Pharmacokinetics of Levetiracetam and Its Enantiomer (R)-α-ethyl-2-oxo-pyrrolidine acetamide in Dogs


Address correspondence and reprint requests to Prof. M. Bialer at Department of Pharmaceutics, School of Pharmacy, Faculty of Medicine, The Hebrew University of Jerusalem, P.O. Box 12065, Jerusalem, Israel. E-mail:


Summary:  Purpose: The new antiepileptic drug, levetiracetam (LEV, ucb LO59), is a chiral molecule with one asymmetric carbon atom whose anticonvulsant activity is highly enantioselective. The purpose of this study was to evaluate and compare the pharmacokinetics (PK) of LEV [(S)-α-ethyl-2-oxo-pyrrolidine acetamide] and its enantiomer (R)-α-ethyl-2-oxo-pyrrolidine acetamide (REV) after i.v. administration to dogs. This is the first time that the pharmacokinetics of both enantiomers has been evaluated.

Methods: Optically pure LEV and REV were synthesized, and 20 mg/kg of individual enantiomers was administered intravenously to six dogs. Plasma and urine samples were collected until 24 h, and the concentrations of LEV and REV were determined by an enantioselective assay. The levels of 2-pyrrolidone-N-butyric acid, an acid metabolite of LEV and REV, were determined by high-performance liquid chromatography (HPLC). The data were used for PK analysis of LEV and REV.

Results: LEV and REV had similar mean ± SD values for clearance; 1.5 ± 0.3 ml/min/kg and volume of distribution; 0.5 ± 0.1 L/kg. The half-life (t1/2) and mean residence time (MRT) of REV (t1/2, 4.3 ± 0.8 h, and MRT, 6.0 ± 1.1 h) were, however, significantly longer than those of LEV (t1/2, 3.6 ± 0.8 h, and MRT, 5.0 ± 1.2 h). The renal clearance and fraction excreted unchanged for LEV and REV were significantly different.

Conclusions: In addition to the enantioselective pharmacodynamics, α-ethyl-2-oxo-pyrrolidine acetamide has enantioselective PK. The enantioselectivity was observed in renal clearance. Because REV has more favorable PK in dogs than LEV, the higher antiepileptic potency of LEV is more likely due to intrinsic pharmacodynamic activity rather than to enantioselective PK.

Levetiracetam, (S)-α-ethyl-2-oxo-pyrrolidine acetamide (LEV, ucb LO59), is a new antiepileptic drug (AED) recently approved by the U.S. Food and Drug Administration (FDA) (1). Levetiracetam is an ethyl analogue of piracetam, a nootropic drug used as a cognition enhancer in the elderly and in the treatment of myoclonus (2).

α-Ethyl-2-oxo-pyrrolidine acetamide possesses a single asymmetric carbon atom, and therefore has two enantiomers (Fig. 1): LEV, which is the (S)-enantiomer, and (R)-α-ethyl-2-oxo-pyrrolidine acetamide (REV). Only the (S)-enantiomer has anticonvulsant activity, and consequently, only this enantiomer has been developed and used as a new AED (3).

Figure 1.

Chemical structures of levetiracetam (LEV), its enantiomer (R)-α-ethyl-2-oxo-pyrrolidine acetamide (REV) and the acid metabolite 2-pyrrolidone-N-butyric acid (PBA).

LEV expresses significant anticonvulsant activity in numerous animal models of epilepsy. It is active against tonic electroconvulsive seizures and kindling, audiogenic seizures, and chemically induced seizures with median effective dose (ED50) values within 5–30 mg/kg, but it is inactive in the traditional maximal electroshock (MES) and metrazol (scMet) tests (4,5). In humans LEV has proven efficacy in clinical trials of patients with partial and generalized seizures (1). LEV also has been suggested to possess antiepileptogenic activity because it slows the development of scMet kindling in mice and amygdala kindling in rats (6).

