Colour Doppler indices of follicular blood flow as predictors of pregnancy after in-vitro fertilization and embryo transfer

C.B. Coulam1, C. Goodman and J.S. Rinehart

The Center for Human Reproduction, 750 N. Orleans Street, Chicago, IL 60610, USA


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Peak systolic velocity (PSV) of individual follicles has been correlated with oocyte recovery, fertilization rate and embryo quality [in women undergoing in-vitro fertilization (IVF) and embryo transfer]. The present study assessed the role of quantitative and qualitative indices of follicular vascularity in predicting pregnancy after IVF and embryo transfer. A total of 106 women undergoing IVF treatment for infertility who were considered to be at risk of failure (>37 years of age, history of low response to gonadotrophin stimulation, or multiple failed IVF cycles) constituted the study group. PSV was measured from the three largest follicles on both the right and left ovaries on the day of human chorionic gonadotrophin (HCG) administration using an Acuson Sequoia with a 4–8 MHz transvaginal probe. The quality of follicular flow was graded from 1 to 4 according to the amount of visible colour flow around the follicle (grade 1 when one-quarter of the follicle, grade 2 when one-half, grade 3 when three-quarters, and grade 4 when the entire follicle was surrounded by colour). Clinical pregnancies resulted in 11 (10%) of the 106 high-risk women. Women who had PSV >= 10 cm/s in at least one follicle on the day of HCG administration more often became pregnant than those with PSV <10 cm/s (P = 0.05). All pregnancies occurred in women with grade 3 or 4 follicular blood flow. Qualitative as well as quantitative measurements of follicular flow predict pregnancy after IVF and embryo transfer.

Key words: colour Doppler/embryo transfer/follicular blood flow/in-vitro fertilization


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Despite advances in reproductive technology, a need for introduction of procedures that will improve the efficacy, and thus reduce the cost, of treating infertile couples by in-vitro fertilization (IVF) and embryo transfer remains. Although many factors contribute to success in IVF programmes, a major role in determining delivery rate is played by the quality of oocytes retrieved (Sauer et al., 1992Go). Transvaginal ultrasonography with colour and pulsed Doppler ultrasound has enabled the assessment of vascular changes in the ovary during the menstrual cycle (Bourne et al., 1991Go; Collins et al., 1991Go; Campbell et al., 1993Go; Kupesic and Kurjak 1993Go; Sladkevicius et al., 1993Go; Dickey, 1997Go). During spontaneously ovulating cycles, an increase in the maximum peak systolic velocity (PSV) in vessels supplying the dominant follicle has been documented (Bourne et al., 1991Go; Collins et al., 1991Go; Campbell et al., 1993Go; Kupesic and Kurjak 1993Go; Dickey, 1997Go). PSV of individual follicles among women undergoing IVF has been shown to be correlated with oocyte recovery (Nargund et al., 1996aGo,bGo; Oyesanya et al., 1996Go), fertilization rate (Nargund et al., 1996aGo), developmental potential of the oocyte (Van Blerkom et al., 1997Go) and quality of the preimplantation embryo (Nargund et al., 1996aGo; Oyesanya et al., 1996Go; Chui et al., 1997Go). A preliminary study (Chui et al., 1997Go) using qualitative assessment of follicular flow suggested that high-grade follicular vascularity is associated with increased pregnancy rate. The aim of the present study is to define the role of quantitative measurements of follicular blood flow in predicting pregnancy following IVF and embryo transfer.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Patients
A total of 106 women undergoing IVF treatment for infertility between April 5, 1997, and January 1, 1998, who were considered by their reproductive endocrinologist to be at increased risk for treatment failure, constituted the study group. Inclusion criteria for consideration of increased risk for treatment failure were as follows: age >37 years, history of low response to gonadotrophin stimulation, and previous IVF failure. A history of low response to gonadotrophin stimulation was diagnosed during human chorionic gonadotrophin (HCG) administration, in a previous IVF cycle, when there were fewer than four follicles between 12 and 22 mm (mean value of two diameters) or when the serum oestradiol concentration, drawn between 6 a.m. and 8 a.m., was <500 pg/ml (1836 SI units). Multiple IVF failures were arbitrarily defined as more than two previous IVF attempts without successful pregnancy.

