- Email: [email protected]
GnRHa flare and IVF pregnancy rates
International Journal of Gynecology and Obstetrics 85 (2004) 36–39
Article
GnRHa flare and IVF pregnancy rates E. Confino*, X. Zhang, R.R. Kazer Section of Reproductive Endocrinology and Infertility, Department of OByGYN, Northwestern University Medical School, Chicago, IL, USA Received 24 December 2002; received in revised form 15 July 2003; accepted 30 July 2003
Abstract Objectives: GnRH agonist administered early in the menstrual cycle (flare) causes an endogenous discharge of FSH and LH. Flare has been used in conjunction with gonadotropin ovarian stimulation for IVF ‘poor responders’. There is an ongoing controversy regarding whether flare protocols improve pregnancy rates in ‘poor responders’. The current study was designed to compare a GnRHa flare protocol with long suppression GnRHa IVF in ‘poor responders’. Methods: Seventy-three newly diagnosed poor responders who failed long GnRHa suppression IVF attempts were compared retrospectively with 128 age-matched IVF patients previously known poor ovarian responders treated with a long GnRHa suppression protocol. ‘Poor responders’ consisted of patients with peak E2 less than 1000 pgyml andyor less than five mature follicles with diameter )15 mm on the day of hCG administration. Student’s ttest was used to analyze the data and the chi-squared test was used to compare fertilization and pregnancy rates. Results: The flare protocol produced higher peak E2 levels (1647"747 vs. 720"258 mIUyml, P-0.05) and a larger number of mature follicles (5.8"2.2 vs. 4.0"1.0 P-0.05) in the study vs. the control group. A 30% pregnancy rate was achieved during this second IVF attempt using GnRHa flare protocol in the study group vs. 37 in the control group (P)0.05, NS). Conclusions: A comparison between the flare protocol group and the age-matched control group of poor ovarian responders subject to down regulation protocol, revealed higher peak E2 levels and more mature follicles, respectively. However, both groups yielded comparable pregnancy rates. The use of high dose gonadotropin treatment in our study groups seems to be the only explanation for their subsequent successful outcome. We concluded that GnRH agonist flare protocol does not result in better IVF outcome compared with long GnRH agonist suppression protocol in IVF poor responders. 䊚 2003 International Federation of Gynecology and Obstetrics. Published by Elsevier Science Ireland Ltd. All rights reserved. Keywords: Flare; High dose gonadotropin treatment; IVF
1. Introduction ‘Poor responders’ are a subset of IVF patients *Corresponding author. Tel.: q1-312-926-8244; fax: q1312-695-4924. E-mail address: [email protected] (E. Confino).
with low follicular yield when subjected to gonadotropin stimulation. Poor responders yield fewer oocytes and lower quality embryos compared with normal responders w1,2x. Many publications use different criteria to define poor responders w3,4x. It is believed that the
0020-7292/04/$30.00 䊚 2003 International Federation of Gynecology and Obstetrics. Published by Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0020-7292(03)00344-8
E. Confino et al. / International Journal of Gynecology and Obstetrics 85 (2004) 36–39
underlying pathophysiology of this condition is a smaller FSH independent pool of recruited oocytes, as well as increased incidence of subcellular damage to the cytoplasm and nucleus of these oocytes. The association with aging, infection, and endometriosis suggests that poor responders constitute a population of patients with a pre-existing genetic tendency to this condition, andyor environmental damage to their oocytes w1,4x. Multiple attempts at maximizing oocyte recruitment have been universally unsatisfactory. The use of high dose gonadotropins to stimulate the ovary was disappointing in some patients w5,6x. The use of high dose gonadotropins to stimulate the ovaries is very costly and clinical evidence did not support its universal application in all poor responders. In extreme cases some authors resorted to in vitro fertilization following natural cycles in poor responders w7x. The use of GnRH agonists in a flare protocol was extensively used to achieve maximum ovarian stimulation during the early FSH-dependent oocyte recruitment stage. Alternative protocols of step-down GnRH agonists and the additional use of human growth hormone did not receive wide acceptance w8x. More recently, the use of GnRH antagonists has been suggested as the preferred ovarian stimulation protocol in normal and poor responders w9x. The purpose of this study was to compare a newly diagnosed group of ‘poor responders’ during their first IVF attempt with a known group of poor responders. Both groups were subjected to high dose gonadotropin treatments. The study group was subjected to GnRHa flare and the control group was subjected to long GnRHa suppression. 