FormalPara Key Points

Clonazepam (CLZ) high doses were used for status epilepticus (SE) in younger patients and was more often associated with intubation for airway protection.

When correcting for potential confounders, varying CLZ doses did not alter outcome in SE.

CLZ doses may be individualized in SE depending on the clinical setting.

1 Background

With an annual incidence of 10–40 per 100 000 population, status epilepticus (SE) is the second most frequent neurological emergency, with a major risk of morbidity and mortality [1]. Its mortality ranges from 5% to 45% according to the cause and its management. Morbidity (most frequently cognitive impairment) can reach 25% [2, 3]. Treatment can vary between small doses of benzodiazepine to prolonged therapeutic coma induction.

Benzodiazepines are the well-established first line treatment in status epilepticus. Midazolam, lorazepam, and diazepam were compared in several randomized trials [4,5,6]. Clonazepam (CLZ) has also a favorable pharmacologic profile for SE treatment (intravenous administration and fast action) [7] and is used in many countries in that setting [8]. It is, however, widely acknowledged that benzodiazepines are frequently underdosed, with potential consequences on their efficacy [8, 9].

In some European countries, CLZ is a commonly used first line treatment. There is, however, no comparative study for this medication. General recommendations about CLZ doses are based on studies conducted in the 1970s with heterogeneous populations and with, at times, very limited samples sizes. Altogether, eight uncontrolled studies with a total of 385 patients reported an overall response of 82% for loading doses ranging from 0.5 mg to 8 mg (most frequently 1–4 mg) [10, 11]. The doses were empirically determined on the basis of clinical experience [12,13,14]. In the initial publication, the authors reported SE termination in 38/39 SE episodes, without unfavorable side effects, after intravenous (IV) CLZ (1–8 mg) [15]. Importantly, for such a frequently used medication, the latest European guidelines do not recommend a specific dose for clonazepam in SE [16]. Doses of 0.015 mg/kg have been recommended as loading doses for the first line treatment [1, 17].

The aim of this study was to determine in what proportion the commonly recommended CLZ doses are used in our current practice, as well as the relationship between doses administered, and outcome in terms of evolution toward refractory SE, number of subsequent required treatments, intubation for airway management and SE treatment, and mortality.

2 Methods

We performed a retrospective analysis of our prospective registry, including all adults (>16 years) with SE episodes treated between February 2016 and February 2021 in Lausanne, Switzerland (CHUV Lausanne University Hospital) [18]. Patients were treated by CLZ, as this molecule is historically used and widely available in Switzerland. The registry was approved by our local ethics committee; need for a specific consent was waived as the study only analyzed anonymized data. Patients that explicitly refused the use of all their data for research purposes were excluded. General consent for use of all clinical data for research is a standard procedure in our hospital. Patients with suspected SE were referred for neurological consultation and electroencephalogram (EEG): both procedures were supervised by A.O.R. or J.N., who proceeded to include all patients with confirmed SE (defined as a single seizure lasting more than five minutes, or multiple seizures without return to baseline) into the registry. Non-convulsive SE required EEG confirmation in all cases, using the Salzburg criteria [19]. Refractory status epilepticus was defined as a status epilepticus that continued despite a treatment with one benzodiazepine and one antiseizure medication (ASM) [18]. EEG was available on a daily basis and monitoring was performed when needed. Episodes occurring in patients younger than 16 years old, or post-anoxic SE, were excluded because of significant differences in physiopathology and prognosis.

For each SE episode, patient characteristics (demographics, body weight), SE features (potentially fatal etiology [20]), STESS prognosis score (including age, consciousness before treatment, worst seizure type, occurrence of previous seizures) [21,22,23], survival at hospital discharge, and treatment characteristics (loading doses, number of treatment after CLZ) were prospectively recorded. Younger patients were defined as ≤ 65 years old and older patients as > 65 years old. End of SE was defined as cessation of seizure activity and clinically determined by the neurologic consulting team based on clinical and EEG findings. Intubation necessity was categorized in intubation for airways protection or intubation for SE treatment (coma induction).

We included every SE episode that was treated with CLZ as a first line treatment. When patients received more than one dose, we considered the cumulative dose. We analyzed different outcome parameters: number of treatment lines after the CLZ, proportion of refractory episodes, intubation for airways protection, intubation for SE treatment, and mortality during hospitalization. We also analyzed the treatment delay (if treated after 1 hour). SE onset was established using witnesses report or as last evidence of good health if no witness was available. For patients with more than one episode, only the last one was considered for the analysis of mortality. We did not consider response to CLZ as last treatment before resolution of the SE, as another ASM is almost always administered as a maintenance treatment. We considered that the loading dose was a “high dose” if the dose received was at least 0.015 mg/kg [1]. We also analyzed doses, considering separately ≥ 80% of the defined high dose (0.015 mg/kg), to account for the patient weight approximation in the emergency setting and for dose rounding.

We planned to perform a sensitivity analysis for the patients that refused the use of their data. We considered hypothetical episodes to account for those missing data, challenging the results with these hypothetical cases. We postulated, for instance, that those patients who received CLZ high doses had good outcomes, to challenge findings that suggested no benefit of high doses up to this point. We did not consider mortality in this analysis as it would not be possible to postulate which proportion could be considered as last episode.

