This randomized, double-blind, placebo-controlled, two-way crossover study aimed to assess the pharmacokinetic interactions between brivaracetam 100 mg/day and a combination oral contraceptive (OC) containing 30 μg ethinylestradiol and 150 μg levonorgestrel. The study was performed in 28 healthy women over five 28-day menstrual cycles: baseline (OC only), two treatment cycles with brivaracetam (50 mg b.i.d.) or placebo coadministered with OC separated by a wash-out cycle (OC only), and a follow-up cycle (OC only). The OC was administered on days 1–21 of each cycle, and brivaracetam or placebo on days 1–28 of the treatment cycles. Pharmacokinetics of ethinylestradiol and levonorgestrel were determined on day 20; brivaracetam morning trough levels on days 20 (with OC) and 29 (without OC) were compared. Cmax (maximum plasma concentration) and AUC (area under the plasma concentration versus time curve) ratios for brivaracetam versus placebo (90% confidence interval [CI]) were 0.96 (0.88–1.04) and 0.90 (0.86–0.95) for ethinylestradiol, and 0.95 (0.91–0.99) and 0.92 (0.88–0.97) for levonorgestrel, within predefined bioequivalence limits (0.80–1.25). Brivaracetam trough levels were similar on days 20 and 29 (ratio 1.08; 90% CI 0.98–1.18). No differences in breakthrough bleeding were seen across the five cycles. It was concluded that there were no interactions between brivaracetam 100 mg/day and the OC.
A PowerPoint slide summarizing this article is available for download in the Supporting Information section here.
Brivaracetam is a novel, high-affinity synaptic vesicle protein 2A (SV2A) ligand in phase III clinical development for epilepsy. In well-controlled trials conducted to date, adjunctive brivaracetam demonstrated efficacy and good tolerability in patients with focal epilepsy.[2-4] Brivaracetam is eliminated primarily by metabolism and urinary excretion.[5-7] The major metabolic pathway involves non–cytochrome P450 (CYP)–dependent hydrolysis of the acetamide group to an acid metabolite. The secondary pathway involves CYP2C19-mediated oxidation to a hydroxy metabolite.
There is evidence that brivaracetam may weakly induce CYP3A4 activity at doses of 400 mg/day or higher, as a small, dose-related increase in 6β-hydroxycortisol/cortisol urinary excretion ratio was observed in healthy male volunteers following 14 days of repeated dosing with brivaracetam 400 and 800 mg/day. In healthy female volunteers, brivaracetam 400 mg/day increased 6β-hydroxycortisol/cortisol by 1.6-fold.
Many fertile women with epilepsy are prescribed a combination oral contraceptive (OC) containing an estrogen (generally ethinylestradiol) and a progestogen (e.g., levonorgestrel). Ethinylestradiol is primarily metabolized via CYP3A4-catalyzed hydroxylation, and levonorgestrel exhibits more complex pharmacokinetics involving strong binding to sex hormone binding globulin, while also displaying sensitivity to CYP inducers. Enzyme-inducing antiepileptic drugs (AEDs) accelerate the metabolism of OCs, thereby decreasing their serum concentration, and increasing the probability of ovulation. Conversely, combination OCs were shown to increase the clearance of some AEDs, such as lamotrigine and valproate. The occurrence of bidirectional interactions between AEDs and hormonal OCs pose the risk of contraceptive failure, or reduced seizure control. A previous study in 24 healthy women showed that brivaracetam 400 mg/day caused a 20–30% reduction in the exposure of ethinylestradiol and levonorgestrel. However, levels of endogenous hormones (estradiol, progesterone, luteinizing hormone, and follicle stimulating hormone) remained normal and similar throughout the two cycles with and without brivaracetam, demonstrating that ovulation did not occur in any individual. Here, we expand on previously published results by assessing the bidirectional pharmacokinetic interactions between brivaracetam 100 mg/day (the middle of the range investigated in phase III trials) and a combination OC containing 30 μg ethinylestradiol and 150 μg levonorgestrel, in healthy women.
