Epilepsia, 43(2):175–182, 2002
Blackwell Publishing, Inc.
© International League Against Epilepsy
Add-on Phenytoin Fails to Prevent Early Seizures
after Surgery for Supratentorial Brain Tumors:
A Randomized Controlled Study
*Antonio De Santis, *Roberto Villani, *Marco Sinisi, †Nino Stocchetti, and ‡Emilio Perucca
*Institute of Neurosurgery, Policlinico IRCCS, University of Milan; †Neurosciences Intensive Care Division, Policlinico IRCCS,
Milan; and ‡Clinical Pharmacology Unit, Department of Internal Medicine and Therapeutics, University of Pavia, Pavia, Italy
Summary: Purpose: To determine the potential effectiveness
of phenytoin (PHT) in preventing early postoperative seizures
in patients undergoing craniotomy for supratentorial brain tu-
mors.
Methods: Two hundred patients requiring elective cranioto-
my for supratentorial brain tumors were randomized to two
groups of equal size, with a prospective, open-label, controlled
design. One group received PHT (18 mg/kg as an intravenous
intraoperative load, followed by additional daily doses aimed at
maintaining serum PHT concentrations within the 10- to 20-
æg/ml range) for 7 consecutive days. In the other group, PHT
was not administered. More than 90% of patients in both
groups continued to take preexisting anticonvulsant medication
(AEDs) with carbamazepine or phenobarbital throughout the
study. The primary efficacy end point was the number of pa-
tients remaining free from seizures during the 7-day period
after the operation.
Results: Of 100 patients allocated to PHT, 13 experienced
The incidence of seizures in patients who underwent
supratentorial surgery for nontraumatic conditions has
been estimated to be ⱖ15–20% (1,2). This has stimu-
lated research into the potential usefulness of pharmaco-
logic prophylaxis, but despite claims to the contrary from
uncontrolled or retrospective studies (3), prospective
controlled trials failed to provide any evidence that an-
ticonvulsants (AEDs) can prevent the onset of postop-
erative epilepsy (4). As a result, the widespread practice
of administering long-term AED prophylaxis after neu-
rosurgical operations (5) cannot be justified on the basis
of available data (6).
Revision accepted December 11, 2001.
Address correspondence and reprint requests to Dr. E. Perucca at
Clinical Pharmacology Unit, Department of Internal Medicine and
Therapeutics, University of Pavia, Piazza Botta 10, 27100 Pavia, Italy.
E-mail: perucca@unipv.it
175
seizures during the 7-day observation period, compared with 11
of 100 patients in the placebo group (p > 0.05). Most seizures
occurred in the first day after surgery in both groups. There
were no differences between groups in the proportion of pa-
tients experiencing more than one seizure, but there was a trend
for generalized seizures to be more common in PHT-treated
patients than in controls (11 vs. five patients, respectively).
Status epilepticus occurred in one patient in the PHT group and
in two patients in the control group. Of the 13 PHT-treated
seizure patients, 11 had serum PHT concentrations within the
target range, and only two had concentrations below range on
the days their seizures occurred.
Conclusions: PHT, given at dosages producing serum con-
centrations within the target range, failed to prevent early post-
operative seizures in patients treated with concomitant AEDs.
Prophylactic administration of PHT cannot be recommended in
these patients. Key Words: Phenytoin—Seizures—Prophy-
laxis—Craniotomy—Crain tumors—Randomized controlled trial.
The possible value of pharmacologic prophylaxis in
preventing early postoperative seizures (i.e., those occur-
ring within 7 days of intervention) in these patients has
been less extensively investigated, and available evi-
dence is controversial. In one study, a short course of
intravenous phenytoin (PHT) was associated with a
marked reduction in the incidence of seizures during the
first 3 days after the operation, even though the effect
failed to reach statistical significance (7). In another
study, North et al. (8) found PHT effective in reducing
the risk of early seizures, but only seizures occurring
between 7 and 72 days after surgery were significantly
reduced. In a third, more recent investigation, neither
PHT nor carbamazepine (CBZ) was found to protect
against early and late postoperative seizures, even though
dosages might have been suboptimal because of failure
to individualize treatment by monitoring serum drug
concentrations (2). Optimization of dosage based on se-
176 A. DE SANTIS ET AL.
rum drug concentrations may be critical for therapeutic
success (4,8), as shown in a recent study in which a
significant reduction in the incidence of early seizures
could be found in head trauma patients treated with
PHT (9).
The highest incidence of seizures after brain surgery is
observed during the initial 48 h, and these seizures can
have serious adverse consequences because they inter-
fere with critical postoperative assessment of the pa-
tient’s state of consciousness, and they may carry a
significant morbidity and an associated risk of progress-
ing to status epilepticus (4,7). Therefore, any treatment
