Pregabalin Drug Interaction Studies: Lack of Effect on the Pharmacokinetics of Carbamazepine, Phenytoin, Lamotrigine, and Valproate in Patients with Partial Epilepsy


Address correspondence and reprint requests to Dr. H.N. Bockbrader at 2800 Plymouth Rd., Ann Arbor, MI 48105, U.S.A. E-mail:

Present address of Dr. Posvar: Amgen, Inc., Thousand Oaks, CA, U.S.A.


Summary: Purpose: Pregabalin (PGB) is an α2-δ ligand with demonstrated efficacy in epilepsy, neuropathic pain, and anxiety disorders. PGB is highly efficacious as adjunctive therapy in patients with refractory partial seizures.

Methods: Given its efficacy as adjunctive therapy, the potential for interaction of PGB with other antiepileptic drugs (AEDs) was assessed in patients with partial epilepsy in open-label, multiple-dose studies. Patients received PGB, 600 mg/day (200 mg q8h) for 7 days, in combination with their individualized maintenance monotherapy with valproate (VPA), phenytoin (PHT), lamotrigine (LTG), or carbamazepine (CBZ).

Results: Trough steady-state concentrations of CBZ (and its epoxide metabolite), PHT, LTG, and VPA were unaffected by concomitant PGB administration. Likewise, PGB steady-state pharmacokinetic parameter values were similar among patients receiving CBZ, PHT, LTG, or VPA and, in general, were similar to those observed historically in healthy subjects receiving PGB alone. The PGB–AED combinations were generally well tolerated. PGB may be added to VPA, LTG, PHT, or CBZ therapy without concern for pharmacokinetic drug–drug interactions.

Pregabalin (PGB) has been approved in the European Union (EU) as an adjunctive therapy for treatment of partial seizures with and without generalized tonic–clonic seizures and for the treatment of peripheral neuropathic pain in adults. Significant dose-related reductions in seizure frequency are seen in as many as three of four patients who receive PGB as part of their antiepileptic drug (AED) therapy, with onset of anticonvulsant activity occurring by the second day of treatment and efficacy being maintained for ≥2 years (1–8).

PGB binds potently to the α2-δ subunit, an auxiliary protein, of Q-type voltage-sensitive calcium channels that are widely distributed throughout the peripheral nervous system and CNS (9–11). Potent binding at this site reduces calcium influx at hyperexcited nerve terminals and, therefore, reduces the release of several neurotransmitters, including glutamate, noradrenaline, and substance P (12–15). These activities and effects result in the anticonvulsant, analgesic, and anxiolytic activity exhibited by PGB. Studies with PGB and a number of structural derivatives indicate that binding at the α2-δ site is required for anticonvulsant, analgesic, and anxiolytic activity in animal models (9,16). PGB is inactive at γ-aminobutyric acid (GABA)A and GABAB receptors; it is not converted metabolically into GABA or a GABA antagonist, and it does not alter GABA uptake or degradation (17,18).

The pharmacokinetics of single and multiple doses of PGB have been assessed in healthy volunteers, in patients with epilepsy, and in patients with impaired renal function. These studies have shown that PGB is not appreciably metabolized, and >90% is renally excreted as unchanged drug (19). PGB clearance is similar in healthy volunteers and patients, and it is unaffected by factors such as sex, race, dose, and menopausal status in women (20,21). Because it is primarily renally excreted, however, its clearance is reduced in patients with renal impairment (20,21).

The potential for drug–drug interactions between AEDs is an important consideration in cases in which patients require multiple agents for optimal seizure control. For example, phenytoin (PHT) and carbamazepine (CBZ) are both substrates, and inducers, of cytochrome P450-3A and have high potential for interaction, whereas valproate (VPA) can increase risk of rash with lamotrigine (LTG) (22–25). The present studies were undertaken to describe the plasma pharmacokinetics of PGB in patients with partial epilepsy maintained on individualized monotherapy with commonly used AEDs, including CBZ, VPA, LTG, and PHT. In each study, the influence of the AEDs on the steady-state pharmacokinetics of PGB was evaluated, as was the influence of PGB on steady-state pharmacokinetics of the other AEDs.


Five, open-label, single-site studies, all similar in design, were conducted in patients with partial epilepsy who were receiving individualized maintenance therapy with CBZ (one study), PHT (one study), LTG (one study), or VPA (two studies combined for analysis). Studies were conducted at Western Infirmary, Epilepsy Unit, Glasgow, Scotland (CBZ, LTG, VPA), Clinical Study Centers, LLC, Little Rock, AR, U.S.A. (PHT), and Aster.Cephac, Paris, France (VPA). Protocols and informed consent forms were approved by the affiliated Institutional Review Boards (West Ethics Committee, Western Infirmary, Glasgow, Scotland; Arkansas Institutional Review Board, Baptist Medical Center Clinical Laboratory, Little Rock, AR, U.S.A.; and CCPPRB de la Pitie-Saltpetriere, Groupe Hospitalier Paris La Pitie-Saltpetriere, Paris, France, respectively) before commencement of each study.


