The Function of the Corpus Luteum of Pregnancy in Ovulatory Dysfunction and Luteal Phase Deficiency*

FERTILITY AND STERILITY Copyright © 1981 The American Fertility Society

Vol. 36, No.1, July 1981 Printed in U.s A.

THE FUNCTION OF THE CORPUS LUTEUM OF PREGNANCY IN OVULATORY DYSFUNCTION AND LUTEAL PHASE DEFICIENCY*

MICHAEL R. SOULES, M.D.t CLAUDE L. HUGHES, JR., B.S. SEZER AKSEL, M.D.* LEE TYREY, PH.D. CHARLES B. HAMMOND, M.D.§

Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, Duke University Medical Center, Durham, North Carolina 27710, and Department of Obstetrics and Gynecology, University of South Alabama School of Medicine, Mobile, Alabama 33617

Relatively little knowledge exists of corpus luteum function in early pregnancy after the successful treatment of ovulatory dysfunction or luteal phase deficiency. To assess the activity of the corpus luteum of such patients, human chorionic gonadotropin (hCG) and 17-hydroxyprogesterone (17-0H-P) levels were determined in serum samples obtained from normal women (44 patients), women with ovulatory dysfunction (10 patients), and women with luteal phase deficiency (7 patients); all determinations were made during conceptive cycles, and sampling continued into the first trimester of pregnancy. There were no statistically significant abnormalities of hCG levels when infertility patients were compared with control patients. According to the premise that 17cOH -P levels reflect corpus luteal function, there appeared to be adequate function in pregnancies after progesterone treatment of luteal phase deficiency. In pregnancies following ovulation induction with clomiphene, the corpus luteum function, on the basis of 17-0H-P levels, was significantly increased in magnitude and duration. These results have clinical implications with regard to supplemental hormone therapy in early pregnancy. Fertil Steril36:31, 1981

Ovulatory dysfunction (either anovulation or oligo-ovulation) and luteal phase deficiency (LPD) are relatively common causes of infertility in female patients. Clomiphene citrate (Clomid; Merrell-National Laboratories, Cincinnati, Ohio)

is frequently used in the treatment of ovulatory dysfunction and occasionally in the treatment of LPD. Progesterone (P) supplementation is a common mode of therapy in the latter case. 1 A pregnancy rate of approximately 40% has been reported with clomiphene citrate treatment in patients with ovulatory difficulties,2 and a pregnancy rate of 50% has been reported with P therapy in patients with LPD. 3 Although substantial clinical research efforts have been expended in achieving treatments effective in establishing pregnancy in such patients, considerably less investigation has been performed on luteal function in similar infertility patients who have achieved pregnancy. This relative lack of interest in corpus luteum function in these patients exists despite the known increased

Received October 21, 1980; revised and accepted February 10, 1981. *Presented at the Thirty-Seventh Annual Meeting of The American Fertility Society, March 14 to 18, 1981, Atlanta, Georgia. tPresent address: Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, University of Washington, Seattle, Washington 98195. *Department of Obstetrics and Gynecology, University of South Alabama School of Medicine. §Reprint requests: Charles B. Hammond, M.D., Box 3143, Duke University Medical Center, Durham, North Carolina 27710.

31

32

SOULES ET AL.

incidence of spontaneous abortion in clomipheneinduced pregnancies. 2 Frequently the clinical approach during early pregnancy in these patients has consisted of conservative follow-up or the empirical use of hormones (human chorionic gonadotropin [hCG] or P) designed to stimulate or supplement the P production of the corpus luteum. The useofprogestins (both natural and synthetic) in early pregnancy is currently controversial in light of the recent studies possibly implicating them as teratogenic agents. 4 In consideration of these issues, the current study was undertaken to investigate ovarian function in the first trimester of pregnancy in infertility patients who became pregnant after treatment for either ovulatory dysfunction or LPD.