that could protect against early seizures could have a
favorable impact on overall clinical outcome. Based on
these considerations, a prospective trial was designed to
determine whether PHT, at dosages individualized on the
basis of serum drug concentrations, prevents seizures oc-
curring in the initial week after neurosurgery. PHT was
selected as a study drug for a number of reasons: (a) it is
available as a parenteral formulation; (b) its use in the
management of acute convulsive disorders is widely es-
tablished with respect to both efficacy and safety (10,11);
(c) unlike barbiturates, it does not produce marked inter-
ference with the level of consciousness and with evalu-
ation of the postoperative course (10); (d) it is the only
AED for which an optimal range of serum concentrations
has been defined clearly (8,12,13); (e) it also is the only
AED that has been found to be effective in preventing
early seizures after acute traumatic brain injury (9,14).
To reduce interpretative problems related to heteroge-
neous underlying pathology, the trial was restricted to
patients operated on for supratentorial brain tumors.
PATIENTS AND METHODS
Patients
Patients requiring elective craniotomy for supratento-
rial brain tumors only at the Institute of Neurosurgery of
the University of Milan were eligible for enrolment ac-
cording to the following inclusion criteria: age between
15 and 75 years; availability of a reliable medical his-
tory; and willingness to provide written informed con-
sent. Criteria for exclusion were pregnancy or lactation;
occurrence of seizures during the 7 days before surgery;
recurrent anginal episodes despite adequate therapy;
second- or third-degree heart block or marked bradycar-
dia; myocardial infarction in the previous 3 months; se-
vere hepatic or renal insufficiency; acute intermittent
porphyria; a history of allergic or idiosyncratic reactions
to hydantoin drugs; a history of drug or alcohol abuse;
ongoing PHT treatment with a total serum PHT concen-
tration >5 ␮g/ml (concentrations below this limit were
considered to be negligible for the purpose of the study).
Patients receiving other AEDs were not excluded be-
cause such treatments have not been found to have any
Epilepsia, Vol. 43, No. 2, 2002
impact on the risk of early postoperative or posttraumatic
seizures, and because the potential value of adjunctive
PHT therapy had not been addressed in previous con-
trolled trials (2,4). The option of evaluating separately a
group of patients not comedicated with other AEDs was
ruled out because, among patients referred to the study
site, only very few were not pretreated with either CBZ
or a barbiturate.
Study design and randomization procedures
The study was carried out according to a prospective
open-label randomized controlled design. Patients were
allocated to two groups by using a random number table.
One group was treated with PHT for 7 days according to
the treatment schedule specified later, whereas the other
group (controls) did not receive PHT treatment. Patients
already receiving AED therapy before surgery continued
treatment at unchanged dosages throughout the entire
study period (if necessary, part of the daily dosage on the
day of surgery was given by nasogastric tube).
Phenytoin administration schedule
Phenytoin (Aurantin; Parke-Davis, U.S.A.) was ad-
ministered at dosages designed to maintain total serum
drug concentrations within the optimal range of 10 to 20
␮g/ml (15). On day 1 (day of surgery), after stabilization
of anesthesia, a loading dose of 18 mg/kg was adminis-
tered, dissolved in normal saline, at a rate ⱕ50 mg/min,
under continuous blood pressure and ECG monitoring. If
hypotension or ECG abnormalities or other side effects
developed, the infusion rate was reduced or interrupted
as appropriate. Subsequent doses of PHT were given
daily and adjusted based on serum PHT concentration,
which was determined on days 2 (24 h after the initial
loading), 3, 5, and 7 before the next PHT dose. PHT
dosage on day 2 was 10 mg/kg intravenously for patients
with a serum drug concentration <10 ␮g/ml, 350 mg
orally for patients with a concentration between 10 and
20 ␮g/ml, and 200 mg orally for patients with a concen-
tration between 20 and 30 ␮g/ml (no PHT was given on
day 2 if its serum concentration exceeded 30 ␮g/ml).
Further dosage adjustments, if required, were made dur-
ing the subsequent days based on results of serum con-
centration measurements. Although the study protocol
required measurement of unbound PHT concentrations
when a major alteration in PHT plasma protein binding
was anticipated (e.g., in the presence of severe hypoal-
buminemia or intake of displacing agents such as val-
proic acid), none of the patients fulfilled the criteria for
unbound concentration monitoring.
Evaluation procedures
Before randomization, on the day before surgery (day
0), each patient underwent a medical history and a physi-
cal and neurologic examination, including assessment of
 PHENYTOIN AND EARLY POSTOPERATIVE SEIZURES
potential side effects from any ongoing treatment. For
patients receiving other AEDs, a blood sample was ob-
tained for the determination of the trough serum concen-
