Serial uterine artery Doppler velocity parameters and human uterine receptivity in IVF/ICSI cycles

Ultrasound Obstet Gynecol 2008; 31: 432 438 Published online 4 February 2008 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/uog.5179 Serial uterine artery Doppler velocity parameters and human uterine receptivity in IVF/ICSI cycles D.A.HOOZEMANS,R.SCHATS,N.B.LAMBALK,R.HOMBURGandP.G.A.HOMPES VU University Medical Center Amsterdam, IVF Center, Amsterdam, The Netherlands KEYWORDS: embryo implantation; ICSI; IVF; serial Doppler ultrasound ABSTRACT Objective To evaluate the predictive value of serial uterine artery Doppler ultrasound for embryo implantation in in-vitro fertilization (IVF)-intracytoplasmic sperm injection (ICSI) cycles. Methods This was a prospective observational study at the VU University Medical Center, Amsterdam. Patients with an indication for IVF or IVF-ICSI according to departmental protocol underwent controlled ovarian hyperstimulation followed by IVF or IVF-ICSI and embryo transfer and had serial Doppler ultrasound performed during this treatment cycle. Patient and cycle characteristics, number of conceptions and ongoing pregnancies and pulsatility index (PI) of both uterine arteries on different cycle days were assessed and results were compared between patients who conceived and those who did not. Results Of the 102 patients enrolled into the study, 83 underwent embryo transfer. Of these, 41 became pregnant and 42 did not (Group 1). Of the 41 pregnancies, 30 were ongoing (Group 2) and 11 miscarried (Group 3). Between Groups 1, 2 and 3, linear regression revealed no significant difference between any of the variables examined except in the quality of transferred embryos. There was no significant difference in the mean PI of the left and right uterine arteries on any day of the cycle, or in the change in PI during the cycle. Receiver operating characteristics curves derived to determine the performance of PI to predict pregnancy outcome supported our findings that the uterine artery PI is not a suitable marker for identifying patients with implantation failure. Multivariate analysis showed no relationship between pregnancy and PI between groups, but it did show a relationship between pregnancy and some patient and cycle characteristics. Conclusion In an unselected group of patients undergoing IVF or IVF-ICSI and embryo transfer, serial Doppler ultrasound examination of the uterine artery does not discriminate between cycles resulting in ongoing pregnany, miscarriage and no pregnancy. Copyright 2008 ISUOG. Published by John Wiley & Sons, Ltd. INTRODUCTION The major limiting factor for success in in-vitro fertilization (IVF)/intracytoplasmic sperm injection (ICSI) treatment cycles is embryo implantation. Despite developments in laboratory techniques and ovarian stimulation protocols, the implantation rate per embryo transferred remains approximately 15% 1. Embryo implantation depends on two parallel and probably interacting processes: embryonic development, which seems most crucial, resulting in an embryo equipped for implantation; and endometrial development, leading to uterine receptivity. Uterine receptivity depends on many biochemical and structural changes in the endometrium during the course of the menstrual cycle 2. Research on such factors has led to the identification of a window of implantation (5 7 days postovulatory) 3,4 outside which the uterus is refractory even to an otherwise normal embryo that has implantation potential. Identification of the window of implantation and its qualitative assessment in the individual patient has major clinical implications. Suitable assessment of uterine receptivity could be achieved through ultrasound because of its noninvasive nature. Various sonographic parameters, such as endometrial echogenicity, uterine contractility and Doppler parameters, have been evaluated as markers of receptivity 5. Doppler ultrasound seems the most popular technique among clinicians because of its reproducibility. Uterine blood flow can be assessed by Doppler ultrasound in the uterine arteries and there is evidence of an association between abnormal uterine artery blood velocities Correspondence to: Dr D. A. Hoozemans, VU University Medical Center Amsterdam, IVF Center, Poli Zuid, PO BOX 7057, 1007 MB Amsterdam, The Netherlands (e-mail: da.hoozemans@vumc.nl) Accepted: 7 June 2007 Copyright 2008 ISUOG. Published by John Wiley & Sons, Ltd. ORIGINAL PAPER

