Introduction

Rapid antiseizure treatment is crucial for persistent, prolonged, or recurrent epileptic seizures. The majority of epileptic seizures spontaneously cease within a few minutes [1]. However, cases that involve prolonged convulsive seizure activity (more than 5 min) or clusters of convulsive seizures without complete remission are defined as convulsive status epilepticus. If left untreated, status epilepticus can result in life-threatening systemic complications [2].

In those cases, rapid administration of antiseizure medications is of great importance but can be challenging in the acute care setting given the background of convulsions and the difficulty in obtaining rapid intravenous access. Benzodiazepines are the recommended first-line therapy for treating status epilepticus [3,4,5]. They exert their antiseizure effects by modulation of inhibitory GABAA receptors, with each agent exhibiting different pharmacokinetic and pharmacodynamic properties [4, 6].

Midazolam administered intranasally may be an alternative when intravenous access is not available [4]. Intranasal absorption occurs mainly through the epithelium via olfactory and trigeminal nerve pathways into the central nervous system [7]. Intranasal application results in rapid peak plasma concentrations (mean at 14 min) and is associated with high bioavailability (mean 83%) [8]. These results are comparable to that after intramuscular administration (mean peak plasma concentration at 25 min; mean bioavailability 87%) [9]. Although clinicians appreciate the convenience of midazolam nasal spray, studies on intranasal midazolam (in-MDZ) administration have primarily focused on pediatric cohorts and its use for sedation purposes [10,11,12,13]. Recently, intramuscular administration of midazolam has gained increasing attention. In a study published after our screening period, a direct comparison with in-MDZ revealed that intramuscular administration was more frequently associated with severe hypotension, which may be a further advantage for in-MDZ administration (especially in out-of-hospital emergency settings without immediate monitoring options) [14]. Another recent retrospective study examined the real-world practice regarding the efficacy of midazolam at different doses and routes of administration [15]. It was shown that the midazolam dosage had an impact on clinical outcomes. It was concluded that intranasal administration might be less effective in terminating status epilepticus compared to intramuscular administration, although the observed differences may also be due to residual confounding associated with the retrospective design.

To date, the efficacy and tolerability of in-MDZ administration for treating epileptic seizures and status epilepticus in adults have not been systematically reviewed. We therefore sought to perform a systematic review regarding the use of in-MDZ administration in adults with epilepsy, epileptic seizures, or status epilepticus and compile information on its efficacy and safety from the literature.

Methods

Registration and Reporting

This systematic review was registered on the International Prospective Register for Systematic Reviews (PROSPERO) and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. The search and study selection were conducted after study registration on November 1, 2022 (PROSPERO study ID: CRD42022369040).

Ethical Standards

This study was conducted in compliance with the ethical standards outlined in the Declaration of Helsinki and its amendments.

Databases and Search Strategy

Two databases (Embase and Medline) were screened by one reviewer (TDD) using the Ovid interface. The search aimed to identify articles in English related to efficacy and tolerability aspects of in-MDZ application as antiseizure treatment in adults. The search period was from January 1, 1985, to October 1, 2022. A predefined search algorithm was used to identify eligible studies (details in Supplemental Text Sect. 1).

Eligibility Criteria and Study Selection

We assessed the study eligibility based on the following criteria, all of which had to be met for a study to be included: (1) study on humans; (2) study that includes adult patients (i.e., ≥ 18 years); (3) study of in-MDZ application in the context of either epilepsy, epileptic seizure(s), or status epilepticus; (4) a randomized clinical trial or observational cohort study (excluding case series); (5) publication written in English language; and (6) study publication between January 1, 1985 and October 1, 2022.

Three reviewers (TDD, DV, and RS) screened the studies for eligibility and manually filtered the retrieved studies to exclude those not fulfilling the previously outlined eligibility criteria. The remaining collaborators reviewed their decisions. The final decision to include studies was made by majority consensus. An online literature management program was used for the screening and selection process (see Data extraction and synthesis).