tration of these drugs by fluorescence polarization
immunoassay on days 1, 2, and 7. For patients random-
ized to PHT treatment, additional samples for the deter-
mination of serum PHT concentration were collected as
specified in the earlier section. Serum sodium and po-
tassium were determined on at least days 1 and 2. No
other specific laboratory tests in addition to those clini-
cally indicated for preoperative and postoperative evalu-
ation were performed.
All patients were closely observed throughout the
7-day study period, and any adverse events, including
seizures, were recorded in especially designed Case Re-
port Forms.
End points and statistical considerations
The primary efficacy end point was the incidence
of seizures during the 7-day postoperative observa-
tion period. Incidence of individual seizure types, total
number of seizures, occurrence of status epilepticus,
and incidence of adverse effects were secondary end
points.
Assuming an incidence of seizures of 15% in the con-
trol group (based on historic data from comparable popu-
lations operated on at the same study site), an overall
sample size of 356 patients was initially targeted, which
would ensure an 80% probability of identifying a reduc-
tion in the incidence of seizures from 15 to 5%, with a p
value of 0.05 (16). The study protocol required an in-
terim analysis to be performed after completion of 200
patients to reassess power calculations and to determine
whether the study should be terminated with a smaller
sample size. Given that the actually observed incidence
of events (seizures) after completion of the first 200 pa-
tients was 13% in the treatment group and 11% in the
control group, conditional power was subsequently esti-
mated to be equal to 0.269 (17). In other words, assum-
ing that the original hypothesis was true, completing the
initially planned sample size would involve a <27%
probability to detect a significant difference between
groups. Based on this analysis, recruitment was stopped
at 200 patients.
Comparability in baseline demographic and clinical
characteristics of patients randomized to the two groups
was assessed by using the Student’s t test for parametric
measures and the ␹2 test for rates and distributions. The
Student’s t test also was used to compare serum PHT
concentrations between patients with seizures and pa-
tients without seizures. The ␹2 test was used to assess
differences in postoperative seizure rates between PHT-
treated patients and controls. Unless indicated otherwise,
values given in text are means ± SD.
177
RESULTS
Characteristics of patients
Of the 200 patients randomized in the trial, 100 (46
male, 54 female) received PHT and 100 (57 male, 43
female) were assigned to the control group. As shown in
Table 1, the two groups were comparable in demo-
graphic characteristics, type and location of tumor and
baseline clinical characteristics.
Of the patients randomised to PHT, 35 had a history of
seizures preoperatively, compared with 32 in the patients
randomized to the control group. Of these patients, 38
had had only one seizure, 15 had fewer than four seizures
in the previous year, and 14 had more than four seizures.
Seizures were classified as partial in 61 cases (with sec-
ondary generalization in 23) and as primarily generalized
tonic–clonic in six. Overall, preoperative seizures were
recorded in 36% of patients with meningioma, 38% of
those with glioma, and 20% of those with metastatic
tumors.
AEDs were prescribed at baseline in 90 patients in the
PHT group and in 95 patients in the control group. In the
PHT group, 77 patients received CBZ (mostly 600 mg/
day), and 13 received phenobarbital (PB; mostly 150
mg/day), with five patients receiving these agents in
combination. For both of these drugs, mean serum con-
centrations on study days 1, 2, and 7 were similar to
those found on day 1 (6.1 ± 1.6 and 5.5 ± 1.7 ␮g/ml vs,
6.9 ± 2.1 ␮g/ml, respectively, for CBZ, and 20.2 ± 4.7
and 21.8 ± 5.3 ␮g/ml vs, 19.6 ± 5.5 ␮g/ml, respectively,
for PB).
Background AED medication among controls was
similar to that recorded in the PHT group. CBZ (mean
serum concentrations, 7.0 ± 2.2, 7.1 ± 2.0, and 6.9 ± 1.7
␮g/ml on days 1, 2, and 7, respectively) and PB (mean
serum concentrations, 19.5 ± 4.3, 19.4 ± 3.5, and 19.3 ±
3.9 ␮g/ml on days 1, 2, and 7, respectively) were pre-
scribed in 79 and 16 patients, respectively, with four
patients receiving these drugs in combination.
For none of the distributions listed in Table 1 there
was a statistically significant difference between the
PHT group and controls.
Surgical intervention and
postoperative complications
Modalities of surgical intervention were almost super-
imposable in the two groups. The intervention required
incision of the cerebral cortex in 115 patients (56 in the
PHT group vs. 59 in controls), dissection of the cortex in
73 (38 vs. 35, respectively), and a transcallous approach
in 12 (six in each group). Macroscopically, complete
removal of the tumor was possible in 155 cases (74 in the
PHT group vs. 81 in controls), whereas in the remaining
cases, removal was subtotal or partial.
Epilepsia, Vol. 43, No. 2, 2002
178 A. DE SANTIS ET AL.