Serial uterine artery flow velocities and IVF-ICSI 433 and infertility 5 7 ; the pulsatility index (PI) is considered best at reflecting uterine artery blood velocity parameters. Using uterine artery PI as a marker of implantation has been found successful by some authors while it has been rejected by others 5 21. Modern Doppler ultrasound techniques, with high-frequency transvaginal probes and color imaging, have far more acceptable intra- and interobserver variation than were possible in the early days of ultrasound 22. The cyclical vascular changes 2 of the endometrium, with angiogenesis and vasodilatation occurring in the implantation window, suggest a time relationship with embryo implantation. Further support for this relationship derives from the observation that known histological and biochemical 12 markers of receptivity correlate with Doppler findings 23. The window of implantation is characterized by a state of increased vascular permeability, massive endometrial angiogenesis 24 and decreased vascular resistance 2. Uterine perfusion is modulated by ovarian hormones; estrogens reduce the vascular impedance both in spontaneous cycles 6,8 and in estrogen and progesterone replacement cycles 9,10. Progesterone increases this effect towards the mid-luteal phase (i.e. the window of implantation), when vascular resistance is minimal and uterine perfusion is maximal 6,11,25. Various other vascular factors, such as prostaglandins and vascular endothelial growth factor, are involved in this monthly vascular endometrial development and their absence leads to subfertility 2. The ultimate proof of this concept of the relationship between vascular changes and successful implantation is the establishment of pregnancy in relation to uterine perfusion. Many authors have addressed the subject of uterine perfusion and its correlation to receptivity. Some found a positive correlation 13 18, while others did not 19,20. However, no previous publication has used serial measurements to assess true physiological changes during a treatment cycle. This seems imperative because the range of PI values between patients, which could lead to overlapping normal and abnormal values, is well known. Previous studies have found that cut-off PI-values for implantation failure range from 2.5 21 to 3.3 15,18.Also, the day of assessment does not seem to be of major importance for the predictive value 23. Therefore, in order to study receptivity sonographically, the patient herself should be used as a reference. In this study we focused on changes in PI in the individual patient during IVF to determine whether serial PI assessments in an IVF or IVF-ICSI cycle reflect a process of endometrial vascular development that predicts the possibility of successful implantation. We hypothesized that a decrease in PI would be observed during the IVF or IVF-ICSI cycle and that this phenomenon would be more pronounced in cycles resulting in pregnancy. To identify confounding factors, such as embryo quality, number of embryos transferred, maternal age, indication for IVF and the stimulation protocol, which might affect the chance of pregnancy directly or indirectly by influencing PI, we used multivariate analysis. METHODS Inclusion and treatment Between March 2002 and December 2003, a total of 102 patients who had an indication for IVF or IVF-ICSI treatment according to the criteria of the division of reproductive medicine of the VU University Medical Center consented to participate in this study. Indications included: idiopathic infertility, either IVF or IVF-ICSI male factor sub fertility (sperm count > 1 million for IVF and < 1 million for IVF-ICSI), and external endometriosis Grade I or II. Inclusion criteria were: age < 39 years, regular menstrual cycle with two normal, nonpolycystic ovaries present at the time of enrollment, and a follicle stimulating hormone (FSH) level < 10 IU/L on cycle day 3. Patients with tubal factor infertility (after pelvic inflammatory disease or with hydrosalpinges), a history of uterine surgery and/or apparent