The outcome of in-vitro fertilization treatment in women with sonographic evidence of polycystic ovarian morphology

Transcription

1 Human Reproduction vol.14 no.1 pp , 1999 The outcome of in-vitro fertilization treatment in women with sonographic evidence of polycystic ovarian morphology Lawrence Engmann 1,2,5, Noreen Maconochie 3, Povilas Sladkevicius 1,4, Jinan Bekir 1, Stuart Campbell 1,4 and Seang Lin Tan 1,2 1 The London Women s Clinic, 115 Harley Street, London, 2 Department of Obstetrics and Gynaecology, Royal Victoria Hospital, McGill University, Montreal, Canada, 3 Department of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine and 4 Department of Obstetrics and Gynaecology, St. George s Hospital Medical School, London, UK 5 To whom correspondence should be addressed at: Department of Obstetrics and Gynaecology, Royal Victoria Hospital, McGill University, 687 Pine Avenue West, Montreal QC H3A 1A1, Canada This study compared the outcome of a course of up to three cycles of in-vitro fertilization (IVF) treatment in 46 women (97 cycles) who had polycystic ovaries (PCO) seen on ultrasound scan, but who had no clinical symptomatology associated with polycystic ovary syndrome, with that of 145 women (332 cycles) who had normal ovarian morphology on ultrasound examination. All 191 women had normal early follicular phase serum follicle stimulating hormone (FSH) concentrations, were less than 40 years of age and used the long protocol of pituitary suppression with gonadotrophinreleasing hormone agonist therapy. On average, the women with PCO produced more follicles, oocytes and embryos than the women with normal ovaries, but the fertilization, cleavage and miscarriage rates were similar. Adjusted for age, the odds of achieving a pregnancy within three cycles of treatment in a woman with PCO were 69% higher than those of a woman with normal ovaries [odds ratio (OR): 1.69, 95% confidence interval (CI) , P 0.05)] and the odds of achieving a live birth were 82% higher (OR: 1.82, 95% CI , P 0.03). There is, therefore, evidence that outcome of IVF treatment for women with PCO seen on ultrasound examination may be better than that for women with normal ovaries. Key words: IVF/normal ovary/outcome/ovarian response/pco Introduction Polycystic ovarian syndrome (PCOS) is a heterogeneous disorder which has clinical, biochemical and ultrasonographic features. At one end of the spectrum of this disorder there is a large group of women who manifest sonographic evidence of polycystic ovaries (PCO) but who do not have any clinical manifestation of the syndrome (Jacobs, 1987; Franks, 1995). Several studies have shown that the pregnancy rate for women with PCOS undergoing in-vitro fertilization (IVF) treatment is comparable with that of women with other causes of infertility (Dor et al., 1990; Dale et al., 1991; Homburg et al., 1993; MacDougall et al., 1993). Every one of these studies, however, have included women at the other end of the spectrum of the disorder, i.e. women with clinical or endocrine manifestations of the syndrome. There are, therefore, no data on the outcome of IVF treatment in women who have PCO diagnosed on ultrasound but who do not have clinical manifestations of the syndrome. The purpose of the present study was to evaluate the outcome of a course of up to three cycles of IVF treatment in women with a variety of indications who had sonographic evidence of PCO, but no clinical symptomatology associated with PCOS, compared with that of women who had normal ovarian morphology on pelvic ultrasonography. Materials and methods Patients Couples who booked a three-cycle package of IVF or intracytoplasmic sperm injection (ICSI) at the London Women s Clinic from 1993 to 1995, in which they were eligible to undergo up to three cycles of IVF treatment within 1 year, were included in the study. This package offers three cycles of treatment, taken within a year, for the price of less than two individual cycles. Any couple presenting at the clinic is allowed to undertake the package, regardless of prognosis. Upon presentation, all couples are given appropriate counselling regarding their chances of pregnancy