Intranasal versus intravenous lorazepam for control of acute seizures in children: A randomized open-label study

Summary

Purpose:  Intravenous lorazepam is considered the drug of first choice for control of acute convulsive seizures. However, resource or personnel constraints necessitate the study of alternative routes and medications. This study compared the efficacy and adverse effects of intranasal versus intravenous lorazepam in children aged 6–14 years who presented with acute seizures.

Methods:  This was a randomized open-label study conducted at an Indian hospital from August 2008 to April 2009. One hundred forty-one consecutive children aged 6–14 years who presented convulsing to the emergency room were included. After stabilization, the children were randomized to receive either intravenous or intranasal lorazepam (0.1 mg/kg, maximum 4 mg). The primary outcome measure was clinical seizure remission within 10 min of drug administration. The study was registered with clinicaltrials.gov (NCT00735527).

Key Findings:  Seventy patients were randomized to receive intravenous and 71 to receive intranasal lorazepam. The patients in the two groups were comparable at baseline. Clinical seizure remission within 10 min of drug administration was found in 80% of the intravenous group as compared to 83.1% of intranasal group. The lower limit of 95% confidence interval for effect size was approximately −9.7%, with an a priori cutoff for noninferiority of −10%.

Significance:  Intranasal administration of lorazepam is not found to be inferior to intravenous administration for termination of acute convulsive seizures in children.

Intravenous (IV) lorazepam (LOR), or IVLOR, is considered the drug of choice for control of acute convulsive seizures. Its efficacy and safety in different settings is supported by available evidence. In the Veterans Affairs study, LOR was significantly superior to phenytoin in a pair-wise comparison (p = 0.002) for control of overt generalized convulsive status epilepticus (SE). No significant differences among the treatments were detected for subtle SE (Treiman et al., 1998). Efficacy of LOR has also been demonstrated in out-of-hospital setting in adults. It had been found that SE was controlled on arrival at the hospital in significantly more patients receiving out-of-hospital benzodiazepines than placebo, with LOR better than diazepam (Alldredge et al., 2001). In childhood SE (n = 31, age 2–18 years) it was found that seizures stopped within 10 min in 60% of patients receiving LOR and within 20 min in 84% (Lacey et al., 1986). The median latency period was 10 min, and the duration of control was at least 3–6 h in 83% and up to 24 h in ∼50% patients. In addition, experience with 300 consecutive parenteral doses of LOR for SE or serial seizures in 77 children and young adults (age 5 week to 25 years) found that LOR stopped the SE in 79% and diminished the intensity of SE in an additional 4% (Crawford et al., 1987). In view of accumulating evidence, the British Pediatric Neurology Association published a guideline regarding use of LOR as first-line agent for SE in children (Appleton et al., 2000). An audit performed after policy change showed LOR to be as efficacious and safe as diazepam (Qureshi et al., 2002).

However, achieving peripheral venous access in a convulsing child is difficult. This difficulty may be increased by resource constraints, lack of trained personnel, and out-of-hospital setting. Hence, alternate routes of administration continue to be studied for administration of benzodiazepines. In a study on intranasal (IN)LOR in healthy adult volunteers, it was found that absorption was faster than intramuscular route and elimination was comparable to other parenteral routes (Wermeling et al., 2001). The only randomized clinical trial using INLOR has been reported by Ahmad et al. (2006) from Malawi, who compared it with their standard of care: intramuscular paraldehyde. It was an open randomized trial in which 160 children aged 2 months to 12 years with seizures persisting for more than 5 min were randomly assigned to receive either INLOR (100 μg/kg, n = 80) or intramuscular paraldehyde (0.2 ml/kg, n = 80). The primary outcome measure was cessation of presenting seizure with one dose of assigned anticonvulsant agent within 10 min of administration. INLOR stopped convulsions within 10 min in 60 episodes treated [75%; absolute risk 0.75, 95% confidence interval (CI) 0.64–0.84], and intramuscular paraldehyde in 49 (61.3%; absolute risk 0.61, 95% CI 0.49–0.72). Seizure recurrence within 24 h was also less in patients who received LOR (10% vs. 14%). None of these differences attained statistical significance. Median time for seizure cessation was comparable in both the groups: 7.5 min for LOR and 8 min for paraldehyde (p = 0.68). No clinically significant cardiorespiratory events were seen in either group (Ahmad et al., 2006).

