YEBEH-04248; No of Pages 12

Epilepsy & Behavior xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Epilepsy & Behavior

journal homepage: www.elsevier.com/locate/yebeh

Review

Nonintravenous midazolam versus intravenous or rectal diazepam

for the treatment of early status epilepticus: A systematic review
with meta-analysis
Francesco Brigo a,b,⁎, Raffaele Nardone b,c, Frediano Tezzon b, Eugen Trinka c,d,e
a
Department of Neurological and Movement Sciences, University of Verona, Italy
b
Department of Neurology, Franz Tappeiner Hospital, Merano, Italy
c
Department of Neurology, Christian Doppler Klinik, Paracelsus Medical University, Salzburg, Austria
d
Centre for Cognitive Neuroscience, Salzburg, Austria
e
Department of Public Health Technology Assessment, UMIT — University for Health Sciences, Medical Informatics and Technology, Hall.i.T., Austria

a r t i c l e i n f o a b s t r a c t

Article history: Background: Prompt treatment of status epilepticus (SE) is associated with better outcomes. Rectal diazepam
Revised 23 February 2015 (DZP) and nonintravenous (non-IV) midazolam (MDZ) are often used in the treatment of early SE instead of
Accepted 24 February 2015 intravenous applications. The aim of this review was to determine if nonintravenous MDZ is as effective and
Available online xxxx safe as intravenous or rectal DZP in terminating early SE seizures in children and adults.
Methods: We searched the Cochrane Central Register of Controlled Trials (CENTRAL), ClinicalTrials.gov, and
Keywords:
MEDLINE for randomized controlled trials comparing non-IV MDZ with DZP (by any route) in patients (all
Diazepam
Meta-analysis
ages) with early SE defined either as seizures lasting N5 min or as seizures at arrival in the emergency depart-
Midazolam ment. The following outcomes were assessed: clinical seizure cessation within 15 min of drug administration,
Seizures serious adverse effects, time interval to drug administration, and time from arrival in the emergency department
Status epilepticus to seizure cessation. Outcomes were assessed using a random-effects Mantel–Haenszel meta-analysis to calcu-
late risk ratio (RR), odds ratio (OR) and mean difference with 95% confidence intervals (95% CIs).
Results: Nineteen studies with 1933 seizures in 1602 patients (some trials included patients with more than one
seizure) were included. One thousand five hundred seventy-three patients were younger than 16 years. For sei-
zure cessation, non-IV MDZ was as effective as DZP (any route) (1933 seizures; RR: 1.03; 95% CIs: 0.98 to 1.08).
No difference in adverse effects was found between non-IM MDZ and DZP by any route (1933 seizures; RR: 0.87;
95% CIs: 0.50 to 1.50). Time interval between arrival and seizure cessation was significantly shorter with non-IV
MDZ by any route than with DZP by any route (338 seizures; mean difference: −3.67 min; 95% CIs: −5.98 to
−1.36); a similar result was found for time from arrival to drug administration (348 seizures; mean difference:
−3.56 min; 95% CIs: −5.00 to −2.11). A minimal difference was found for time interval from drug administra-
tion to clinical seizure cessation, which was shorter for DZP by any route than for non-IV MDZ by any route
(812 seizures; mean difference: 0.56 min; 95% CIs: 0.15 to 0.98 min). Not all studies reported information on
time intervals. Comparison by each way of administration failed to find a significant difference in terms of clinical
seizure cessation and occurrence of adverse effects. The only exception was the comparison between buccal MDZ
and rectal DZP, where MDZ was more effective than rectal DZP in terminating SE but only when results were
expressed as OR (769 seizures; OR: 1.78; 95% CIs: 1.11 to 2.85; RR: 1.15; 95% CIs: 0.85 to 1.54). Only one study
was entirely conducted in an adult population (21 patients, aged 31 to 69 years), showing no difference in effi-
cacy or time to seizure cessation after drug administration between intranasal MDZ and rectal DZP.
Conclusions: Non-IV MDZ is as effective and safe as intravenous or rectal DZP in terminating early SE in children
and probably also in adults. Times from arrival in the emergency department to drug administration and to
seizure cessation are shorter with non-IV MDZ than with intravenous or rectal DZP, but this does not necessarily
result in higher seizure control. An exception may be the buccal MDZ, which, besides being socially more accept-
able and easier to administer, might also have a higher efficacy than rectal DZP in seizure control.

This article is part of a Special Issue entitled Status Epilepticus.

© 2015 Elsevier Inc. All rights reserved.

⁎ Corresponding author at: Department of Neurological and Movement Sciences, University of Verona, Piazzale L.A. Scuro, 10, 37134 Verona, Italy. Tel.: +39 0458124174; fax: +39
0458124873.
E-mail address: dr.francescobrigo@gmail.com (F. Brigo).

http://dx.doi.org/10.1016/j.yebeh.2015.02.030
1525-5050/© 2015 Elsevier Inc. All rights reserved.

