Bi-directional communication between the follicle
and oocyte is essential for regulating both follicle
and oocyte development (
At present, follicle diameter and serum estradiol
levels are used as the primary markers for determining
the maturity of the follicle and oocyte prior to
oocyte retrieval and
Zegers-Hochschild et al. reported that stabilization
of follicular growth 24 hours prior to ovulation and
a short interval (i.e., 24 hours or less) between the
luteinizing hormone (LH) surge and ovulation was
associated with conception during the natural menstrual
A decrease in the number of follicles occupying
the ovarian reserve occurs with age (
It is not practical, at present, to monitor the serial growth of individually identified follicles in women undergoing assisted reproduction. The ovarian stimulation phase length (SPL) is a user-friendly way for clinicians to monitor the overall growth of all follicles and serves as an indirect measure of mean follicular growth rate. Determination of an optimal SPL that results in competent oocytes and successful IVF outcomes would allow clinicians to tailor patient stimulation protocols to maximize the chance of a successful treatment. The objective of this study was to determine whether SPL influences follicle, oocyte, embryo and pregnancy outcomes in women of different ages undergoing IVF. We hypothesized that short or long stimulation phases would be associated with suboptimal follicle, oocyte, embryo and pregnancy outcomes. We further hypothesized that advanced age would be associated with a shorter stimulation phase and suboptimal follicle, oocyte, embryo and pregnancy outcomes.
It is not practical or ethical to conduct a prospective study to evaluate the effects of different ovarian stimulation phase lengths on human IVF outcomes. Therefore, a retrospective cohort study was conducted for patients that underwent IVF at the ARTUS Fertility Center in the Department of Obstetrics, Gynecology and Reproductive Science at the University of Saskatchewan. Ethical approval was obtained from the University of Saskatchewan’s Biomedical Research Ethics Board.
Women with a history of poor response to ovarian stimulation, polycystic ovarian syndrome, those undergoing superovulation (SO) or ICSI treatment cycles, or who had converted from SO to IVF were not included. In addition, frozen embryo transfer cycles, oocyte donor cycles and couples with male factor infertility were excluded from the study. A total of 148 charts were reviewed. A total of 14 women were further excluded from the study for the following reasons: inadequate follicular response (n=12), premature ovulation (n=1) and unavailable participant semen sample (n=1). The resulting sample consisted of 134 women.
All patients were cared for by the same clinician. Follicular development was monitored every 2-3 days during ovarian stimulation using high-resolution transvaginal ultrasonography (multi-frequency 5-9 MHz curvilinear transducer, SONIX OP Ultrasound System, Ultrasonix Medical Corporation, Burnaby, BC Canada). All follicles ≥2 mm were measured and recorded at each ultrasound examination. The diameters of follicles <10 mm were measured in a single plane. For all follicles >10 mm, mean maximal follicle diameter was calculated as the mean of the follicle length and width in the widest plane of section.
All patients were suppressed with oral contraception (OC; Marvelon, Organon, Canada), prior to ovarian stimulation to synchronize their cycles (n=134). Patients received either a gonadotropinreleasing hormone (GnRH) agonist [Suprefact, Sanofi-Aventis (n=56)] for 5 days before discontinuing OCs and continuing until the day of hCG (i.e., long GnRH agonist protocol) or a GnRH antagonist [(Cetrotide, EMD Serono, Inc. (n=62); Orgalutran, Organon, Canada (n=16)] beginning on day 6 of the follicular phase and continuing until the day of hCG (i.e., fixed protocol). Recombinant FSH [Gonal F, Serono, Canada, Inc. (n=100); Puregon, Organon, Canada (n=18); Bravelle, Ferring Pharmaceuticals, Inc. (n=11); or Menopur, Ferring Pharmaceuticals, Inc. (n=5)] was administered daily, beginning on day 2 or 3 after menses. When > 3 follicles reached >17 mm, 10,000 IU hCG (Profasi, Serono Canada Inc.; Chorionic Gonadotropin, Pharmaceutical Partners of Canada) was administered to induce final maturation of the follicles. Blood was drawn at each monitoring visit during ovarian stimulation to measure serum estradiol concentrations. The SPL was defined as the time period from the start of follicle stimulating hormone (FSH) administration to the day of hCG administration.
