Acta Obstet Gynecol Scand Suppl 152: 25 - 31, 1990

EFFICACY A N D CLINICAL PROFILE OF A NEW ORAL CONTRACEPTIVE CONTAINING NORGESTIMATE U.S. clinical trials Stephen L. Corson From the Philadelphia Fertility Institute, Pennsylvania Hospital, University of Pennsylvania School o f Medicine, Philadelphia, Pennsylvania

Abstract. The search for a highly selective progestin that exerts a potent and targeted progestational response with minimal or absent androgenic effect has paralleled investigation into the possible impact of these agents on cardiovascular disease in women. Such a progestin, norgestimate (NGM), in a dose of 250 pg, has been combined with ethinyl estradiol (EE) 35 pg in a new oral contraceptive (OC), Ortho-Cyclen or Cilest. Results of t w o long-term, multicentric clinical trials demonstrate that this formulation is comparable in efficacy to the norgestrel-containing OC Lo/Ovral. There were no statistically significant differences in pregnancy rate, and both OCs were welltolerated in a large and diverse study population. In several areas, however, the inherently lower androgenicity of the norgestimate OC produced clinical changes compared with the norgestrel formulation. These changes were primarily evident in the more natural menstrual patterns with the norgestimate OC, its less severe impact on the endometrium, and, most important, its positive impact on lipoprotein metabolism. NGM/EE consistently produced statistically significant increases in high density lipoprotein (HDL) and concomitant improvement in the ratio of low density lipoprotein (LDL) to HDL. By cycle 24, this highly predictive parameter of atherosclerotic risk had decreased 7.7% in the NGM/EE group. Conversely, the norgestrel-containing formulation resulted in statistically significant decreases in HDL and increases in the LDL/HDL ratio: by cycle 24, these patients showed an 18.5% increase in the LDL/HDL ratio. All between regimen comparisons of mean changes in the LDL/HDL ratio were statistically significant, from baseline through cycles 3, 6, 12 and 24.

Early observations in lipoprotein research included the finding that patients with coronary heart disease had lower levels of high-density lipoprotein (HDL) cholesterol than matched healthy controls (1). The role of HDL in reducing atherogenic risk was further delineated in the classic report of Miller and Miller (2), and subsequent population studies (3-6) conducted in the late 1970s. During the past decade, extensive investigation has identified not only the major risk factor for cardiovascular (CV) disease, low-density lipoprotein (LDL) cholesterol, but further clarified the cardioprotection provided by HDL. A finding of major health consequence has been the correlation between alterations in lipoprotein metabolism and the use of certain oral contraceptives (OCs). With continued study, it is becoming clear that even small changes in lipid levels may have a profound

impact on CV risk. This is particularly important considering the large population of OC users and the impact that even small changes in CV risk may have in this group. Changes in lipoprotein metabolism are attributed to both the estrogen and progestin components in OCs (6). Estrogen decreases LDL and increases HDL and HDL,; most progestins now in clinical use have the opposite effect. The extent of these potentially deleterious effects of progestins on lipoprotein metabolism is modulated by the individual potency and dose of estrogen and progestin in a combination OC and on the inherent androgenic activity of the progestin (7-9). Hence, the search for a highly selective progestin that elicits a potent and targeted progestational response without androgenic effect has been a vigorous one. According to numerous preclinical assays, norgestiActa Obstet Gynecoi Scand Suppi 152

26

S. L. Corson

mate demonstrates a unique activity profile, and achieves the criteria of high specificity and selectivity. While it is strongly progestational and anti-estrogenic, norgestimate is relatively free of androgenic activity as measured by in vitro androgen receptor binding, stimulation of ventral prostate growth, and in vivo affinity for human sex hormone binding globulin (SHBG) (10).

