- Email: [email protected]
Different influence of incongruent follicular development on in vitro fertilization–embryo transfer and gamete intrafallopian transfer pregnancy rates
FERTILITY AND STERILITYt VOL. 70, NO. 6, DECEMBER 1998 Copyright ©1998 American Society for Reproductive Medicine Published by Elsevier Science Inc. Printed on acid-free paper in U.S.A.
Different influence of incongruent follicular development on in vitro fertilization– embryo transfer and gamete intrafallopian transfer pregnancy rates Talia Eldar-Geva, M.D., Ph.D.,*† Philip J. M. Lowe, M.B., Ch.B.,* Vivien MacLachlan, B.Sc.,* Luk Rombauts, M.D., Ph.D.,‡ and David L. Healy, Ph.D.*† Monash In Vitro Fertilization, and Monash University, Melbourne, Victoria, Australia
Received March 6, 1998; revised and accepted July 9, 1998. Presented in part at the XVI Annual Meeting of The Fertility Society of Australia, Adelaide, South Australia, December 2– 4, 1997. Reprint requests and present address: IVF Unit, Shaare-zedek Medical Center, P.O. Box 3235, Jerusalem Israel 91031 (FAX: 972-2-6540114; E-mail: [email protected]). * Monash IVF. † Department of Obstetrics and Gynecology, Monash University. ‡ Present address: Medical Center for Fertility Diagnostics and IVF-ET, Leuven, Belgium. 0015-0282/98/$19.00 PII S0015-0282(98)00338-0
Objective: To compare the influence of incongruent (asymmetric) follicular development on treatment outcome in IVF-ET and GIFT cycles. Design: A retrospective comparative study. Setting: Tertiary referral center for infertility. Patient(s): Five hundred forty-three consecutive assisted reproduction cycles (428 IVF-ET and 115 GIFT) in 422 infertile patients. Intervention(s): Controlled ovarian hyperstimulation (COH) and IVF-ET or GIFT. Main Outcome Measure(s): The incongruity ratio as a parameter of the asymmetry in follicular development and pregnancy rate (PR). Result(s): For GIFT cycles, the PRs were 37.8% and 15.7% in cycles with congruent and incongruent follicular development, respectively. However, for IVF-ET cycles, the PR was not affected by incongruent follicular development: 28.2% and 29.0%, respectively. An inverse relationship was observed between the degree of incongruity and the estimated probability of pregnancy in GIFT cycles but not in IVF-ET cycles. Neither the side of the dominant ovary nor the degree of incongruity were consistent in consecutive cycles. Conclusion(s): Incongruent follicular development during COH has a significantly negative influence on the outcome of GIFT cycles but not on the outcome of IVF-ET cycles. The reason for this difference is not clear. We recommend considering IVF-ET instead of GIFT if incongruent follicular development occurs. (Fertil Sterilt 1998;70:1039 – 43. ©1998 by American Society for Reproductive Medicine.) Key Words: GIFT, IVF-ET, follicular development, incongruity ratio, clinical pregnancy rate, probability of pregnancy
The distribution of developing follicles in both ovaries during controlled ovarian hyperstimulation (COH) for assisted reproductive technology (ART) is sometimes asymmetric or incongruent. Recently, we showed that incongruent follicular development during COH had a marked negative influence on the outcome of GIFT cycles (1). The mechanism responsible for this observation was not clear. Factors, such as the side of gamete transfer and oocyte number or oocyte quality, did not account for the observed differences. Because no other reason could be found, it was speculated that the receptivity of the endometrium might be impaired (1). If so, we would expect to find a similar negative influ-
ence of incongruent follicular development on the pregnancy rate (PR) of IVF-ET cycles. The aim of the present study was to evaluate the influence of incongruent follicular development on ART outcome in patients undergoing IVF-ET compared with patients undergoing GIFT treatment.
MATERIALS AND METHODS Patients Five hundred forty-three consecutive ART cycles in 422 patients treated at Monash IVF (Melbourne, Victoria, Australia) between January 1996 and April 1997 were retrospectively analyzed. Institutional review board approval 1039
was not required because the study included only a retrospective analysis of ART results. The inclusion criteria were the diagnosis of tubal, idiopathic or male factor infertility. Patients with polycystic ovarian disease, endometriosis, a history of ovarian surgery or absence of an ovary were not included. The indications for GIFT treatment were idiopathic (85%) or male factor infertility (15%). The indications for IVF-ET were idiopathic (27%), tubal (27%), male factor (35%), or a combination of male and tubal infertility (11%). Male factor infertility was defined as the presence of one of the following abnormalities or a combination of them: sperm count of ,20 3 106/mL; progressive motility of ,25%; abnormal forms of .85%; presence of sperm antibodies of .50%. If abnormalities were present, semen was considered suitable for GIFT if .1.2 3 106 motile sperm could be recovered after a Percoll density gradient centrifugation procedure. We analyzed treatment outcome in three groups of patients according to female diagnosis and type of treatment: idiopathic-GIFT, idiopathic-IVF, and tubal-IVF.