TABLE 1. Characteristics of the patients randomized in the trial

Phenytoin-treated Controls
group (n ⳱ 100) (n ⳱ 100)
Age (yr, median and range) 59 (18–74) 56 (19–75)
Sex (male/female) 46/54 57/43
Level of consciousness preoperatively (n of patients)
Fully alert 76 80
Psychomotor slowing 22 19
Coma 2 1
Neurologic handicap (n of patients)
Minor (not affecting working activities) 65 64
Moderate (affecting working ability) 32 30
Severe (requiring assistance for daily routines) 3 6
Preoperative CT scan (n of patients)
Enhancement of cerebral mass after injection
of iodinated contrast medium 93 91
Moderate to marked shift of median
brain structures 32 30
Histology of tumor (n of patients)
Glioma 49 46
Meningioma 39 42
Metastatic carcinoma 5 5
Other 7 7
Site of tumor (n of patients)
Frontal 36 34
Temporal 21 26
Parietal 21 20
Sellar 7 10
Occipital lobe 5 2
Deep structures 4 2
Intraventricular 3 3
Tentorial 3 3
History of seizures preoperatively (n of patients) 35 32
Preexisting antiepileptic drug therapy (n of patients) 90 95

CT, computed tomography

Serious postoperative complications, apart from sei-
zures, were recorded in 19 patients who received PHT
and in 14 controls. These included hematoma requiring
reintervention (12 patients taking PHT and nine con-
trols), perilesional edema requiring reintervention (two
patients in each group), and hematoma or edema man-
aged conservatively (five patients taking PHT and three
controls). Two patients died, one in the PHT group as a
result of abdominal complications, and one in the control
group of respiratory insufficiency.
Early postoperative seizures in the two study
groups and relation to serum drug concentration
Of the 200 patients included in the study, 24 (12%)
experienced one or more seizures during the 7-day ob-
servation period. Of these, 13 were treated with PHT,
and 11 were controls.
There were no apparent differences between groups in
the proportion of patients experiencing more than one
seizure and in time of seizure occurrence, with most
patients experiencing seizures within the first 24 h, usu-
ally during waking from anesthesia ⱖ5 h after termina-
tion of the PHT infusion (Table 2). Generalized seizures
tended to be more common in PHT-treated patients than
in controls (11 vs. five, respectively). Generalized con-
vulsive status developed in one patient in the control
group, whereas partial status with motor symptoms oc-
curred in two patients, one in each group. Total number
of seizures was slightly higher in the PHT group (21
compared with 15 among controls).
In 17 cases, seizures were terminated by administering
intravenous diazepam (DZP). Two patients were admin-
istered intramuscular PB, and four (two in each group)
were treated with intravenous PHT.
Mean serum PHT concentrations are summarized in
Table 3. On the assessment days, serum concentrations
were within the target range (10–20 ␮g/ml) in 65 to 85%
of patients. Mean PHT concentrations in patients with
seizures did not differ from those in patients without
seizures (13.8 ± 3.8 ␮g/ml vs. 12.2 ± 4.1␮g/ml, respec-
tively). Among the 13 seizure patients in the PHT-treated
group, 11 had serum PHT concentrations within the tar-
get range, and only two had concentrations below target
on the days when the seizures occurred.
AEDs other than PHT were taken by 11 of 13 seizure
patients in the PHT group and by each of the 11 seizure
patients in the control group. Serum concentrations of
associated AEDs did not differ significantly between pa-
tients with seizures and patients without seizures (data
not shown).