Data Extraction and Synthesis

Authors, year of publication, participating country(ies), study design, number of patients treated with in-MDZ, information on the underlying condition, applied midazolam dosage, comparators, and outcomes of interest (as outlined in the Outcomes of interest section) were extracted and archived by DV using the freely available SRDR + (Systematic Review Data Repository) online tool. Two reviewers (TDD and RS) subsequently reviewed and sorted the extracted data according to the study design. Disagreements were resolved through majority agreement. Corresponding authors were contacted and asked for unreported or missing data.

Outcomes of Interest

The primary outcome was the efficacy of in-MDZ administration in terms of seizure termination (defined as suppressed clinical seizure activity) and recurrence (after first in-MDZ application with subsequent suppressed clinical seizure activity). The secondary outcome was the tolerability, reported side effects, or adverse events of in-MDZ administration. The latter was considered exploratory because we recorded all reported side effects in temporal relation to the in-MDZ administration and (if applicable) midazolam with different administration routes or other benzodiazepines.

Risk of Bias Assessment

Risk of bias (ROB) assessment was conducted using the Cochrane Risk of Bias Tool for randomized trials and the Risk of Bias in Non-Randomized Studies of Interventions tool. Both tools are standardized and include a systematic assessment of the study design, implementation, and reporting of results.

Evidence Rating

Studies comparing in-MDZ administration with a control intervention (another benzodiazepine[s], midazolam with another administration route, or placebo) were assigned an evidence rating using the Grading of Recommendations Assessment, Development, and Evaluation approach [16]. The secondary outcome, which was exploratory in nature, did not undergo rating.

Statistics

Descriptive statistical analyses were performed to obtain relative frequencies (percentages and standard deviations [SDs]). For the certainty rating, risk ratios and absolute effects were calculated. STATA version 17.0 (StataCorp LLC, College Station, TX) was used for all analyses.

Results

Study Selection and Characteristics Of Included Studies

The screening of 185 articles identified via two databases (Embase and Medline) yielded 12 studies (including one study [17] identified during a detailed review process) with a total of 929 patients (without rigorous stratification by age group) treated with in-MDZ that addressed aspects of efficacy and tolerability of in-MDZ application for antiseizure treatment in adults (Fig. 1).

Fig. 1
figure 1

Study flow diagram. *These seven articles have been added, of which three had to be excluded because they only included healthy study participants

Full size image

Most of the included studies (66.7%) were published after 2018 [17,18,19,20,21,22,23,24]. The only two randomized trials [18, 19] were multicentric (with their main site in the United States) and accounted for 22.1% of the overall patient cohort with a total of 205 patients who received in-MDZ (Table 1). The remaining ten cohort studies, except for the study by Wheless et al. [20], were all monocentric and mostly retrospective. Among all retrospective studies, the majority were conducted in Germany, followed by other European countries (Netherlands and Switzerland), North America, and Australia (Fig. 2).

Table 1 Characteristics of the included studies with patients with status epilepticus, epileptic seizure, or known epilepsy
Full size table
Fig. 2
figure 2

Included studies according to country of origin and over time. The geographical distribution refers to the country of the main site (as indicated by the host institution of the corresponding author). The point size corresponds to the number of studies. The Khartis software (Sciences Po Cartography Laboratory, 2017) was used to create the visualizations

Full size image

Successful treatment was mainly defined clinically as suppressed seizure activity within 10 to 15 min [18, 20, 22, 25]. The exception was two studies [17, 26] that did not explicitly define what was considered a successful treatment response (in some studies, successful treatment implied a subsequent absence of seizures for 2 to 6 h [18, 20, 25]). For seizure recurrence, the majority of studies applied a time frame of 6 h [17,18,19,20] and/or 24 h [17, 24, 27], with one study confining recurrence to a very early recurrence (within 1 h after first application) [28]. Data on seizure termination and recurrence after the first administration of in-MDZ were available for 8 of 12 studies (66.7%) (Table 1).

