Phenytoin and postoperative epilepsy

Transcription

1 J Neurosurg 58: , 1983 Phenytoin and postoperative epilepsy A double-blind study J. BRIAN NORTIt, F.R.A.C.S., ROBERT K. PENHALL, F.R.A.C.P., AttMAD HANIEH, F.R.A.C.S., DEREK B. FREWIN, F.R.A.C.P., AND WILLIAM B. TAYLOR, PH.D. Departments of Neurosurgery and Clinical Pharmacology, Royal Adelaide Hospital, and the Department of Statistics, University of A de[aide, Adelaide, South Australia, Australia ~/ A double-blind trial of phenytoin therapy following craniotomy was performed to test the hypothesis that phenytoin is effective in reducing postoperative epilepsy. A significant reduction in the frequency of epilepsy was observed in the group receiving the active drug up to the 10th postoperative week. Half of the seizures occurred in the first 2 weeks and two-thirds within 1 month of cranial surgery. High rates of epilepsy were observed after surgery in patients with meningioma, metastasis, aneurysm, and head injury. Routine prophylaxis with phenytoin (in a dosage of 5 to 6 mg/kg/day) would seem to be indicated, particularly in high-risk patients and, where possible, this treatment should be started 1 week preoperatively. Seizure control is best when therapeutic levels of phenytoin are maintained. KEY WORDS " epilepsy 9 grand mal seizure phenytoin Dilantin epilepsy prophylaxis postoperative epilepsy seizure prophylaxis C LINICAL practice with anti-epileptic prophylaxis after craniotomy varies widely from unit to unit. Despite a large body of evidence suggesting that epilepsy can be minimized, there is no generally accepted pharmacological regimen for postoperative seizure prophylaxis. Since the introduction ofphenytoin in 1938, it has become the most popular anticonvulsant agent because of its efficacy and relative freedom from serious side effects?,11 However, a prospective and randomized study to establish the efficacy of prophylactic postoperative therapy has not yet been reported. 9,12 In an attempt to overcome this, we have conducted a double-blind trial of phenytoin to investigate whether it is effective in reducing postoperative epilepsy. The preliminary results of our investigation have already been published/ Protocol Clinical Material and Methods Patients undergoing a supratentorial operation (burr hole, craniectomy, or osteoplastic flap procedures) at the Royal Adelaide Hospital from January, 1978, to April, 1980, were randomly allocated to receive either phenytoin or an identical placebo. The trial was approved by the Research and Ethics Committee of the Royal Adelaide Hospital. Patients with preoperative epilepsy and those who had previously received any anti-epileptic drug were excluded. Those with cerebral abscess were also excluded because of a strong tendency to epilepsy. 7 Treatment was started in the recovery room with intravenous phenytoin (250 mg twice a day), and then continued with oral medication (100 mg three times a day initially) for 12 months, unless a complication intervened. Serum drug levels were monitored weekly for inpatients, and every 2 months for outpatients, with an enzyme-multiplied immunoassay technique. The drug levels were known only to a clinical pharmacologist, who modified dosage regimens to maintain the phenytoin level within the therapeutic range of 40 to 80 #mol/liter (10 to 20 mg/liter). Blood samples were also obtained from patients on the placebo for "drug monitoring" and "dosage adjustments." If a patient suffered a seizure, a blood sample was taken for determination of the phenytoin level, and the patient was withdrawn from the trial. Placebo-treated patients 672 J. Neurosurg. / Volume 58 / May, 1983