MATERIALS AND METHODS Norgestimate (NGM) in a dose of 250 pg was combined with 35 pg ethinyl estradiol (EE) in a new oral contraceptive OrthoCyclen or Cilest (NGM/EE), and was compared in clinical use with a reference preparation containing norgestrel (NG) 300 pg plus EE 30 pg (Lo/Ovral). Results from two multicenter controlled clinical trials are combined in this report. They represent findings in 1,473 women (736 and 737 for NGM and NG, respectively) over 19,078 menstrual cycles (9,351 and 9,727 for NGM and NG, respectively). Both studies involved the administration of medication for 24 menstrual cycles, with a 90-120 day posttherapy evaluation. They were conducted at 28 sites throughout the United States to reduce demographic bias, and were representative of a large and diverse population of healthy young women. Subject Selection. Sexually active women between 18 and 38 years old with histories of regular menstrual cycles were eligible for inclusion into the study. Other inclusion criteria were good general health and Class 1 or I1 cervical cytologic smear. All participants agreed to refrain from other nonhormonal methods of contraception during the study. Exclusion criteria included Class 111-V cervical smear and any acute or chronic condition that would contraindicate the use of OCs, such as cervical dysplasia, undiagnosed abnormal vaginal bleeding, thrombophlebitis or thromboembolic disorders, pulmonary embolism or cerebral apoplexy, cardiovascular disease, liver impairment, known or suspected cancer, neurovascular eye lesions or serious visual disturbance. Study Protocol. Subjects were randomly assigned to receive NGM/EE or NG/EE for 24 treatment cycles. Each regimen involved a course of 21 days per cycle on-pill and seven days per cycle off-pill. On admission, all subjects underwent a thorough history, and physical and gynecologic examination including cervical cytologic testing. Laboratory analyses, including hemoglobin and hematocrit, liver and thyroid function tests, lipid profile, and fasting glucose were performed prior to entry and again after 3,6, 12, and 24 cycles of therapy. Glucose tolerance tests were to be performed at cycles 3, 12, and 24 in 20% of subjects in the larger of the two clinical studies. Participants were required to complete a monthly diary card, documenting pill intake, bleeding patterns, and side effects. Upon entry into the study, subjects provided information including age, ethnic group, incidence of smoking, menstrual history, number of previous pregnancies, and prior OC use. The latter was categorized according to the time interval between the last OC use (or pregnancy) and the first study tablet Acta Obstet Gynecol Scand Suppl152

as (1) postpartum, (2) direct switch, (3) indirect switch, and (4) fresh subjects. Efficacy Assessments. Prevention of pregnancy was the primary efficacy objective, with menstrual cycle control as a secondary parameter. Strict and conservative criteria for determining pregnancy were followed, and all suspected or questionable pregnancies, as well as those determined by serum assay only, were carefully analyzed. Based upon the detection sensitivity of pregnancy testing methods employed, even early cycle pregnancies ending in spontaneous abortion were identified. The efficacy of the two study formulations was statistically compared using the log rank test and the Gehan generalized Wilcoxon test. Pregnancy rates were determined by standard Life Table analysis as well as by computation of the Pearl Index. Menstrual cycle control was assessed in terms of (1) the number of days of early withdrawal flow, defined as any bleeding (requiring tampons or pads) or spotting (not requiring protection) on-therapy beginning during the pill-taking interval and continuing without interruption into the pill-free interval; (2) the number of days of withdrawal flow that is continuous with pill-free bleeding and continues after the next cycle of pill taking has begun; (3) the number of days of breakthrough bleeding or spotting-i.e., bleeding occurring during the pill-taking interval that is not categorized as early withdrawal flow or withdrawal flow continuing from the previous pill cycle; and (4) duration of menses. The incidence of failed-withdrawal cycles in which there was no bleeding during the pill-free interval and of amenorrhea, defined as two consecutive cycles without bleeding or spotting, were also evaluated. Study Cutoff: This study was terminated on December 31, 1986, regardless of the duration of individual patient exposure, to facilitate submission of study results to regulatory agencies for approval to market the NGM/EE formulation.

RESULTS In terms of baseline characteristics, subjects in the NGM/EE and NG/EE groups were comparable in age, ethnic distribution (Table I), weight, and incidence of smoking. There were no major differences in amount of pretherapy menstrual flow or in breakthrough bleeding, which in this population had Table I. Population characteristics

Parameter Mean age (yrs) Ethnic Group White Black Hispanic Oriental Other

All subjects

Fresh

NGMIEE NG/EE

NGM/EE

NG/EE

24.6

24.3

24.5

24.0

75.6 10.0 13.8 0.3 0.3

69.7 15.6 13.9 0.4 0.4

75.8 9.7 13.1 0.8 0.4

(YO): 73.1 12.1 13.7 0.7

0.4

U. S. clinical trials Table 11. Breakthrough bleeding or spotting: all cycles Cycle

NGM/EE

(%I

NG/EE (%)

1- 6 7-12 13-24

11.3 6.6 5.1

10.6 7.3 6.3

Table Ill. Duration of m e n s e s in days NGM/EE Cvcle ~

1 1- 6 7-12 13-24

NG/EE No.