LEV has been regarded as a new AED with ideal pharmacokinetics (PK) (7). In humans LEV has a very low potential for drug–drug interactions (7), and it is eliminated mainly by renal excretion with a fraction excreted unchanged (fe) of ∼66%(8,9). LEV has a renal clearance (CLr) of 0.6 ml/min/kg, indicating that the drug is renally excreted by glomerular filtration and tubular reabsorption. In humans, 24% of the LEV dose is biotransformed to 2-pyrrolidone-N-butyric acid (PBA, Fig. 1), which is subsequently excreted in urine (7). The metabolism of LEV in other species has not yet been reported. The volume of distribution of LEV in humans is ∼0.5–0.7 L/kg, a value close to total body water, and the plasma protein binding of LEV in humans is <10%(7).

Several new AEDs [e.g., vigabatrin (VGB), pregabalin, remacemide, and 10-hydroxycarbazepine (MHD)] are chiral compounds (8). Enantioselectivity is, therefore, an important issue in design and development of new AEDs. Enantiomers often exhibit pronounced differences in their PK as well as in their pharmacodynamic properties (10). The pharmacologic effects are determined, on the one hand pharmacodynamically, by the interaction of a drug with a particular enzyme or receptor, and on the other hand, pharmacokinetically by the access of a drug to the site of action (11). Enantioselective PK can lead to enantioselectivity in efficacy, toxicity, and in clinical outcome in general. VGB is marketed as a racemic mixture although only the (S)-enantiomer possesses anticonvulsant activity (12). Whereas pregabalin [(S)-3-isobutyl GABA] is more potent than its (R)-enantiomer in in vitro binding assays, the enantioselectivity is much more pronounced in vivo in rodents, probably due to enantioselective transport or disposition of the compound (13). Oxcarbazepine is a prochiral new AED that undergoes presystemic enantioselective metabolism to its chiral active entity, MHD. In animal (rodent) models, (S)-MHD is slightly more potent as an anticonvulsant than the (R)-enantiomer, but the enantioselective PK of MHD in humans suggest that one of the enantiomers (S-MHD) may be clinically superior (14). Because of the impact of enantioselective PK on drug performance, it is important to investigate the enantioselectivity in the PK of LEV and REV and to evaluate possible enantioselectivity and chiral inversion that might affect the antiepileptic activity of the drug in vivo.

Enantiomerization (chiral inversion) of several chiral drugs, such as thalidomide (15), oxazepam (16), and (R)-ibuprofen (17), has been previously reported. In all of these compounds, the chiral center is in α-position to a carbonyl group (in ibuprofen to carboxylic acid, and in thalidomide and oxazepam, to an amide), which allows tautomerization and consequently chiral inversion. Enantiomerization can be spontaneous (16), or it can proceed by metabolic interconversion (17). In LEV the chiral center is in α-position to the carbonyl of the butyramide, and therefore chiral inversion appears possible.

The evaluation of enantioselective PK is important in assessing the mechanism of action of chiral drugs like LEV. Because the mechanism of action of LEV is currently unknown, the in vitro intrinsic pharmacodynamics of LEV cannot be evaluated. The enantioselectivity in pharmacodynamics can be used as a powerful tool in finding the site of action of LEV, and consequently, it is important to evaluate the possibility of PK-based enantioselective pharmacodynamics.

The purpose of this research was to investigate the PK of LEV and REV enantiomers, to determine their PK parameters after intravenous administration to dogs, and to evaluate the enantioselective disposition and possible in vivo chiral inversion between LEV and REV. This study also evaluated whether the enantioselective pharmacodynamics have a PK basis.



