Abstract

Background: Previous trials have suggested lorazepam may be superior to diazepam as first‐line treatment of convulsive status epilepticus (CSE), with improved seizure outcome, and a lower incidence of side‐effects. Many published guidelines however still recommend diazepam.

Aim: To compare the efficacy, safety and cost of lorazepam compared to diazepam, in adults with CSE.

Design: Retrospective case note audit.

Methods: Cases of CSE were retrospectively identified over two 18‐month periods either side of the introduction of a new management protocol in May 1997, in which lorazepam 4 mg i.v. was substituted for diazepam 10 mg i.v. as first‐line treatment for CSE. Diagnostic codes for all admissions and casualty presentations of patients over 16 years of age were examined for primary or secondary codes including ‘epilepsy’, ‘fits’ or ‘status epilepticus’. Medical records and casualty notes were reviewed to identify CSE cases. Treatment groups were compared using ANOVA and a Tukey post hoc analysis. Treatment success was defined as cessation of seizures without recurrence in the subsequent 12 h.

Results: In both premonitory and established CSE, both drugs were equally effective at terminating seizures, but significantly fewer seizure recurrences followed lorazepam, and fewer repeat doses were needed. There were no differences in reported adverse events or in drug costs.

Discussion: We recommend that lorazepam be the first‐line therapy in preference to diazepam in adults with CSE.

Introduction

Convulsive status epilepticus (CSE) is recognized as requiring prompt diagnosis and treatment, as seizure duration is an important prognostic factor.1,2 Traditionally, parenteral benzodiazepines have been used as first‐line therapy, and a number of studies3–6 have suggested that lorazepam is superior to diazepam as first‐line therapy, with improved seizure outcome and less respiratory depression. However many published guidelines still recommend diazepam,7,8 and there are few data about the use of lorazepam in routine clinical practice in adults, and no published cost analysis data. We report a comparison of the efficacy, safety and cost of lorazepam and diazepam, undertaken as part of a London teaching hospital audit on the management of CSE in adults.

Methods

Cases of CSE were retrospectively identified over two 18‐month periods either side of the introduction of a new management protocol (May 1997) (Figure 1). The main change was the recommendation of lorazepam (4 mg i.v., repeated up to 2 times) in place of diazepam (10 mg i.v., repeated up to 3 times) as first‐line treatment for CSE. Phenobarbitone was also recommended in preference to phenytoin, as cardiac monitoring was not easily available on all medical wards, and it was felt this might delay second‐line treatment in some instances. The protocol was endorsed by all the Neurology Consultant staff at the hospital, approved with the hospital pharmacy and local statutory committees, and then widely distributed and publicized during two months prior to May 1997. All relevant senior staff (Medical and Accident & Emergency Consultants) were personally informed by letter, along with a copy of the protocol. All junior staff were given a ‘filofax’‐sized copy for their own use, and it was included in the hospital formulary and junior doctors handbook given to all new staff. The protocol was also presented along with a detailed rationale at the hospital medical grand round, and separately to Accident & Emergency junior doctors. Subsequent to May 1997, new staff continued to receive the same documents, and copies of the protocol were displayed prominently in Accident & Emergency Department. Teaching sessions to junior doctors (Medical and Accident & Emergency SHOs), occurring every 6 months, continued to reinforce the protocol.

The diagnostic codes for all admissions and casualty presentations of patients over 16 years of age admitted between 1.10.95–31.3.97 and 1.6.97–31.11.98 were examined, and patients with primary or secondary codes including ‘epilepsy’, ‘fits’ or ‘status epilepticus’ identified. Medical records and casualty notes of these cases were reviewed by a single assessor (HC), and those with premonitory (p)CSE or established (e)CSE were identified. Premonitory SE (pCSE) was defined as new onset convulsive seizures lasting more than 5 min, two or more seizures without recovery of consciousness in between, or an increased frequency of convulsive seizures in patients with known epilepsy of sufficient magnitude to warrant urgent medical intervention, as judged by the assessing clinician. Established CSE (eCSE) was defined as convulsive seizures, either continuously or recurrently without regaining consciousness, lasting more than 30 min.9,10 Patients subsequently considered to have non‐epileptic attacks were excluded from the study. Demographic, clinical, and treatment details from all identified patients were recorded anonymously on a computer database for subsequent analysis. The treatment groups were compared using ANOVA and a Tukey post hoc analysis. Treatment success was defined as cessation of seizures without recurrence in the subsequent 12 h. The outcomes following diazepam or lorazepam as early therapy were compared using χ2 and a two‐sided Fishers exact test or Pearson's test. Values of p<0.05 were considered significant.