Davs ~

~

~

5.54 5.20 4.93 4.90

70 1 3409 2413 3110

No.

Davs ~

~

4.87 4.63 4.39 4.23

705 3619 2564 3157

occurred in 21.4%-24.3% of women a t some time prior to treatment. The number of previous pregnancies and the incidence of prior O C use were similar in both groups. Continuation rates in these two groups were similar. Of the 736 women who used NGM/EE in the two clinical trials combined, 382 (51.9%) completed 12 cycles and 198 (26.9%) completed 24 cycles of treatment. Of the 737 women who used NG/EE, 397 (53.9%) completed 12 cycles and 192 (26.1070) completed 24 cycles. All subjects were discontinued on the study cutoff date as noted above. This abrupt termination limited the potential length of exposure for some patients.

Contraception Efficacy. There were no statistically significant differences between the two drug groups with respect to pregnancy rates. Statistical significance between regimens was assessed on the basis of useefficacy rates. In the two studies combined, the theoretical Pearl indices (pregnancies among women who complied with the pill-taking regimen) were 0.64 for NGM/EE versus 0.62 for NG/EE. Failures due to poor patient compliance resulted in Pearl Indices of 0.39 and 0.24 for NGM/EE and NG/EE, respectively.

21

Overall, the combined pregnancy rates included eight pregnancies in 9,351 cycles of NGM/EE (Life Table rate: 1.02 pregnancied100 woman-years of use; Pearl Index: 1.03) versus seven pregnancies over 9,727 cycles of NG/EE (Life Table rate: 0.70; Pearl Index: 0.86).

Cycle Control. In general, the incidence of bleeding or spotting is high at the beginning of a pill cycle because subjects may initiate tablets during the last few days of withdrawal bleeding. The incidence of reported bleeding decreases by tablets 3-8, when it gradually increases until the end of the pill-taking interval. Bleeding or spotting is typically higher during the first few cycles on OCs, and tends to stabilize with continued use. The daily incidence of bleeding was similar with the two treatment regimens. Early withdrawal flow was reported in 10.9% of cycles 1-6 with NGM/EE and 5 . 5 % with NG/EE. Breakthrough bleeding or spotting occurred in 11.3% of cycles 1-6 with NGM/EE versus 10.6% with NWEE. Low levels of early withdrawal flow were reported with both regimens, in 7.1% of cycles 7-12 with NGM/EE, versus 3.2% of NG/EE, decreasing by cycles 13-24 to 5.4% and 2.3% for NGM/EE and NG/EE patients, respectively. Both groups had comparable levels of breakthrough bleeding or spotting (Table 11), reported in 6.6% of cycles 7-12 with NGM/EE versus 7.3% with NG/EE. Here, too, the incidence of breakthrough bleeding or spotting decreased with time, and by cycles 13-24 was noted in 5.1% and 6.3% of NGM/EE and NG/EE groups respectively. The duration of menstrual bleeding on therapy with NGM/EE remained almost unchanged during therapy (Table 111), compared with the natural cycle. In NGM/EE subjects, menses lasted 4.9-5.5 days, compared with 4.2-4.9 days in the NG/EE group. The incidence of single withdrawal episodes missed was 1.1% and 0.7% of cycles 1-6 and 7-12, respectively, in women using NGM/EE compared with 1.8% and 1.5% for NG/EE subjects (Table IV). During cycles 13-24, only 0.4% of NGM/EE cycles had missed

Table IV. Amenorrheic cycles NGM/EE

NG/EE

Cycle

%

No. valid cycles

No. amenorrheic valid cycles

%

No. valid cycles

No. arnenorrheic valid cycles

1 1- 6 7-12 13-24

0.7 1.1 0.7 0.4

701 3409 241 3 31 10

5 37 16 13

1.7 1.8 1.5 1.4

705 361 9 2564 31 57

12 64 39 43

Acta Obstet Gynecol Scand Suppl152

28

S. L. Corson

withdrawal flow versus 1.4% of NG/EE cycles. Amenorrhea was defined statistically as the absence of bleeding or spotting for two consecutive cycles. No subject on NGM/EE developed amenorrhea by this definition as compared with 0.8% of NG/EE patients. In the 90-120 day posttreatment evaluation period absence of menstruation was reported in 1.4% of NGM/EE and 1.9% of NG/EE subjects.