Treatment regimens for ovarian stimulation were determined by the referring reproductive endocrinologist (n = 9) and were based on the woman's age and previous response to treatment. Treatment regimens included standard protocols and stop lupron protocols. The standard treatment protocol utilized luteal long-acting gonadotrophin-releasing hormone agonist (GnRHa; TAP Pharmaceuticals Inc.; Deerfield, IL, USA) down-regulation followed by gonadotrophin stimulation. The standard protocol initiated gonadotrophin therapy with 300 IU of follicle stimulating hormone (FSH) (Serono Laboratories; Randolph, MA, USA) per day in a split dose. The patients who had responded inadequately to the standard protocol in an earlier cycle initiated their gonadotrophin stimulation with either 450 IU FSH per day or 600 IU FSH per day in split doses. For some patients, the long-acting GnRHa was stopped after menses and then gonadotrophin stimulation was begun (stop lupron protocol). IVF was performed in the usual way using conscious i.v. sedation for an ultrasound-guided vaginal oocyte retrieval. Embryo transfer was performed vaginally on the second day after retrieval (day 0 was day of retrieval). Embryos were assigned a subjective quality score patterned after Veeck (1988), with 1 being a very poor quality, non-viable embryo and 5 being of optimal quality. The number of clinical pregnancies (gestational sac visible on transvaginal ultrasound at 3 weeks after embryo transfer) was recorded. A successful IVF cycle was defined as documentation of clinical pregnancy.

Colour Doppler imaging
Transvaginal colour Doppler imaging and pulsed Doppler spectral analysis were performed on the day of, but prior to, HCG administration. All scans were performed using an Acuson Sequoia (Acuson, Mountain View, CA, USA) with a 4–8 MHz transvaginal probe. Spectral analysis was performed using colour Doppler mapping on the three largest follicles from each of the right and left ovaries. The largest follicles were determined by the average of the maximum transverse and anteroposterior diameters. The spatial peak temporal average intensity for B-mode and colour Doppler imaging was <80 mW/cm2, which is within the safety limits recommended by the Bioeffects Committee of the American Institute of Ultrasound in Medicine (AIUM, 1993). The vascularity of each study follicle was graded from 1 to 4 according to the percentage of the follicular circumference in which flow was visible at the time of the scan. The grading system has been previously reported (Chui et al., 1997Go) and was as follows: grade 0 = no flow visible; grades 1, 2, 3 and 4 respectively had <25%, 26–50%, 51–75% and 76–100% of follicular circumference in which flow was identified. For all colour Doppler mapping, the following presets were held constant: spatial time resolution, S1; edge, –1; persistence, 2; post-processing, V:2; gate size, 1; and filter, 1. Scale was set at 0.025 m/s.

Spectral analysis was performed with the filter set under 50 Hz and sample volume range of 1.0–1.2 mm2. Flow velocity waveforms were used to calculate the highest PSV. The angle of the probe was adjusted until a flow velocity waveform with the maximum peak systolic shifted frequency was obtained and the value recorded.

Experimental design and statistical analysis
Women undergoing IVF for treatment of infertility had quantitative and qualitative indices of follicular vascularity obtained on the day of HCG administration. PSV was measured and quality of flow was graded from the three largest follicles on both the right and left ovaries. Mean PSV of follicular flow were compared between women experiencing clinical pregnancies and those that did not cycle using an unpaired Student's t-test. Previous studies had indicated that a follicle with a PSV >=10 cm/s had a 70% chance of producing a good quality embryo compared to a 14% chance if PSV was <10 cm/s (Nargund et al., 1996aGo,bGo). We therefore compared clinical pregnancy rates among women possessing at least one follicle demonstrating a PSV >=10 cm/s and grade 3–4 pattern of flow with those who did not, using Fisher's exact test. For continuous variables (Table IGo), a logit regression was performed to compare pregnant and non-pregnant patients. P < 0.05 was regarded as significant.