The Institutional Animal Care and Use Committee of the Faculty of Medicine of the Hebrew University approved the study. The experiments were performed on six male mongrel dogs weighing 15–25 kg. The dogs were housed in the animal farm of the Hebrew University and brought to the animal facility of the school at 2-week intervals for crossover experiments. On two separate occasions, each dog was administered, in a randomized crossover design, 20 mg/kg of LEV and REV (dissolved in 5 ml saline) via an indwelling catheter. The dose of 20 mg/kg was chosen according to ED50 values of LEV in rodents (5) and to obtain plasma concentrations similar to clinically relevant plasma concentrations in humans. To compare the PK of LEV and REV, the same dose of both enantiomers was administered. Dogs were fed with regular commercial dog food 6 h after the drug injection and had free access to water during the whole experiment. Venous blood samples (5 ml) were collected from the cephalic vein at specified times after drug administration (0, 5, 10, 20, and 45 min and 1, 1.5, 2, 2.5, 3, 4, 5, 6, 8, 10, 12, and 24 h) via an indwelling catheter into heparinized test tubes. Plasma was separated by centrifugation at 3,000 g for 15 min and stored at −20°C until analysis. Urine was collected at 1 and 2 h and every 2 h thereafter until 12 h after dosing and then at 24 h using urinary catheter, the volume was measured and samples stored at −20°C.


Methanol and acetonitrile were purchased from Lichrosolv, Merck, Germany. Double-distilled ultrapure water was used throughout the study. C18 solid-phase extraction (SPE) cartridges (Bond Elut, 3 ml/500 mg) were obtained from Varian Associates (Harbor City, CA, U.S.A.). The diol SPE cartridges (3 ml/500 mg) were purchased from Lida (Kenosha, WI, U.S.A.). Phosphoric acid and triethylamine were purchased from Sigma Chemical Co. (St. Louis, MO, U.S.A.), and ethyl acetate and petroleum ether were purchased from Frutarom (Haifa, Israel). All the chemicals were of the highest grade.

Synthesis of LEV and REV

LEV and REV were synthesized by previously reported methods (18). In brief, the synthesis of LEV was started from l-methionine that was esterified to its methyl ester and then amidated to obtain l-methionine amide. This amide was treated with KOH, tetrabutylammonium bromide, and 4-chlorobutyryl chloride to yield (S)-α-[2-(methylthio)ethyl]-2-oxo-pyrrolidine acetamide. Desulfurization of the obtained acetamide produced LEV with a specific rotation of [α]D25−87.0° (c = 1, acetone) (enantiomeric excess, 94%). REV was synthesized analogously to LEV starting from d-methionine methyl ester and obtained with a specific rotation of [α]D25+88.9° (c = 1, acetone) (enantiomeric excess, 97%).

Synthesis of the metabolite 2-pyrrolidone-N-butyric acid

2-pyrrolidone-N-butyric acid was synthesized from α-ethyl-2-oxo-pyrrolidine acetamide by acidic hydrolysis. Racemic α-ethyl-2-oxo-pyrrolidine acetamide (300 mg) was dissolved in 3 ml of concentrated H2SO4, and 500 mg NaNO2 was added slowly with cooling and mixing. The reaction mixture was sealed and incubated for 4 days at 37°C. The reaction mixture was extracted twice with chloroform (50 ml), dried over Na2SO4, filtered, and evaporated to dryness. Crystallization from ethyl acetate yielded 210 mg of white crystals with a melting point of 121–123°C. 1H-NMR (300 MHz, CDCl3): δ 0.918–0.967 (t, 3H), 1.63–1.81 (m, 1H), 2.03–2.11 (m, 3H), 2.45–2.50 (t, 2H), 3.34–3.41 (m, 1H), 3.45–3.57 (m, 1H), 4.58–4.64 (dd, 1H).