The in-MDZ doses administered mostly ranged from 2.5 to 10 mg per single dose [17,18,19,20,21,22, 24,25,26,27,28], with weight-adapted dosing reported in only one study [23] and an allowed maximum dose of 20 mg. Repetitive doses were permitted explicitly in five studies [17, 18, 20, 22, 23, 26]; in the remaining studies, repetitive administration was not permitted or specified.

Of a total of 12 included studies, two studies compared the efficacy of in-MDZ administration for seizure termination with placebo [18, 19], two studies compared the efficacy of in-MDZ administration versus no in-MDZ application within the same individuals [24, 27], and four studies compared the efficacy of in-MDZ administration with other administration routes and/or other benzodiazepines (rectal diazepam [25], intranasal diazepam [22], different regimens [intravenous midazolam, intramuscular midazolam, rectal, intravenous, or intramuscular diazepam] [23], and intravenous lorazepam [17]). One study had a comparator that was not further specified (mentioned as “previously used alternative emergency medication” [28]), and the remaining three studies [20, 21, 26] had no comparator.

Efficacy of in-MDZ Administration for Antiseizure Treatment In Adults

A mean of 72.7% seizures stopped after the first administration of in-MDZ, with a standard deviation (SD) of 18% (Fig. 3). This compares to various other treatments: 70% success with placebo in one randomized study [18], 89–100% with rectal diazepam in two studies [23, 25], 63% with intranasal diazepam [22], and 57% with intravenous midazolam [23]. The highest success rate with in-MDZ was observed in studies in which a dose of 10 mg was administered or at least recommended [25, 26]. After the first in-MDZ treatment, seizures recurred in 36.5% of cases (SD 15.9%). For other treatments, recurrence rates were 61–63% with placebo [18], 60–75% without midazolam [24, 27], 0–9.5% with rectal diazepam [23, 25], and 43% with intravenous midazolam [23].

Fig. 3
figure 3

Successful seizure termination and reported side effects related to intranasal midazolam administration. *Mean value. **Percentages based only on one available study for each comparator. #Total number of patients treated with available data on successful seizure termination (no stratification by age group possible due to lack of information). IN-DZP intranasal diazepam, IN-MDZ intranasal midazolam, IV-MDZ intravenous midazolam

Full size image

Tolerability of Intranasal Midazolam Administration for Antiseizure Treatment in Adults

Nine of 12 (75%) studies reported side effects in detail. The most commonly reported adverse events after in-MDZ administration were dizziness (three studies [20, 22, 25]; mean 23.5% [SD 38.6%]), confusion (one study [17]; 17.4%), local irritation (eight studies [18,19,20,21,22, 24, 25, 27]; mean 16.6% [SD 9.6%]), sedation (eight studies [17,18,19,20,21,22, 24, 25]; mean 12.7% [SD 9.7%]), and respiratory difficulties (three studies [17, 24, 27]; mean 5.7% [SD 9.8%]; Fig. 3). Less common adverse events included nausea (three studies [19, 20, 24]; mean 4.9% [SD 4.1%]), abnormal product taste (three studies [18,19,20]; mean 4% [SD 2.2%]), suicidal ideation (one study [19]; 3.6%), headache (six studies [18,19,20, 22, 24, 25]; mean 2.6% [SD 2.7%]), convulsions (one study [20]; 3%), and restlessness (one study [25]; 2%). In the two studies with (administered or at least recommended) a higher in-MDZ dose (10 mg), there was no excess of respiratory depression reported [25, 26]. However, in one of the two studies, which reported adverse effects in detail, a relatively high proportion of patients reported drowsiness (two of three study participants) [25]. In addition, one study [17] reported that 21.7% of patients treated with in-MDZ had tongue bites compared to 7.4% in the intravenous lorazepam group (not statistically significant). Only three studies [17, 22, 25] provided data on reported side effects with either midazolam administered via different routes or other benzodiazepines.

