Subjects and methods

Twenty children suffering from status epilepticus admitted to the paediatric ward from November 1993 to November 1995 were included in this study. Eleven were already taking various antiepileptic drugs.

Status epilepticus was diagnosed according to the criteria of Engel, namely continuous seizures for 30 minutes or longer or several seizures occurring with impairment of consciousness between seizure activity.2 Refractory status epilepticus was the diagnosis if the seizures continued, despite at least two doses of diazepam intravenously or rectally in succession followed by phenytoin sodium/phenobarbitone or both 20 mg/kg given over 30 minutes as an infusion, or failure to respond to the latter alone or in combination. The duration of status before midazolam therapy was approximate, based on the history obtained from the patient’s attendants and the referring physician’s notes. Electroencephalography (EEG) was not used for the diagnosis. It was, however, performed after the seizures had been controlled to monitor the electrical suppression of seizure discharge. Continuous EEG monitoring was not available. However, it was used to diagnose non-convulsive status epilepticus at onset. All children underwent computed tomography scanning of the brain and other relevant investigations.

All 20 children received intravenous midazolam at 0.15 mg/kg as a bolus followed by a constant infusion starting at 1 μg/kg/min up to 5 μg/kg/min increasing by 1 μg/kg/min every 15 minutes until complete control of seizures was achieved. The optimum rate of infusion at which seizure control was achieved was maintained for a period of 24 hours. Subsequently the midazolam infusion rate was gradually decreased (by 1 μg/kg/min every two hours) until tapering was completed.

Variables such as age, weight, sex, history of seizures, underlying diseases, and the time required for control of seizures were carefully recorded for each patient. Vital parameters including respiratory rate, heart rate, and blood pressure were documented. The oxygen saturation of each child was monitored continuously by pulse oximetry.

The children were also monitored for the development of adverse effects of benzodiazepines including hypotension, hypoxia, and respiratory depression. In order to exclude electrolyte and metabolic disturbances as a cause of the seizures, blood samples were taken on admission and at 24 hours to measure circulating sodium, potassium, calcium, glucose, and magnesium concentrations.

Results

Of the 20 children with status epilepticus admitted to our high dependency care unit, 15 were boys (table 1). The mean age was 4.07 years (range 2 months to 13 years). Eleven children had a history of seizures and were already taking antiepileptic drugs, which included various combinations of sodium valproate (n = 9), carbamazepine (n = 3), phenobarbitone (n = 2), phenytoin sodium (n = 6), clonazepam (n = 2), ethosuximide (n = 1), vigabatrin (n = 1), and lamotrigine (n = 1). Eight children had idiopathic epilepsy, three acute purulent meningitis, three acute meningoencephalitis, and the remainder had various vascular or degenerative lesions of the brain.

The type of status presented in the children as follows: generalised tonic-clonic (n = 13), partial seizure status (partial motor (n = 2), complex partial (n = 2)), myoclonic astatic (n = 1), and Lennox-Gastaut status (myoclonic + tonic; n = 2). Myoclonic seizures were seen as a combination of myoclonus with generalised tonic-clonic status in three others. Twelve patients had refractory status epilepticus (table 2). Their seizures continued for more than 30 minutes after administration of diazepam, followed by phenytoin sodium/phenobarbitone or both intravenously. Eight children being followed up in the outpatient department who were on various antiepileptic drugs were given midazolam infusion alone.

Complete arrest of seizures was achieved with midazolam infusion in all but one of the 20 children (table 3). The non-responder had Batten’s disease. The mean time taken to control status in refractory status epilepticus was 64.6 minutes (range 15–240 minutes) and in established status epilepticus, 34.3 minutes (range 10–60 minutes). However, the mean time between the start of midazolam infusion and total cessation of seizures in all patients was 0.90 hours (54 minutes). The mean infusion rate of midazolam required to control the seizures completely was 2 μg/kg/min (range 1–5 μg/kg/min). Fifteen of the patients were controlled with 2 μg/kg/min or less (table 4). In two of the children there was a transient fall in oxygen saturation (to 90%), as demonstrated by pulse oximetry. However, there was no associated hypotension or change in heart rate. No active intervention was needed except oxygen by mask (2 litres/minute for two hours).

Both these children had received intravenous phenobarbitone (20 mg/kg) from the peripheral hospital three hours before transfer. All the vital parameters of the other 18 children remained well within normal limits. No child required endotracheal intubation or mechanical ventilation. The electrolyte and glucose levels were within normal limits in all the children at the time of admission and 24 hours later. All children regained consciousness at a mean time of 5.1 hours after discontinuation of the midazolam infusion.

