SESSION I. PHARMACOKINETICS OF ORAL CONTRACEPTIVE STEROIDS Chairman: K. Fotherby

Pharmacokinetics of ethinyl estradiol and mestranol Joseph W. Goldzieher, MD, and Steven A. Brody, MD Houston, Texas Pharmacokinetally, a 50 f,Lg oral dose of mestranol (which itself is inactive) is bioequivalent to a 35 f,Lg dose of ethinyl estradiol. Physiologically, mestranol ranges from 50% to 100% of the activity of ethinyl estradiol, depending on the endpoint chosen. Compounds such as these, which are metabolized with a first-pass effect and are enterohepatically recirculated, demonstrate large interindividual and intraindividual variability in their pharmacokinetics. Thus a given dose of ethinyl estradiol in one person may produce an effect equivalent to a substantially larger (or smaller) dose in another person. This wide variability confounds efforts to establish tight dose-response relationships, a pOint rarely considered in clinical or epidemiologic studies of these compounds. The circulating levels of ethinyl estradiol sulfates may be higher than those of free ethinyl estradiol itself. It has been thought that these sulfates represent a "reservoir" of ethinyl estradiol. Our studies show that this idea is untenable because the half-life of the sulfates is not long enough for such an effect. Differences in the pharmacokinetics of ethinyl estradiol and mestranol have been observed in studies of various populations. The reality of these group differences is affirmed by analyses of urinary metabolite patterns. (AM J OSSTET GVNECOL 1990;163:2114-9.)

Key words: Ethinyl estradiol, mestranol, pharmacokinetics All oral contraceptives currently in use rely on ethinyl estrogens for their estrogenic component. The reason for this is the greatly increased pituitary-inhibiting activity relative to other estrogenic effects, that is imparted by the ethinyl group. In addition, studies have shown that the ethinyl estrogens act synergistically with the 19-nor progestins with respect to gonadotropin suppression. Whether a synergism of non-ethinyl estrogens with these progestins exists is uncertain. This amplification is the basis on which the contraceptive effectiveness of the very low-dose formulations rests, and the reason why the dosage in the original "sequential" regimens could not be lowered to currently desired levels. Some of the adverse symptomatic effects of the estrogens are clearly dose related. Therefore it becomes important to have an in-depth understanding of their pharmacokinetics.

Mestranol In vitro receptor studies l have demonstrated that mestranol is an inactive compound that becomes biologically active on conversion to ethinyl estradiol (EE). The degree of conversion varies with different species. From the Department of Obstetrics and Gynecology, Baylor College of Medicine. Reprint requests: Joseph W. Goldzieher, MD, Department of Obstetrics and Gynecology, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030. 6/0/23775

2114

Published human studies of the pharmacokinetics of EE derived from mestranol generally suffer from the small number of subjects used (usually < 10) and from insufficient sampling frequency. Thus the general impression of a 50% conversion rate of mestranol to EE is based on rather uncertain grounds. In view of what is known today about the magnitude of interindividual and intraindividual variation with these compounds, a reinvestigation is clearly needed. We 2 have recently examined the pharmacokinetic parameters derived from analysis of plasma EE levels after administration of a single dose of three bioequivalent norethindrone (1 mg)/EE (35 fLg) formulations from different manufacturers, as well as three norethindrone (1 mg)/mestranol (50 fLg) formulations from the same manufacturers. The protocol was designed as an open-label, three-way crossover study. Each subject received a two-tablet dose of a drug within the first 6 days of three consecutive menstrual cycles. This double dose was used to improve assay reliability. The drugs were assigned in a randomized sequence so that all three drugs were tested in each subject at a giveh estrogen dose. Twenty-four women took each of these EE-35 formulations; another 27 took the three mestranol-50 formulations. Some of the derived phar~ macokinetic parameters are shown in Table I. The data clearly show the delay in achieving maximum plasma EE levels derived from mestranol compared with EE as expected. Interestingly, the peak plasma concentration (e max ) values for EE derived from

Pharmacokinetics of ethinyl estrogens

Volume 163 Number 6, Part 2

2115

Table I. Pharmacokinetic parameters for EE after oral administraion of single two-tablet doses of each oral contraceptive Formulation Variable

Dose

AVC (pg-hr/ml)

1I35EE 1/50ME 1I35EE 1/50ME 1/35EE ]/50ME

Cmax.