MATERIALS AND METHODS

Study Population (Table 1). The normal group (group A) consisted of 44 women of reproductive age who conceived after spontaneous ovulation and subsequently completed full-term pregnancies. Only patients with uncomplicated singleton pregnancies were included in the normal group. Group B comprised 10 women with ovulatory dysfunction who conceived and completed normal full-term singleton pregnancies after ovulation induction with clomiphene citrate. In this group with oligo-ovulation and anovulation, no major additional infertility factors were found and there were no identifiable, specific etiologic factors responsible for their ovulatory dysfunction prior to initiation of ovulation induction. * Clomiphene citrate was administered in doses varying from 250 mg to 500 mg (total) over a 5-day course. Group C consisted of seven women of reproductive age who conceived and completed normal fullterm singleton pregnancies after substitutional progestational therapy of previously diagnosed LPD. These women initially presented as infertility patients and, after evaluation, were found to have inadequate maturation of endometrial elements on two timed endometrial biopsies in two consecutive menstrual cycles. The methods of evaluation and therapy of LPD as applied here *The endocrine evaluations of these patients generally included thyroid function tests, determinations of serum andrOgen levels and serum prolactin levels, radiographs of the sella turcica, and progesterone challenge tests. The infertility evaluation consisted of timed endometrial biopsy, semen analysis, postcoital test, and hysterosalpingography.

July 1981 have been described previously.3 Briefly stated, these WOJIlen conceived after initiating the use of P vaginal suppositories, 25 mg twice daily during the luteal phase of their menstrual cycles. After pregnancy was confirmed, these patients received weekly injections of 250 mg of 17 -hydroxyprogesterone (17-0H-P) caproate (Delalutin; E. R. Squibb and Company, Princeton, N. J.) until 12 weeks' gestation. By chance, no patients received their first injections until 8 weeks of gestation, and received only progesterone vaginal suppositories until that time. Therefore, the 17-0H-P levels for group C in weeks 5 to 7 should' reflect only endogenous hormone. It is unknown to what extent, if any, the injections of 17-0H-P caproate in weeks 8 to 12 contributed to the endogenous levels of 17-0H-P. Group D consisted of three patients who were infertile· secondary to LPD as diagnosed in the same manner as in group C. However, these patients were treated with clomiphene citrate in doses of 250mg to 500 mg (total) for a 5-day course given in routine fashion. They received no progesterone therapy before or after pregnancy was diagnosed. Endometrial biopsies were used to document the adequacy of therapy as was done in group C. Individual physician discretion, not patient differences, resulted in the treatment of these patients with clomiphene citrate instead of progesterone suppositories. These women in group D completed normal full-term singleton pregnancies. For groups A through D, the normalcy of the pregnancies and the gestational age were determined by thorough review of the charts during the postpartum interval. Throughout the study, gestational age was measured in weeks from the patient's last menstrual period. On occasion, the gestational age was appropriately modified by review of the clinical course of pregnancy (e.g., McDonald measurements, ultrasound) and estimated gestational age of the infant afterdelivery. During the first 13 weeks of pregnancy, one or more venous blood samples were obtained from each patient for analysis of hCG and 17-0H-P. The samples were allowed to clot and were centrifuged, and the serum was separated and stored at - 20° C until assayed. The samples were assayed for hCG over a 6-month period on random routine clinical assays; the samples were analyzed for 17OH-P in three separate large assays. The hormonal data from groups A through D were placed for comparative purposes in three divisions according to gestational age-weeks 5 through 7, 8 through 10, and 11 through 13 (Table 1).