View this table:
[in this window]
[in a new window]
 
Table I. Comparison of selected patient characteristics between women who became pregnant and those who did not in the same cycle as PSV measurement (mean ± SD)
 

    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
A total of 106 women aged 26–42 years undergoing 114 cycles of IVF for treatment of infertility had 565 follicles studied using colour Doppler flow. All the women were considered to be at risk for treatment failure for the following reasons: age >37 years (n = 35, 33%), history of low response to gonadotrophin stimulation (n = 43, 41%) and history of IVF failure (n = 62, 59%). Some women had more than one of the above criteria; hence, the sum of the percentages totals more than 100%. Clinical pregnancies resulted in 11 (10%) of the 106 high-risk women. Eight (73%) of the pregnancies ended in delivery of healthy infants and three (27%) aborted.

Of the 565 follicles studied, five were grade 0, 87 were grade 1, 138 were grade 2, 194 were grade 3, and 141 were grade 4. The results demonstrated that follicular vascularity grading was independent of follicular size. All pregnancies occurred in women with grade 3 or 4 follicular blood flow.

When women who conceived were compared with those who did not, no significant differences in mean age, number of days of stimulation, or follicular PSV were observed (Table IGo). Women who subsequently became pregnant had a greater number of oocytes retrieved and embryos transferred than women who did not become pregnant. Women who received four or more embryos per transfer had a significantly higher pregnancy rate than those who had fewer than four embryos transferred (33 versus 3%, P = 0.002). However, if fewer than four follicles were measured for PSV, the rate of pregnancy in that cycle was the same as for four or more follicles measured, i.e. if at least one follicle had a PSV >=10 cm/s and was grade 3–4 (P = 0.1) (Table IIGo).


View this table:
[in this window]
[in a new window]
 
Table II. Comparison of total number of follicles measured and subsequent pregnancy rate that same cycle
 
The number of follicles with grade 3–4 flow within individual patients is compared with pregnancy rates for those cycles in Table IIIGo. Pregnancy rate among women with one, two, three, four, five and six follicles classified as grade 3–4 was similar. The association between maximum follicular PSV and subsequent pregnancy in that cycle is shown in Table IVGo. Among the 11 pregnancy cycles, 10 (91%) had a threshold value of grade 3–4 flow with PSV >=10 cm/s in at least one follicle. Thus, a threshold value of >=10 cm/s for PSV in at least one follicle with grade 3–4 flow detected 91% of pregnancies with a false positive rate of 9%. There was a 15% (10/67) chance of clinical pregnancy per patient [13% (10/76) per cycle] in this high-risk population if the maximum follicular PSV was >=0 cm/s as compared to 3% if the maximum follicular PSV was <10 cm/s (P = 0.05). A threshold value of >=10 cm/s for PSV in at least one grade 3–4 follicle predicted a pregnancy with a sensitivity of 91%, specificity of 36%, positive predictive value of 13%, and negative predictive value of 97%.


View this table:
[in this window]
[in a new window]
 
Table III. Pregnancy rates per number of follicles with grade 3–4 flow
 

View this table:
[in this window]
[in a new window]
 