Epilepsia, Vol. 43, No. 2, 2002

 PHENYTOIN AND EARLY POSTOPERATIVE SEIZURES 179

TABLE 2. Occurrence of early postoperative seizures in the alert patients (22.0 vs. 9.6%, respectively), and in those
two study groups with contrast medium–associated lesion enhancement at
Phenytoin-treated Controls preoperative computed tomography (CT) scan than in
group (n ⳱ 100) (n ⳱ 100) those without enhancement (19.0 vs. 11.5%). The degree
Number of patients with seizures of neurologic deficits before operation did not appear to
Any number of seizures 13 11 have any impact on risk of early postoperative seizures.
One seizure only 6 5 In the 33 patients with postoperative complications, the
Two seizures 3 2
Three seizures 3 2 incidence of seizures was only slightly higher than that
Status epilepticus 1 2 recorded in patients without complications (15 vs. 11%).
Timing of first postoperative seizure In particular, none of the postoperative findings, with
(n of patients)
Waking from anesthesia 8 6 special reference to CT imaging, appeared to be useful in
Later on day 1 0 3 predicting the risk for seizures. In none of the observed
Days 2 to 7 5 2 seizure events was a clear-cut metabolic precipitant (e.g.,
Type of postoperative seizures
(n of patients) hypoglycemia, major electrolyte disturbances, or admin-
Simple partial 2 6 istration of proconvulsant drugs) identified.
Partial secondarily generalized 4 3
Generalized without apparent Other adverse events
focal onset 7 2 No life-threatening adverse events that could be as-
Total number of seizures
All seizures 21 15 cribed to PHT were observed during the trial. In 14 pa-
Simple partial 4 9 tients, the initial loading dose was associated with
Partial secondarily generalized 6 4 development of hypotension, which in two cases was
Generalized without apparent
focal onset 11 2 sufficiently severe to require a reduction of the admin-
istered dosage. During continuation of treatment on days
1 to 7, a total of 13 additional adverse events was re-
Incidence of seizures in relation to potential corded. These consisted of severe hypotension (two
risk factors events), mild to moderate hypotension (10 events) and
The frequency of postoperative seizures in relation to mild alteration in the level of consciousness (three
a number of potential risk factors is summarized in Table events). All adverse events were reversible without ac-
4. Because of the small number of patients in each sub- tion, except for a reduction in PHT infusion rate (or
group, only a descriptive analysis is presented. The risk discontinuation of infusion) in patients with hypotension.
of early postoperative seizures showed no obvious rela- No adverse effects ascribed to concomitant medication
tion to age or to a history of preoperative seizures, nor were recorded.
did it show major differences between glioma and me-
ningioma patients. There was a tendency for seizures to DISCUSSION
be less common in patients with temporal tumors com-
pared with patients with extratemporal tumors, and in Although the effectiveness of PHT in preventing early
patients whose operation involved cortical dissection posttraumatic seizures has been demonstrated (9), evi-
compared with cortical incision. Four of the 12 patients dence on its potential usefulness as prophylaxis against
who underwent a transcallous approach experienced sei- seizures occurring shortly after craniotomy is unsubstan-
zures. Seizures occurred more frequently in patients with tiated. In particular, each of the four randomized studies
incomplete removal of tumor than in those who under- previously conducted produced somewhat different re-
went complete removal. sults. Because the majority of postoperative seizures are
With respect to potential risk factors not listed in known to occur within 72 h (4), Lee et al. (7) adminis-
Table 4, seizures occurred more frequently in patients tered PHT intravenously for 3 consecutive days to 189
with preoperative psychomotor slowing than in fully patients, the first dose being given before completion of

TABLE 3. Distribution of serum phenytoin concentrations during the study period

Serum phenytoin Percentage of patients with concentration
concentration
Day (mean ± SD, ␮g/ml) <10 ␮g/ml 10–20 ␮g/ml >20 ␮g/ml
2 12.3 ± 4.1 23% 72% 5%
3 12.9 ± 3.6 17% 81% 2%
5 13.0 ± 3.7 12% 85% 3%
7 11.6 ± 3.9 31% 65% 4%

Day 1 corresponds to the day of surgery.