Vol. 36, No.1

FUNCTION OF CORPUS LUTEUM OF PREGNANCY

33

TABLE 1. Study Population WIt 5-7 Group and description

A: Normal patients (control) B: Ovulatory dysfunction, clomiphene treatment C: Luteal phase deficiency, progesterone treatment D: Luteal phase deficiency, clomiphene treatment

WkS-lO

heG

17-0H-P

Wk 11-13

heG

heG

17-0H-P

Patients

Samples

Patients

Samples

Patients

Samples

Patients

21

38

21

38

23

50

23

9

16

9

13

6

15

6

7

12

7

10

3

6

3

3

5

3

3

Samples

17-0H-P Samples

Patients

Samples

Patients

50

19

24

19

24

15

6

12

6

12

2

3

2

6a

3a

aPatients receiving 17-0H-P caproate, 250 mg weekly, when samples were obtained.

Hormone Assays. Serum samples were assayed for hCG in duplicate using a modified double-antibody l3-subunit radioimmunoassay method. 5 Purified hCG (whole molecule, provided by Robert Canfield, Ph.D., Columbia University, New York, N. Y.) was radioiodinated with 1251 and used as the assay tracer. The standard employed was the 2nd IS-hCG (World Health Organization); results are expressed in international units per milliliter. Rabbit antiserum (SB6) generated against the l3-subunit of hCG was generously provided by Dr. J. L. Vaitukaitis and has been described elsewhere. 6 The levels of 17-0H-P were analyzed in duplicate serum samples by a radioimmunoassay modification of the assay procedure described by Strott and Lipsett7 and Stripp et al. 8 as performed by the Medical Endocrinology Laboratory at Duke University Medical Center. The intra-assay coefficients of variation were 4% and <10%; the interassay coefficiehts of variation were 6% and <10% for hCG and 17-0H-P assays, respecti vely. Statistics. One-way analysis of variance with partitioning of the sum of squares was used to test for (1) significant differences among group means, (2) linear regression, and (3) deviation from regression among groups (comparison of slopes).

simultaneously with the marked decline in 17OH-P levels (Fig. 1). When mean hCG levels in group B (Clomid), group C (LPD-P), and group D (LPD-Clomid) in weeks 5 to 7 of gestation were compared with those of group A, there were no significant differences (Fig. 2). In weeks 5 to 7 for groups A, B, and C, a significant linear increase in hCG over time was demonstrated. The change in hCG over time (comparison of slopes) was not significantly different among these groups. Although it appeared that the mean hCG levels in clomiphene-treated patients (groups B and D) were lower than those in the control patients in weeks 5 to 7, the difference was not significant (P = 0.05). In weeks 8 to 10, the mean hCG level in group B was significantly lower than that of the control group, although there were no differences among groups at weeks 11 to 13 (Fig. 2). 4000

80 70

e 3000 ;

60

.!:

~

.... z 0

.... '"....V> ....

50

2000 ~

~ 40

11.

>-

>< 0

30 20

In normal pregnant patients (group A), the 17OH-P values exhibited maximal levels at 5 to 6 weeks of gestation with a gradual decline over the remainder of the first trimester of pregnancy (Fig. 1). The hCG levels in group A plateaued at 8 to 9 weeks of gestation and declined thereafter (Fig. 1). The hCG levels demonstrated a sharp increase

'">'"~ 0

RESULTS 1000

10 4

6

7

8 WEEKS

9

10

11

12

FIG. 1. The mean hCGand 17-0H-P levels (± 1 SEM)for44 normal women in the first trimester of spontaneous singleton pregnancies. - - . , hCG; .i-.i, 17-0H-P.

SOULES ET AL.

34

10

FIG. 2. The comparative mean heG levels for groups A through D at three time intervals in the first trimester of pregnancy. The actual mean is listed at the top of each bar graph with 1 SEM indicated above. **, A significant difference from group A at P .;; 0.025.

