|A comparison of lorazepam and diazepam as initial therapy in convulsive status epilepticus
|Cock H R, Schapira A H
This is a critical abstract of an economic evaluation that meets the criteria for inclusion on NHS EED. Each abstract contains a brief summary of the methods, the results and conclusions followed by a detailed critical assessment on the reliability of the study and the conclusions drawn.
Lorazepam was compared with diazepam as first-line treatment for convulsive status epilepticus (CSE). The dose of lorazepam was 4 mg intravenously (i.v.), repeated up to 2 times. The dose of diazepam was 10 mg i.v., repeated up to 3 times.
Economic study type
The study population comprised patients aged over 16 years, who were admitted to the hospital with primary or secondary codes "epilepsy", "fits" or "status epilepticus". Patients who were subsequently considered to be suffering from non-epileptic attacks were excluded from the study.
The setting was secondary care. The economic study was carried out in London, UK.
Dates to which data relate
Cases of CSE were retrospectively identified over two 18-month periods (1 October 95 - 31 March 97 and 1 June 97 to 3 November 98), either side of the introduction of a new treatment protocol (May 1997), which recommended using lorazepam in place of diazepam. The resource use data were collected during the same time. The price year was not reported.
Source of effectiveness data
The evidence for the final outcomes was derived from a single study.
Link between effectiveness and cost data
The costing was undertaken retrospectively on the same sample of patients as that used in the effectiveness study.
No sample size was determined in the planning phase of the study. In addition, no power calculations were performed retrospectively. A total of 720 epilepsy coded episodes were identified, of which 590 (82%) were retrieved. The other 130 (18%) were excluded due to missing notes. From the examined records, 90 episodes of CSE were identified during the study period, 37 with established CSE and 53 with premonitory CSE. In 13 (all premonitory CSE) of these 90 episodes no benzodiazepine was given, while a further 5 did not receive it as part of the initial treatment. Thus, the lorazepam and diazepam outcomes were compared on the basis of the remaining 72 episodes.
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). From the 17 patients (10 males and 7 females) receiving it as first dose, 6 were diagnosed with established CSE and 11 with premonitory CSE. The mean age of this group was 46.9 years (range: 17 - 84). In the entire study period, 96 doses of diazepam were given to 56 patients (55 as first dose, 31 as second and 10 as third, which seem to have included the 9 patients who also received lorazepam as first treatment). From the 55 patients (34 males and 21 females) receiving diazepam as first benzodiazepine treatment, 31 were diagnosed with established CSE and 24 with premonitory CSE. The mean age of this group was 51.8 years (range: 20 - 88).
The study was a single-centred, case-control study carried at the Royal Free and University College Medical School, London).
Analysis of effectiveness
All of the patients included in the study were accounted for in the analysis. The primary health outcomes used in the analysis were:
the number of treatment successes after initial benzodiazepine treatment;
the number of complications, including sedation and hypotension); and
the number of patients requiring second-line treatment (given where initial benzodiazepine therapy failed to achieve control in established cases).
Treatment success was defined as seizure cessation without recurrence over 12 hours. The number of complications or clinical differences was obtained from case notes. The treatment groups were compared using analysis of variance and Tukey post hoc analysis. They were shown to be comparable in terms of their clinical characteristics.
When lorazepam was the first benzodiazepine give, treatment success was significantly more likely than for diazepam (9/17 doses lorazepam versus 14/55 doses diazepam), (p=0.042).
There was no difference in the likelihood of either drug terminating the seizures (11/17 doses lorazepam versus 41/45 doses diazepam). However, following initial seizure control, seizure recurrence within the subsequent 12 hours was significantly less likely to occur after lorazepam (2/11 doses versus 31/45 doses diazepam), (p=0.005, Pearson's).
There were no significant differences between the treatment groups in the occurrence of documented complications, such as hypotension, sedation or respiratory depression, following benzodiazepine administration.
Four patients, two treated with diazepam and two with lorazepam, all with established CSE, died within one month of presentation.
Second-line therapy with phenobarbitone, phenytoin or other, did not differ significantly between the two groups.
In this study, i.v. lorazepam was associated with a significantly better seizure outcome than diazepam, with no associated increase in adverse events. Even if both drugs were equally effective at terminating seizures, diazepam was associated with significantly more seizure recurrences requiring repeated doses.
Measure of benefits used in the economic analysis
The measure of health benefit used in the economic analysis was the number of treatment successes. This was determined as seizure cessation without recurrence over 12 hours.
The resource quantities and the unit costs were reported separately. The direct costs included were those of the hospital. Only the hospital costs of lorazepam and diazepam were included in the analysis. The unit costs of these were estimated from the authors' setting. Since all of the costs were incurred during a short time, discounting was unnecessary and was not conducted. The study only reported the unit cost for the two drugs. The price year was not stated.