(pg/ml) Tmax

(hr)

Kel

I

1I35EE ]/50ME

1089 996 171 175 1.3

± ± ± ± ± 1.7 ±

OrthoNovum

I

NOYr£pt 570 454 57 57 0.5 0.7 P = 0.008 0.08 ± 0.07 0.13 ± 0.10

993 940 178 175 1.3 1.8

± ± ± ± ± ±

P=

0.09 ± 0.18 ±

P=

270 494 45 69 0.6 0.8 0.02 0.07 0.13 0.04

I

Norinyl 1024 983 174 175 1.4 2.0

± ± ± ± ± ±

454 429 59 50 0.6 0.8 P = 0.003 0.11 ± 0.07 0.13 ± 0.10

Total 1036 963 174 175 1.3 1.9

± ± ± ± ± ±

483* 544t 67 72 0.5 0.8 P = < 0.0001 0.09 ± 0.09 0.15 ± 0.12

Reproduced with permission from Brody]A, Turkes A, Goldzieher ]W. Contraception 1989;40:269-84. Values shown are mean ± SD for area under the plasma concentration/time curve (AVC o.,4) by the trapezoidal rule; peak plasma concentration (C max ); time to peak concentration (T m ,,); and the elimination rate constant (1(,,1). *n = 24 x three trials. tn = 27 x three trials.

Table II. Interindividual variability of EE plasma concentrations after administration of two 35 Il-g EE and two 50 Il-g mestranol pills Parameter EE (35 fLg EE pills) AVC o/24 (pg-hr/ml) C max (pg/ml) EE from mestranol (50 fLg mestranol pills) AVC o.2 • (pg-hr/ml) Cm" (pg/ml)

Mean ± SD

Range

CV (%)

1036 ± 483 174 ± 67

284-2498 55-311

47 39

963 ± 544 175 ± 72

215-2122 67-391

57 41

Data from Brody]A, Turkes A, Goldzieher ]W. Contraception 1989;40:269-84. CV, Coefficient variation.

Table III. Range of pharmacokinetic values for EE reported in the literature tl;2 (at) tl;, ([3) tl;2 (Ka) Bioavailability K!2 K2! KIO Tmax

0.5-2.4 hr 13.1-27.0 hr 0.2-0.4 hr 0.38-0.48 0.182-0.249/hr 0.101-0.245/hr 0.193-0.309/hr 1-2 hr

Estimate of tl;, (u) has a wide range and needs further documentation. Ranges of values indicated are derived from Newburger and Goldzieher.!6 tl;2 (u) is the half-life of drug disposition from the central to the peripheral compartment. tl;2 ([3) is the elimination half-life. tl;2 (Ka) is the half-life of drug delivery to the central compartment, where Ka is the transfer rate constant in this direction. K!2 is the transfer rate constant from the central to the peripheral compartment; K2! is the transfer rate constant in the reverse direction. KIO is the same as the elimination rate constant (KeI)' Bioavailability is calculated by dividing AVC after oral administration by AVC after intravenous administration with the same dose of drug. T m.x is the time of maximum plasma concentration of drug. The theoretic model for these pharmacokinetic parameters is described in Goldzieher et al. 1I

2116

Goldzieher and Brody

December 1990

Am

E c5

'6

...... .-.. ........ -

300

~., 250

w 200 >. c:::

:c

Highest response Lowest response Mean (n=24) ±1 SO

CV

Dose: 2 tabs p.o. at time=O

150

W 100 ~II>

~

Table IV. Intraindividual variability of EE plasma concentration AUC o2 , after the administration of two 35 f1g EE and 50 /Lg mestranol pills: three single-dose trials

400

g, 350 Q.

....... .........

50 ~~s:::::::;;~~ ~~ 0

1 2

6 8 10 12 24 Hours Following Drug Administration

4

J Obstet Gynecol

30

...• 48

Fig. 1. Plasma EE contrations after administration of N orceptE1/35. A single dose of two tablets consisting of EE 35 fJ.g/norethindrone I mg was given by mouth at time = O. The mean ± SD is depicted along with highest and lowest individual responses. (Reproduced with permission from Brody SA, Turkes A, Goldzieher jW. Contraception 1989;40:26984.)