Epilepsia, Vol. 43, No. 2, 2002

180 A. DE SANTIS ET AL.

TABLE 4. Descriptive analysis of the rate of postoperative after surgery, and the probability of developing epilepsy
seizures in relation to potential risk factors in the subsequent 2 years was equally unaffected by
Incidence of postoperative treatment. Although in the latter study, serum drug levels
Variable seizures (%) were not monitored in all patients, the initial dosing
Age (yr) schedule used for PHT (15 mg/kg intravenously for ⱖ24
15–35 18.2 (n ⳱ 22) h before surgery, and 300 mg daily thereafter) would be
36–55 11.1 (n ⳱ 54)
56–75 11.3 (n ⳱ 124) expected to have produced serum concentrations in the
Epileptic seizures before operation optimal range in the majority of patients.
Present 13.4 (n ⳱ 67) In view of the conflicting results reviewed earlier, it is
Absent 11.3 (n ⳱ 133)
Type of tumor no surprise that management practices differ widely in
Glioma 8.6 (n ⳱ 81) craniotomy patients. Many authors advise against pro-
Meningioma 13.7 (n ⳱ 95) phylactic drug treatment (2,6,19,20), others suggest that
Other 16.7 (n ⳱ 24)
Site of tumor (%) PHT may be indicated for the immediate postoperative
Frontal lobe 16.0 (n ⳱ 69) period (7,18,21,22). More remarkably, the practice of
Temporal lobe 4.2 (n ⳱ 42)
Parietal lobe 9.5 (n ⳱ 42)
prescribing long-term AED prophylaxis remains wide-
Other 16.7 (n ⳱ 42) spread, either in all patients at large or in specific patient
Surgical intervention subgroups (23,24), despite lack of adequate evidence that
Cortical incision 13.0 (n ⳱ 115)
Cortical dissection 6.8 (n ⳱ 73)
postoperative and posttraumatic epilepsy can be pre-
Transcallous 33.3 (n ⳱ 12) vented by treatment (4,6,9,14,25). The extent to which
Extent of tumor removal AED prophylaxis is practiced by neurologists and neu-
Complete 10.0 (n ⳱ 155)
Incomplete 17.8 (n ⳱ 45) rosurgeons in Italy is best illustrated by the observation
that >90% of the patients undergoing craniotomy in the
present study had been prescribed long-term treatment
with either CBZ or a barbiturate, although only one third
supratentorial surgery. The indication for surgery was
of these patients had experienced a seizure.
highly heterogeneous in this population, and 87 patients
Our study differs from previous randomized trials in at
had severe traumatic head injury, a relevant observation
least three aspects. First, inclusion was restricted to pa-
in view of the demonstrated effectiveness of PHT in
tients operated on for supratentorial brain tumors. This
preventing seizures after trauma (9). Overall, only three
immediate or early postoperative seizures were recorded eliminated the confounding effect of head trauma and
in these patients compared with 13 seizures in a group of reduced heterogeneity in underlying conditions that char-
185 control patients randomized to placebo, and the acterized most of the studies conducted to date, even
number of patients with seizures also was lower in the though it should be acknowledged that the type and lo-
PHT group (two vs. nine in controls), even though these cation of the tumor, together with the surgical technique,
differences failed to reach statistical significance. A also may significantly affect the risk of postoperative
smaller trial in which 63 patients operated on for tumor seizures (1,2,4,8,26), as confirmed by the descriptive
were randomized to PHT, PB, or no treatment also sug- analysis performed in the overall population. A second
gested that treatment could protect against early-onset feature of our study was the intensive use of drug-level
seizures, by the effect was not statistically significant monitoring, aimed at producing and maintaining serum
(18). Although a statistically significant reduction in the PHT levels within the commonly accepted target range
risk of early postoperative seizures by PHT could be of 10 to 20 ␮g/ml. Although these levels were not
demonstrated by North et al. (8) in a placebo-controlled achieved at all sampling times in all patients, this ap-
trial in a total of 281 patients (including 34 with head proach resulted in average PHT concentrations well
injury), no protective effect could be demonstrated in the within target, and the proportion of patients with PHT
first week of treatment, and only seizures occurring be- concentrations within range on any given day was much
tween 7 and 72 days after surgery were reduced signifi- higher than that recorded in the trial by Lee et al. (7), in
cantly by PHT. Possibly the ineffectiveness of PHT in which only 59.8% of patients had concentrations be-
the first week of therapy in this study could be explained tween 10 and 20 ␮g/ml. A third important aspect in
by the fact that treatment was delayed until the patient which our trial differed from previous studies is that PHT
entered the recovery room, and the loading dosage of was mostly used adjunctively to other AEDs. Although
PHT (250 mg intravenously, twice daily) was probably there is no evidence that pretreatment with CBZ or PB is
insufficient to achieve therapeutic blood levels during effective in preventing early postoperative seizures, and
the first critical days after surgery. In a fourth study in a the incidence of seizures in our control group receiving
total of 276 patients, neither PHT nor CBZ reduced sig- these drugs did not differ from that recorded in other
nificantly the incidence of seizures within the first week series of craniotomy patients receiving no treatment (3),