Hence, there is a need to compare INLOR with the more widely accepted standard IVLOR for control of seizures. This study was conducted to compare IN versus IV administration of LOR in children aged 6–14 years presenting with acute seizure regarding efficacy and adverse-event profile.

Methods

This was a randomized open-label noninferiority study that included children from the pediatric emergency service of a tertiary-level Indian teaching hospital from August 2008 to April 2009.

Participants

All children aged 6–14 years who presented convulsing to the emergency room (ER) or developed a seizure during an ER stay were eligible for inclusion. Exclusion criteria were known hypersensitivity to any benzodiazepine, child having received any parenteral antiepileptic drug (AED) within 1 h prior to enrollment, presence of severe cardiorespiratory compromise, presence of cerebrospinal fluid rhinorrhea, and upper respiratory tract infection sufficiently severe to preclude IN administration (Fig. 1).

Figure 1.   Flow of patients in the study.

Interventions

Initial priority was given to stabilization of the airway, breathing, and circulation. While the child was being assessed and medications drawn up, the guardian was informed of the procedure, and as detailed a description as feasible of the trial was provided in the given circumstances and verbal consent was obtained. Informed consent after a fuller explanation was undertaken with the guardians as soon as the child was stable. All guardians gave written informed consent for their children to participate. Randomization was done using blocks of variable length. Opaque, sealed envelopes containing allocation of consecutive patients were kept in the ER.

Patients were randomized to receive either INLOR or IVLOR, 0.1 mg/kg (0.05 ml/kg) to a maximum of 4 mg. In either case, 0.05 ml/kg was drawn up in a syringe. If the patient was randomized to receive INLOR, it was directly instilled into any one nostril, with the patient in a supine position, drop by drop over 30–60 s. If the patient was to receive IVLOR, attempt was made to achieve peripheral venous access as soon as possible and LOR was given as rapid push. The formulation of LOR used for both IN and IV arms contained lorazepam British Pharmacopeia (BP) 2 mg/ml and benzyl alcohol 2% v/v. This formulation was available in the pediatric emergency service on a routine basis and was uniformly obtained from a single pharmaceutical company.

Patients were attached to a monitor for continued measurements of heart rate (HR), oxygen saturation (SpO₂) and noninvasive blood pressure (NIBP). A chart was established for periodic documentation of respiratory rate, any respiratory distress, and any seizure recurrence. These values were noted every 5 min for first 20 min, and then every 10 min for a total of 60 min. The time of drug administration was taken as 0 min and all measurements made with that reference.

Patients were assessed for persistence of clinically visible motor seizure activity at the end of 10 min, if seizures were persisting or recurred after an initial period of remission, Phenytoin was given at a dose of 20 mg/kg diluted in normal saline at a rate not exceeding 1 mg/kg/min. If there was no seizure recurrence or persistence, it was taken as a positive primary outcome and patients were observed for 1 h and then given maintenance AED and managed as usual.

Because the study involved recruitment of patients from emergency service, casualty medical officers (CMOs) were regularly apprised of the protocol every week during study period. The monitoring and drug administration were initiated by them and taken over by the principal investigator (PI) as soon as feasible. All time points were measured using a single stop watch installed in ER. At the protocol stage, it was clarified that if a patient randomized to receive IV LOR could not have IV access obtained within 10 min, then such a patient would be managed on an individual basis as per discretion of the treating CMO in consonance with PI. Such patients would be included in the analysis.

Outcomes

Primary outcome measure was cessation of all clinically visible seizure activity within 10 min of drug administration. Secondary outcomes included persistent cessation of seizure activity for 1 h, time to achieve IV access after arrival in the ER, time from drug administration to termination of seizure(s), development of hypotension (defined as fall of ≥20 mm Hg systolic and/or ≥10 mm Hg diastolic pressure) within 1 h of drug administration; and development of significant respiratory depression requiring assisted ventilation.

Sample size and statistical methods

As shown from previous studies, IVLOR has an approximate efficacy of 80% (Crawford et al., 1987). Expected difference between test and control group was set at 10%. Using an allocation ratio of 1, alpha error to 0.05, and power of the study as 80%, “N-QUERY” software returned a sample size estimate of 70 patients for each arm.

Data was entered into a MS Excel spread sheet (Microsoft Corp., Redmond, WA, U.S.A.). Proportions were compared using the chi square test. Medians were compared using the Mann-Whitney test. SPSS version 10 (SPSS Inc, Chicago, IL, U.S.A.) was used for calculations.

