- Email: [email protected]
Treatment-independent pregnancies after cessation of gonadotropin ovulation induction in women with oligomenorrhea and anovulatory menses*
Vol. 50, No. 1, July 1988 Printed in U.S.A.
FERTILITY AND STERILITY Copyright
1988 The American Fertility Society
Treatment-independent pregnancies after cessation of gonadotropin ovulation induction in women with oligomenorrhea and anovulatory menses*
Suk-Yee Lam, M.R.C.O.G.t Gordon Baker, M.D. Roger Pepperell, M.D. James H. Evans, M.D. Department of Obstetrics & Gynaecology, University of Melbourne, Parkville, Victoria, Australia
Life-table analysis was performed for the cumulative spontaneous pregnancy rate (CSPR) of 56 patients with oligomenorrhea and anovulatory cycles who had been treated with gonadotropin for ovulation induction between 1963 and 1985. Twenty-seven had at least one spontaneous pregnancy, giving rise to a CSPR of 66.4% (95% confidence limit [CL] 42.4% to 90.4%) at 115 months for the first spontaneous pregnancy, which is significantly lower than the cumulative induced pregnancy rate (CIPR) of 88.6% at 23 months for the first course of gonadotropin therapy (P < 0.0001). This fertility potential was not affected by the baseline estrogen and follicle-stimulating hormone levels, diagnosis, result of gonadotropin therapy, and age and menstrual pattern during exposure to spontaneous pregnancy by Cox regression analysis. More multiple births occurred in the induced pregnancies than in the spontaneous pregnancies (P = 0.005). Fertil Steril 50:26, 1988
Few studies have reported on spontaneous pregnancy after gonadotropin-induced pregnancy, while the fertility potential of those who have failed to conceive with this mode of therapy has never been reported. GemzelP quoted 8 spontaneous pregnancies in 101 amenorrheic women after human pituitary gonadotropin (hPG)-induced pregnancy. Taymor and Thompson 2 reported 8 spontaneous pregnancies occurring 3 to 30 months after human menopausal gonadotropin (hMG)-induced pregnancy. None of the 24 amenorrheic women conceived spontaneously, while 7 of the 10 patients with polycystic ovarian disease (PCOD) and 1 of the 4 patients with anovulatory cycles did
Received January 8, 1988; revised and accepted March 17, 1988. * Supported in part by the National Health and Medical Research Council of Australia grant 870421. t Reprint requests: Suk-Yee Lam, M.R.C.O.G., Department of Obstetrics and Gynaecology, University of Melbourne, Parkville, Victoria 3052, Australia.
26
Lam et al.
Post gonadotropin fer tility potential
so. Spadoni et al. 3 followed up 50 hMG-treated women and found that 18 had become pregnant spontaneously: 10 with polycystic ovaries, 4 with anovulatory cycles, 3 with unexplained infertility, and 1 with postpartum amenorrhea-galactorrhea. However, they included patients with unexplained infertility in the early part of the study, and the incidences of spontaneous pregnancy in the groups with successful and unsuccessful gonadotropin therapy were not differentiated. Ben-Rafael et al. 4 with the use_of life-table analysis reported a cumulative spontaneous pregnancy rate of 30.4% in 141 patients 5 years after hMG-induced pregnancy, 19.2% for those with low endogenous gonadotropin and estrogen levels (group 1) and 38.6% for those with normal gonadotropin levels and distinct endogenous estrogen activity (group 2). Group 2 included 6 patients with unexplained infertility. Among the group 2 patients, spontaneous pregnancy occurred in only 2 of the 17 amenorrheic women (11.8%) but in 22 of the 67 patients with oligomenorrhea or anovulatory cycles (32.8%). All Fertility and S terility
spontaneous pregnancies occurred in women who had return of menstruation, especially if the periods were regular. These studies suggested that the fertility potential after gonadotropin-induced pregnancy is much more favorable in patients who have oligomenorrhea and anovulatory cycles compared with that of amenorrheic women. In this study, we aim to analyze the spontaneous pregnancy rate in all patients with oligomenorrhea and anovulatory cycles who have been treated with gonadotropin ovulation induction in our hospital whether or not such therapy was successful. Factors that may affect this fertility potential such as diagnosis and baseline estrogen and gonadotropin levels have also been evaluated. MATERIALS AND METHODS
The medical records of all the women who had been treated with gonadotropin induction of ovulation at the Royal Women's Hospital between 1963 and December 1985 were reviewed. The treatment schedule and monitoring procedures have been widely reported and have not changed significantly during the last 24 years. 5-7 HPG (Commonwealth Serum Laboratories of Australia, Victoria) was used exclusively until it was withdrawn in 1984 because of the report of possible association of Creutzfeldt-Jakob disease with human growth hormone therapy in the United States8 and United Kingdom, 9 and hMG (Humegon, N.V. Organon, Oss, Holland) was used in treatments for the remainder of the study period. Of the 300 women treated with gonadotropin for ovulation induction, 87 had oligomenorrhea (defined as periods more than 6 weeks but less than 6 months apart) or anovulatory cycles (defined as cycles less than 6 weeks apart and urinary pregnandiol < 2 mg/24 hours in the midluteal phase of the cycle) before treatment. All patients had received gonadotropin treatment because ovulation did not occur during treatment with clomiphene citrate (CC) or because, although ovulation did occur during treatment with CC, pregnancy did not result. Sixty-two (71.2%) were contacted by letter or phone or attended for interview, while the other 19 had moved and could not be traced despite exhaustive effort. Six of those contacted were not exposed to pregnancy because they were still breast-feeding after a gonadotropin-induced pregnancy or had received hormonal contraception or replacement shortly after stopping gonadotropin therapy; 56 were thus available for the study of the Vol. 50, No. 1, July 1988