Epilepsia, Vol. 43, No. 2, 2002

 PHENYTOIN AND EARLY POSTOPERATIVE SEIZURES
our results may not necessarily be applicable to patients
in whom no background medication is present.
Under these study conditions, PHT was clearly inef-
fective in affording a clinically significant protection
against early postoperative seizures. Neither the total
number of patients experiencing seizures during the
7-week treatment, nor the number of patients with im-
mediate postoperative seizures, nor the total number of
seizures recorded during the entire observation period
showed even a trend to be reduced in the PHT treatment
group. By contrast, there was paradoxically a trend for
the incidence of generalized seizures to be more common
in PHT-treated patients than in controls (11 vs. five,
respectively). These results, which are based on the
sample size originally planned for interim analysis, led
us to conclude that it would have been unethical to con-
tinue recruitment, and the calculations of conditional
power supported this conclusion. Although no life-
threatening adverse effects were observed during the
study, four patients experienced severe hypotension dur-
ing PHT infusions, and the lack of even a trend for a
beneficial effect of treatment hardly justified exposing
additional patients to the risk of side effects. Admittedly,
the limited sample size achieved in the study does not
allow us to exclude a protective effect of a smaller mag-
nitude than that used for power calculations, but such an
effect would be clinically unimportant in view of the
relatively low frequency of the event to be prevented,
and the consequently high number of patients that would
need to be treated unnecessarily to prevent a single event.
The ineffectiveness of PHT in the present study could
not be explained by underlying differences between
treated patients and controls, because the two groups
were balanced in terms of baseline clinical characteris-
tics, surgical procedures, and incidence of postoperative
complications. Equally, the occurrence of seizures in the
PHT-treated group could not be ascribed to inadequate
dosing, as illustrated by the observation that 11 of the 13
seizure patients in the treated group had serum PHT con-
centrations within the target range on the days when
seizures occurred. Moreover, the proportion of patients
with serum PHT concentrations within target range in the
first 2 days (72–81%) was greater than that observed
over the same period by Lee et al. (7) in the only trial that
suggested a marked protective effect against immediate
or early postoperative seizures. Based on our findings,
we can conclude that either PHT is ineffective per se in
protecting against early postoperative seizures, as sug-
gested by Foy et al. (2), or that it fails to confer any
additional benefit beyond that which might be afforded
by underlying AED medication.
Whatever the mechanisms underlying these findings,
our results provide strong support for the view that short-
term PHT prophylaxis has no useful role in the routine
management of patients undergoing craniotomy for su-
181
pratentorial tumors (2,19,20), particularly when other
AEDs have been already prescribed. Whether prophy-
laxis could be indicated in specific subgroups of patients
at high risk for seizures can be addressed only in a large
multicenter study.
Acknowledgment: We thank Dr. Erminio Bonizzoni for his
valuable assistance in the statistical analysis of data.
REFERENCES
1. Foy PM, Copeland GP, Shaw MDM. The incidence of postopera-
tive seizures. Acta Neurochir 1981;55:253–64.
2. Foy PM, Chadwick DW, Rajgopalan N, et al. Do prophylactic
anticonvulsant drugs alter the pattern of seizures after craniotomy?
J Neurol Neurosurg Psychiatry 1992;55:753–7.
3. Matthew E, Sherwin AL, Welner SA, et al. Seizures following
