Changes in arrhythmia profile and heart rate variability during abrupt withdrawal of antiepileptic drugs. Implications for sudden death

Sudden unexpected death (SUD) has been associated with low or undetectable concentrations of antiepileptic drugs in patients with epilepsy suggesting that a sudden fall in plasma levels of these drugs might be a critical factor for the occurrence of SUD. We studied the changes in arrhythmia profile and heart-rate variability, during abrupt withdrawal of carbamazepine and phenytoin treatment in 10 patients with side effects on these drugs. Continuous ECG recording and daily measurements of drug plasma concentrations were performed from the last day of steady-state treatment and the following 4 days. Three patients had a 10-fold increase in ventricular premature beats. In addition, there was a significant reduction in heart-rate variability, assessed over 24 hours, in both the time (SDNN index, P = 0.03) and frequency domains from days 1-5. In the frequency domain analysis there was a significant reduction in total power (P = 0.01), very-low-frequency power (P = 0.004) and in low-frequency (LF) power (P = 0.01). Similar reductions in heart-rate variability and increases in ventricular automaticity have been associated with increased mortality in other patient groups. Two factors that might contribute to the increased rate of SUD in patients with epilepsy have thus been identified.

Cause-specific mortality in epilepsy: a cohort study of more than 9,000 patients once hospitalized for epilepsy

PURPOSE

We studied overall and cause-specific mortality rates in a large cohort of patients with epilepsy compared with mortality rates of the general population in the same geographic area.

METHODS

The cohort consisted of all patients (N = 9,061) aged > 15 years admitted with a diagnosis of epilepsy for inpatient care in Stockholm during the years 1980-1989. All patients were followed in the National Cause-of-Death Register, from which the causes of death were obtained, until December 31, 1992. Thus, 53,520 person-years were observed. Mortality rates were compared with those of the general population of Stockholm.

RESULTS

We observed 4,001 deaths in the cohort, compared with an expected number of 1,109 deaths in the general population. This yielded a standardized mortality ratio (SMR) of 3.6 [95% confidence interval (CI) 3.5-3.7]. Although highest in the younger patients, the SMR was significantly increased in all age groups. The excess mortality rate in the cohort was due to a wide range of causes of death, including malignant neoplasms [SMR 2.6 (2.4-2.8)], diseases of the circulatory system, [SMR 3.1 (3.0-3.3)], diseases of the respiratory system [SMR 4.0 (3.6-4.5)], diseases of the digestive system [SMR 5.1 (4.4-5.8)], and injuries and poisoning [SMR 5.6 (5.0-6.3)].

CONCLUSIONS

Our results demonstrate that this large subgroup of patients with a diagnosis of epilepsy, once hospitalized and discharged, is a population at risk, with an excess mortality rate due to several different causes.

Lamotrigine in pregnancy and lactation: a case report

PURPOSE

We investigated the effect of pregnancy on the kinetics of lamotrigine (LTG), passage of LTG over the placenta and the excretion of the drug in breast milk.

METHODS

We used high-performance liquid chromatography to determine concentrations of LTG in plasma and in breast milk in a woman who was treated with LTG monotherapy during pregnancy and lactation.

RESULTS

Plasma levels of LTG decreased as pregnancy progressed. The ratio of dose to plasma concentration was 5.8 times higher at delivery and 3.6 times higher in late pregnancy as compared with 5 months postpartum, suggesting enhanced clearance of LTG during pregnancy. The concentration ratio of umbilical cord to mother's plasma was 1.2 indicating extensive passage of LTG over the placenta. The LTG plasma concentration in the newborn was still 48 h after birth similar to the plasma levels of the mother at delivery and in the umbilical cord. The ratio of milk to plasma concentration was 0.6 2 weeks after delivery and the plasma concentration in the breast-fed child was 25% of the mother's plasma levels. No adverse effects were observed in the newborn.

CONCLUSIONS

The kinetics of LTG may be influenced by pregnancy to such a degree that dose adjustments may be indicated. Due to an extensive passage of LTG into breast milk, and a slow elimination in the newborn, LTG concentrations in the nursed infant may reach levels at which pharmacological effects can be expected.