after one and multiple cycles of treatment and the financial risks of undertaking a three-cycle package, but the decision whether or not to use the package is entirely the decision of the couple. The three cycles had to be taken within a period of 1 year with the end-point being an ongoing pregnancy beyond 12 weeks of gestation, that is, the end of the first trimester. The package was therefore no longer valid should a pregnancy continue beyond 12 weeks but was subsequently lost through miscarriage, stillbirth or other obstetric complications. If, however, the pregnancy failed within the first 12 weeks of pregnancy, the patient was then eligible to continue using the remainder of the package. Only data from women undergoing IVF treatment for the first time who were less than 40 years of age and who had a normal early follicular phase serum follicle stimulating hormone (FSH) concentration of 10 IU/l were included in the analyses. The couples had various causes of infertility, but all the female partners had regular menstrual cycles and no symptoms of hyperandrogenism. Women whose male partners had severe male factor infertility (i.e spermatozoa with progressive motility, 90% abnormal spermatozoa), or who had failed fertilization in a previous IVF treatment cycle, were offered ICSI. All the women underwent transvaginal ultrasonography to assess uterine and ovarian morphology on day 2 or day 3 of the menstrual cycle, and they were then divided into two groups, depending on whether they had ultrasonic evidence of PCO. PCO were diagnosed European Society of Human Reproduction and Embryology 167

2 L.Engmann et al. if the ultrasound scan showed 10 or more cysts measuring 2 8 mm in diameter arranged peripherally around a dense core of stroma or scattered through an increased amount of stroma (Adams et al., 1985). Treatment protocol All the women used our standard long protocol of pituitary suppression with gonadotrophin-releasing hormone (GnRH) agonist, followed by administration of urinary gonadotrophins for ovarian stimulation (Tan et al., 1992). The standard starting dose of urinary gonadotrophins was 2 to 4 ampoules ( IU FSH activity) per day depending on the patient s age, basal serum FSH concentrations and presence or absence of PCO on ultrasound assessment. This standard starting regimen then varied in succeeding cycles depending on the patient response. Monitoring of follicular growth was achieved with serial ultrasound scans and the dose of gonadotrophin adjusted according to follicular response. The management of the IVF treatment cycle was as previously described (Engmann et al., 1998). Clinical pregnancy was defined as positive urine β-human chorionic gonadotrophin (HCG) test with ultrasonic evidence of a gestational sac. Fresh embryo transfer was not offered to patients who had serum oestradiol concentrations pmol/l on the day of HCG administration, or oestradiol concentrations between pmol/l with more than 15 oocytes retrieved because of the increased risk of ovarian hyperstimulation syndrome (OHSS). All embryos for these patients were cryopreserved and transferred in a subsequent cycle. Patients with OHSS were classified into three groups: mild, moderate and severe (Schenker, 1993). Statistical analysis Mean values of continuous variables estimated across all cycles were calculated using linear regression analysis, assessing statistical significance using the F-test (Draper and Smith, 1981). For each woman and cycle, the fertilization rate, cleavage rate and implantation rate (number of gestational sacs, assessed by ultrasound at 6 weeks gestation, out of number of embryos transferred), was calculated. Maximum likelihood estimates of these rates, averaged over all cycles and women, were calculated using weighted least squares regression, and statistical significance was then assessed using the F-test (Draper and Smith, 1981). Characteristics estimated at one point in time, such as age at the first cycle or cause of infertility, were summarized and compared using means, standard deviations and t-tests, or percentage distributions and χ 2 tests, as appropriate (Altman, 1991). The effect of