Brain, Vol. 122, No. 3, 441-448,
March 1999
© 1999 Oxford University Press
Article |
Does withdrawal of different antiepileptic drugs have different effects on seizure recurrence?
Further results from the MRC Antiepileptic Drug Withdrawal Study
.
Correspondence to:
Professor David Chadwick, Department of Neurological Science, The Walton Centre for Neurology and Neurosurgery, Lower Lane, Liverpool L9 7LJ, UK
 |
Abstract |
|---|
 Top Abstract IntroductionMethodResultsDiscussionAppendixReferences |
|---|
One thousand and thirteen patients, in remission of epilepsy for at least 2 years, were randomized to continued therapy or slow withdrawal over 6 months and were followed up for a median period of 5 years. At the time of randomization 83% of patients were receiving monotherapy with carbamazepine (237 patients), phenobarbitone/primidone (72 patients), phenytoin (184 patients) or valproate (228 patients) in low doses, and plasma levels were below the usual optimal range. The most important factor determining seizure recurrence was continued therapy, which was the case for barbiturates, phenytoin and valproate. There was no significant difference for patients taking carbamazepine at randomization, because of a low rate of recurrence in those withdrawing treatment. The difference between carbamazepine and other drugs was not explained by differences in covariate prognostic factors. There was no evidence that withdrawal of phenobarbitone was associated with withdrawal seizures. These data provide unique evidence for the effectiveness of standard antiepileptic drugs as monotherapy. The results for carbamazepine may be open to a number of interpretations.
epilepsy; antiepileptic drugs; drug withdrawal; seizures
|
Notes |
|---|
* For affiliations see the Appendix
|
Introduction |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionAppendixReferences |
|---|
In many conditions the efficacy of a particular intervention or drug treatment can be examined by randomized clinical trials comparing the intervention with no intervention (blinded or unblinded) in a group of patients receiving no other treatment at the time of randomization. This has not been so in epilepsy, one of a number of conditions where it has not been ethically acceptable to randomize patients to a trial where one of the arms involves no active treatment (Rothman and Michels, 1994
). For this reason, efficacy studies of new antiepileptic drugs have usually used add-on, placebo-controlled designs in which patients maintained on, but refractory to, existing drug therapy are randomized to the addition of either placebo or a new potential drug (Commission on Antiepileptic Drugs of the International League against Epilepsy, 1989). The interpretation of such studies is potentially confounded by both pharmacokinetic and pharmacodynamic interactions. A consequence of these difficulties is that new drugs are not immediately licensed by regulatory authorities for administration as monotherapy, since add-on trials clearly do not address the issue of efficacy when a drug is used as the sole treatment. Even standard drugs (carbamazepine, valproate, phenytoin and barbiturates), which are licensed and widely prescribed as monotherapy, have never been assessed in placebo-controlled monotherapy studies.
It is only with the introduction in the last decade of a new generation of antiepileptic drugs, including vigabatrin, lamotrigine, gabapentin and topiramate, that regulatory agencies have had to address the issue of licensing antiepileptic drugs for monotherapy. In the United Kingdom lamotrigine has recently obtained a monotherapy licence as a result of randomized clinical trials in newly diagnosed patients with epilepsy where comparisons were made between lamotrigine and other standard antiepileptic drugs (carbamazepine and phenytoin) (Steiner and Yuen et al., 1994; Brodie et al., 1995). The UK Regulatory Agency appears to have accepted that broadly equivalent efficacy with standard treatments is an acceptable criterion for a monotherapy licence. This has not been accepted, however, by the US Federal Drugs Administration, which has taken the intellectually more rigorous but less pragmatic stance that equivalence may be as easily interpreted as equal ineffectiveness as equal effectiveness (Temple, 1982; Leber, 1989). However, the Federal Drugs Administration has demanded that comparative monotherapy studies comparing one drug with another must show a difference in favour of the novel agent in order for it to be accepted for monotherapy. More general acceptance of the European view might, of course, be aided by any studies of the efficacy of older standard antiepileptic drugs in which clear differences could be demonstrated between the standard drug and no-treatment in-patients with epilepsy. Unfortunately, few such studies exist. Three studies, only one of significant size, have compared antiepileptic drug treatment with no treatment in adults and children presenting with a single spontaneous seizure [Camfield et al., 1989; Chandra, 1992; First Seizure Trial Group (FIR.S.T. Group), 1993]. The First Seizure Study was a pragmatic trial in which any one of a number of drugs could be used when subjects were randomized to drug treatment, so that even here the efficacy of individual drugs in preventing seizure recurrence cannot easily be investigated. Apart from this, an exhaustive search of >350 randomized clinical trials in epilepsy has failed to reveal any that compare monotherapy with a standard antiepileptic drug with no treatment or placebo in patients with a history of seizures or epilepsy (Marson and Chadwick, 1996).