Patients of either sex with a history of epilepsy but who were otherwise healthy [determined by medical history, physical examination, electrocardiogram (ECG), vital signs, and clinical laboratory measurements] were eligible for all studies, provided they were 18 to 65 years old and weighed ≥50 kg. Additionally, only patients stabilized on AED monotherapy were eligible. Stable therapy was demonstrated by trough plasma AED concentrations being within therapeutic ranges (3 to 8 μg/ml for CBZ; 2–15 μg/ml for LTG; 1.0–2.0 μg/ml for free PHT; and 50–100 μg/ml for VPA) and not varying >20% during two separate determinations ∼1 week apart in the 14 days before PGB administration. Women of child-bearing potential or lactating women, patients with a history of significant adverse reaction to gabapentin (GBP), and patients with unstable seizure disorders were not eligible. All patients gave written informed consent before participation in the studies.

Study designs and procedures

All patients were maintained on their individualized AED monotherapy before, during, and after PGB administration. In all but the PHT study, PGB (200 mg as 2 × 100-mg capsules) was administered in the fasting state with 250 ml water every 8 h for 7 days (600 mg/day, days 1 through 7) followed by a single dose on day 8.

Administration of PGB with food had been shown to decrease Cmax by ∼25–30%, although overall exposure was unchanged (26,27). It was hypothesized that if tolerability were a function of Cmax, then coadministration of PGB with food might improve the tolerability of PGB. Therefore in the PHT study only, PGB doses were administered with food and 250 ml of water. Patients who had been taking food with their generic PHT doses were instructed to continue doing so throughout the PGB phase of the study, and patients who had not been taking food with their generic PHT doses were instructed to continue with this regimen.

In all studies, patients fasted for 8 h before each morning drug dose, or before each clinical laboratory assessment, or both; for 2 h before all other doses; and for 2 h after each dose. On day 8, patients remained fasting until after the 4-h blood collection. Strenuous exercise was prohibited during the period of the study and, as a precautionary measure, driving a motor vehicle or operating heavy machinery was prohibited during, and for 1 to 2 days after, completion of PGB dosing.

Blood samples (5 or 10 ml) for determination of PGB concentrations were collected before the morning PGB dose on days 1, 2, 3, 4, 6, 7, and 8, and at 0.5, 1, 1.5, 2, 3, 4, 6, 8, 12, 16, 24, 36, and 48 h after dosing on day 8. Blood samples (5 ml) for determination of trough plasma AED concentrations were collected before the morning dose before (two samples ∼1 week apart, no earlier than day −14, and on day 1 before the first pregabalin dose), during (days 2, 3, 4, 6, 7, and 8), and after (days 9 and 10) PGB administration.

Analysis of plasma samples

All blood samples were collected into heparinized glass tubes, and after centrifugation, plasma samples were transferred to plastic tubes and stored frozen at −20°C until assayed. Plasma samples were shipped on dry ice to Bioassay Laboratory, Inc. (Houston, TX, U.S.A.) for determinations of PGB and of CBZ (and CBZ 10,11-epoxide); to Phoenix International (Montreal, Quebec, Canada) for determinations of total PHT and LTG; and to PPD Development (Richmond, VA, U.S.A.) for determination of VPA. PGB was determined by a validated, high-performance liquid chromatography (HPLC) assay with UV detection (350 nm) and a lower limit of quantification (LOQ) of 0.05 μg/ml (28). CBZ and CBZ 10,11-epoxide were determined by a validated HPLC assay with UV detection (218 nm) and lower LOQs of 0.1 μg/ml and 0.01 μg/ml, respectively. Total PHT and LTG were determined by validated liquid chromatography tandem mass spectrometry (LC/MS/MS) assays with lower LOQs of 0.02 μg/ml and 0.03 μg/ml, respectively. VPA was determined by a gas chromatography assay with an LOQ of 2.0 μg/ml.