[Epilepsy and pregnancy. Balancing between risks to the mother and child]

Pregnancy may be especially problematic for the epileptic woman, obstetric complications tend to be more frequent, and seizure control and the pharmacokinetics of anticonvulsants may be affected. The risk of seizures is particularly high during labour and delivery-almost 10-fold higher than at other times during pregnancy. As uncontrolled generalised tonic-clonic seizures may be hazardous to both gravida and fetus, the use of anticonvulsants to prevent their occurrence is to be recommended during pregnancy even though all anticonvulsant drugs are potential teratogens. There is a 2- to 3-fold increase in the risk of birth defects in conjunction with fetal exposure to these drugs. Although the mechanisms mediating the teratogenic effects have not been identified, interference with folate metabolism, formation of toxic metabolites and drug-induced fetal hypoxia have been suggested. Despite the incompleteness of our knowledge, some recommendations can be made for the management of pregnant women with epilepsy. Pre-pregnancy counselling is important. Epileptic women contemplating pregnancy need to be informed of the pros and cons, and any major change in anticonvulsant therapy should be made before conception. Monotherapy is preferable, using the drug appropriate to seizure type and epilepsy syndrome at the lowest dosage and serum level that protects against tonic-clonic seizure. The clinical situation needs to be assessed and drug levels need to be monitored more frequently during pregnancy than otherwise. Patients on anticonvulsant treatment during pregnancy also need to be informed of the possibility of antenatal diagnosis. The use of new anticonvulsant drugs during pregnancy represents a particular challenge, since available clinical data may be insufficient to indicate their teratogenic potential Such a drug should be used in pregnancy only if essential to obtain seizure control. Moreover, the outcome of all such pregnancies needs to be carefully documented.

Carbamazepine treatment induces the CYP3A4 catalysed sulphoxidation of omeprazole, but has no or less effect on hydroxylation via CYP2C19

AIMS

Omeprazole has been shown previously to be metabolized by the two cytochrome P450 isoforms CYP2C19 (hydroxylation) and CYP3A4 (sulphoxidation). The objective of this study was to test the inducibility of these enzymes by carbamazepine (CBZ).

METHODS

Omeprazole was given as a single oral dose before and after 3 weeks of treatment of five patients with CBZ (400-600 mg daily).

RESULTS

Mean area under the plasma concentration vs time curve (AUC) between 0 and 8 h after drug intake, decreased by about 40% for omeprazole and its hydroxy metabolite and increased for its sulphone metabolite, but the changes were not statistically significant. The ratio of the AUCs of omeprazole and its sulphone, used as an index of CYP3A4 activity, decreased in all patients (P = 0.052), while there was no change in the omeprazole/hydroxyomeprazole AUC ratio used as an index for CYP2C19 activity. There was a significant decrease in the mean ratio of the AUC of the hydroxy and sulphone metabolites from 2.58 to 0.93 (P = 0.046) with a mean difference of 1.79 (95% CI: 0.07 to 3.50) showing that the induction was more pronounced for CYP3A4 than for CYP2C19.

CONCLUSIONS

CBZ induces CYP3A4, but not, or to a lesser extent, CYP2C19. The induction of the sulphoxidation of omeprazole by CBZ seems to have no major clinical implication.

[Drug analysis during epilepsy treatment is valuable]

Therapeutic drug monitoring (TDM) of antiepileptic agents has been an invaluable aid in efforts to individualise epilepsy treatment since the 1970s. In few series, however, have attempts been made to assess the impact of TDM on the outcome of epilepsy therapy. The value of TDM has often been questioned in recent years, mainly due to the poor results obtained with its inappropriate use. Guidelines for the use of TDM of antiepileptic agents are therefore presented and discussed in this review. The usefulness of TDM varies according to the drug monitored, depending on its pharmacokinetic properties and on whether correlation is known to exist between the serum concentration of the drug and its effect. TDM is a prerequisite for the safe use of phenytoin, for example, but is of very limited value in connection with vigabatrin treatment. Of other novel antiepileptics, TDM would probably be useful in connection with lamotrigine treatment, and perhaps in connection with topiramate or gabapentin treatment. However, the rational use of TDM is hampered by the lack of properly designed concentration-effect studies of the new antiepileptic drugs.

