Articles
`
`Lancet 2006; 367: 1591–97
`See Comment page 1555
`Department of Paediatrics,
`College of Medicine, University
`of Malawi, Malawi (S Ahmad
`FFAEM, J C Ellis MRCPCH,
`H Kamwendo CCMA,
`Prof E Molyneux FFAEM); and
`Paediatric Intensive Care Unit,
`Birmingham Children’s
`Hospital NHS Trust,
`Birmingham, West Midlands,
`UK (S Ahmad)
`Correspondence to:
`Dr Shafi que Ahmad
`doctorshaf@hotmail.com
`
`Effi cacy and safety of intranasal lorazepam versus
`intramuscular paraldehyde for protracted convulsions in
`children: an open randomised trial
`
`Shafi que Ahmad, Jane C Ellis, Hastings Kamwendo, Elizabeth Molyneux
`
`Summary
`Background In sub-Saharan Africa, rectal diazepam or intramuscular paraldehyde are commonly used as fi rst-line
`anticonvulsant agents in the emergency treatment of seizures in children. These treatments can be expensive and
`sometimes toxic. We aimed to assess a drug and delivery system that is potentially more eff ective, safer, and easier to
`administer than those presently in use.
`
`Methods We did an open randomised trial in a paediatric emergency department of a tertiary hospital in Malawi.
`160 children aged over 2 months with seizures persisting for more than 5 min were randomly assigned to receive
`either intranasal lorazepam (100 μg/kg, n=80) or intramuscular paraldehyde (0·2 mL/kg, n=80). The primary outcome
`measure was whether the presenting seizure stopped with one dose of assigned anticonvulsant agent within 10 min
`of administration. The primary analysis was by intention-to-treat. This study is registered with ClinicalTrials.gov,
`number NCT00116064.
`
`Findings Intranasal lorazepam stopped convulsions within 10 min in 60 (75%) episodes treated (absolute risk 0·75, 95%
`CI 0·64–0·84), and intramuscular paraldehyde in 49 (61·3%; absolute risk 0·61, 95% CI 0·49–0·72). No clinically
`important cardiorespiratory events were seen in either group (95% binomial exact CI 0–4·5%), and all children fi nished
`the trial.
`
`Interpretation Intranasal lorazepam is eff ective, safe, and provides a less invasive alternative to intramuscular
`paraldehyde in children with protracted convulsions. The ease of use of this drug makes it an attractive and preferable
`prehospital treatment option.
`
`Introduction
`Acute protracted convulsions in children are one of the
`most common medical emergencies in sub-Saharan
`Africa. In areas with endemic malaria and a heavy
`burden of
`infectious diseases,
`including bacterial
`meningitis, and where resources are scarce, the incidence
`of seizures is many times greater than in well-resourced
`countries.1 Timely
`interventions are necessary
`to
`maintain cardiorespiratory function, stop the seizure,
`and diagnose and treat the underlying disorder. Failure
`to provide such interventions could lead to a protracted
`seizure episode that is more diffi cult to control,
`increasing
`the
`likelihood of death or
`long-term
`neurological sequelae.2,3
`In the prehospital setting in developed countries, rectal
`diazepam is the most commonly used fi rst-line agent in
`childhood seizures. However, concerns over social
`acceptability and convenience of rectal administration
`have led to a search for alternatives. Two randomised
`trials have shown buccally-administered midazolam is
`better than, or at least as eff ective as, rectal diazepam in
`childhood seizures.4,5
`In developing countries, the context of care and the
`challenges faced in medical emergencies are diff erent
`from those in developed countries. Patients often present
`late in an illness because of an absence of organised
`
`www.thelancet.com Vol 367 May 13, 2006
`
`prehospital emergency services, the cost of transport,
`and the distances to travel. Many seizures at presentation
`are protracted and need several doses of anticonvulsant
`agents to be controlled. The risks of benzodiazepine-
`induced respiratory depression are substantial if the
`drug is given in excessive doses or in combination with
`other sedative agents, especially when provision of
`oxygen therapy might be scarce, or skills and resources
`for short-term ventilatory assistance are unavailable.6
`Seizures are frequently due to acute central nervous or
`severe systemic
`infections, whereas
`in developed
`countries protracted febrile convulsions or poorly
`controlled
`idiopathic recurrent seizures are most
`common.5 The risk of recurrent seizures is great, and
`seizures might take place on wards that are inadequately
`staff ed
`to monitor and manage
`them. Resource
`limitations reduce the availability of equipment such as
`appropriately sized cannulae, and so emergency therapy
`is often delivered by non-intravenous routes.