Statistics The last US before hCG was used to determine the side of the “dominant” ovary by comparing the summed follicular diameters (.10 mm) in both ovaries. To estimate the degree of discrepancy between both ovaries, the “incongruity ratio” was calculated, as previously described (1). Briefly, the sum of all follicular diameters of .10 mm in the dominant ovary divided by the sum of all follicular diameters of .10 mm in the contralateral ovary. We showed that the incongruity ratio remains virtually unchanged during the last 2 days of follicular development before administration of hCG (1). All patients in this study had their last US a maximum of 2 days before hCG administration. SAS software system version 6.12 (SAS Institute, Cary, NC) was used for statistical analysis of the results. We used the x2 test and Mann-Whitney U test where appropriate. Some results were generated from logistic regression, after statistical outliers were removed to improve the overall model fit. Unless stated otherwise, values are expressed as means 6 95% confidence intervals.
Methods Controlled ovarian hyperstimulation protocols used in our institute previously were described (2). Briefly, patients were down-regulated with 0.5 mg/d of the GnRH-agonist nafarelin (Synarel; Searle, Sydney, New South Wales, Australia), starting in the luteal phase (long protocol) or on day 2 of the follicular phase (flare or boost protocol). Multiple follicular development was initiated with use of individually adjusted doses of purified urinary FSH (Metrodin; Serono, Melbourne, Victoria, Australia). The ovarian response was monitored with serum E2 levels and transvaginal ultrasonography. Follicular diameters were taken as the mean of two perpendicular axes, one being the maximal diameter in that plane. Patients were given 5,000 IU of hCG IM (Profasi; Serono) when at least three follicles of .17 mm were present and the serum E2 level was appropriately rising. Oocytes were retrieved transvaginally 36 hours later under general anesthesia. In GIFT cycles, one to four oocytes were transferred laparoscopically to one tube together with 150,000 motile spermatozoa purified on a Percoll density gradient. Routine IVF or intracytoplasmic sperm injection (ICSI) and embryo culture were used as indicated, and one to four embryos were transferred 48 –72 hours after egg retrieval. Patients were given 2,000 IU of hCG IM 4 and 8 days after oocyte pick-up for luteal support. If ovarian hyperstimulation syndrome was anticipated, vaginal progesterone pessaries (200 mg two times per day) were given instead of hCG. Clinical pregnancies were confirmed with serial serum b-hCG measurements starting 16 days after oocyte pick-up and the detection of at least one gestational sac on vaginal ultrasound (US) at 6 weeks’ gestation. 1040 Eldar-Geva et al.
RESULTS The mean age of the patient population was 34.7 years (range, 22– 44 years). The median cycle number was 2 (range, 1–12). The clinical PR per transfer in this patient group was 27.7%. The patients who have had “congruent” follicular development, namely, in whom follicular development was similar in both ovaries (IR of ,1.5), were compared with those with “incongruent” follicular development, namely, patients with marked differences between both ovaries (IR of $1.5) (Fig. 1). Overall, incongruent follicular development occurred in 57% of the cycles. In GIFT cycles, incongruent follicular development was associated with a significantly lower clinical PR per transfer compared with congruent follicular development—15.7% and 37.8%, respectively (P,.01, x 2 test). However, in IVF-ET cycles incongruent follicular development had no influence on PR, whether the cause of infertility was idiopathic (clinical PRs were 31.9% for congruent and 32.2% for incongruent follicular development) or tubal (clinical PRs per transfer were 23.6% and 22.2%, respectively). Patients with idiopathic or male infertility with incongruent follicular development had significantly higher PR after IVF-ET treatment (37 of 116 [32.2%]) than after GIFT treatment (11 of 70 [15.7%]) (P,.01, x 2 test). In all groups the proportions of patients who had the flare versus the long down-regulation protocol were distributed equally. No difference between groups was observed for the number of days between the last US and the administration of hCG. Other factors known to influence PR were analyzed in Table 1. We found no statistically significant difference between congruent and incongruent follicular development
Incongruent follicular development in ART
Vol. 70, No. 6, December 1998
FIGURE 1 Clinical pregnancy rates per transfer for patients with congruent (incongruity ratio of ,1.5) and incongruent (incongruity ratio of $1.5) follicular development in the three groups according to female diagnosis and type of treatment. Numbers above the bars indicate number of transfers per group. h 5 congruent; ■ 5 incongruent. p P,.01 versus congruent in idiopathic-GIFT, x2 test.