The MRC Antiepileptic Drug Withdrawal Study (1991) compared continued treatment with antiepileptic drugs with the slow withdrawal of these drugs in 1013 randomized patients who had a history of epilepsy and were seizure-free for a period of at least 2 years at the time of randomization. Since 83% of randomized patients were maintained on monotherapy with a variety of antiepileptic drugs, the design of the trial allows a unique examination of continued monotherapy with standard antiepileptic drugs versus drug withdrawal.
|
Method |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionAppendixReferences |
|---|
One hundred and twenty-five clinicians from 45 centres in the UK and Europe recruited 1021 patients with a history of two or more definite seizures and who had been free of all seizures for at least 2 years while taking antiepileptic drugs. The patients were recruited between February 1984 and June 1988. Eight patients were subsequently excluded from follow-up and analysis because of ineligibility. More details of the methods used in the original study have been published (MRC Antiepileptic Drug Withdrawal Study Group, 1991) in an initial report, which deals with follow-up until the first 6 months of 1989. We have subsequently continued follow-up of this cohort to 1991 or beyond (773 patients) and until the end of 1990 (206 patients); only 14 patients were lost prior to 1990. Median follow-up from randomization is 5 years (25th and 75th percentiles, 3.9 and 5.9, respectively), giving a total follow-up of 5000 patient-years. In total, 244 (48%) of the 510 patients randomized to slow withdrawal have now experienced seizures during follow-up compared with 165 (33%) of 503 patients allocated to continued treatment. Antiepileptic drug treatment at the time of randomization is shown in Table 1
.
|
Comparison of the two treatment policies within the first three monotherapy groups appears feasible, and the combination of phenobarbitone and primidone yields a further moderately sized group. The main outcome examined is the time from randomization to first seizure recurrence. Results are presented as actuarial (Kaplan–Meier) survival curves; the influence of prognostic factors was examined using the Cox proportional hazards regression model.
|
Results |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionAppendixReferences |
|---|
As a first step, an examination was undertaken of the changes in drug therapy from the onset of remission to the time of randomization, an interval of at least 2 years. This revealed some changes both in the drugs used and in the doses administered between these two points in time; the results are summarized in Table 2
. Overall, 87% of patients were maintained on the same drug at remission and randomization, and 79% of these were maintained at the same dose; 13% had undergone dose reduction and 8% an increase. For those undergoing drug changes between remission and randomization, and on monotherapy at randomization, 34% had a straight swap of one drug for another, 56% a reduction from two drugs to one, and 10% a reduction from three drugs to one. We have excluded from further study the 102 patients who changed drug between the onset of remission and randomization (as well as the four patients whose drugs at remission were not known), to create a group which had been treated consistently with the same antiepileptic drug over an extended period and for whom differences between the two treatment policies can be directly related to the drug being taken at the time of randomization.
|
The number of patients, their demographic and clinical characteristics, numbers of patients maintained on the same drug between the onset of remission and randomization, and numbers of patients randomized to no withdrawal and slow withdrawal, are shown in Table 3
. Figure 1 shows actuarial seizure-free curves for both randomized groups within the four different drug groups. It can be seen that there are differences between the two treatment policies within each drug group that are of the expected magnitude for all drug groups other than carbamazepine. The hazard ratios for seizure recurrence for the four drug groups, shown in Table 4, indicate that the hazard ratio for carbamazepine is closer to unity than the others over the first 12 months, and lower than unity over the second 12 months.
|
|
|
This unexpected result with carbamazepine, failing to show a statistically significant difference between a policy of continued treatment and withdrawal, by contrast to the other drug groups, was reinforced by an apparent interaction between drug and treatment policy in Cox proportional hazards regression models which included the four drug groups separately [
2(3) = 7.4, P = 0.06] or contrasted carbamazepine with the other three [2(1) = 3.1, P = 0.08]. The source of this interaction merits further investigation.