Statistical analysis of costs
The costs were treated as point estimates (i.e. the data were deterministic).
No indirect costs were included in the analysis.
No sensitivity analyses were performed.
Estimated benefits used in the economic analysis
The estimated benefits used in the economic analysis were 9 successes for lorazepam and 14 successes for diazepam.
The cost per 4 mg of lorazepam was 0.78, compared to a cost per 10 mg of diazepam of 0.45.
Synthesis of costs and benefits
The estimated benefits and costs were combined as the cost per successful outcome. The cost per success was 1.46 with diazepam and 1.47 with lorazepam.
Intravenous (i.v.) lorazepam was associated with a significantly better seizure outcome than diazepam, with no associated increase in cost or adverse events.
CRD COMMENTARY - Selection of comparators
The use of diazepam as the comparator was justified. You should decide if this is a widely used health technology in your own setting.
Validity of estimate of measure of effectiveness
The analysis used a case-control study. The authors acknowledged that this type of study design has the potential for introducing bias, and hence has the potential to affect their results. Further, the cases were retrospectively identified over two different time frames, either side of a new treatment protocol recommending lorazepam, with virtually all doses of this drug being prescribed in the second time period. This fact may have confounded the results, as improvements in health technologies and health care management could have resulted in the benefits seen in the use of lorazepam.
The authors also pointed out that, in attempting to retrieve notes up to one year after the data collection ended, 18% of the notes were never examined. Thus, the possibility that this had introduced some bias into the study could not be ruled out. However, the authors also mentioned that, out of the 590 examined, only 15% included a CSE episode. Hence, it was very likely that the majority of missing notes was not relevant, and missing notes were equally distributed among the two study periods.
The use of a randomised controlled trial would have been a much better option since, if properly conducted, this design tends to reduce the potential for confounding and bias to affect the results. The patient groups were shown to be comparable in terms of their clinical characteristics. Appropriate statistical techniques were used to test for statistically significant differences between the two groups.
Validity of estimate of measure of benefit
The estimation of benefits was obtained directly from the effectiveness analysis. This choice of estimate was justified.
Validity of estimate of costs
Since only the costs of the two drugs were included in the analysis, not all of the relevant categories of costs were included. For example, the costs of hospitalisation, other drugs for recurrent CSE episodes, and staff were not incorporated. The non inclusion of such costs may have led to an underestimation of the cost-effectiveness of lorazepam, as it was found to be more effective, and would therefore have had fewer complications and lower hospitalisation rates. The costs and the quantities were reported separately, hence enhancing generalisability to other settings. Resource use was obtained directly from the examined records. The unit costs were obtained from the authors' setting, but no statistical analysis of prices was performed, hence the uncertain reliability of these estimates. The costs were not discounted, which was appropriate since they were incurred during a short time. The price year was not reported, making the results of any possible reflationary exercises questionable.
The authors made appropriate comparisons with three prospective studies comparing lorazepam with diazepam as first-line therapy in status epilepticus. All of these studies concluded that lorazepam was superior to diazepam, but none of them assessed subsequent seizure recurrence as in the present study. The issue of generalisability to other settings was not explicitly addressed. The authors did not appear to have presented their results selectively. However, it would have been desirable had they clearly reported the health outcomes, and how the results reported in table 2 of the paper translated into the results described in the main text. The authors' conclusions reflected the scope of the analysis as all cases of CSE (i.e. premonitory CSE and established CSE) were considered. The authors reported no further limitations to their study.
Implications of the study
The authors recommended that, where venous access is possible, lorazepam should be the first-line treatment for those with CSE in place of diazepam, both in and out of hospital. The authors also highlighted the need for further investigations to address the potential role of rectal administration of lorazepam.
Cock H R, Schapira A H. A comparison of lorazepam and diazepam as initial therapy in convulsive status epilepticus. QJM 2002; 95(4): 225-231
Other publications of related interest
Leppik IE, Derivan AT, Homan RW, et al. Double-blind study of lorazepam and diazepam in status epilepticus. JAMA 1983;249:1452-4.
Appleton R, Sweeney A, Choonara I, et al. Lorazepam versus diazepam in the acute treatment of epileptic seizures and status epilepticus. Developmental Medicine and Child Neurology 1995;37:682-8.
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.
Subject indexing assigned by NLM
Adolescent; Adult; Aged; Aged, 80 and over; Anticonvulsants /economics /therapeutic use; Chi-Square Distribution; Costs and Cost Analysis; Diazepam /economics /therapeutic use; Female; Humans; Lorazepam /economics /therapeutic use; Male; Middle Aged; Retrospective Studies; Status Epilepticus /drug therapy
Date bibliographic record published
Date abstract record published