The study was approved by the institutional ethics committee. Permission was obtained from Drug Controller General of India for use of INLOR in the preceding age group. The study was registered with Clinical Trials Registry of India (CTRI/2008/091/000135) and clinicaltrials.gov (NCT00735527). The World Health Organization (WHO) Universal Trial Reference Number is 104817617-14082008229360.

Results

One hundred forty-one children were randomized, 70 to the IV and 71 to the IN arm, respectively (Fig. 1). The groups were comparable at baseline with respect to relevant variables (Table 1).

Table 1.   Baseline comparisons for demographic and pertinent clinical variables between intravenous and intranasal groups

Variable	Intravenous	Intranasal	p-Value
GTCS, generalized tonic–clonic seizure.
No. enrolled	70	71
Age, years (mean ± SD)	8.63 ± 2.18	8.97 ± 2.68	0.410
Gender (M, F)	36, 34	41, 30	0.451
Duration of current seizure (%)
<10 min	42 (60)	47 (66.2)	0.446
≥10 min	28 (40)	24 (33.8)
Semiology (%)
GTCS	35 (50)	23 (32.39)	0.095
Partial	33 (47.14)	44 (61.97)
Others/unclear	2 (2.86)	4 (5.63)
Prior chronic AED use (%)	28 (40)	26 (36.62)	0.680

The primary outcome measure was cessation of all clinically visible seizure activity within 10 min of drug administration. IVLOR attained such seizure control in 56 (80%) of 70 patients, whereas INLOR administration in 59 (83.1%) of 71 [relative risk (RR) 0.963, 95% CI 0.823–1.127, p = 0.635]. The lower bound 95% CI for effect size was −9.698 with an a priori cut off of −10% for this noninferiority trial. Furthermore, IVLOR maintained seizure remission for the trial period of 1 h in 41 patients (58.57%) compared to 44 patients receiving INLOR (61.97%) (RR=0.945, 95% CI 0.723–1.236, p = 0.680) (Table 2).

Table 2.   Selected primary and secondary efficacy outcomes in patients assigned to receive intravenous and intranasal lorazepam

Outcome measure	Intravenous (n = 70)	Intranasal (n = 71)	Effect size % (95% CI)	RR (95% CI)	p-Value
Clinical seizure remission within 10 min of drug administration	56 (80.00%)	59 (83.10%)	3.1 (−9.698 to 15.898)	0.963 (0.823–1.127)	0.635
Persistent cessation of seizure activity for 1 h (did not require rescue medication)	41 (58.57%)	44 (61.97%)	3.4 (−12.746 to 19.546)	0.945 (0.723–1.236)	0.680
Time from drug administration to termination of seizure(s) (min) median, (IQR)	3 (4.25)	3 (3.00)			0.900

The time taken for clinical seizure remission after drug administration varied from 1–36 min (median 3 min) for IVLOR and 1–25 min (median 3 min) for INLOR (p = 0.998) (Table 2). None of the trial patients in either group developed significant hypotension. Two patients in the IV group and one patient in the IN group required assisted ventilation (p = 0.989). The median time to achieve peripheral IV access in the ER in a convulsing child was found to be 4 min (range 1–25 min).

IV and IN administration of LOR for seizure control was further analyzed depending on seizure semiology and prior AED use. Of the patients randomized to IV and IN arms, 35 and 23, respectively, were having generalized seizures. In these patients, IVLOR and INLOR achieved seizure remission within 10 min in about 74% and 70%, respectively (RR 1.068, 95% CI 0.765–1.490, p = 0.694) and maintained seizure control for 1 h in 43% and 61%, respectively (RR 0.724, 95% CI 0.425–1.165, p = 0.180). Median time taken from drug administration to seizure cessation was 4 min in both the arms. Similarly, in the IV and IN arms, respectively, 33 and 44 patients were found to be having partial seizures. All patients with partial seizures had focal motor seizures. In these patients, seizure cessation was achieved within 10 min in 91% patients in both the IV and IN arms. IVLOR and INLOR maintained seizure remission for 1 h in 74% and 66% of patients, respectively (RR 1.195, 95% CI 0.907–1.576, p = 0.216). Median time to achieve seizure cessation was 3 min in both groups.

