A life-table analysis of pregnancy yield in fixed low-dose menotropin therapy for patients in whom clomiphene citrate failed to induce ovulation

Vol. 37 PP 639-644, May 1982 Printed in U.8A.

A life-table analysis of pregnancy yield in fixed low-dose menotropin therapy for patients in whom clomiphene citrate failed to induce ovulation

Alvin F. GQldfarb, M.D.* Sheldon Schlaff, M.D. Michael L. Mansi, D.O.t Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology and the Breskman Endocrine Laboratory, Pennsylvania Hospital, University of Pennsylvania Sclwol of Medicine, Philadelphia, Pennsylvania

Forty-nine patients in whom clomiphene citrate failed to induce ovulation were treated for 177 cycles with a fixed low dosage of menotropin. Among these 49 patients, there were 24 pregnancies. Among these pregnancies were two that were multiple and three spontaneous abortions. In only one treatment cycle was there a hyperstimulation syndrome. These patients were divided into three clinical groups: the secondary amenorrheic patient, the oligo-amenorrheic patient, and the patient with poor corpus luteum function. There was no statistically significant difference in the pregnancy rate per month among all groups during the first three treatment cycles (average value, 0.07). However, there was a statistically significant improvement in the pregnancy rate per month in the group with secondary amenorrhea and the group with poor corpus luteum in the last three treatment cycles, as compared with the first three treatment cycles (P = 0.05; average value, 0.75). The oligo-amenorrheic patients, on the other hand, during the last 3 months of treatment, had no statistically significant increase in the pregnancy rate per month. These data suggest that menotropin therapy may have a priming effect. These data do not fit the currently accepted model of a constant pregnancy rate per month for all patients. The data suggest that caution should be exercised before combining patient groups when evaluating the results of menotropin therapy. Fertil Steril 37:639, 1982

The evaluation of the effect of a disease process or medical regimen leading to conception in an infertile population can be made more useful by the use of life-table analysis. 1 Life-table analysis allows for the calculation of a monthly conception rate and takes into account new patients who Received June 15, 1981; revised and accepted January 26, 1982. *Reprint requests: Alvin F. Goldfarb, M.D., Gynecologic Endocrine Associates P.C., Suite 100, St. James House, 13th and Walnut Streets, Philadelphia, Pennsylvania 19107. tReproductive Endocrine Fellow, Department of Obstetrics and Gynecology, Pennsylvania Hospital.

48

enter the study as well as dropouts from the study. This rate should be independent in a homogeneous population at the time of entrance into the study. In addition, derivation of a cumulative pregnancy rate from the monthly conception rate permits the determination of the clinical time necessary for appropriate treatment. With .this information, the physician can properly counsel the couple concerning a particular therapeutic regimen. It is the purpose of this paper to (1) review the effects of a fixed low dosage of menotropin used for the management of Clomid (Merrell-National Laboratories, Cincinnati, OH) failure patients;

Table 1. Prevwus Reproductive History of Study Group Secondary Oligo·amen- Poor coramenorrhea orrhea pus luteum function

Nulligravida Previous pregnancy Spontaneous abortion Term delivery

3 7 2 9

4 6 5 4

14

15 11 9

(2) establish the conception rates for such prob-

lems; and (3) evaluate the therapeutic safety of such a program. MATERIALS AND METHODS

Forty-nine patients who had failed to achieve a pregnancy after five treatment cycles with up to 200 mg per day of Clomid or Clomid plus human chorionic gonadotropin (hCG) served as a population base for this study. These patients were put into three clinical groups: (1) patients with secondary amenorrhea, defined as the absence of menses for more than 150 days2; (2) patients with oligo-amenorrhea, defined as menses occurring 60 to 120 days apart2; (3) patients with poor corpus luteum function, based on a thermogenic shift of 0.8 0 F or greater on the basal body temperature (BBT) chart for 10 days or less or an endometrial biopsy specimen out of phase by more than 48 hours. 3 ,4 These patients were 21 to 36 years of age, with a mean age of 27.5 years. Each had tubal patency established by either hysterosalpingogram or laparoscopy with indigo carmine perfusion through the reproductive tract. Each of these individuals had normal basal follicle-stimulating hormone (FSH) , luteinizing hormone (LH), and prolactin (PRL) determinations. All patients were euthyroid chemically and clinically and had no evidence of androgen excess as determined by plasma dehydroepiandrosterone sulfate (DHEAS) levels and plasma testosterone (T) levels. There was also no history of severe psychologic stress, weight loss, strenuous physical activity, or jogging. The reproductive history of this group of patients is seen in Table 1. Each patient was treated through six treatment cycles. The protocol called for 6 months of low-dose menotropin therapy; however, a number of patients did not complete the total therapeutic regimen (pregnancies and dropouts). They were given two ampules of Pergonal (Serono Laboratories, Braintree, MA) (75 FSH units and 75 LH units per ampule) on a daily basis from day 5

