 1992; Eicher et al., 2005; Hall et al., 2010). We did not notice increased bleeding in our patients, including in DB who underwent laparotomy and bowel resection. Hypothermia increases infection risk (Compagnoni et al., 2008; Laupland et al., 2012; Seamon et al., 2012). One catheter-related urinary tract infection occurred in our cohort. Mild hypothermia commonly causes shivering, which necessitates neuromuscular blockade. It may also cause hypotension mainly during cooling or rewarming (Sessler, 2009). Indeed, a therapeutic hypothermia trial in children with traumatic brain injury associated hypothermia with increased incidence of hypotension during rewarming (Hutchison et al., 2008). Hypotension during rewarming (and cooling) is thought to result from changes in systemic vascular resistance and from uncontrolled diuresis (Polderman, 2009; Sessler, 2009). In our case series, we anticipated hypotension by monitoring arterial blood pressure and urine output during cooling/rewarming. We also minimized hypotension likelihood by changing temperature slowly, in some cases following a protocol established for an unrelated clinical trial of therapeutic hypothermia in pediatric cardiac arrest (Moler, 2009). We did not appreciate clinically significant hypotension during cooling or rewarming in our patients.
Generalization of our observations is limited by several factors. First, this is a retrospective case series. Second, hypothermia use and initiation timing varied significantly due to current lack of standardized protocols. Variability also existed in target temperature and in hypothermia duration. Absence of standardized hypothermia protocols likely explains implementation variability (Fink et al., 2010). Despite variability, hypothermia improved seizure burden in all five patients. Nevertheless, three patients had poor outcomes. RSE carries substantial risk of morbidity and mortality in and of itself (Barberio et al., 2012), and in all five cases, hypothermia was used as a rescue therapy after multiple agents failed. Whether earlier implementation of hypothermia would improve neurologic outcomes in RSE remains unknown. Future studies of hypothermia in SE/RSE will require development of standardized protocols for implementation, patient selection, target temperature, and therapy duration. Although this case series demonstrates potential efficacy of hypothermia in treating pediatric RSE, prospective multicenter studies are required.