ovarian morphology on probability of pregnancy, live birth or miscarriage was estimated using odds ratios (OR), by means of logistic regression analysis, adjusting for cycle number and age (Cox, 1972; Breslow and Day, 1980; D Agostino et al., 1990). Statistical significance was assessed using the likelihood ratio test. Because of the association between observations taken from cycles undertaken by the same woman, it was inappropriate to calculate standard errors for OR and mean values using standard methods, where there is an assumption of independence between observations. A robust method was therefore used in all regression analyses to calculate correct standard errors (in the measurement sense), based on the sandwich estimate of the standard error (Huber, 1967; Rogers, 1993). Cumulative conception and live birth rates were analysed using the life-table approach [Kaplan Meier estimates with corresponding Greenwood 95% confidence intervals (CI)], which accounts for the variable number of cycles undertaken by the women (Kalbfleisch and Prentice, 1980). In accordance with life-table methods, calculation of cumulative conception and live birth rates for the first course of treatment used data on all treatment cycles leading to the first clinical pregnancy and first live birth respectively. 168 Table I. Characteristics at first cycle of women with polycystic ovaries (PCO) and normal ovaries a Characteristics PCO Normal ovaries (n 46) (n 145) Mean age in years (SD) 32.6 (2.9) 33.6 (3.4) Mean duration of infertility in years (SD) 3.7 (2.1) 4.3 (2.7) Cause of infertility (n %) Male factor 17 (37) 44 (30) Tubal disease 11 (24) 26 (18) Endometriosis 3 (7) 13 (9) Unexplained 12 (26) 37 (26) Multiple 3 (7) 25 (17) a No significant difference in any result. All P-values quoted are two-sided, and values below 0.05 were taken to indicate statistical significance. All analyses were performed using Stata statistical package (Statacorp, 1996) Results A total of 191 women undergoing a total of 429 cycles fulfilled the study criteria. They comprised 46 women with PCO who underwent a total of 97 IVF treatment cycles and 145 women with normal ovaries who underwent a total of 332 cycles. There were no significant differences between the two groups with respect to age, duration of infertility and cause of infertility (Table I). ICSI was performed in nine cycles (9.3%, six women) among women with PCO and 34 cycles (10.2%, 16 women) among women with normal ovaries. Three (6.5%) women with PCO and 23 (15.9%) women with normal ovaries failed to complete all three cycles of treatment, within the agreed period of 1 year, for reasons other than achieving a live birth from the IVF treatment: two women with PCO (4.3%) and four with normal ovaries (2.8%) achieved natural conception between cycles which progressed to a live birth. The remaining one woman with PCO (2.2%) and 19 women with normal ovaries (13.1%) dropped out for reasons unrelated to outcome, such as divorce. The unsuccessful IVF attempts of these 26 women are included in all analyses. No cycles were abandoned among women with PCO, whilst among the women with normal ovaries, seven (2.1%, six women) cycles were abandoned on account of poor ovarian response. Two women with PCO had all their embryos cryopreserved because they were deemed to have an increased risk of developing OHSS. The ovarian response to stimulation is shown in Table II. On average, women with PCO required 16.4 less ampoules of gonadotrophin and 0.8 fewer days of ovarian stimulation compared with women with normal ovaries. Furthermore, on average, women with PCO produced 7.5 more follicles, 3.2 more viable oocytes and 1.3 more embryos compared with those who had normal ovaries. There was no significant difference in fertilization rate between women with PCO and those with normal ovaries (59.4% versus 60.2% respectively, P 0.74, adjusted for age, cycle and woman). Failed fertilization occurred in five cycles (5.1%, four women) among women with PCO and 18 cycles (5.4%, 17 women) among women with normal ovaries. There