`The ideal fi rst-line anticonvulsant agent would be one
`that can be given safely and easily at a primary health-
`care facility. The anticonvulsant should be quick acting,
`have minimum cardiorespiratory side-eff ects, have a
`longlasting eff ect, and be inexpensive.
`In view of its favourable pharmacokinetics and
`potential practical advantages, we wished to assess the
`AQUESTIVE EXHIBIT 1144 Page 0001
`
`1591
`
`

`

`Articles
`
`effi cacy and safety of intranasal delivery of lorazepam
`compared with intramuscular paraldehyde, which is our
`fi rst-line anticonvulsant agent in the treatment of acute
`seizures in children.
`
`results were available to those who wished to know the
`HIV status of their child after full counselling before and
`after the test was done. The trial was approved by the
`research and ethics committee of the College of Medicine,
`University of Malawi (P03/04/248).
`
`Methods
`Study design
`Patients
`Children arriving to the emergency department who
`Between July, 2004, and June, 2005, we did an open
`were convulsing or developed convulsions within the
`randomised trial comparing the effi cacy and safety of
`department were immediately taken to the resuscitation
`intranasal lorazepam with intramuscular paraldehyde in
`room. All convulsing patients were managed according to
`consecutive patients with acute seizures presenting to
`advanced paediatric life support guidelines8 up to the point
`the paediatric emergency department at the Queen
`Elizabeth Central Hospital, a teaching and referral
`of choice of anticonvulsant agent. Children diagnosed as
`hospital in Blantyre, southern Malawi. Participants were
`hypoglycaemic (bedside glucostix ≤2·6 mmol/L, Roche
`children aged between 2 months and 12 years who
`Diagnostics, Boehringer Mannheim, Germany) had
`presented with generalised convulsions continuing for a
`intravenous or intra-osseous access established and were
`minimum of 5 min. Generalised convulsions were
`given glucose. Those with continuing seizures were then
`defi ned as the presence of rhythmic twitching of the
`randomly allocated to receive the named therapeutic
`arms, legs, trunk, or facial muscles, tonic eye deviation,
`options. Blocked randomisation was done in advance by a
`or nystagmoid eye jerking in a comatose child. Exclusion
`computer that randomly generated a table of numbers in
`criteria were any child who had received an anticonvulsant
`batches of ten, and treatment allocations were sealed in
`agent within an hour of presentation, whose seizure had
`unmarked identical envelopes. Investigators were masked
`stopped with cooling or correction of hypoglycaemia, or
`to these allocations before the point of patient treatment.
`who had features consistent with hepatic or hypertensive
`Study randomisation envelopes were kept in a so-called
`encephalopathy or organophosphate poisoning. The
`seizure control tray along with emergency drugs and
`study was done in the resuscitation room at the paediatric
`equipment in the resuscitation room. All children who
`emergency department of the Queen Elizabeth Central
`attended the emergency department were either weighed
`Hospital during its opening hours of 0730–1700, when it
`on arrival or had their weight estimated by use of a locally
`is operated by medical and nursing staff trained in
`modifi ed Broselow tape.9
`paediatric emergency medicine.