Pharmacokinetics of valproic acid in patients with juvenile myoclonic epilepsy on monotherapy

Two different doses of sodium valproate (VPA), 500 mg b.i.d. and 1,000 mg b.i.d. as enteric-coated tablets, were used in this randomized, double-blind, cross-over monotherapy study of 16 patients with juvenile myoclonic epilepsy. Observation time was 6 months on each dose and included admittance for a 12-h serum concentration-time curve. There was a nonlinear relation between dose and concentration, with a negative deviation from the linear relation for total concentration and a positive deviation for the unbound fraction. Clearance for total concentration increased during high-dose treatment, but intrinsic clearance did not differ between doses. We measured the variation of repeated total and unbound VPA concentrations in up to 6 monthly samples on each dose. The coefficient of variation was 20.7% for total and 29.9% for unbound concentration on the lower dose, and 16.5% for total and 28.5% for unbound concentration on the higher dose. This difference between doses is not statistically significant. There was good correlation between the concentration taken before morning dose and AUC for one dose interval, especially during high dose, but the morning concentration was not the trough level. We conclude that the pharmacokinetic requirements for therapeutic drug monitoring of VPA are established.

Monitoring free extracellular valproic acid by microdialysis in epileptic patients

This report describes the first experiences with monitoring of valproic acid pharmacokinetics in patients with epilepsy by subcutaneous microdialysis. We could demonstrate good correspondence between total concentration in plasma and free concentration in plasma and dialysate. In this study the dialysate concentrations were not corrected for in vivo recovery because the aim was to explore the possibilities of the method in routine clinical work rather than to estimate absolute concentrations in extracellular fluid. In one patient, microdialysis was continued for 3 days without problems. The dialysis probe with a small portable pump could be carried without any notable interference with daily activities of patients in the ward. No side effects were noted. The results encourage further studies along two lines: to investigate the free-concentration effect relation for valproic acid and to monitor drug concentration under nonhospital conditions in order to study the interaction between daily activities and drug concentration.

Phenytoin causes phalangeal hypoplasia in the rabbit fetus at clinically relevant free plasma concentrations

New Zealand White rabbits were treated orally with 0 (controls), 50, 100, or 150 mg/kg phenytoin on days 14-16 of pregnancy. Total and free plasma concentrations of phenytoin were determined in maternal plasma 2, 6, and 24 hr after the final dose in all animals. In addition, after administration of 150 mg/kg maternal plasma concentrations were also determined after 12 and 48 hr, the concentrations in amniotic fluid after 6 hr, and those in fetal tissue 6 and 24 hr after the final treatment. A high degree of plasma protein binding was observed in maternal blood. Treatment with 50 mg/kg resulted in free plasma concentrations of up to 5.0 mumol/l during the 24 hr period following the final dose. Significantly higher free plasma concentrations were observed at the two higher dose levels; up to 9.7 mumol/l at 100 mg/kg and 12.7 mumol/l at 150 mg/kg. Digital hypoplasia was not seen in the control group or the animals treated with 50 mg/kg. However, treatment with 100 mg/kg resulted in hypoplasia in a single or a few digits in approximately 50% of the fetuses, and 150 mg/kg provoked hypoplasia in almost all digits in all fetuses. These results show that even though the doses which caused digital defects in rabbits are much higher than those used therapeutically, the resulting free concentrations of phenytoin are similar to those which are associated with the same type of defects in humans. These data indicate that the pharmacologically induced fetal hypoxia/ischemia and vascular disruption preceding malformations of this type, which were observed in a previous study in rabbits, may be of human relevance.

Electrophysiological evaluation of the sodium-channel blocker carbamazepine in healthy human subjects

Carbamazepine (CBZ) is a sodium-channel blocker used mainly for the treatment of epileptic seizures and neuralgias. It may impair the function of the cardiac conduction system in susceptible patients, but its electrophysiological effects have not been thoroughly assessed in the normal heart, which was the aim of the present study. Ten healthy volunteers, mean age 32 years, underwent two electrophysiological investigations at baseline and three at different dose levels of CBZ. The transesophageal atrial stimulation technique was used to evaluate sinus node function, refractoriness of the atrial myocardium, atrioventricular conduction, and ventricular depolarization and repolarization (as reflected by the QRS, JT, and QT intervals) at spontaneous rhythm and after atrial pacing. Atropine was administered to facilitate 1:1 conduction and assessment of rate-dependent effects. At the highest CBZ dose (800 mg/day), which gave plasma concentrations within the upper therapeutic range, the PQ interval was mildly prolonged (151 vs. 159 msec; p < 0.01). In addition, the shortening of the JT interval normally seen at higher pacing rates was counteracted by high-dose CBZ, as demonstrated by a lower mean slope of the regression line after atropine and CBZ than after atropine alone (0.17 vs. 0.20; p < 0.05). No other effects were detected. At therapeutic levels CBZ had minimal effects on the healthy conduction system, supporting its safe use in the absence of cardiac disease.