Figure 1. 

Protocol for management of status epilepticus in adults.

Figure 1. 

Protocol for management of status epilepticus in adults.

Results

At total of 720 coded episodes were identified, from which medical records were retrieved for 590 (82%). From the examined records, there were 90 episodes of CSE identified (37 eCSE and 53 pCSE) during the study period. In 13/90 episodes (all pCSE), no benzodiazepine was given, and a further five did not receive either i.v. lorazepam or diazepam as part of their initial treatment (four chlormethiazole, one midazolam). Comparisons between lorazepam and diazepam outcomes have been made based on the remaining 72 episodes (Figure 2). Details of these cases are given in Table 1, subgrouped according to initial benzodiazepine treatment. There were no significant differences in clinical characteristics between the treatment groups.

In the entire study period, 34 doses of lorazepam were given to 26 patients (17 as first dose of benzodiazepine given, 13 as second, and 4 as third), all but one dose being after the new protocol; 96 doses of diazepam were given to 56 patients (55 as first dose, 31 as second and 10 as third); 9 patients received both diazepam and lorazepam during initial benzodiazepine treatment. Second‐line therapy (Table 2) was only given where initial benzodiazepine therapy had failed to achieve control in established cases. This did not differ significantly between the two groups.

Where lorazepam was the first benzodiazepine given, treatment success was significantly more likely than for diazepam (9/17 doses lorazepam vs. 14/55 doses diazepam; p=0.042). There was no difference in the likelihood of either drug terminating the seizures (11/17 lorazepam vs. 41/55 diazepam), but following initial seizure control, seizure recurrence within the subsequent 12 h was significantly less likely to occur after lorazepam (2/11 vs. 31/45 diazepam; p=0.005, Pearson's).

For any individual dose of lorazepam (whether given as first, second or third dose of benzodiazepine, with prior use of diazepam in some cases) the incidence of seizure recurrence was again significantly reduced (7/34 doses vs. 53/96 doses diazepam; p=0.0006). In keeping with this, where a second or third dose of benzodiazpine was required, this occurred significantly less often after lorazepam than after diazepam (8/31 doses vs. 49/86 diazepam; p=0.0034). Separate analysis of eCSE and pCSE cases did not change the conclusions, although numbers were then too small reach significance (data not shown).

There was no significant difference in the occurrence of documented complications such as hypotension, sedation or respiratory depression following benzodiazepine administration between the treatment groups (Table 2). The frequencies of other complications, as documented in the notes during the patients admission, are also given in Table 2. Four patients, two treated with diazepam and two with lorazepam, all with eCSE, died within 1 month of presentation, giving a mortality of 11% in this group. Associated morbidities in these patients included encephalitis (1), cerebrovascular disease (2), and a subdural haematoma (1). Two of the fatalities also had a history of alcohol abuse. All the surviving patients were discharged from hospital back to their prior residence, but further follow‐up data on level of function/disability were not obtainable from this retrospective study.

Prior to the introduction of the new protocol, all but one (treated by HC) of the cases received only diazepam as first‐line treatment. After May 1997, 43% of cases (19/44) received lorazepam as the first hospital‐administered benzodiazepine, in keeping with the protocol. This included 12 patients who had been given diazepam by paramedics/care staff prior to arrival at the hospital. However, the remaining cases were not managed according to the protocol and in these diazepam was the first hospital‐administered benzodiazepine.

The hospital cost of lorazepam was £0.78 for 4 mg, compared to £0.45 for 10 mg for diazepam at the time of the study. However, when used as the first benzodiazepine, because lorazepam was significantly more likely to achieve seizure control, the cost per successful outcome was no different between the two drugs (lorazepam 17 doses at £0.78/dose, 9 successes=£1.47/success; diazepam 55 doses at £0.45/dose, 14 successes=£1.46/success).

Figure 2. 

Case retrieval and classification.

Figure 2. 

Case retrieval and classification.

