EDITOR:

Phenytoin is generally prescribed to patients with supratentorial tumours to decrease the risk of seizures. Earlier studies showed that plasma phenytoin concentration may not be in the therapeutic range despite continued therapy [Reference Yeh, Dhir and Green1]. During craniotomy for a supratentorial tumour, an intraoperative loading dose of phenytoin is generally used to prevent postoperative seizures [Reference Lee, Lui and Chang2,Reference Mathew, Sherwin and Welner3]. In our institution, it has been common practice not to administer phenytoin intraoperatively. To understand the consequences of our practice of withholding intraoperative phenytoin, we measured perioperative serum phenytoin concentration in a group of patients undergoing supratentorial tumour surgery. We also tried to determine the factors influencing postoperative serum phenytoin concentrations.

Twenty-five adult patients (ASA I or II) of either sex, receiving phenytoin for a period not less than 7 days before supratentorial surgery, were studied after institutional approval and informed consent. On the day of surgery, 300 mg of phenytoin was administered either orally or intravenously 4 h before surgery. The anaesthetic technique comprised of induction with thiopentone (5–6 mg kg⁻¹), tracheal intubation facilitated by a muscle relaxant and maintenance with either isoflurane or propofol. Intraoperative analgesia was provided by fentanyl. Serum phenytoin concentration was measured before induction, immediately after surgery and 24 h after surgery. The assay, performed by a chemiluminescence technique using an Immulite Assay Kit^® (DPC^®; Los Angeles, CA, USA), permitted the measurement of total phenytoin concentration. The following parameters were recorded in all patients: duration of anaesthesia and surgery, volume of crystalloids, colloids and blood products administered, volume of urine output and blood loss, and occurrence of immediate postoperative seizures.

A repeated-measures analysis of variance (ANOVA) with Bonferroni’s test was used to find out significant differences among the preinduction, immediate postoperative and delayed postoperative serum phenytoin concentrations. Study variables in patients with therapeutic and subtherapeutic concentrations of phenytoin were compared by one-way ANOVA for continuous data and a χ ²-test for categorical variables. Pearson’s test was used to correlate preinduction phenytoin concentration and its decrease in the immediate postoperative period. Logistic regression analysis was used to determine the independent predictors of immediate postoperative subtherapeutic serum phenytoin concentration. A P value of <0.05 was considered significant.

There were 17 male and 8 female patients in the study. Their age was 38 ± 12 yr and body weight was 56 ± 11 kg. Twelve patients had a preoperative history of seizures. The preinduction serum phenytoin concentration was highly variable among the patients (range 2.5–37.3 μg mL⁻¹ (95% CI = 9.8–17.8 μg mL⁻¹)). Despite continuous medication until the morning of surgery, 11 patients (44%) had a subtherapeutic concentration of serum phenytoin (normal range 10–20 μg mL⁻¹) in the preinduction sample. Serum phenytoin concentration was significantly lower in the immediate postoperative sample compared with the preinduction sample (9.5 ± 7.0 vs. 13.8 ± 9.4 μg mL⁻¹; P < 0.001). The concentration increased significantly in the delayed postoperative sample (11.8 ± 8.0 μg mL⁻¹; P < 0.001). The decrease in phenytoin concentration in the immediate postoperative sample correlated with its preinduction value (P < 0.01; r = 0.8). Seizures within 24 h of surgery occurred in two patients. Only one of these patients had a preoperative history of seizures. Serum phenytoin concentration was within the therapeutic range (16.3 and 10.4 μg mL⁻¹) in both the patients.

In the immediate postoperative period, serum phenytoin concentration was in a subtherapeutic range in 15 out of the 25 study patients. On univariate analysis, the variables that were significantly different between the therapeutic and subtherapeutic groups were the patient’s gender, preinduction phenytoin concentration, blood loss, blood transfusion, duration of surgery and duration of anaesthesia (Table 1). Of these, preoperative phenytoin level, intraoperative blood transfusion and the duration of surgery/anaesthesia were found to be the independent predictors of low serum phenytoin concentration in the immediate postoperative period (P < 0.05).

Table 1 Study variables in patients with subtherapeutic and therapeutic postoperative phenytoin concentration.

Data are mean ± SD or number of patients.

The value of prophylactic administration of antiepileptic drugs in patients with brain tumours remains controversial. Some authors claim a significant decrease in the incidence of seizures in the postoperative period [Reference Lee, Lui and Chang2,Reference Mathew, Sherwin and Welner3] with preoperative antiepileptic therapy. However, in one study, phenytoin doses aimed at maintaining serum phenytoin concentrations in the 10–20 μg mL⁻¹ range did not decrease the incidence of postoperative seizures [Reference De Santis, Villani, Sinisi, Stocchetti and Perucca4]. One meta-analysis showed a statistically insignificant reduction of postoperative convulsions with prophylactic anticonvulsant and suggested the need for further investigation of the issue [Reference Kuijlen, Teernstra and Kessels5].

In our study, preoperative serum phenytoin concentration was highly variable with 44% of the patients having subtherapeutic concentrations as has been reported earlier [Reference Yeh, Dhir and Green1]. Widely variable clearance, as is known to occur even in normal individuals, could probably be the reason for this variation. Another possible cause is the interaction between dexamethasone and phenytoin as has been reported earlier [Reference Lackner6]; all our patients had been receiving dexamethasone for several days before surgery. There was a linear correlation between preoperative serum phenytoin concentration and its decrease in the intraoperative period, which is possibly related to the first-order kinetics of phenytoin.

We found that preinduction serum phenytoin concentration, need for blood transfusion and the duration of surgery/anaesthesia were independent predictors of low phenytoin concentration in the immediate postoperative period. A simple process of dilution caused by blood and fluid replacement might be responsible for the decrease in phenytoin concentration. Long-duration surgery might have caused increased excretion and lower serum concentration. In a recent study by Yeh and colleagues [Reference Yeh, Dhir and Green1], less than 50% of the patients had a therapeutic level of serum phenytoin and the predictors of low serum phenytoin concentration were the same as in our study.

Despite withholding the intraoperative phenytoin dose, only two patients had postoperative seizures within 24 h. The small sample size in the study prevents drawing serious conclusions regarding the incidence of perioperative seizures with and without intraoperative loading. Given the subtherapeutic phenytoin concentration in a major proportion of our patients, it is advisable to administer an intraoperative dose of phenytoin to achieve the therapeutic level. Whether these additional doses cause toxic levels of phenytoin in patients in whom the metabolic pathways are already saturated also remains to be seen. A more rational approach would be to decide the dosing based on serum phenytoin levels measured preoperatively. It would also be interesting to study the influence of perioperative corticosteroids on serum phenytoin levels.