`the named
`Study envelopes were opened and
`Most children attending the emergency department
`medication was given by the study investigators (SA or
`live within the Blantyre region, and are not referred by a
`EM). For administration of lorazepam, 100 μg/kg was
`doctor. Fewer patients attend the paediatric emergency
`drawn up in a 1 mL syringe (4 mg/mL vial, Ativan, Wyeth-
`department outside the usual opening hours; however,
`Ayerst, Philadelphia, USA), attached to a mucosal
`patient characteristics in terms of diagnostic spectrum
`atomisation device (Wolfe Tory Medical, Salt Lake City,
`and seizure duration are similar to those who arrive
`Utah, USA) and squirted rapidly into a nostril. To attain
`during the day. The emergency department is a dedicated
`maximum bioavailability of intranasal lorazepam, we
`children’s emergency unit that sees at least 75 000 patients
`kept the child’s head in the recovery position while
`a year, of whom around 30% need admission. About
`placing the atomisation device fi rmly into one nostril and
`1000 children are triaged as priority one (needing
`directing it upwards during administration. Intramuscular
`immediate emergency care) every year, and managed in
`paraldehyde, 0·2 mL/kg (10 mL vial, Medicopharma UK,
`the resuscitation room.
`Romford, UK) was drawn up into a syringe and delivered
`Staff of the paediatric department were fully informed
`immediately into the buttock or thigh after sterilisation
`of our study before it began. Rapid seizure control was
`of the injection site.
`our fi rst priority. Informed consent for the participants
`During seizure activity, humidifi ed oxygen from an
`was obtained in accordance with the code of federal
`oxygen concentrator was provided through a mask or
`regulations section 50.24 and 50.25 for clinical research
`nasal cannula if the patient arrived with or developed an
`operations (the ethical regulation on emergency trials).7
`oxygen saturation reading of 92% or less. Oxygen
`saturation was monitored continuously for 30 min and
`While the child was being assessed and medications
`recorded on arrival, every 2 min until 10 min after arrival,
`drawn up, the guardian was informed of the procedures,
`then every 5 min until 30 min. Blood pressure was
`and detailed descriptions of care took place in real time
`measured and recorded according to the same time
`so as not to delay treatment. Informed consent after a full
`schedule
`after
`administration of
`the
`assigned
`explanation was requested from the guardians as soon as
`anticonvulsant agent (Drager Dialog 2000 monitor type
`the child was stable. All guardians gave written informed
`no 5706200, Medizintechnik GbmH, Lubeck, Germany,
`consent for their children to participate. Anonymous
`with a Nellcor pulse oximeter sensor, 800 series,
`HIV spot testing was done and was not linked to any
`Pleasanton, CA, USA). The end of a seizure was defi ned
`patient identifi cation data in the fi nal analysis. The
`AQUESTIVE EXHIBIT 1144 Page 0002
`
`www.thelancet.com Vol 367 May 13, 2006
`
`1592
`
`

`

`Articles
`
`Airway position,
`high-flow oxygen,
`check glucose
`
`Screen for eligibility
`
`Randomisation
`
`Lorazepam intranasal
`100 μg/kg
`
`Paraldehyde
`intramuscular 0·2 mL/kg
`
` If child still fitting 10 minutes after study treatment give rescue therapy
`
`Paraldehyde intramuscular (0·2 mL/kg)
`First-line rescue agent
`
`If child still fitting at 20 minutes after study treatment give
`
`Phenobarbitone intravenous/intraosseous (15 mg/kg)
`Second-line rescue agent
`
`If child still fitting at 30 minutes after study treatment give
`
`Phenytoin intravenous/intraosseous (18 mg/kg)
`Third-line rescue agent
`
`Figure 1: Study treatment algorithm
`
`as cessation of all visible convulsive activity. Treatment
`was regarded as successful if one dose of the assigned
`treatment stopped the presenting seizure within 10 min
`of administration. If the seizure continued beyond
`10 min, the treatment was judged ineff ective and a rescue
`regimen was followed (fi gure 1). All children had a thick
`blood fi lm taken for detection of malaria parasitaemia.
`Lumbar puncture with cerebrospinal fl uid analysis was
`done if there were no contraindications (such as
`irregular respirations, posturing, or other signs of
`acute raised intercranial pressure). Plasma electrolytes
`were analysed with a point of care device (i-STAT1
`handheld portable clinical analyser, Abott Laboratories,
`IL, USA). HIV-1 and HIV-2 test cards (Determine, Abbott
`Laboratories, Tokyo, Japan) were used to assess HIV
`serological
`status. After
`leaving
`the emergency
`department, all patients were admitted to a ward for
`high-dependency care.