Analysis of plasma and urine concentration of LEV and REV

Before analysis, the samples were allowed to reach room temperature. LEV and REV in plasma and urine were analyzed by a previously reported enantioselective gas-chromatographic (GC) method (18). In brief, the C18 SPE cartridges were conditioned with 2 ml of methanol followed by 2 ml of water. 500 μl of plasma, or 100 μl of urine was applied to the cartridge. The cartridge was allowed to run dry and washed with 500 μl of water. The analytes were eluted by using 1 ml of methanol, and the internal standard, N-dimethyl-valproylglycinamide (10 μg for plasma samples and 40 μg for urine samples) was added. Thereafter the sample was evaporated to dryness. The dry residue was dissolved to 250 μl (plasma samples) or 1 ml (urine samples) of methanol, and 1 μl was injected to the gas chromatograph. The separation was performed by using a chiral GC column, and the concentration of LEV or REV was quantified by using selected ion monitoring and ion-trap mass spectrometer (Finnigan MAT GCQ) (18). The within-batch coefficient of variation (CV%) was, in plasma samples, 5.6–7.2% (LEV) and 3.1–6.0 (REV), and in urine samples, 5.8–7.0% (LEV) and 5.1–6.3% (REV). The between-batch CV% was, in plasma samples, 8.1–14.6% (LEV) and 6.0–15.0% (REV), and in urine samples, 8.4–11.3% (LEV) and 6.6–11.7% (REV), respectively. The limit of quantification (LOQ) for both enantiomers was 0.4 mg/L and 100 mg/L in plasma and urine, respectively.

Assay of urine concentration of 2-pyrrolidine-N-butyric acid

The metabolite, 2-pyrrolidine-N-butyric acid (PBA), was assayed in urine by the following method: 100 μl of urine was mixed with 500 μl of ethyl acetate and 2 ml of acetonitrile. The diol SPE cartridges were conditioned by using 4 ml of ethyl acetate/acetonitrile 1:1 solution and 2 ml petroleum ether. The samples were applied to the cartridge, and the eluate was collected. The dry SPE cartridge was further eluted with 5 ml ethyl acetate. The combined eluate was evaporated to dryness and dissolved to 100 μl mobile phase; 10 μl was injected to the HPLC. The separation was performed by using Shimadzu HPLC system (model 10A) and a C18 column (LiChroCART, LiChrospher 250-4 with 5-μm particle size), and the compounds were detected by using a Shimadzu model SPD 10A UV detector set at 215 nm. Mobile phase consisted of 85% 0.75 mM phosphate buffer titrated to pH 2.5 with triethylamine and 15% acetonitrile. The flow rate was 0.5 ml/min. Concentrations were calculated by using calibration curves constructed by plotting the analyte peak area as a function of analyte concentration using spiked samples of concentrations 15, 30, 60, 125, 250, 500, 750, 900, and 1,000 mg/L. The precision of the method was <15% in all analyzed concentrations, and the LOQ was defined at the lowest analyzed concentration, 15 mg/L.

Pharmacokinetic analysis

PK parameters were calculated by the classic noncompartmental methods based on the statistical moment theory (19). The area under the plasma concentration-versus-time curve (AUC) was calculated by the log-linear trapezoidal method with extrapolation to infinity. The mean residence time (MRT) was calculated from the quotient AUMC/AUC, where AUMC is the area under the concentration–time product versus time curve from zero to infinity. The clearance (CL) was calculated from the quotient Dose/AUC, and the volume of distribution at steady state (Vss), from the product of CL * MRT. The CLr of LEV and REV was calculated as the ratio of cumulative amount of each enantiomer excreted unchanged in urine (Ae) for 24 h after dosing and the AUC of each enantiomer during the same time interval (Ae/AUC). The fraction of the dose excreted as PBA was calculated by dividing the cumulative amount of the acid metabolite excreted within 24 h (Me) after dosing, by the total dose (D) of the enantiomer administered (Me/D).

Statistical analysis

All data are given as mean ± standard deviation (SD). Statistical significance of differences between the PK parameters of the two enantiomers were calculated by the Wilcoxon signed-rank test for paired rEPIicates. A p value ≤0.05 was considered significant.


Mean plasma concentration-versus-time curves for LEV and REV are presented in Fig. 2. Figure 2 also depicts the cumulative amount excreted in urine of both enantiomers and the cumulative amount of PBA excreted in urine as a function of time after administration of LEV or REV. The PK parameters obtained for LEV and REV are presented in Table 1. The total body clearance and volume of distribution of the enantiomers were not significantly different. All other PK parameters were significantly different. The half-life of REV was significantly longer than that of LEV. Consequently, the S/R enantiomeric ratio in plasma concentrations decreased from 1.6 at 5 min after drug administration to 0.8 at 12 h after dosing.