Pharmacokinetic parameters

Noncompartmental methods (WinNonlin version 2.1) were used in the determination of PGB pharmacokinetic parameters, which were assessed from each patient's plasma PGB Concentration × Time data on day 8. Pharmacokinetic parameters determined for PGB were peak plasma concentration (Cmax); time from dosing to when Cmax was reached (Tmax); area under the concentration–time curve from 0 to 8 h (AUC[0-8]) calculated by the linear trapezoidal method; minimum plasma concentration (Cmin) calculated as the average of the 0- and 8-h determinations on day 8; terminal elimination rate constant (λz), the absolute value of slope of linear regression of the natural logarithm (ln) of concentration on time during the terminal phase of concentration–time profile; and terminal elimination half-life (T½) calculated by 0.693/λz. Plasma concentrations of AEDs (and of CBZ 10,11-epoxide) were determined from samples taken before dosing (i.e., Cmin) before, during, and after PGB administration.

Safety assessments

Safety was assessed by physical examination, clinical laboratory assessments, monitoring vital signs, and recording of 12-lead ECG and 2-min rhythm strips before, during, and after PGB administration. Mean change in QTc interval from baseline to day 8 was calculated. All adverse events reported by patients were recorded, and their relation to treatment was determined by the investigator.

Statistical analysis

Pharmacokinetic data for PGB are presented as mean and percentage coefficient of variation (%CV). Concentrations of AEDs determined during PGB administration were compared with those observed before and after PGB for clinically meaningful differences. Log-transformed AED concentrations were compared by a repeated measures analysis of variance (ANOVA) using a model incorporating patient and treatment (before, during, or after PGB) effects. The 90% confidence intervals (CIs) for the ratio of treatment least-squares (LS) means were calculated; treatment differences were considered inconsequential if the 90% CI for ratio of treatment means lay entirely within the 80–125% range. PGB pharmacokinetic parameter estimates derived in this study were visually inspected and compared with historical values in healthy volunteers to assess similarity (29). No formal statistical tests were performed to compare pharmacokinetic parameter estimates across studies.


Study populations

Demographic information regarding patients in each of the studies is summarized in Table 1. Data from all patients (14 receiving CBZ; 11, PHT; 12, LTG; 16, VPA) were included in the demographic and safety analyses for the respective studies. Data from seven patients (two receiving CBZ; four, VPA; one, PHT) who did not complete the studies were not included in the pharmacokinetic analyses for the respective studies.

Table 1. Demographic characteristics of study populations
ParameterMaintenance therapy
(n = 14)
(n = 11)
(n = 12)
(n = 16)
Sex (M/F)8/610/17/59/7
Age [yr, mean (range)]43 (22–62)43 (18–65)41 (26–61)37 (18–58)  
Weight [kg, mean (range)] 79.2 (52.3–123)   90.8 (66.8–135.9) 79.1 (56.5–114)73.5 (51.9–87.9)
Mean daily dosage (range)592.9 mg/d350 mg/d287.5 mg/d1.34 g/d
(300–1,000 mg/d)(100–500 mg/d)(100–600 mg/d)     (1.0–2.0 g/d)

Pharmacokinetics of pregabalin in patients with partial epilepsy

Plasma Concentration × Time curves for PGB (day 8) in patients receiving AEDs are shown in Fig. 1. Pharmacokinetic parameter values are summarized in Table 2. Overall, plasma Concentration × Time curves and pharmacokinetic parameter values for PGB were similar among patients receiving the different AEDs.

Figure 1.

Mean pregabalin plasma Concentration × Time profiles during administration of pregabalin (200 mg every 8 h; day 8) in patients maintained on carbamazepine, phenytoin, lamotrigine, or valproate.

Table 2. Pharmacokinetics of pregabalin in patients with partial epilepsy maintained on antiepileptic drugs
Pregabalin parameterMaintenance therapy
(n = 12)
Mean (%CV)
(n = 12)
Mean (%CV)
(n = 12)
Mean (%CV)
(n = 10)
Mean (%CV)
  1. Ratio is based on comparison with data from 11 human volunteers receiving 200 mg pregabalin every 8 h for 22 days. Ratio was calculated as 100 × (patient value/volunteer value) by using the following volunteer values: Cmax= 8.52 μg/ml; Tmax= 0.909 h; AUC[0-8]= 41.7 μg/h/ml; and T½= 6.27 h (26).