In patients receiving AED chronically on an outpatient basis, seizure remission within 10 min of drug administration was achieved in 75% and 88% patients, respectively, receiving IV (n = 28) and IN (n = 26) LOR (RR 0.848, 95% CI 0.657–1.094, p = 0.203). Seizure remission was maintained for 1 h in 50% and 54% patients, respectively, in the IV and IN arms (RR 0.929, 95% CI 0.556–1.552, p = 0.777). In both groups, median time to achieve seizure control was found to be 3 min. Efficacy of INLOR was further compared in patients with (n = 26) and without (n = 45) prior chronic AED use. Seizure remission within 10 min was attained in 88% and 84% patients, respectively, of those receiving and not receiving AED on a chronic basis (p = 0.691) and maintained for 1 h in 54% and 62%, respectively (p = 0.954).

There was one protocol violation in the IV arm (failure to establish IV access within 10 min) and none in the IN arm. Outcome measures were applied to all patients. Fidelity of measurements was checked by comparing the number of actual observations against as planned. All clinical outcome measures (observation for seizure activity at 10 and 60 min, time from drug administration to seizure remission, time to achieve IV access) were recorded in all patients. A total of 677 BP measurements (mean 4.8 per patient) could be performed against a promise of 1,269 (9 per patient). Observations for respiratory rate and distress could be made 996 times (7 per patient) of 1,269 (9 per patient).

Discussion

This study supports the hypothesis that INLOR is an acceptable alternative to conventional IVLOR for control of seizures in children with respect to efficacy and safety. To the best of our knowledge this is the first study comparing INLOR to the standard IV route.

Primary efficacy outcome

INLOR achieved clinical seizure remission within 10 min of administration in 83% of patients versus 80% in the IV group. Hence, with respect to seizure control in children aged 6–14 years, INLOR is not inferior to the current standard of practice IVLOR. The INLOR arm of the study by Ahmad et al. (2006) included 80 children aged 2 months to 12 years. The presenting seizure stopped within 10 min of INLOR administration in 75% of their patients, which is comparable to our figure of 83%. LOR in childhood SE is most efficacious in the 5–11 year age group, with 92% having successful seizure control (Crawford et al., 1987). Our study population corresponded to this age group, whereas those of Ahmad et al. (2006) also included infants and young children, in whom the efficacy of LOR is relatively modest (72–76% successful seizure control; Crawford et al., 1987). Another contributor to this difference could be etiologic profile of seizures. In sub-Saharan Africa the most common cause of SE was identified as cerebral malaria (Ahmad et al., 2006), whereas, in our study, it was neurocysticercosis (49%, data not presented).

This study has used the time of drug administration as reference time or 0 min. This is primarily because, whereas it was expected that establishing an IV access would take a few min, IN administration was planned to be done in <1 min. Hence, if other alternatives for reference time were used, such as decision to treat, the results would have been artificially in favor of IN administration.

Secondary efficacy outcomes

In the present study, continued clinical seizure cessation for 1 h was observed in 62% in the IN arm versus 59% in the IV arm, implying that INLOR is not inferior to IVLOR regarding duration of action. Prolonged duration of action for IVLOR is well established (Schmidt, 2002); however, there is some evidence for similar long duration of action of INLOR also. In a pharmacokinetic study of healthy adult volunteers, INLOR showed a second delayed peak attributed putatively to inadvertent oral absorption, which likely contributed to prolonged efficacy (Wermeling et al., 2001). In the experience of Ahmad et al. (2006), 10% of the patients in the INLOR arm (n = 80) failed to achieve seizure remission within 1 h.

Median time to achieve seizure remission in both the INLOR and IVLOR arms was 3 min. Ahmad et al. (2006) found median time for seizure cessation of 7.5 min[(interquartile range (IQR) 4.5–11.5 min]. This is difficult to explain, but probably results from differences in age and etiologic profile of our patients and lack of use of the atomizer device by us. However, our experience agrees with data from healthy adults for which peak bioavailability was achieved within 5 min (Wermeling et al., 2001).

Adverse drug effects

No patients in the present study developed significant hypotension within 1 h of drug administration. Ahmad et al. (2006) have also found a median decrease in systolic pressure of 7 mm Hg and diastolic pressure of 7.5 mm Hg, with no requirement for intervention. One patient in the present study from the IN arm required intubation and assisted ventilation. This patient had convulsive SE of >30 min prior to reporting to the ER and was hypoxic and had suboptimal respiratory status at presentation. Therefore, it is difficult to attribute respiratory compromise entirely to INLOR administration. In the IV arm, two patients required assisted ventilation. In the study by Ahmad et al. (2006), two children receiving INLOR had SpO₂ of 92% for 10 min, but responded to oxygen supplementation. Because our sample size was insufficient to capture enough adverse effects, this study was underpowered to conclude meaningfully with respect to the difference in the incidence of adverse effects between the IN and IV arms.