Table IV. Association between maximum follicular PSV at time of HCG administration and pregnancy
 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Analysis of results from the present study demonstrates a direct relationship between establishment of clinical pregnancy after IVF and embryo transfer and follicular blood flow expressed by values of PSV. A previous study showed that follicular vascularity measured by qualitative grading predicted pregnancy after IVF and embryo transfer (Chui et al., 1997Go). In that study (Chui et al., 1997Go), pregnancies were confined to women whose embryos were derived from follicles exhibiting grade 3 and 4 vascularity and with pregnancy rates per embryo transfer of 12.5 and 61.5% respectively. The present study confirmed the previous finding of a relationship between pregnancy and qualitative grading of follicular vascularity and extended the observation to include quantitative measurements of follicular blood flow. All pregnancies occurred in women with grade 3 and 4 follicular vascularity and 91% of pregnancies occurred with follicular PSV >=10 cm/s. Women who had at least one follicle with a PSV of >=10 cm/s at the time of HCG administration had a significantly higher pregnancy rate than women with maximum follicular PSV <10 cm/s. A threshold value of PSV >=10 cm/s predicted pregnancy with high sensitivity (91%) and negative predictive value (97%), but low specificity (36%) and positive predictive value (13%), suggesting that reasons not associated with follicular flow are associated with lack of pregnancy in this high-risk population. However, even in a population at high risk for IVF failure, women with follicular flow of PSV >=10 cm/s and grade 3–4 had a 13% pregnancy rate that cycle (Table IVGo). This finding is consistent with previous studies (Nargund et al., 1996aGo,bGo) that reported a significant relationship between follicular PSV within a given follicle and the recovery of an oocyte and the subsequent production of an embryo. The probability of producing a high-grade embryo was 75% if the follicular PSV was >=10 cm/s (Nargund et al., 1996bGo). Transfer of high-grade embryos has been associated with higher pregnancy rates than transfer of low-grade embryos (Staessen et al., 1992Go).

The potential role of colour Doppler imaging in the overall assessment of folliculogenesis in women undergoing IVF and embryo transfer has been controversial (Dickey, 1997Go). Previous studies have involved the use of mean values for the PSV from many follicles (Balakier and Stronnell, 1994), the maximum PSV from serial monitoring (Tekay et al., 1995Go) and the PSV from individual follicles (Nargund et al., 1996aGo,bGo). When mean values for follicular PSV (Balakier and Stronnell, 1994) and maximum PSV from serial monitoring (Tekay et al., 1995Go) were evaluated, no differences in values between conception and non-conception cycles were noted. However, when individual follicles, oocytes, and preimplantation embryos were studied, a significant difference in PSV was found in conception compared with non-conception cycles. Thus, the discrepant findings can be explained by the experimental designs. The studies that minimize the between-follicle differences (Balakier and Stronnell, 1994; Tekay et al., 1995Go) yield no-difference results and those that detailed individual follicles showed significant results (Nargund et al., 1996aGo,bGo). In fact, in the present study and in the studies of individual follicles, no differences in mean value of follicular PSV were observed between women conceiving and those who did not. Thus, it is the individual follicle(s)/oocyte(s), not the cohort, that determines successful outcome. While it is true that the more follicles/oocytes available, the higher the probability that one of these will have the developmental potential to generate a viable pregnancy, it also holds true that if only one follicle possessing the developmental potential is available, the probability of pregnancy approaches 100%. The concept of quality rather than quantity is further emphasized when the number of follicles with grade 3–4 flow in an individual was compared to pregnancy rates that cycle (Table IIIGo). Pregnancy rates were similar in cycles with at least one follicle demonstrating grade 3–4 flow and in cycles with six follicles. Even in women who had fewer than four follicles measured at the time of HCG administration, 33% (3/9) became pregnant in that cycle if at least one grade 3–4 follicle had a PSV >=10 cm/s (Table IIGo). The key is to be able to identify follicles that hold the potential to generate viable pregnancies.