Exploratory statistical analysis was performed using univariable analyses to investigate the association between loading doses and clinical response. Chi-squared, Fisher, Mann-Whitney U, and Spearman tests were applied, as required for univariable analyses. A multivariable stepwise backward binary logistic regression was applied for identification of variables, including CLZ dose, associated with the clinical response (as defined before), and adjusting for potential confounders, such as relevant SE outcome predictors [SE severity (STESS), potentially fatal cause and demographics]. Multivariable linear regression was similarly used to analyze CLZ dose as a continuous variable. We calculated odds ratios (ORs) and 95% confidence intervals (CIs). Statistical analyses were performed using SPSS version 23.0 (IBM corp., Armonk, NY).

3 Results

We collected 485 SE episodes in the period of interest, of which 283 involved CLZ as first line therapy. We excluded 17 patients who refused authorization for the use of their data for research. CLZ doses were missing for 15 episodes, which were also excluded. A total of 251 SE episodes were therefore included in our study, with 225 SE considered for mortality analysis (taking into account the last episode per patient). SE characteristics are summarized in Table 1. In the total cohort, the median loading dose of CLZ was 0.010 mg/kg. High doses (≥ 0.015 mg/kg) were administered in 21.9% SE episodes, and 43.8% of patients received doses that were > 80% of 0.015 mg/kg.

Table 1 Characteristics of status epilepticus (SE) episodes of the cohort and differences between episodes in which clonazepam (CLZ) high doses were or were not administered
Full size table

We analyzed all outcome parameters (except mortality) in the whole cohort, using a multivariable analysis of all parameters in Table 1. Comparing SE episodes in which high doses were given with episodes where low doses were given, younger age (OR 1.02 per year, 95% CI 1.01, 1.05, p = 0.008), lesser weight (OR 1.04, 95% CI 1.02, 1.07, p = 0.001), and greater need of intubation for airways protection (OR 3.1, 95% CI 1.36, 6.98, p = 0.007) were the only independent factors associated with high doses.

Distribution of CLZ loading doses according to age is shown in Fig. 1, and distribution according to weight is shown in Fig. 2. When considering CLZ loading dose in mg/kg as a continuous variable, younger age (estimate per year: 7.8 × 10−5, 95% CI 1.8 × 10−6, 1.4 × 10−4, p = 0.012), lesser weight (per kg: 1.1 × 10−4, 95% CI 3.6 × 10−5, 1.8 × 10−4, p = 0.003), early SE treatment (0.002, 95% CI 1.4 × 10−5, 0.004, p = 0.03), intubation for airways protection (0.005, 95% CI 0.002, 0.007, p < 0.001) and intubation for SE treatment (0.006, 95% CI 0.003, 0.009, p < 0.0001) were independently associated with increasing doses. The same variables were associated with CLZ loading dose > 80% of 0.015 mg/kg. Outcome parameters, such as number of treatment lines after the CLZ and proportion of refractory episodes, were not associated with CLZ dose.

Fig. 1
figure 1

Distribution of clonazepam (CLZ) dose (in mg/kg) according to age, dichotomized in 65 years (median) or less versus more than 65 years. Patients being more than 65 years old received lower CLZ doses (p = 0.005, U test). Median is shown with a horizontal black bar in each group

Full size image
Fig. 2
figure 2

Distribution of clonazepam (CLZ) dose (in mg/kg) according to weight, dichotomized in ≤ 70 kg (median) and > 70 kg. Patients weighing 70 kg or more received lower, weight-adjusted CLZ doses (p < 0.00001). Median is shown with a horizontal black bar in each group

Full size image

Analyzing mortality as an outcome (thus only considering the last SE episodes), mortality was not correlated with the dose of CLZ; only STESS (OR 1.74, 95% CI 1.34, 2.26, p < 0.0001) and potentially fatal etiology (OR 4.4, 95% CI 1.7, 11.45, p < 0.002) were independently associated with death during hospitalization.

Performing a sensitivity analysis with 17 hypothetical cases, there was no difference in terms of SE refractoriness between episodes in which high doses were given (38/72, 52.7%) and episodes in which low doses were given (126/196, 64.3%, p = 0.09, chi-squared test). There was also no difference in terms of intubation for airways control (13/72, 18.1% versus 22/196, 12.6%, p = 0.14, chi-squared test) or intubation for SE treatment (10/72, 13.9% versus 22/196, 11%, p = 0.55, chi-squared test).

4 Discussion

Our data confirm that CLZ loading underdosing in SE is common, since the majority of patients received doses clearly below the commonly recommended dose of 0.015mg/kg. Underdosing was more common in elderly and overweighted subjects. This underdosing was, however, not a major determinant of the outcome in SE.