50 /Lg of mestranol are identical with those derived from 35 /Lg of EE, both of which were given along with 1 mg of norethindrone. The mean area under the plasma concentration/time curve from 0 to 24 hours (AUC n. e,) values for EE derived from 50 /Lg of mestranol are clearly lower than those derived from 35 /Lg of EE as shown, but the very large intersubject variability keeps this difference from being statistically significant. From these data it may be concluded that EE from mestranol would yield about 70% of the C max value of a similar dose of EE. Based on AUC o. e4 , mestranol is somewhat less than 70% as bioavailable as EE. It appears that the usual estimate of 50% is probably low in terms of AUC, The interindividual variation of plasma EE levels derived from EE or mestranol was similar; the coefficient of variation of AUC for EE was 47%; that for EE from mestranol was 57%. The coefficient of variation of C max for EE was 39%; for EE derived from mestranol it was 41 % (Table II), These pharmacokinetic data are not necessarily concordant with pharmacodynamic observations. Studies of human endometrial response and measurements of hepatic synthesis of certain proteins suggest a potency of about 50% for mestranol compared with EE."·' However, other studies of human endometrial response have found that the two compounds are bioequivalent over the range of 50 to 100 /Lg/day. Furthermore, studies of the effects on plasma gonadotropins, cortisol, testosterone, and androstenedione and their binding globulins and the effects on carbohydrate and lipid metabolism also found bioequivalence. These studies were conducted with estrogens alone rather than with commercial oral contraceptive formulations. It must be added, however, that these studies were not carried out on the numbers of patients we require in light of con-

EE (35 fJ.g EE pills) EE from mestranol (50 fJ.g mestranol pills)

(%J

41 42

Range*

-66%-+71% -70%-+92%

Data from Brody .lA, Turkes A, Goldzieher .JW. Contraception 1989;40:269-84. CV, Coefficient of variation. *The percent difference of high and low values from the mean of three trials per subject.

temporary knowledge of interindividual variability.n.1O These findings are highly relevant to epidemiologic studies, which frequently analyze data sets in terms of the quantity of estrogen in contraceptive formulations without regard to the chemical nature of the estrogen. Thus some "high-dose" mestranol pills have a lower biologically active estrogen content than do some "lowdose" EE pills. 2 EE

The literature on the pharmacokinetics of EE is substantial. There are large discrepancies in the values reported for the various parameters. These are probably because of limitations in sampling frequency, difficulties in measuring of plasma EE levels at clinically relevant dosages, and the effects of enterohepatic recirculation. The large interindividual variation seen with these drugs is a major confounding factor, accounting for the discrepant findings reported. Analysis of these difficulties has been presented elsewhere." Our best estimate of the relevant pharmacokinetic parameters is shown in Table III. As Table III indicates, the half-lives for EE (tI;2-J 0-

w

'"

9

.

,.

~

'"

:I:

A280nm

N

"'

~

oJ

0W

~

!£

N

W

jw-

~

'"

~

"'0

:I:

NATURAL - d-homo

ON

~w

~

N

"'

~

~

w

:I:

9

l.

'"

I, "

10

20

30

40

50

60

70

80

90

100

110

120

130

140

150

RETENTION TIME (MINUTES-FRACTION NO.)

Linear gradient 6 -158 minutes 2.25 % Isopropanol in heptane

15 % Isopropanol in heptane

1

10

Fig. 2. Pattern of reference EE and non-ethinyl estrogens on a high-performance liquid chromatography. Chromegaprep Diol column, 9.6 mm by 50 cm, 10 f1; flow-7.5 mUmin; pressure, 1000 psi; linear gradient, 2.25 f1 to 15% isopropanol in heptane in 158 minutes; detection by ultraviolet absorption at 280 nm. (Reproduced with permission from Butterworth-Heinemann. Williams MC, Goldzieher, Jw. Steroids 1980;36:255-82.)

Table V. Representative plasma AUC o24 values (pg-hr/ml) for EE and norethindrone after repeated trials in the same persons Ethynyl Estradiol Subject No.

Norcept

I

Ortho-Novum

2 9 12 13 20

1002 252 1892 1563 657

431 499 637 539 1594

104 106 3 16 25

215 1355 768 1033 433

1230 854 963 839 289

1

Norethindrone Nonnyl

Norcept

I + 35 EE formulations 855 29.2 569 32.2 1390 79.7 98.9 934 2122 90.3 I + 50 ME formulations 658 30.3 1080 175.0 696 109.8 818 58.5 331 68.1

1

Ortho-Novum

I

Nonnyl

23.1 55.7 77.2 169.9 86.3

49.3 46.8 91.7 127.2 88.0

73.7 33.9 86.6 38.4 62.1

120.5 15.6 31.6 109.0 62.5

Reproduced with permission from Brody JA, Turkes A, Goldzieher JW. Contraception 1989;40:269-84.

both the 35 f.Lg EE and the 50 f.Lg mestranol preparations, there was a large intraindividual coefficient of variation: 41% to 42%. Within individual subjects the range of AUC varied enormously, from -70% to + 92% of the mean of the three trials. Typical examples of this variability are shown in Table V. The unexpectedly large intraindividual variance, which showed

no product-specific tendency, is apparent from the table. Moreover, the rank order of values for AUC r[ in the three trials per individual does not correlate with the rank order of the three results for AUC"n (norethindrone), a point that has already been made by Fotherby.13 A prevailing opinion has been that the conjugates of

2118

Goldzieher and Brody

December 1990 Am J Obstel Gynecol

W W

W

6000 4000

SRI LANKA 3000

4000

:;; 0

E thynylestrodtol

..!-

",I Q.