Eligible participants were healthy, premenopausal women, aged 18–40 years (inclusive), nonpregnant and nonlactating, with a body weight of at least 50 kg and a body mass index (BMI) of 18–29 kg/m2. All participants had normal, regular menstrual patterns with no or minimal intermenstrual spotting, had a normal Papanicolaou (Pap) smear within 1 year prior to screening (Pap<3), were taking a combination OC containing 30 μg ethinylestradiol and 150 μg levonorgestrel (Minidril® or equivalent, e.g., Ludeal® or Microgynon 30®) for at least two consecutive menstrual cycles prior to start of study, and were willing to use additional barrier contraception during the study.
Excluded were women with gynecologic disorders, irregular bleeding patterns, or more than minimal intermenstrual spotting within 2 months prior to start of study. Also excluded were women with a history of anorexia; history or presence of significant cardiovascular, respiratory, hepatic, renal, gastrointestinal, endocrinologic, neurologic, or psychiatric disorders; chronic or acute illness; drug addiction; excessive alcohol or caffeine intake; and smoking. No other medication apart from the OC and occasional paracetamol was allowed within 14 days prior to start of study (2 months in the case of enzyme-inducing medication).
This phase I, randomized, double-blind, placebo-controlled, two-way crossover, multiple oral dose interaction study (N01282; EudraCT: 2007-000768-26) of brivaracetam (100 mg/day) and a combination OC (30 μg ethinylestradiol and 150 μg levonorgestrel) took place at two centers in France, between August 2007 and April 2008. The study was conducted over five menstrual cycles, each consisting of 28 days. After a 3-week screening period, data on demographics, medical history, and physical (including gynecological) examination were collected during a baseline period with OC only (Cycle 1). Eligible participants were randomized to receive either brivaracetam 100 mg/day or placebo in addition to the OC during two crossover treatment phases (Cycles 2 and 4), separated by a wash-out period with OC only (Cycle 3). The second treatment cycle was followed by a follow-up period with OC only (Cycle 5). Breakthrough bleeding/spotting events and tolerability were recorded in individual diaries from the first day of Cycle 1 until the last day of Cycle 5.
Participants took OC tablets on days 1–21 of each cycle at approximately 8 a.m.. Brivaracetam (100 mg/day) or matching placebo was administered twice daily in equal doses (50 mg b.i.d.) on days 1–28 of Cycles 2 and 4, at approximately 8 a.m. and 8 p.m. On days 18–28, morning and evening doses of brivaracetam were taken at least 1 h after a standard meal, in an upright (sitting) position, with 240 ml of water, following a strict administration interval of 12 h. On days 1–17, there were no meal restrictions, and a 2-h interval either side of the expected administration time was considered acceptable.
This study was conducted in accordance with the International Conference on Harmonisation Guidance on Good Clinical Practice and the U.S. Food and Drug Administration and European Medicines Agency guidance on the investigation of drug interactions. The protocol was approved by an independent ethics committee and all participants provided written informed consent to participate.
During Cycles 2 and 4, blood samples were taken for measurement of ethinylestradiol and levonorgestrel prior to the morning dose and at predefined intervals up to 24 h post-dose, on day 20. Blood samples were taken for measurement of brivaracetam prior to the morning dose on days 3, 8, 14, 18, 20, and 29 (day 1 of the next cycle), and, additionally, at predefined intervals up to 12 h post-dose on day 20. A strict 12-h interval was maintained between the evening administration of brivaracetam or placebo on days 13, 17, 19, and 28 and blood sampling time on the following day. Plasma was separated in a refrigerated centrifuge (approximately +4°C) within 30 min after sampling, at 1600 g for 10 min, and was subsequently stored at −20°C until analysis. Plasma concentrations of brivaracetam, ethinylestradiol, and levonorgestrel were measured as described previously.
Breakthrough bleeding episodes were recorded during all cycles using personal diaries. Safety and tolerability were assessed by observed or spontaneously reported adverse events (AEs), clinical laboratory tests, and vital signs.