through day 9 of the cycle. As they became periovulatory, as evidenced by vaginal hormonal cytologic studies,5 cervical score,6 and on occasion serum estradiollevels,7 sequential hCG (5000 U) was added (Fig. 1). Each patient was required to keep a BBT chart. They were seen on three occasions during each treatment cycle. At each visit they had a routine pelvic examination, vaginal smear for hormonal cytologic studies, cervical score, and evaluation of the temperature charts. The life table (Table 2) was constructed as shown in Figure 2. If the model of a constant pregnancy rate holds, then the log (1 - cumulative pregnancy rate) is plotted versus months, one should obtain a straight line. The slope of this line is equal to log (1 - P). A straight line plot indicates intragroup homogeneity. It should be stressed that any deviation from this theoretic model may make interpretation more difficult, since it could indicate a nonhomogeneous group, or a nonconstant pregnancy rate per month.

RESULTS

Forty-nine patients were treated for 177 cycles with Pergonal in the regimen described above. Figure 1 illustrates the typical treatment program and the points of monitoring in this population. A total of 24 pregnancies were achieved, which represents a combined cumulative pregnancy rate for the 6 months of study of 75%. This represents all patients, including dropouts during varying treatment months. There were 6 pregnancies in the secondary amenorrhea group; 3 pregnancies in the oligo-amenorrhea group; and 15 pregnancies in the poor corpus luteum function group. Among these 24 pregnancies there were 19 single term pregnancies (79%), 2 premature labors associated with multiple gestations (1

~~ 98.6

T!r ~~ p

..... : ...

I

I

7

~ERIOD p p p P P BEGINS

14 21 VAGINAL SMEAR CERVICAL MUCUS ESTRADIOL 5,000 UNITS OF HCG

28

35 POSITIVE BETA SUB UNITS

p. 150 FSH U OF PERGONAL

DAYS

Figure 1 Patient S. B. conceived during the first treatment cycle. She was seen on days 7,14, and 21 of the cycle.

49

Table 2. Life Table No. of cycles

No. of patients

Pre~ancies

1 2 3

10 7 6

4

5

3 1 0 2

ac ·eved

Lost to foUow:up

Patient months of treatment

Pregnancy rate per month

Cumulative proba· bility of pregnancy

0.35 0.15 0 0.80

0.35 0.45 0.45 0.89

0.11 0 0.15 0 0.22 0

0.11 0.11 0.24 0.24 0.41 0.41

0.11 0.13 0 0.54 0 1.0

0.11 0.23 0.23 0.64 0.64 1.0

Secondary amenorrhea 0 8.5 0 6.5 1 5.5 3 2.5

5 6 1

10

2 3 4 5 6

8 7 6 5 4

1 2 3 4

29 26 21 20

5

6

6

4

1 0 1 0 1 0 3 3 0 7 0 2

Oligo-amenorrhea 1 9.0 1 7.5 0 6.5 1 5.5 0 4.5 0 4.0 Poor corpus luteum function

0 2 1 7 2 2

set of triplets and 1 set of sextuplets), and 3 first trimester abortions. The sextuplets went to 34 weeks' gestation, and the quadruplets survived. In this experience, there was a multiple pregnancy rate of 8.3% and a spontaneous abortion rate of 12.5%. The three spontaneous abortions occurred in two patients with poor corpus luteum function and one patient with secondary amenorrhea. The multiple pregnancies occurred in a patient with poor corpus luteum function and another patient with oligo-amenorrhea. In only one cycle was there a hyperstimulation syndrome associated with ascites, and it occurred in a patient with secondary amenorrhea. Table 2 represents the life-table data of the entire study group. Pregnancies are expressed as A) Patient/months of treatment

Number of patients

27.5 23.5 20.5 13.0 5.0 2.0

both the pregnancy rate per month and the cwriulative probability of pregnancy. The normal pregnancy rate from several studies8 • 9 is 17% to 21% per month with unprotected intercourse and no attention paid to ovulatory information. However, our patients had significant attention paid to their ovulatory data including BBT charts, vaginal hormonal cytology, cervical mucus scoring, and estradiol determinations. A population derived from an artificial insemination by donor (AID) group could more clearly approximate a control population for pregnancy rate determination. One such study indicates a pregnancy rate of approximately 25%.10 A pregnancy rate of 17% to 25% was thus used as a guide but not as statistical limits for this study.