3 Outcome of IVF in women with PCO Table II. Characteristics of ovarian response in women with pco and those with normal ovaries a Characteristics PCO Normal ovaries P-value b (n 97) (n 332) Mean (SE) a Mean (SE) a Duration of ovarian stimulation (days) 11.1 (0.3) 11.9 (0.2) 0.03 No. of ampoules of HMG 25.6 (1.7) 42.0 (1.8) Total no. of follicles 20.6 (1.3) 13.1 (0.4) No. of follicles ( 14 mm) 10.2 (0.5) 7.8 (0.2) No. of follicles (10 14 mm) 6.2 (0.4) 3.4 (0.1) Endometrial thickness (mm) 10.8 (0.3) 10.7 (0.2) NS Peak oestradiol concentration (pmol/l) 8254 (476) 6899 (245) 0.01 No. of viable oocytes 11.7 (0.6) 8.5 (0.3) Embryos produced 5.9 (0.5) 4.6 (0.2) 0.01 No. of embryos transferred 2.3 (0.1) 2.3 (0.1) NS NS not significant; HMG human menopausal gonadotrophin. a Averaged over all cycles undertaken. Standard errors (SE) corrected for correlation between outcomes of cycles undertaken by the same woman (see methods). b P-value for difference between means. Table III. Relative probability (odds ratios, OR) of clinical pregnancy or live birth in women with polycystic ovaries (PCO) compared with that of women with normal ovaries a Clinical pregnancy Live birth Ovarian morphology OR (95% CI) P-value OR (95% CI) P-value Normal ovaries PCO 1.69 ( ) ( ) 0.03 a Adjusted for age at first cycle and cycle number. Standard errors, used in calculating confidence interval (CI), corrected for correlation between outcomes of cycles undertaken by the same woman (see methods). was no significant difference in cleavage rate for women with PCO (91%) compared with women with normal ovaries (89.4%) (P 0.44, adjusted for age, cycle and woman). Failed cleavage occurred in none of the cycles in women with PCO and three cycles (0.9%, three women) among women with normal ovaries. Though slightly higher among women with PCO, the implantation rates were not significantly different between the two groups, at an average of 18.6% versus 11.7% (P 0.12, adjusted for age, cycle and woman). Ten cycles (10.3%, nine women) in women with PCO resulted in moderate (nine cycles, 9.3%) or severe (one cycle, 1.0%) OHSS while one cycle (0.3%, one woman) in women with normal ovaries resulted in moderate OHSS. The odds of a woman with PCO conceiving were 77% higher than those of a woman with normal ovaries [OR: 1.77, 95% CI, , P 0.03). After adjusting for age this crude OR was only slightly less at 1.69 (95% CI , P 0.05) (Table III). Similarly, the odds of a woman with PCO having a live birth were 90% higher than those of a woman with normal ovaries (OR: 1.90, 95% CI ; P 0.02). Adjusting for age reduced this odds ratio to 1.82 (95% CI , P 0.03). There was no significant difference in the clinical miscarriage rate between women with PCO and those with normal ovaries [6.9% (two pregnancies) versus 11.1% (seven pregnancies); P 0.51, adjusted for age, cycle and woman]. The cumulative conception rate after three cycles of treat- ment was 70.8% (95% CI 54.6% 85.3%) among women with PCO and 54.7% (95% CI 44.8% 65.3%) among women with normal ovaries (Table IV). The cumulative live birth rate after three cycles of treatment was 66.5% (95% CI ) in women with PCO compared with 46.9% (95% CI ) in women with normal ovaries (Table IV). Discussion The results of this study suggest that the odds of achieving a live birth, after undergoing a course of up to three cycles of treatment, in women who have PCO seen on ultrasound examination but who have no clinical manifestation associated with PCOS are 82% higher than those of women with normal ovaries. Previous studies have shown that pregnancy rates for women with PCOS undergoing IVF treatment are comparable with those of women with tubal factor infertility (Dor et al., 1990; Dale et al., 1991; Homburg et al., 1993). This is contrary to the findings of this study which suggests that the outcome of IVF treatment in women with isolated PCO morphology is better than that for women with normal ovaries, with a cumulative live birth rate of 66.5% compared with 46.9% for women with normal ovaries after three cycles of treatment. However, these previous studies included only women with PCOS (Dor et al., 1990; Urman et al., 1992; Homburg et al., 1993) or a heterogeneous group of women which included both PCO and