`The primary outcome measure was whether the
`presenting seizure stopped or not with one dose of
`assigned anticonvulsant agent within 10 min of
`administration. We chose this time for further action to
`remain consistent with current standard treatment
`algorithms for acute paediatric convulsions.8,10 We also
`obtained data for several secondary outcome measures:
`these included time between seizure onset and drug
`administration, time between opening the study envelope
`and drug administration, time from drug administration
`to cessation of convulsion, frequency of episodes needing
`two or more rescue anticonvulsant agents, development
`
`www.thelancet.com Vol 367 May 13, 2006
`
`of hypotension (5 mm Hg or more reduction from that at
`enrolment for systolic and diastolic pressures) or hypoxia
`(oxygen saturation ≤92%) for 30 min after drug
`administration, and whether seizures recurred within
`24 h of termination of the presenting convulsion. Final
`diagnosis of the presenting convulsion was made once
`all the available diagnostic and demographic data had
`been assessed. All patients were followed-up until
`discharge or death. Patients needing two or more rescue
`agents represent a group with seizures that were diffi cult
`to control, and so this outcome measure is also useful in
`indicating those patients likely to need further resources
`in terms of drugs and high dependency nursing care.
`
`Statistical analysis
`We anticipated, from previous clinical experience, that
`about 25% of children in the paraldehyde group would
`need a second dose of anticonvulsant after 10 min. We
`were unable to accurately anticipate the results with
`intranasal lorazepam, but using the fi ndings reported in
`a Cochrane review of the use of rectal lorazepam, we
`assumed that 5% would need a second dose of
`anticonvulsant.11 Using these estimations, we calculated
`that a total sample size of 154 people was needed to
`measure a 20% diff erence in primary outcome, with 90%
`power (β) and 5% signifi cance (α). Data were entered on
`study proforma and double entered into a Microsoft
`Access fi le. We used Stata version 8 for analysis of the
`data. We used
`the following summary statistics:
`Wilcoxon’s rank sum test to compare medians, Fisher’s
`exact test to look for associations between variables, and
`χ2 tests for absolute risk. The primary analysis was on an
`intention-to-treat basis.
`
`Role of the funding source
`The sponsor of the study had no role in the study design,
`data collection, data analysis, data interpretation, or
`writing of the report. The corresponding author had full
`access to all the data in the study and had fi nal
`responsibility for the decision to submit for publication.
`
`Results
`All 160 children randomised were followed up with no
`protocol violations by the end of the study. Of the
`42 children excluded, most (29 episodes) stopped
`convulsing before randomisation and their convulsions
`were attributed to fever or hypoglycaemia (fi gure 2).
`Between treatment groups, children were similar for sex,
`age, median seizure duration pretreatment, median time
`to drug delivery, underlying cause, and HIV status.
`Seizures were exclusively due to acute brain infection
`secondary to cerebral malaria or bacterial meningitis in
`51 (64%) of 80 children in the intranasal lorazepam
`group, and 53 (66%) of 80 children in the intramuscular
`paraldehyde group (table 1).
`The presenting seizure stopped within 10 min of the
`assigned study drug in three-quarters of children in the
`AQUESTIVE EXHIBIT 1144 Page 0003
`
`1593
`
`

`

`Articles
`
`202 screened
`
`160 randomised
`
`42 not recruited
` 18 febrile convulsions stopped
` spontaneously
` 11 hypoglycaemic convulsions
` stopped with glycaemic correction
` 5 other anticonvulsant within 1 hour
` 4 less than 2 months of age
` 1 primary idiopathic epilepsy
` stopped spontaneously
` 1 hypertensive encephalopathy
` 1 acute fulminant hepatic failure
` 1 organophosphate poisoning
`
`80 assigned lorazepam
`
`80 assigned paraldehyde
`
`80 completed trial
`
`80 completed trial
`
`Figure 2: Trial profi le
`
`92% after treatment. In both these children oxygen
`saturations remained below 92% for 10 min and needed
`supplemental oxygen only. In the paraldehyde group,
`one child de-saturated to 92% after two doses of
`paraldehyde and one dose of phenobarbitone. This child
`needed oxygen therapy for 2 h to maintain normal
`oxygen
`saturations. Hypoxaemia
`preceded
`the
`administration of allocated treatment in eight cases,
`which was attributed to severe malarial anaemia (four
`cases), co-existent pneumonia (two), near drowning
`(one) and lung aspiration of gastric contents before
`presentation at hospital (one). The 95% exact binomial
`CI for the proportion of cardiorespiratory events is
`0–4·5% in each group.