Figure 2.

Mean plasma concentrations of levetiracetam (LEV) (A) and its enantiomer (R)-α-ethyl-2-oxo-pyrrolidine acetamide (REV) (B) obtained after individual administration of each enantiomer (20 mg/kg) to six dogs. Cumulative amount excreted in urine of LEV and REV (C) and cumulative amount excreted in urine of the metabolite 2-pyrrolidone-N-butyric acid (PBA) after administration of LEV and REV (D). Data points are presented as an average of six dogs and standard deviation. ▪, administration of LEV; □, administration of REV.

Table 1.  Pharmacokinetic parameters obtained for LEV and REV
 S-enantiomerR-enantiomerP Valuea
  • a

    Wilcoxon signed-ranks test (NS, not significant).

  • b

    Fraction of the dose excreted as PBA.

  • CL, clearance; fe, fraction excreted; MRT, mean residence time; CLr, renal clearance; Me, metabolite excreted; Vss, steady-state volume.

CL (ml/min/kg)1.5 ± 0.31.5 ± 0.17NS
Vss (L/kg)0.45 ± 0.130.51 ± 0.11NS
t1/2 (h)3.6 ± 0.84.3 ± 0.8≤0.05
MRT (h)5.0 ± 1.26.0 ± 1.1≤0.05
CLr (ml/min/kg)0.69 ± 0.161.02 ± 0.14≤0.05
fe (%)50 ± 571 ± 10≤0.05
Me/Db (%)4.5 ± 3.72.2 ± 2.1NS

The CLr and fraction excreted unchanged of REV were significantly greater than those of LEV. The fraction of the dose excreted as the acid-metabolite PBA was greater after the administration of LEV than after the administration of REV. The fraction excreted unchanged was 50% and 71% for LEV and REV, respectively. The glomerular filtration rate in the dog is 4.0–6.13 ml/min/kg (20,21). In this study the renal clearance was found to be 0.69 ml/min/kg and 1.02 ml/min/kg for LEV and REV, respectively, indicating enantioselective tubular reabsorption. The fraction of the dose excreted as PBA after administration of LEV was higher (4.5%) than that after administration of REV (2.2%), but the difference was not statistically significant. The observed urine concentrations of PBA in dog's urine were in the range of 17–407 mg/L, with most of the samples having concentrations between 200 and 300 mg/L after administration of LEV. After administration of REV, the urine concentrations of PBA were in the range of 15–270 mg/L, with most of the samples with a concentrations of 100–200 mg/L.

After the administration of pure LEV and REV enantiomers separately to dogs, the plasma and urine samples were analyzed by using an enantioselective assay, but none of the second enantiomer could be detected in the samples. This confirmed that no chiral inversion occurs in vivo in the case of LEV or REV.


Biologic macromolecules are very specific to spatial arrangements, and consequently, enantioselective interactions are frequently encountered. At present, there are no reports in the literature concerning enantioselective PK of LEV, and except our assay (18), there are no published enantioselective methods for LEV and REV analysis needed to explore possible chiral inversion. This study confirmed that the PK of α-ethyl-2-oxo-pyrrolidine acetamide is enantioselective. Enantioselectivity is more likely to be found in organ-specific parameters, such as renal or hepatic clearance, than in whole-body parameters like total body clearance and volume of distribution (22). This study is in agreement with this principle. The total body clearance of the LEV and REV enantiomers was not different, but the CLr was enantioselective. The fact that there was no significant difference in the total-body clearance indicates that in the dog, there may also be enantioselective metabolism that counteracts the enantioselectivity in CLr. This difference in metabolic clearance could not result only from the formation clearance of the acidic metabolite PBA, the major metabolite of LEV in humans, because the fraction of the dose excreted as PBA was rather minor (2–4%) in the dog. Consequently, in the dog, there must be an additional metabolite(s) whose formation clearance is enantioselective. The fraction of LEV metabolized to the corresponding acid is significantly smaller in dogs than in humans (24%). This finding might be of toxicologic importance when extrapolating data from dogs to humans, and the toxicity profile of the acid might need further evaluation. Enantioselective active tubular secretion has been reported for few drugs, such as 10-hydroxynortriptyline (23), ofloxacin (24), salbutamol (25), and pindolol (26), but enantioselective tubular reabsorption is not well known. Both processes can be active, saturable, or carrier-mediated and, consequently, enantioselective. The CLr of LEV and REV is smaller than the glomerular filtration rate, and therefore tubular reabsorption is indicated. Because the renal clearance is enantioselective, it is likely to be due to enantiospecific tubular reabsorption.