PCOS (MacDougall et al., 1993). In addition, a variety of ovarian stimulation protocols were used in previous studies, where gonadotrophins were either used alone (Dor et al., 1990) or with, or without, GnRH agonist therapy (Urman et al., 1992; MacDougall et al., 1993). In this study, we have included only women with sonographic evidence of PCO who did not have any clinical manifestation of PCOS. Moreover, all the women used the same stimulation protocol, which involved the use of GnRH agonist to achieve pituitary desensitization in combination with gonadotrophins. It is essential to distinguish between women with PCO and those with PCOS, as they represent different ends of the spectrum of this heterogeneous disorder, and may in fact have different outcomes to IVF treatment. 169

4 L.Engmann et al. Table IV. Cumulative conception and live birth rates in women with PCO and women with normal ovaries Group Cycle No. of Pregnancies Live births % Cumulative % Cumulative no. women a (1st) (n) conception rate live birth rate (95% CI) (95% CI) PCO ( ) 35.2 ( ) ( ) 46.7 ( ) ( ) 66.5 ( ) Normal ovaries ( ) 18.6 ( ) ( ) 32.9 ( ) ( ) 46.9 ( ) a For conception analyses, the number of women undertaking cycles leading to first pregnancy were: PCO 46, 28, 20; normal ovaries 145, 104, 71. A greater number of cycles were undertaken for live birth analyses due to miscarriage of first ( second) pregnancy. One exception was first cycle of live birth analysis for women with normal ovaries, where one woman was excluded because of unknown pregnancy outcome. The response to ovarian stimulation of this group of women is consistent with previously published data on women with PCOS (Homburg et al., 1993) and women with isolated PCO morphology undergoing ovum donation (Wong et al., 1995). Women with PCO required fewer ampoules of gonadotrophins for ovarian stimulation and produced more follicles and viable oocytes than women with normal ovaries. The incidence of OHSS in this study and in the PCO group is similar to that reported in other studies (MacDougall et al., 1993). Our findings also suggest that women with PCO who are on the long protocol of GnRH agonist therapy have similar fertilization and miscarriage rates to those who have normal ovaries. This is consistent with the findings of studies involving patients with isolated PCO morphology undergoing ovum donation (Wong et al., 1995) and patients with PCOS using a combination of gonadotrophins and GnRH agonist therapy (Homburg et al., 1993a). However, women with PCOS undergoing IVF using gonadotrophins without GnRH agonist therapy have been found to have lower fertilization rates (Dor et al., 1990; Homburg et al., 1993) and higher miscarriage rates (Homburg et al., 1993a) compared with women with tubal factor infertility. Hypersecretion of luteinizing hormone (LH) has been implicated as the cause of reduced fertilization and high miscarriage rates in women with PCOS (Howles et al., 1986; Homburg et al., 1993a) and it has been suggested that normalization of LH concentrations with GnRH agonist therapy improves the fertilization rate (Homburg et al., 1993) and reduces miscarriage rate (Homburg et al., 1993a). Balen et al. (1993) reported previously that the miscarriage rate was higher in women with PCOS than in women with normal ovaries when treated with ovarian stimulation protocols without the GnRH agonist therapy. However, they found that women with PCOS had comparable miscarriage rates with women with normal ovaries when treated with the long protocol of GnRH agonist therapy, and they postulated that this was due to the normalization of LH concentrations when the long protocol was used. The reason why women with isolated PCO morphology undergoing IVF treatment with coexistent causes of infertility perform better than women with normal ovaries is probably because they produce more oocytes, but of comparable quality and fertilization rates. Consequently, there is a wider choice of embryos to select for transfer in these women, thereby resulting in a higher chance of conception. A previous study 170 on the in-vitro development and metabolic activity of preimplantation embryos