`The recorded number of deaths from underlying causes
`was small and much the same between the randomised
`groups. 75 children had a pretreatment-seizure duration
`of less than 2 h, eight of whom died (absolute risk 0·1,
`95% CI 0·04–0·2). Of the 85 patients with a pretreatment-
`seizure duration of greater than 2 h, 20 died (absolute
`risk 0·23, 95% CI 0·15–0·34; RR 0·45, 95% CI 0·21–0·86,
`p=0·03). 19 children were found to be infected with HIV
`at presentation. Of these children, eight were male,
`11 female, and the median age was 18 months (IQR 7–48).
`In HIV-infected children, eight of 19 had an underlying
`diagnosis of acute bacterial meningitis versus 13 of 141 in
`children not
`infected with HIV (4·57, 2·18–9·56
`p=0.0001). The proportion of deaths was also greater in
`those with HIV infection. Seven deaths occurred in the
`19 HIV-infected children compared with 21 of 141 non-
`infected (RR 2·51, 95% CI 1·23–5·10, p=0·02). There
`was no signifi cant relation between HIV status and
`allocated treatment effi cacy or safety.
`
`intranasal lorazepam group and in about 60% of those
`on intramuscular paraldehyde (table 2). The median time
`for the presenting seizure to stop after drug administration
`did not diff er between groups (table 2). However, the
`number of children needing two or more rescue
`anticonvulsant agents was signifi cantly lower in the
`intranasal lorazepam group than in the intramuscular
`paraldehyde group. Logistic regression analysis showed
`that only the diff erence in treatment was signifi cant, and
`that none of the diagnoses were statistically important in
`aff ecting the need for two or more rescue agents either
`alone or in interaction with the intranasal lorazepam
`Discussion
`group or the intramuscular paraldehyde group (5% level
`of signifi cance, odds ratio 6·33, 95% CI 1·64–24·45,
`Intranasal lorazepam stopped three-quarters of the
`p>0·007). Only eight of 80 (10%) children in the intranasal
`presenting seizures in less than 10 min. Intramuscular
`paraldehyde was eff ective in two-thirds of children within
`lorazepam group and 11 of 80 (14%) children in the
`10 min. This diff erence was not signifi cant, although
`intramuscular paraldehyde group had a
`further
`signifi cantly
`fewer patients
`receiving
`intranasal
`convulsion within 24 h of presentation (table 2).
`lorazepam needed two or more rescue anticonvulsant
`Median oxygen saturation at enrolment was 98% (IQR
`agents than did those who received paraldehyde. There
`95–99) in the intranasal lorazepam group and 99%
`were no clinically important cardiorespiratory events,
`(97–99) in the intramuscular paraldehyde group. After
`suggesting both treatment options were safe. There were
`treatment the median readings were 99% (97–100) and
`fewer recurrent seizures in the intranasal lorazepam
`99% (97–99), respectively. In the intranasal lorazepam
`group but the diff erence was not signifi cant.
`group, there were 15 children in whom systolic blood
`In several children, the severity of tonic-clonic activity
`pressure fell by at least 5 mm Hg, with a median
`diminished over time after initial treatment, and the
`reduction of 7 mm Hg (range 5–20 mm Hg) and
`initially allocated drug could well have taken more than
`12 children in whom diastolic blood pressure fell by at
`10 min to eff ectively stop the presenting seizure. Our
`least 5 mm Hg with a median of 7·5 mm Hg
`study endpoint, the visible cessation of all convulsive
`(5–16 mm Hg). In the intramuscular paraldehyde group,
`activity, was tightly defi ned and therefore waiting beyond
`there were 16 children with a systolic blood pressure
`10 min in selected children would have introduced
`reduction of at least 5 mm Hg with a median of
`performance bias. In
`the absence of continuous
`6·5 mm Hg (5–10 mm Hg) and four children with a
`electroencephalogram (EEG) monitoring, we did not
`diastolic blood pressure reduction of at least 5 mm Hg,
`think that breaking the study protocol was safe. Overall,
`median 6·5 mm Hg (5–20 mm Hg). In the lorazepam
`the number of seizures controlled within 20 min
`group, two children had a fall in oxygen saturation to
`AQUESTIVE EXHIBIT 1144 Page 0004
`
`www.thelancet.com Vol 367 May 13, 2006
`
`1594
`
`

`

`Articles
`
`(intramuscular
`therapy
`rescue
`requiring fi rst-line
`the
`intranasal
`ten episodes
`in
`paraldehyde) was
`lorazepam group and 11 episodes in the intramuscular
`paraldehyde group.