Time to achieve peripheral venous access

It took a median of 4 min (range 1–25 min) to achieve peripheral venous access. This shows that even in a well-equipped, tertiary-level ER with skilled residents, valuable time of up to 25 min may be lost in struggling for peripheral venous access in a convulsing child. Intranasal administration, however, is virtually instantaneous and does not require any special skill.

Effect of seizure semiology

INLOR achieved seizure control within 10 min of administration in 78% of patients with generalized and 82% of those with partial seizures. Continued seizure cessation for 1 h was seen in 57% of patients with generalized and 66% of those with partial seizures, who received INLOR. There are no comparable data in literature regarding efficacy of INLOR in different seizure types. Although our study was not sufficiently powered to establish equivalence of IN and IV routes for administration of LOR in these subgroups, it provides preliminary support to the hypothesis that INLOR may be an acceptable alternative to IVLOR for different seizure types.

Prior chronic AED use

There is a justifiable concern that prior AED use might influence the efficacy of benzodiazepines for seizure control. In patients on prior AEDs, IVLOR achieved seizure cessation within 10 min in 75% and maintained it for 1 h in 54% as compared to INLOR, which achieved seizure cessation within 10 min in 81% and maintained remission in 62%. This finding supports pharmacokinetic data showing comparable bioavailability with IV and IN administration of LOR (Wermeling et al., 2001). A more relevant question is whether the efficacy of INLOR differs in patients on or not on chronic AEDs. Seizure cessation with INLOR was observed in 81% of patients on AED as compared to 84% of those not on them. Although sufficient power is lacking to draw valid conclusions, these suggest that INLOR may be an acceptable alternative to IVLOR in these patients.

Limitations

The present study has several inadvertent limitations. Video electroencephalography (EEG) would have been ideal for monitoring of response to LOR administration. It would have allowed the use of clinical as well as electroencephalographic seizure remission as outcome end points. However, video-EEG is not available with pediatric neurology service at our institute. An atomizer device was not used for instillation of INLOR, again because of nonavailability. Sample size estimation was done keeping in view the primary efficacy outcome, and this study is underpowered for other outcomes or for subgroup analysis. This is a feasibility issue, as including a sufficient number of patients to draw meaningful conclusions about adverse drug effects or subgroups would have greatly increased the sample size.

Conclusion

INLOR (0.1 mg/kg, maximum 4 mg) is not inferior to IVLOR for termination of seizures in children aged 6–14 years. In addition, there is preliminary suggestion that INLOR may not be inferior to IVLOR for continued seizure remission for 1 h, time to achieve seizure cessation, and the incidence of clinically significant cardiorespiratory adverse events. In cognizance of the above limitations, the results of present study need to be interpreted with caution. There is a need to compare INLOR against IVLOR in an out-of-hospital setting and against other benzodiazepines and routes of administration.

Disclosure

None of the authors has any conflict of interest to express. The trial medication was available in the pediatric emergency service of our institution, procured as per routine procedures. None of the authors had any financial transactions with the pharmaceutical company providing this medication.

We confirm that we have read the Journal’s position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.

Authors’ Contributions

Ravindra Arya: principal investigator, design, conduct and analysis of the study, and drafting of the manuscript; Sheffali Gulati and Madhulika Kabra: critical review of the manuscript; Jitendra K. Sahu: inputs to design of the study and reviewing the manuscript; Veena Kalra: design of the study, critical review of manuscript.

Role of Funding Agency

No funding agency was involved. Trial medication was available in the pediatric emergency of our hospital on routine supply.

Ethics Committee Approval

The study was approved by institutional ethics committee via letter no. Ref A-113/24.09.2007; dated 02.07.2008. Permission was obtained from Drug Controller General of India for use of INLOR in the above age group. The study was registered with Clinical Trials Registry of India (CTRI/2008/091/000135) and clinicaltrials.gov (NCT00735527). World Health Organization (WHO) Universal Trial Reference Number is 104817617-14082008229360.