endometrial pathology (polyps, submucous myoma or synechia) were excluded, as were patients with clinically relevant systemic diseases, such as diabetes mellitus, ulcerative colitis, Crohn s disease, connective tissue diseases or hypertension. Smoking and a body mass index > 28 kg/m 2 were also exclusion criteria. Each patient was included only once. Patients were stimulated according to our standard long agonist protocol, starting with 30 µg ethinylestradiol and 150 µg levonorgestrel (Microgynon 30, Schering, Mijdrecht, The Netherlands) administered orally on cycle day 3 and continued until cycle day 24. From cycle day 17 we started treatment with 0.1 mg triptorelin-acetate (Decapeptyl, Ferring, Hoofddorp, The Netherlands) daily until the day of ovum pick-up. After withdrawal bleeding started, controlled ovarian hyperstimulation was commenced from cycle day 3 with recombinant FSH (follitrophin beta, Puregon, Organon, OSS, The Netherlands) in an individually adjusted dose that depended on age and antral follicle count. Patients were monitored with regular transvaginal ultrasound examinations and urinary hcg 10 000 IU/ampoule (Profasi, Serono, Denhaag, The Netherlands) was administered to induce final follicular maturation according to protocol 26. Transvaginal ultrasound-guided follicular aspiration was performed 36 h after hcg administration and the number of oocytes was assessed under a microscope. Fertilization was achieved by either IVF or ICSI and the number and of embryos was noted. Embryo quality was assessed by experienced IVF laboratory technicians using a score based on the number of blastomeres; an overall score of 1 4 was given based on the scoring system described by Roseboom et al. 26 and Steer et al. 27. At 2 3 days after ovum pick-up, a maximum of two embryos was placed into the uterine cavity. The luteal phase was supported by 100 mg Progestan (Organon) three times daily until a pregnancy test was done 14 16 days after ovum pick-up. Pregnancy was the primary endpoint. Clinical pregnancy was defined by serum hcg > 50 IU/L on day 14 16 post follicle aspiration, increased hcg levels 1 week after initial assessment and the appearance of a gestational

434 Hoozemans et al. sac on ultrasound. Ongoing pregnancy was defined by positive fetal cardiac activity at 12 weeks gestation. Ultrasound Transvaginal sonographic measurements were performed by two observers (R.S. and D.A.H.). Uterine artery Doppler flow velocity measurements were carried out with the patient in a standard lithotomy position, using an Aloka SSD 4000 Prosound Pure HD ultrasound machine (Aloka, Tokyo, Japan) equipped with a 7.5- MHz transvaginal probe. The observations were made on specific cycle days as indicated in Table 1 and were carried out on mornings to avoid fluctuations due to the normal circadian rhythm of uterine blood flow. We assessed the PI after validation of the inter- and intraobserver variation (intraclass correlation coefficient (ICC), 0.92 (95% CI, 0 0.94) for intraobserver variation and ICC, 0.95 (95%CI, 3 0.96) for interobserver variation). The mean PI of left and right uterine arteries was used for analysis. Statistical analysis Univariate analysis was performed on the PI on the different assessment days, comparing absolute PI values in relation to pregnancy outcome. Univariate analysis was also used to relate the changes in PI values in the different outcome groups. Multivariate analysis (logistic regression model) was used to investigate the influence of possible confounders on the chances of pregnancy, using PI, maternal age, endometrial thickness on day of hcg injection, treatment (IVF/ICSI), IVF indication, primary or secondary infertility and total embryo quality as explanatory variables and pregnancy as the primary outcome variable. The power calculation indicated a sample size of 100 (alpha 5, estimated averages, 2.00 for pregnant and 3.00 for non-pregnant and estimated 75% non-pregnant and 25% pregnant). The test characteristics of uterine artery PI on different cycle days for embryo implantation were expressed by receiver operating characteristics (ROC) curves and the areas under the curves (AUC). The study was conducted according to the declaration of Helsinki revised in 1983 and in