In comparing the mean 17-0H-P levels in normal pregnant patients (group A) at weeks 8 to 10 and 11 to 13 with those at weeks 5 to 7, there was a significant (P < 0.005) decrease at both time intervals (Fig. 3). For the clomiphene-treated patients in group B, when the 8- to 10- and 11- to 13-week intervals were compared with the 5- to 7-week interval, a significant decrease (P < 0.01) in the mean 17-0H-P levels was delayed until 11 to 13 weeks (Fig. 3). Figure 3 also illustrates that the mean 17-OH-P levels in the clomiphene-treated patients in group B were clearly and significantly elevated over those of the control group throughout the first trimester of pregnancy. The 17-0H-P values were significantly elevated in the clomiphene-treated patients of group D as well, whose 17-0H-P levels were analyzed only for the 5- to 7-week interval. The 17-OH-P levels in women with LPD treated with P supplementation (group C) were not different from those in the normal group during any of the periods sampled (Fig. 3). I~ is unknown to what degree, if any, therapy with 17-0H-P caproate in weeks 8 to 12 contributed to the serum levels of 17-0H-P (group Conly).

July 1981 duces progesterone but negligible quantities of 17-0H-P, as it lacks 17-hydroxylase enzymatic activity.12-14 Therefore, the determination of peripheral serum levels of 17-0H-P serves as a useful marker of corpus luteum activity during pregnancy. 17-0H-P has no known significant activity during human pregnancy. (Note: The synthetic esterification of the 17a-hydroxyl group of 17-0HP [e.g., Delalutinl results in a hormone with progestational activity.) A strong positive correlation exists between the corpus luteal secretion of P and 17-0H-P, during both the luteal phase of the menstrual cycle15 and early pregnancy.16, 17 In pregnancy, there is a transient decline in P levels at 5 to 9 weeks which parallels a decrease in 17-0H-P levels. A subsequent increase in P levels at 8 to 10 weeks reflects placental activity.16, 18 Therefore, the corpus luteum appears to be the primary source of progesterone during pregnancy until 5 to 8 weeks of gestation, at which time the placenta assumes that role. hCG is thought to be a principal determinant of corpus luteum function during pregnancy. Exogenous hCG (in large doses) can prolong corpus luteum activity when administered during the luteal phase of the menstrual cycle,19 in the first trimester of· pregnancy,20 .and postpartum. 21 In early human pregnancy the corpus luteum has luteinizing hormone-hCG membrane receptors. 22 Nevertheless, Figure 1 demonstrates declining corpus luteum function (17-0H-P levels) despite increasing levels of hCG. This apparent paradox

DISCUSSION

As determined by studies in women undergoing first-trimester therapeutic abortions, an active corpus luteum is necessary for pregnancy maintenance until approximately 6 to 7 weeks gestation. 9 The corpus luteum of pregnancy can synthesize both progesterone and 17-0H_P.1O,11 In contrast to the corpus luteum, the placenta pro-

FIG. 3. The comparative mean 17-0H-P levels for groups A through D at three time intervals in the first trimester of pregnancy. The actual mean is listed at the top of each bar graph with 1 SEM indicated above. ****, A significant difference from group A at P .;; 0.005.