Red cell folate levels in pregnant epileptic women

Red cell folate concentrations were determined in 74 epileptic women in early pregnancy in a prospective study. All patients were treated continuously with antiepileptic drugs since before conception. The most frequently used drugs were carbamazepine (n = 39) and phenytoin (n = 26). Sixty-four patients (86%) were on monotherapy. Blood samples for red cell folate and antiepileptic drug concentrations were drawn before folate supplementation. Red cell folate levels in patients, 468 nmol.l-1, did not differ from those in non-epileptic, drug-free, pregnant women, 416 nmol.l-1 or from those in non-pregnant age-matched healthy controls, 412 nmol.l-1. No correlation was found between red cell folate concentrations and doses or plasma levels of phenytoin or carbamazepine.

Single dose pharmacokinetics of carbamazepine-10,11-epoxide in patients on lamotrigine monotherapy

The pharmacokinetics of a single oral dose of carbamazepine-10,11-epoxide (CBZ-E, 100 mg) were compared in 10 patients on chronic monotherapy with lamotrigine (LTG, 200-300 mg/day) and in 10 drug-free healthy control subjects. CBZ-E pharmacokinetic parameters in LTG-treated patients were found to be similar to those observed in controls (half-life: 7.2 +/- 1.6 vs 6.1 +/- 0.9 h; apparent oral clearance: 110.8 +/- 53.1 vs 120.5 +/- 29.9 ml/h/kg; apparent volume of distribution: 1.08 +/- 0.37 vs 1.04 +/- 0.25 l/kg respectively; means +/- s.d.). These data indicate that, contrary to previous suggestions, LTG has no effect on the metabolic disposition of CBZ-E.

Ethosuximide enantiomers in pregnancy and lactation

A gas-chromatographic method was used to determine plasma concentrations of the ethosuximide enantiomers during three pregnancies in two epileptic women, in samples from the umbilical cord in one, and in breast milk in two patients. The ratio of the two enantiomers of ethosuximide ranged from 1.00 to 1.36 and was apparently unaffected by pregnancy, passage into breast milk, and transfer over the placenta. Hence, determination of total ethosuximide concentrations appear to be sufficient for therapeutic drug monitoring during pregnancy and lactation.

Carbamazepine and cigarette smoking induce differentially the metabolism of codeine in man

The inducibility of codeine metabolism by carbamazepine (CBZ) and cigarette smoking has been investigated. A single oral dose of 25 mg of codeine was given to seven epileptic patients before and after 3 weeks' regular CBZ treatment (400-600 mg per day). Codeine was also given to nine volunteers who were heavy smokers (20 cigarettes per day) and to nine non-smokers as controls. All subjects were found to be extensive metabolizers of codeine by O-demethylation. Urine was collected over 8 h following codeine intake. Codeine and the metabolites were analysed with HPLC. CBZ significantly increased the urinary excretion of the N-demethylated metabolite, norcodeine (NC) which led to a significant decrease in the metabolic ratio (MR) for N-demethylation. The O-demethylation was not significantly altered. The excretion of normorphine, an active metabolite formed through both O- and N-demethylation of codeine increased by almost three-fold after CBZ treatment. Contrary to CBZ treatment, cigarette smoking slightly but significantly induced the glucuronidation of codeine as shown by a decreased MR for glucuronidation in the smokers, while the O- and N-demethylations were not significantly changed as indicated by the similar MRs in smokers and in non-smokers. These results suggest that CBZ and cigarette smoking selectively induce different metabolizing enzymes. The polymorphic O-demethylation is relatively stable to these factors.