ROB in Included Studies

Table 2 summarizes the ROB assessment for the included studies. All included randomized trials were rated at low ROB for the randomization process, the effect of assignment to the intervention group, handling of missing outcome data, outcome measurement, and reporting of results. However, one study [18] was rated as of concern for the effect of adherence to the assigned intervention. The overall ROB was considered low because both randomized trials had a low ROB rating in most domains assessed.

Table 2 Risk of bias assessment
Full size table

The included nonrandomized studies [17, 20,21,22,23,24,25,26,27,28] were all rated as of some concern for risk of confounding and selection bias. The ROB in classifications of interventions was considered low in all studies. In contrast, the ROB for deviations from the planned interventions was rated as low in only five of nine studies. For the remaining studies, there was at least some concern about the ROB in this domain. All but one study (the latter was rated as of concern) were classified as low ROB because of missing data. Eight studies were rated as having a high ROB regarding the measurement of outcomes. All studies were considered low risk in the domain of bias risk in the selection of reported results. Overall, the ROB was graded as moderate for the nonrandomized studies because most studies raised at least some concern about the overall ROB.

Discussion

Rapid and easy administration of benzodiazepines can be challenging in the acute care setting in patients with prolonged epileptic seizures, seizure clusters, or status epilepticus. From an application perspective, in-MDZ administration appears to be particularly promising. This route of administration seems less invasive compared to intravenous, intramuscular, or intraosseous routes and less stigmatizing compared to rectal administration and is thus of potential clinical relevance, particularly regarding an administration by a layperson. To date, to the best of our knowledge, no systematic review has examined the existing global evidence on the efficacy and tolerability of in-MDZ administration for antiseizure treatment in adults.

Within the framework of this systematic review, we have compiled the available evidence from the international literature, including 12 identified studies published within the last 38 years, with most published over the previous 5 years. Most of the included studies provided data regarding the proportion of successful seizure termination and the recurrence of seizures after the first in-MDZ administration. The intranasal administration led to a mean seizure termination in nearly three of four patients. Compared with the reported successful seizure termination rates with other benzodiazepines or midazolam administered via alternative routes (except for rectal midazolam), in-MDZ administration did not appear to be inferior to other benzodiazepine alternatives (i.e., intravenously administered midazolam, diazepam given via intranasal route). It should be noted that the few studies in which comparatively higher doses (10 mg per single dose) were administered or at least recommended had the highest rates of successful seizure termination, which might suggest a dose–response relationship. The reported higher rates of successful seizure termination (and lower recurrence rates) under rectal diazepam might be attributable, in part, to the longer half-life of diazepam [29]. This could be relevant because the definition of successful seizure termination was linked to seizure absence (over a variable period) immediately following the first dose in some studies. The high rate of spontaneous seizure termination in the placebo group is in line with evidence from the literature that most epileptic seizures are self-limiting and last less than a few minutes [1]. However, the evidence for the efficacy of in-MDZ is overall low because only five studies (of which only one was a randomized controlled trial) compared in-MDZ administration to a comparator (Supplemental Table 1). From a pharmacokinetic perspective, it should be noted that rectal administration might not be as efficient compared to most other routes of administration because of slower absorption rates through the mucosa compared to all other routes of administration [4]. In addition, rectal emergency treatment is not universally accepted and is associated with adverse psychosocial effects, such as embarrassment, increased stigmatization, social fear, and inconvenience in administration [30]. Further points that should be considered when assessing efficacy in terms of seizure termination is that first-line doses of administrated in-MDZ varied by up to fourfold across the included studies, and successful seizure termination was not defined uniformly across the individual studies. Considerable evidence suggests that underdosing of benzodiazepines is one of the main reasons for the failure of seizure termination in the context of status epilepticus and should therefore be considered when interpreting these results [31,32,33]. Furthermore, there is a relatively wide range of seizure termination rates, ranging from 52% to almost 97%. This observed variation could be partially explained by the inhomogeneous cohorts and the lack of standardization regarding reported seizure termination and recurrence rates. Given that most studies were not randomized, this may be relevant because baseline characteristics that may have influenced the response to benzodiazepine acute therapy (such as underlying seizure etiology and subsequent different concomitant antiepileptic treatments) may have been unequally distributed. These concerns are reinforced by the ROB assessment of the nonrandomized studies, which attests to some concerns with most of the included studies, especially regarding the outcome assessments.

