CONTRACEPTION

PHARMACOKINETIC COMPARISON OF TWO TRIPHASIC CONTRACEPTIVE FORMULATIONS CONTAINING LEVONORGESTREL AND ETHINYLESTRADIOL

ORAL

Frank Z. Stanczyk, PhD Rogerio A. Lobo, MD Soong T. Chiang, PhD Theodore B. Woutersz, MD Department of Obstetrics and Gynecology University of Southern California School of Medicine and Women's Hospital, Los Angeles County/USC Medical Center Los Angeles, California and Wyeth-Ayerst Laboratories Philadelphia, Pennsylvania Abstract The pharmacokinetics of levonorgestrel and (LNG) ethinylestradiol (EE2) were determined in 24 women (aged 21 to 35 years), following the administration of a single tablet from the second preparations (Triphasil (RPh~n~ o,fin~r~i~~~~~;fn'a~~i~~~~~~ contained 0.075 mg of LNG and 0.040 mg of EE2. The data were compared to the pharmacokinetics of LNG and EE2 obtained following the oral administration of a hydroalcoholic solution (standard) containing the same steroids and dose. The study consisted of a randomized design in which the three formulations were administered to each of the 24 subjects in a three-period crossover pattern. Blood samples were taken at frequent intervals after dosing. Serum levels of LNG and EE2 were measured by specific radioimmunoassays. The results show that both LNG and EE2 in the Triphasil and Trinordiol tablets are bioequivalent with respect to rate and extent of absorption. Furthermore, LNG, but not EE in both tablet formulations was bioequivalent to the so ?‘ ution dose. The serum concentration-time profiles for the three formulations showed that the range of mean peak 'levels was 2.3-2.8 ng/ml for LNG and 116-159 pg/ml for EE2. These levels were achieved within 2 hours in the majority of subjects. The ranges of mean values calculated for the areas under the curves were 15-16 ng.hr/ml for LNG and 10531390 pg.hr/ml for EE . The ranges of mean values calculated for other pharmaco t?inetic parameters were: volume of distribution: ING - 1.6-1.8 L/kg, EE2 - 7.7-9.1 L/kg; clearance: LNG - 84-88 ml/hr/kg, EE - 0.67-0.99 ml/hr/kg; half-life: LNG - 13-15 hr, EE2 - 7-1 3 hrs. Submitted for publication August 18, 1989 Accepted for publication September 25, 1989

JANUARY199OVOL.41NO.1

39

CONTRACEPTION

Introduction In recent years, an attempt has been made to reduce dose in combination oral further the progestogen contraceptives by development of three-phase, low-dose varying proportions of preparations containing the It has been shown that progestogen and estrogen components. a triphasic formulation, containing levonorgestrel (LNG) and (marketed in the Unite ;thh;Xy;;@pdi~;d(E~) Trinordiol('))Sta% "E Europe as predictable, efficacious, well-tolerated and accepied means contraception (1). There is, however, no of oral information on the pharmacokinetics of low-dose LNG (0.075 mg) which is used in combination with EE2 (0.040 mg) during and the second phase (5 days) of the Il-day Triphasil Trinordiol regimen. The objectives of the present study were: a) to determine the pharmacokinetics of LNG and EE2 in women, following oral administration of a single second phase Triphasil or Trinordiol tablet: b) to compare these data with the LNG and EE2 pharmacokinetics determined following oral administration of a hydroalcoholic solution (standard) containing the same steroids and doses as the tablets: c) to determine whether Triphasil and Trinordiol are bioeguivalent with respect to the second phase tablets. Materials

and Methods

Study Population Twenty-four healthy women ranging in age from 21 to 35 years (mean + SE; 31 + 1.0 years) were studied. They were within 10% of ideal body weight for age and height, and their weights ranged between 48 and 71 kg (mean + SE, 57 f. 1 Excluded from participation were subjects who had any kg). contraindication to the use of oral contraceptives, or had been pregnant within 60 days, or had used reproductive hormones within 15 days or any other drug within 7 days prior to the study. None of the subjects smoked when admitted to the study or during the study. The subjects underwent a pretreatment laboratory screening and physical examination to determine the adequacy of their state of health. Institutional review board approval was obtained, and written informed consent was given by each participant, Study Design Following an overnight, absolute fast of 12 hours, each subject was administered either a single tablet of Triphasil (second phase: Wyeth Laboratories, Philadelphia, PA) or Trinordiol (second phase: Wyeth-Phanna, Germany), or 30 ml of a hydroalcoholic solution (standard), each containing 0.075 mg of LNG and 0.040 mg of EE2. The standard solution