To look at the differences between the four drug groups more consistently the models above have been extended to include six of the factors previously identified as influencing outcome (MRC Antiepileptic Drug Withdrawal Study Group, 1993): age at randomization (<16, >16 years); history of generalized or secondarily generalized tonic–clonic seizures; history of myoclonic seizures; EEG at randomization (normal, abnormal); period seizure-free; and a history of seizures after the start of treatment. (The number of antiepileptic drugs at randomization was not included in this analysis, since patients were on monotherapy.) The interactions between drug and treatment policy remained whether the drug groups other than carbamazepine were included separately [2(3) = 8.5, P = 0.037] or combined [2(1) = 4.8, P = 0.028].
Factors other than those previously identified as having prognostic importance could influence the outcome differentially for the individual drugs. Table 3
provides a summary of demographic and clinical information collected prior to randomization by treatment policy and drug group. Although there are obvious differences between the different drug groups (e.g. in age at entry, percentages with generalized seizures only and percentages previously attempting withdrawal), there are few obvious imbalances between the two randomized policies within drug groups. Extension of the previous models to adjust for the extensive features presented in Table IV of our initial report (MRC Antiepileptic Drug Withdrawal Study Group, 1991) again failed to remove the interactions [2(3) = 8.2, P = 0.043 and 2(1) = 4.7, P = 0.31]. Addition of a variable to encode dosage changes between remission and randomization (coded 1 for higher dose at remission, 2 for no change in dosage and 3 for higher dosage at randomization) produced no change to these conclusions and only slight evidence that the risk of seizure recurrence increased with increasing difference between dosage at randomization and remission [odds ratio (95% CL) = 1.24 (0.93, 1.65)]. Estimates of actuarial proportions seizure-free by randomized treatment and drug group derived from a Cox model including the six important prognostic factors, treatment policy, drugs and drug x treatment policy interactions showed that the main source of the unexpected finding for carbamazepine appears to be the lower seizure recurrence rate in patients randomized to withdrawal of carbamazepine. This result is essentially confirmed by examination of the Kaplan–Meier curves (Fig. 1) (which avoid the assumption of proportional hazards), where the proportions seizure-free 4 years from randomization in the slow withdrawal group are 54, 61, 49 and 66 for valproate, phenobarbitone, primidone, phenytoin and carbamazepine, respectively, the last value being the same as that in the valproate group randomized to no withdrawal.
The difference between carbamazepine and the other drug groups does not appear to derive from different policies within the slow withdrawal groups: 81% of patients randomized to withdraw carbamazepine actually stopped the drug before seizure recurrence or within 2 years of entry (those remaining seizure-free) compared with 18% who reduced dosage but did not stop, and just 3% who never reduced drugs. These percentages were identical to those in the other four drug groups combined. In patients randomized to continue carbamazepine, 72% did so (up to seizure recurrence or 2 years from entry in those seizure-free), whereas 8% reduced dosage but did not stop, and 19% stopped altogether; these figures compare with 78, 9 and 13% in the other four drug groups combined and suggest possible overestimation of the risk of seizure recurrence in the carbamazepine group compared with the other drug groups.