Vol. 36, No.1

FUNCTION OF CORPUS LUTEUM OF PREGNANCY

has been previously demonstrated and probably reflects either a finite life-span, down-regulation ofluteinizing hormone-hCG,23 or other unknown control mechanisms for the corpus luteum of pregnancy. In this study there was a great deal of individual variation in hCG levels among patients in all the groups throughout the first trimester of pregnancy. Nevertheless, in comparing hCG levels in the first trimester of pregnancy in women who conceived after treatment of ovulatory dysfunction or LPD with those of normal pregnant women, no substantial differences were observed (Fig. 2). In viable singleton pregnancies, there is no reason to expect a difference between spontaneous and induced conceptions in the ability of the trophoblast to synthesize hCG. The decrease in the mean hCG level at 8 to 10 weeks' gestation in group B with ovulatory dysfunction is of questionable clinical significance as these same patients simultaneously had a significantly increased mean 17 -OH-P level, in agreement with an apparent lack of effectiveness of hCG on 17-0H-P levels at this stage of gestation, as shown in Figure 1. There appeared to be a trend in the clomiphene-induced pregnancies, as shown in Figures 2 and 3, toward lower hCG levels in association with increased 17-0H-P levels as compared with those of normal patients. Although hCG is thought to be a principal determinant of corpus luteum function, in these instances there does not appear to be a strong relationship. Therefore, since patients who conceive after therapy (clomiphene or progesterone) for ovulatory dysfunction and LPD have normal levels of hCG which only loosely correlate with luteal function, there appears to be little rationale for exogenous hCG supplementation of these pregnancies. Corpus luteum function is increased in magnitude and duration in pregnancies conceived after clomiphene citrate induction of ovulation on the basis of the premise that 17-0H-P levels during early pregnancy reflect the corpus luteum function. This situation suggests that the functional capabilities of the corpus luteum of pregnancy are determined prior to ovulation. Previous studies have demonstrated (1) increased luteal phase P levels after clomiphene therapy24 and (2) increased P and 17-0H-P levels in early pregnancy after ovulation induction with exogenous gonadotropins. 18 In this study, the ovary responded to clomiphene in the same manner in patients with either ovulatory dysfunction or LPD. Since P levels are highly correlated with 17-0H-P levels during early pregnancy, there does not appear to be

35

any necessity for P supplementation in pregnancies after clomiphene therapy. (This statement assumes that adequate luteal phase function has been documented as an integral part of clomiphene therapy.) The mean 17 -OH-P level for patients in group C (LPD) , who conceived after P therapy, appeared to be slightly decreased at 5 to 7 weeks of gestation; however, this was not statistically significant. It was difficult to assess corpus luteum function in these patients by 17-0H-P levels during weeks 8 to 13 since they were receiving supplemental therapy with 17-0H-P caproate at that time. From the data in Figure 3, it would appear that continued progestin supplementation of the corpus luteum from the luteal phase into pregnancy in patients with LPD may be unnecessary. It has been presumed, but not established, that a deficient corpus luteum during the menstrual cycle is invariably deficient in pregnancy. Endogenous hCG may be sufficient in some cases to rescue the corpus luteum. This situation needs further study by measuring P levels in addition to hCG and 17-0H-P in both treated and control groups. Therefore, we would recommend continued use of progestin supplementation in early pregnancy in these patients until more definitive data are available. Although our data indicate that corpus luteum function in early pregnancy in patients with LPD is increased in patients treated with clomiphene versus those treated with P supplementation, they do not necessarily imply that clomiphene is a superior mode of therapy. The published pregnancy rates are higher with progesterone treatment than with clomiphene citrate treatment,25 and an increased incidence of LPD has been found in clomiphene-induced cycles. 26 Acknowledgments. The authors gratefully acknowledge the assistance of Harold Lebovitz, M.D., Medical-Endocrinology Laboratories, Duke University Medical Center, in performing the 17 -OH -P assays, and the secretarial assistance of Mrs. Patricia Dennos.

REFERENCES 1. Jones GS: The luteal phase defect. Fertil Steril 27:351,

1976 2. Behrman SJ, Kistner RW [Editors]: Induction of ovulation with clomiphene citrate. In Progress in Infertility, Second Edition. Boston, Little, Brown and Co, 1975, p 518 3. Soules MR, Wiebe RH, Aksel S, Hammond CB: The diagnosis and therapy of luteal phase deficiency. Fertil Steril 28:1033, 1977

36

SOULES ET AL.