accordance with the research guidelines of our institute. The study was approved by the institute s review board and patients gave their written informed consent. RESULTS One hundred and two patients were enrolled into the study. Seventeen patient cycles were cancelled because of low response, prevention of ovarian hyperstimulation syndrome or fertilization failure. One patient withdrew her informed consent before starting ovarian hyperstimulation. One patient had to be excluded from the study due to violation of the protocol (it was later discovered that she smoked). Thus, 83 patients underwent embryo transfer of at least one embryo. Of these, 42 patients had a negative pregnancy test 15 days after ovum pick-up (Group 1), and 41 patients conceived. In the 41 pregnant patients there were 30 ongoing pregnancies, determined by ultrasound at 12 weeks gestation (Group 2), and 11 miscarriages/biochemical pregnancies (Group 3). The miscarriages, of which there were four, occurred at 7, 7, 8 and 9 weeks. Because of their low number they were not evaluated separately. It is possible, however, that these had a different etiology from the biochemical pregnancy losses, miscarriages being caused probably by embryonic factors (chromosomal abnormalities) and biochemical pregnancy losses having a more endometrial-driven etiology. Table 2 shows the patient characteristics of the participants and their treatment cycle characteristics. There were no significant differences between pregnant and non-pregnant patients with respect to age, cause of infertility, primary or secondary infertility or number of previous IVF attempts. The multivariate model revealed a correlation between pregnancy and: age (inverse correlation) and number and quality of embryos (positive correlation) and subfertility factor (data not shown). The individual treatment cycle characteristics did not yield discriminating factors for successful implantation. The only significant factor found was the embryo quality as assessed by light microscopy. In all patient groups there was a decrease in PI towards the mid-luteal phase, indicating a decrease in resistance of the uterine circulation, which would possibly allow better perfusion of the uterus (Table 3). However, there was an increase in PI at 15 days after ovum pick-up in Table 1 Timing of pulsatility index measurements during in-vitro fertilization/intracytoplasmic sperm injection treatment cycle, with corresponding physiological stages Phase of treatment cycle Cycle day (mean (range)) Event Start down-regulation cycle 3 (1 4) Start (basal value) Day of hcg administration 12 (10 14) Maximum E2 levels, pre-lh surge Day of ovum pick-up 14 (12 16) Ovulation, start luteal phase Day of embryo transfer 17 (14 19) Embryo enters uterus and process of hatching is initiated Ovulation day + 7 9 24 (20 25) Onset of nidation and trophoblast invasion Day of pregnancy test 29 (27 33) Implantation complete E2, estradiol; hcg, human chorionic gonadotropin; LH, luteinizing hormone.

Serial uterine artery flow velocities and IVF-ICSI 435 Table 2 Patient and in-vitro fertilization (IVF)-intracytoplasmic sperm injection (ICSI) treatment cycle characteristics according to study group Group Variable 1: not pregnant (n = 42) 2: ongoing pregnancy (n = 30) 3: biochemical pregnancy (n = 7)/ miscarriage (n = 4) Age (years) 33.5 ± 3.6 33.6 ± 3.3 33.1 ± 4.0 Idiopathic factor 13 (31) 8 (27) 3 (27) Male factor IVF 16 (38) 12 (40) 1 (9) Male factor ICSI 12 (29) 8 (27) 5 (45) Endometriosis Grade I or II 1 (2) 2 (7) 2 (18) Primary infertility 32 (76) 20 (67) 8 (73) Secondary infertility 10 (24) 10 (33) 3 (27) Number of previous cycles 4 ± 0.97 0.77 ± 0.90 1.45 ± 1.64 Total FSH dose received (IU) 2428 ± 720 2214 ± 517 2455 ± 643 Number of stimulation days 12.2 ± 2.14 12.17 ± 1.68 11.60± 1.35 Number of oocytes retrieved 13.67 ± 6.38 14.93 ± 6.48 13.45 ± 6.87 Number of embryos formed 7.45 ± 4.16 8.03 ± 4.58 6.80 ± 5.11 Score of transfer embryos 4.36 ± 1.14* 5.1 ± 1* 4.3 ± 1.62* Timing of embryo transfer: 2 days after ovum pick-up 7 (17) 3 (10) 1 (9) 3 days after ovum pick-up 35 (83) 27 (90) 10 (91) Endometrial thickness (mm) 9.95 ± 1.78 10.10 ± 1.99 10.14 ± 6 Values are mean ± SD or n (%). ANOVA or Chi-square tests were used as applicable. *P < 5. FSH, follicle stimulating hormone. Table 3 Mean (SD) pulsatility index (PI) values in different study groups undergoing in-vitro fertilization/intracytoplasmic sperm injection Group Timing of PI measurement 1: not pregnant (n = 42) 2: ongoing pregnancy (n = 30) 3: biochemical pregnancy/ miscarriage (n = 11) P Start down-regulation cycle 2.18 (7) 2.32 (0.50) 2.18 (8) 0.51 Day of hcg administration 2.11 (7) 2.19 (0.56) 2.15 (0.38) 0.78 Day of ovum pick-up 2.08 (0.38) 2.13 (0.38) 2.15 (1) 2 Day of embryo transfer 1.84 (0) 2.04 (0.59) 1.98 (0.37) 0 Ovulation day + 7 9 1.72 (3) 1.75 (1) 1.73 (0.36) 0.94 Day of pregnancy test 1.89 (7) 1.85 (0) 1.87 (0.50) 0.94 hcg, human chorionic gonadotropin. pregnant as well as in non-pregnant patients. There was no significant difference between the groups in the change (decrease) in PI from the start of the cycle towards the day of embryo transfer and beyond (Table 4). Table 5 presents the results of the multivariate logistic regression analysis. Odds ratios for PI were not significant, indicating that the amount of decrease in PI did not discriminate between cycles that resulted in pregnancy and cycles that did not. This was regardless of the other variables in the model. ROC curves on different cycle days indicated the poor test characteristics (sensitivity and specificity) of PI as an implantation marker (Figure 1), with areas under the curve of between 1 and 0.53. DISCUSSION The success of human embryo implantation depends on maternal and embryonic factors and their interactions. To assess uterine receptivity one must take various factors into account. The main factor determining success, identified by various studies (reviewed by Hoozemans et al. 2 ), is probably embryo quality as determined by light microscopy. Our results confirmed this. We tried to increase the sensitivity of PI for implantation by performing serial measurements in the same patient, to assess true physiological changes involved in the endometrial vascular development cycle. Because vascular development is imperative for a receptive endometrium, its assessment by Doppler ultrasound might be predictive for implantation. We must now reject this hypothesis because there were no significant differences in PI on any of the cycle days and in the change in PI through the treatment cycle. We observed a wide range of PI values that were distributed almost equally between pregnant patients, non-pregnant patients and patients with a failed pregnancy (Tables 3 and 4). It is remarkable that in our

436 Hoozemans et al. Table 4 Mean (SD) changes in pulsatility index (PI) values between cycle days in different study groups undergoing in-vitro fertilization/intracytoplasmic sperm injection Group Period over which PI changed 1: not pregnant (n = 42) 2: ongoing pregnancy (n = 30) 3: biochemical pregnancy/ miscarriage (n = 11) P Baseline: start of down-regulation cycle 2.18 (7) 2.32 (0.50) 2.18 (8) 0.51 To day of hcg administration 8 (0.52) 0.14 (9) 4 (0.57) 4 To day of ovum pick-up 0.11 (7) 0 (7) 4 (0.57) 8 To day of embryo transfer 0.35 (0.51) 0.30 (0.78) 0.31 (0.70) 2 To ovulation day + 7 9 6 (0.53) 0.55 (0.56) 6 (7) 0.79 To day of pregnancy test 0.30 (0.59) 7 (3) 0.32 (0.96) 0.55 Values were considered significant if P < 5. (a) (b) (c) (d) (e) (f) Figure 1 Receiver operating characteristics curves for uterine artery pulsatility index measurements during in-vitro fertilization-intracytoplasmic sperm injection treatment cycles, with assessment: (a) on day of start of the cycle (area under the curve (AUC) = 1), (b) at end of stimulation phase (AUC = 3), (c) at ovum pick-up (AUC = 4), (d) at embryo transfer (AUC = 1), (e) at mid-luteal phase (AUC = 6) and (f) on day of pregnancy test, 14 16 days after ovum pick-up (AUC = 0.53). univariate analysis the only significant difference in patient characteristics or cycle characteristics between pregnant patients and non-pregnant patients was embryo score of the transferred embryos. It is well known that female age, for example, is a strong determining factor in IVF success rates, but we could not substantiate this, probably due to the small size of our study group. The small size of our study group might also be responsible for the absence of significant differences in uterine vascular impedance values. Another possible explanation is that, although vascular endometrial development is crucial for implantation, it is not so crucial in the stage that we assessed. This might be because during the window of implantation the embryo is