The PK of LEV has been studied in humans after oral administration and in rats after i.p. administration (27), but PK in dogs has not been reported in the literature. Because of the lack of data after i.v. administration to any species, the clearance of LEV has not been previously reported. The comprehensive understanding of the pharmacokinetics of an enantiomer requires intravenous administration of both enantiomers. Because REV cannot be administered to humans and a crossover study could not be carried out in rats, the dog was used in this study. As the PK of LEV was found to be linear in rats within a dose range of 20–80 mg/kg (27), and in humans, within a dose range of 1,000–3,000 mg/day (14–42 mg/kg daily), it was assumed that the PK of LEV and REV are linear in the dog. Consequently a single-dose study was found to describe satisfactorily the PK of LEV and REV.

The t1/2 of LEV in rats after i.p. administration was 1.8–3 h (6,27), whereas in humans, the t1/2 is 7–8 h and increases in the elderly (7). The half-life observed in dogs (3.6 h for LEV) was between the values for rats and humans. The volume of distribution of LEV in humans was approximated to 0.5–0.7 L/kg (7), a value similar to the total body water (TBW). The Vss observed in the current study (0.5 l/kg) was slightly smaller than the TBW of a dog. The results are in good agreement with the data reported for humans after oral administration and indicate that LEV distribution in humans and dogs is similar. Like Vss, the CLr of LEV in dogs and humans was quite similar. In humans the CLr was also about one third of the glomerular filtration rate (7), similar to that observed in the dog for LEV. The fe of LEV and the fraction of the dose excreted as the metabolite PBA is very similar in dogs and in children. In children, the fe of LEV is 52%, and the fraction metabolized (fm) to PBA, 9%(7). However, the fraction of the dose excreted as the metabolite PBA in dogs (4.5%) after administration of LEV was much smaller than that observed in adult humans (24%), showing more extensive amide–acid biotransformation in humans than in dogs. This resembles the phenomena observed in dogs and humans after administration of valpromide (VPD), an amide derivative of valproic acid (VPA). In dogs the fm of VPD to VPA was 30–40%, whereas in humans, it was 80–90%(28). LEV might also be regarded as a heterocyclic analogue of VPD that contains the substitution of 2-oxo-pyrrolidine in the β-position of the molecule. Thus, like aliphatic analogues of VPD (29,30), it undergoes metabolic hydrolysis to its corresponding acid PBA. The PK of REV has not been previously reported in any species, and this is also the first time that the PK of both enantiomers, LEV and REV, has been reported after i.v. administration. This study demonstrated that the PK of the enantiomers of (R)- and (S)-α-ethyl-2-oxo-pyrrolidine acetamide in dogs is enantioselective. It also showed that the enantioselectivity in secondary PK parameters is a consequence of enantioselectivity in the organ-specific parameters, such as CLr and metabolic clearance. The inactive enantiomer, REV, has more favorable PK in the dog, as reflected by its longer half-life. The primary PK parameters (CL and Vss) were, however, similar, and therefore the enantiospecificity in the pharmacodynamic activity is likely to be due to a enantioselective mechanism of action or receptor binding rather than a consequence of PK differences between the two enantiomers. This study also demonstrates that LEV and REV do not undergo chiral inversion in vivo in dogs.

Acknowledgments: We thank Kamal Amarnay for his skillful technical assistance. This study was supported by the Ministry of Science of Baden-Württemberg, Germany (grant III-3-H3-99-71/97).