Pharmacokinetics in children

There is more limited information related to the disposition of benzodiazepines in children; these studies are detailed in table 3. The ontogeny of the CYP-dependent metabolism pathways is likely to alter benzodiazepine clearance. CYP activity is low at birth, increasing to supranormal adult levels at age 2–3 years before gradually declining to adult levels by age 4.13 In particular, these differences explain the much longer elimination half-life seen in preterm infants.

How quickly do they work?

The rapid entry of benzodiazepines to the central nervous system and other tissues with a good blood supply is commensurate with their short distribution half-lives. Following this initial distribution, they are redistributed into less well-perfused tissues; a redistribution which is most rapid for those benzodiazepines with the greatest lipophilicity.14

Route of administration clearly has a major impact on the speed of onset of anticonvulsant properties for benzodiazepines and this will be dealt with later. However, after intravenous administration of 10 mg diazepam and 4 mg of lorazepam, onset of action ranged from immediate to 10 min (median 2 min) for the former and immediate to 15 min (median 3 min) for the latter in a double-blind trial of both drugs in adult patients in status epilepticus.15 In an uncontrolled report of the use of midazolam in 20 adult patients presenting to an emergency department with seizures (12 in status epilepticus), all seizure activity was stopped in less than 1 min.16 It is clear therefore, that once these benzodiazepines are present within the intravascular space, their effect is rapid.

There is relatively sparse data available to attempt to quantify the minimum plasma concentrations at which each of the benzodiazepines is effective in suppressing seizure activity. For diazepam, a plasma concentration of between 200 and 600 μg/l is thought to be adequate.17 18 For lorazepam, the figures are less certain, although a plasma concentration of 20–30 μg/l improved seizure control in patients with intractable partial seizures.19 For midazolam, a therapeutic plasma concentration has not been defined. However, there appears significant intraindividual variability of threshold plasma concentrations for each benzodiazepine, beyond which brain seizure activity will cease. There are several reasons for this. Animal studies indicate that neuronal surface expression of benzodiazepine-sensitive GABA_A receptors is reduced by continuing electrical activity,20 suggesting that the longer a seizure persists, the higher the plasma concentration of benzodiazepine required to stop it. Since most studies recruit a heterogeneous group of patients having seizures of varying durations, this introduces inherent variability when trying to define a therapeutic plasma concentration. This heterogeneity of patient and cause for prolonged seizures contributes to variability in other ways: evidence is emerging that individual pathologies may increase benzodiazepine pharmacoresistance in seizures.21 Other methodological shortcomings play their part: when considered, the establishment of minimal therapeutic concentrations is almost always a more minor objective of efficacy trials and therefore robust studies are limited.

How long do they work for?

Length of anticonvulsant effect for the different benzodiazepines has been generally determined from non-comparative studies. In spite of its long elimination half-life, diazepam appears to have a short duration of anticonvulsant effect of about 15–30 min,22 with a risk of recurrence of seizures in about 50% of patients.23 This is probably related to its rapid redistribution as a result of its high lipophilicity. By comparison, lorazepam is associated with a much lower risk of seizure recurrence with duration of anticonvulsant effect of 24–48 h.22 This is incongruous with its short elimination half-life and reflects its relatively low lipid solubility, preventing redistribution and causing longer persistence in the intravascular space compared with diazepam.

Midazolam has a short duration of action. In a retrospective Japanese case series, of 162 children whose seizures were terminated by a bolus of intravenous midazolam, only 28 remained seizure-free, with a further 119 requiring a continuous midazolam infusion for seizure control, and 15 having recurrent seizures despite infusion.24 This is related to its short elimination half-life combined with its lipid solubility causing rapid redistribution from the intravascular space, followed by rapid excretion, after its initial anticonvulsant effect.

Side effects

Although generally well tolerated, as a class, the benzodiazepines have a number of adverse effects. Drowsiness, confusion, amnesia and hangover effects are all commonly reported during routine use. However, the most relevant of these relating to the treatment of prolonged seizures is the risk of respiratory depression leading to a requirement for intubation and ventilatory support. In a North London cohort of children treated for convulsive status epilepticus, more than two doses of benzodiazepines were associated with respiratory depression (OR=2.9, 95% CI 1.4 to 6.1).25

In practice, lorazepam has been reported to be associated with less respiratory depression than diazepam, with one prospective comparison in children reporting a rate of 4% for lorazepam compared with 21% for diazepam when given intravenously.26 This is confirmed by other retrospective case series.27 28