Pharmacokinetic and statistical calculations
Assuming a residual variability of 21%, a sample size of 22 evaluable subjects was needed to demonstrate the lack of interaction for both ethinylestradiol and levonorgestrel with a power of at least 80%, provided that the true test/control means ratio was within 0.95 and 1.05. Considering the length of the study and taking into account possible drop-outs, the sample size was set at 28 subjects.
Pharmacokinetic parameters were calculated by noncompartmental methods using WinNonLin® version 5.2 (Pharsight Corporation, Sunnyvale CA, U.S.A.). For ethinylestradiol, levonorgestrel, and brivaracetam, maximum plasma concentration (Cmax), time to Cmax (tmax), area under the plasma concentration versus time curve over the dosing interval (AUCτ, where τ is 24 h for ethinylestradiol and levonorgestrel and 12 h for brivaracetam), and plasma clearance at steady state (CLss/F) were determined.
The effect of brivaracetam on ethinylestradiol and levonorgestrel pharmacokinetic parameters was assessed using analysis of variance (ANOVA) after logarithmic transformation, with treatment, cycle, and sequence as fixed-effect terms, and participants as random-effect term. Geometric least-squares means (LSMs) for AUCτ and Cmax, and the ratio (brivaracetam vs. placebo) of LSMs along with 90% confidence intervals (CIs) were calculated. Lack of pharmacokinetic interaction between brivaracetam and ethinylestradiol and levonorgestrel was concluded if the 90% CI for the brivaracetam versus placebo LSM ratio was fully contained within the bioequivalence range of 0.80–1.25 for AUCτ and Cmax.
The effect of the OC on brivaracetam was assessed using ANOVA, by comparing brivaracetam morning trough levels on day 20 (with OC) and day 29 (after the OC-free week).
Statistical calculations were performed using SAS® version 9.1.3 (SAS Institute Inc., Cary, NC, U.S.A.).
Twenty-eight participants were randomized. Overall, 23 participants completed the study, three discontinued due to AEs, one had a protocol deviation (OC intake omission for more than 12 h), and one withdrew consent.
Participants had a mean age (range) of 29.1 years (18.7–39.2 years), mean weight of 62.0 kg (52.0–77.5 kg), and mean BMI of 22.7 kg/m2 (19.3–27.6 kg/m2). None of the participants were current smokers or reported drinking alcohol, and 20 participants reported drinking caffeinated beverages, within the limits of the exclusion criteria. Based on their medical history, before the start of the study six participants had been pregnant, of whom two had had an elective abortion.
Ethinylestradiol and levonorgestrel
Ethinylestradiol and levonorgestrel geometric mean plasma concentration versus time profiles and pharmacokinetic parameters on day 20 were similar when the OC was coadministered with brivaracetam or placebo (Fig. 1, Table 1). There was an increase in the apparent plasma clearance of ethinylestradiol and levonorgestrel (+10.1% and +7.5%, respectively), and a corresponding decrease in AUCτ (−9.7% and −7.4%, respectively) during the brivaracetam cycle compared with the placebo cycle. However, for both ethinylestradiol and levonorgestrel, 90% CIs for AUCτ and Cmax geometric LSM ratios (brivaracetam vs. placebo) were entirely contained within the predefined bioequivalence range of 0.80–1.25 (Table 1).
Table 1. Pharmacokinetic parameters of ethinylestradiol and levonorgestrel when coadministered with brivaracetam or placebo
AUCτ, area under the plasma concentration versus time curve over the dosing interval; CI, confidence interval; CLSS/F, plasma clearance at steady state; Cmax, maximum plasma concentration.
Geometric mean (% coefficient of variation).
Point estimate (90% CI) for brivaracetam versus placebo geometric least-squares means ratio derived from analysis of variance.
AUCτ (pg h/ml)
AUCτ (ng h/ml)
Mean trough plasma concentrations indicated that brivaracetam reached steady-state by day 3. On day 20 of 100 mg/day dosing, brivaracetam reached a geometric mean Cmax (coefficient of variation) of 2.6 μg/ml (14.1%) at 1 h post-dose (median tmax range of 0.5–2 h); geometric mean AUCτ was 18.8 μg h/ml (13.6%), and CLss/F was 0.72 ml/min/kg (13.6%). Trough brivaracetam levels on day 29, after the OC-free week, were not significantly different from those on day 20, the last day of the cycle (geometric means: 988 vs. 1062 ng/ml; point estimate of ratio: 1.08; 90% CI: 0.98–1.18; p > 0.05).