Cmin (μg/ml)  3.21 (24.7)3.90 (31.6)3.55 (19.5)3.44 (23.6)
Cmax (μg/ml)  8.20 (50.7)9.80 (32.8)8.99 (18.5)5.81 (14.0)
Tmax (h)  1.45 (32.9)1.08 (67.7)1.10 (23.3)2.65 (37.8)
AUC[0-8] (μg/h/ml)38.3 (29.0)48.1 (33.8)  46.9 (18.9)  35.8 (18.6)  
T1/2 (h)  7.22 (45.0)6.21 (26.6)7.81 (19.5)6.79 (17.0)
Ratio115 99.0125108

PGB was rapidly absorbed. Time to peak plasma concentrations ranged between 1.08 and 2.65 h. Peak plasma concentrations of PGB ranged between 5.81 and 9.80 μg/ml, and exposure, based on AUC[0-8], ranged between 35.8 and 48.1 μg/h/ml. The mean elimination half-life ranged between ∼6 and 8 h. Of the four maintenance therapies tested, the combination with PHT resulted in the longest Tmax and lowest Cmax observed for PGB, but exposure (AUC[0-8]) and elimination (T½) were not affected (see Table 2).

Because patients were maintained on their AED therapy during the study, no opportunity arose to compare the PGB pharmacokinetic profiles in the same patients without AEDs. For this reason, data from healthy volunteers (30) were used to provide a comparison (see Table 2). Pharmacokinetic profiles for PGB in patients receiving CBZ, LTG, and VPA were similar to those in healthy subjects receiving PGB alone (see Table 2). In patients receiving PHT, the rate of PGB absorption based on Tmax values appeared to be slower than that in healthy subjects receiving PGB alone. Mean PGB Cmax and AUC[0-8] values with PHT were ∼30% and 14% lower, respectively, relative than those reported for the healthy volunteers.

Effect of pregabalin on antiepileptic drugs

PGB administration did not significantly affect the pharmacokinetics of CBZ, LTG, VPA, or PHT. Trough (Cmin) concentrations of AEDs before, during, and after treatment with PGB are shown in Fig. 2. Least squares (LS) mean trough concentrations before, during, and after PGB dosing are shown in Table 3, along with the ratio (and 90% CIs) of treatment LS means. Overall, concentrations of CBZ and its 10,11-epoxide, and of PHT, LTG, and VPA were not significantly different across the study period. Compared with values obtained before PGB dosing, no clinically important differences were noted in plasma concentrations of any of the AEDs during and after dosing with PGB. With one exception (VPA), the respective 90% CI for ratio of LS means lay entirely within the 80–125% range (see Table 3). For VPA, the 90% CI for ratio of LS means was within the 80–125% range when PGB was coadministered. After discontinuing PGB, the 90% CI for ratio of LS means was slightly outside the range (74.9–106); however, this difference is not considered clinically important. Likewise, plasma concentrations after the last PGB dose were similar to those during PGB dosing, and the 90% CIs for the respective ratio of means lay within the expected range. Based on similar steady-state CBZ epoxide trough plasma concentrations in patients maintained on CBZ therapy before, during, and after PGB administration, the extent of CBZ metabolism to CBZ epoxide appears to be unaltered by PGB. LS mean CBZ 10,11-epoxide concentrations were 0.643 μg/ml before, 0.683 μg/ml during, and 0.655 μg/ml after PGB dosing.

Figure 2.

Mean (SD) premorning dose concentrations of carbamazepine (CBZ; solid circles) and CBZ epoxide (open circles) (A) and of total phenytoin (B), lamotrigine (C), and valproate (D) before (screen to day 1 before the first pregabalin dose), during (days 2, 3, 4, 6, 7, and 8), and after (days 9 and 10) pregabalin administration (200 mg every 8 h).

Table 3. Effect of pregabalin on steady-state concentrations of antiepileptic drugs
Dose period/ComparisonMaintenance therapy
Carbamazepine (n = 12)Phenytoin (n = 10)Lamotrigine (n = 12)Valproate (n = 12)
LS MeanRatio (90% CI)LS MeanRatio (90% CI)LS MeanRatio (90% CI)LS MeanRatio (90% CI)
  1. Ratio of LS means, expressed as percentage; comparisons are with values before pregabalin administration.

Before pregabalin7.8212.53.9570.1
During pregabalin8.06103 (96.8–110)12.5100 (94.1–106)3.5990.9 (85.0–97.3)66.494.7 (83.4–108)
After pregabalin7.99102 (94.6–111)12.8102 (95.7–107)3.6993.2 (85.5–102)62.689.3 (74.9–106)


Most of the patients in these studies reported one or more adverse events. Most commonly these were somnolence and dizziness, which occurred irrespective of maintenance therapy (Table 4). Approximately 95% of reported adverse events were considered mild to moderate in severity and transient in nature. Investigators assessed most of the adverse events as treatment emergent. These events were most likely related to PGB, as they are consistent with the safety profile for PGB reported in clinical trials (5–7), although possible pharmacodynamic interaction effects could not be entirely ruled out. Six patients withdrew from the studies because of adverse events. Five patients (three maintained on VPA and two on CBZ) had seizures, which investigators considered only possibly related to study medication, and one patient receiving VPA had a myocardial infarction, which was considered unrelated to treatment. Other safety evaluations, including physical examination, clinical laboratory assessments, and vital signs showed no clinically significant changes. Mean QTc interval changes before and after (day 8) dosing with PGB also showed no clinically relevant changes in patients maintained on CBZ (−1.6 ms), PHT (0.5 ms), LTG (−1.4 ms), or VPA (−4.5 ms).