4. Chez RA: Proceedings of the symposium, "Progesterone, Progestins, and Fetal Development." Fertil Steril 30:16, 1978 5. Tyrey L, Hammond CB: The hCG-j3-subunit assay: potential error in hCG measurement related to choice of labeled antigen. Am J Obstet GynecoI125:160, 1976 6. Vaitukaitis JL, BraunsteinGD, Ross GT: A radioimmunoassay which specifically measures human chorionic gonadotropin in the presence of human luteinizing hormone. Am J Obstet Gynecol 113:751, 1972 7. Strott CA, Lipsett MB: Measurement of17 -hydroxyprogesterone in human plasma. J Clin Endocrinol Metab 28: 1426, 1968 8. Stripp B, Taylor AA, Bartter FC, Gilette JR, Loriaux DL, Easley R, Menard RH: Effect of spironolactone on sex hormones in man. J Clin Endocrinol Metab 41:777,1975 9. Csapo AI, Pulkkinen MO, Ruttner B, Sauvage JP, Wiest WG: The significance of the human corpus luteum in pregnancy maintenance. Am J Obstet GynecoI112:1061, 1972 10. Mikhail G, Allen WM: Ovarian function in human pregnancy. Am J Obstet Gynecol 99:308, 1967 11. Tulchinsky D, Simmer HH: Sources of plasma 17 -hydroxyprogesterone in human pregnancy. J Clin Endocrinol Metab 35:799, 1972 12. Jungmann RA, Schweppe JS: Biosynthesis and metabolism of neutral steroids by human midterm placenta and fetal liver. J Clin Endocrinol Metab 27:1151, 1967 13. Pion R, Jaffe R, Erikson G, Wiquist N, Diczfalusy E: Studies on the metaliolism of C-21 steroids in the human foeto-placental unit. Acta Endocrinol (Kbh) 48:234, 1965 14. Palmer R, Blair JA, Eriksson G, Diczfalusy E: Studies on the metabolism of C-21 steroids in the human foeto-placental unit. Acta Endocrinol (Kbh) 53:407, 1966 15. Ross GT: Pituitary and gonadal hormones in women during spontaneous and induced ovulatory cycles. Recent Prog Horm Res 26:1, 1970

July 1981 16. Mishell DR, Thorneycroft IH, Nagata Y, Takaaki M, Nakamura RM: Serum gonadotropin and steroid patterns in early human gestation. Am J Obstet Gynecol 117:884, 1973 17. Tulchinsky D, Hobol C: Plasma hCG, E 1 , E 2, E 3 , progesterone, and 17-0H-progesterone in human pregnancy. Am J Obstet Gynecol 117:884, 1973 18. Yoshimi F, Strott CA, Marshall JR, Lipsett MB: Corpus luteum function in early pregnancy. J Clin Endocrinol Metab 29:225, 1969 19. Stock RJ, Josimovich JB, Kosar B, Klopper A, Wilson GR: The effect of hCG and hPL on steroidogenesis in the corpus luteum. J Obstet Gynaecol Br Commonw 78:549, 1971 20. Garner PR, Armstrong DT: The effect of hCG and E2 on the maintenance of the human corpus luteum of early pregnancy. Am J Obstet Gynecol 128:469, 1977 21. LeMaire WJ, Conly PW, Moffett A, Spellacy WN, Cleveland WW, Savard K: Function of the corpus luteum during the puerperium: its maintenance by exogenous human chorionic gonadotropin. Am J Obstet Gynecol 110: 612, 1971 22. Rao CV, Sanfilippo J, Carman FR: Gonadotropin receptors in human corpus lutea of term pregnancies. Am J Obstet Gynecol 132:581, 1978 23. Midgeley AR, Sadler WA [Editors]: Hormone regulation of hormone receptors in ovarian follicular development. In Ovarian Follicular Development and Function. New York, Raven Press, 1979, p 225 24. Wu CH: Plasma hormones in clomiphene citrate therapy. Obstet Gynecol 49:443, 1977 25. Andrews WC: Luteal phase defects. Fertil Steril 32:501, 1979 26. Jones GS, Maffezzol RD, Strott CA, Kaplan G, Ross GT: Pathophysiology of reproductive failure after clomipheneinduced ovulation. Am J Obstet Gynecol108:847, 1970