Table VI. Percent change from baseline in mean HDL: all subjects Cycle

3 6 12 24

NGMiEE

NGiEE

7.1 " 7.4" 8.6" 9.1"

-7.7" -6.4" -7.4" -5.8"

* Two-sided t-test significant at the 5 % level. Data reported in part by Chapdelaine et al (111

Side'Effects and Safety. The adverse events reported by 5% or more of subjects were as follows. NGM/EE group: headache (29.5%), upper respiratory infection (15.1%), flu (7.3%), and mastodynia (6.4%); NG/EE group: headache (29.6%), upper respiratory infection (17.0%), and flu (5.2%). While the incidence of headache reported in both groups may seem high, these figures represent subjects who had ever reported headache during the two-year studies. Additionally, although mastodynia developed in 6.4% of NGM/EE subjects versus 3.3% of the NG/EE group, proportionately fewer episodes were judged severe by subjects taking NGM/EE. Discontinuation because of weight gain occurred in 1.0% of NGM/EE subjects (7 of 736) and in 1.4% of NG/EE subjects (10 of 737). Nausea and vomiting occurred in 24.4% of NGM/EE subjects at cycle 1. These effects dramatically decreased during the study, and reported for only 2.4% of all cycles 13-24. In subjects taking Lo/Ovral, 20.4% reported nausea and vomiting at cycle 1, but noted this complaint at only 3.4% of cycles 13-24. No unusual or clinically significant adverse findings were noted in hematology, liver or thyroid function in either group. While blood pressure fluctuated throughout the study, all mean changes from baseline were small and clinically insignificant in both sets of subjects.

ments of total serum cholesterol, LDL, HDL, triglycerides and the LDL/HDL ratio at baseline, and at cycles 3, 6, 12, and 24. LDL and HDL levels for the larger of the two studies combined here have been reported by Chapdelaine et a1 (11). The parameter considered most sensitive to atherosclerotic risk, the LDL/HDL ratio was markedly superior in NGM/EE as compared with NG/EE subjects ( ~ ~ 0 . 0 (Table 5) V). The percent change in the mean LDL/HDL ratio from pretherapy baseline among NGM/EE subjects ranged from a 4.1% decrease at cycles 3 and 6, to a 7.7% decrease at cycle 24. All of these reductions in LDL/HDL were statistically significant at all intervals measured. The percent change in the mean LDLIHDL ratio for NG/EE subjects, on the other hand, varied from a 14.2% increase at cycle 6 to an 18.5% increase at cycle 24. All increases were statistically significant. Most of these changes were attributed to alterations in HDL. The NGM/EE formulation consistently led to statistically significant increases in HDL, compared with statistically significant decreases in HDL in the NG/EE group (Table VI). Elevations in LDL were small and statistically significant only at cycles 3 and 6 for NGM/EE subjects, but were greater and statistically significant at all cycles measured in the NG/EE group (Table VII).

Effects on Lipid Metabolism. The effects of both OCs

Effects on Glucose Metabolism. The effects of both

on lipoprotein metabolism were compared by assess-

OC formulations on carbohydrate metabolism were compared by assessments on fasting blood glucose and 3-hour glucose tolerance tests at baseline and at cycles 3, 12 and 24. Fasting glucose was also measured at cycle 6. In NGM/EE subjects, mean changes in fasting glucose levels were small and clinically insignificant. The percent change in mean levels in NGM/EE subjects was 0.1 at cycle 12 and 2.3 at cycle 24, compared with an increase of 1.0 at cycle 12 and a statistically significant increase of 4.6 at cycle 24 in NG/EE subjects. The 3-hour glucose tolerance test was performed in only one of the two clinical trials reported here. For

Table V. Percent change from baseline in mean LDL/HDL ratio: all subjects Cycle

NGMiEE

NG/EE

3 6 12 24

-4.1 * -4.1 * -6.4' -7.7"

15.1" 14.2" 17.2" 18.5"

* Two-sided t-test significant at the 5 % level. Data reported in part by Chapdelaine et al (11). Aria Obsiei Gynecol S c a d Suppl IS2

U S. clinical trials

29

both treatment groups, mean values were very similar to baseline at cycles 3 and 12 and slightly elevated at cycle 24. These changes were minimal and clinically neutral.