in GIFT cycles in patient’s age, the number of previous treatment cycles, maximal serum E2 levels, number of eggs retrieved, FSH cumulative dose, number of eggs transferred, and the fertilization rate of supernumerary eggs. However, in IVF-ET cycles, both the idiopathic and the tubal infertility groups, incongruent follicular development was accompanied by significantly lower number of eggs retrieved (P,.05; Mann-Whitney U test). No statistically significant differences were found for any of the other parameters in Table 1. Using logistic regression models we estimated the probability of achieving a clinical pregnancy in an ART cycle by incongruity ratio (Fig. 2). For GIFT cycles, there was a highly significant negative influence of incongruity ratio on the estimated probability of pregnancy, (P 5 0.0001; x 2 test). However, in IVF cycles, changes in incongruity ratio were not accompanied by significant changes in the estimated probability of pregnancy for both tubal and idiopathic infertility. For example, for a patient having incongruity ratio of 3, the probability of pregnancy for GIFT treatment was 16%, compared with 32% for IVF-ET treatment. Combined analysis showed a statistically significant difference between the GIFT curve and both IVF curves (P,.01). There was no statistically significant difference between FERTILITY & STERILITYt
patients with congruent and incongruent follicular development, in each group separately, with regard to the abortion rates (for congruent and incongruent follicular development, respectively, 0 of 17 and 0 of 11 in patients undergoing GIFT treatment, 5 of 21 and 6 of 16 in patients undergoing IVF-ET for tubal factor infertility, and 11 of 48 and 7 of 37 in patients undergoing IVF-ET for idiopathic infertility. In all GIFT cycles, the transfer of gametes was limited to one uterine tube. We analyzed the effect of the side of gamete transfer on PR in cycles with incongruent follicular development. Of the 70 cycles in this group, gametes were transferred to the tube on the dominant side (ipsilateral) in 40 cycles and to the tube on the contralateral side in 30 cycles. The PRs for ipsilateral (6 of 40 [15.0%]) and contralateral (5 of 30 [16.7%]) GIFT were not statistically significantly different. To evaluate the consistency of ovarian response to hyperstimulation, we compared the incongruity ratios in consecutive cycles in 98 patients who had more than one cycle during the study period. In the first treatment cycle 54 patients had incongruent follicular development. Of these patients, 19 had a similar response to ovarian stimulation in the following cycle, namely, incongruent follicular development with dominance of the same ovary. Sixteen additional 1041
TABLE 1 Cycle characteristics for patients with congruent and incongruent follicular development. Type of cycle
GIFT for idiopathic infertility
Variable Age (y) No. of treatment cycles Maximum E2 levels (pg/mL) No. of eggs retrieved Cumulative dose of FSH (IU) No. of eggs/no. of embryos transferred Percentage of patients who underwent ICSI Fertilization rate (%)* Incongruity ratio
IVF-ET for tubal factor infertility
IVF-ET for idiopathic infertility
Congruent (n 5 45)
Incongruent (n 5 70)
Congruent (n 5 70)
Incongruent (n 5 90)
Congruent (n 5 117)
Incongruent (n 5 151)
33.7 6 1.1 2.2 6 0.6 1,844 6 307 9.4 6 1.2 22.7 6 4.2 2.6 6 0.2
34.1 6 1.1 2.1 6 0.5 1,652 6 240 8.0 6 1.2 25.0 6 3.5 2.4 6 0.2
49 6 13 1.2 6 0.04
50 6 9 5.5 6 2.7†
36.2 6 0.9 4.1 6 0.7 2,195 6 521 10.8 6 1.5 27.2 6 3.5 2.4 6 0.2 27.8 58 6 5 1.2 6 0.03
35.5 6 1.0 3.5 6 0.6 1,655 6 205 8.5 6 1.0 31.3 6 4.4 2.4 6 0.2 28.9 60 6 4 6.4 6 2.9†
34.7 6 1.2 2.5 6 0.5 1,738 6 181 10.2 6 1.1 29.0 6 3.2 2.4 6 0.1 56 56 6 5 1.2 6 0.03
34.0 6 0.9 2.8 6 0.5 1,624 6 134 8.2 6 0.8 30.0 6 3.3 2.4 6 0.1 57.7 56 6 5 8.5 6 3.5†
Note: Values are means 6 95% confidence intervals unless otherwise indicated. * For GIFT 5 fertilization rates of supernumerary eggs. † P,.05 (incongruent versus congruent).
patients had incongruent follicular development again but with dominance of the contralateral ovary. Nineteen patients had incongruent follicular development in their second ART cycle. Of the 44 patients who had congruent follicular development in the first cycle, 17 had a similar response in the consecutive cycle, and 27 had incongruent follicular development. Overall, only 36 women (36.7%) had the same congruent or incongruent response to ovarian stimulation in their two consecutive ART cycles.
influence of incongruent follicular development during COH is not clear. One speculation was that dysfunction of the ovary in the nondominant side (expressed by an inferior response to COH) may be accompanied by dysfunction of the tube in the same side. In our previous study we found that GIFT to the uterine tube ipsilateral or contralateral to the dominant ovary yielded comparable PRs. However, cycles with congruent and incongruent follicular development were analyzed to-
DISCUSSION FIGURE 2
The aim of our study was to explore the mechanism responsible for the marked adverse effect of incongruent follicular development on GIFT PR, as demonstrated in our previous study (1). In that study we found that factors, such as side of gamete transfer, oocyte number, and oocyte quality, did not account for the observed influence. Because no other reason could be found, it was speculated that the receptivity of the endometrium might be impaired. We reasoned that if incongruent follicular development had its negative influence on PR through impairing endometrial receptivity, the same negative influence should be demonstrated in IVF-ET cycles.
Estimated probability of pregnancy by incongruity ratio for the three groups of patients, using logistic regression models, after statistical outliners were removed to improve the overall model fit. ■ 5 idiopathic-IVF; D 5 tubal-IVF; E 5 idiopathic-GIFT. Combined analysis showed a statistically significant difference between the GIFT curve and either IVF curves (P,.01).