Intravenous route

It is clear that the intravenous route is the most reliable and rapid way to administer benzodiazepines in prolonged seizures. In the UK, guidelines for the management of prolonged seizures in children29 suggest that of the benzodiazepines, lorazepam is the preferred initial intravenous drug, supported by data showing an equivalent efficacy with diazepam, a longer duration of action and a lower reported incidence of respiratory depression.26 30 However, its major drawback is the perceived need for intravenous access, the achievement of which can be difficult in small children, and is unlikely to be achieved out of hospital. The apparent need for rapid termination of prolonged seizures would mandate a treatment that could be administered by parents, caregivers or paramedics early in the course of presentation. This has led to the investigation of alternative routes of administration of benzodiazepines and a number of studies have examined plasma concentrations in children after administration via non-intravenous routes. These are summarised in table 4. The lack of a defined threshold therapeutic plasma concentration for each of the benzodiazepines means that although each of the drugs appear to be absorbed via these routes, comparative trial data are required to interpret their relative efficacy.

The results of comparative trials in children are summarised in table 5. Six were considered in a Cochrane review of the management of tonic–clonic convulsions in children,31 but only four of those listed were incorporated and the other two32 33 were excluded because they included some patients over the age of 16 years32 and included seizure types other than tonic–clonic.32 33 All the trials listed are of variable quality—it is beyond the remit of this review to consider this in detail; however, differences in the detail with which each study reported definitions of termination of seizure activity, seizure recurrence and adverse effects—for example, respiratory depression, complicate direct comparisons. In addition, comparing one drug given intravenously with another given by a transmucosal route is probably not fair and makes interpretation more fraught, particularly when trying to select the best drug for prehospital use. For this reason, the data for the studies have been grouped according to administration route.

Rectal route

The rectal route is a well-tolerated route of administration which, in the case of diazepam, has been used widely. It can provide rapid and reliable absorption. However, variable efficacy has been reported. Recent comparative studies demonstrate treatment success in 27–88.5% (table 5).This may be a reflection of differences in the aetiology of the seizures.

Lorazepam is also well tolerated when administered by the rectal route. Its rate of absorption appears to be slow with a t_max of 67.5 ± 42.1 min in healthy adult volunteers.34 Only one small trial has studied its efficacy, demonstrating that of six children treated, all respondedsuccessfully.26

Midazolam has been shown to be rapidly absorbed from the rectum in children.35 36 However, no efficacy data have been published.

Intranasal route

The intranasal route for delivery of benzodiazepines has been the focus of significant interest in recent years, particularly with regard to midazolam. For diazepam, some pharmacokinetic data exist: after intranasal administration of diazepam to nine adult volunteers, a t_max of 18 ± 11 min was observed with bioavailability of 50.4 ± 23.3%,37 but no trial appears to have been carried out in clinical practice.

Intranasal lorazepam has recently been demonstrated to be an effective, safe, non-invasive treatment for prolonged seizures in children in Malawi, using a dose of 100 µg/kg in an open randomised trial compared with intramuscular paraldehyde.38

A significant number of clinical studies have been carried out in children relating to the efficacy of intranasal midazolam for seizures (table 5). Equivalent efficacy has been reported in head-to-head comparisons with intravenous diazepam39 40 and it appears to be better than rectal diazepam.31 33

Buccal route

The buccal or sublingual route is growing in popularity for the delivery of benzodiazepines for acute seizures, with buccal midazolam replacing rectal diazepam as the preferred treatment in some centres when the intravenous route is not available. It is more socially acceptable than the rectal route and seems to be preferred to the intranasal route as it may provide less variable absorption.41

An audit of the administration of the rectal formulation of diazepam into the buccal cavity in children suggests possible efficacy,42 but there is little other published evidence.

For midazolam, however, the route has been examined extensively. In children with malaria, a median t_max of 10 min (range 5–40) has been reported with 87% bioavailability compared with intravenous administration43 and there is good evidence of its clinical efficacy. In head-to-head comparisons with rectal diazepam, buccal midazolam has demonstrated equivalent or superior efficacy.32 44 45

The pharmacokinetics of lorazepam following buccal administration have been characterised in healthy adult volunteers. After administration of a 2 mg sublingual tablet of lorazepam, median t_max was 2.25 h with systemic availability of 98.2% compared with the intravenous route.46 This study recognised that C_max and t_max could be improved by using a preparation of lorazepam that was more easily absorbed, since the in vitro dissolution rate of the tablet was reported as 84% in 1 h. There is little clinical experience, however, with the use of buccal lorazepam in seizures. Only one report, an uncontrolled case series, could be identified. In that study, 10 children with intractable epilepsy and serial seizures were treated with sublingual tablets of lorazepam.47 Eight children responded and had no subsequent seizures and two had a partial response. The lorazepam was felt to be effective 5–30 min after administration.