Please cite this article as: Brigo F, et al, Nonintravenous midazolam versus intravenous or rectal diazepam for the treatment of early status
epilepticus: A systematic review with meta-analysis, Epilepsy Behav (2015), http://dx.doi.org/10.1016/j.yebeh.2015.02.030
2 F. Brigo et al. / Epilepsy & Behavior xxx (2015) xxx–xxx
1. Introduction
Status epilepticus (SE) can be regarded as the most extreme and se-
vere form of seizure activity, being associated with high morbidity and
mortality [1]. In clinical practice, SE has been traditionally defined as
epileptic activity persisting for more than 30 min or as two or more se-
quential seizures without full interictal recovery [2]. However, over the
years, this timeframe has been progressively shortened to the pragmatic
definition of 5 min because of the seriousness of the condition and the
urge to treat it as early as possible [3]. Its prompt treatment can prevent
death or irreversible brain damage. In fact, early treatment is associated
with lower morbidity and mortality, fewer drugs required in hospitals,
and shorter overall seizure duration [4,5].
There are, however, several factors, including education regarding
seizure emergencies and transferring of patients to the hospital, that
may hinder prompt treatment, resulting in a significant treatment
delay. Hence, prehospital management of SE might be beneficial pro-
vided that administered drugs are effective in terminating seizures,
safe, and easy to use.
Diazepam (DZP) and midazolam (MDZ) are commonly used in
the treatment of early (stage I) SE. Midazolam is a water-soluble benzo-
diazepine, which may be administered by different routes: intravenous,
intramuscular, buccal, and intranasal. Conversely, DZP can be admin-
istered either intravenously or per rectum. Rectal DZP is the most
common drug used in the prehospital management of early SE in
Spain [6] and possibly also in other countries. However, its administra-
tion is most of the time socially unacceptable. Furthermore, its adminis-
tration requires the removal of clothes and positioning the patient
appropriately, which may result in relevant treatment delay. The same
limitation holds true for intravenous administration of DZP or other
drugs such as lorazepam, which requires the placement of an intrave-
nous access.
Hence, MDZ, which can be administered by different and more
practical routes (buccal, intranasal, intramuscular), has emerged as an
alternative to drugs administered by intravenous or rectal route, such
as lorazepam or DZP [7,8].
The aim of this systematic review was to determine if nonintravenous
(non-IV) MDZ is as effective and safe as intravenous or rectal DZP in ter-
minating early SE in children and adults. Furthermore, we aimed to
evaluate whether non-IV MDZ administration is faster than intravenous
or rectal DZP administration and, if so, whether this “time gain” results in
higher seizure control.
2. Methods
This review was guided by a written prespecified protocol de-
scribing research questions, review methods, and a plan for data extrac-
tion and synthesis. The protocol is available at: http://www.crd.york.ac.
uk/PROSPERO/DisplayPDF.php?ID=CRD42015016179.
2.1. Criteria for considering studies for this review
We included randomized controlled trials (RCTs), blinded or un-
blinded. Uncontrolled and nonrandomized trials were excluded. We in-
cluded patients of any age diagnosed with early (stage I) SE defined
either as seizures lasting N 5 min [3] or as seizures at arrival to the emer-
gency department.
We considered all trials in which non-IV MDZ used as a first-line
agent in monotherapy was compared with DZP (first-line drug given
singly) by any route. The following outcomes were considered:
Efficacy
▪ The number of status epilepticus episodes which were termi-
nated within 15 min after drug (MDZ or DZP) administration
or before emergency medical service support arrived (only for
studies conducted in prehospital settings);
Please cite this article as: Brigo F, et al, Nonintravenous midazolam versus intravenous or rectal diazepam for the treatment of early status
epilepticus: A systematic review with meta-analysis, Epilepsy Behav (2015), http://dx.doi.org/10.1016/j.yebeh.2015.02.030
▪ Time from arrival at the emergency department to drug admin-
istration (or time from seizure initiation to drug administration
for studies conducted in prehospital settings);
▪ Time from drug administration to clinical seizure cessation; and
▪ Time from arrival at the emergency department to clinical sei-
zure cessation (or time from seizure initiation to clinical seizure
cessation for studies conducted in prehospital settings).
Tolerability and safety
▪ The number of patients experiencing serious adverse effects
(respiratory depression or hypotension).
2.4. Search methods for identification of studies
A comprehensive review of the literature of computerized databases
as well as searches to find unpublished trials were performed to mini-
mize publication bias. The following electronic databases and data
sources were searched:
1. MEDLINE (January 1966–20th of January 2015), accessed through
PubMed;
2. Cochrane Central Register of Controlled Trials (CENTRAL) (Issue 12,
The Cochrane Library, December 2014) (accessed 20th of January
2015); the following search strategy was adopted: ((“Status
Epilepticus”[Mesh] OR “status epilepticus” OR seizur*) AND midazo-
lam). All resulting titles and abstracts were evaluated, and any rele-
vant article was considered. There were no language restrictions;
3. ClinicalTrials.gov (available at: https://clinicaltrials.gov/; accessed
20th of January 2015); the following search strategy for this database
was adopted: ((“Status Epilepticus” OR seizure OR seizures) AND
midazolam). There were no language restrictions;
4. Handsearching of the references quoted in the identified trials;
5. Contact with pharmaceutical companies (Viropharma and Accord
Healthcare) to identify unpublished trials or data missing from arti-
cles (January 2015); and
6. Contact with authors and known experts to identify any additional
data.
3. Data collection and analysis
3.1. Study selection
Retrieved articles were independently assessed for inclusion by two
review authors; any disagreement was resolved through discussion.
3.2. Quality assessment
Trials were scrutinized, and the methodological quality of all in-
cluded studies was evaluated. Quality assessment included the fol-
lowing aspects of methodology: study design, definition and clinical
relevance of outcomes, type of control, method of allocation conceal-
ment, total study duration, completeness of follow-up, intention-to-
treat analysis, data concerning adverse effects, risk of bias, and conflict
of interests. The randomized trials were judged on the reported method
of allocation concealment and on the risk of bias as outlined in the
Cochrane Handbook for Systematic Reviews of Interventions Version
5.1.0 [updated March 2011] [9]. We also evaluated whether authors
disclosed their conflict of interest and whether pharmaceutical compa-
nies sponsored the studies.
3.3. Data extraction
The following trial data were extracted: main study author and age
of publication; country; definition of SE applied in the study; type of
participants (children and/or adults); total number, age, and sex of
 F. Brigo et al. / Epilepsy & Behavior xxx (2015) xxx–xxx
participants for each treatment group; seizure type; intervention details
(dose, route of administration); definition of successful treatment
adopted in each trial; proportion of seizures controlled after drug ad-
ministration in each treatment group; time from arrival at the ED to
drug administration; time from drug administration to clinical seizure
cessation; time from arrival at the ED to clinical seizure cessation; and
proportion of serious adverse effects (respiratory depression and/or
hypotension) in each group.
3.4. Data analysis
We sought data on the number of participants in the treatment
groups and with each outcome in the articles.
Provided that we thought that it was clinically appropriate and no
important clinical and methodological heterogeneity was found, we
planned to summarize results in a meta-analysis.
The trials comparing the same drugs were combined. Pooled risk ra-
tios were determined using Mantel–Haenszel random-effects models.
Data were stratified into subgroups comparing the following:
1. Non-IV MDZ by any route versus DZP by any route;
2. Intranasal MDZ versus DZP by any route;
3. Buccal MDZ versus DZP by any route;
4. Intramuscular MDZ versus DZP by any route;
5. Intranasal MDZ versus intravenous DZP;
6. Intranasal MDZ versus rectal DZP;
7. Buccal MDZ versus intravenous DZP;
8. Buccal MDZ versus rectal DZP;
9. Intramuscular MDZ versus intravenous DZP; and
10. Intramuscular DZP versus rectal DZP.
Where study data were available, we assessed the mean differences
in times between arrival at the ED and drug administration (non-IV
MDZ by any route versus DZP by any route) or clinical seizure cessation
and between drug administration and cessation of seizure activity.
Dichotomous outcomes (clinical seizure cessation and occurrence of
serious adverse effects) were analyzed by calculating risk ratio (RR) for
each trial, with the uncertainty in each trial being expressed using 95%
confidence intervals (CIs). For each outcome, a weighted treatment ef-
fect across trials was calculated. Odds ratio (OR) were also calculated.
To evaluate consistency across study results, results expressed as OR
were compared with results expressed as RR.
Continuous data (time intervals from arrival to the emergency
department to drug administration or clinical seizure cessation) were
analyzed by calculating the mean difference for each trial, with the un-
certainty in each study being expressed using 95% CIs.
Homogeneity among study results was evaluated using a standard
chi-squared test, combined with the I2 statistics, and the hypothesis
of homogeneity was rejected if the p-value was less than 0.10. The
interpretation of I2 for heterogeneity was made as follows: 0–25%
represents low heterogeneity, 25–50% moderate heterogeneity, 50–75%
substantial heterogeneity, and 75–100% high heterogeneity [10]. Trial
outcomes were combined to obtain a summary estimate of effect (and
the corresponding CI) using a random-effects model. A random-effects
model is considered more conservative than a fixed-effect model since
it takes into account the variability between studies, thus leading to
wider CIs.
Statistical analyses were undertaken with the Review Manager soft-
ware developed by the Cochrane Collaboration (5.3).
4. Results
4.1. Description of studies (Tables 1 and 2)
The search strategy described earlier yielded 771 results (671
MEDLINE, 84 CENTRAL, and 16 ClinicalTrials.gov). The pharmaceutical
companies Viropharma and Accord Healthcare were contacted (January
Please cite this article as: Brigo F, et al, Nonintravenous midazolam versus intravenous or rectal diazepam for the treatment of early status
epilepticus: A systematic review with meta-analysis, Epilepsy Behav (2015), http://dx.doi.org/10.1016/j.yebeh.2015.02.030
3
2015), but no additional unpublished trials were found. After excluding
duplicate publications (38), case reports and case series, letters, re-
views, and uncontrolled and nonrandomized trials, 30 studies potential-
ly includible were provisionally selected. No additional unpublished
study was identified. We excluded 11 studies after reading the full pub-
lished papers (a list of these articles and reasons for exclusion are re-
ported in Supplementary material 1). Thus, nineteen trials [11–29]
contributed to this review: the earliest was published in 1997 and the
most recent in 2014 (Fig. 1).
All studies were conducted in children except one which included
both children and young adults [12] and one which included only adults
[23]. All studies were conducted in hospital settings (emergency depart-
ments) except for one RCT where caregivers administered MDZ or DZP
at home if seizures lasted more than 5 min [24].
All studies had two arms and used DZP (intravenous or rectal) as the
active comparator. One study adopted a cross-over design [23]. Six dif-
ferent comparisons were available (intranasal MDZ versus IV DZP, intra-
nasal MDZ versus rectal DZP, buccal MDZ versus intravenous DZP,
buccal MDZ versus rectal DZP, intramuscular MDZ versus intravenous
DZP, and intramuscular MDZ versus rectal DZP).
4.2. Risk of bias in included studies (Table 3)
All studies were described as RCTs. Two studies, however, involved a
systematic, nonrandom approach (sequence generation was generated
by a rule based on day of admission, which was not further explained)
[14,16]. Four studies used blockrandomization [20,24,26,29]. Other
studies used a random number table [13,15,17,21,22,25,27,28] or gener-
ated the sequence of randomization by shuffling envelopes [12,19].
Authors of three trials did not specify how random sequence was gener-
ated [11,18,23].
Allocation concealment was adequate in five RCTs [13,15,20,24,27]
and probably adequate in three studies [12,19,29] where it was not re-
ported whether an opaque envelope containing the name of the drug to
be administered was used. One study explicitly specified that allocation
was not concealed from attending staff [17]. The other studies did not
describe in sufficient detail the concealment of allocations prior to
assignment.
In five studies, investigators masked to therapeutic assignment
assessed the outcomes [13,15,20,24,27]. In one study [20], the study
team was not aware of which treatment a patient received, but they
were aware of the treatment code to which a patient was assigned,
potentially introducing bias. In one RCT, attending physicians, research
assistants, and patients/caregivers were blinded to the drug to be pre-
scribed until written consent was obtained [24]. In all other studies,
except six which were specified to be not blinded [22,24–26,28,29]
and one which was defined as single-masked without further specifica-
tions [19], blinding of participants and personnel was not explicitly
reported. No trial, except one [20], specified whether similar-looking
comparison drugs were used. However, the ‘hard’ outcomes chosen
in all studies are not probably to be influenced by lack of blinding. As
a consequence, all studies have a low risk of performance and detection
bias.
In ten studies, authors failed to disclose conflict of interests [11,12,
14–16,18,19,21,22,26], so it was not possible to evaluate a possible
sponsorship/funding by pharmaceutical companies. The other studies
explicitly reported the conflict of interest of authors involved, and
none of them was funded/sponsored by pharmaceutical companies.
4.3. Effects of interventions and meta-analysis outcomes: quantitative data
synthesis
Nineteen studies with 1933 seizures in 1602 patients (some trials in-
cluded patients with more than one episode) were included in the
quantitative data synthesis. Eighteen out of the 19 studies included in
this systematic review were conducted in children (one of these 18
4 F. Brigo et al. / Epilepsy & Behavior xxx (2015) xxx–xxx