It is surprising that we found that incongruent follicular development during COH had no influence on the PR of IVF-ET cycles despite being accompanied by a significantly lower number of eggs retrieved (Table 1). Nevertheless, in the present study, we confirmed in a larger number of cycles that incongruent follicular development had a marked negative influence on the outcome of GIFT cycles. The reason for this discordance between GIFT and IVF-ET about the 1042 Eldar-Geva et al.
Incongruent follicular development in ART
Vol. 70, No. 6, December 1998
gether (1). In the present study we further analyzed the effect of the side of GIFT on PR in the 70 cycles with incongruent follicular development. Our results confirmed that there was no benefit to transferring to the tube ipsilateral to the dominant ovary and hence rejected the above hypothesis. Furthermore, incongruent follicular development during COH was not infrequent (57% in the present study; 56% in our previous study). Tubal pathology was not associated with a higher incidence of incongruent follicular development compared with idiopathic infertility (Table 1). Neither the side of the dominant ovary nor the type of ovarian response (congruent or incongruent) was consistent in the subsequent cycle. Moreover, incongruent follicular development during COH for IVF-ET, even with a very high degree of incongruity between both ovaries (IR .5), was not associated with a decreased probability of pregnancy (Fig. 2). However, because incongruent follicular development in GIFT cycles tended to occur in older patients who took higher cumulative doses of FSH but had lower maximal serum E2 concentrations and lower number of eggs retrieved (Table 1), we cannot exclude the possibility that asymmetrical distribution of growing follicles is a manifestation of poor follicular development during COH. The same inclination were presented in our previous study (1). The ability to select the best embryos for transfer in IVF-ET cycles may “compensate” for other adverse effects. Unfortunately, in this retrospective study, day 3 FSH levels were not applicable for most of the patients; hence, we could not investigate this question further. The exact mechanism responsible for the discrepancy between the influence of incongruent follicular development on GIFT and IVF-ET outcomes remains unclear. Gamete intrafallopian transfer is still an important and successful treatment for patients with idiopathic or mild male factor infertility (3–5). The cumulative pregnancy and live birth rates after three GIFT cycles in a nonselected large group of patients (1,628 women) treated in our program were 49.6% and 38.8%, respectively (3). This compared favorably with similar studies on cumulative pregnancy rate (30.5%– 38.7%) (6 –9) and live birth rate (28.3%) (9) after IVF-ET. No large prospective randomized trial comparing the results of GIFT with IVF-ET for this group of patients has ever been reported in the literature. The implication from our study is that the incongruity
FERTILITY & STERILITYt
ratio is of clinical value in assessing the patient’s probability of pregnancy. We recommend considering changing the treatment from GIFT to IVF-ET if incongruent follicular development occurs during COH for ART. If the patients insist on GIFT treatment, they should consider the possibility of canceling their treatment cycle and try again in subsequent cycles. In summary, incongruent follicular development during COH had a significant negative influence on the outcome of GIFT cycles but not on the outcome of IVF-ET cycles. The reason for this difference is not clear. The side of the dominant ovary and the degree of incongruity were not repeated in consecutive cycles. We recommend considering IVF-ET instead of GIFT if incongruent follicular development occurs.
Acknowledgments: The authors thank all the clinical, nursing, embryology, and administrative staff of Monash IVF, and Monash US for women, Epworth Programme, for their invaluable help. They also thank P. McCloud, Ph.D., and G. Simmons, M.Sc. from the Department of Mathematics, Monash University, for statistical analysis of the data.
References 1. Rombauts L, Webster D, Wood CE, Healy DL. Incongruent follicular development reduces gamete intrafallopian transfer pregnancy rate. Fertil Steril 1996;66:987–90. 2. MacLachlan V, Besanko M, O’Shea F, Wade H, Wood C, Trounson A, et al. A controlled study of luteinizing hormone releasing hormone agonist (buserelin) for the induction of folliculogenesis before in vitro fertilization. N Engl J Med 1989;320:1233–7. 3. Rombauts L, Dear M, Breheny S, Healy DL. Cumulative pregnancy and live birth rates after gamete intra-fallopian transfer. Hum Reprod 1997; 12:1338 – 42. 4. Tournaye H, Camus M, Ubaldi F, Clasen K, Van Steirtegham A, Devroe´y P. Is there still an important role for tubal transfer procedures? Hum Reprod 1996;11:1815– 8. 5. Bulletti C. Debating tubal transfer in assisted reproductive technologies. Hum Reprod 1996;11:1820 –2. 6. Dor J, Seidman DS, Ben-Shlomo I, Levran D, Ben-Rafael Z, Mashiach S. Cumulative pregnancy rates following in-vitro fertilization: the significance of age and infertility aetiology. Hum Reprod 1996;11:425– 8. 7. Alsalili M, Yuzpe A, Tummon I, Parker J, Martin J, Daniel S, et al. Cumulative pregnancy rates and pregnancy outcome after in-vitro fertilization: .5000 cycles at one centre. Hum Reprod 1995;10:470 – 4. 8. Haan G, Bernardus RE, Hollanders HMG, Leerentveld BA, Park FK, Naaktgeboren N. Selective drop-out in successive in-vitro fertilization attempts: the pendulum danger. Hum Reprod 1991;6:939 – 43. 9. Tan SL, Royston P, Campbell S, Jacobs HS, Betts J, Mason B, et al. Cumulative conception and live birth rates after in-vitro fertilisation. Lancet 1992;339:1390 – 4.