Vertebral arch nonfusion and juvenile myoclonic epilepsy

Neural tube defects (NTD) are known to occur at a higher rate in pregnancies of women with epilepsy. Antiepileptic drugs (AEDs), notably valproate (VPA) and carbamazepine (CBZ), have been identified as risk factors, but a familial aggregation of this condition also occurs in the absence of pharmacologic teratogens. Spina bifida occulta, defined as a nonsymptomatic nonfusion of vertebral arches, has been suggested to be genetically determined, with an increased prevalence in patients with primary generalized epilepsy, and that the presence of this trait in fetal development can be enhanced pharmacologically to produce NTD such as meningomyelocele. In this study, plain abdominal radiographs were obtained from 56 patients with juvenile myoclonic epilepsy (JME) and 56 age- and sex-matched controls. The radiographs were presented in a random order to an unbiased radiologist. No difference in prevalence of vertebral arch nonfusion (VAN) was noted between the two groups. Even if it has no increased frequency in patients with epilepsy, however, VAN is a common radiologic finding, and its relation to symptomatic neural tube defects should be clarified in future studies.

Disposition of carbamazepine and phenytoin in pregnancy

Free and total plasma concentrations of phenytoin (PHT) and carbamazepine (CBZ) and its active metabolite carbamazepine-10, 11-epoxide (CBZ-E) were determined in a prospective study of 86 pregnant epileptic women. The pharmacokinetics of PHT and CBZ during the three trimesters were compared with kinetics at least 10 weeks postpartum. Plasma clearance and unbound CBZ clearance were slightly decreased during the last trimester. Total and free plasma CBZ-E concentrations did not change significantly during pregnancy. Plasma PHT clearance, on the other hand, increased from the first trimester. A less pronounced increase was observed for clearance of unbound PHT; the increase was statistically significant only during the third trimester.

Epilepsy and pregnancy: a prospective study of seizure control in relation to free and total plasma concentrations of carbamazepine and phenytoin

Seizure control and plasma concentrations of antiepileptic drugs (AEDs) were determined in a prospective, population-based study of 93 pregnancies (cases) of 70 patients with epilepsy. Seventy-seven cases were treated with monotherapy, which in 70 cases consisted of carbamazepine (CBZ) or phenytoin (PHT). Dosage was kept constant unless poor seizure control prompted an increase. Plasma concentrations were determined at monthly intervals throughout pregnancy and compared with baseline levels obtained at least 10 weeks postpartum. Both free and total CBZ and PHT concentrations were analyzed. Seizure frequency during pregnancy for the group as a whole was not different as compared with the 9 pregestational months and was unaltered or improved in 85% of cases. Total CBZ concentration was slightly lower during the third trimester as compared with baseline, whereas free concentration was unchanged. In contrast, PHT levels decreased steadily as pregnancy progressed. Total plasma concentration was 39% of baseline during the third trimester, whereas free PHT concentration decreased far less, being 82% of baseline level during the third trimester. No clear-cut relation could be demonstrated between seizure control and plasma concentrations, which may be explained by the limited changes in free AED concentrations and the small number of cases with an increased seizure frequency. Our results indicate that total plasma concentrations may be misleading and that monitoring of free concentrations, in particular of PHT concentrations, may be advantageous during pregnancy.

Serum prolactin response to thyrotropin-releasing hormone during status epilepticus

Determination of serum prolactin can help distinguish between epileptic and pseudo-epileptic attacks since generalized tonic-clonic and complex partial seizures frequently are accompanied by a transient rise in prolactin. In status epilepticus, however, serum prolactin levels are well within the normal range: cellular depletion due to the prolonged seizure activity has been suggested as a mechanism for this finding. The control of prolactin secretion is complex. Among several possible regulators, inhibitory dopamine and stimulatory thyrotropin-releasing hormone (TRH) may take part in the regulation of prolactin levels in connection with epileptic activity. There may be subpopulations of prolactin-producing cells that react differently in response to various regulators. A dopamine receptor blocker given during status epilepticus brings forth a distinct increase in prolactin levels. In order to add to the understanding of prolactin changes in connection with status epilepticus, we injected TRH i.v. during status epilepticus in seven consecutive patients. All patients had prolactin levels within the normal range (< 25 micrograms/l) before injection of TRH which resulted in at least a two-fold increase in prolactin levels. Our results contradict the hypothesis of cellular depletion of prolactin in connection with status epilepticus. The mechanism behind prolactin values within the normal range after prolonged seizure activity remains unknown.