in patients with PCOS undergoing IVF treatment, using a combination of gonadotrophins and GnRH agonist therapy, showed that these women had embryos with less fragmentation which cleaved faster, cavitated earlier and had more cells at the blastocyst stage than embryos from women with tubal disease (Hardy et al., 1995). Furthermore, recent studies using colour Doppler ultrasound have shown that women with PCO have a higher ovarian stromal blood flow velocity prior to commencement of gonadotrophin therapy than women with normal ovaries (Zaidi et al., 1995; Engmann et al., 1999a). It has also been shown that women with PCO have higher serum vascular endothelial growth factor concentrations in the early follicular phase compared with those of women with normal ovaries, which may explain the increased vascularity in women with PCO (Agrawal et al., 1998). Doppler studies have also shown that ovarian stromal blood flow velocity measured before commencing gonadotrophin stimulation is predictive of good ovarian response (Zaidi et al., 1996; Engmann et al., 1999b) and successful outcome of IVF treatment (Engmann et al., 1999b). Ovarian blood flow plays an important role during ovulation, and animal studies have suggested that increased follicular vascularity may be a primary determinant of follicular dominance and that dominant follicles have an increased uptake of serum gonadotrophins (McNatty et al., 1981; Zeleznik et al., 1981). Increased ovarian stromal blood flow velocity may therefore be associated with an increased delivery of gonadotrophins to the target cells for stimulation of follicular growth resulting in the production of more oocytes. There are several implications of this study. The finding on transvaginal ultrasound of polycystic ovarian morphology in women who do not have clinical manifestations of PCOS and who are undergoing IVF treatment because of other coexistent infertility factors is a favourable prognostic factor. This group of women have a very good, if not better, chance of achieving pregnancy and live birth after IVF treatment using the long GnRH agonist protocol, and this is important for counselling of these patients. They may therefore be favourable candidates for oocyte donation (Wong et al., 1995), although this has to be approached with caution because of the possible hereditary nature of the condition (Franks, 1995). Since these women exhibit an exaggerated response to gonadotrophin therapy, the

5 Outcome of IVF in women with PCO major problem remains the increased risk of OHSS. It is therefore important that every woman undergoing IVF has a baseline ultrasound scan to assess ovarian morphology before she commences treatment. Acknowledgements The authors are grateful to the medical, nursing and scientific staff of the London Women s Clinic who were involved in the care of the patients studied. References Adams, J., Franks, S., Polson, D.W. et al (1985) Multifollicular ovaries: clinical and endocrine features and response to pulsatile gonadotrophin releasing hormone. Lancet, ii, Agrawal, R., Sladkevecius, P., Engmann, L. et al. (1998) Serum vascular endothelial growth factor concentrations and ovarian stromal blood flow are increased in women with polycystic ovaries. Hum. Reprod., 13, Altman, D.G. (1991) Practical Statistics for Medical Research. Chapman & Hall, London. Balen, A.H., Tan, S.L., MacDougall, J. and Jacobs, H.S. (1993) Miscarriage rates following in-vitro fertilization are increased in women with polycystic ovaries and reduced by pituitary desensitization with buserelin. Hum. Reprod., 8, Breslow, N.E. and Day, N.E. (1980) Statistical Methods In Cancer Research, Volume 1. The Analysis Of Case-Control Studies. IARC, Lyon. Cox, D.R. (1972) Regression models and life tables (with discussion). J. R.. Statist. Soc. B., 34, D Agostino, R.B., Lee, M., Belanger, A.J. et al. (1990) Relation of pooled logistic regression to time dependent Cox regression analysis: the Framingham heart study. Statist. Med., 9, Dale, P.O., Tanbo, T., Abyholm, T. (1991) In-vitro fertilization in infertile women with the polycystic ovarian syndrome. Hum. Reprod., 6, Dor J., Shulman, A., Levran, D. et