NIGERIA

5000

E thynylestrodlol

:;; Q.

2

3000

0

W W

2000

I

.,'?

WW

1000

'f!

.:;

W W

1000

W

w

I

I

?

0

~ 20

GO

80

100

120

140

20

IGO

40

FRACTION NUMBER

80

60

FRACTION

~ 100

120

140

IGO

140

160

NUMBER

w

15,000

UNITED STATES

sooo

UNITED STATES

Ethynylestradlol

4000

E thynylestrodlOI

,;;

",I

w w

:;; 3000

0

0

~

Q.

10,000 W

",I

,;;

.

:;; Q.

,;;

0

"'

2000

,;; 5000

W W

~

~

I

? W

1000

oW WW

II

00

N

20

40

60

80

100

120

140

160

FRACTION NUMBER

20

40

60

80

100

;,

~

"'"' 120

FRACTION NUMBER

Fig. 3. Representative patterns of urinary mestranol and EE metabolites in women from Sri Lanka, Nigeria and the United States. High-performance liquid chromatography Chromegaprep Diol column, 9.6 mm by 50 cm; 10 flo; flow-7.5 mllmin; pressure, 1000 psi; linear gradient, 2.25% to 15% isopropanol in heptane in 158 minutes, scintillation counting of I-minute fractions. Results from the two U.S. women illustrate the wide range of oxidative metabolism seen in this population. (Reproduced with permission from Butterworth-Heinemann. Williams MC, Goldzieher ]W. Steroids 1980;36:255-82.)

EE, specifically the 3-sulfates, have relatively long halflives and through enterohepatic recirculation might be deconjugated to provide a slow-release reservoir of EE. We have synthesized the various sulfates and have carried out clinical trials with these substances. A detailed analysis of the pharmacokinetic parameters has been published. 14 In summary, only 3.4% of intravenously administered and 11.4% of orally administered 17sulfate appeared in the blood as free EE; with the 3sulfate the conversion rates were 13.7% and 20.7%, respectively, which suggests that the sulfates are not important reservoirs. Moreover, the half-lives of free and sulfoconjugated EE are similar, ranging from 8.8 to 11.2 hours. Finally, the tII2 of the 17-sulfates after intravenous administration was similar to that of free EE. As expected, much more individual variation was encountered with oral administration than with the intravenous route. This phenomenon was apparent for both the 3- and 17 -sulfates, as reflected in measurement of the total circulating EE sulfates.

Ethnic differences

Several investigators 15- l7 have studied the pharmacokinetics of contraceptive steroids in various populations. Substantial differences in bioavailability have been noted. In comtemporary perspectives, the sample sizes studies and the blood sampling schedules used raise questions about the impact of interindividual variability as a possible confounding factor in the interpretation of these data. Although evidence of quantitative differences between populations is perhaps now less certain than was formerly thought, qualitative studies unequivocally support the conclusion that these population differences are real. In published studies l8 we have analyzed the pattern of urinary conjugates (glucuronides and sulfates) of EE after oral administration of radioactively labeled material to women in the United States, Sri Lanka, and Nigeria. In the three populations studied, the proportions of glucuronides and sulfoconjugates were similar: about 70% and 18%, respectively. However, the patterns of glucuronide con-