2. Materials and methods The study was designed as a retrospective casecontrolled study performed on IVF patients from January 1999 and through July of 2001. All pregnancies were followed to term and their obstetric outcomes verified. IRB consent was waived because the data were obtained from an anonymous database reported to the CDC as described in the patient IVF consent form. Peak estradiol levels of less than 1000 pgyml andyor less than five mature follicles with a
37
Table 1 Average age, duration of infertility and pre IVF diagnosis in the study populationa
No. of patients Age Months infertile Diagnosis x Ovarian reserve Unexplained Male factor
GnRHa flare
Age-matched, non-flare of poor responders
73 37.3"4.1 27.5"26.6
128 34.7"3.8 30.1"15.9
32 31 10
28 76 24
a No statistically significant differences were found among the groups.
diameter of more than 15 mm on the day of hCG administration defined ‘poor responders’. Table 1 demonstrates the demographics of this patient population. Day 3 FSH levels were similar in both groups. The control group constituted of 128 IVF age-matched known poor responders treated with long GnRHa suppression. Both groups were similar in their demographics as depicted in Table 1. All patients received daily 450–600 units of rFSH, 10 000 units of hCG and progesterone in oil (25 mg twice a day) luteal support. The flare group received 0.1 mg leuprolide acetate daily injection during their FSH stimulation days. The control group was suppressed first using daily injections of leuprolide acetate. The mean number of FSH ampoules per patient was similar in both groups. 3. Results Table 2 demonstrates IVF results of the flare protocol poor responders group and the down regulation poor responders control group. The only statistically significant difference between the groups were higher peak E2 levels (1647"747 vs. 720"258, P-0.05) and a larger number of mature follicles (5.8"2.2 vs. 4.0"1.0, P-0.05) in the flare group. Both groups demonstrated comparable oocyte fertilization rates, a similar number of day 3 eight-cell embryos, and comparable clinical pregnancy and miscarriage rates. Improved oocyte number and corresponding higher means estradiol levels in the GnRHa poor responders flare group did not result in better
E. Confino et al. / International Journal of Gynecology and Obstetrics 85 (2004) 36–39
38
Table 2 IVF and pregnancy outcomes in the flare and GnRHa stimulation protocols in poor responders Flare
Age-matched, non-flare of poor responders
73 12.2"2.6 2277"1022 1647"747 5.8"2.2 9.4"5.0 55"28 40"30 3.1"1.2 30 21
128 12.5"3.4 2285"485 877"312* 3.8"2.2 6.7"5.0* 36"26 36"25 2.8"1.0 37 22
No. of cycles No. of days of stimulation Total dose of FSH Peak E2 No. of follicles )16 mm No. of eggs retrieved Fertilization rate (%) No. of 8-cells on Day 3 (%) No. of embryos transferred Clinical pregnancy rate (%) Miscarriage rate (%) *P-0.05.
pregnancy rates compared with age-matched long GnRHa suppressed poor responders. 4. Discussion The only irrefutable method to evaluate a stimulation protocol is the use of prospective randomized patient assignment to a treatment plan. During the randomization of IVF patients attention must be paid to adequate control for age, the underlying infertility diagnosis, variations in the amount of medications used, patient selection criteria such as comparable day 3 FSH levels w10x and other demographic variables. However, prospective study designs frequently result in unforeseen differences in one or more variables. A retrospective case control study has the advantage of producing a well-matched control group avoiding patients who mismatch the study group. We believe that our patient selection has been close to ideal. Both groups were identified as ‘poor responders’ based on standard criteria w11x. The study group was identified following previous IVF attempt in which a low number of follicles was recruited using recombinant FSH injection in a standard dose of 225 IU per day. The age-matched control group in this study was identified as a poor responder group during either previous IVF cycles performed elsewhere or during empirical therapy using gonadotropins and intrauterine inseminations. This selection method is not expected to create bias.