spontaneous pregnancy rates. For each patient, details about contraceptive practice, subsequent treatment for infertility, menstrual pattern, and the occurrence of spontaneous or induced pregnancies after completion of gonadotropin therapy were obtained. Life-table analysis (Kaplan Meier) 10 was performed to analyze the cumulative spontaneous pregnancy rate (CSPR) for the first spontaneous pregnancy for all the patients irrespective of whether the gonadotropin therapy was successful or not. The time used in the analysis, the period of exposure to pregnancy, was calculated in months from the date when the last gonadotropin treatment was ceased if it was unsuccessful. If the treatment was successful, period of exposure to pregnancy started from the date of abortion or the date of delivery, and if the patient breast-fed, when breast-feeding was stopped. If breast-feeding was continued for more than 1 year, only 1 year was taken into account because prolonged lactation has a definite but not unlimited effect on the return of fertility after childbirth. 11 The time that the couple were practicing contraception and the duration of subsequent infertility treatment and induced pregnancy were also excluded. Period of exposure to pregnancy was counted up to the date when the patient was contacted, or the date of sterilization (tubal ligation or vasectomy of husband), or the date of hysterectomy, or when she reached 45 years of age, whichever was the earliest. The life-table curve for the CSPR was compared with that for the cumulative induced pregnancy rate with gonadotropin therapy (CIPR). Only the first course of treatment was used. A course of treatment was continued until the patient became pregnant or the treatment was given up. It might be made up of one or more cycles of therapy. As specific radioimmunoassay for {j-subunit human chorionic gonadotropin (fj-hCG) was not available for the patients treated before 1978, biochemical pregnancies were not included for either the CSPR or the CIPR. Log rank test was used for the difference between CSPR and CIPR. The chance of conception for patients under gonadotropin therapy and the fertility potential after cessation of therapy were also analyzed by monthly fecundity rate (f) (number of pregnancies per women-months exposure)P The Cox regression method was used to analyze the factors that might influence the CSPR. These included the baseline estrogen level before gonadotropin therapy (low if urine total estrogen < 10 #Lg/24 hours), the baseline follicule-stiinulating Lam et al. Post gonadotropin fertility potential
27
hormone (FSH) level (low if urinary FSH < 5 mouse units/24 hours or serum FSH by radioimmunoassay< 1 IU/L), whether the patient had oligomenorrhea or anovulatory cycles, whether such menstrual disturbance was caused by polycystic ovarian disease (PCOD), whether the gonadotropin therapy was successful or not, age at the start of period of exposure to pregnancy, and menstrual pattern after cessation of therapy (amenorrhea, oligomenorrhea, or regular menses). A patient was labeled as having PCOD if she had follicular phase serum luteinizing hormone (LH) between 13 and 25 IU /L; a raised LH:FSH ratio of 3 or more; and histologic, laparoscopic, or ultrasonic evidence of the disease. However, histologic criteria were used before the availability of serum FSH and LH testing. The pregnancy outcome of all the gonadotropin-induced and spontaneous pregnancies in this group of patients was analyzed by Fisher's exact test.
Table 1 Life-Table Cumulative Pregnancy Rate Induced with Gonadotropin Therapy (CIPR; n = 62) and Spontaneous Pregnancy Rate After Cessation of Gonadotropin Therapy (CSPR; n = 56) No. of cycles
No. of patient
CPR"
95%CU
Patient Group 1. Gonadotropin-induced pregnancies
0.353 0.239-0.467 3 44 0.363-0.623 29 0.493 5 0.585 0.424-0.746 7 15 0.772 0.648-0.894 .9 10 2 0.886 23 2. Spontaneous pregnancies after cessation of gonadotropin period of exposure to pregnancy in months 12 24 36 60 84 108 115 a
b
37 31 23 19 10 6 5
0.240 0.348 0.439 0.465 0.580 0.580 0.664
0.120-0.360 0.213-0.517 0.287-0.591 0.301-0.629 0.382-0.778 0.324-0.836 0.424-0.904
CPR, cumulative pregnancy rate. CL, confidence limit.
RESULTS Life-table curves for CSPR of the first spontaneous pregnancy after cessation of gonadotropin therapy and CIPR of the first course of gonadotropin treatment for women with oligomenorrhea or anovulatory cycles are shown in Figure 1 and Table 1. Thirty-nine of the 62 patients contacted had become pregnant during gonadotropin treatment, giving rise to a CIPR of 88.6% at the 23rd cycle of treatment, while 27 of the 56 patients exposed to pregnancy after stopping gonadotropin therapy 1{)
0·9 CIPR {n-621
0·8 0·7
•
c:
5"' i"
0..
c
.Q
g
0·5
·-. ..--
CSPR (n-561
•••
0
0::
•
0·6
0-4 0·3
0·2
..•
0
••
0·1
0
20
40
60
80
100
120
140
Months
Figure 1 Life-table curves for cumulative induced pregnancy rate and cumulative spontaneous pregnancy rate.