Table 1
Characteristics of included studies.

Study Country Definition of status Participants Midazolam group Diazepam group Seizure type
epilepticus M/F ratio M/F ratio
Age Age

Intranasal midazolam versus intravenous diazepam

Lahat et al. [13] Israel Febrile seizures N 10 min Children 21 patients 23 patients Febrile convulsive (tonic, clonic,
M/F: 13/8 M/F: 12/11 tonic–clonic) seizures
16 (6–38) 18 (6–40)
(median (range), mo) (median (range), mo)
Mahmoudian Iran Seizures at arrival to the Children 70 patients (both groups) Generalized tonic–clonic seizures,
and Zadeh [15] emergency department M/F: NR simple partial seizures, complex partial
2 mo–15 ys (both groups, range) seizures, and myoclonic seizures
Javadzadeh Iran Seizures at arrival to the Children 30 patients 30 patients Simple febrile seizures, complex
et al. [26] emergency department M/F: 16/14 M/F: 17/13 febrile seizures, generalized
2.3[1.5] 2.5[1.2] tonic–clonic seizures, partial
(median [IQR], ys) (median [IQR], ys) seizures
with 2nd generalization,
Lennox–Gastaut syndrome,
infantile spasms, and undetermined
Thakker and India Seizures N 10 min Children 27 patients 23 patients Generalized tonic–clonic seizures,
Shanbag [27] M/F: 15/12 M/F: 12/11 simple partial seizures, complex
3.84 ± 2.93 3.97 ± 3.33 partial seizures, and subtle
(mean ± SD, ys) (mean ± SD, ys) convulsions

Intranasal midazolam versus rectal diazepam

Fişgin et al., Turkey Seizures at arrival to the Children 23 patients 22 patients Febrile and nonfebrile convulsions
2002 [14] emergency department M/F: 8/15 M/F: 11/11 (simple focal, after focal secondarily
3.80 (mean, ys) 2.02 (mean, ys) generalized, generalized
tonic–clonic, generalized tonic,
myoclonic)
Bhattacharyya India Seizures at arrival to the Children 23 patients 23 patients Simple partial seizures, generalized
et al. [19] emergency department M/F: 13/10 M/F: 16/7 tonic–clonic seizures, myoclonic
60.47 ± 45.35 74.53 ± 38.29 seizures, others (e.g., absence,
(mean ± SD, mo) (mean ± SD, mo) atonic seizures)
de Haan et al. The Seizures N 5 min Adults 21 patients (both groups) Tonic seizures, tonic–clonic seizures,
[23] Netherlands M/F: 8/13 (both groups) and complex partial seizures
40.3 ± 10.5 (31–69)
(both groups) (mean ± SD (range), ys)
Holsti et al. [24] USA Seizures N 5 min Children 50 patients 42 patients Seizures (any type)
M/F: 24/26 M/F: 22/30
5.6 [2.5–0.7] 6.9 [3.8–10.8]
(median [IQR], ys) (median [IQR], ys)

Buccal midazolam versus intravenous diazepam

Talukdar and India Seizures at arrival to the Children 60 patients 60 patients Partial and generalized tonic,
Chakrabarty emergency department M/F: 42/18 M/F: 40/20 clonic, and tonic–clonic seizures
[21] 30.4 ± 38.1 45.3 ± 45.4
(mean ± SD, mo) (mean ± SD, mo)
Tonekaboni Iran Seizures at arrival to the Children 32 patients 60 patients Tonic, tonic–clonic, and atonic
et al. [25] emergency department M/F: 14/18 M/F: 37/23 seizures
18.4 ± 10.3 17.1 ± 10.1
(mean ± SD, mo) (mean ± SD, mo)

Buccal midazolam versus rectal diazepam

Scott et al. [12] UK Seizures at arrival of Children/adults 14 patientsa 14 patientsa Tonic–clonic seizures, complex
paramedics M/F: N/R (both groups) partial seizures, absences,
14.5 ± 4.13 (5–19) (both groups) myoclonic seizures, atonic seizures,
(mean ± SD (range), ys) and tonic seizures
Baysun et al. [16] Turkey Febrile seizures at arrival Children 23 patients 20 patients Generalized tonic–clonic,
to the emergency M/F: 11/12 M/F: 11/9 generalized tonic, simple partial,
department 3.87 ± 3.39 (2 mo–12 ys) 2.85 ± 3.13 (4 mo–9 ys) and complex partial seizures
(mean ± SD (range), ys) (mean ± SD (range), ys)
McIntyre et al. United Seizures at arrival to the Children 92 patients 110 patients Convulsive seizures
[17] Kingdom emergency department M/F: 59/50 (no. of episodes) M/F: 64/46
2[1–5] (median [IQR], ys) (no. of episodes)
3 [1–6] (median [IQR], ys)
Mpimbaza et al. Uganda Seizures at arrival to the Children 165 patients 165 patients Generalized, focal, and febrile
[20] emergency department M/F: 84/81 M/F: 82/83 convulsions
or N5 min 17.0[10.5–30.0] 18.0[11.5–36.0]
(median [IQR], mo) (median [IQR], mo)
Ashrafi et al. [22] Iran Seizures at arrival to the Children 49 patients 49 patients Generalized tonic–clonic seizures,
emergency department M/F: 26/23 M/F: 32/17 myoclonic seizures, focal tonic
or N5 min 24 (median, mo) 48 (median, mo) seizures, and focal clonic seizures

Intramuscular midazolam versus intravenous diazepam

Chamberlain USA Seizures N 10 min Children 13 patients 11 patients Generalized tonic–clonic seizures
et al. [11] M/F: 8/5 M/F: 9/2 and focal motor seizures
42 (9–165) 39 (3–112)
(median (range), mo) (median (range), mo)