Careful analysis of the demographics of the study group and the control group (Table 1) did not reveal any statistically significant differences. The patients in both groups demonstrated comparable age, underlying diagnosis, and day 3 FSH levels. The only difference between the treatment group and the control group was the use of GnRH agonist flare in the study group vs. the use of GnRH agonist suppression in the control group. The only difference in patient selection for the study group is that they all failed for the first time their IVF attempt using long GnRHa suppression protocol. The control group of known poor responders were all diagnosed during previous failed IVF in a different program or during gonadotropinyIUI treatment prior to their conversion to IVF. It is therefore reasonable to believe that selection bias in this study is unlikely to exist. Elimination of GnRHa suppression during IVF resulted in a larger number of oocytes and corresponding higher estradiol levels in our study group compared with the long GnRHa suppression control group (Table 2). This result should intuitively result in better pregnancy rates. However, pregnancy rates were similar in the study and the control group. Miscarriage rates did not reveal any statistically significant difference in the study vs. the control group. These data suggest that while GnRH flare protocol may result in the recruitment, growth and retrieval of more oocytes, the protocol does not alter the most important IVF outcome: preg-
E. Confino et al. / International Journal of Gynecology and Obstetrics 85 (2004) 36–39
nancy rates and miscarriage rates. These results are plausible because the ultimate outcome of IVF is dependent largely on oocyte quality and not necessarily on oocyte number. The same principle is revealed when GnRH antagonist are used in poor responders to eliminate pituitary suppression during the late phases of follicular growth w12– 14x. Evaluating multiple strategies for improving ovarian response of poor responders during IVF seems to be unresolved at this point w15x. It seems that once a maximum dose of FSH is administered to patients, the only determinant of IVF success is the quality of the patient’s oocytes. Even though in our study oocyte number and peak estradiol levels reached statistical significance when the GnRHa flare protocol was evaluated compared with the GnRHa suppression protocol, this statistical significance did not translate into a biological significance reflected in higher pregnancy rates. We conclude that GnRHa flare protocol does not offer a clinically significant advantage in ‘poor responders’ in a patient population similar to ours.
w6 x
w7 x
w8 x
w9 x
w10x
w11x
w12x
References w1x Van Rysselberge M, Puissant F, Barlow P, et al. Fertility prognosis in IVF treatment of patients with cancelled cycles. Hum Reprod 1989;4:663 –666. w2x Jenkins JM, Davies DW, Devonport H, et al. Comparison of ‘poor’ responders with ‘good’ responders using a standard buserelinyhuman menopausal gonadotropin regime for in vitro fertilization. Hum Reprod 1991;6:918 –921. w3x Karande V, Gleicher N. A rational approach to the management of low responders in in vitro fertilization. Hum Reprod 1999;14:1744 –1748. w4x Pellicer A, Lightman A, Diamond MP, et al. Outcome of in vitro fertilization in women with low response to ovarian stimulation. Fertil Steril 1987;47:812 –815. w5x Van Hooff MH, Alberda AT, Huisman GJ, et al. Doubling the human menopausal gonadotropin dose in the course of an in vitro fertilization treatment cycle in low
w13x
w14x
w15x
39