al. (1990) The treatment of patients with polycystic ovarian syndrome by in-vitro fertilization and embryo transfer: a comparison of results with those of patients with tubal infertility. Hum. Reprod., 5, Draper, N., Smith, H. (1981) Applied Regression Analysis. Second edition. John Wiley & Sons, New York. Engmann, L., Shaker, A., White, E. et al. (1998) A prospective randomized study to assess the clinical efficacy of gonadotrophins administered subcutaneously and intramuscularly. Hum. Reprod., 13, Engmann, L., Sladkevicius, P., Agrawal, R. et al. (1999a) The pattern of changes in ovarian stromal and uterine artery blood flow velocities during IVF treatment and its relationship with outcome of the cycle. Ultrasound Obstet. Gynecol., in press. Engmann, L., Sladkevicius, P., Agrawal, R. et al. (1999b) The value of ovarian stromal blood flow velocity measurement after pituitary suppression in the prediction of ovarian responsiveness and outcome of IVF treatment. Fert. Steril., in press. Franks, S. (1995) Polycystic ovary syndrome. N. Eng. J. Med., 333, Hardy, K., Robinson, F.M., Paraschos T. et al. (1995) Normal development and metabolic activity of preimplantation embryos in vitro from patients with polycystic ovaries. Hum. Reprod., 10, Homburg, R., Berkowitz, D., Levy, T. et al. (1993) In vitro fertilization and embryo transfer for the treatment of infertility associated with polycystic ovary syndrome. Fertil. Steril., 60, Homburg, R., Levy, T., Berkovitz, D. et al. (1993a) Gonadotropin-releasing hormone agonist reduces the miscarriage rate for pregnancies achieved in women with polycystic ovarian syndrome. Fertil. Steril., 59, Howles, C.M., Macanamee, M.C., Edwards, R.G. et al. (1986) Effect of high tonic levels of luteinising hormone on outcome of in-vitro fertilisation. Lancet, ii, Huber, P.J. (1967) The behaviour of maximum likelihood estimates under non-standard conditions. In Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability. Berkeley, CA, University of California Press, Los Angeles, 1, Jacobs, H.S. (1987) Polycystic ovaries and polycystic ovary syndrome. Gynecol. Endocrinol., 1, Kalbfleisch J.D. and Prentice, R.L. (1980) The Statistical Analysis Of Failure Time Data. John Wiley & Sons, New York. MacDougall, M.J., Tan, S.L., Balen, A. and Jacobs, H.S. (1993) A controlled study comparing patients with or without polycystic ovaries undergoing invitro fertilization. Hum. Reprod., 8, McNatty, K.P., Dobson, C., Gibb, M. et al. (1981) Accumulation of luteinising hormone, oestradiol and androstenedione by sheep ovarian follicles in vivo. J. Endocrinol., 91, Rogers, W.H. (1993) Sg 17: Regression standard errors in clustered samples. Stata Tech. Bull., 13, Schenker, J.S. (1993) Prevention and treatment of ovarian hyperstimulation. Hum. Reprod., 8, Statacorp (1996) Stata Statistical Software; Release 5.0. Stata Corporation, Texas, USA. Tan, S.L., Kingsland, C., Campbell, S. et al. (1992) The long protocol of administration of GnRHa is superior to the short protocol of ovarian stimulation for in-vitro fertilisation. Fertil. Steril., 57, Urman, B., Fluker, M.R., Yuen, B.H. et al. (1992) The outcome of in vitro fertilization and embryo transfer in women with polycystic ovary syndrome failing to conceive after ovulation induction with exogenous gonadotropins. Fertil. Steril., 57, Wong, I.L., Morris, R.S., Lobo, R.A. et al. (1995) Isolated polycystic morphology in ovum donors predicts response to ovarian stimulation. Hum Reprod., 10, Zaidi, J., Barber, J., Kyei-Mensah, A. et al. (1996) Relationship of ovarian stromal blood flow at baseline ultrasound to subsequent follicular response in an in vitro fertilisation program. Obstet. Gynecol., 88, Zaidi, J., Campbell, S., Pittroff, R. et al. (1995) Ovarian stromal blood flow in women with polycystic ovaries a possible new marker for diagnosis? Hum. Reprod., 10, Zeleznik, A.J., Sciler, H.M. and Reichert, L.E. (1981) Gonadotropin-binding sites in the Rhesus monkey ovary: role of the vasculature in the selective distribution of human chorionic gonadotropin to the preovulatory follicle. Endocrinology, 109, Received on May 15, 1998; accepted on September 30,