Please cite this article as: Brigo F, et al, Nonintravenous midazolam versus intravenous or rectal diazepam for the treatment of early status
epilepticus: A systematic review with meta-analysis, Epilepsy Behav (2015), http://dx.doi.org/10.1016/j.yebeh.2015.02.030
 F. Brigo et al. / Epilepsy & Behavior xxx (2015) xxx–xxx
Table 1 (continued)
Study Country Definition of status Participants
epilepticus
Intramuscular midazolam versus intravenous diazepam
Shah and India Seizures at arrival to the Children
Deshmukh emergency department
[18]
Portela et al. Brazil Seizures at arrival to the Children
[29] emergency department
Intramuscular midazolam versus rectal diazepam
Momen et al. Iran Convulsive generalized Children
[28] seizures at arrival to the
emergency department
(N5 min)
F: female, IQR: interquartile range, M: male, mo: months, NR: not explicitly reported, SD: standard deviation, and ys: years.
a
This study included 18 patients; some patients had more than one seizure and were assigned to both treatments (they are, therefore, represented in both groups).
studies included also young adults [12]) resulting in a total of 1573 pa-
tients younger than 16 years.
Details of each meta-analytic comparison are reported in Supple-
mentary material 2.
4.4. Non-IV MDZ versus DZP (all routes of administration)
For seizure cessation, non-IV MDZ was as effective as DZP (any
route) (RR: 1.03; 95% CIs: 0.98 to 1.08) (Fig. 2a). No difference in ad-
verse effects was found between non-IM MDZ and DZP by any route
(RR: 0.87; 95% CIs: 0.50 to 1.50) (Fig. 2b). After excluding the two
studies which included also adults [12,23] (sensitivity analysis), no dif-
ferences were found between non-IV MDZ and DZP by any route both
for seizure cessation (RR: 1.03; 95% CIs: 0.98 to 1.09) and for occurrence
of adverse effects (RR: 0.87; 95% CIs: 0.50 to 1.50).
Time interval between arrival and seizure cessation was signifi-
cantly shorter with non-IV MDZ by any route than with DZP by any
route (mean difference: − 3.67 min; 95% CIs: − 5.98 to − 1.36 min)
(Fig. 3c); a similar result was found for time from arrival to drug admin-
istration (mean difference: −3.56 min; 95% CIs: −5.00 to −2.11 min)
(Fig. 3a). A minimal difference was found for time interval from drug
administration to clinical seizure cessation, which was shorter for DZP
by any route than for non-IV MDZ by any route (mean difference:
0.56 min; 95% CIs: 0.15 to 0.98 min) (Fig. 3b); this statistically signifi-
cant difference has, most likely, too low a magnitude to be clinically
relevant.
Because of lack of data and differences in reporting time values
(mean, median, SD, range, interquartile range), meta-analytic compari-
sons evaluating time intervals between arrival and drug administration
or seizure cessation were not performed for each route of administra-
tion. However, we reported in Table 2 the time intervals and the degree
of statistical significance between the different groups from each study
reporting this information.
4.5. Intranasal, buccal, and muscular MDZ versus intravenous and rectal
DZP
Comparison by each route of administration (intranasal MDZ versus
DZP by any route, buccal MDZ versus DZP by any route, intramuscular
MDZ versus DZP by any route, intranasal MDZ versus intravenous
DZP, intranasal MDZ versus rectal DZP, buccal MDZ versus intravenous
DZP, buccal MDZ versus rectal DZP, intramuscular MDZ versus intrave-
nous DZP, and intramuscular MDZ versus rectal DZP) failed to find a sig-
nificant difference in terms of clinical seizure cessation and occurrence
Please cite this article as: Brigo F, et al, Nonintravenous midazolam versus intravenous or rectal diazepam for the treatment of early status
epilepticus: A systematic review with meta-analysis, Epilepsy Behav (2015), http://dx.doi.org/10.1016/j.yebeh.2015.02.030
5
Midazolam group Diazepam group Seizure type
M/F ratio M/F ratio
Age Age
115 patients (both groups) Generalized tonic–clonic seizures,
M/F: NR focal convulsions, tonic convulsions,
1 mo to 12 ys (both groups) and clonic convulsions
16 patients 16 patients Febrile and nonfebrile convulsive
M/F: 12/4 M/F: 8/8 seizures
46.4 ± 53.1 45.0 ± 49.2
(mean ± SD, mo) (mean ± SD, mo)
13.9 (10.7–70.0) 14.3 (4.4–90.8)
(median (range), mo) (median (range), mo)
50 patients 50 patients Generalized convulsive seizures
M/F: 31/19 M/F: 27/23
2 ± 1.1 (4 mo–15 ys) 2.5 ± 1.4
(mean ± SD (range), ys) (5 mo–13 ys)
(mean ± SD (range), ys)
of adverse effects. The only exception was the comparison between
buccal MDZ and rectal DZP, where MDZ was more effective than rectal
DZP in terminating SE but only when results were expressed as OR
(OR: 1.78; 95% CIs: 1.11 to 2.85; RR 1.15; 95% CIs 0.85 to 1.54) (Fig. 4).
The superiority of buccal MDZ over rectal DZP in terms of seizure control
was observed even after performing a sensitivity analysis by excluding
one study that also included adult patients [12] (OR: 1.67; 95% CIs:
0.89 to 3.14). However, this superiority was not evident with results
expressed as RR (RR 1.12; 95% CIs 0.81 to 1.56). Details of each meta-
analytic comparison are reported in Supplementary material 2.
5. Discussion
In this systematic review, we used systematic and explicit methods
to identify, select, and critically appraise studies and to extract and ana-
lyze data with a meta-analysis. A meta-analysis is the statistical combi-
nation of results from two or more separate studies (pairwise
comparisons of interventions), allowing an increase in statistical
power and an improvement in precision and sometimes permitting
questions to be answered that were not posed by individual studies
and to settle controversies arising from conflicting claims.
In the present meta-analysis, we found that non-IV MDZ is as effec-
tive and safe as intravenous or rectal DZP in terminating early SE. With
the exception of buccal MDZ, which was more effective than rectal DZP,
no other route of administration of MDZ was found to be more effective
than intravenous or rectal DZP but only when results are expressed as
OR. However, this absence of evidence of a difference in efficacy and
safety is not evidence that there is no difference [30] as it is probable
that for some comparisons, sample size is smaller than what would be
required to detect a clinically important benefit. Hence, it is likely that
some subanalyses lacked sufficient statistical power to detect a differ-
ence given the relatively small number of studies included in meta-
analytic comparisons.
The only differences in results expressed as OR and as RR were found
for the comparison between buccal MDZ and rectal DZP. Rectal MDZ
was found to be more effective than rectal DZP but only when results
were expressed as OR. This finding suggests that the difference in effica-
cy between the two drugs, although significant, might be small. The pos-
sible higher efficacy of buccal MDZ compared to rectal DZP may be
explained by the facts that transmucosal MDZ is rapidly effective [31,
32] and the sublingual absorption might be more complete and rapid
than after rectal administration [33], although no pharmacokinetics
study directly comparing the two drugs has been published so far. How-
ever, the mouth and the rectum have similar pH and surface areas rich