responders: a randomized study. Hum Reprod 1993;8:369 –373. Karande VC, Jones GS, Veeck LL, et al. High-dose follicle-stimulating hormone stimulation at the onset of the menstrual cycle does not improve the in vitro fertilization outcome in low responder patients. Fertil Steril 1990;53:486 –494. Feldman B, Seidman DS, Levron J, Bider D, Shulman A, Shin S, Dor J. In vitro fertilization following natural cycles in poor responders. Gynecol Endocrinol 2001;15:328 –334. Dor J, Seidman DS, Amudai E, et al. Adjuvant growth hormone therapy in poor responders to in vitro fertilization: a prospective randomized placebo-controlled double-blind study. Hum Reprod 1995;10:40 –43. Akman MA, Erden HF, Tosun SB, et al. Addition of GnRH antagonist in cycles of poor responders undergoing IVF. Hum Reprod 2000;15:2145 –2147. Brown JR, Liu HC, Sewitch KF, et al. Variability of day 3 follicle-stimulating hormone levels in eumenorrheic women. J Reprod Med 1995;40:620 –624. Scott RT Jr., Hofmann GE, Oehninger S, et al. Intercycle variability of day 3 follicle stimulating hormone levels and its effect on stimulation quality in in vitro fertilization. Fertil Steril 1990;54:297 –302. Karande V, Morris R, Rinehart J, et al. Limited success using the ‘flare’ protocol in poor responders in cycles with low basal follicle-stimulating hormone levels during in vitro fertilization. Fertil Steril 1997;67:900 –903. Albano C, Smitz C, Tournaye H, et al. Luteal phase and clinical outcome after human menopausal gonadotrophinygonadotrophin releasing hormone antagonist treatment for ovarian stimulation in in vitro fertilizationy intracytoplasmic sperm injection cycles. Hum Reprod 1999;14:1426 –1430. Ganirelix Dose-Finding Study Group. A double-blind, randomized, dose-finding study to assess the efficacy of the gonadotrophin-releasing hormone antagonist Ganirelix (org 37462) to prevent premature luteinizing hormone surges in women undergoing ovarian stimulation with recombinant follicle stimulating hormone (Puregon). Hum Reprod 1998;13:3023 –3031. Surrey ES, Schoolcraft WB. Evaluating strategies for improving ovarian response of the poor responder undergoing assisted reproductive techniques. Fertil Steril 2001;75:226 –227.
Article
GnRHa flare and IVF pregnancy rates E. Confino*, X. Zhang, R.R. Kazer Section of Reproductive Endocrinology and Infertility, Department of OByGYN, Northwestern University Medical School, Chicago, IL, USA Received 24 December 2002; received in revised form 15 July 2003; accepted 30 July 2003
Abstract Objectives: GnRH agonist administered early in the menstrual cycle (flare) causes an endogenous discharge of FSH and LH. Flare has been used in conjunction with gonadotropin ovarian stimulation for IVF ‘poor responders’. There is an ongoing controversy regarding whether flare protocols improve pregnancy rates in ‘poor responders’. The current study was designed to compare a GnRHa flare protocol with long suppression GnRHa IVF in ‘poor responders’. Methods: Seventy-three newly diagnosed poor responders who failed long GnRHa suppression IVF attempts were compared retrospectively with 128 age-matched IVF patients previously known poor ovarian responders treated with a long GnRHa suppression protocol. ‘Poor responders’ consisted of patients with peak E2 less than 1000 pgyml andyor less than five mature follicles with diameter )15 mm on the day of hCG administration. Student’s ttest was used to analyze the data and the chi-squared test was used to compare fertilization and pregnancy rates. Results: The flare protocol produced higher peak E2 levels (1647"747 vs. 720"258 mIUyml, P-0.05) and a larger number of mature follicles (5.8"2.2 vs. 4.0"1.0 P-0.05) in the study vs. the control group. A 30% pregnancy rate was achieved during this second IVF attempt using GnRHa flare protocol in the study group vs. 37 in the control group (P)0.05, NS). Conclusions: A comparison between the flare protocol group and the age-matched control group of poor ovarian responders subject to down regulation protocol, revealed higher peak E2 levels and more mature follicles, respectively. However, both groups yielded comparable pregnancy rates. The use of high dose gonadotropin treatment in our study groups seems to be the only explanation for their subsequent successful outcome. We concluded that GnRH agonist flare protocol does not result in better IVF outcome compared with long GnRH agonist suppression protocol in IVF poor responders. 䊚 2003 International Federation of Gynecology and Obstetrics. Published by Elsevier Science Ireland Ltd. All rights reserved. Keywords: Flare; High dose gonadotropin treatment; IVF
1. Introduction ‘Poor responders’ are a subset of IVF patients *Corresponding author. Tel.: q1-312-926-8244; fax: q1312-695-4924. E-mail address: [email protected] (E. Confino).