28
Lam et al. Post gonadotropin fertility potential
had at least one spontaneous pregnancy giving l"ise to a CSPR of 66.4% at 9~ years of exposure for the first spontaneous pregnancy. The maximum period of exposure to pregnancy was 130 months. The difference between the CSPR and CIPR was significant (P < 0.0001). The monthly fecundity rate (f) for the first course of gonadotropin therapy was 0.133 at 1 year (95% confidence limits 0.091 to 0.175) and 0.144 at 23 cycles of treatment (95% confidence limits 0.098 to 0.189). The mean number of months required to achieve conception was 6.9. The monthly fecundity rate for the first spontaneous pregnancy after cessation of treatment was 0.026 at 1 year (95% confidence limits 0.019 to 0.033) and 0.012 at 9~ years (95% confidence limits 0.010 to 0.014). The mean number of months required to achieve conception was 83.3. The patients under gonadotropin therapy had an 11-fold better chance of conceiving in a given cycle . Thirty-five of the 39 pregnancies conceived during the first course of gonadotropin therapy occurred in the 56 women who were subsequently exposed to spontaneous pregnancy. Eighteen of these 35 women subsequently conceived spontaneously (CSPR 48.6% at 3 years and 67.3% at 9~ years) compared with 9 of the 21 patients who have failed gonadotropin therapy (CSPR 35% at 3 years and 58.2% at 9~ years). The difference in CSPR was not statistically significant at 3 years and at 9~ years between the two groups. Fertility and Sterility
Table 2
Factors Analyzed in Cox Regression
Low baseline estrogen level Normal baseline estrogen Low baseline FSH level" Normal baseline FSH Type of ovulatory disorder Oligomenorrhea Anovulatory cycles Diagnosis Polycystic ovaries (PCOD) NoPCOD Successful gonadotropin therapy Unsuccessful gonadotropin therapy Age at start of exposure of preg.b Menstrual pattern after gonadotropin Amenorrhea Oligomenorrhea Regular cycles
Spontaneous pregnancy (n = 27)
No spontaneous pregnancy (n = 29)
7 20 4 16
6 23 1 19
24 3
23 6
4 23
3 26
18
17
9
12
30, 23-40
33, 29-38
5 8 14
11
6 12
• Not available from the record of 17 patients. b Median, range.
The factors analyzed in the Cox regression are shown in Table 2. None of these factors was found to be significantly related to the occurrence of spontaneous pregnancy subsequent to gonadotropin therapy. However, the presence of PCOD approached significance in its association with a higher spontaneous rate (P = 0.063). Of the 27 patients who had become pregnant spontaneously subsequent to gonadotropin therapy, 12 had more than 1 spontaneous pregnancy (8 had 2, 3 had 3, and 1 had 4) and 13 (48.1%) practiced contraception or were sterilized after the spontaneous pregnancies. Two of these 43 spontaneous pregnancies were ongoing. The pregnancy outcomes of the remaining 41 spontaneous pregnancies were compared with the 61 gonadotropininduced pregnancies from all courses of treatment in Table 3. All spontaneous pregnancies were singleton, while 10 of the 46 live births induced by gonadotropin were twins (P < 0.01). Although the pregnancy outcome appeared worse in the gonadotropin-induced pregnancy group than in the spontaneous pregnancy group, none of the other aspects shown in the table was significantly different between the two groups. Four children with congenital abnormalities were born in each group, and all survived. For the gonadotropin-induced pregnancy group, one male infant had Down's syndrome and another had hypospadias, while a pair of twins had Vol. 50, No. 1, July 1988
strawberry angiomata. For the spontaneous pregnancy group, the abnormalities were tetralogy of Fallot, deformed legs, inguinal hernia, and thalassemia minor, respectively. DISCUSSION
Ideally, all clinical trials should be controlled and randomized. However, it will be more acceptable and justifiable to assign patients to the control group if there is a reasonable chance of spontaneous cure. Treatment-independent spontaneous pregnancies have been reported in couples suffering from various causes of infertility .13- 16 We report here a cumulative spontaneous pregnancy rate of about 44% 3 years after stopping gonadotropin therapy in oligomenorrheic patients or those with anovulatory cycles. The pregnancy rate reached a maximum of 66.4% at 9! years, this being higher than that reported in previous studies. Even when the life-table analysis for CSPR is repeated to include patients whom we have not been able to contact (and assuming that they have not had a subsequent spontaneous pregnancy), the CSPR (33.2% at 3 years and 48.4% at 9! years) is not significantly different statistically from the one we reported. However, the CSPR is still significantly lower than the CIPR associated with gonadotropin therapy, which is 58.5% at 6 months and 77.2% at 1 year. Women receiving gonadotropin therapy have an 11-fold better chance of conceiving in a given cycle. Although Ben Rafael et al. 4 found a lower CSPR in the group of patients with low baseline gonadotrophin and estrogen levels, in our study the subsequent spontaneous pregnancy rate is the same whether the patient had low or normal baseline gonadotropin or estrogen levels. However, it needs to be stressed that the patients in our study had Table 3 Comparison of Pregnancy Outcome of Gonadotropin-Induced and Spontaneous Pregnancies
Live birth Singleton Twins Induced abortion Spontaneous abortion First trimester Second trimester Ectopic pregnancy Stillbirth Neonatal deaths Congenital abnormalities
Gonadotropin induced pregnancies