`Children who needed two or more rescue anticonvulsant
`agents were more likely to have been initially allocated to
`receive
`intramuscular paraldehyde
`than
`intranasal
`lorazepam. As a measure of effi cacy, this diff erence is
`notable. It might indicate poorer availability of blood-
`borne paraldehyde circulating to the brain, as a result of
`slower absorption and systemic release when given
`intramuscularly, or intranasal lorazepam might have a
`more favourable synergistic eff ect with fi rst-line rescue
`therapy than intramuscular paraldehyde. Overall, 19 (12%)
`of 160 patients had further seizures within 24 h of initial
`convulsion, of whom ten (53%) children had cerebral
`malaria. 16 (84%) of 19 patients with seizure recurrence
`had not received second or third-line rescue anticonvulsant
`agents. These data suggest that prophylactic therapy
`should be considered in selected patients. Another
`randomised
`controlled
`study of phenobarbitone
`prophylaxis has been done in childhood cerebral malaria
`in a similar setting (Kilifi , Kenya) to our study. It showed
`a reduction in seizure frequency but with an unacceptable
`doubling of mortality.12 This fi nding is important, and
`needs further study.
`second-line
`or
`the fi rst-line
`Paraldehyde
`is
`anticonvulsant agent in many countries in sub-Saharan
`Africa because of its favourable safety and effi cacy
`profi le.10,13,14 Paraldehyde acts within 5–10 min and can
`last up to 8 h. The major disadvantages of paraldehyde
`are a potential for local injury when given intramuscularly,
`incompatibility with plastics, and substantial cost. At
`around US$30 per 10 mL vial, the cost of the drug alone
`for one intramuscular dose at 0·2 mL/kg for a child
`weighing 10 kg is $6.15 This amount compares poorly
`with the cost of lorazepam, which at roughly $1 per
`4 mg/mL vial for the same sized child as one intranasal
`dose at 100 μg/kg
`is $0·25.15 Paraldehyde given
`intramuscularly is preferred over the rectal route since it
`is absorbed quicker, is easier to administer, and avoids
`the risk of bowel irritation or perforation, potential
`complications if mixed inappropriately or given in
`decomposed form.16 In
`those allocated
`to receive
`
`Demographic characteristics
`Sex
`Boy
`Girl
`Age (months, median, IQR)
`Seizure duration before treatment (min, median, IQR)
`Time to drug delivery once study envelope opened (seconds, median, IQR)
`HIV infected
`Cause of seizure (fi nal diagnosis)*
`Cerebral malaria
`Protracted febrile convulsion
`Metabolic derangement (serum Na <130 or >145 mmol/L)
`Alone
`Co-existing
`Acute bacterial meningitis
`Primary idiopathic epilepsy
`Drug toxic side-eff ect
`Pre-existing neurological abnormality
`Aseptic meningitis
`Post head injury
`Hypoxic brain injury
`Tuberculosis meningitis
`Cerebral abscess
`
`Lorazepam
`(n=80)
`
`Paraldehyde
`(n=80)
`
`44 (55%)
`36 (45%)
`18·5 (9–33)
`127 (66·5–220)
`60 (60–60)
`8 (10%)
`
`43 (54%)
`37 (46%)
`19 (10·5–36)
`120 (35·5–252)
`60 (60– 60)
`11 (14%)
`
`39 (49%)
`16 (20%)
`
`5 (6%)
`7 (9%)
`12 (15%)
`3 (4%)
`2 (2·5%)
`1 (1%)
`1 (1%)
`1 (1%)
`1 (1%)
`1 (1%)
`1 (1%)
`
`44 (55%)
`15 (19%)
`
`6 (7·5%)
`6 (7·5%)
`9 (11%)
`3 (4%)
`4 (5%)
`2 (2·5%)
`1 (1%)
`0
`0
`0
`0
`
`Data are number (%) unless otherwise stated. *Some children had two or more co-existing conditions.