with low follicular yield when subjected to gonadotropin stimulation. Poor responders yield fewer oocytes and lower quality embryos compared with normal responders w1,2x. Many publications use different criteria to define poor responders w3,4x. It is believed that the
0020-7292/04/$30.00 䊚 2003 International Federation of Gynecology and Obstetrics. Published by Elsevier Science Ireland Ltd. All rights reserved. doi:10.1016/S0020-7292(03)00344-8
E. Confino et al. / International Journal of Gynecology and Obstetrics 85 (2004) 36–39
underlying pathophysiology of this condition is a smaller FSH independent pool of recruited oocytes, as well as increased incidence of subcellular damage to the cytoplasm and nucleus of these oocytes. The association with aging, infection, and endometriosis suggests that poor responders constitute a population of patients with a pre-existing genetic tendency to this condition, andyor environmental damage to their oocytes w1,4x. Multiple attempts at maximizing oocyte recruitment have been universally unsatisfactory. The use of high dose gonadotropins to stimulate the ovary was disappointing in some patients w5,6x. The use of high dose gonadotropins to stimulate the ovaries is very costly and clinical evidence did not support its universal application in all poor responders. In extreme cases some authors resorted to in vitro fertilization following natural cycles in poor responders w7x. The use of GnRH agonists in a flare protocol was extensively used to achieve maximum ovarian stimulation during the early FSH-dependent oocyte recruitment stage. Alternative protocols of step-down GnRH agonists and the additional use of human growth hormone did not receive wide acceptance w8x. More recently, the use of GnRH antagonists has been suggested as the preferred ovarian stimulation protocol in normal and poor responders w9x. The purpose of this study was to compare a newly diagnosed group of ‘poor responders’ during their first IVF attempt with a known group of poor responders. Both groups were subjected to high dose gonadotropin treatments. The study group was subjected to GnRHa flare and the control group was subjected to long GnRHa suppression. 2. Materials and methods The study was designed as a retrospective casecontrolled study performed on IVF patients from January 1999 and through July of 2001. All pregnancies were followed to term and their obstetric outcomes verified. IRB consent was waived because the data were obtained from an anonymous database reported to the CDC as described in the patient IVF consent form. Peak estradiol levels of less than 1000 pgyml andyor less than five mature follicles with a
37
Table 1 Average age, duration of infertility and pre IVF diagnosis in the study populationa
No. of patients Age Months infertile Diagnosis x Ovarian reserve Unexplained Male factor
GnRHa flare
Age-matched, non-flare of poor responders
73 37.3"4.1 27.5"26.6
128 34.7"3.8 30.1"15.9
32 31 10
28 76 24
a No statistically significant differences were found among the groups.