Spontaneous pregnancies
(n = 61)
(n = 41)
36 10 0
34 0 1
5 6 3 1 1 4
3 1 1 1 0 4
Lam et al. Post gonadotropin fertility potential
29
oligomenorrhea or anovulatory cycles and, therefore, presumably "normal" estrogen levels. Ben Rafael et al. 4 also reported that spontaneous pregnancies occurred only in patients with a return of menses, while five of the spontaneous pregnancies in our patients occurred before the return of a period. PCOD appeared to have a favorable fertility potential after gonadotropin therapy in previous studies,2•3 but we failed to confirm this. This may be attributed to changing diagnostic criteria for PCOD from the purely clinical diagnosis of obesity, hirsutism, menstrual disturbance, and infertility to the more recent concept of a continuous spectrum of clinical disorders of varying severity associated with chronic anovulation and an increased LH-toFSH ratio. The diagnostic criteria for PCOD in the previous studies2•3 were not defined. The increased pregnancy rate in the seven patients with PCOD in our study was almost significant. Because our study includes all patients treated with gonadotropin from 1963 and bioassay for total gonadotropin in the urine for the earlier patients is a relatively crude method and does not differentiate between FSH and LH, some of these early patients who were not labeled as having PCOD may be included in this category by more modern diagnostic criteria. However, despite the fact that there were 56 patients with oligomenorrhea and anovulatory cycles in this study, there is still a possibility that the failure to find other significant differences may result from insufficient statistical power because too few patients were able to be included. 17 While the fertility potential of patients who have failed gonadotropin therapy has not been established previously, we have shown in this study that, at least for patients with oligomenorrhea and anovulatory cycles, it is the same as for patients who have conceived with gonadotropin treatment. This has important implications in relation to the advice being given to patients who elect to discontinue gonadotropin therapy under such circumstances. The prognosis of this group of women may be better than we think. It is premature to conclude that the spontaneous pregnancy occurring after gonadotropin therapy is genuinely treatment-independent. The CSPR in this group of oligomenorrheic and anovulatory women in our study is not too low. In the past, these women were usually told that the prospect of subsequent spontaneous pregnancy was poor, especially before the return of regular menses. As gonadotropin ovulation induction is a highly efficacious therapy with a significantly higher number of multiple pregnancies, we propose that contra30
Lam et al. Post gonadotropin fertility potential
ceptive measures or sterilization should be advised if the gonadotropin treatment was successful and the patient has completed her family to save her from the surprise of unplanned pregnancies. REFERENCES 1. Gemzell CA: Treatment of female and male sterility with
human gonadotropins. Acta Obstet Gynecol Scand [Suppl] 48:17,1969 2. Taymor ML, Thompson IE: Spontaneous pregnancy following gonadotropin induced ovulation and conception. Am J Obstet Gynecol113:901, 1972 3. Spadoni LR, Cox DW, Smith DC: Use of human menopausal gonadotropin for the induction of ovulation. Am J Obstet Gynecol120:988, 1974 4. Ben-Rafael Z, Mashiach S, Oelsner G, Farine D, Lunenfeld B, Serr DM: Spontaneous pregnancy and its outcome after human menopausal gonadotropin/human chorionic gonadotropin-induced pregnancy. Fertil Steril 36:560, 1981 5. Townsend SL, Brown JB, Johnstone JW, Adey FD, Evans JH, Taft HP: Induction of ovulation. J Obstet Gynaecol Br Comm 73:529, 1966 6. Brown JB, Evans JH, Adey FD, Taft HP, Townsend L: Factors involved in the induction of fertile ovulations with human gonadotrophins. J Obstet Gynaecol Br Comm 76:289, 1969 7. Brown JB, Pepperell RJ, Evans JH: Disorders of ovulation. In The Infertile Couple, Edited by RJ Pepperell, B Hudson, C Wood. Edinburgh, Churchill Livingstone, 1987, p 45 8. Kock TK, Berg BO, De Armond SJ, Gravina RF: Creutzfeldt-Jakob disease in a young adult with idiopathic hypopituitarism. Possible relation to the administration of cadaveric human growth hormone. N Eng! J Med 313:731, 1985 9. Powell-Jackson J, Weller RO, Kennedy P, Preece MA, Whitcombe EM, Newsom-Davis J: Creutzfeldt-Jakob disease after ac;lministration of human growth hormone. Lancet 2:244, 1985 10. Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. Journal of American Statistics Association 53:457, 1958 11. Brown JB, Harrisson P, Smith MA: A study of returning fertility after childbirth and during lactation by measurement of urinary oestrogen and pregnanediol excretion and cervical mucus production. J Biosoc Sci Suppl 9:5, 1985 12. Cramer DW, Walker AM, Schiff 1: Statistical methods in evaluating the outcome of infertility therapy. Fertil Steril 32:80, 1979 13. Collins JA, Wrixon W, Janes LB, Wilson EH: Treatmentindependent pregnancy among infertile couples. N Eng! J Med 309:1201, 1983 14. Baker HWG, Burger HG, De Kretser DM, Hudson B, Rennie GC, Straffon WGE: Testicular vein ligation and fertility in men with varicoceles. Br Med J 291:1697, 1985 15. Baker HWG: Requirement for controlled therapeutic trials in male infertility. Clin Reprod Fertil4:13, 1986 16. Ben-Rafael Z, Mashiach S, Dor J, Rudak E, Goldman B: Treatment-independent pregnancy after in vitro fertilization and embryo transfer trial. Fertil Steril 45:564, 1986 17. Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, Mantel N, McPherson K, Peto J, Smith PG: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. Br J Cancer 34:585, 1976 Fertility and Sterility