2 Phenytoin and postoperative seizures who developed a seizure were then treated with phenytoin. The data from 281 patients (140 on active medication and 141 on placebo) are presented. Statistical Analysis The results of the present study have been analyzed in accordance with the statistical methodology proposed by Peto, et al., 1~ which is designed for clinical trials that study the length of time to an untoward event. The method was proposed by a group of British and American statisticians, and was developed out of collaborative clinical trials in leukemia and other cancers. They found that the techniques of life table graphs and log-rank p values are accurate, sensitive, and simple ways to interpret the data from such trials. In this analysis, the "survival time" for each of our patients is given as the number of days from randomization at trial inception to the incidence of the first seizure, or to 365 days in the case of seizure-free patients. Other instances of "survival times" of less than 365 days occurred where the patient died or was withdrawn from the trial because of an adverse drug reaction or because of a change of anticonvulsant treatment. Upon experiencing a seizure, the patient was removed from further consideration. The information thus obtained has been subjected to an analysis by actuarial methods to provide estimates of the "survival rates" together with the extent of exposure to risk. The latter forms the basis of risk assessment, since the risk of a seizure at any specified time depends upon the number of seizures observed and the number of patients exposed to risk. The differences between groups were tested by the corrected chi-square test. A probability of less than 5% between groups was accepted as a significant difference. Results This study comprised 281 patients. A more detailed analysis of their group data is presented in Table 1. A total of 44 seizures were observed in the two groups: 18 occurred in the group on active therapy and 26 in the group treated with placebo. Life tables for the active drug, placebo, and combined groups are shown in Table 2 and Fig. 1. The decrease in patients at risk is not strictly compatible with the seizures observed due to departure of individuals from the trial for causes other than epilepsy (Table 1). Therefore, at each seizure time, the estimated percentage of survivors on a life table basis has been calculated, together with the extent of exposure to risk, using the method suggested by Peto, et al.1~ The time relationship between operation and the occurrence of the first seizure is shown in Fig. 2. Eighteen seizures (45%) were observed in the 1st postoperative week, and 28 (64%) in the 1st month. Calculating the total exposure to risk from the life table, we would expect there to be seizures in TABLE 1 Trial data for active drug- and placebo-treated patients Parameter Phenytoin Placebo Group Group no. of cases male:female ratio 1.37:1 1.47:1 age (mean yrs + SD) 46.7 _ no. completing 12 months deaths lost to follow-up review 2 4 medication discontinued 23 7 medication changed 5 18 focal seizures 9 10 generalized seizures 9 16 drug withdrawn due to 12 3 adverse reaction the active drug group and in the placebo group during the 12-month period of the trial. However, we observed 18 patients with epilepsy on active phenytoin treatment and 26 in the placebo group. The chi-square comparison (1.12) of these two groups indicates that the apparent discrepancy between the groups is not significant. In view of this finding, it is unlikely that phenytoin reduces the incidence of epilepsy over the whole 365-day period when the group receiving active drug therapy is considered in its entirety. In our preliminary communication, 8 attention was drawn to the high incidence of epilepsy in the 1st month after surgery. On this basis and considering the apparent rapid fall in survival within the first 30 days after the operation, as evidenced from the combined life table, the exposure to risk over the separate periods of 0 to 30 days and 31 to 365 days has been obtained (Table 3). A significant difference was observed between the active and placebo groups in the first period but not in the second. A further analysis of the efficacy of treatment in relation to postoperative time was obtained by comparing active drug- and placebo-treated groups in terms of seizure incidence at various stages of the 365 days (Table 4). There was no significant difference between the two groups in the first 6 days, but the difference was significant for the period from 7 to 72 days, when the active drug-treated group had a lower incidence of epilepsy. For the period 72 to 365 days, no significant difference was observed between the active and placebo patients, although fewer seizures were observed in the patients on active therapy. Of the 140 patients receiving the active drug, 98 had their dose increased from the initial value of 100 mg three times daily. Seven patients had their dose decreased and 35 patients were not changed. Compliance with therapy (as judged by serum phenytoin levels within the therapeutic range) was seen in 81% of the patients. In those patients completing 12 months in the study, the dosage of phenytoin required to reach therapeutic levels ranged from 3.5 to 6.9 mg/kg (clothed weight)/day, with the average dose for the J. Neurosurg. / Volume 58 / May,

3 J. B. North, et al. TABLE 2 Life table data for the active drug and placebo groups* Phenytoin Group Placebo Group Time Calculated Calculated (days) Seizures No. at Life Table Exposure to Seizures No. at Life Table Exposure to l I l totalseizures * The time (in days) at which seizures occurred during the trial, the number of patients remaining at each time point and their respective exposure to risk are shown. This analysis has been performed using the method of Peto, et al.'~ "" = risk of seizures. group being 5.35 mg/kg/day. Side effects necessitating withdrawal from the trial were seen in 12 patients: eight developed rashes, one involuntary movements, one hirsutism, one headache, and one discomfort of the face. Of the 141 placebo-treated patients, 83 had TABLE 3 Incidence of seizures in the active drug and placebo groups Factor Time Postop (days) phenytoin group seizures observed 8 10 seizures expected placebo group seizures observed 20 6 seizures expected chi-square p value < NS* * NS = not significant. their "dose" increased, while 58 had no change in "medication." It was also interesting to note that three individuals were withdrawn due to "drug-induced" side effects: a rash in one, dizziness in one, and nausea in one. The frequency of seizures observed in the different conditions for which operations were performed is recorded in Table 5. High rates of epilepsy were observed following operations for aneurysm, head injury, and meningioma. When the active and placebo groups were compared in these "high-risk" categories, there was a significantly lower incidence of epilepsy in the active drug group at 30 days and at 343 days (Table 6). However, when patients suffering from metastasis were included in this analysis, the difference at 343 days between the active drug and placebo groups fell below significance (X 2 = 2.51). The 122 patients who completed 12 months on our trial were followed to determine the incidence of seizures during the period of 12 to 24 months after 674 J. Neurosurg. / Volume 58 / May, 1983