Serial uterine artery flow velocities and IVF-ICSI 437 Table 5 Results of logistic regression analysis Variable Odds ratio 95% CI Change in PI to: Day of hcg administration 1.36 0.38 1.44 Day of ovum pick-up 1.31 0.90 1.55 Day of embryo transfer 1.52 4 1.88 Ovulation day + 7 9 3 8 9 Day of pregnancy test 1 7 2.01 Odds ratio was considered significant if a value of 1 did not fall within the 95% CI. In the model, pregnancy was the outcome and pulsatility index (PI), age, endometrial thickness on day of human chorionic gonadotropin (hcg) injection, treatment (in-vitro fertilization (IVF)-intracytoplasmic sperm injection (ICSI)), IVF indication, primary or secondary infertility and total embryo quality were explanatory variables. a known anaerobic organism which is best cultured invitro in a low-oxygen atmosphere 28. Whether/when this state changes in the intrauterine phase of treatment is not known. In fact, we did see in follow-up studies of these patients that PIs remained low (or even tended to decrease a little further) in the first trimester of pregnancy (data not shown). Around the time of hatching and attachment of the embryo, the local circumstances in the uterus are optimal, but it seems that to a great extent factors involving the embryo itself determine whether implantation will take place. In fact, the uterus is perhaps more limiting than it is facilitating to implantation. It is well known that endometrium is not essential for embryo implantation, as proved by the occurrence of ectopic pregnancies 29.The uterus is actually the only female organ that protects the mother actively from the invasive growth and potentially destructive trophoblast. One final possible explanation of our findings, with the uterine artery being some distance from the site of implantation and not supplying the endometrium exclusively, is that the assessment of uterine artery PI was not sufficiently sensitive to assess the microvascular events of embryo implantation. In conclusion, we have shown that there does not seem to be a place for PI assessment as a routine non-invasive marker of implantation. Future research should always compensate for known confounders and assess whether there is a role for PI in a more selected population, looking in more detail in and around the window of implantation. Investigation of the vascular process of embryo implantation by assessing known markers of angiogenesis and vascular growth factors combined with ultrasound will hopefully lead to greater understanding of this process. REFERENCES 1. Assisted reproductive technology in the United States: 1997 results generated from the American Society for Reproductive Medicine/Society for Assisted Reproductive Technology Registry. Fertil Steril 2000; 74: 641 653. 2. Hoozemans DA, Schats R, Lambalk CB, Homburg R, Hompes PG. Human embryo implantation: current knowledge and clinical implications in assisted reproductive technology. Reprod Biomed Online 2004; 9: 692 715. 3. Psychoyos A. Hormonal control of ovoimplantation. Vitam Horm 1973; 31: 201 256. 4. Psychoyos A. Hormonal control of uterine receptivity for nidation. J Reprod Fertil Suppl 1976; 25: 17 28. 5. Fanchin R. Assessing uterine receptivity in 2001: ultrasonographic glances at the new millennium. Ann N Y Acad Sci 2001; 943: 185 202. 6. Steer CV, Campbell S, Pampiglione JS, Kingsland CR, Mason BA, Collins WP. Transvaginal colour flow imaging of the uterine arteries during the ovarian and menstrual cycles. Hum Reprod 1990; 5: 391 395. 7. Goswamy RK, Williams G, Steptoe PC. Decreased uterine perfusion a cause of infertility. Hum Reprod 1988; 3: 955 959. 8. Goswamy RK, Steptoe PC. Doppler ultrasound studies of the uterine artery in spontaneous ovarian cycles. Hum Reprod 1988; 3: 721 726. 9. de Ziegler D, Bessis R, Frydman R. Vascular resistance of uterine arteries: physiological effects of estradiol and progesterone. Fertil Steril 1991; 55: 775 779. 