FERTILITY AND STERILITY Copyright
1988 The American Fertility Society
Treatment-independent pregnancies after cessation of gonadotropin ovulation induction in women with oligomenorrhea and anovulatory menses*
Suk-Yee Lam, M.R.C.O.G.t Gordon Baker, M.D. Roger Pepperell, M.D. James H. Evans, M.D. Department of Obstetrics & Gynaecology, University of Melbourne, Parkville, Victoria, Australia
Life-table analysis was performed for the cumulative spontaneous pregnancy rate (CSPR) of 56 patients with oligomenorrhea and anovulatory cycles who had been treated with gonadotropin for ovulation induction between 1963 and 1985. Twenty-seven had at least one spontaneous pregnancy, giving rise to a CSPR of 66.4% (95% confidence limit [CL] 42.4% to 90.4%) at 115 months for the first spontaneous pregnancy, which is significantly lower than the cumulative induced pregnancy rate (CIPR) of 88.6% at 23 months for the first course of gonadotropin therapy (P < 0.0001). This fertility potential was not affected by the baseline estrogen and follicle-stimulating hormone levels, diagnosis, result of gonadotropin therapy, and age and menstrual pattern during exposure to spontaneous pregnancy by Cox regression analysis. More multiple births occurred in the induced pregnancies than in the spontaneous pregnancies (P = 0.005). Fertil Steril 50:26, 1988
Few studies have reported on spontaneous pregnancy after gonadotropin-induced pregnancy, while the fertility potential of those who have failed to conceive with this mode of therapy has never been reported. GemzelP quoted 8 spontaneous pregnancies in 101 amenorrheic women after human pituitary gonadotropin (hPG)-induced pregnancy. Taymor and Thompson 2 reported 8 spontaneous pregnancies occurring 3 to 30 months after human menopausal gonadotropin (hMG)-induced pregnancy. None of the 24 amenorrheic women conceived spontaneously, while 7 of the 10 patients with polycystic ovarian disease (PCOD) and 1 of the 4 patients with anovulatory cycles did
Received January 8, 1988; revised and accepted March 17, 1988. * Supported in part by the National Health and Medical Research Council of Australia grant 870421. t Reprint requests: Suk-Yee Lam, M.R.C.O.G., Department of Obstetrics and Gynaecology, University of Melbourne, Parkville, Victoria 3052, Australia.
26
Lam et al.
Post gonadotropin fer tility potential
so. Spadoni et al. 3 followed up 50 hMG-treated women and found that 18 had become pregnant spontaneously: 10 with polycystic ovaries, 4 with anovulatory cycles, 3 with unexplained infertility, and 1 with postpartum amenorrhea-galactorrhea. However, they included patients with unexplained infertility in the early part of the study, and the incidences of spontaneous pregnancy in the groups with successful and unsuccessful gonadotropin therapy were not differentiated. Ben-Rafael et al. 4 with the use_of life-table analysis reported a cumulative spontaneous pregnancy rate of 30.4% in 141 patients 5 years after hMG-induced pregnancy, 19.2% for those with low endogenous gonadotropin and estrogen levels (group 1) and 38.6% for those with normal gonadotropin levels and distinct endogenous estrogen activity (group 2). Group 2 included 6 patients with unexplained infertility. Among the group 2 patients, spontaneous pregnancy occurred in only 2 of the 17 amenorrheic women (11.8%) but in 22 of the 67 patients with oligomenorrhea or anovulatory cycles (32.8%). All Fertility and S terility
spontaneous pregnancies occurred in women who had return of menstruation, especially if the periods were regular. These studies suggested that the fertility potential after gonadotropin-induced pregnancy is much more favorable in patients who have oligomenorrhea and anovulatory cycles compared with that of amenorrheic women. In this study, we aim to analyze the spontaneous pregnancy rate in all patients with oligomenorrhea and anovulatory cycles who have been treated with gonadotropin ovulation induction in our hospital whether or not such therapy was successful. Factors that may affect this fertility potential such as diagnosis and baseline estrogen and gonadotropin levels have also been evaluated. MATERIALS AND METHODS
The medical records of all the women who had been treated with gonadotropin induction of ovulation at the Royal Women's Hospital between 1963 and December 1985 were reviewed. The treatment schedule and monitoring procedures have been widely reported and have not changed significantly during the last 24 years. 5-7 HPG (Commonwealth Serum Laboratories of Australia, Victoria) was used exclusively until it was withdrawn in 1984 because of the report of possible association of Creutzfeldt-Jakob disease with human growth hormone therapy in the United States8 and United Kingdom, 9 and hMG (Humegon, N.V. Organon, Oss, Holland) was used in treatments for the remainder of the study period. Of the 300 women treated with gonadotropin for ovulation induction, 87 had oligomenorrhea (defined as periods more than 6 weeks but less than 6 months apart) or anovulatory cycles (defined as cycles less than 6 weeks apart and urinary pregnandiol < 2 mg/24 hours in the midluteal phase of the cycle) before treatment. All patients had received gonadotropin treatment because ovulation did not occur during treatment with clomiphene citrate (CC) or because, although ovulation did occur during treatment with CC, pregnancy did not result. Sixty-two (71.2%) were contacted by letter or phone or attended for interview, while the other 19 had moved and could not be traced despite exhaustive effort. Six of those contacted were not exposed to pregnancy because they were still breast-feeding after a gonadotropin-induced pregnancy or had received hormonal contraception or replacement shortly after stopping gonadotropin therapy; 56 were thus available for the study of the Vol. 50, No. 1, July 1988