4 Phenytoin and postoperative seizures 100-~ 9 Phenytoin 140 Patients Placebo 141 Patients 9 All Patients ;o DAYS FIG. 1. Graph showing postoperative epilepsy in craniotomy from life table data for the subjects who participated in the trial. The horizontal axis has been terminated at 250 days for reasons of scale. No seizures were observed between Days 133 and 209, and none after Day 343. operation, during which time none were receiving anti-epileptic medication. The first seizure occurred in 11 patients during this second 12-month period; seven of these patients had received active phenytoin during the 12 months of the trial and four had been given placebo. Review of the serum phenytoin levels at the time of (days) served Obepilepsy in the 18 patients on active drug therapy 1 2 reveals an interesting trend. Nine of the 18 subjects 3 2 experienced generalized convulsions and eight of these nine had subtherapeutic levels (< 40/~mol/liter 5 4 or 10 mg/liter) of phenytoin. The remaining nine 7 4 patients had focal seizures, and seven of these patients 9 4 had a therapeutic phenytoin level (40 to 80/zmol/liter 13 5 or 10 to 20 mg/liter) at the time of their seizure Discussion 65 9 The data presented in this report substantiate our preliminary finding s that phenytoin significantly re duces the incidence of postcraniotomy epilepsy in comparison with placebo. This protection seems to start at the 7th postoperative day, and continues until about Day 72. The effect of phenytoin is maximal in the 2nd postoperative week (Table 4). TABLE 4 Progressive analysis of data in the active drug and placebo groups* Phenytoin Group Placebo Group Time Postop Seizures Seizures Seizures Seizures Chi- Ex- Ob- Ex- Square pected served pected p Value NS NS NS < < < < < < < NS NS NS NS NS * No seizures were observed after Day 343. NS = not significant. J. Neurosurg. / Volume 58 / May,

5 J. B. North, et al. TABLE 5 Frequency of postoperative epilepsy after various neurosurgical procedures* Phenytoin Placebo Reason for Total Group Group Group Surgery No. % No. % No. % aneurysm 9/ / /27 22 head injury total 17/ / /55 22 with clot 14/ / /40 23 meningioma 4/ / /9 22 metastasis 5/ /6 50 2/7 29 sellar tumor 1 / / /7 glioma 4/ / /16 6 VA shunt 0/25 0/15 0/10 other 4/ / /10 30 * Figures are given as number of patients with epilepsy/number of patients studied. Follow-up period was 12 months. VA = ventriculoatrial. The duration of protection afforded by phenytoin has relevance to the period for which the drug should be given postoperatively. Our data suggest that a period of 2 to 3 months would seem to be a reasonable recommendation, since three-quarters of the seizures had occurred within 3 months (Fig. 2) and a significant reduction in seizures was observed until the 72 to 82 day mark (Table 4). It is somewhat more difficult to make a recommendation from our data as to whether phenytoin should be given for longer periods in patients having operations for aneurysm, head 18 TABLE 6 Analysis of seizure data in high-risk patients* Factor Time Postop (days) phenytoin group seizures observed 5 8 seizures expected placebo group seizures observed seizures expected chi-square p value < 0.05 < 0.05 * High-risk patients included those with aneurysms, head injury, or meningioma. injury, and meningioma (Table 5). The life table analysis used in this series is not well suited to comparing the relatively small subgroups in the "highrisk" patients, but for these a treatment period of 12 months is probably warranted (Table 6). The incidence of seizures in our series is similar to that recorded by others. 1,3-5 Our rates refer to the 12- month period immediately after operation, but this period varies in other series. In our patients receiving placebo, the rates were remarkably similar (22% to 23%) in those patients suffering from aneurysm, head injury, and meningioma. We were surprised at the low incidence (6%) of seizures in the 17 patients with sellar tumors (mainly chromophobe adenoma and craniopharyngioma). This subfrontal operation usually requires retraction of the frontal lobe, but Foy, et al., 4 also recorded a low incidence with this procedure. We did not observe any seizures in patients after >. 16 a_ iii 14 D. w -i- 12 P D ACTIVE PHENYTOIN -]PLACEBO (R 10 P Z P 8 ~L IL O 6 r m :f 4 Z WEEKS POST-OPERATIVE FIG. 2 The time relationship between operation and occurrence of seizures in the active drug and placebo groups. 676 J. Neurosurg. / Volume 58 / May, 1983