1043
Different influence of incongruent follicular development on in vitro fertilization– embryo transfer and gamete intrafallopian transfer pregnancy rates Talia Eldar-Geva, M.D., Ph.D.,*† Philip J. M. Lowe, M.B., Ch.B.,* Vivien MacLachlan, B.Sc.,* Luk Rombauts, M.D., Ph.D.,‡ and David L. Healy, Ph.D.*† Monash In Vitro Fertilization, and Monash University, Melbourne, Victoria, Australia
Received March 6, 1998; revised and accepted July 9, 1998. Presented in part at the XVI Annual Meeting of The Fertility Society of Australia, Adelaide, South Australia, December 2– 4, 1997. Reprint requests and present address: IVF Unit, Shaare-zedek Medical Center, P.O. Box 3235, Jerusalem Israel 91031 (FAX: 972-2-6540114; E-mail: [email protected]). * Monash IVF. † Department of Obstetrics and Gynecology, Monash University. ‡ Present address: Medical Center for Fertility Diagnostics and IVF-ET, Leuven, Belgium. 0015-0282/98/$19.00 PII S0015-0282(98)00338-0
Objective: To compare the influence of incongruent (asymmetric) follicular development on treatment outcome in IVF-ET and GIFT cycles. Design: A retrospective comparative study. Setting: Tertiary referral center for infertility. Patient(s): Five hundred forty-three consecutive assisted reproduction cycles (428 IVF-ET and 115 GIFT) in 422 infertile patients. Intervention(s): Controlled ovarian hyperstimulation (COH) and IVF-ET or GIFT. Main Outcome Measure(s): The incongruity ratio as a parameter of the asymmetry in follicular development and pregnancy rate (PR). Result(s): For GIFT cycles, the PRs were 37.8% and 15.7% in cycles with congruent and incongruent follicular development, respectively. However, for IVF-ET cycles, the PR was not affected by incongruent follicular development: 28.2% and 29.0%, respectively. An inverse relationship was observed between the degree of incongruity and the estimated probability of pregnancy in GIFT cycles but not in IVF-ET cycles. Neither the side of the dominant ovary nor the degree of incongruity were consistent in consecutive cycles. Conclusion(s): Incongruent follicular development during COH has a significantly negative influence on the outcome of GIFT cycles but not on the outcome of IVF-ET cycles. The reason for this difference is not clear. We recommend considering IVF-ET instead of GIFT if incongruent follicular development occurs. (Fertil Sterilt 1998;70:1039 – 43. ©1998 by American Society for Reproductive Medicine.) Key Words: GIFT, IVF-ET, follicular development, incongruity ratio, clinical pregnancy rate, probability of pregnancy
The distribution of developing follicles in both ovaries during controlled ovarian hyperstimulation (COH) for assisted reproductive technology (ART) is sometimes asymmetric or incongruent. Recently, we showed that incongruent follicular development during COH had a marked negative influence on the outcome of GIFT cycles (1). The mechanism responsible for this observation was not clear. Factors, such as the side of gamete transfer and oocyte number or oocyte quality, did not account for the observed differences. Because no other reason could be found, it was speculated that the receptivity of the endometrium might be impaired (1). If so, we would expect to find a similar negative influ-
ence of incongruent follicular development on the pregnancy rate (PR) of IVF-ET cycles. The aim of the present study was to evaluate the influence of incongruent follicular development on ART outcome in patients undergoing IVF-ET compared with patients undergoing GIFT treatment.
MATERIALS AND METHODS Patients Five hundred forty-three consecutive ART cycles in 422 patients treated at Monash IVF (Melbourne, Victoria, Australia) between January 1996 and April 1997 were retrospectively analyzed. Institutional review board approval 1039
was not required because the study included only a retrospective analysis of ART results. The inclusion criteria were the diagnosis of tubal, idiopathic or male factor infertility. Patients with polycystic ovarian disease, endometriosis, a history of ovarian surgery or absence of an ovary were not included. The indications for GIFT treatment were idiopathic (85%) or male factor infertility (15%). The indications for IVF-ET were idiopathic (27%), tubal (27%), male factor (35%), or a combination of male and tubal infertility (11%). Male factor infertility was defined as the presence of one of the following abnormalities or a combination of them: sperm count of ,20 3 106/mL; progressive motility of ,25%; abnormal forms of .85%; presence of sperm antibodies of .50%. If abnormalities were present, semen was considered suitable for GIFT if .1.2 3 106 motile sperm could be recovered after a Percoll density gradient centrifugation procedure. We analyzed treatment outcome in three groups of patients according to female diagnosis and type of treatment: idiopathic-GIFT, idiopathic-IVF, and tubal-IVF.