spontaneous pregnancy rates. For each patient, details about contraceptive practice, subsequent treatment for infertility, menstrual pattern, and the occurrence of spontaneous or induced pregnancies after completion of gonadotropin therapy were obtained. Life-table analysis (Kaplan Meier) 10 was performed to analyze the cumulative spontaneous pregnancy rate (CSPR) for the first spontaneous pregnancy for all the patients irrespective of whether the gonadotropin therapy was successful or not. The time used in the analysis, the period of exposure to pregnancy, was calculated in months from the date when the last gonadotropin treatment was ceased if it was unsuccessful. If the treatment was successful, period of exposure to pregnancy started from the date of abortion or the date of delivery, and if the patient breast-fed, when breast-feeding was stopped. If breast-feeding was continued for more than 1 year, only 1 year was taken into account because prolonged lactation has a definite but not unlimited effect on the return of fertility after childbirth. 11 The time that the couple were practicing contraception and the duration of subsequent infertility treatment and induced pregnancy were also excluded. Period of exposure to pregnancy was counted up to the date when the patient was contacted, or the date of sterilization (tubal ligation or vasectomy of husband), or the date of hysterectomy, or when she reached 45 years of age, whichever was the earliest. The life-table curve for the CSPR was compared with that for the cumulative induced pregnancy rate with gonadotropin therapy (CIPR). Only the first course of treatment was used. A course of treatment was continued until the patient became pregnant or the treatment was given up. It might be made up of one or more cycles of therapy. As specific radioimmunoassay for {j-subunit human chorionic gonadotropin (fj-hCG) was not available for the patients treated before 1978, biochemical pregnancies were not included for either the CSPR or the CIPR. Log rank test was used for the difference between CSPR and CIPR. The chance of conception for patients under gonadotropin therapy and the fertility potential after cessation of therapy were also analyzed by monthly fecundity rate (f) (number of pregnancies per women-months exposure)P The Cox regression method was used to analyze the factors that might influence the CSPR. These included the baseline estrogen level before gonadotropin therapy (low if urine total estrogen < 10 #Lg/24 hours), the baseline follicule-stiinulating Lam et al. Post gonadotropin fertility potential
27
hormone (FSH) level (low if urinary FSH < 5 mouse units/24 hours or serum FSH by radioimmunoassay< 1 IU/L), whether the patient had oligomenorrhea or anovulatory cycles, whether such menstrual disturbance was caused by polycystic ovarian disease (PCOD), whether the gonadotropin therapy was successful or not, age at the start of period of exposure to pregnancy, and menstrual pattern after cessation of therapy (amenorrhea, oligomenorrhea, or regular menses). A patient was labeled as having PCOD if she had follicular phase serum luteinizing hormone (LH) between 13 and 25 IU /L; a raised LH:FSH ratio of 3 or more; and histologic, laparoscopic, or ultrasonic evidence of the disease. However, histologic criteria were used before the availability of serum FSH and LH testing. The pregnancy outcome of all the gonadotropin-induced and spontaneous pregnancies in this group of patients was analyzed by Fisher's exact test.
Table 1 Life-Table Cumulative Pregnancy Rate Induced with Gonadotropin Therapy (CIPR; n = 62) and Spontaneous Pregnancy Rate After Cessation of Gonadotropin Therapy (CSPR; n = 56) No. of cycles
No. of patient
CPR"
95%CU
Patient Group 1. Gonadotropin-induced pregnancies
0.353 0.239-0.467 3 44 0.363-0.623 29 0.493 5 0.585 0.424-0.746 7 15 0.772 0.648-0.894 .9 10 2 0.886 23 2. Spontaneous pregnancies after cessation of gonadotropin period of exposure to pregnancy in months 12 24 36 60 84 108 115 a
b
37 31 23 19 10 6 5
0.240 0.348 0.439 0.465 0.580 0.580 0.664
0.120-0.360 0.213-0.517 0.287-0.591 0.301-0.629 0.382-0.778 0.324-0.836 0.424-0.904
CPR, cumulative pregnancy rate. CL, confidence limit.
RESULTS Life-table curves for CSPR of the first spontaneous pregnancy after cessation of gonadotropin therapy and CIPR of the first course of gonadotropin treatment for women with oligomenorrhea or anovulatory cycles are shown in Figure 1 and Table 1. Thirty-nine of the 62 patients contacted had become pregnant during gonadotropin treatment, giving rise to a CIPR of 88.6% at the 23rd cycle of treatment, while 27 of the 56 patients exposed to pregnancy after stopping gonadotropin therapy 1{)
0·9 CIPR {n-621
0·8 0·7
•
c:
5"' i"
0..
c
.Q
g
0·5
·-. ..--
CSPR (n-561
•••
0
0::
•
0·6
0-4 0·3
0·2
..•
0
••
0·1
0
20
40
60
80
100
120
140
Months
Figure 1 Life-table curves for cumulative induced pregnancy rate and cumulative spontaneous pregnancy rate.