Pregnancies achieved each month

+

Dropouts each month

2

B) Pregnancy rate per month (P)

Number of pregnancies each month Patient months of treatment

C) Cumulative probability of pregnancy for any month n 1 - (1 - pregnancy ratemonth 1) (1 - pregnancy ratemonth 2)

50

=

... (1 -

pregnancy ratemonth n)

Figure 2 The formula used for calculation of life-table data.

100

>-

iii~ m>-

,~

.... - ..... POOR C.L. FUNCTION _ _ OLIGOAMENORRHEA

~

-.J~

.... - ... 2" AMENORRHEA

/

80

<{~

OU

o:Z

60

Cl.<{

Z

LU(!) >LU

40

-0:

cundability (F) in the group with secondary amenorrhea was 0.26 (limits 0.05 < F < 0.47). F in the group with oligo-amenorrhea was 0.08 (limits 0 < F < 0.7). F in the group with poor corpus luteum function was 0.16 (limits 0.08 < F < 0.25). However, evaluation of the probability ofpregnancy (slopes, Fig. 4) indicate a statistically significant difference in the patients with secondary

~Cl.

<{

-.JIL..

=>0 :i!: =>

1.0

20

O.A

U

0.6

2

3

4

5

6

MONTHS Figure 3 Cumulative probability of pregnancy derived from the life table data is plotted versus months of treatment. A normal pregnancy rate of 17% to 25% per month (see T test) was used as a guide to calculate the limits as shown.

0.4

>-

u

z z

«

0.2

(!)

2° AMENORRHEA

w

Figure 3 illustrates the cumulative probability of pregnancy plotted versus months of treatment. The patients with secondary amenorrhea and poor corpus luteum function were able to achieve near normal cumulative pregnancy rates. However, the oligo-amenorrheic patients maintained a significantly lower cumulative probability of pregnancy throughout the study. Since it seemed that the pregnancy rate in later months was higher than earlier months, we tested the model for a constant pregnancy rate, using the plot shown in Figure 4. A straight line would indicate a constant pregnancy rate indicating a homogeneous population. A monthly probability of pregnancy for the oligo-amenorrheic group of 8.2% (P = 0.082) is derived, which is constant and linear over the 6-month study period. However, patients with poor corpus luteum function and secondary amenorrhea had conception data that best fit nonlinear or possibly a bilinear plot as shown. In the group with poor corpus luteum function the monthly probability of pregnancy was 9% (P = 0.09) for the first 3 months of treatment and then increased to 90% (P = 0.90) for the next 3 months of treatment. Similarly, in the secondary amenorrheic group, the monthly probability of pregnancy starts at 18% per month (P = 0.18) for the first 3 months and in the 4th month increased to include all patients in this study group. There was no significant difference among the three groups using the average pregnancy rate per month including all months of treatment. Fe-

a:: a..

O.IL-___ L_ _

~

____

~

__

~

_ _ _ _L __ __ L_ _

~

LL

o

0.8

>~

...J

m « m o

0.4

a:: a..

w

> ~ «

POOR 0.1 L-___ L_ _

CORPUS LUTEUM FUNCTION ~

_ _ _ _~_ _~_ _~~_ _~_ _~

...J

::>

~

::>

u

06

I CI

0.4

o

0.2

OLIGOAMENORRHEA 01 L-__

o

~

__

~

____

2

~

3

__

~

____L -__-L__

4

5

6

~

7

MONTHS Figure 4 The cumulative pregnancy rate data derived from the life table was replotted to analyze the model of a constant pregnancy rate per month. This was achieved with a low pregnancy rate per month in the oligo-amenorrheic patients, but with bilinear functions in the two other groups. Log (l - cumulative pregnancy rate) is graphed versus months of treatment. P = pregnancy rate per month derived from the graph data.