diameter of more than 15 mm on the day of hCG administration defined ‘poor responders’. Table 1 demonstrates the demographics of this patient population. Day 3 FSH levels were similar in both groups. The control group constituted of 128 IVF age-matched known poor responders treated with long GnRHa suppression. Both groups were similar in their demographics as depicted in Table 1. All patients received daily 450–600 units of rFSH, 10 000 units of hCG and progesterone in oil (25 mg twice a day) luteal support. The flare group received 0.1 mg leuprolide acetate daily injection during their FSH stimulation days. The control group was suppressed first using daily injections of leuprolide acetate. The mean number of FSH ampoules per patient was similar in both groups. 3. Results Table 2 demonstrates IVF results of the flare protocol poor responders group and the down regulation poor responders control group. The only statistically significant difference between the groups were higher peak E2 levels (1647"747 vs. 720"258, P-0.05) and a larger number of mature follicles (5.8"2.2 vs. 4.0"1.0, P-0.05) in the flare group. Both groups demonstrated comparable oocyte fertilization rates, a similar number of day 3 eight-cell embryos, and comparable clinical pregnancy and miscarriage rates. Improved oocyte number and corresponding higher means estradiol levels in the GnRHa poor responders flare group did not result in better
E. Confino et al. / International Journal of Gynecology and Obstetrics 85 (2004) 36–39
38
Table 2 IVF and pregnancy outcomes in the flare and GnRHa stimulation protocols in poor responders Flare
Age-matched, non-flare of poor responders
73 12.2"2.6 2277"1022 1647"747 5.8"2.2 9.4"5.0 55"28 40"30 3.1"1.2 30 21
128 12.5"3.4 2285"485 877"312* 3.8"2.2 6.7"5.0* 36"26 36"25 2.8"1.0 37 22
No. of cycles No. of days of stimulation Total dose of FSH Peak E2 No. of follicles )16 mm No. of eggs retrieved Fertilization rate (%) No. of 8-cells on Day 3 (%) No. of embryos transferred Clinical pregnancy rate (%) Miscarriage rate (%) *P-0.05.
pregnancy rates compared with age-matched long GnRHa suppressed poor responders. 4. Discussion The only irrefutable method to evaluate a stimulation protocol is the use of prospective randomized patient assignment to a treatment plan. During the randomization of IVF patients attention must be paid to adequate control for age, the underlying infertility diagnosis, variations in the amount of medications used, patient selection criteria such as comparable day 3 FSH levels w10x and other demographic variables. However, prospective study designs frequently result in unforeseen differences in one or more variables. A retrospective case control study has the advantage of producing a well-matched control group avoiding patients who mismatch the study group. We believe that our patient selection has been close to ideal. Both groups were identified as ‘poor responders’ based on standard criteria w11x. The study group was identified following previous IVF attempt in which a low number of follicles was recruited using recombinant FSH injection in a standard dose of 225 IU per day. The age-matched control group in this study was identified as a poor responder group during either previous IVF cycles performed elsewhere or during empirical therapy using gonadotropins and intrauterine inseminations. This selection method is not expected to create bias.
Careful analysis of the demographics of the study group and the control group (Table 1) did not reveal any statistically significant differences. The patients in both groups demonstrated comparable age, underlying diagnosis, and day 3 FSH levels. The only difference between the treatment group and the control group was the use of GnRH agonist flare in the study group vs. the use of GnRH agonist suppression in the control group. The only difference in patient selection for the study group is that they all failed for the first time their IVF attempt using long GnRHa suppression protocol. The control group of known poor responders were all diagnosed during previous failed IVF in a different program or during gonadotropinyIUI treatment prior to their conversion to IVF. It is therefore reasonable to believe that selection bias in this study is unlikely to exist. Elimination of GnRHa suppression during IVF resulted in a larger number of oocytes and corresponding higher estradiol levels in our study group compared with the long GnRHa suppression control group (Table 2). This result should intuitively result in better pregnancy rates. However, pregnancy rates were similar in the study and the control group. Miscarriage rates did not reveal any statistically significant difference in the study vs. the control group. These data suggest that while GnRH flare protocol may result in the recruitment, growth and retrieval of more oocytes, the protocol does not alter the most important IVF outcome: preg-
E. Confino et al. / International Journal of Gynecology and Obstetrics 85 (2004) 36–39
nancy rates and miscarriage rates. These results are plausible because the ultimate outcome of IVF is dependent largely on oocyte quality and not necessarily on oocyte number. The same principle is revealed when GnRH antagonist are used in poor responders to eliminate pituitary suppression during the late phases of follicular growth w12– 14x. Evaluating multiple strategies for improving ovarian response of poor responders during IVF seems to be unresolved at this point w15x. It seems that once a maximum dose of FSH is administered to patients, the only determinant of IVF success is the quality of the patient’s oocytes. Even though in our study oocyte number and peak estradiol levels reached statistical significance when the GnRHa flare protocol was evaluated compared with the GnRHa suppression protocol, this statistical significance did not translate into a biological significance reflected in higher pregnancy rates. We conclude that GnRHa flare protocol does not offer a clinically significant advantage in ‘poor responders’ in a patient population similar to ours.