28
Lam et al. Post gonadotropin fertility potential
had at least one spontaneous pregnancy giving l"ise to a CSPR of 66.4% at 9~ years of exposure for the first spontaneous pregnancy. The maximum period of exposure to pregnancy was 130 months. The difference between the CSPR and CIPR was significant (P < 0.0001). The monthly fecundity rate (f) for the first course of gonadotropin therapy was 0.133 at 1 year (95% confidence limits 0.091 to 0.175) and 0.144 at 23 cycles of treatment (95% confidence limits 0.098 to 0.189). The mean number of months required to achieve conception was 6.9. The monthly fecundity rate for the first spontaneous pregnancy after cessation of treatment was 0.026 at 1 year (95% confidence limits 0.019 to 0.033) and 0.012 at 9~ years (95% confidence limits 0.010 to 0.014). The mean number of months required to achieve conception was 83.3. The patients under gonadotropin therapy had an 11-fold better chance of conceiving in a given cycle . Thirty-five of the 39 pregnancies conceived during the first course of gonadotropin therapy occurred in the 56 women who were subsequently exposed to spontaneous pregnancy. Eighteen of these 35 women subsequently conceived spontaneously (CSPR 48.6% at 3 years and 67.3% at 9~ years) compared with 9 of the 21 patients who have failed gonadotropin therapy (CSPR 35% at 3 years and 58.2% at 9~ years). The difference in CSPR was not statistically significant at 3 years and at 9~ years between the two groups. Fertility and Sterility
Table 2
Factors Analyzed in Cox Regression
Low baseline estrogen level Normal baseline estrogen Low baseline FSH level" Normal baseline FSH Type of ovulatory disorder Oligomenorrhea Anovulatory cycles Diagnosis Polycystic ovaries (PCOD) NoPCOD Successful gonadotropin therapy Unsuccessful gonadotropin therapy Age at start of exposure of preg.b Menstrual pattern after gonadotropin Amenorrhea Oligomenorrhea Regular cycles
Spontaneous pregnancy (n = 27)
No spontaneous pregnancy (n = 29)
7 20 4 16
6 23 1 19
24 3
23 6
4 23
3 26
18
17
9
12
30, 23-40
33, 29-38
5 8 14
11
6 12
• Not available from the record of 17 patients. b Median, range.
The factors analyzed in the Cox regression are shown in Table 2. None of these factors was found to be significantly related to the occurrence of spontaneous pregnancy subsequent to gonadotropin therapy. However, the presence of PCOD approached significance in its association with a higher spontaneous rate (P = 0.063). Of the 27 patients who had become pregnant spontaneously subsequent to gonadotropin therapy, 12 had more than 1 spontaneous pregnancy (8 had 2, 3 had 3, and 1 had 4) and 13 (48.1%) practiced contraception or were sterilized after the spontaneous pregnancies. Two of these 43 spontaneous pregnancies were ongoing. The pregnancy outcomes of the remaining 41 spontaneous pregnancies were compared with the 61 gonadotropininduced pregnancies from all courses of treatment in Table 3. All spontaneous pregnancies were singleton, while 10 of the 46 live births induced by gonadotropin were twins (P < 0.01). Although the pregnancy outcome appeared worse in the gonadotropin-induced pregnancy group than in the spontaneous pregnancy group, none of the other aspects shown in the table was significantly different between the two groups. Four children with congenital abnormalities were born in each group, and all survived. For the gonadotropin-induced pregnancy group, one male infant had Down's syndrome and another had hypospadias, while a pair of twins had Vol. 50, No. 1, July 1988
strawberry angiomata. For the spontaneous pregnancy group, the abnormalities were tetralogy of Fallot, deformed legs, inguinal hernia, and thalassemia minor, respectively. DISCUSSION
Ideally, all clinical trials should be controlled and randomized. However, it will be more acceptable and justifiable to assign patients to the control group if there is a reasonable chance of spontaneous cure. Treatment-independent spontaneous pregnancies have been reported in couples suffering from various causes of infertility .13- 16 We report here a cumulative spontaneous pregnancy rate of about 44% 3 years after stopping gonadotropin therapy in oligomenorrheic patients or those with anovulatory cycles. The pregnancy rate reached a maximum of 66.4% at 9! years, this being higher than that reported in previous studies. Even when the life-table analysis for CSPR is repeated to include patients whom we have not been able to contact (and assuming that they have not had a subsequent spontaneous pregnancy), the CSPR (33.2% at 3 years and 48.4% at 9! years) is not significantly different statistically from the one we reported. However, the CSPR is still significantly lower than the CIPR associated with gonadotropin therapy, which is 58.5% at 6 months and 77.2% at 1 year. Women receiving gonadotropin therapy have an 11-fold better chance of conceiving in a given cycle. Although Ben Rafael et al. 4 found a lower CSPR in the group of patients with low baseline gonadotrophin and estrogen levels, in our study the subsequent spontaneous pregnancy rate is the same whether the patient had low or normal baseline gonadotropin or estrogen levels. However, it needs to be stressed that the patients in our study had Table 3 Comparison of Pregnancy Outcome of Gonadotropin-Induced and Spontaneous Pregnancies
Live birth Singleton Twins Induced abortion Spontaneous abortion First trimester Second trimester Ectopic pregnancy Stillbirth Neonatal deaths Congenital abnormalities
Gonadotropin induced pregnancies
Spontaneous pregnancies
(n = 61)
(n = 41)
36 10 0
34 0 1
5 6 3 1 1 4
3 1 1 1 0 4
Lam et al. Post gonadotropin fertility potential
29