40

JANUARY

1990 VOL. 41 NO. 1

CONTRACEPTION

prepared and furnished by Wyeth Laboratories in individual vials. The solution consisted of 7 ml of 95% (V/V) ethyl alcohol, which was used to dissolve the steroids, and 23 ml of distilled water. Each subject's vial was rinsed with 25 ml of water, and the rinse was administered to the subject. An additional 100 ml of water was taken by each subject to eliminate any after-taste. The subjects taking the tablets drank 155 ml of water concomitantly. After the subjects took the medication, an absolute fast was sustained for an additional 4 hours. Thereafter, the subjects were permitted the following diet: a light lunch, a mid-afternoon snack, and a hot dinner. Ad libitum use of food and/or drink was prohibited until 10 hours post-dosing. The subjects were generally ambulatory on study days, and were not permitted to engage in any strenuous or athletic activities.

was

The study consisted of a randomized open single oral dose design in which the 3 formulations were administered to each of the 24 subjects in a 3-period crossover pattern. Drugs were given in random fashion according to a code key based upon a computer-generated table of random numbers. The subjects were assigned to one of 3 seguelae groups and received a different formulation at 8:00 am on either cycle day 7, 8 or 9 in each of 3 consecutive menstrual cycles. The schedule of drug administration for this three-period crossover study was based on a Latin-square design. Blood samples were drawn immediately before drug administration (0 hours) and at 0.5, 1, 1.5, 2, 4, 6, 8, 12, 24, 48, 72 and 96 hours after administration. Each sample consisted of approximately 12 ml of venous blood which was collected from an antecubital vein by means of an indwelling non-anticoagulated red heparin-lock needle into top The blood was allowed to clot for 60 vacutainer tubes. minutes and was centrifuged. The serum was transferred to new tubes and then frozen rapidly and stored at -20° C. Assays Levonorgestrel and ethinylestradiol were measured in serum using radioimmunoassays (RIAs) that were developed and validated in our laboratory. Our previously described LWG assay method $2) was modified slightly. The assay procedure Germany) instead now utilizes H-LWG (Schering AG, B@in, of LWG-3-(O-carboxymethyl)oximino- I-iodohistamine, and procedural losses are followed by addition of tritiated internal standard (1000 dpm of 3H-LNG in 0.1 ml of assay buffer) to each sample prior to the extraction step, instead of using an external standard. present LNG criteria for the The reliability radioimmunoassay show the following: Logit B/Be vs log dose

JANUARY

1990 VOL. 41 NO. 1

41

CONTRACEPTION

standard curves were linear between 3.9 and 500 pg per tube of LNG. The lowest amount of LNG per RIA tube which reduced the number of cpm of labeled LNG bound at 0 mass by 2 Buffer and serum blanks standard deviations was 3.9 pg. included in each RIA were indistinguishable from 0. Based on an average procedural recovery of 70% and the factors accounting for serum and extract aliguots used in the assay, these results indicate that the limit of detection by this RIA is 0.11 ng per milliliter of serum. Accuracy of the LRG RIA was assessed by recovery experiments in which various amounts of ING were added to control serum, yielding serum concentrations of 0.5, 1, 3, and 5 ng/ml. The LNG levels were then measured by RIA. Twelve determinations were made at each concentration. Linear regression analysis of the LNG concentrations measured (y) versus the concentrations of LNG added (x) yielded the equation, y = 1.02 x - 0.12, with a correlation coefficient of r = 0.999. Measurement of LNG at concentrations of 0.5 and 3 ng/ml gave an intraassay coefficient of variation of 7.0 and 6.3%, respectively, for 3 determinations of each concentration. The interassay coefficients of variation for LNG concentrations of 1 and 3 were 13.9 and 12.4%, respectively, for 12 ng/ml determinations of each concentration. Since the previous LNG RIA (using the iodinated radioligand) was shown to be highly specific, we compared LNG values measured in a set of serum samples to those measured by the present LNG RIA (using the tritiated radioligand) in the same set of Linear regression analysis of 14 pairs of serum samples. LRG levels measured by the two methods gave a correlation coefficient of r = 0.960. BE2 levels were measured in l-2 ml aliguots of serum by specific radioimmunoassay following diethyl ether zxtraction and celite column chromatography (3). Procedural losses were monitored by addition of dependent internal standard (3H-EE2) to each sample prior to extraction. An antiserum against ethinylestradiol-7-(3-thiopropionic acid)$ovine serum albumin (4) was employed in conjunction with H-EE2 in the RIA. The limit of reliable detection of EE2 in this assay was 20 pg/ml. The appropriate baseline BE2 value was subtracted from each sample BE2 value. Phannacokinetic