Developmental competence of a human oocyte is heterogeneous within cohorts of human oocytes. This fact is demonstrated most clearly by IVF programmes in which apparently normal metaphase II (MII)-stage oocytes resist penetration by competent spermatozoa (Bedford and Kim, 1993Go), fail to progress after penetration (Asch et al., 1995Go), or arrest development during the preimplantation stages (Edwards, 1986Go; Osborn and Moor, 1988Go; Van Blerkom, 1993Go; Van Blerkom et al., 1994Go). Metabolic differences between morphologically equivalent MII-stage oocytes have been reported and include oxygen consumption and ATP content (Magnusson et al., 1986Go; Van Blerkom, 1993Go). In addition, follicular fluid concentrations of oxygen (Gosden and Byat-Smith, 1986Go; Fischer et al., 1992Go; Van Blerkom, 1996Go; Van Blerkom et al., 1997Go) and vascular endothelial growth factor (Van Blerkom et al., 1997Go) have been associated with the developmental competence of the corresponding oocyte. The dissolved oxygen content of the follicle has been shown to be related to colour Doppler qualitative patterns (Van Blerkom et al., 1997Go). The current findings of an association between follicular PSV and subsequent clinical pregnancy are consistent with the hypothesis that follicular blood flow may be associated with events essential for successful reproduction. The mechanisms controlling follicular blood flow are unknown as is the link between vascularity and pregnancy. Whatever the mechanisms, the implications in IVF programmes are far reaching.

The knowledge gained from our study may have use in future management of women undergoing IVF and embryo transfer treatment. Pregnancy rates would be expected to be increased by selecting for transfer of embryos derived from follicles with PSV >=10 cm/s. Conversely, pregnancy rates would be expected to decrease in cycles where the follicular PSV are universally poor. It may be appropriate to counsel these women as to the advisability of proceeding with oocyte retrieval. The net effect of the above actions is an increased pregnancy rate per embryo transfer. Currently, improved pregnancy rates after IVF and embryo transfer have come at a cost of an increased frequency of higher order gestations (Faber, 1997Go). If embryos with high developmental competence can be identified and transferred, then a small number of such replaced embryos will not reduce pregnancy rates, but rather the probability of multiple gestations. Prospective clinical trials are necessary to test these hypotheses.


    Notes
 
1 To whom correspondence should be addressed Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
American Institute of Ultrasound in Medicine (AIUM) (1993) Bioeffect and Safety of Diagnostic Ultrasound. Laurel, MD.

Asch, R., Simerly, C., Ord, T. et al. (1995) The stages at which human fertilization arrests: microtubule and chromosome configurations in inseminated oocytes which failed to complete fertilization and development in humans. Hum. Reprod., 10, 1897–1906.[Abstract]

Balakier, H. and Stronell, R.D. (1994) Colour Doppler assessment of folliculogenesis in in vitro fertilization patients. Fertil. Steril., 62, 1211–1216.[ISI][Medline]

Bedford, J.M. and Kim, H.H. (1993) Sperm/egg binding patterns and oocyte cytology in a retrospective analysis of fertilization failures in vitro. Hum. Reprod., 8, 453–463.[Abstract]

Bourne, T.H., Jurkovic, D., Waterstone, J. et al. (1991) Follicular blood flow during human ovulation. Ultrasound Obstet. Gynecol., 1, 53–59.[ISI][Medline]

Campbell, S., Bourne, T.H., Waterstone, J. et al. (1993) Transvaginal colour flow imaging of periovulatory follicle. Fertil. Steril., 60, 433–438.[ISI][Medline]

Chui, D.K.C., Pugh, N.D., Walker, S.M. et al. (1997) Follicular vascularity – the predictive value of transvaginal power Doppler ultrasonography in an in-vitro fertilization programme: a preliminary study. Hum. Reprod., 12, 191–196.[ISI][Medline]

Collins, W., Jurkovic, D., Bourne, T. et al. (1991) Ovarian morphology, endocrine function and intra-follicular blood flow during the periovulatory period. Hum. Reprod., 6, 319–324.[Abstract]

Dickey R.P. (1997) Doppler ultrasound investigation of uterine and ovarian flow in infertility and early pregnancy. Hum. Reprod. Update, 3, 467–502.[Abstract/Free Full Text]

Edwards, R.G. (1986) Causes of early embryonic loss in human pregnancy. Hum. Reprod., 1, 185–198.[ISI][Medline]

Faber, K. (1997) IVF in the US: multiple gestations, economic competition, the necessity of excess. Hum. Reprod., 12, 1614–1616.[ISI][Medline]