10. Achiron R, Levran D, Sivan E, Lipitz S, Dor J, Mashiach S. Endometrial blood flow response to hormone replacement therapy in women with premature ovarian failure: a transvaginal Doppler study. Fertil Steril 1995; 63: 550 554. 11. Scholtes MC, Wladimiroff JW, van Rijen HJ, Hop WC. Uterine and ovarian flow velocity waveforms in the normal menstrual cycle: a transvaginal Doppler study. Fertil Steril 1989; 52: 981 985. 12. Steer CV, Tan SL, Dillon D, Mason BA, Campbell S. Vaginal color Doppler assessment of uterine artery impedance correlates with immunohistochemical markers of endometrial receptivity required for the implantation of an embryo. Fertil Steril 1995; 63: 101 108. 13. Sterzik K, Grab D, Sasse V, Hutter W, Rosenbusch B, Terinde R. Doppler sonographic findings and their correlation with implantation in an in vitro fertilization program. Fertil Steril 1989; 52: 825 828. 14. Steer CV, Campbell S, Tan SL, Crayford T, Mills C, Mason BA, Collins WP. The use of transvaginal color flow imaging after in vitro fertilization to identify optimum uterine conditions before embryo transfer. Fertil Steril 1992; 57: 372 376. 15. Coulam CB, Bustillo M, Soenksen DM, Britten S. Ultrasonographic predictors of implantation after assisted reproduction. Fertil Steril 1994; 62: 1004 1010. 16. Serafini P, Batzofin J, Nelson J, Olive D. Sonographic uterine predictors of pregnancy in women undergoing ovulation induction for assisted reproductive treatments. Fertil Steril 1994; 62: 815 822. 17. Zaidi J, Pittrof R, Shaker A, Kyei-Mensah A, Campbell S, Tan SL. Assessment of uterine artery blood flow on the day of human chorionic gonadotropin administration by transvaginal color Doppler ultrasound in an in vitro fertilization program. Fertil Steril 1996; 65: 377 381. 18. Cacciatore B, Simberg N, Fusaro P, Tiitinen A. Transvaginal Doppler study of uterine artery blood flow in in vitro fertilization-embryo transfer cycles. Fertil Steril 1996; 66: 130 134. 19. Tekay A, Martikainen H, Jouppila P. Blood flow changes in uterine and ovarian vasculature, and predictive value of transvaginal pulsed colour Doppler ultrasonography in an invitro fertilization programme. Hum Reprod 1995; 10: 688 693. 20. Yang JH, Wu MY, Chen CD, Jiang MC, Ho HN, Yang YS. Association of endometrial blood flow as determined by a modified colour Doppler technique with subsequent outcome of in-vitro fertilization. Hum Reprod 1999; 14: 1606 1610. 21. Chien LW, Tzeng CR, Chang SR, Chen AC. The correlation of the embryo implantation rate with uterine arterial impedance

438 Hoozemans et al. in in vitro fertilization and embryo transfer. Early Pregnancy 1995; 1: 27 32. 22. Steer CV, Williams J, Zaidi J, Campbell S, Tan SL. Intraobserver, interobserver, interultrasound transducer and intercycle variation in colour Doppler assessment of uterine artery impedance. Hum Reprod 1995; 10: 479 481. 23. Schild RL, Holthaus S, d Alquen J, Fimmers R, Dorn C, van Der Ven H, Hansmann M. Quantitative assessment of subendometrial blood flow by three-dimensional-ultrasound is an important predictive factor of implantation in an in-vitro fertilization programme. Hum Reprod 2000; 15: 89 94. 24. Smith SK. Angiogenesis and implantation. Hum Reprod 2000; 15 (Suppl 6): 59 66. 25. Tan SL, Zaidi J, Campbell S, Doyle P, Collins W. Blood flow changes in the ovarian and uterine arteries during the normal menstrual cycle. Am J Obstet Gynecol 1996; 175: 625 631. 26. Roseboom TJ, Vermeiden JP, Schoute E, Lens JW, Schats R. The probability of pregnancy after embryo transfer is affected by the age of the patient, cause of infertility, number of embryos transferred and the average morphology score, as revealed by multiple logistic regression analysis. Hum Reprod 1995; 10: 3035 3041. 27. Steer CV, Mills CL, Tan SL, Campbell S, Edwards RG. The cumulative embryo score: a predictive embryo scoring technique to select the optimal number of embryos to transfer in an invitro fertilization and embryo transfer programme. Hum Reprod 1992; 7: 117 119. 28. Bavister B. Oxygen concentration and preimplantation development. Reprod Biomed Online 2004; 9: 484 486. 29. Lenglet JE, Bekker MN, Akkerman C, Bolte AC, van Vugt JM. Prenatal ultrasound and MRI predict placental localization in a combined intrauterine and extrauterine twin pregnancy. Prenat Diagn 2006; 26: 376 378.