Safety and tolerability
No obvious differences in breakthrough bleeding patterns were observed between different treatments or cycles. In particular, the incidences of prolonged menstrual bleeding (days 1–2) during the brivaracetam and placebo cycles (4/25 [16.0%] vs. 5/27 [18.5%]) and during the cycles following brivaracetam or placebo coadministration (3/24 [12.5%] vs. 4/24 [16.7%]) were comparable. The most frequently reported treatment-emergent AEs (TEAEs) during the brivaracetam versus placebo cycle were the following: dizziness (10/25 [40.0%] vs. 0/27 [0%]), somnolence (7/25 [28.0%] vs. 1/27 [3.7%]), headache (6/25 [24.0%] vs. 3/27 [11.1%]), and asthenia (5/25 [20.0%] vs. 1/27 [3.7%]).
TEAEs were generally mild or moderate in intensity. Three participants had TEAEs leading to permanent discontinuation of the study drug (dizziness and asthenia, rectal hemorrhage, and ankle sprain). There were no serious AEs or deaths in the study. No clinically significant changes in laboratory measurements or vital signs were recorded.
This two-way crossover study evaluated the pharmacokinetic interactions between brivaracetam 100 mg/day and a widely used combination OC containing 30 μg ethinylestradiol and 150 μg levonorgestrel, in healthy women. The brivaracetam dose selected for this study was within the expected therapeutic range (50–200 mg/day). Twenty-eight healthy women, with baseline characteristics representative of a fertile female population, were enrolled.
Plasma concentration versus time curves for ethinylestradiol and levonorgestrel were similar when the OC was coadministered with brivaracetam or placebo. For both ethinylestradiol and levonorgestrel, 90% CIs for AUCτ and Cmax geometric LSM ratios derived from ANOVA were entirely contained within the accepted range for lack of interaction (0.80–1.25).
Furthermore, the OC did not modify brivaracetam steady-state trough levels. Assessment of brivaracetam morning trough levels during OC coadministration (day 20) and after 1 week of OC wash out (day 29) demonstrated the absence of pharmacokinetic interaction of the OC on brivaracetam.
In healthy women, ethinylestradiol significantly reduced CYP2C19-dependent hydroxylation of omeprazole. Because brivaracetam is known to be partly metabolized by CYP2C19, it could be hypothesized that its metabolism may be altered by ethinylestradiol. However, the absence of detectable change in brivaracetam trough plasma concentrations after 1 week of OC wash out is consistent with previous reports that CYP2C19-dependent hydroxylation is a minor metabolic pathway for brivaracetam.
Breakthrough bleeding patterns showed no clinically significant abnormalities and the absence of changes in incidence across the five observed cycles and between the placebo and brivaracetam treatments. In particular, there was no change in the bleeding and spotting pattern when the OC was coadministered with brivaracetam or placebo. Brivaracetam 100 mg/day was well tolerated.
The findings of this study indicate that steady-state dosing of brivaracetam 100 mg/day had no clinically relevant effect on the disposition of either levonorgestrel or ethinylestradiol in healthy female participants. Furthermore, trough levels of brivaracetam were not affected by coadministration of the OC.
This study was sponsored by UCB Pharma. UCB Pharma was involved in the design and conduct of the study, and collection, management, and analysis of the data. The authors thank Muriel Boulton, MSc (UCB Pharma) for statistical assistance and Ioana Dumitrescu, PhD (QXV Communications, Macclesfield, UK) who provided writing support that was funded by UCB Pharma.
Conflicts of Interest
Armel Stockis and Shikiko Watanabe are employees of UCB Pharma. Nicolas Fauchoux is an employee of Biotrial, a contract research organization that received fees from UCB Pharma for performing the clinical trial. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.
Armel Stockis, PhD, is senior director CNS clinical pharmacology at UCB Pharma, Braine-l'Alleud, Belgium.