Repetition of all the analyses reported above (time to seizure recurrence) for time to first generalized (tonic–clonic) seizure yielded similar conclusions. Indeed, in the Cox models the drug group x randomized treatment interactions were yet more statistically significant though based on fewer events [e.g. in models adjusting for the six important prognostic factors, 2(3) = 9.9, P = 0.020 and 2(1) = 6.4, P = 0.012, respectively].
|
Discussion |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionAppendixReferences |
|---|
It is now widely recognized that the majority of patients who develop epilepsy have a mild disorder characterized by relatively few seizures followed by prolonged periods of remission following the introduction of antiepileptic drugs. Thus, Annegers et al. (1979) showed that ~70% of patients were in a remission that had lasted 5 years by some 20 years after the diagnosis and initial treatment of epilepsy, a result supported by Cockerell et al. (1995). Studies of the withdrawal of antiepileptic drugs from such populations have been reviewed recently by Berg and Shinnar (1994). On average, 40–50% of individuals on treatment and in remission of epilepsy will have further seizures on the withdrawal of treatment, indicating that such populations do have a significant continuing predisposition to seizures. The MRC Antiepileptic Drug Withdrawal Study (1991) is the first and largest randomized study to compare the policies of continued treatment with a slow, phased withdrawal of antiepileptic drugs in such populations. Eighty per cent of patients recruited to this study were taking a single antiepileptic drug at randomization, and most were taking low doses and had plasma levels below the usual optimal range on admission to the trial. In spite of this the results clearly demonstrated that even low doses of antiepileptic drugs are the most important single factor affecting the risk of further seizures. This indicates the fallacy of defining a lower limit to `optimal' ranges of plasma levels. Many patients with mild epilepsy respond to plasma levels that are optimal for them but which lie below supposed population-based `optimal' ranges.
It is of interest to examine the outcomes of withdrawal of individual drugs in this study, as they have varying mechanisms of actions, which might affect outcomes differently (White, 1997). Previous studies have either not examined recurrence rates for individual drugs or have lacked the power to detect differences that might exist (Berg and Shinnar, 1994). It has been suggested that barbiturate antiepileptic drugs are more likely to be associated with withdrawal seizures (Buchthal et al., 1968) a hypothesis which we should be in an ideal position to test. It has also been suggested that some drugs may not simply suppress seizures but could also modify the natural history of epilepsy. Thus, in the kindling model of epilepsy, drugs active at voltage-dependent sodium channels (phenytoin and carbamazepine) do not appear to influence the kindling phenomenon in rodents, while drugs active at GABAergic and glutaminergic receptors have this capability (McNamara, 1986). We should note that the study was not designed to address these specific issues, so that examination of individual drug effects should be used only to create hypotheses, not as proof of concept. We are in effect conducting subgroup analyses and making indirect comparisons between drugs, as patients were selected on clinical grounds for the drug they were taking at randomization.
The current analysis of our data shows that the overall results for all patients in the study are applicable to those who at randomization were receiving monotherapy with phenytoin, valproate or the barbiturate antiepileptic drugs. Our data do not support the contention that barbiturates are more likely to be associated with withdrawal seizures. We found no evidence of an increased hazard ratio for seizures in the first 12 months following randomization, and the hazard ratio remains elevated through to 4 years following withdrawal.
No significant differences between randomized policies were found for carbamazepine, a drug which is currently regarded as a first-line treatment for newly diagnosed patients with epilepsy, particularly in those where a clear diagnosis of partial epilepsy can be made (Chadwick, 1997). We have investigated this inconsistency by making adjustments for the major prognostic and other factors influencing the outcome in the withdrawal study, and after adjustment for these it still appears impossible to explain the lack of difference between continued treatment and withdrawal of carbamazepine. While the risk of seizures on continued treatment with carbamazepine is similar to that with other antiepileptic drugs, the risk of seizures on withdrawal of carbamazepine appears to be significantly less than with other antiepileptic drugs.
There are a number of possibilities that could explain such a finding.
- (i) The doses and plasma levels in the carbamazepine group could have been less than the minimal effective dose and plasma level to a greater degree than for other antiepileptic drugs. There is, however, no evidence that this was the case: indeed, only one-third of the carbamazepine-treated patients had plasma levels at entry that were below the usual optimal range, compared with one-half to three-quarters of the patients on other drugs.