Statistics The last US before hCG was used to determine the side of the “dominant” ovary by comparing the summed follicular diameters (.10 mm) in both ovaries. To estimate the degree of discrepancy between both ovaries, the “incongruity ratio” was calculated, as previously described (1). Briefly, the sum of all follicular diameters of .10 mm in the dominant ovary divided by the sum of all follicular diameters of .10 mm in the contralateral ovary. We showed that the incongruity ratio remains virtually unchanged during the last 2 days of follicular development before administration of hCG (1). All patients in this study had their last US a maximum of 2 days before hCG administration. SAS software system version 6.12 (SAS Institute, Cary, NC) was used for statistical analysis of the results. We used the x2 test and Mann-Whitney U test where appropriate. Some results were generated from logistic regression, after statistical outliers were removed to improve the overall model fit. Unless stated otherwise, values are expressed as means 6 95% confidence intervals.
Methods Controlled ovarian hyperstimulation protocols used in our institute previously were described (2). Briefly, patients were down-regulated with 0.5 mg/d of the GnRH-agonist nafarelin (Synarel; Searle, Sydney, New South Wales, Australia), starting in the luteal phase (long protocol) or on day 2 of the follicular phase (flare or boost protocol). Multiple follicular development was initiated with use of individually adjusted doses of purified urinary FSH (Metrodin; Serono, Melbourne, Victoria, Australia). The ovarian response was monitored with serum E2 levels and transvaginal ultrasonography. Follicular diameters were taken as the mean of two perpendicular axes, one being the maximal diameter in that plane. Patients were given 5,000 IU of hCG IM (Profasi; Serono) when at least three follicles of .17 mm were present and the serum E2 level was appropriately rising. Oocytes were retrieved transvaginally 36 hours later under general anesthesia. In GIFT cycles, one to four oocytes were transferred laparoscopically to one tube together with 150,000 motile spermatozoa purified on a Percoll density gradient. Routine IVF or intracytoplasmic sperm injection (ICSI) and embryo culture were used as indicated, and one to four embryos were transferred 48 –72 hours after egg retrieval. Patients were given 2,000 IU of hCG IM 4 and 8 days after oocyte pick-up for luteal support. If ovarian hyperstimulation syndrome was anticipated, vaginal progesterone pessaries (200 mg two times per day) were given instead of hCG. Clinical pregnancies were confirmed with serial serum b-hCG measurements starting 16 days after oocyte pick-up and the detection of at least one gestational sac on vaginal ultrasound (US) at 6 weeks’ gestation. 1040 Eldar-Geva et al.
RESULTS The mean age of the patient population was 34.7 years (range, 22– 44 years). The median cycle number was 2 (range, 1–12). The clinical PR per transfer in this patient group was 27.7%. The patients who have had “congruent” follicular development, namely, in whom follicular development was similar in both ovaries (IR of ,1.5), were compared with those with “incongruent” follicular development, namely, patients with marked differences between both ovaries (IR of $1.5) (Fig. 1). Overall, incongruent follicular development occurred in 57% of the cycles. In GIFT cycles, incongruent follicular development was associated with a significantly lower clinical PR per transfer compared with congruent follicular development—15.7% and 37.8%, respectively (P,.01, x 2 test). However, in IVF-ET cycles incongruent follicular development had no influence on PR, whether the cause of infertility was idiopathic (clinical PRs were 31.9% for congruent and 32.2% for incongruent follicular development) or tubal (clinical PRs per transfer were 23.6% and 22.2%, respectively). Patients with idiopathic or male infertility with incongruent follicular development had significantly higher PR after IVF-ET treatment (37 of 116 [32.2%]) than after GIFT treatment (11 of 70 [15.7%]) (P,.01, x 2 test). In all groups the proportions of patients who had the flare versus the long down-regulation protocol were distributed equally. No difference between groups was observed for the number of days between the last US and the administration of hCG. Other factors known to influence PR were analyzed in Table 1. We found no statistically significant difference between congruent and incongruent follicular development
Incongruent follicular development in ART
Vol. 70, No. 6, December 1998
FIGURE 1 Clinical pregnancy rates per transfer for patients with congruent (incongruity ratio of ,1.5) and incongruent (incongruity ratio of $1.5) follicular development in the three groups according to female diagnosis and type of treatment. Numbers above the bars indicate number of transfers per group. h 5 congruent; ■ 5 incongruent. p P,.01 versus congruent in idiopathic-GIFT, x2 test.