6 Phenytoin and postoperative seizures shunting procedures, although Foy, et al., 4 reported a rate of 22% with this operation which they partly explained by the use of Myodil or a Jacques catheter. It has been suggested that a period of anticonvulsant treatment may retard the "maturation" of epileptic foci and therefore that it would have a beneficial effect in the longer term. In other words, the question of interest is whether giving an anticonvulsant agent for a time after craniotomy will result in a reduction in the expected incidence of epilepsy when the administration of the drug is discontinued? Although our trial was not designed to test this hypothesis, we found no support for it; we found a trend for more seizures during the 2nd postoperative year in those patients who had received active phenytoin during the initial 12 months (seven patients) when compared to the placebo group (four patients). We did not find a significant difference in the 1st postoperative week between the incidence of seizures in the phenytoin-treated patients compared to that in patients receiving placebo. We started phenytoin treatment (intravenously) in the recovery room for all cases including elective operations. Phenytoin usually takes 7 to 10 days to reach steady-state kinetics in view of its relatively long half-life. Therefore it would seem appropriate to start phenytoin treatment at least a week before operation when this is feasible, so as to provide adequate cover at the time of surgery. Such a regimen might then be expected to further reduce the incidence of epilepsy in the 1st postoperative week, and this would significantly increase the protection afforded by phenytoin. The dose range of phenytoin of 3.5 to 7 mg/kg/day, which resulted in therapeutic serum levels, is in keeping with the data of others. 6,9 The average dose of 5.35 mg/kg/day was obtained from 37 patients who had completed 12 months of therapy, and who were weighed clothed. The results of the serum phenytoin levels taken at the time of the seizure suggest that patients maintained in the therapeutic range appear to be protected against generalized epilepsy. However, focal seizures did occur in seven of nine patients with a therapeutic phenytoin level. This may suggest that the hydantoins are of somewhat greater value in the prophylaxis against generalized, rather than focal seizures? Conclusions For patients who are to have elective surgery for tumor, aneurysm (including arteriovenous malformation), or head injury, phenytoin should be started 7 to 10 days preoperatively in oral doses of 5 to 6 mg/ kg body weight/day. This period should allow phenytoin to reach a steady state on the day before operation, when the first blood level should be checked, and dosage adjustments made accordingly. In patients who are to have emergency surgery, it would seem appropriate to use an intravenous loading dose of phenytoin (12 mg/kg) in order to achieve an adequate plasma concentration. Phenytoin is continued post- operatively, by the oral route if possible, otherwise intravenously. Blood levels should be monitored weekly for inpatients and every 2 months for outpatients. The length of phenytoin therapy should be 2 to 3 months, but those patients with aneurysm, arteriovenous malformations, head injury, or meningioma should be treated for 12 months. Other factors such as the patients' occupation, prognosis, and likely drug compliance also need consideration in the selection of patients for anti-epileptic therapy and its duration. Patients having shunt operations would not appear to require phenytoin. Acknowledgments We thank Parke Davis Proprietary, Ltd., for their assistance which included supply of phenytoin and placebo medication. Mrs. Alice Grote gave valuable help throughout the trial. We thank Drs. Carney, Reilly, and Simpson for allowing us to study their patients. The Pharmacy Department at the Royal Adelaide Hospital assisted in the conduct of the trial. References 1. Cabral RJ, King TT, Scott DF: Epilepsy after two different neurosurgical approaches to the treatment of ruptured intracranial aneurysm. J Neurol Neurnsurg Psychiatry 39: , Clark PPR: Prevention of postoperative epilepsy. Lancet 1:650, 1980 (Letter) 3. Fabinyi GCA, Artiola-Fortuny L: Epilepsy after craniotomy for intracranial aneurysm. Lancet 1: , 1980 (Letter) 4. Foy PM, Copeland GP, Shaw MDM: The incidence of postoperative seizures. Acta Nenrochir 55: , Gautier-Smith PC: Parasagittal and Falx Menlngiomas. London: Butterworths, 1970, pp Hooper WD, Tyrer JH, Eadie M J: Plasma diphenylhydantoin levels in Australian adults. Anst NZ J Med 4: , Legg N J, Gupta PC, Scott DF: Epilepsy following cerebral abscess. A clinical and EEG study of 70 patients. Brain 96: , North JB, Penhall RK, Hanieh A, et al: Postoperative epilepsy: a double-blind trial of phenytoin after craniotomy. Lancet 1: , Olanow CW, Finn AL: Phenytoin: pharmacokinetics and clinical therapeutics. Nenrosnrgery 8: , Peto R, Pike MC, Armitage P, et al: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. II. Analysis and examples. Br J Cancer 35:1-39, Reynolds EH: Drug treatment of epilepsy. Lancet 2: , Royal College of Physicians of London: Clinicopathological conference: Diagnostic difficulties in a case of polypharmacy. Br Med J 1: , 1977 (see Thomas PK, p 216) Manuscript received September 4, Address reprint requests to: J. Brian North, F.R.A.C.S., Department of Neurosurgery, Royal Adelaide Hospital, North Terrace, Adelaide, South Australia 5000, Australia. J. Neurosurg. / Volume 58 / May,