(ii) The carbamazepine group could have contained a number of patients in whom carbamazepine would not be an effective drug. It is recognized that carbamazepine may, in some instances, exacerbate idiopathic generalized epilepsies, particularly by increasing absence or myoclonic seizures (Snead and Hosey, 1985
). Inevitably, the retrospective classification of seizures in the withdrawal study may have been inaccurate. A higher proportion of patients taking carbamazepine and phenytoin were classified as having partial seizures (~50%) compared with those taking valproate or barbiturates (~25%). The majority of patients were classified as having tonic–clonic seizures, but in relatively few instances was there certainty whether they were generalized tonic–clonic seizures or secondarily generalized tonic–clonic seizures. We do note that a relatively high proportion of EEGs in carbamazepine monotherapy patients showed generalized abnormalities at the time of randomization. This could have had two possible effects. First, we might have expected to find a higher seizure recurrence rate in patients on carbamazepine randomized to continued therapy. This was not the case. Secondly, the requirement for 2 years' remission on carbamazepine before randomization might have preselected a subgroup of patients with generalized seizures, who have a much better outlook because remission occurred in spite of treatment with an inappropriate drug. This would give rise to a lower relapse rate on carbamazepine withdrawal.
(iii) Failure to detect a difference for carbamazepine could be due to a combination of chance and a relatively low power (type 2 error). The number of patients experiencing seizure recurrence in the carbamazepine, phenobarbitone/primidone, phenytoin and valproate groups was 80, 22, 77 and 96, respectively. The three larger groups consequently have similar power to detect a difference between randomized policies. The 95% confidence intervals of the hazard ratio for carbamazepine do not exclude continued treatment being twice as effective or 15% worse in preventing seizures than withdrawal.
(iv) The risk of recurrence in the group with no withdrawal from carbamazepine may have been overestimated.
(v) Carbamazepine may be less effective in suppressing seizures than the other monotherapy drugs. This seems unlikely; there are now a number of studies, which have compared carbamazepine monotherapy with monotherapy with other antiepileptic drugs in different populations of patients (Mattson et al., 1985
, 1992; Richens et al., 1994; Heller et al., 1995; Kälviäinen et al., 1995; Verity et al., 1995; Chadwick et al., 1996, 1997). No study has shown that carbamazepine is inferior in terms of efficacy to any other antiepileptic drug in populations that include newly diagnosed patients and those failing on other monotherapies. Indeed, the unequivocal findings of superiority of 600 mg/day carbamazepine over vigabatrin and low doses of gabapentin in previously untreated populations of patients provide clear proof of efficacy for the drug (Chadwick et al., 1996, 1997). Furthermore, in populations of patients with epilepsy that was more resistant to treatment, withdrawal of carbamazepine was more likely to be associated with increased seizure frequency than the withdrawal of other drugs (Duncan et al., 1990).
(vi) It is possible that the carbamazepine group contained a larger number than other groups of patients, having a higher rate of spontaneous remission, or that there could be differences in the long-term effects of different antiepileptic drugs. If carbamazepine had an effect in producing a permanent long-term cure (i.e. influencing the long-term prognosis) while the other drugs only suppressed seizures but did not modify long-term prognosis, this could explain our observations. It is generally believed that carbamazepine is ineffective in preventing kindling in rodents compared with valproate and phenobarbitone (McNamara, 1986
), which raises doubts about this latter hypothesis. It should, however, be noted that some reports exist of species-specific differences in kindling and that carbamazepine might be effective in blocking kindling in subhuman primates (Wada, 1977).
What, then, are the implications of this data for studies of monotherapy? We had expected that this analysis of our data would show a consistent difference between continued treatment and slow treatment withdrawal for all the drug groups examined. Had this been the case we could then have strengthened the argument (admittedly by the use of `historical controls') that a demonstration of equivalence, when rigorously defined (Jones et al., 1996) between two antiepileptic drugs (one standard and one novel), could be interpretable as evidence of effectiveness of the new agent in monotherapy. Unfortunately, carbamazepine is likely to be the major comparator used in these studies. Such actively controlled studies have already been undertaken comparing carbamazepine with lamotrigine, gabapentin and vigabatrin (see above). Our failure to find differences in outcome between patients randomized to continuation of or slow withdrawal from carbamazepine in our patients with mild remote epilepsy means that continued questioning of the equivalence issue will be required. Our data strongly support proof of efficacy in suppressing seizures for phenytoin, barbiturates and valproate in populations of people with mild epilepsy (both partial and generalized) in remission. The findings for carbamazepine will require further study to determine a satisfactory explanation.