in GIFT cycles in patient’s age, the number of previous treatment cycles, maximal serum E2 levels, number of eggs retrieved, FSH cumulative dose, number of eggs transferred, and the fertilization rate of supernumerary eggs. However, in IVF-ET cycles, both the idiopathic and the tubal infertility groups, incongruent follicular development was accompanied by significantly lower number of eggs retrieved (P,.05; Mann-Whitney U test). No statistically significant differences were found for any of the other parameters in Table 1. Using logistic regression models we estimated the probability of achieving a clinical pregnancy in an ART cycle by incongruity ratio (Fig. 2). For GIFT cycles, there was a highly significant negative influence of incongruity ratio on the estimated probability of pregnancy, (P 5 0.0001; x 2 test). However, in IVF cycles, changes in incongruity ratio were not accompanied by significant changes in the estimated probability of pregnancy for both tubal and idiopathic infertility. For example, for a patient having incongruity ratio of 3, the probability of pregnancy for GIFT treatment was 16%, compared with 32% for IVF-ET treatment. Combined analysis showed a statistically significant difference between the GIFT curve and both IVF curves (P,.01). There was no statistically significant difference between FERTILITY & STERILITYt
patients with congruent and incongruent follicular development, in each group separately, with regard to the abortion rates (for congruent and incongruent follicular development, respectively, 0 of 17 and 0 of 11 in patients undergoing GIFT treatment, 5 of 21 and 6 of 16 in patients undergoing IVF-ET for tubal factor infertility, and 11 of 48 and 7 of 37 in patients undergoing IVF-ET for idiopathic infertility. In all GIFT cycles, the transfer of gametes was limited to one uterine tube. We analyzed the effect of the side of gamete transfer on PR in cycles with incongruent follicular development. Of the 70 cycles in this group, gametes were transferred to the tube on the dominant side (ipsilateral) in 40 cycles and to the tube on the contralateral side in 30 cycles. The PRs for ipsilateral (6 of 40 [15.0%]) and contralateral (5 of 30 [16.7%]) GIFT were not statistically significantly different. To evaluate the consistency of ovarian response to hyperstimulation, we compared the incongruity ratios in consecutive cycles in 98 patients who had more than one cycle during the study period. In the first treatment cycle 54 patients had incongruent follicular development. Of these patients, 19 had a similar response to ovarian stimulation in the following cycle, namely, incongruent follicular development with dominance of the same ovary. Sixteen additional 1041
TABLE 1 Cycle characteristics for patients with congruent and incongruent follicular development. Type of cycle
GIFT for idiopathic infertility
Variable Age (y) No. of treatment cycles Maximum E2 levels (pg/mL) No. of eggs retrieved Cumulative dose of FSH (IU) No. of eggs/no. of embryos transferred Percentage of patients who underwent ICSI Fertilization rate (%)* Incongruity ratio
IVF-ET for tubal factor infertility
IVF-ET for idiopathic infertility
Congruent (n 5 45)
Incongruent (n 5 70)
Congruent (n 5 70)
Incongruent (n 5 90)
Congruent (n 5 117)
Incongruent (n 5 151)
33.7 6 1.1 2.2 6 0.6 1,844 6 307 9.4 6 1.2 22.7 6 4.2 2.6 6 0.2
34.1 6 1.1 2.1 6 0.5 1,652 6 240 8.0 6 1.2 25.0 6 3.5 2.4 6 0.2
49 6 13 1.2 6 0.04
50 6 9 5.5 6 2.7†
36.2 6 0.9 4.1 6 0.7 2,195 6 521 10.8 6 1.5 27.2 6 3.5 2.4 6 0.2 27.8 58 6 5 1.2 6 0.03
35.5 6 1.0 3.5 6 0.6 1,655 6 205 8.5 6 1.0 31.3 6 4.4 2.4 6 0.2 28.9 60 6 4 6.4 6 2.9†
34.7 6 1.2 2.5 6 0.5 1,738 6 181 10.2 6 1.1 29.0 6 3.2 2.4 6 0.1 56 56 6 5 1.2 6 0.03
34.0 6 0.9 2.8 6 0.5 1,624 6 134 8.2 6 0.8 30.0 6 3.3 2.4 6 0.1 57.7 56 6 5 8.5 6 3.5†
Note: Values are means 6 95% confidence intervals unless otherwise indicated. * For GIFT 5 fertilization rates of supernumerary eggs. † P,.05 (incongruent versus congruent).
patients had incongruent follicular development again but with dominance of the contralateral ovary. Nineteen patients had incongruent follicular development in their second ART cycle. Of the 44 patients who had congruent follicular development in the first cycle, 17 had a similar response in the consecutive cycle, and 27 had incongruent follicular development. Overall, only 36 women (36.7%) had the same congruent or incongruent response to ovarian stimulation in their two consecutive ART cycles.
influence of incongruent follicular development during COH is not clear. One speculation was that dysfunction of the ovary in the nondominant side (expressed by an inferior response to COH) may be accompanied by dysfunction of the tube in the same side. In our previous study we found that GIFT to the uterine tube ipsilateral or contralateral to the dominant ovary yielded comparable PRs. However, cycles with congruent and incongruent follicular development were analyzed to-
DISCUSSION FIGURE 2
The aim of our study was to explore the mechanism responsible for the marked adverse effect of incongruent follicular development on GIFT PR, as demonstrated in our previous study (1). In that study we found that factors, such as side of gamete transfer, oocyte number, and oocyte quality, did not account for the observed influence. Because no other reason could be found, it was speculated that the receptivity of the endometrium might be impaired. We reasoned that if incongruent follicular development had its negative influence on PR through impairing endometrial receptivity, the same negative influence should be demonstrated in IVF-ET cycles.
Estimated probability of pregnancy by incongruity ratio for the three groups of patients, using logistic regression models, after statistical outliners were removed to improve the overall model fit. ■ 5 idiopathic-IVF; D 5 tubal-IVF; E 5 idiopathic-GIFT. Combined analysis showed a statistically significant difference between the GIFT curve and either IVF curves (P,.01).
It is surprising that we found that incongruent follicular development during COH had no influence on the PR of IVF-ET cycles despite being accompanied by a significantly lower number of eggs retrieved (Table 1). Nevertheless, in the present study, we confirmed in a larger number of cycles that incongruent follicular development had a marked negative influence on the outcome of GIFT cycles. The reason for this discordance between GIFT and IVF-ET about the 1042 Eldar-Geva et al.