Table 1 

Clinical details of cases of CSE according to initial benzodiazepine treatment


 
LZP only
 
DZP only
 
Both
 
Total number of episodes 17 46  9 
Number eCSE:pCSE  6:11 25:21  6:3 
Mean age in years (range) 46.9 (17–84) 51.8 (20–88) 51.8 (25–78) 
Sex (M:F) 10:7 30:16  4:5 
Known epileptics (n (%)) 12 (70) 37 (80)  8 (88) 
Identified aetiology/precipitants* (n (%)) 10 (59) 33 (71)  6 (66) 
   Alcohol  5 (29) 18 (39)  3 (33) 
   Poor compliance  3 (18) 13 (28)  4 (44) 
   CNS infection  1 (6)  1 (2)  0 (0) 
   Cerebrovascular disease  0 (0)  2 (4)  0 (0) 
   Other**  4 (23)  8 (17)  1 (11) 
Mean duration prior to first treatment*** (range) 45 (30–90) 59 (15–315) 45 (10–120) 

 
LZP only
 
DZP only
 
Both
 
Total number of episodes 17 46  9 
Number eCSE:pCSE  6:11 25:21  6:3 
Mean age in years (range) 46.9 (17–84) 51.8 (20–88) 51.8 (25–78) 
Sex (M:F) 10:7 30:16  4:5 
Known epileptics (n (%)) 12 (70) 37 (80)  8 (88) 
Identified aetiology/precipitants* (n (%)) 10 (59) 33 (71)  6 (66) 
   Alcohol  5 (29) 18 (39)  3 (33) 
   Poor compliance  3 (18) 13 (28)  4 (44) 
   CNS infection  1 (6)  1 (2)  0 (0) 
   Cerebrovascular disease  0 (0)  2 (4)  0 (0) 
   Other**  4 (23)  8 (17)  1 (11) 
Mean duration prior to first treatment*** (range) 45 (30–90) 59 (15–315) 45 (10–120) 

LZP, lorazepam; DZP, diazepam; *Some patients had more than one identified factor. **Includes systemic infection; subdural haemorrhage; toxic; metabolic; multiple sclerosis. ***eCSE cases only (some DZP patients were treated prior to arrival in A&E).

Table 2 

Outcome following initial benzodiazepine treatment


 
LZP alone
 
DZP alone
 
Both
 
Number of episodes 17 46 
Treatment success* 11 28 
Complications**    
Sedation  5  8 
Hypotension  0  4 
Other  1  3 
Second‐line treatment    
Phenobarbitone  3  1 
Phenytoin  0 11 
Other  3  6 
Second‐line treatment success  5 11 
Required ITU admission  2  3 
Other complications    
Pneumonia  1  2 
Sepsis  0  1 
Other***  2  2 

 
LZP alone
 
DZP alone
 
Both
 
Number of episodes 17 46 
Treatment success* 11 28 
Complications**    
Sedation  5  8 
Hypotension  0  4 
Other  1  3 
Second‐line treatment    
Phenobarbitone  3  1 
Phenytoin  0 11 
Other  3  6 
Second‐line treatment success  5 11 
Required ITU admission  2  3 
Other complications    
Pneumonia  1  2 
Sepsis  0  1 
Other***  2  2 

*As defined in the text, seizure cessation without recurrence over 12 h. **Complications/clinical changes as documented in the notes following benzodiazepine administration. Factors other than the drug (post‐icthal changes, cause of status) may also have contributed. ***Included: respiratory arrest (1), respiratory depression (1), atrial fibrillation (1), gastrointestinal haemorrhage (1), raised intracranial pressure (2). Some patients had more than one complication.

Discussion

In this retrospective study, i.v. lorazepam was associated with significantly better seizure outcome than diazepam, with no associated increase in cost or adverse events. Both drugs were equally effective at terminating seizures, but diazepam was associated with significantly more seizure recurrence requiring repeated doses, in keeping with the known pharmacokinetics. Diazepam is highly lipid‐soluble, and rapidly crosses the blood‐brain barrier. However, it is also rapidly redistributed from into peripheral stores, with a short distribution half‐life. Repeated doses may compensate for this, but can lead to saturation of the peripheral compartments and unpredictable sudden rises in serum levels.8 Lorazepam is less lipid‐soluble with a longer duration of action, and less risk of cardiorespiratory depression, although repeat doses have been associated with tachyphylaxis. However, in practical terms this has little significance, as where two doses of a benzodiazepine have failed to achieve control in CSE, further doses are not indicated and second‐line therapy (e.g. phenobarbitone, phenytoin) should be instituted.