DISCUSSION Results of these two clinical trials indicate that the combination OC formulation consisting of norgestimate at a dose of 250 pg with ethinyl estradiol 35 pg is comparable in efficacy to a marketed reference drug, Lo/Ovral. There were no statistically significant differences in pregnancy rates between the two agents, as assessed by log rank test and Gehan generalized Wilcoxori test, or in associated side effects or cycle control. These findings confirm the following results of earlier phase I1 studies on suppression of ovulation and alterations in cervical mucus (11,12).

Endocrine Profile. A Phase I1 randomized clinical trial was undertaken to compare the effects of NGM/EE and NG/EE on the endocrine system (11). Forty healthy ovulating women 18 to 40 years of age were randomly assigned to two treatment groups of 20 subjects each. Participants were analyzed for six cycles: one pretherapy control cycle, four on-therapy cycles, and one posttherapy cycle. Treatment consisted of four cycles of NGM/EE or NG/EE, both taken on a 28-day regimen of 21 days per cycle on medication and 7 days per cycle off medication. Serum hormone determinations were obtained during three cycles: the pretherapy control cycle, cycle 4 on-therapy, and posttherapy. Determinations included: estrone, estradiol, progesterone, follicle stimulating hormone (FSH), luteinizing hormone (LH), prolactin, growth hormone, thyroid stimulating hormone, androstanedione, dehydroepiandrosterone sulfate, cortisol, and SHBG. All subjects completed six cycles of study, except for one Lo/Ovral patient who discontinued treatment after one on-therapy cycle due to reported transient visual loss and migraine headache. Table VII. Percent change from baseline in mean LDL: all subjects NGMiEE

NGiEE

3

3.5*

6

3.4" 0.8 0.5

6.7" 6.0" 8.1 * 10.3"

Cycle

12 24

* Two-sided t-test significant at the 5 % level. Data reported in part by Chapdelaine e t al 111)

Pre-therapy

Cycle 4

Post-therapy

Fig. 1. Mean progesterone levels in NGM/EE subjects.

NGM/EE caused a statistically significant decrease in the average level of L H from pretherapy to cycle 4, with a rapid posttherapy return to higher levels. There were no statistically significant differences between regimens, and both were highly effective in reducing the levels of LH. Mean progesterone levels-a highly sensitive marker of ovulation-were uniformly low in both regimens, and were extremely reduced at cycle days 17/18 and 21/22 compared with pretherapy values (Fig. 1). There were no statistically significant between-drug differences in the change from pretherapy to cycle 4. Mean SHBG levels increased from pretherapy to cycle 4 in both regimens, but in the NGM/EE group, this increase was large and statistically significant, whereas in the NG/EE group it was small and not statistically significant (Fig. 2). When SHBG values were compared between the groups, a highly statistically significant difference (p=0.0007) was detected in the mean SHBG change from pretherapy to cycle 4. NGM/EE showed a greater SHBG increase than NG/EE. Postlreatment SHBG levels were significantly elevated compared with pretherapy values in each group. These findings clearly reflect the clinical difference in overall androgenicity between the two formulations, and provide further support for preclinical assay findings.

Cervical Mucus Scores. Although the primary mechanism of OC action is inhibition of ovulation, alterations in cervical mucus may provide a hostile environActa Obstet Gynecol Scand Suppi152

30

S. L. Corson

160 140

120

100 SHBG

(nM)

80

60 40

20 0 Pre-therapy Ortho-Cyclen

Cycle 4

LolOvral 0

ment for sperm migration and penetration. A Phase I1 study (12) was conducted in five healthy, ovulatory women to assess the effects of NGM/EE on cervical mucus and resultant ovulation. Following two pretherapy cycles, serial cervical mucus evaluations were performed during one cycle on study medication. Ovulatory status was documented by basal body temperature and luteal phase progesterone levels before and during therapy. Cervical mucus was evaluated for ferning, spinnbarkeit, viscosity, quantity, and cell count, with each parameter scored on a scale of zero to three (Moghissi system) (13). This score is directly proportional to the ovulatory quality of the mucus, with 15 an optimal quality rating, and less than 10, unfavorable. NGM/EE caused an inhibition of ferning and spinnbarkeit, an increase in viscosity and cellularity, and a decrease in quantity, indicative of cervical mucus deterioration. In all five subjects, cervical mucus scores decreased from baseline to unfavorable or “hostile” levels during treatment. Basal body temperatures and serum progesterone values further demonstrated ovulation suppression in women taking NGM/EE.