Discussion

If normally adequate doses of diazepam, phenytoin, and phenobarbitone fail to terminate seizures in status epilepticus, the condition is then considered to be refractory.7 In a review article, Shorvon divided status epilepticus into early, established, and refractory.8 Refractory is the description if the seizures continue for 60–90 minutes after the initiation of therapy.

Midazolam, a 1,4-benzodiazepine agent of the group of 1,2-annelated benzodiazepines, is a water soluble compound which penetrates the central nervous system rapidly and has a short elimination half life of 1.5–3.5 hours.9 It is commonly used as an amnestic and anxiolytic agent and for operative induction and sedation of critically ill patients.10 The finding of anticonvulsant efficacy in animal studies11 was followed by anecdotal reports of the success of intravenous midazolam in terminating seizures and status epilepticus in humans.6 12

In our study, midazolam infusion controlled seizures in all cases of established status epilepticus and 11of 12 cases of refractory status epilepticus, giving an overall 95% success rate, which is almost equal to the results of Rivera et al.13 One patient with refractory status epilepticus was confirmed to have Batten’s disease, a progressive neurodegenerative disorder, and failed to respond. The mean rate of infusion of midazolam required to control seizures was 2.12 μg/kg/min in refractory status epilepticus and 1.79 μg/kg/min in established status epilepticus (table 3). These results are comparable with those reported for adult patients.12 14 The maximum midazolam dose given to achieve control was 5 μg/kg/min, which is much lower than the 18 μg/kg/min used in an earlier study.13 The mean time required to control the seizure was 34.3 minutes (10–60 minutes) in established status epilepticus and 64.6 minutes (15–240 minutes) in refractory status epilepticus, which is very similar to the 0.78 hours (range 15 minutes to 4.5 hours) found by Rivera et al.13 All types of status epilepticus were treated with midazolam. The majority of our cases (75%) were convulsive, with five cases (numbers 10, 11, 13, 14 and 15, see table 1) that were non-convulsive. About 25% of cases of status epilepticus are non-convulsive,15 16 and aggressive management has been recommended to terminate clinical and EEG detected seizure activities.16 17 The mean infusion rate of midazolam required to control seizures was 1.9 μg/kg/min and the duration was 31 minutes.

It is not clear how midazolam works, when diazepam and phenytoin/phenobarbitone have failed to control seizures.18 The more rapid rate of arrival at the receptors concerned with termination of seizures may be the critical factor.18 It acts more rapidly, is safer and more effective.19 Several other antiepileptic mechanisms of midazolam do not distinguish it from diazepam or lorazepam.20

During the acute phase of seizure disorder, midazolam was more effective and safer for the control of seizures than comparable doses of diazepam.21 Because of the effective control of status epilepticus with midazolam, eight patients who had status epilepticus as the result of drug default were given midazolam alone to control the condition in addition to the defaulted oral antiepileptic drug (given via a nasogastric tube).

Drugs such as pentobarbitone, paraldehyde, and isoflurane have been used to treat status epilepticus with variable degrees of success. Pentobarbitone, which is commonly used in the treatment of the condition, is a general anaesthetic and its use is frequently accompanied by myocardial depression and hypotension.1 As it is a respiratory depressant, its use at high dose for the control of seizures often has to be accompanied by endotracheal intubation and mechanical ventilation. Paraldehyde, which is no longer recommended,22 has serious side effects such as pulmonary haemorrhage, pulmonary oedema, and renal and liver toxicity. Isoflurane, an inhalational general anaesthetic agent, although effective in controlling refractory seizures, produces considerable respiratory depression and muscle relaxation, often warranting endotracheal intubation and mechanical ventilation.23

Ghilain and coworkers found a 15% rate of occurrence of mild hypotension and a 10% frequency of bradycardia when 0.2 mg/kg midazolam was given intramuscularly to adult patients with seizures.24 Two children in our series developed transient mild hypoxia. As the oxygen saturation fell to 90%, the children needed oxygen by mask for two hours. Without any further intervention the oxygen saturation returned to normal. The vital parameters of the remaining 18 children were well within normal limits. No child had to be intubated or mechanically ventilated, and there were no electrolyte abnormalities.

Conclusion

The results of our study using midazolam at 1–5 μg/kg/min as a constant intravenous infusion after a bolus dose of 0.15 mg/kg, substantiate the suggestion that midazolam infusion is an effective and safe therapeutic approach for the management of childhood status epilepticus including the refractory condition. It can be used alone in status epilepticus resulting from drug default. In addition, the dose of midazolam can be conveniently titrated to the requirement of each child. Adverse effects such as respiratory depression are not seen when the drug is given alone or with phenytoin. Midazolam also has the pharmacokinetic advantage over other commonly used drugs in having a shorter duration of action. We suggest that midazolam should be the drug of choice for treatment of childhood status epilepticus.