The analysis of all factors associated with CLZ dose showed that generalized convulsive SE and early (< 1 h) treatment were significantly more common in SE treated with high doses. These two factors, features of more severe SE forms, did not prove to be independent predictors when corrected for age. SE semiology and treatment delay are probably partially age related: non-convulsive and late-diagnosed SE are common in elderly patients [24, 25]. Early treatment (< 1 h) was, however, independently associated with increasing CLZ dose when considered as a continuous variable. This may be explained by previous findings, suggesting that beyond a few hours, the prognosis of non-convulsive SE is not determined by its duration [26]. The risks of a late aggressive treatment may thus be less justifiable. Reluctance to administer high doses probably originates in the perceived risk of poor tolerance in this age group; adverse events of benzodiazepines are several fold more common in patients over 65 years old [27,28,29]. CLZ high dose was indeed associated with an increased proportion of intubation for airway protection, as shown previously in our center [30]. Although intubation for SE has recently been suggested to be safe [31], admission in intensive care unit is associated overall with a higher short-term mortality, irrespective of the patient baseline characteristics, in elderly patients [32]. It therefore may be understandable that treating clinicians may be cautious when administering CLZ loading doses.

An increasing CLZ loading dose was not associated with a better outcome in terms of treatment response (number ASMs subsequently required, as well as progression to refractory SE) or mortality. This finding also holds true when correcting for epidemiological factors of poor prognosis and SE severity. If increasing CLZ loading dose is not associated with better outcome, it seems linked with more adverse events. Increasing CLZ loading dose (as a continuous variable) was also associated with an increased risk of subsequent need for therapeutic coma. This association should probably not be interpreted as a higher CLZ loading dose leading to a more refractory SE. It is also unlikely that patients with a more severe SE in the first place received a higher CLZ loading dose on purpose and then required therapeutic coma, as this association was independent from initial SE severity. The mostly likely explanation is that once patients were intubated for airway protection, refractory SE would have been treated at that stage with therapeutic coma [31] to maximize efficacy, rather than merely adding other ASMs. We recorded intubation reasons as mutually exclusive, which may be a simplification. We considered intubation for airway protection when intubation was performed only for that purpose. Should the patient have required therapeutic coma after intubation for airway protection, this was categorized only as intubation for therapeutic coma.

These results contrast with studies correlating SE outcome with lorazepam loading doses and showing a benefit of higher doses in terms of progression towards refractory SE [33, 34], while other studies did not find a difference [35]. Importantly, studies exclusively on CLZ [33, 36] did not find such association. Comparison between CLZ and lorazepam in terms of dosing strategy is probably not fair. Loading doses given in an emergency setting are probably mostly based on packaging of the treatment and weakly adapted according to the patient weight. Cautious clinicians when administering loading doses may routinely prescribe half of the medication ampule. This was indeed the median dose (0.5 mg) in the group that received lower doses, also explaining why patients who were overweight were more frequently underdosed. This strategy would leave the possibility to complete the full dose later if the first dose was tolerated and SE was still not controlled. This titration seems even more common with lorazepam, for which most patients received 2–3 doses [37]. Considering ampule content very pragmatically, lorazepam ampules of 4 mg represent a lower weight-adjusted dose (0.1 mg/kg for a 40 kg patient) than CLZ 1 mg ampules (0.015 mg/kg for a 67 kg patient), making lorazepam underdosing more likely to happen.

There are several evidences that show that generally increasing ASMs dose to the greatest tolerated dose is counterproductive. In SE, there are multiple examples where administering maximal dose and reaching higher drug level brings little benefit [37, 38]. Similar findings arose in chronic epilepsy, where no improved efficacy is found when titrating ASMs beyond a certain dose and plasma level [39, 40]. ASM doses administered in trials (such as the ESETT trial) are maximized to allow a fair comparison between the tested agents, but they are beyond what was needed for the SE treatment [37]. CLZ dose recommendations in SE are not based on trials, but on the experience derived from old, heterogeneous cohort studies. As with other ASMs, recommended CLZ loading dose may be beyond the dose needed to reach full efficacy. Our findings suggest that personalizing (decreasing) CLZ loading dose, especially in older patients, is not associated with loss of efficacy and may decrease the rate of adverse events.

These results should, however, be interpreted in the light of several limitations. The retrospective nature of study is obviously a major limitation, even though data were recorded prospectively. A trial randomizing for different doses would be required to fully assess this issue. The monocentric aspect of this study might hamper generalizability; the profile and outcome of the studied patients correspond, however, to other SE cohorts [41]. We chose to use the number of ASMs and refractoriness (more than two ASMs needed) as surrogate for the response to CLZ; this remains an estimation and cannot account for the possibility of synergistic effects. ASMs received after CLZ were also heterogeneous, but are unlikely to have led to differences in the number of treatments received, as there are no major differences in efficacy between commonly used second-line ASMs [42]. Treatment exposure was not assessed by measure of CLZ plasma levels, which might reflect inter-individual pharmacokinetic variability or drug interactions; CLZ plasma levels were not available at the time of this study in our hospital.

5 Conclusions

Administration rate of CLZ high doses for SE is low, especially in elderly patients. CLZ high doses are associated with a higher rate of intubation for airway protection. Overall dosing strategy is likely to be more dependent on medication packaging than the patient weight. Differing CLZ loading doses did not alter outcome in SE. Our results suggest that CLZ dose in SE should be individualized, depending on the clinical setting.