`
`Table 1: Baseline characteristics of treated children
`
`intramuscular paraldehyde, 61% had seizure control
`within 10 min. Our anticipated control of 75% of seizures
`by one dose of paraldehyde was an educated guess after
`years of experience of using this drug in sub-Saharan
`Africa. Our results, however, suggest intramuscular
`paraldehyde was not as eff ective as we had anticipated,
`and we acknowledge that our estimate might have been
`overoptimistic.
`Midazolam administered buccally or nasally is eff ective
`for initial seizure control;4,5,17,18 however, in view of its
`shorter duration of action (3–4 h) than lorazepam (12–18 h)
`we were concerned midazolam might not act for long
`enough to prevent seizure recurrence. Such a short-lived
`eff ect is a major concern in settings in which seizures are
`
`Lorazepam
`(n=80)
`
`Paraldehyde
`(n=80)
`
`Absolute risk (95% CI)
`
`Presenting seizure stopped within 10 min
`Seizures needing two or more rescue
`anticonvulsant agents
`Seizure recurrence within 24 h
`Died
`Time for fi t to stop (min, median, IQR)
`
`60 (75%)
`8 (10%)
`
`8 (10%)
`15 (19%)
`7·5 (4·5–11·5)
`
`49 (61%)
`21 (26%)
`
`11 (14%)
`13 (16%)
`8 (5–21)
`
`Data are number (%) unless otherwise stated. *Wilcoxon’s rank-sum test.
`
`Table 2: Primary and secondary outcome measures
`
`Lorazepam
`
`0·75 (0·64–0·84)
`0·1 (0·04–0·19)
`
`0·1 (0·04–0·19)
`0·19 (0·11–0·29)
`··
`
`Paraldehyde
`
`0·61 (0·49–0·72)
`0·26 (0·17–0·37)
`
`0·14 (0·07–0·23)
`0·16 (0·08–0·26)
`··
`
`p
`
`0·06
`0·007
`
`0·46
`0·68
`0·06*
`
`www.thelancet.com Vol 367 May 13, 2006
`
`AQUESTIVE EXHIBIT 1144 Page 0005
`
`1595
`
`

`

`Articles
`
`often caused by continuing pathological changes in the
`central nervous system. Lorazepam is eff ective in acute
`seizure control and has few clinically important adverse
`eff ects when given intravenously.19,20 In the absence of
`intravenous access, buccal or intranasal routes for
`lorazepam are appealing alternatives. We chose the
`intranasal route for delivery because in our clinical
`practice we see children with pronounced salivation. We
`do not know if excessive salivation can cause dilution or
`delayed absorption, but did not want to have that doubt if
`the buccal route was unsuccessful. In similar previous
`studies4,5 undertaken in the UK, many children had coryza
`and upper respiratory tract infections and the researchers
`did not wish that to aff ect the results of absorption if the
`intranasal route was used. There were 11 convulsive
`episodes in our study in which children had evidence of
`nasal congestion, six in the intranasal lorazepam group
`and fi ve in the intramuscular paraldehyde group. All
`cases responded to their allocated treatment within
`10 min. We used an atomisation device (connected to a
`1 mL syringe) that aerosolised the treatment solution.