Fischer, B., Künzel, W., Kleinstein, J. and Gips, H. (1992) Oxygen tension in follicular fluid falls with follicle maturation. Eur. J. Obstet. Gynecol. Reprod. Biol., 43, 39–43.[ISI][Medline]

Gosden, R.G. and Byat-Smith, J.G. (1986) Oxygen concentration gradient across the ovarian follicular epithelium: model, predictions and implications. Hum. Reprod., 1, 65–68.[Abstract]

Kupesic, S. and Kurjak, A. (1993) Uterine and ovarian perfusion during the periovulatory period assessed by transvaginal colour Doppler. Fertil. Steril., 60, 439–443.[ISI][Medline]

Magnusson, C., Hillensjo, T., Hamberger, L. and Nilsson, L. (1986) Oxygen consumption by human oocytes and blastocysts grown in vitro. Hum. Reprod., 1, 183–184.[Abstract]

Nargund, G., Bourne, T., Doyle, P. et al. (1996a) Associations between ultrasound indices of follicular blood flow, oocyte recovery and preimplantation embryo quality. Hum. Reprod., 11, 109–113.[Abstract]

Nargund, G., Doyle, P.E., Bourne, T.H. et al. (1996b) Ultrasound derived indices of follicular blood flow before HCG administration and prediction of oocyte recovery and preimplantation embryo quality. Hum. Reprod., 11, 2512–2517.[Abstract]

Osborn, J. and Moor, R. (1988) An assessment of the factors causing embryonic loss after fertilization in vitro. J. Reprod. Fertil., 36, 59–72.

Oyesanya, O.A., Parsons, J.H., Collins, W.P. and Campbell, S. (1996) Prediction of oocyte recovery rate by transvaginal ultrasonography and colour Doppler imaging before human chorionic gonadotropin administration in in vitro fertilization cycles. Fertil. Steril., 65, 806–809.[ISI][Medline]

Sauer, M.V., Paulson, R.J. and Lobo, R.A. (1992) Reversing the natural decline in human fertility: an extended clinical trial of oocyte donation to women of advanced reproductive age. J. Am. Med. Assoc., 268, 1275–1279.[Abstract]

Sladkevicius, P., Valentin, L. and Marsal, K. (1993) Blood flow in the uterine and ovarian arteries during the normal menstrual cycle. Ultrasound Obstet. Gynecol., 3, 199–208.[ISI][Medline]

Staessen, C., Camus, M., Bollen, N. et al. (1992) The relationship between embryo quality and the occurrence of multiple gestation. Fertil. Steril., 3, 626–630.

Tekay, A., Martikainen, H. and Jouppila, P. (1995) Blood flow changes in uterine and ovarian vasculature and predictive value of transvaginal pulsed colour Doppler ultrasonography in an in vitro fertilization programme. Hum. Reprod., 10, 688–693.[Abstract]

Van Blerkom, J. (1993) Development of human embryos to the hatched blastocyst stage in the presence or absence of a monolayer of Vero cells. Hum. Reprod., 8, 1525–1539.[Abstract]

Van Blerkom, J. (1996) The influence of intrinsic and extrinsic factors on the developmental potential of chromosomal normality of the human oocyte. J. Soc. Gynecol. Invest., 3, 3–11.[ISI][Medline]

Van Blerkom, J., Davis, P.W. and Merriam, J. (1994) A retrospective analysis of unfertilized and presumed parthogenetically activated human oocytes demonstrates a high frequency of sperm penetration. Hum. Reprod., 9, 2381–2388.[Abstract]

Van Blerkom, J., Antczak, M. and Schrader, R. (1997) The developmental potential of the human oocyte is related to the dissolved oxygen content of follicular fluid: association with vascular endothelial growth factor levels and perifollicular blood flow characteristics. Hum. Reprod., 12, 1047–1055.[ISI][Medline]

Veeck, L.L. (1988) Oocyte assessment and biological performance. Ann. N. Y. Acad. Sci., 541, 259–274.[ISI][Medline]

Submitted on January 5, 1999; accepted on May 11, 1999.