Carbamazepine induced bradycardia--a problem in general or only in susceptible patients? A 24-h long-term electrocardiogram study

Carbamazepine is a first line drug in the treatment of epilepsy and trigeminal neuralgia, but may exert negative chronotropic and dromotropic effects on the cardiac conduction system. Bradyarrhythmias of different types and severity have been described, especially in the elderly, but the prevalence of arrhythmias in a larger group of carbamazepine treated patients is unknown. Forty-eight patients, 40 years of older, on continuous carbamazepine treatment because of various neurologic disorders were investigated by interview, physical examination, 12-lead surface electrocardiogram, and 24-h long-term electrocardiogram recording. The prevalence of bradyarrhythmias was compared with that in an age-stratified reference group. There was no differences between the two groups, either in the number or the duration of pauses or in the type of pauses. In conclusion, carbamazepine does not increase the risk of bradyarrhythmias in the vast majority of patients.

Nonconvulsive status epilepticus in adults: thirty-two consecutive patients from a general hospital population

We studied all adult patients who between 1984 and 1989 were initially diagnosed at our hospital as having nonconvulsive status epilepticus. Thirty-two patients fulfilled the criteria, which included ictal EEG recordings. The annual incidence was 1.5 in 100,000 inhabitants. The median age at onset of status was 51 years. Ten patients had status as their first epileptic manifestation, but most patients had a previous history of epilepsy. Median duration of epilepsy at onset of status was 4 years. Fourteen patients had focal ictal seizure activity on EEG and thus met the criteria for complex partial status. Eighteen patients had generalized seizure activity on EEG, but only 6 of these had a history of absence epilepsy or juvenile myoclonic epilepsy. None had Lennox-Gastaut syndrome. The clinical features of status in the remaining 12 patients were in some respects similar to those of the patients with complex partial status. We hypothesize that the EEG seizure activity in these patients may have been generalized from an initial focus.

Impairment of carbamazepine-10, 11-epoxide elimination by valnoctamide, a valpromide isomer, in healthy subjects

The effect of the valpromide isomer valnoctamide (VCD, 200 mg three times daily for 8 days), an over-the-counter tranquillizer, on the elimination kinetics of a single oral dose of carbamazepine-10, 11-epoxide (CBZ-E, 100 mg) was investigated in healthy subjects. During VCD treatment, the half-life of CBZ-E was prolonged significantly compared with control (19.7 +/- 6.7 h vs 6.9 +/- 2.0 h, means +/- s.d., P less than 0.01), and its oral clearance decreased four-fold (from 109.6 +/- 30.7 to 28.8 +/- 11.1 ml h-1 kg-1, P less than 0.01). These findings indicate that VCM, like valpromide, strongly inhibits epoxide hydrolase in vivo.

Serum prolactin response to metoclopramide during status epilepticus

Transient elevation of serum prolactin frequently follows generalised tonic-clonic and complex partial seizures. However, the levels of prolactin during status epilepticus are not increased above the normal range. Exhaustion of central prolactin supplies has been proposed as a possible mechanism for the absence of prolactin increase during status epilepticus. To test this hypothesis we injected intravenous metoclopramide (10 mg) in eight consecutive patients with status epilepticus. One patient had generalised tonic-clonic status epilepticus. Seven patients had EEG-verified non-convulsive status epilepticus, consisting of one typical absence status, one atypical absence status and five complex partial status epilepticus. Metoclopramide raised the mean (SD) prolactin levels at least five-fold in all patients, from 5.8 (8.0) micrograms/l to 87.0 (39.0) micrograms/l, within 60 minutes after the injection. Thus the mechanism for low prolactin values in status epilepticus is not cellular depletion of stored prolactin, but more likely an altered regulation, presumably induced by prolonged seizure activity.

Cytochrome-P450-dependent hydroxylation in cluster headache

Two isozymes of the cytochrome-P450-dependent drug oxidizing system exhibit polymorphism. Five to 10% of a Caucasian population are deficient in debrisoquine-hydroxylase activity and about 3% in mephenytoin-hydroxylase activity (poor metabolizers). We tested the hypothesis of a possible over-representation of poor metabolizers among patients with cluster headache. The individual metabolic capacity was determined in 30 cluster headache patients after administration of a test dose of 10 mg of debrisoquine and 100 mg of mephenytoin. Two patients (6.7%) were poor metabolizers of debrisoquine and one (3.3%) a poor metabolizer of mephenytoin. This was no different from the rate of poor metabolizers, 7.1% and 3.3% respectively, in a reference panel of healthy Swedish volunteers.