Pharmacokinetics of ethinyl estrogens

Volume 163 Number 6, Part 2

jugates seemed to differ. Further evidence of metabolic differences between populations can be demonstrated by examination of the oxidative metabolites of EE itself. High-performance liquid chromatography profiling permits high-resolution identification of the various metabolites of EE (Fig. 2). There was a consistent difference, with Nigerian women (n = 10) displaying the least, Sri Lanka women (n = 10) intermediate, and American women (n = 6) the highest degree of oxidative metabolism, as evidenced by multiple chromatographic peaks identifying distinct metabolic products (Fig. 3). Such findings rule out the possibility that the observed pharmacokinetic differences between ethnic groups were from the confounding effects of interindividual variation. More important, they raise interesting questions regarding ethnic differences in the enterohepatic metabolism of EE. Further studies will be needed to address the role of nutritional factors, intestinal bacterial flora, genetic steroid-metabolic differences, and other as yet undiscovered influences on the metabolism of these compounds. REFERENCES I. Kappus H, Bolt HM, Remmer H. Affinity of ethynylestradiol and mestranol for the uterine estrogen receptor and for the microsomal mixed function oxidase of the liver.] Steroid Biochem 1973;4:121-8. 2. Brody SA, Turkes A, Goldzieher ]W. Pharmacokinetics of three bioequivalent norethindrone I mestranol-50f.\-g and three norethindrone/ethynyl estradiol-35f.\-g OC formulations: are "low-dose" pills really lower? Contraception 1989;40:269-84. 3. Delforge ]P, FerinJ. A histometric study of two estrogens: ethinylestradiol and its 3-methyl-ester derivative (mestranol); their comparative effect upon the growth of the human endometrium. Contraception 1970;1:57-63. 4. Teter ], Stupnicki R. A comparative study of the estrogenic potential of two synthetic estrogens (mestranol and ethinylestradiol). Acta Cytol 1971; 15: 167-70. 5. Goldzieher ]W, Maqueo M, Chenault CB, Woutersz TB. Comparative studies of the ethynyl estrogens used in oral contraceptives. I. Endometrial response. AM] OBSTET GvNECOL 1975; 122:615-8. 6. Goldzieher ]W, de la Pena A, Chenault CB, Cervantes A.

7.

8.

9.

10.

II. 12.

13. 14. 15. 16. 17. 18.

2119

Comparative studies of the ethynyl estrogens used in oral contraceptives. III. Effect on plasma gonadotropins. AM ] OB5TE1' GVNECOL 1975;122:625-36. Goldzieher ]W, Chenault CB, de la Pena A, Dozier TS, Kraemer DC. Comparative studies of the ethynyl estrogens used in oral contraceptives: effects with and without progestational agents on plasma cortisol and cortisol binding in humans, baboons, and beagles. Ferti! Steril 1977;28:1182-90. Goldzieher ]W, Chenault CB, de la Pena A, Dozier TS, Kraemer DC. Comparative studies of the ethynyl estrogens used in oral contraceptives: effects with and without progestational agents on plasma androstenedione, testosterone, and testosterone binding in humans, baboons, and beagles. Fertil Steril 1978;29:388-96. Goldzieher ]W, Chenault CB, de la pena A, Dozier TS, Kraemer DC. Comparative studies of the ethynyl estrogens used in oral contraceptives. VI. Effects with and without progestational agents on carbohydrate metabolism in humans, baboons, and beagles. Fertil Steril 1978;30: 146-53. Goldzieher ]W, Chenault CB, de la Pena A, Dozier TS, Kraemer DC. Comparative studies of ethynyl estrogens used in oral contraceptives. VII. Effects with and without progestational agents on ultracentrifugally fractionated plasma lipoproteins in humans, baboons, and beagles. Fertil Steril 1978;30:522-33. Newburger], Goldzieher ]W. Pharmacokinetics of ethynyl estradiol: a current view. Contraception 1985;32:3344. Back D], Breckenridge AM, Crawford FE, Maciver M, Orme ML'E, Rowe PH. Interindividual variation and drug interactions with hormonal steroid contraceptives. Drugs 1981;21:46-61. Fotherby K. Pharmacokinetics of ethynyloestradiol in humans. Methods Find Exp Clin Pharmacol 1982;4:13141. Goldzieher ]W, Mileikowsky G, Newburger ], Dorantes A, Stavchansky SA. Human pharmacokinetics of ethynyl estradiol3-sulfate and 17-sulfate. Steroids 1988;51 :63-79. Goldzieher ]W, Dozier TS, de la Pena A. Plasma levels and pharmacokinetics of ethynyl estrogens in various populations. I. Ethynylestradiol. Contraception 1980;21:1-16. Fotherby K, Akpoviroro ], Abdel-Rahman HA, et al. Pharmacokinetics of ethynyloestradiol in women from different populations. Contraception 1981 ;23:487-96. Fotherby K. Variability of pharmacokinetic parameters for contraceptive steroids. ] Steroid Biochem 1983; 19:817-20. Williams MC, Goldzieher ]W. Chromatographic patterns of urinary ethynyl estrogen metabolites in various populations. Steroids 1980;36:255-82.