TH may be applied with external devices and/or core cooling methods [infusion of cold fluids through peripheral or central approaches (Holzer, 2010)]. Deep hypothermia (15–22°C) leads to marked greater complications as compared to mild TH (30–35°C; Harris et al., 2002). Adverse effects may be related to the cooling devices or the low temperature itself. In 41 clinical trials using mild TH on postanoxic encephalopathy, complications related to the cooling device involved only 29 (1%) of 3,133 patients, with bleeding, infection, vein thrombosis, or pulmonary edema (Holzer, 2010). Conversely, in the few major comparative clinical trials in the same clinical setting, side effects were recorded in 223 of 300 patients treated with TH (74%) versus 201 of 285 controls (71%; p = 0.31; Holzer, 2010). These included pneumonia (probably related to mechanical ventilation), hyperglycemia, cardiac arrhythmias, and electrolyte disturbances.
A pioneer description published >25 years ago on three children with generalized SE successfully treated with TH (30–31°C) and thiopental (Orlowski et al., 1984) was followed by only a handful reported cases. Four adult patients with SE due to limbic encephalitis (two patients), hepatic encephalopathy (one patient), and of unknown origin (one patient) were treated with TH (31–35°C) together with midazolam; SE was controlled in all, but two subjects died subsequently (Corry et al., 2008). Shivering had to be controlled with neuromuscular blockade; other side effects included venous thrombosis and pulmonary embolism. One additional adult patient with cryptogenic SE was treated with TH (34°C) and thiopental, developing paralytic ileus that required emergency intestinal resection; she survived (Cereda et al., 2009). Finally, an infant with SE due to hemimegalencephaly had his ongoing seizures controlled by hypothermia at 35–36°C, together with ketamine; later, hemispherectomy was performed (Elting et al., 2010). Reported TH durations are markedly variable, ranging between 20 h and several days; clinical data are summarized in Table 1. Taking these reports into account, it seems that TH may help controlling ongoing seizures, but that its efficacy seems mostly to be only transient, that is, seizures tend to recur as soon as TH is reversed. TH represents thus a therapeutic option in severely refractory SE that may mostly help to bridge the time awaiting for the underlying etiology to be controlled, in line with several other anecdotic alternatives, such as volatile anesthetics, ketamine, lidocaine, verapamil, magnesium, paraldehyde, etomidate, ketogenic diet, immunomodulatory therapies, electroconvulsive treatment, vagus nerve stimulation, repetitive transcranial magnetic stimulation, or cerebrospinal fluid–air exchange (reviewed in Rossetti, 2009a). To achieve this goal, these options may be used in succession, alternating with common anesthetics (midazolam, propofol, barbiturates).
It has been increasingly recognized that postanoxic SE does not necessarily imply a poor outcome. In this setting, recent observations suggest that SE occurring during TH, mostly as a “seizure suppression” EEG pattern, represents a situation of extreme brain damage, and patients are extremely unlikely to survive (Rossetti et al., 2010b; Thenayan et al., 2010); conversely, SE arising after rewarming may be treated with antiepileptic drugs (AEDs) if the electroencephalography (EEG) background is reactive, and somatosensory evoked potentials and brainstem reflexes are preserved (Rossetti et al., 2009b). In those cases (representing about 10% of patients with postanoxic SE) survival with reasonable functional outcome may be obtained (Rossetti et al., 2010a). This suggests that TH (together with moderate midazolam or propofol doses) is sufficient to transitorily control “benign” postanoxic SE forms (illustrating its antiepileptic properties), whereas it does not prevent a dismal outcome in more severe forms.