The definition of eCSE used in this study is widely accepted. Our definition of premonitory status includes the more recently proposed operational definition introduced in order that appropriate treatment is instituted without delay.11 We also included patients presenting with an increased frequency or severity of habitual seizures in the pCSE group, as urgent intervention in such patients can prevent the development of eCSE.8

We accept that retrospective case analysis from diagnostic codes is liable to inaccuracies, and confounded further by problems with note retrieval. Of the cases coded as epilepsy, even after repeated attempts to retrieve notes up to 1 year after the data collection period ended, 18% of notes were never examined. This is comparable to retrieval figures in other retrospective studies in the UK12 and abroad.13 We cannot rule out the possibility that this has introduced some bias into the study. However, given that of the 590 notes that were examined, only 90 (15%) included a CSE episode, it is likely that the majority of the missing notes were also not relevant, and missing notes were equally distributed in the two study periods. In addition, we also independently examined casualty records in the study period, and notes for patients identified by this route were all subsequently retrieved. The annual incidence of SE is estimated to be around 20/100 000/year, including non‐convulsive SE, and with the majority of cases in children.14 We identified 90 episodes of convulsive status in adults over a 3‐year period, from a regional population of 500 000, giving an annual adult incidence of 6/100 000, which is in keeping with the expected incidence. The mortality rate (11%) amongst eSE cases is also in keeping with published figures.8 Thus we believe case identification to be as near complete as possible.

Although the lorazepam group included more patients with pCSE (11/17) than the diazepam group (21/46), and the mean time to first treatment was shorter (Table 1, last row), these differences were not significant. The ‘time to first treatment’ data were also in part skewed by a few cases with very prolonged delays before treatment (up to 315 minutes in the diazepam group), whereas in many instances diazepam was actually given earlier than lorazepam, as it is often administered by paramedics prior to arrival in A&E. Thus these factors do not account for the greater success of treatment with lorazepam compared to diazepam, although they may have contributed to it.

There have been three previous prospective studies directly comparing lorazepam with diazepam as first‐line therapy in SE. In the first of these, a randomized double‐blind trial in 81 episodes, lorazepam achieved seizure control in 89% of episodes compared to 76% with diazepam.3 In keeping with our findings, this was largely accounted for by a significant reduction in seizure recurrence following lorazepam, with both drugs equally effective at initial seizure termination. In a later open prospective randomized study in 102 children with SE, a single dose of lorazepam controlled convulsions in 76% of cases, compared to 51% with diazepam.15 This study also documented a significantly lower incidence of respiratory depression (3% lorazepam vs. 15% diazepam), and included rectal administration of the solution lorazepam where venous access was not possible. Finally, in a randomized double‐blind trial in 205 episodes, given intravenous therapy out of hospital by paramedics, seizures had terminated on arrival to A&E in 59% given lorazepam, 43% given diazepam and 21% given placebo. After adjustment for covariates, the odds ratio for termination of seizures was 1.9 in the lorazepam group compared to the diazepam group, (95%CI 0.8–4.4), with no difference in the rates of circulatory or respiratory complications (10% both treatments). The authors in all studies concluded that lorazepam was superior to diazepam with higher rates of seizure termination. None assessed subsequent seizure recurrence, which as supported by known pharmacokinetics and our study data, provides further evidence in favour of lorazepam. It was not possible in our retrospective study to accurately document time from benzodiazepine therapy to initial seizure termination, but this did not differed significantly in the studies described above.