Table 4. Most common adverse events reported in the clinical pharmacology studies involving concomitant pregabalin administration
Adverse event*Maintenance Therapy
(n = 14)
(n = 11)
(n = 12)
(n = 16)
  1. *Treatment-emergent, all causality.

 Number of patients
Dry mouth47


The general findings of these studies are that PGB may be safely combined with commonly used AEDs (CBZ, PHT, LTG, and VPA) without concern for pharmacokinetic interactions. In general, the pharmacokinetics of PGB were similar in patients maintained on these different AED therapies and were consistent with those observed in healthy subjects receiving the same PGB dose regimen over a 3-week period. Overall, PGB was rapidly absorbed, reaching peak concentrations of ∼6 to 10 μg/ml in 1 to 3 h. Thereafter, it was eliminated with a T½ of 6 to 8 h.

In patients who were being maintained on PHT, alterations in the pharmacokinetics of PGB were noticeable: absorption, based on the Tmax of PGB, was somewhat slower, and Cmax and AUC[0-8] were lower compared with those reported for healthy volunteers who received PGB alone. Although differences in body weight between the two studies (79.2 vs. 90.8 kg) might explain some of the differences observed in Cmax and AUC[0-8] parameter values, they cannot, however, explain the changes observed in Tmax or the magnitude of change in Cmax. The changes in the pharmacokinetic parameter values more likely reflect the fact that all patients were given PGB with food. The changes in Tmax (1.7 h) and Cmax (32% reduction) observed in this study are consistent with those observed in PGB food-effect studies in which Tmax was delayed 1 to 2.5 h and Cmax was reduced 25–31% (26,27). (The 14% change in AUC[0-8] was not considered clinically relevant.) Therefore the changes observed in the pharmacokinetics of PGB during coadministration with PHT likely reflect an effect of food on PGB rather than a drug–drug interaction between PGB and PHT. The observations that PGB pharmacokinetics are unaffected by CBZ, VPA, LTG, and PHT are consistent with the population pharmacokinetic analyses involving data from phase I through III studies which showed that the pharmacokinetics of PGB were similar among healthy volunteers and patients with epilepsy and unaffected by the aforementioned AEDs.

PGB did not alter the steady-state concentrations of other AEDs. Plasma concentrations of CBZ and its epoxide, PHT, LTG, and VPA were stable during PGB treatment and did not differ from those observed either before or after PGB dosing. These findings are consistent with population pharmacokinetic analyses of PGB as add-on therapy in placebo-controlled, double-blind clinical trials involving patients with refractory partial seizures. In these analyses, steady-state plasma concentrations of CBZ, PHT, LTG, and VPA, as well as phenobarbital and topiramate, were no different before and during PGB treatment (based on 90% CI of ratio of means being in the 80–125% range) (31).

In general, patients maintained on CBZ reported more adverse events than did patients maintained on other AEDs while taking PGB. CBZ has also been associated with a higher incidence of adverse events than either VPA or LTG in newly diagnosed patients taking their first AED (32). CBZ carries a low risk of QTc abnormalities in psychiatric patients (33). PGB administration had no relevant effect on QTc interval in patients maintained on CBZ or any of the other AEDs used in these studies. The incidence of spontaneously reported adverse events was lowest in the VPA study. Although the differences in adverse-event reporting rates across the AEDs studied may be related to a chance, a pharmacodynamic interaction for any of the reported adverse events could not be entirely ruled out in these small studies. Nevertheless, most of the adverse events were considered mild to moderate in severity and transient in nature. Because tolerability has been shown to contribute as much as efficacy to the overall effectiveness of AED therapy (32), these results are supportive of the use of PGB as adjunctive therapy in patients with refractory partial epilepsy. In summary, the lack of drug–drug interactions between PGB and other commonly used AEDs will benefit both patients and prescribers of such medications.


Acknowledgment:  We are very grateful to Dr. Joyce Willetts of Palanquin Corporation and Ms. Beth McGee of Embryon for their technical assistance. This research was supported by Pfizer Global Research Division, Pfizer, Inc., Ann Arbor, MI 48105.