|
Appendix |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionAppendixReferences |
|---|
Study participants in the UK: Dr G. Barton (Aylesbury); Dr E. Hicks, Dr V. H. Patterson, Dr M. W. Swallow (Belfast); Dr A. Bush, Dr J. A. Finnegan, Dr A. Gupta, Dr A. Sharma (Birmingham); Dr K. T. Thomas (Blackburn); Dr I. T. Ferguson, Dr C. B. Karki, Dr R. M. Walters (Bristol); Dr J. Morrow, Professor A. Richens, Dr S. Wallace, Dr S. J. Wroe (Cardiff); Dr R. Godwin-Austen (Derby); Dr A. Davidson, Dr D. L. W. Davidson, Dr R. Roberts (Dundee); Dr J. K. Brown, Dr R. E. Cull, Dr A. McInnes, Dr P. Sandercock (Edinburgh); Dr C. Gardner-Thorpe (Exeter); Dr W. F. Durward, Dr I. Melville, Dr M. Thomas (Glasgow); Dr M. A. Barrie (Harlow); Dr J. E. Rees, Dr M. Rice-Oxley (Haywards Heath); Dr A. M. Butterfill, Dr N. Fraser (Hereford); Dr M. D. Rawson (Hull); Dr C. H. Hawkes (Ipswich); Dr A. I. Mukhtar (Kettering); Dr E. G. S. Spokes (Leeds); Dr B. Kendall, Dr P. Millac, Dr J. R. Moore, Dr I. F. Pye (Leicester); Dr J. W. Garry (Lincoln); Dr J. Andrews, Dr L. D. Blumhardt, Dr M. Hand, Professor F. Harris, Dr P. MacFarlane, Dr T. McKendrick, Dr N. Marlow, Dr P. Minchom, Dr I. Morgan, Dr J. R. Roberts, Dr L. Rosenbloom, Dr S. R. Sadik, Dr J. Sills, Dr W. B. Spry, Dr P. Stutchfield, Dr A. P. Thomson, Dr A. J. Williams, Dr E. H. Yousif, Dr T. D. Yuille (Liverpool); Dr P. Chesterman, Dr C. R. A. Clarke, Dr F. Clifford Rose, Dr P. Crawford, Dr C. Dellaportas, Dr R. Elwes, Dr T. D. C. Fox, Dr D. Hall, Dr A. Heller, Dr N. Legg, Dr P. Monro, Dr G. D. Perkin, Dr E. H. Reynolds, Dr C. de Silva, Dr T. Steiner, Dr S. Wilson (London); Dr W. J. K. Cumming, Dr R. W. Newton, Dr M. Noronha, Dr D. I. Shepherd (Manchester); Dr P. Newman, Dr M. Saunders (Middlesbrough); Dr D. Bates, Dr N. E. F. Cartlidge, Dr D. Gardner-Medwin, Dr D. W. A. Milligan (Newcastle-upon-Tyne); Dr D. R. Knight (Northampton); Dr R. Greenhall, Dr N. Hyman, Dr E. Spalding (Oxford); Dr P. I. Tomlin (Preston); Dr G. Venables (Sheffield); Dr A. W. Pantlin (South Ockendon); Dr B. Crossley, Dr C. Tyrie (Southampton); Dr H. G. Boddie, Dr R. A. Cooper, Dr S. L. Manawadu, Dr R. P. Murphy (Stoke-on-Trent); Dr A. C. Butler (Stourbridge); Dr P. Cleland (Sunderland); Dr C. A. R. Bainton (Torquay); Dr L. Loizou (Wakefield); Dr G. P. McMullin (Warrington); Dr J. Platt (Whitehaven); Dr R. Corston (Wolverhampton).