`Aerosolisation compared with dripping a solution into
`the nasal cavity with a syringe has improved particle
`distribution, and targets a larger surface area of olfactory
`epithelium, thus increasing benzodiazepine maximum
`plasma concentration from about 50% to 80% of that
`reached when given intravenously.21 We used an undiluted
`lorazepam 4 mg/mL solution, thus gaining maximum
`concentration while keeping volumes low. We did not
`fi nd access to the nasal cavity diffi cult with the mucosal
`atomisation device, even in the youngest children. In no
`children was there evidence of nasal discharge of the drug
`or aspiration when this technique was used. We
`acknowledge that the number of children who had
`protracted seizures with evidence of upper respiratory
`tract infection was small in our study. Whether in a
`developed or resource poor setting, in view of the
`diff erences in patient characteristics, there are valid
`theoretical concerns over the routes of absorption. We
`need to compare benzodiazepines given nasally, buccally,
`and via intravenous routes to see if all routes are as safe
`and eff ective as each other.
`and
`intraneuronal
`both
`Neural
`connections,
`extraneuronal between the olfactory epithelium and the
`brain provide a unique pathway for the non-invasive
`delivery of
`therapeutic agents
`to
`the CNS. The
`intraneuronal pathway involves axonal transport and
`needs hours to days for drugs to reach the brain, whereas
`the extraneuronal pathway is thought to rely on bulk fl ow
`transport through perineural channels, delivering drugs
`directly to the brain within min.22 Experimental study
`with nerve growth factor has shown direct entry into the
`brain via extraneuronal pathways involving olfactory and
`trigeminal sensory nerves, thus bypassing the slower
`penetration of the blood-brain barrier.23 We postulate a
`similar mechanism for intranasal lorazepam’s rapid
`entry and delivery to cerebral tissue.
`
`A large number of patients in our study had continuous
`seizures for longer than 2 h before treatment. Longlasting
`continuous seizures or those refractory to treatment and
`continuing for 2 h or more have a mortality of between 0%
`and 32% in well-resourced settings. Worse outcomes are
`associated with an acute symptomatic cause or progressive
`encephalopathy, multifocal or generalised abnormalities
`on initial EEG, age under 12 months, and duration of
`seizure pretreatment of 2 h.24–27 In our study, 28 (17·5%) of
`160 children died. This high rate of death is likely to
`indicate a protracted seizure duration pretreatment and
`incidence of progressive encephalopathy predominantly
`of infective origin.
`In comparison with previous studies in an emergency
`setting, which have been heavily weighted with febrile
`convulsions and known epilepsies, we treated more
`convulsions due to infectious causes. This group might
`be perceived to be more diffi cult to control, but we were
`nevertheless encouraged by the success of intranasal
`lorazepam. However, we acknowledge the limitations of
`clinical assessment alone in defi ning cessation of seizure
`activity, and that we were unable to exclude subclinical
`persisting electroconvulsive activity in our setting.
`Lorazepam delivered intranasally fulfi ls many of the
`criteria for an ideal combination of drug and delivery
`system for the treatment of protracted seizures in a cost-
`restrained setting. Its favourable cost compared with
`paraldehyde and its effi cacy and safety should make it a
`preferred option within the benzodiazepine family before
`and during hospital care.
`Contributors
`S Ahmad, J C Ellis, H Kamwendo and E Molyneux designed the study
`and cared for patients clinically. S Ahmad, J Ellis, and E Molyneux
`collected and entered the data. S Ahmad maintained the database.
`S Ahmad and E Molyneux analysed the data. All investigators
`contributed to the writing of the fi nal draft of the report.
`
`Confl ict of interest statement
`We declare that we have no confl ict of interest.
`
`Acknowledgments
`This study was supported by an academic grant from the College of
`Emergency Medicine (UK). Mucosal atomisation devices were supplied
`at no cost by Wolfetory Medical, Salt Lake City, Utah, USA. We thank the
`medical, nursing, and laboratory staff who cared for the patients in this
`study, Mavuto Mukaka of the Wellcome Trust Research Laboratories,
`Blantyre, who helped with statistical analyses, and the children and
`families who participated.
`
`2
`
`References
`1 Molyneux EM, Walsh AL, Forsyth H, et al. Dexamethasone
`treatment in childhood bacterial meningitis in Malawi: a
`randomised controlled trial. Lancet 2002; 360: 211–18.
`Shorvon S. The Management of status epilepticus.
`J Neurol Neurosurg Psychiatry 2001; 70 (suppl 2): 22–27.
`3 Knudsen FU. Rectal administration of diazepam in solution i