The results from our study support and extend the findings of previous trials, and demonstrate that lorazepam is superior to diazepam not only in the trial setting, but also in routine clinical practice, and at no added overall cost. Despite the introduction of evidence‐based protocols, at least in this hospital, clinicians have been slow to adopt this drug and diazepam is still widely used. There have been no published studies of guideline adherence the management of status epilepticus. However, the problems of guideline implementation in medicine are widely recognized and documented both within hospital16,17 and community practice.18 In addition, guidelines for the management of status epilepticus have been in the literature for many years, yet other retrospective studies have confirmed that management often fails to reflect these guidelines. In a study of 26 cases transferred to a neurological intensive care unit, 68% had had inadequate loading with phenytoin prior to transfer.19 Similarly, in a recent report from the USA, 28% of 184 cases had had delayed or inappropriate initial drug treatment, and even where an early appropriate antiepileptic was given, the dose was inadequate in 76%.20

There have been few systematic studies of factors contributing to poor guideline adherence, but some are recognized. In our study, the fact that diazepam, a more familiar and readily available alternative21 to lorazepam, was always available in Accident & Emergency is a likely contributor, and not one that can be overcome given the numerous other indications for diazepam. Similarly the high turnover of treating (largely junior) medical staff requires very regular reinforcement of guidelines if they are to be maintained. As part of reinforcing adherence, feedback from this audit has already been presented at the hospital. We are continuing to supply to protocol copies to all new medical staff, and to include teaching on the topic in their educational programs, but recognize in light of our experience that this alone is insufficient. The use of guideline‐based ‘care‐management teams’ has been shown to be more efficient than simple guideline dissemination in the management of chronic conditions such as heart failure,17 but this is not practical for emergency situations. We suggest in this scenario, regular education and reinforcement must be directed at the Accident & Emergency staff, and must include the nursing and pharmacy staff, particularly given their lower turnover compared to junior medical staff. In this way it is hoped that even if the new junior doctor asks for diazepam, they might be reminded at the time that there is a more suitable alternative. Were lorazepam to become more routinely used by paramedics prior to arrival in hospital, this might also increase familiarity with the drug among medical staff.

Whether nationally agreed guidelines (such as have been published for the management of CSE in children)22 would improve adherence, is not known. It has been suggested that if guidelines are developed nationally, this leaves hospitals to focus their energies on local adaptation, dissemination, implementation and evaluation.23 We agree that such national guidelines, such have been agreed for the management of other medical emergencies (e.g. cardiac arrest, myocardial infarction) would be beneficial, but further study of methods to improve implementation, and outcome evaluation, is also needed.

We recommend that where venous access is possible, lorazepam should be the first‐line treatment for premonitory and established status epilepticus in place of diazepam, both in and out of hospital. In patients without venous access, further study is needed to address the potential role of rectal administration of lorazepam, although alternatives such as buccal midazolam are also under evaluation and appear promising.24

Address correspondence to Dr H.R. Cock, Department of Clinical and Experimental Epilepsy, Institute of Neurology, Queen Square, London WC1N 3BG. e‐mail: h.cock@ion.ucl.ac.uk

Grateful thanks to the Royal Free Hospital Audit department for assistance with case identification and notes retrieval. Thanks also to Hilary Watts (Statistician, Institute of Neurology) for advice on statistical analysis, and Professor Simon Shorvon (Institute of Neurology) for comments on the manuscript.

References

1
Aicardi J, Chevrie JJ. Convulsive status epilepticus in infants and children.
Epilepsia
 
1970
;
11
:
187
–97.
2
Barois A, Estournet B, Baron S, Levyalcover M. Long‐term outcome in prolonged status epilepticus: review of 29 cases of status epilepticus exceeding 24 hours.
Annales De Pediatrie
 
1985
;
32
:
621
–6.
3
Leppik IE, Derivan AT, Homan RW, Walker J, Ramsay RE, Patrick B. Double‐blind study of lorazepam and diazepam in status epilepticus.
JAMA
 
1983
;
249
:
1452
–4.
4
Appleton R, Sweeney A, Choonara I, Robson J, Molyneux E. Lorazepam versus diazepam in the acute treatment of epileptic seizures and status epilepticus.
Dev Med Child Neurol
 
1995
;
37
:
682
–8.
5
Treiman DM, Meyers PD, Walton NY, Collins JF, Colling C, Rowan AJ, Handforth A, Faught E, Calabrese VP, Uthman BM, Ramsay RE, Mamdani MB. A comparison of four treatments for generalized convulsive status epilepticus. Veterans Affairs Status Epilepticus Cooperative Study Group.
N Engl J Med
 
1998
;
339
:
792
–8.
6
Alldredge BK, Gelb AM, Isaacs SM, Corry MD, Allen F, Ulrich S, Gottwald MD, O'Neil N, Neuhaus JM, Segal MR, Lowenstein DH. A comparison of lorazepam, diazepam and placebo for the treatment of out‐of‐hospital status epilepticus.
N Engl J Med
 