Regarding reported side effects, in-MDZ administration appeared safe and associated mainly with only moderate accompanying symptoms, such as dizziness, sedation, and local irritation. The latter seems specific to the intranasal route of administration but may not be necessarily midazolam-related, although comparisons with other intranasally administered benzodiazepines were scarce (with only one study investigating intranasal diazepam administration [22]). This is supported by the latest results of the randomized IN-MIDAZ study (published after the screening period) [14]. The study investigated the efficacy and safety of in-MDZ and intramuscularly administered midazolam for the termination of epileptic seizures (clinical and electroencephalographic) in 130 study participants (both pediatric and adult patients). Local side effects were seen more frequently in the intranasal group, with hypotension occurring more commonly as a severe adverse effect in the intramuscular group. In terms of efficacy, there was a significant advantage for intramuscular application, although the mean clinical seizure termination time was still less than 2 min (53.9 ± 25.8 s vs. 104.3 ± 66.4 s; p = 0.002). The authors concluded that in-MDZ represents a useful option for seizure interruption.

The extent to which a dose dependence (i.e., 5- vs. 10-mg in-MDZ single dose) of the reported side effects may exist cannot be conclusively assessed. There is weak evidence that higher doses may be associated with increased side effects, such as drowsiness [25]. However, the study suggesting this did not report excess respiratory depression [25]. Of the three studies that recorded respiratory difficulties, one showed a decline in oxygen saturation (< 90%; without necessary intubation in any patient) [24], whereas the other two reported no recorded respiratory difficulties [17, 27]. It should be noted that these three studies were retrospective and thus without a placebo control. Respiratory failure is not always related to the use of benzodiazepines in seizing patients, as respiratory depression and failure have been reported as leading complications (80%) in patients with convulsive status epilepticus [34] and are independent predictors of death in this context [35]. Dizziness and sedation have been described previously in association with other midazolam administration routes and are also described for other benzodiazepines [4, 36, 37]. Yet in the studies included in this systematic review, there were no reports of adverse effects under alternative benzodiazepine treatments in most cases.

Limitations include the following:

  1. 1.

    That most included studies were nonrandomized. This results in an inherent risk of systematic bias, especially for retrospective designs (e.g., recall bias in retrospective questioning about side effects).

  2. 2.

    The high heterogeneity of the studies considered for this systematic review. This applies, for instance, to the age of the patients included in the study. The age cutoff (pediatric versus adult) was not always clearly stated, and outcomes were not always stratified by age group.

  3. 3.

    The different definitions of efficacy outcomes used.

  4. 4.

    The largely missing blinding for therapy allocation is relevant in the context of adverse event reporting, which possibly introduced a bias in the reported frequency of adverse events.

Strengths of this review are the following:

  1. 1.

    The consideration of control or comparative interventions consisting of the administration of other benzodiazepines or midazolam with other administration routes, which places the efficacy and safety assessment of in-MDZ in the context of other commonly used acute therapies and thus contributes to better comparability.

  2. 2.

    That this review, despite the aforementioned heterogeneity of the included studies, provides the most comprehensive overview of the available evidence on the efficacy and tolerability of in-MDZ for antiseizure treatment in adults to date. Given the limited data available on this topic, we consider the investigation of in-MDZ application in the context of either epilepsy, epileptic seizure(s), or status epilepticus to be justified despite differences in prognosis.

Conclusions

In summary, this systematic review provides evidence for the safe and effective use of in-MDZ against epileptic seizures in adults. However, a high ROB concerning outcome measurement was noted in 8 of 12 included studies, which needs to be considered when interpreting the efficacy of in-MDZ compared to other benzodiazepines or routes of administration. In this regard, this study provides supportive evidence for current clinical practice and calls for further prospective studies with larger sample sizes and a lower ROB.