w6 x
w7 x
w8 x
w9 x
w10x
w11x
w12x
References w1x Van Rysselberge M, Puissant F, Barlow P, et al. Fertility prognosis in IVF treatment of patients with cancelled cycles. Hum Reprod 1989;4:663 –666. w2x Jenkins JM, Davies DW, Devonport H, et al. Comparison of ‘poor’ responders with ‘good’ responders using a standard buserelinyhuman menopausal gonadotropin regime for in vitro fertilization. Hum Reprod 1991;6:918 –921. w3x Karande V, Gleicher N. A rational approach to the management of low responders in in vitro fertilization. Hum Reprod 1999;14:1744 –1748. w4x Pellicer A, Lightman A, Diamond MP, et al. Outcome of in vitro fertilization in women with low response to ovarian stimulation. Fertil Steril 1987;47:812 –815. w5x Van Hooff MH, Alberda AT, Huisman GJ, et al. Doubling the human menopausal gonadotropin dose in the course of an in vitro fertilization treatment cycle in low
w13x
w14x
w15x
39
responders: a randomized study. Hum Reprod 1993;8:369 –373. Karande VC, Jones GS, Veeck LL, et al. High-dose follicle-stimulating hormone stimulation at the onset of the menstrual cycle does not improve the in vitro fertilization outcome in low responder patients. Fertil Steril 1990;53:486 –494. Feldman B, Seidman DS, Levron J, Bider D, Shulman A, Shin S, Dor J. In vitro fertilization following natural cycles in poor responders. Gynecol Endocrinol 2001;15:328 –334. Dor J, Seidman DS, Amudai E, et al. Adjuvant growth hormone therapy in poor responders to in vitro fertilization: a prospective randomized placebo-controlled double-blind study. Hum Reprod 1995;10:40 –43. Akman MA, Erden HF, Tosun SB, et al. Addition of GnRH antagonist in cycles of poor responders undergoing IVF. Hum Reprod 2000;15:2145 –2147. Brown JR, Liu HC, Sewitch KF, et al. Variability of day 3 follicle-stimulating hormone levels in eumenorrheic women. J Reprod Med 1995;40:620 –624. Scott RT Jr., Hofmann GE, Oehninger S, et al. Intercycle variability of day 3 follicle stimulating hormone levels and its effect on stimulation quality in in vitro fertilization. Fertil Steril 1990;54:297 –302. Karande V, Morris R, Rinehart J, et al. Limited success using the ‘flare’ protocol in poor responders in cycles with low basal follicle-stimulating hormone levels during in vitro fertilization. Fertil Steril 1997;67:900 –903. Albano C, Smitz C, Tournaye H, et al. Luteal phase and clinical outcome after human menopausal gonadotrophinygonadotrophin releasing hormone antagonist treatment for ovarian stimulation in in vitro fertilizationy intracytoplasmic sperm injection cycles. Hum Reprod 1999;14:1426 –1430. Ganirelix Dose-Finding Study Group. A double-blind, randomized, dose-finding study to assess the efficacy of the gonadotrophin-releasing hormone antagonist Ganirelix (org 37462) to prevent premature luteinizing hormone surges in women undergoing ovarian stimulation with recombinant follicle stimulating hormone (Puregon). Hum Reprod 1998;13:3023 –3031. Surrey ES, Schoolcraft WB. Evaluating strategies for improving ovarian response of the poor responder undergoing assisted reproductive techniques. Fertil Steril 2001;75:226 –227.