oligomenorrhea or anovulatory cycles and, therefore, presumably "normal" estrogen levels. Ben Rafael et al. 4 also reported that spontaneous pregnancies occurred only in patients with a return of menses, while five of the spontaneous pregnancies in our patients occurred before the return of a period. PCOD appeared to have a favorable fertility potential after gonadotropin therapy in previous studies,2•3 but we failed to confirm this. This may be attributed to changing diagnostic criteria for PCOD from the purely clinical diagnosis of obesity, hirsutism, menstrual disturbance, and infertility to the more recent concept of a continuous spectrum of clinical disorders of varying severity associated with chronic anovulation and an increased LH-toFSH ratio. The diagnostic criteria for PCOD in the previous studies2•3 were not defined. The increased pregnancy rate in the seven patients with PCOD in our study was almost significant. Because our study includes all patients treated with gonadotropin from 1963 and bioassay for total gonadotropin in the urine for the earlier patients is a relatively crude method and does not differentiate between FSH and LH, some of these early patients who were not labeled as having PCOD may be included in this category by more modern diagnostic criteria. However, despite the fact that there were 56 patients with oligomenorrhea and anovulatory cycles in this study, there is still a possibility that the failure to find other significant differences may result from insufficient statistical power because too few patients were able to be included. 17 While the fertility potential of patients who have failed gonadotropin therapy has not been established previously, we have shown in this study that, at least for patients with oligomenorrhea and anovulatory cycles, it is the same as for patients who have conceived with gonadotropin treatment. This has important implications in relation to the advice being given to patients who elect to discontinue gonadotropin therapy under such circumstances. The prognosis of this group of women may be better than we think. It is premature to conclude that the spontaneous pregnancy occurring after gonadotropin therapy is genuinely treatment-independent. The CSPR in this group of oligomenorrheic and anovulatory women in our study is not too low. In the past, these women were usually told that the prospect of subsequent spontaneous pregnancy was poor, especially before the return of regular menses. As gonadotropin ovulation induction is a highly efficacious therapy with a significantly higher number of multiple pregnancies, we propose that contra30
Lam et al. Post gonadotropin fertility potential
ceptive measures or sterilization should be advised if the gonadotropin treatment was successful and the patient has completed her family to save her from the surprise of unplanned pregnancies. REFERENCES 1. Gemzell CA: Treatment of female and male sterility with
human gonadotropins. Acta Obstet Gynecol Scand [Suppl] 48:17,1969 2. Taymor ML, Thompson IE: Spontaneous pregnancy following gonadotropin induced ovulation and conception. Am J Obstet Gynecol113:901, 1972 3. Spadoni LR, Cox DW, Smith DC: Use of human menopausal gonadotropin for the induction of ovulation. Am J Obstet Gynecol120:988, 1974 4. Ben-Rafael Z, Mashiach S, Oelsner G, Farine D, Lunenfeld B, Serr DM: Spontaneous pregnancy and its outcome after human menopausal gonadotropin/human chorionic gonadotropin-induced pregnancy. Fertil Steril 36:560, 1981 5. Townsend SL, Brown JB, Johnstone JW, Adey FD, Evans JH, Taft HP: Induction of ovulation. J Obstet Gynaecol Br Comm 73:529, 1966 6. Brown JB, Evans JH, Adey FD, Taft HP, Townsend L: Factors involved in the induction of fertile ovulations with human gonadotrophins. J Obstet Gynaecol Br Comm 76:289, 1969 7. Brown JB, Pepperell RJ, Evans JH: Disorders of ovulation. In The Infertile Couple, Edited by RJ Pepperell, B Hudson, C Wood. Edinburgh, Churchill Livingstone, 1987, p 45 8. Kock TK, Berg BO, De Armond SJ, Gravina RF: Creutzfeldt-Jakob disease in a young adult with idiopathic hypopituitarism. Possible relation to the administration of cadaveric human growth hormone. N Eng! J Med 313:731, 1985 9. Powell-Jackson J, Weller RO, Kennedy P, Preece MA, Whitcombe EM, Newsom-Davis J: Creutzfeldt-Jakob disease after ac;lministration of human growth hormone. Lancet 2:244, 1985 10. Kaplan EL, Meier P: Nonparametric estimation from incomplete observations. Journal of American Statistics Association 53:457, 1958 11. Brown JB, Harrisson P, Smith MA: A study of returning fertility after childbirth and during lactation by measurement of urinary oestrogen and pregnanediol excretion and cervical mucus production. J Biosoc Sci Suppl 9:5, 1985 12. Cramer DW, Walker AM, Schiff 1: Statistical methods in evaluating the outcome of infertility therapy. Fertil Steril 32:80, 1979 13. Collins JA, Wrixon W, Janes LB, Wilson EH: Treatmentindependent pregnancy among infertile couples. N Eng! J Med 309:1201, 1983 14. Baker HWG, Burger HG, De Kretser DM, Hudson B, Rennie GC, Straffon WGE: Testicular vein ligation and fertility in men with varicoceles. Br Med J 291:1697, 1985 15. Baker HWG: Requirement for controlled therapeutic trials in male infertility. Clin Reprod Fertil4:13, 1986 16. Ben-Rafael Z, Mashiach S, Dor J, Rudak E, Goldman B: Treatment-independent pregnancy after in vitro fertilization and embryo transfer trial. Fertil Steril 45:564, 1986 17. Peto R, Pike MC, Armitage P, Breslow NE, Cox DR, Howard SV, Mantel N, McPherson K, Peto J, Smith PG: Design and analysis of randomized clinical trials requiring prolonged observation of each patient. Br J Cancer 34:585, 1976 Fertility and Sterility