Study participants in Europe: Dr B Pedersen (Denmark); Dr O. Dulac (France); Dr M. Conran (Ireland); Dr P. Zagnoni, Dr G. Zaccara, Dr A. van Lierde, Professor F. Viani (Italy); Dr J. Overweg (Netherlands).
|
Acknowledgments |
|---|
The study was supported by a grant from the Medical Research Council.
* For affiliations see the Appendix
|
References |
|---|
|
TopAbstractIntroductionMethodResultsDiscussionAppendixReferences |
|---|
Annegers JF, Hauser WA, Elverback LR. Remission of seizures and relapse in patients with epilepsy. Epilepsia 1979; 20: 729–37.[Web of Science][Medline]
Berg AT, Shinnar S. Relapse following discontinuation of antiepileptic drugs: a meta-analysis [see comments]. Neurology 1994; 44: 601–8. Comment in: ACP J Club 1994; 121 Suppl 2: 53.
Brodie MJ, Richens A, Yuen AW. Double-blind comparison of lamotrigine and carbamazepine in newly diagnosed epilepsy [published erratum appears in Lancet 1995; 345: 662] [see comments]. Lancet 1995; 345: 476–9. Comment in: Lancet 1995; 345: 1300–2.[Web of Science][Medline]
Buchthal F, Svensmark O, Simonsen H. Relation of EEG and seizures to phenobarbital in serum. Arch Neurol 1968; 19: 567–72.
Camfield P, Camfield C, Dooley J, Smith E, Garner B. A randomized study of carbamazepine versus no medication after a first unprovoked seizure in childhood [see comments]. Neurology 1989; 39: 851–2. Comment in: Neurology 1989; 39: 1645–6.
Chadwick D. Rational drug therapy of epilepsy. In: Porter RJ, Chadwick D, editors. The epilepsies 2. Boston: Butterworth-Heinemann; 1997. p. 247–66.
Chadwick DW, Roi L, Kennedy KM. Vigabatrin (Sabril) as a first-line monotherapy in newly diagnosed epilepsy: a double-blind comparison with carbamazepine [abstract]. Epilepsia 1996; 37 Suppl 4: 6.
Chadwick D, Anhut H, Murray G, Greiner M, Alexander J. Gabapentin monotherapy in patients with newly diagnosed epilepsy: results of a double-blind fixed-dose study comparing three dosages of gabapentin and open-label carbamazepine [abstract]. Epilepsia 1997; 38 Suppl 3: 34.
Chandra B. First seizure in adults: to treat or not to treat? [published erratum appears in Clin Neurol Neurosurg 1992; 94: 89] Clinical Neurol Neurosurg 1992; 94 Suppl: S613.
Cockerell OC, Johnson AL, Sander JW, et al. Remission of Epilepsy: Results from the National General Practice Study of Epilepsy. Lancet 1995; 346: 140–4.[Web of Science][Medline]
Commission on Antiepileptic Drugs of the International League against Epilepsy. Guidelines for clinical evaluation of antiepileptic drugs. Epilepsia 1989; 30: 400–8.[Web of Science][Medline]
Duncan JS, Shorvon SD, Trimble MR. Discontinuation of phenytoin, carbamazepine, and valproate in patients with active epilepsy. Epilepsia 1990; 31: 324–33.[Web of Science][Medline]
First Seizure Trial Group (FIR.S.T. Group). Randomized clinical trial on the efficacy of antiepileptic drugs in reducing the risk of relapse after a first unprovoked tonic–clonic seizure [see comments]. Neurology 1993; 43: 478–83. Comment in: Neurology 1994; 44: 781–2.
Heller AJ, Chesterman P, Elwes RD, Crawford P, Chadwick D, Johnson AL, et al. Phenobarbitone, phenytoin, carbamazepine, or sodium valproate for newly diagnosed adult epilepsy: a randomised comparative monotherapy trial. J Neurol Neurosurg Psychiatry 1995; 58: 44–50.
Jones B, Jarvis P, Lewis JA, Ebbutt AF. Trials to assess equivalence: the importance of rigorous methods [see comments] [published erratum appears in BMJ 1996; 313: 550]. BMJ 1996; 313: 36–9. Comment in: BMJ 1996; 313: 3–4.