2001
;
345
:
631
–7.
7
Dodson WE, DeLorenzo RJ, Pedley TA, Shinnar S, Treiman DM, Wannamaker BB. Treatment of convulsive status epilepticus: recommendations of the epilepsy foundation of America working group on status epilepticus.
JAMA
 
1993
;
270
:
854
–9.
8
Shorvon SD. Emergency treatment of epilepsy: acute seizures, serial seizure clusters and status epilepticus. In:
Handbook of Epilepsy Treatment
 . 1st edn. Oxford, Blackwell Science,
2000
:
173
–94.
9
Shorvon SD. Clinical forms of status epilepticus. In:
Status Epilepticus—Its Clinical Features and Treatment in Children and in Adults
 . Cambridge, Cambridge University Press,
1994
:
34
–137.
10
Lowenstein DH, Bleck T, Macdonald RL. Its time to revise the definition of status epilepticus.
Epilepsia
 
1999
;
40
:
120
–2.
11
Lowenstein DH. Status epilepticus: an overview of the clinical problem.
Epilepsia
 
1999
;
40
:
S3
–8.
12
Parker L, Cole M, Craft AW, Hey EN. Neonatal vitamin K administration and childhood cancer in the north of England: retrospective case‐control study.
Br Med J
 
1998
;
316
:
189
–93.
13
Brinton LA, Persson I, Boice JD, McLaughlin JK, Fraumeni JF. Breast cancer risk in relation to amount of tissue removed during breast reduction operations in Sweden.
Cancer
 
2001
;
91
:
478
–83.
14
Hesdorffer DC, Logroscino G, Cascino G, Annegers JF, Hauser WA. Incidence of status epilepticus in Rochester, Minnesota, 1965–84.
Neurology
 
1998
;
50
:
735
–41.
15
Appleton R, Sweeney A, Choonara I, Robson J, Molyneux E. Lorazepam versus diazepam in the acute treatment of epileptic seizures and status epiletpicus.
Develop Med Child Neurol
 
1995
;
37
:
682
–8.
16
Marshall C, Ramaswamy P, Bergin FG, Rosenberg IL, Leaper DJ. Evaluation of a protocol for the non‐operative management of perforated peptic ulcer.
Br J Surg
 
1999
;
86
:
131
–4.
17
Costantini O, Huck K, Carlson MD, Boyd K, Buchter CM, Raiz P, Cooper GS. Impact of a guideline‐based disease management team on outcomes of hospitalized patients with congestive heart failure.
Arch Intern Med
 
2001
;
161
:
177
–82.
18
Loeb M, Simor AE, Landry L, McGeer A. Adherence to antibiotic guidelines for pneumonia in chronic‐care facilities in Ontario.
Clin Invest Med
 
2001
;
24
:
304
–10.
19
Walker MC, Howard RS, Smith SJ, Miller DH, Shorvon SD, Hirsch NP. Diagnosis and treatment of status epilepticus on a neurological intensive care unit.
Q J Med
 
1996
;
89
:
913
–20.
20
Cascino GD, Hesdorffer D, Logroscino G, Hauser WA. Treatment of nonfebrile status epilepticus in Rochester, Minnesota, from 1965 through 1984.
Mayo Clin Proc
 
2001
;
76
:
39
–41.
21
Fijn R, Lenderink AW, Egberts ACG, Brouwers JRBJ, De Jong‐Van DenBerg L. Assessment of indicators for hospital drug formulary non‐adherence.
Eur J Clin Pharmacol
 
2001
;
57
:
677
–84.
22
Appleton R, Choonara I, Martland T, Phillips B, Scott R, Whitehouse W. The treatment of convulsive status epilepticus in children.
Arch Dis Childhood
 
2000
;
83
:
415
–19.
23
Renvoize EB, Hampshaw SM, Pinder JM, Ayres P. What are hospitals doing about clinical guidelines?
Quality Health Care
 
1997
;
6
:
187
–91.
24
Scott RC, Besag FMC, Neville BGR. Buccal midazolam and rectal diazepam for treatment of prolonged seizures in childhood and adolescence: a randomised trial.
Lancet
 
1999
;
353
:
623
–6.