epilepticus: A systematic review with meta-analysis, Epilepsy Behav (2015), http://dx.doi.org/10.1016/j.yebeh.2015.02.030
Please cite this article as: Brigo F, et al, Nonintravenous midazolam versus intravenous or rectal diazepam for the treatment of early status
Table 2
Efficacy and safety outcomes in included studies.
Study
Intranasal midazolam versus intravenous diazepam
Lahat et al. [13]
Mahmoudian and
Zadeh [15]
Javadzadeh et al. [26]
Thakker and Shanbag [27]
Intranasal midazolam versus rectal diazepam
Fişgin et al. 2002
[14]
Bhattacharyya et al. [19]
de Haan et al. [23]
Holsti et al. [24]
Buccal midazolam versus intravenous diazepam
Talukdar and
Chakrabarty [21]
Tonekaboni et al. [25]
6
Interventions Definition of successful Clinical seizure Time from arrival Time from drug Time from arrival Serious adverse
treatment cessation in the emergency administration to in the emergency effects (hypotension
department to clinical seizure department to clinical or respiratory
drug administration cessation seizure cessation depression)
IN MDZ 0.2 mg/kg Seizure cessation within MDZ: 23/26a MDZ: 3.5 ± 1.8 MDZ: 3.1 ± 2.2 MDZ: 6.1 ± 3.6 MDZ: 0/26b
(max dose: 10 mg) vs. 10 min of drug administration DZP: 24/26a DZP: 5.5 ± 2.0 DZP: 2.5 ± 1.9 DZP: 8.0 ± 4.1 DZP: 0/26b
IV DZP 0.3 mg/kg (mean ± SD, min) (mean ± SD, min) (mean ± SD, min)
(max dose: 10 mg) p b 0.001 p b 0.001 p b 0.001
IN MDZ 0.2 mg/kg vs. Seizure cessation within MDZ: 35/35 MDZ: 3.58 ± 1.68c MDZ: 0/35
IV DZP 0.2 mg/kg 10 min of drug administration DZP: 35/35 DZP: 2.94 ± 2.62 DZP: 0/35
F. Brigo et al. / Epilepsy & Behavior xxx (2015) xxx–xxx
(mean ± SD, min)
p = 0.007
IN MDZ 0.2 mg/kg vs. Seizure cessation within MDZ: 30/30 MDZ: 3.16 ± 1.24 MDZ: 3.16 ± 1.24 MDZ: 0/30
IV DZP 0.3 mg/kg 10 min of drug administration DZP: 30/30 DZP: 2.16 ± 1.02 DZP: 6.42 ± 2.59 DZP: 0/30
(mean ± SD, min) (mean ± SD, min)
p = 0.001 p b 0.001
IN MDZ 0.2 mg/kg vs. Seizure cessation within MDZ: 18/27 MDZ: 3.37 ± 2.46 MDZ: 3.01 ± 2.79 MDZ: 6.67 ± 3.12 MDZ: 0/27
IV DZP 0.3 mg/kg 10 min of drug administration DZP: 15/23 DZP: 14.13 ± 3.39 DZP: 2.67 ± 2.31 DZP: 17.18 ± 5.09 DZP: 1/23
(mean ± SD, min) (mean ± SD, min) (mean ± SD, min)
p = 0.001 p = 0.05 p = 0.001
IN MDZ 0.2 mg/kg vs. Seizure cessation within MDZ: 20/23 MDZ: 0/23
rectal DZP 0.3 mg/kg 10 min of drug administration DZP: 13/22 DZP: 0/22
IN MDZ 0.2 mg/kg vs. Seizure cessation within MDZ: 89/92a MDZ: 50.6 ± 14.1 MDZ: 116.7 ± 126.9 MDZ: 0/92b
rectal DZP 0.3 mg/kg 10 min of drug administration DZP: 85/96a DZP: 68.3 ± 55.1 DZP: 178.6 ± 179.4 DZP: 1/96b
(mean ± SD, s) (mean ± SD, s)
p = 0.002 p = 0.005
IN MDZ 10 mg vs. Seizure cessation within MDZ: 50/61 MDZ: 4.6 ± 3.4 MDZ: 0/61
rectal DZP 10 mg 15 min of drug administration DZP: 56/63 DZP: 4.3 ± 3.4
(mean ± SD, min) DZP: 0/63
p = 0.6
IN MDZ 0.2 mg/kg Seizure cessation Seizure cessation MDZ: 5.0 [4.0–7.0] MDZ: 3.0 [1.0–10.0] MDZ: 10.5[7.0–18.0] MDZ: 4/50
(max dose: 10 mg) vs. (time not reported) before emergency DZP: 5.0 [4.0–8.0] DZP: 4.3 [2.0–14.5] DZP: 12.5[7.0–30.0] DZP: 1/42
rectal DZP0.3 to 0.5 mg medical ervices (median [IQR], min) (median [IQR], min) (median [IQR], min)
(max dose: 20 mg) arrived p = 0.57 P = 0.09 p = 0.25
MDZ: 42/50
DZP: 34/42
Buccal MDZ 0.2 mg/kg Seizure cessation within MDZ: 51/60 MDZ: 0.97 ± 0.23 MDZ: 1.69 ± 0.93 MDZ: 2.39 ± 1.04 MDZ: 0/60
vs. IV DZP 0.3 mg/kg 5 min of drug administration DZP: 56/60 DZP: 2.07 ± 0.84 DZP: 1.13 ± 0.5 DZP: 2.98 ± 1.01 DZP: 0/60
(mean ± SD, min) (mean ± SD, min) (mean ± SD, min)
p b 0.001 p b 0.001 p = 0.004
Buccal MDZ 2.5 mg Seizure cessation within MDZ: 22/32 MDZ: 5.68 ± 2.39 MDZ: 7/32b
(age 6–12 months), 10 min of drug administration DZP: 42/60 DZP: 4.52 ± 2.68 DZP: 13/60b
5 mg (1–4 years), (mean ± SD, min)
7.5 mg (5–9 years), p = 0.09
and 10 mg (≥10 years)
vs. IV DZP 0.3 mg/kg
 Buccal midazolam versus rectal diazepam
epilepticus: A systematic review with meta-analysis, Epilepsy Behav (2015), http://dx.doi.org/10.1016/j.yebeh.2015.02.030
Please cite this article as: Brigo F, et al, Nonintravenous midazolam versus intravenous or rectal diazepam for the treatment of early status

Scott et al. [12] Buccal MDZ 10 mg Seizure cessation within MDZ: 30/40a MDZ: 2[1–4] MDZ: 4–5 MDZ: 0/40b
vs. rectal DZP 10 mg 10 min of drug administration DZP: 23/39a DZP: 2[1–3] DZP: 4–5 DZP: 0/39b
(median [IQR], min) (median, min)
p = 0.81 p=1
Baysun et al. [16] Buccal MDZ 0.25 mg/kg vs. Seizure cessation within MDZ: 18/23 MDZ: 0/23
rectal DZP 0.5 mg/kg 10 min of drug administration DZP: 17/20 DZP: 1/20
(children ≤ 5 years) or
0.3 mg/kg (children ≥ 6 years)
McIntyre et al. [17] Buccal MDZ 0.5 mg/kg vs. Seizure cessation within MDZ: 61/109a MDZ: 8 [5–20] MDZ: 5/109b
rectal DZP 0.5 mg/kg 10 min of drug administration, DZP: 30/110a DZP: 15 [5–31] DZP: 7/110b
no recurrence within 1 h Initial seizure control: (median [IQR], min)
MDZ: 71/109a p b 0.001
DZP: 45/110a
Mpimbaza et al. [20] Buccal MDZ 0.5 mg/kg vs. Seizure cessation within MDZ: 115/165 MDZ: 4.75[3.02–6.52] MDZ: 2/165
rectal DZP 0.5 mg/kg 10 min of drug administration, DZP: 94/165 DZP: 4.35 [2.72–6.58] DZP: 2/165
no recurrence within 1 h Initial seizure control: (median [IQR], min)
MDZ: 125/165 p = 0.518
DZP: 114/165
Ashrafi et al. [22] Buccal MDZ 0.3–0.5 mg/kg Seizure cessation within 5 min MDZ: 49/49 MDZ: 2 MDZ: 4 MDZ: 0/49
versus rectal DZP 0.5 mg/kg of drug administration, DZP: 40/49 DZP: 3 DZP: 5 DZP: 0/49
no recurrence within 1 h Seizure control (median, min) (median, min)