Kälviäinen R, Äikiä M, Saukkonen AM, Mervaala E, Riekkinen PJ Sr. Vigabatrin vs carbamazepine monotherapy in patients with newly diagnosed epilepsy. A randomized, controlled study [see comments]. Arch Neurol 1995; 52: 989–96. Comment in: Arch Neurol 1995; 53: 477–8.
Leber PD. Hazards of inference: the active control investigation. Epilepsia 1989; 30 Suppl 1: S57–63.
Marson AG, Chadwick DW. How easy are randomized controlled trials in epilepsy to find on Medline? The sensitivity and precision of two Medline searches. Epilepsia 1996; 37: 377–80.[Web of Science][Medline]
Mattson RH, Cramer JA, Collins JF, Smith DB, Delgado-Escueta AV, Browne TR, et al. Comparison of carbamazepine, phenobarbital, phenytoin, and primidone in partial and secondarily generalized tonic–clonic seizures. N Engl J Med 1985; 313: 145–51.[Abstract]
Mattson RH, Cramer JA, Collins JF. A comparison of valproate with carbamazepine for the treatment of complex partial seizures and secondarily generalized tonic–clonic seizures in adults. The Department of Veterans Affairs Epilepsy Cooperative Study No. 264 Group [see comments]. N Engl J Med 1992; 327: 765–71. Comment in: N Engl J Med 1992; 328: 207–9.[Abstract]
McNamara JO. Kindling model of epilepsy. [Review]. Adv Neurol 1986; 44: 303–18.[Medline]
Medical Research Council Antiepileptic Drug Withdrawal Study Group. Prognostic index for recurrence of seizures after remission of epilepsy. BMJ 1993; 306: 1374–8.
Medical Research Council Antiepileptic Drug Withdrawal Study Group. Randomized study of antiepileptic drug withdrawal in patients in remission. Lancet 1991; 337: 1175–80.[Web of Science][Medline]
Richens A, Davidson DL, Cartlidge NE, Easter DJ. A multicentre comparative trial of sodium valproate and carbamazepine in adult onset epilepsy. Adult EPITEG Collaborative Group. J Neurol Neurosurg Psychiatry 1994; 57: 682–7.
Rothman KJ, Michels KB. The continuing unethical use of placebo controls [see comments]. N Engl J Med 1994; 331: 394–8. Comment in: N Engl J Med 1995; 332: 60–2.
Snead OC 3rd, Hosey LC. Exacerbation of seizures in children by carbamazepine. N Engl J Med 1985; 313: 916–21.[Abstract]
Steiner TJ, Yuen AWC. Comparison of lamotrigine and phenytoin monotherapy in newly diagnosed epilepsy [abstract]. Epilepsia 1994; 35 Suppl 8: 31.
Temple R. Government viewpoint on clinical trials. Drug Information J 1982; 1: 10–7.
Verity CM, Hosking G, Easter DJ. A multicentre comparative trial of sodium valproate and carbamazepine in paediatric epilepsy. The Paediatric EPITEG Collaborative Group. Dev Med Child Neurol 1995; 37: 97–108.[Web of Science][Medline]
Wada JA. Pharmacological prophylaxis in the kindling model of epilepsy. [Review]. Arch Neurol 1977; 34: 389–95.
White HS. Mechanisms of antiepileptic drugs. In: Porter RJ, Chadwick D, editors. The epilepsies 2. Boston: Butterworth-Heinemann; 1997. p. 1–30.
Received July 29, 1998. Revised October 21, 1998. Accepted November 9, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S Hadjikoutis and P E M Smith Approach to the patient with epilepsy in the outpatient department Postgrad. Med. J., July 1, 2005; 81(957): 442 - 447. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Kale and E. Perucca Revisiting phenobarbital for epilepsy BMJ, November 20, 2004; 329(7476): 1199 - 1200. [Full Text] [PDF]
|
|
|
|
 |
|
 When and how to stop antiepileptic drugs in adults DTB, June 1, 2003; 41(6): 41 - 43. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. W. Britton Antiepileptic Drug Withdrawal: Literature Review Mayo Clin. Proc., December 1, 2002; 77(12): 1378 - 1388. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||