CON CLUS 10N Whereas both the NGM/EE and NG/EE formulations were well-tolerated in all studies reported here, lack of androgenicity with the norgestimate formulation appeared to offer the potentially important Acta Obstet Gynecol Scand Suppl152

Post-therapy

Fig. 2. Sex hormone binding globulin (SHBG) levels. Data adapted from Chapdelaine et al (11).

therapeutic advantage of a favorable lipid profile. The LDL/HDL ratio shifted in a salutary direction in NGM/EE subjects, and adversely in the NG/EE group. The beneficial change can be attributed largely to significant increases in the cardioprotective factor HDL, seen in NGM/EE subjects. The clinical implications of Ortho-Cyclen’s metabolically positive effects on HDL cholesterol are difficult to ascertain. There are as yet no objective data linking a specific OC with the development of atherosclerotic cardiovascular disease. Such documentation will require additional years of OC use before a meaningful epidemiologic study can be undertaken, and results may be confounded by changes in formulations over that time. Nevertheless, since atherosclerosis is known to begin in childhood (14), and since even small changes in lipoproteins can affect CV disease risk (5,15), it seems reasonable when considering an OC for long-term use in healthy young women to select one that is not associated with adverse lipid effects.

REFERENCES 1. Barr DP, Russ EM, Eder HA. Protein-lipid relationships

in human plasma. 11. In atherosclerosis and related conditions. Am J Med 1951; 11:480-493. 2. Miller GJ, Miller NE. Plasma high-density-lipoprotein concentration and development of ischaemic heart disease. Lancet 1975; 1:16-19. 3. Gordon T, Castelli WP, Hjortland MC, et al. High density lipoprotein as a protective factor against coronary heart disease: The Frarningham Study. Am J Med 1977; 62:707-714.

!,l S. clinical trials 4. Miller NE, Thelle DS, Forde OH, Mjos OD. The Tromso Heart Study. High-density lipoprotein and coronary heart-disease: a prospective case-control study. Lancet 1977; 1:965-968. 5. Castelli WP, Doyle JT, Gordon T, et al. HDL cholesterol and other lipids in coronary heart disease: The cooperative lipoprotein phenotyping study. Circulation 1977; 55:767-772. 6. LaRosa JC. The varying effects of progestins on lipid levels and cardiovascular disease. Am J Obstet Gynecol 1988; 158:1621--1629. 7. Wahl PW, Walden CE, Knopp RH, et al. Effect of estrogen/progestin potency on lipid/lipoprotein cholesterol. N Engl J Med 1983; 303:862-867. 8. Bradley DD, Wingerd J, Petitti DB, et al. Serum highdensity-lipoprotein cholesterol in women using oral contraceptives, estrogens and progestins. N Engl J Med 1978; 299117-20. 9. Knopp RH, Walden CE, Wahl PW, et al. Effects of oral contraceptives on lipoprotein triglyceride and cholesterol: Relationship to estrogen and progestin potency. Am J Obstet Gynecol 1982; 142725-731. 10. Phillips A. The selectivity of a new progestin. Acta Obstet Gynecol Scand (this supplement). 11. Chapdelaine A, Desmarais J-L, Derman RJ. Clinical

12.

13. 14.

15.

31

evidence of the minimal androgenic activity of norgestimate. Int J Fertil 1989; 34:347-352. Hull ME, Moghissi KS: Effects of norgestimate (0.250 rng) in combination with ethinyl estradiol(O.035 mg) on cervical mucus. Advances in Contraception 1986; 2:71. Moghissi KS. The cervix in infertility. Clin Obstet Gynecol 1979; a 22:27-41. Newman WP, Freedman DS, Voors AW, et al. Relation of serum lipoprotein levels and systolic blood pressure to early atherosclerosis. N Engl J Med 1986; 314:138144. Lipid Research Clinics Program. The Lipid Research Clinics Coronary Primary Prevention Trial Results: 11. The relationship of reduction in incidence of coronary heart disease to cholesterol lowering. JAMA 1984; 251: 365-374.

Address for correspondence: Stephen L. Corson, M.D. Director Philadelphia Fertility Institute 815 Locust Street Philadelphia, PA 19107 USA

Acta Obsrer Gynecol Scand Suppl15.2