F. Brigo et al. / Epilepsy & Behavior xxx (2015) xxx–xxx

within 8 min: p b 0.001 p b 0.001
MDZ: 49/49
DZP: 49/49

Intramuscular midazolam versus intravenous diazepam

Chamberlain et al. [11] IM MDZ 0.2 mg/kg Seizure cessation within MDZ: 12/13 MDZ: 3.3 ± 2.0 MDZ: 4.5 ± 3.0 MDZ: 7.8 ± 4.1 MDZ: 0/13
(max dose: 7 mg) vs. 10 min of drug administration DZP: 10/11 DZP: 7.8 ± 3.2 DZP: 3.4 ± 2.0 DZP: 11.2 ± 3.6 DZP: 0/11
IV DZP 0.3 mg/kg (mean ± SD, min) (mean ± SD, min) (mean ± SD, min)
(max dose: 10 mg) p = 0.001 p = 0.32 p = 0.047
Shah and Deshmukh [18] IM MDZ 0.2 mg/kg vs. Seizure cessation within MDZ: 45/50 MDZ: 97.22 (15–240) MDZ: 0/50
IV DZP 0.2 mg/kg 5 min of drug administration DZP: 29/31 DZP: 250.35 (90–300) DZP: 0/31
(without IV access) (without IV access) DZP: 0/34
DZP: 25/34 p b 0.005
(with IV access) DZP: 119.44 (1–270)
(with IV access)
p = 0.17
(mean (range), s)
Portela et al. [29] IM MDZ 0.5 mg/kg Seizure cessation within MDZ: 14/16 MDZ: 2.8 ± 1.5 MDZ: 4.4 ± 0.5 MDZ: 7.3 ± 1.4 MDZ: 1/16
(max dose: 10 mg) versus 5 min of drug administration DZP: 14/16 DZP: 7.4 ± 4.1 DZP: 3.3 ± 0.8 DZP: 10.6 ± 3.9 DZP: 2/16
IV DZP 0.5 mg/kg (mean ± SD, min) (mean ± SD, min) (mean ± SD, min)
(max dose: 15 mg) p = 0.001 p b 0.001 p = 0.006

Intramuscular midazolam versus rectal diazepam

Momen et al. [28]
IM MDZ 0.3 mg/kg vs. Seizure cessation within Initial seizure control: MDZ: 65(54–95) MDZ: 66(24–245) MDZ: 127 (83–320) MDZ: 0/50
rectal DZP 0.5 mg/kg 10 min of drug administration, MDZ: 48/50 DZP: 65(55–190) DZP: 130 (45–600) DZP: 243 (115–725) DZP: 0/50
no recurrence within 1 h DZP: 47/50 (median (range), s) (median (range), s) (median (range),s)
None of the patients p b 0.017 p b 0.001 p b 0.001
showed recurrence
within 1 h

DZP: diazepam, h: hour, IM: intramuscular, IN: intranasal, IQR: interquartile range, IV: intravenous, MDZ: midazolam, and SD: standard deviation.
a
Expressed as the number of episodes; some patients experienced more than one seizure.
b
Expressed as the number of episodes; some patients experienced more than one adverse effect.
c
Time required inserting the intravenous line not included.

7
8 F. Brigo et al. / Epilepsy & Behavior xxx (2015) xxx–xxx

through the portal circulation with consequent biotransformation of

Fig. 1. Study flow diagram.
in blood supply which are responsible for drug absorption into the sys-
temic circulation and avoidance of first-pass metabolism. The same ad-
vantage occurs after intranasal administration, which prevents passing
MDZ through hydroxylation by hepatic microsomal oxidative mecha-
nisms. Furthermore, buccal MDZ as well as intranasal MDZ may be ad-
ministered more rapidly than rectal DZP without the need to remove
clothing and position the patient appropriately. The same limitation
holds true for intravenous administration of DZP, which requires the
placement of an intravenous access, leading to a delay before drug ad-
ministration [18]. In addition, gaining access to an intravenous route re-
quires some expertise and may be difficult in patients with convulsive
seizures or in infants.
The faster administration of non-IV MDZ (administered by buccal,
intranasal, or intramuscular route) may explain why, in this systematic
review, times from arrival in the emergency department to drug admin-
istration and to seizure cessation were shorter with non-IV MDZ than
with intravenous or rectal DZP.
The motto “time is brain” applies not only for stroke but also for SE as
the overall duration of seizure activity represents a primary determinant
of outcome [34]. Transferring a patient with SE from home to hospital
by ambulance may result in delayed treatment with reduced chances
of successful response to a single drug [4]. Hence, the availability of effec-
tive drugs, which can be administered easily by nonmedical staff in
prehospital settings, is welcome. However, to be clinically relevant,
shorter time to drug administration should result also in higher seizure
control. The present meta-analysis shows that the faster administration
of non-IV MDZ does not necessarily result in higher seizure control. How-
ever, the absence of evidence that a difference exists does not mean that
a difference would not be realized if the sample size had been larger [35].
Our review, however, suggests that the magnitude of difference between
non-IV MDZ and intravenous or rectal DZP may be not so high.
Diazepam is more widely available throughout the world than either
lorazepam or MDZ. Rectal DZP administration is a valuable option to treat
emergency seizure situations when the intravenous route is not feasible,
for example, in a number of prehospital settings particularly in the ab-
sence of skilled health-care personnel. However, there is a trend for rectal
DZP to be substituted by buccal or intranasal MDZ, which is socially more
accepted. The present review suggests that buccal MDZ, besides being
more acceptable and easy to use, is also more effective than rectal DZP
for seizure control. Hence, whenever possible, the administration of buc-
cal MDZ should be preferred over rectal DZP. This conclusion is further
supported by the fact that in four RCTs specifically assessing this aspect,
buccal MDZ [20,22] and intranasal MDZ [23,24] were found to be signifi-
cantly preferable (overall satisfaction, easiness of use) than rectal DZP.
Eighteen out of the 19 studies included in this systematic review
were conducted in children (1573 patients younger than 16 years).

Table 3
Risk of bias in included studies.

Study Random sequence Allocation Blinding of participants Blinding of outcome Incomplete outcome Sponsored by
generation concealment and personnel assessment data (attrition bias) pharmaceutical
(selection bias) (selection bias) (performance bias) (detection bias) company

Chamberlainet al. [11] Unclear Unclear Unclear Low Low Unclear

Scott et al. [32] Low Low Low Low Low Unclear
Lahat et al. [13] Low Low Low Low Low No
Fişgin et al. 2002 [14] High Unclear Low Low Low Unclear
Mahmoudian and Zadeh [15] Low Low Low Low Low Unclear
Baysun et al. [16] High Unclear Low Low Low Unclear
McIntyre et al. [17] Low High Low Low Low No
Shah and Deshmukh [18] Unclear Unclear Low Low Low Unclear
Bhattacharyya, 2006 Low Low Low Low Low Unclear
Mpimbaza et al. [20] Unclear Low Low Low Low No
Talukdar and Chakrabarty [21] Low Unclear Low Low Low Unclear
Ashrafi et al. [22] Low Unclear Low Low Low Unclear
de Haan et al. [23] Unclear Unclear Low Low Low No
Holsti et al. [24] Unclear Low Low Low Low No
Javadzadeh et al. [26] Unclear Unclear Low Low Low Unclear
Tonekaboni et al. [25] Low Unclear Low Low Low No
Thakker and Shanbag [27] Low Low Low Low Low No
Momen et al. [28] Low Unclear Low Low Low No
Portela et al. [29] Unclear Low Low Low Low No

Please cite this article as: Brigo F, et al, Nonintravenous midazolam versus intravenous or rectal diazepam for the treatment of early status
epilepticus: A systematic review with meta-analysis, Epilepsy Behav (2015), http://dx.doi.org/10.1016/j.yebeh.2015.02.030
 F. Brigo et al. / Epilepsy & Behavior xxx (2015) xxx–xxx 9

Fig. 2. Non-IV MDZ versus DZP (all routes of administration). (a) Clinical seizure cessation after drug administration and (b) serious adverse effects.