intracranial seizures: incidence in the first post-operative week.
Can J Neurol Sci 1980;7:285–90.
4. Kuijlen JM, Teernstra OP, Kessels AG, et al. Effectiveness of
antiepileptic drug prophylaxis used in supratentorial craniotomies:
a meta-analysis. Seizure 1996;5:291–8.
5. De Santis A. Diffusion of antiepileptic prophylaxis in Italy. In:
Murri L, Parenti G, Annegers JF, eds. Pharmacological prophy-
laxis of postraumatic epilepsy. Padua: Piccin Edizioni, 1991:95–
112.
6. Perucca E. Principles of drug treatment. In: Shorvon SD, Dreifuss
F, Fish D, et al., eds. The treatment of epilepsy. Oxford: Blackwell
Science, 1996:162–8.
7. Lee ST, Lui TN, Chang CN, et al. Prophylactic anticonvulsants for
prevention of immediate and early postcraniotomy seizures. Surg
Neurol 1989;31:361–4.
8. North JB, Penhall RK, Hanieh A, et al. Phenytoin and post-
operative epilepsy: a double-blind study. J Neurosurg 1983;58:
672–7.
9. Temkin NR, Dikmen SS, Wilensky AJ, et al. A randomized,
double-blind study of phenytoin for the prevention of postraumatic
seizures. N Engl J Med 1990;323:497–502.
10. Olanow CW, Finn AL. Phenytoin: pharmacokinetics and clinical
therapeutics. Neurosurgery 1981;8:112–7.
11. Treiman DM. Treatment of status epilepticus. In: Engel J Jr, Ped-
ley TA, eds. Epilepsy: a comprehensive textbook. Philadelphia:
Lippincott-Raven, 1997:1317–23.
12. Perucca E, Richens A. Antiepileptic drugs: clinical aspects. In:
Richens A, Marks V, eds. Therapeutic drug monitoring. Edin-
burgh: Churchill-Livingstone, 1983:320–48.
13. Glauser TA, Pippenger CE. Controversies in blood-level monitor-
ing : reexamining its role in the treatment of epilepsy. Epilepsia
2000;41(suppl 8):S6–15.
14. Iudice A, Murri L. Pharmacological prophylaxis of post-traumatic
epilepsy. Drugs 2000;59:1091–9.
15. Cranford RE, Leppik IE, Patrick B, et al. Intravenous phenytoin:
clinical and pharmacokinetic aspects. Neurology 1978;28:874–80.
16. Fleiss J. Statistical methods for rates and proportions. New York:
J. Wiley and Sons, 1973.
17. Proschan MA, Hunsberger SA. Designed extension of studies
based on conditional power. Biometrics 1995;51:1315–24.
18. Franceschetti S, Binelli S, Casazza M, et al. Influence of surgery
and antiepileptic drugs on seizures symptomatic of cerebral tu-
mours. Acta Neurochir 1990;103:47–51.
19. O’Laoire SA. Epilepsy following neurosurgical intervention. Acta
Neurochir 1990;50:52–4.
20. Hayashi T, Hadeishi H, Kawamura S, et al. Postoperative anticon-
vulsant prophylaxis for patients treated for cerebral aneurysms.
Neurol Med Chir 1999;39:828–33.
21. Tsolaki M, Yannacou-Peftoulidou M, Grammaticos P, et al. Phe-
nytoin plasma levels after intraoperative administration, for the
prevention of post-craniotomy seizures. Acta Neurochir 1987;84:
36–8.
22. Baker CJ, Prestigiacomo CJ, Solomon RA. Short-term periopera-
Epilepsia, Vol. 43, No. 2, 2002
182 A. DE SANTIS ET AL.
tive anticonvulsant prophylaxis for the surgical treatment of low-
risk patients with intracranial aneurysms. Neurosurgery 1995;37:
863–70.
23. Notani M, Kawamura H, Amano K, et al. The incidence of post-
operative epilepsy and prophylactic anticonvulsants in patients
with intracranial aneurisms. No Shinkei Geka 1984;12(suppl 3):
269–74.
24. Beenen LF, Lindeboom J, Kasteleijn-Nolst Trenite DG, et al. Com-
parative double-blind clinical trial of phenytoin and sodium val-
Epilepsia, Vol. 43, No. 2, 2002
proate as anticonvulsant prophylaxis after craniotomy: efficacy,
tolerability, and cognitive effects. J Neurol Neurosurg Psychiatry
1999;67:474–80.
25. Temkin NR, Dikmen SS, Anderson GD, et al. Valproate therapy
for prevention of post-traumatic seizures: a randomized trial. J
Neurosurgery 1999;91:593–600.
26. Fukamachi A, Koizumi H, Nukui H. Immediate postoperative sei-
zures: incidence and computed tomographic findings. Surg Neurol
1985;24:671–6.