and Statistical

Analyses

Analysis of the LNG and EE2 data obtained for each subject was carried out using a non-compartmental pharmacokinetic model, which included an assumed first-order absorption process (5). The following equation was used to fit both LNG and EE2 data: - X,t C = Ale

-k,t + A2:X2t

- (A1+A2)e

JANUARY

1990VOL. 41 NO. 1

CONTRACEPTION

In this equation, C represents the serum concentration of ING or EE at a given time, t; Al and A2 are the hybrid coefficien % s; X and > are the initial and terminal disposition rates', respecttvely: k, is the absorption rate constant. The following pharmacokinetic parameters were computed: (a) maximum serum concentration (Cmax); (b) time when C was achieved (tmax); (c) k which was determined from B% above pharmacokinetic equa9'. ion; (d) mean absorption time (tabs)l which was estimated from the equation, tabs = l/k + is the lag time before absorption (6); Te) a9 ea under the serum concentration-time curve (AUC) calculated by the log-trapezoidal rule to the last sampling time (96h); (f) apparent clearance (CL/F, where F is the unknown systemic availability), calculated from the ratio of dose/AUC; (g) terminal phase disposition rate constant (AZ); (h) apparent terminal phase volume of distribution (Vd), which was determined by the equation, = Cl/F; (i) terminal phase half-life, which was VdXz calculated from the equation, (j) relptive which ;td2 = 0:693/x,; bioavailability (Fr), obtained by the ratio of AUCtablet/ AUCsolution' Mean serum and EE LNG concentrations and pharmacokinetic parameter values Zor the three different formulations were compared by analysis of variance of the log-transformed values using a 3-period crossover design. Results Serum Concentrations

and Pharmacokinetics

of LNG

The serum concentration-time profile of LNG following administration of the tablets or solution in the 24 women is shown in Table I and Figure 1. Mean baseline LNG concentrations were similar for all 3 formulations and were less than 0.03 ng/ml. At the 30-minute sampling interval, the LNG levels obtained following administration of the 2 different significantly tablets were lower than the corresponding levels attained with the solution dose. Also, the Trinordiol tablet gave significantly lower LNG levels compared to the Triphasil tablet and the solution dose at the l- and 1.5hour sampling intervals, respectively. pharmacokinetic The multi-exponential equation was found to fit the serum LNG data 2atisfactorily (Figure 1) (R ) value of approximately reaching a mean goodness-of-fit Comparison of the 3 0.97 for all three formulations. formulations with respect to the mean pharmacokinetic parameters of LNG (Table II) showed significantly lower and significan%a : for Trinordiol than for the solution, higher t,,, and tabs for both tablets than for the solution.

JANUARY

1990VOL. 41 NO. 1

43

CONTRACEPTION

means and ranoes of serum levonorcestrel (LNGI Table I. Geometric concentrations, following the administration of a single oral‘dosh of LNG (0.075 mg) in combination with ethinyleitradiol (0.040 mg) as a tablet fTriohasi.1 or Trinordioll and as a solution in 24 women

Time After Rose (h)

Geometric Means and Ranges of Serum LNG Concentrations (n&ml) Paired Solution Comva&aons* 3riDhasi.l Trinordiol

0.5

1.52 (O-21-3.50)

1.16 (0.20-4.36)

2.67 (1.24-5.36)

A=B