Consequently, the generalizability of results of this systematic review is
restricted to children. Only one study was entirely conducted in an adult
population (21 patients, aged 31 to 69 years), showing no difference in
efficacy or time to seizure cessation after drug administration between
intranasal MDZ and rectal DZP [23].
5.1. Comparison with a previous systematic review with meta-analysis
A systematic review previously compared non-IV MDZ with DZP by
any route in early SE seizures in children and adults, pooling data from
several RCTs into a meta-analysis [36]. Six studies with 774 subjects
were included. Midazolam by any route was superior to DZP by any
route for seizure cessation (RR: 1.52; 95% CIs: 1.27 to 1.82). Non-IV
MDZ was found to be as effective as intravenous DZP (RR: 0.79; 95%
CIs: 0.19 to 3.36) and buccal MDZ more effective than rectal DZP in
achieving seizure control (RR: 1.54; 95% CIs: 1.29 to 1.85). Time for
drug administration was faster for MDZ than for DZP (mean difference:
2.46 min; 95% CIs: 1.52 to 3.39 min), whereas times between drug
administration and seizure cessation were similar between the two
drugs.
Our systematic review confirms these results, especially with regard
to superiority of buccal MDZ over rectal DZP in seizure control and in
shorter interval between arrival at the emergency department and
drug administration for non-IV MDZ. Conversely, our review did not
show any difference in terms of seizure cessation between non-IV
MDZ by any route and DZP by any route. This difference may be due
to the fact that our review differs from that conducted by McMullan
et al. [36] in several aspects: (1) obviously, our review includes studies
which were not considered in the previous review as they were not yet
published [22–29]; (2) the choice of the efficacy outcomes is slightly dif-
ferent as we chose to consider clinical seizure cessation within 15 min of
drug administration instead of failure to achieve seizure cessation as
made by McMullan et al.; (3) our meta-analytic comparisons were con-
ducted using a random-effects model, and not a fixed-effect model, as

Please cite this article as: Brigo F, et al, Nonintravenous midazolam versus intravenous or rectal diazepam for the treatment of early status
epilepticus: A systematic review with meta-analysis, Epilepsy Behav (2015), http://dx.doi.org/10.1016/j.yebeh.2015.02.030
10 F. Brigo et al. / Epilepsy & Behavior xxx (2015) xxx–xxx

Fig. 3. Non-IV MDZ versus DZP (all routes of administration). (a) Time from arrival to drug administration, (b) time from drug administration to clinical seizure cessation, and (c) time from
arrival to clinical seizure cessation.

we preferred a more conservative statistical model to take into account
the variability between studies; and (4) we analyzed the quality assess-
ment of potentially eligible studies focusing more on the methodologi-
cal quality of the trial conduction than on the accuracy and quality of
reporting study results.
5.2. Exploration of heterogeneity
The results of the present meta-analysis should be read with caution
mainly because of the considerable statistical heterogeneity found in
most meta-analytic comparisons that is indicative of inconsistency
in the results of included studies. The term “statistical heterogeneity”
describes the degree of variation in the effect estimates from a set of
studies and indicates the presence of variability among studies beyond
the amount expected due solely to the play of chance.
Such a statistical heterogeneity may be explained both by differ-
ences in clinical characteristics of study participants (clinical heteroge-
neity) and by different drug regimens (methodological heterogeneity)
adopted.
Regarding clinical heterogeneity, potentially relevant aspects to be
taken into account are differences in age of patients, seizure duration,
seizure type, and seizure etiology.
Repeating pooled analyses on clinical seizure cessation and occur-
rence of adverse effects by excluding the RCTs with adult patients [12,
23], we did not find a relevant reduction in statistical inconsistency
among studies, suggesting that age alone does not explain the
variability in study results. It is unlikely that minimal time differences
in seizure duration (5 min versus 10 min) may represent a relevant
source of heterogeneity. Conversely, the lack of individual patient
data on seizure type and etiology (both relevant prognostic factors)
prevented us from performing a more detailed sensitivity analysis
taking into account these aspects.
Regarding seizure type, most studies included patients with dif-
ferent types of convulsive seizures, and in some cases, studies did not
specify in further detail the clinical semiology of seizures. One RCT re-
ported data on seizure control according to seizure types and found
no difference between buccal MDZ and intravenous DZP for generalized
tonic–clonic seizures (88.9% versus 90.2%, respectively; p = 0.559)
and tonic seizures (100% versus 100%, respectively; p = 0.999). In
another study from India, buccal MDZ was less effective than intrave-
nous DZP in controlling “clonic partial seizures” (63.6% versus 100%,
respectively; p = 0.01) [21]. These findings suggest that MDZ or DZP
might be more effective in some subtypes of convulsive seizures. How-
ever, this hypothesis needs to be tested in further studies with larger
sample size.
Regarding methodological heterogeneity, potentially relevant as-
pects to be considered are differences in randomization/allocation
concealment and blinding; sponsorship by pharmaceutical