Incongruent follicular development in ART
Vol. 70, No. 6, December 1998
gether (1). In the present study we further analyzed the effect of the side of GIFT on PR in the 70 cycles with incongruent follicular development. Our results confirmed that there was no benefit to transferring to the tube ipsilateral to the dominant ovary and hence rejected the above hypothesis. Furthermore, incongruent follicular development during COH was not infrequent (57% in the present study; 56% in our previous study). Tubal pathology was not associated with a higher incidence of incongruent follicular development compared with idiopathic infertility (Table 1). Neither the side of the dominant ovary nor the type of ovarian response (congruent or incongruent) was consistent in the subsequent cycle. Moreover, incongruent follicular development during COH for IVF-ET, even with a very high degree of incongruity between both ovaries (IR .5), was not associated with a decreased probability of pregnancy (Fig. 2). However, because incongruent follicular development in GIFT cycles tended to occur in older patients who took higher cumulative doses of FSH but had lower maximal serum E2 concentrations and lower number of eggs retrieved (Table 1), we cannot exclude the possibility that asymmetrical distribution of growing follicles is a manifestation of poor follicular development during COH. The same inclination were presented in our previous study (1). The ability to select the best embryos for transfer in IVF-ET cycles may “compensate” for other adverse effects. Unfortunately, in this retrospective study, day 3 FSH levels were not applicable for most of the patients; hence, we could not investigate this question further. The exact mechanism responsible for the discrepancy between the influence of incongruent follicular development on GIFT and IVF-ET outcomes remains unclear. Gamete intrafallopian transfer is still an important and successful treatment for patients with idiopathic or mild male factor infertility (3–5). The cumulative pregnancy and live birth rates after three GIFT cycles in a nonselected large group of patients (1,628 women) treated in our program were 49.6% and 38.8%, respectively (3). This compared favorably with similar studies on cumulative pregnancy rate (30.5%– 38.7%) (6 –9) and live birth rate (28.3%) (9) after IVF-ET. No large prospective randomized trial comparing the results of GIFT with IVF-ET for this group of patients has ever been reported in the literature. The implication from our study is that the incongruity
FERTILITY & STERILITYt
ratio is of clinical value in assessing the patient’s probability of pregnancy. We recommend considering changing the treatment from GIFT to IVF-ET if incongruent follicular development occurs during COH for ART. If the patients insist on GIFT treatment, they should consider the possibility of canceling their treatment cycle and try again in subsequent cycles. In summary, incongruent follicular development during COH had a significant negative influence on the outcome of GIFT cycles but not on the outcome of IVF-ET cycles. The reason for this difference is not clear. The side of the dominant ovary and the degree of incongruity were not repeated in consecutive cycles. We recommend considering IVF-ET instead of GIFT if incongruent follicular development occurs.
Acknowledgments: The authors thank all the clinical, nursing, embryology, and administrative staff of Monash IVF, and Monash US for women, Epworth Programme, for their invaluable help. They also thank P. McCloud, Ph.D., and G. Simmons, M.Sc. from the Department of Mathematics, Monash University, for statistical analysis of the data.
References 1. Rombauts L, Webster D, Wood CE, Healy DL. Incongruent follicular development reduces gamete intrafallopian transfer pregnancy rate. Fertil Steril 1996;66:987–90. 2. MacLachlan V, Besanko M, O’Shea F, Wade H, Wood C, Trounson A, et al. A controlled study of luteinizing hormone releasing hormone agonist (buserelin) for the induction of folliculogenesis before in vitro fertilization. N Engl J Med 1989;320:1233–7. 3. Rombauts L, Dear M, Breheny S, Healy DL. Cumulative pregnancy and live birth rates after gamete intra-fallopian transfer. Hum Reprod 1997; 12:1338 – 42. 4. Tournaye H, Camus M, Ubaldi F, Clasen K, Van Steirtegham A, Devroe´y P. Is there still an important role for tubal transfer procedures? Hum Reprod 1996;11:1815– 8. 5. Bulletti C. Debating tubal transfer in assisted reproductive technologies. Hum Reprod 1996;11:1820 –2. 6. Dor J, Seidman DS, Ben-Shlomo I, Levran D, Ben-Rafael Z, Mashiach S. Cumulative pregnancy rates following in-vitro fertilization: the significance of age and infertility aetiology. Hum Reprod 1996;11:425– 8. 7. Alsalili M, Yuzpe A, Tummon I, Parker J, Martin J, Daniel S, et al. Cumulative pregnancy rates and pregnancy outcome after in-vitro fertilization: .5000 cycles at one centre. Hum Reprod 1995;10:470 – 4. 8. Haan G, Bernardus RE, Hollanders HMG, Leerentveld BA, Park FK, Naaktgeboren N. Selective drop-out in successive in-vitro fertilization attempts: the pendulum danger. Hum Reprod 1991;6:939 – 43. 9. Tan SL, Royston P, Campbell S, Jacobs HS, Betts J, Mason B, et al. Cumulative conception and live birth rates after in-vitro fertilisation. Lancet 1992;339:1390 – 4.
1043