51

amenorrhea and the patients with poor corpus luteum function (P = 0.05) between the first 3 months of treatment and the later months of treatment. There was no statistically significant difference between the first 3 months of treatment in all the groups studied; therefore, it appears that a large increase in pregnancy rate was achieved in later months in two study groups, and the model for a constant pregnancy rate does not apply in two of these groups. DISCUSSION

Pregnancy is the ultimate goal in treating individuals with ovulatory defects in infertility. It becomes necessary, however, for us to find homogeneous clinical groups that can be evaluated from a predictive statistical standpoint. For the most part, life-table analysis and proper clinical groupings of patients have not been used to evaluate infertile populations and their response to treatment. Usually data from an infertility population is combined together, independent of the possible intergroup variation. It was for this reason that we elected to divide ovulatory defects into the three clinical categories, as stated earlier, to analyze possible differences. In addition, during the last 3 months of treatment, there was no statistically significant difference between poor corpus luteum function and secondary amenorrhea. Life-table analysis is a convenient method of studying such homogeneous groupings within an infertility population. It allows one to define a starting point as well as an end point in the treatment. It also allows one to calculate a monthly conception rate while taking into account new patients and dropouts from the study. Pregnancy was used as an end point for the statistical analysis. The data for the group with secondary amenorrhea and poor corpus luteum function represent the unexpected finding of nonlinear graphs. There are several explanations for this phenomenon. These include the possibility that in the group with secondary amenorrhea and in the group with poor corpus luteum function a priming effect is needed for the ovary to respond in adequate fashion. This is evidenced by the fact that in the first 3 months the pregnancy rate is lower than in the last 3 months. It may take three cycles of menotropin priming before adequate corpus luteum function is developed and the endometrium is properly prepared for nidation. This

52

concept is supported by a previous review of anovulatory status evaluated following delivery, lactation, and spontaneous abortion. l l These authors noted that in the puerperium it took three to four cycles following delivery or the cessation of lactation for good corpus luteum function to develop. This raises the question of whether or not some ovulatory defects require several treatment cycles of priming. This hypothesis may indicate that previous cycles influence the response of treatment in subsequent cycles. Another possibility is that the individual groups were not a homogeneous population. However, if this were the case, one would expect the higher pregnancy rate per month to occur first followed by a group having a lower pregnancy rate per month. We find, however, in the cases of poor corpus luteum function and secondary amenorrhea groups, the lower probability was achieved, first followed by the higher probability of pregnancy. In addition, there may have been some technical problem in the administration of the drug that may account for the nonlinear graph. This possibility seems remote and certainly did not occur in the oligo-amenorrheic patients who were treated in an identical fashion. We recognize that considerably more patients will be necessary to fully substantiate any of these hypotheses. Clearly, the initial pregnancy rate for all groups of patients were not statistically different. However, further treatment of two groups revealed increases in the pregnancy rate. It is to be stressed that obtaining an average pregnancy rate with Pergonal treatment is not clinically or statistically valid and may lead to errors of interpretation, depending on the type of patients in each group. This regimen, however, is safe and allows for a normal pregnancy rate in patients treated for secondary amenorrhea and poor corpus luteum function. The oligo-amenorrheic patients, on the other hand, were more resistant to therapy, and one must anticipate a longer treatment program when counseling the couple with this particular ovulatory defect.

REFERENCES 1. Cramer DW, Walker AM, Schiff I: Statistical methods in evaluating the outcome of infertility therapy. Fertil Steril 32:80,1979 2. WHO Classification: Moscow, 1975 3. Jones GES, WoodruffDJ, Moszkowski E: The inadequate luteal phase. Am J Obstet Gynecol 83:363, 1962

4. Noyes RW, Hertig A, Rock J: Dating the endometrial biopsy. Fertil Steril 1:3, 1950 5. Goldfarb AF, MacMillan RM: Predictability of responses to menotropin therapy by cytohormonology. Int J Fertil 14:16,1969 6. Insler V, Melmed H, Eichenbremer I, Serr DM, Lunenfeld B: The cervical score: a simple semiquantitative method for monitoring of the menstrual cycle. Int J Gynecol Obstet 10:223, 1972 7. Wu CH: Monitoring of ovulation induction. Fertil Steril 30:617,1978

8. Tietze C, Guttmacher AF, Rubin S: Time required for conception in 1927 planned pregnancies. Fertil Steril 1: 338, 1950 9. Westoff CF, Potter RG, Sagi PC, Mishler EG: Family growth in metropolitan America. Princeton, NJ, Princeton University Press, 1961, p 15 10. Corson SL: Factors affecting donor artificial insemination success rates. Fertil Steril 33:415, 1980 11. Ritter M, Poindexter AN, Buttram VC, Besch PK: Inadequate luteal phase-postpartum human female model. Fertil Steril 29:235, 1978

53