Please cite this article as: Brigo F, et al, Nonintravenous midazolam versus intravenous or rectal diazepam for the treatment of early status
epilepticus: A systematic review with meta-analysis, Epilepsy Behav (2015), http://dx.doi.org/10.1016/j.yebeh.2015.02.030
 F. Brigo et al. / Epilepsy & Behavior xxx (2015) xxx–xxx
Fig. 4. Buccal MDZ versus rectal DZP. Clinical seizure cessation after drug administration. Results are expressed as Odds Ratio.
companies; definition of SE (seizures lasting N 5 min, seizures lasting
N10 min, seizures at arrival to the emergency department); drug
dosage; and intervention settings (hospital or prehospital). Ran-
domization was adequate in the great majority of studies, and over-
all, the methodological quality of studies was good. Even if some
studies were not blinded and in others, blinding was not explicitly
reported, the ‘hard’ outcomes chosen in all studies (clinical seizure
cessation, serious adverse effects) were not probably influenced by
lack of blinding. As a consequence, all studies have a low risk of per-
formance and detection bias. However, because of the limited infor-
mation available, we were not able to assess to what extent these
factors may impact on inconsistency among trial results.
6. Conclusions
Non-IV MDZ is as effective and safe as intravenous or rectal DZP in
terminating early SE in children. Times from arrival in the emergency
department to drug administration and to seizure cessation are
shorter with non-IV MDZ than with intravenous or rectal DZP, but
this does not necessarily translate into higher seizure control. An ex-
ception may be the buccal MDZ, which, besides being socially more
acceptable and easier to administer, might also have higher efficacy
than rectal DZP in seizure control. The generalizability of these re-
sults is restricted to children as no sufficient data are available for
adults.
Several areas require attention in future research in the treat-
ment of SE: investigators should use acceptable and uniform defini-
tions of SE (including different stages of SE) and clear and uniform
methods of data presentation to facilitate meta-analysis [37].
Datasets with individual patient data (raw data) should be made
publicly available by trial authors to facilitate subgroup meta-
analyses for relevant prognostic aspects such as etiology and seizure
type.
Please cite this article as: Brigo F, et al, Nonintravenous midazolam versus intravenous or rectal diazepam for the treatment of early status
epilepticus: A systematic review with meta-analysis, Epilepsy Behav (2015), http://dx.doi.org/10.1016/j.yebeh.2015.02.030
11
Acknowledgments
We thank Piero Bonzano and Vincenzo De Simone (Accord
Healthcare) and Alberto Liuti (ViroPharma) for seeking information
on unpublished randomized controlled trials on buccal midazolam or
studies in progress.
Conflict of interest
FB received speakers' honoraria from UCB Pharma.
ET has acted as a paid consultant to Bial, Biogen Idec, Eisai, Ever
Neuropharma, Medtronics, Takeda, Upsher-Smith, and UCB; has re-
ceived speakers' honoraria from Bial, Boehringer, Eisai, GL Lannacher,
and UCB Pharma; and has received research funding from Biogen
Idec, Merck, Novartis, Red Bull, UCB Pharma, the European Union,
FWF (Österreichischer Fond zur Wissenschaftsförderung), and
Bundesministerium für Wissenschaft und Forschung.
RN and FT declare no conflicts of interest.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.yebeh.2015.02.030.
References
[1] DeLorenzo RJ, Pellock JM, Towne AR, Boggs JG. Epidemiology of status epilepticus.
J Clin Neurophysiol 1995;12:316–25.
[2] Epilepsy Foundation of America's Working Group on Status Epilepticus. Treatment
of convulsive status epilepticus. Recommendations of the Epilepsy Foundation of
America's Working Group on Status Epilepticus. JAMA 1993;270:854–9.
[3] Lowenstein DH, Bleck T, Macdonald RL. It's time to revise the definition of status
epilepticus. Epilepsia 1999;40:120–2.
[4] Alldredge BK, Wall DB, Ferriero DM. Effect of prehospital treatment on the outcome
of status epilepticus in children. Pediatr Neurol 1995;12:213–6.
12 F. Brigo et al. / Epilepsy & Behavior xxx (2015) xxx–xxx
[5] Silbergleit R, Lowenstein D, Durkalski V, Conwit R, NETT Investigators. Lessons from
the RAMPART study—and which is the best route of administration of benzodiaze-
pines in status epilepticus. Epilepsia 2013;54(Suppl. 6):74–7.
[6] Sánchez-Carpintero R, Camino R, Smeyers P, Raspall-Chaure M, Martínez-Bermejo A,
Ruiz-Falcó ML, et al. Use of benzodiazepines in prolonged seizures and status epilep-
ticus in the community. An Pediatr (Barc) 2014;81:400.e1–6.
[7] Silbergleit R, Durkalski V, Lowenstein D, Conwit R, Pancioli A, Palesch Y, et al. Intra-
muscular versus intravenous therapy for prehospital status epilepticus. N Engl J Med
2012;366:591–600.
[8] Welch RD, Nicholas K, Durkalski-Mauldin VL, Lowenstein DH, Conwit R, Mahajan PV,
et al. Intramuscular midazolam versus intravenous lorazepam for the prehospital
treatment of status epilepticus in the pediatric population. Epilepsia 2015. http://
dx.doi.org/10.1111/epi.12905.
[9] Higgins JPT, Altman DG, Sterne JAC. Chapter 8: assessing risk of bias in included
studies. In: Higgins JPT, Green S, editors. Cochrane Handbook for Systematic Reviews
of Interventions Version 5.1.0 (updated March 2011)The Cochrane Collaboration;
2011 [Available at: www.cochrane-handbook.org (accessed 21st January 2015)].
[10] Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-
analyses. BMJ 2003;327:557–60.
[11] Chamberlain JM, Altieri MA, Futterman C, Young GM, Ochsenschlager DW, Waisman
Y. A prospective, randomized study comparing intramuscular midazolam with intra-
venous diazepam for the treatment of seizures in children. Pediatr Emerg Care 1997;
13:92–4.
[12] Scott RC, Besag FM, Neville BG. Buccal midazolam and rectal diazepam for treatment
of prolonged seizures in childhood and adolescence: a randomised trial. Lancet
1999;353:623–6.
[13] Lahat E, Goldman M, Barr J, Bistritzer T, Berkovitch M. Comparison of intranasal
midazolam with intravenous diazepam for treating febrile seizures in children:
prospective randomised study. BMJ 2000;321:83–6.
[14] Fişgin T, Gurer Y, Teziç T, Senbil N, Zorlu P, Okuyaz C, et al. Effects of intranasal
midazolam and rectal diazepam on acute convulsions in children: prospective
randomized study. J Child Neurol 2002;17:123–6.
[15] Mahmoudian T, Zadeh MM. Comparison of intranasal midazolam with intravenous
diazepam for treating acute seizures in children. Epilepsy Behav 2004;5:253–5.
[16] Baysun S, Aydin OF, Atmaca E, Gürer YK. A comparison of buccal midazolam and
rectal diazepam for the acute treatment of seizures. Clin Pediatr (Phila) 2005;44:
771–6.
[17] McIntyre J, Robertson S, Norris E, Appleton R, Whitehouse WP, Phillips B, et al. Safety
and efficacy of buccal midazolam versus rectal diazepam for emergency treatment
of seizures in children: a randomised controlled trial. Lancet 2005;366:205–10.
[18] Shah I, Deshmukh CT. Intramuscular midazolam vs intravenous diazepam for acute
seizures. Indian J Pediatr 2005;72:667–70.
[19] Bhattacharyya M, Kalra V, Gulati S. Intranasal midazolam vs rectal diazepam in acute
childhood seizures. Pediatr Neurol 2006;34:355–9.
[20] Mpimbaza A, Ndeezi G, Staedke S, Rosenthal PJ, Byarugaba J. Comparison of buccal
midazolam with rectal diazepam in the treatment of prolonged seizures in
Ugandan children: a randomized clinical trial. Pediatrics 2008;121:e58–64.
Please cite this article as: Brigo F, et al, Nonintravenous midazolam versus intravenous or rectal diazepam for the treatment of early status
epilepticus: A systematic review with meta-analysis, Epilepsy Behav (2015), http://dx.doi.org/10.1016/j.yebeh.2015.02.030
[21] Talukdar B, Chakrabarty B. Efficacy of buccal midazolam compared to intravenous
diazepam in controlling convulsions in children: a randomized controlled trial.
Brain Dev 2009;31:744–9.
[22] Ashrafi MR, Khosroshahi N, Karimi P, Malamiri RA, Bavarian B, Zarch AV, et al. Effi-
cacy and usability of buccal midazolam in controlling acute prolonged convulsive
seizures in children. Eur J Paediatr Neurol 2010;14:434–8.
[23] de Haan GJ, van der Geest P, Doelman G, Bertram E, Edelbroek P. A comparison of
midazolam nasal spray and diazepam rectal solution for the residential treatment
of seizure exacerbations. Epilepsia 2010;51:478–82.
[24] Holsti M, Dudley N, Schunk J, Adelgais K, Greenberg R, Olsen C, et al. Intranasal
midazolam vs rectal diazepam for the home treatment of acute seizures in pediatric
patients with epilepsy. Arch Pediatr Adolesc Med 2010;164:747–53.
[25] Tonekaboni SH, Shamsabadi FM, Anvari SS, Mazrooei A, Ghofrani M. A comparison of
buccal midazolam and intravenous diazepam for the acute treatment of seizures in
children. Iran J Pediatr 2012;22:303–8.
[26] Javadzadeh M, Sheibani K, Hashemieh M, Saneifard H. Intranasal midazolam com-
pared with intravenous diazepam in patients suffering from acute seizure: a ran-
domized clinical trial. Iran J Pediatr 2012;22:1–8.
[27] Thakker A, Shanbag P. A randomized controlled trial of intranasal-midazolam versus
intravenous-diazepam for acute childhood seizures. J Neurol 2013;260:470–4.
[28] Momen AA, Azizi Malamiri R, Nikkhah A, Jafari M, Fayezi A, Riahi K, et al. Efficacy
and safety of intramuscular midazolam versus rectal diazepam in controlling status
epilepticus in children. Eur J Paediatr Neurol 2014. http://dx.doi.org/10.1016/j.ejpn.
2014.11.007.
[29] Portela JL, Garcia PC, Piva JP, Barcelos A, Bruno F, Branco R, et al. Intramuscular
midazolam versus intravenous diazepam for treatment of seizures in the pediatric
emergency department: a randomized clinical trial. Med Intensiva 2014. http://dx.
doi.org/10.1016/j.medin.2014.04.003.
[30] Ranganathan P, Pramesh CS, Buyse M. Common pitfalls in statistical analysis: “no
evidence of effect” versus “evidence of no effect”. Perspect Clin Res 2015;6:62–3.
[31] Schwagmeier R, Alincic S, Striebel HW. Midazolam pharmacokinetics following
intravenous and buccal administration. Br J Clin Pharmacol 1998;46:203–6.
[32] Scott RC, Besag FM, Boyd SG, Berry D, Neville BG. Buccal absorption of midazolam:
pharmacokinetics and EEG pharmacodynamics. Epilepsia 1998;39:290–4.
[33] Papillard S. Safety and efficacy of buccal midazolam vs rectal diazepam for emergen-
cy treatment of seizures in children: a randomised controlled trial. Arch Pediatr 2005;
12:1644–5.
[34] Towne AR, Pellock JM, Ko D, DeLorenzo RJ. Determinants of mortality in status
epilepticus. Epilepsia 1994;35:27–34.
[35] Altman DG, Bland JM. Absence of evidence is not evidence of absence. Br Med J
1995;311:485.
[36] McMullan J, Sasson C, Pancioli A, Silbergleit R. Midazolam versus diazepam for the
treatment of status epilepticus in children and young adults: a meta-analysis. Acad
Emerg Med 2010;17:575–82.
[37] Brigo F, Storti M, Del Felice A, Fiaschi A, Bongiovanni LG. IV valproate in generalized
convulsive status epilepticus: a systematic review. Eur J Neurol 2012;19:1180–91.