Oocyte retrievals were performed approximately 35 hours following hCG administration. Oocytes were inseminated (standardized sperm concentration of 80-120 × 106 sperm/mL) 4 hours following retrieval. Fertilization was assessed the next day and daily embryo monitoring was conducted thereafter. Fertilization rate was defined as the number of normally fertilized oocytes (release of the second polar body and two pronuclei) out of the total number of oocytes collected. Cleavage rate was characterized by the number of zygotes that underwent division on day 2 out of the total number of oocytes fertilized (day 0 = day of oocyte retrieval). Blastocyst rate was defined as the number of blastocysts that developed out of the total number of oocytes fertilized. An appropriate number of embryos [1 embryo (n=13), 2 embryos (n=100) or 3 embryos (n=27)] were transferred 3 or 5 days later; based on patient age, diagnosis, previous treatment success, and number and quality of embryos available for transfer. A serum β-hCG test was performed 2 weeks post-retrieval. Chemical pregnancy was defined as a positive serum pregnancy test 2 weeks following oocyte retrieval. Clinical pregnancy was documented ultrasonographically as a positive fetal heartbeat at 8-12 weeks of gestation. Live birth data were not available due to the inability to consistently obtain birth outcome results from patients.
SPL was considered the independent variable. Pregnancy rate (chemical and clinical) was the primary dependent outcome variable. Secondary outcome variables included: number of follicles > 6, 10 and 15 mm, serum estradiol concentration on the day of hCG, number of oocytes collected, fertilization rate, cleavage rate, blastocyst rate and endometrial thickness. Statistical significance was set at p<0.05.
SPL data were stratified in the following manner: 1) <10 days, 2) 10-12 days, and 3) >12 days. Patient demographic characteristics were compared among the three SPL groups using one-way analyses of variance tests and Scheffe post-hoc tests for continuous variables and chi-square tests for categorical variables (SPSS version 17.0; SPSS Inc., Chicago, IL, USA). Associations between SPL and primary/ secondary outcomes were evaluated using multivariate linear and logistic regression (SPSS Version 17.0; SAS Version 9.2). Age, FSH start day, FSH start dose, FSH regimen and GnRH agonist/antagonist use were included in the multivariate regression models to evaluate these variables as potential confounders or covariates.
No differences in body mass index (BMI), gravidity,
parity or age were detected between women
with short, medium or long SPL (p>0.05,
Comparison of patient demographics in women with short (<10 days), moderate (10-12 days) and long (>12 days) stimulation phase lengths
|Overall||<10 days (n=18)||10 to 12 days (n=101)||>12 days (n=15)||P value|
|25.3 ± 0.5||25.2 ± 1.4||25.6 ± 0.6||23.9 ± 1.3||0.6|
|1.5 ± 0.1||1.9 ± 0.5||1.3 ± 0.2||1.8 ± 0.5||0.4|
|0.7 ± 0.1||0.6 ± 0.2||0.6 ± 0.1||1.1 ± 0.3||0.2|
|33.8 ± 0.4||33.8 ± 1.1||33.8 ± 0.4||33.6 ± 1.1||1.0|
a, b Within rows, values with common superscripts are not different (p>0.05).
Associations between the stimulation phase length (SPL) and number of follicles that developed to ≥ 6, 10 and 15 mm (A-C, respectively), serum estradiol concentrations (D) and the number of oocytes retrieved (E). The solid line on each graph represents the best fit regression line.
Associations between the stimulation phase length (SPL) and endometrial thickness (ET), fertilization rate (%), cleavage rate (%) and blastocyst rate (%). No regression lines are shown because no associations were detected.
Confounding variables for the associations between stimulation phase length (SPL) and IVF outcomes
|P-value||SE (β)||Estimate (β)||Model|
|Confounder=FSH start dose||-0.02||0.005||<0.0001|
|Confounder=FSH start dose||-0.02||0.004||0.0002|
|Confounder=FSH start dose||-0.008||0.003||0.01|
|Confounder=FSH start dose||-0.009||0.004||0.021|
|Confounder=GnRH agonist vs. antagonist||-3561||607.8||<0.0001|
The following variables were not found to be associated with SPL (SPL, SPL2; p>0.05): endometrial thickness, fertilization rate, cleavage rate, blastocyst rate, chemical pregnancy rate and clinical pregnancy rate. Results were obtained using multivariate polynomial linear regression analyses.
SPL = Stimulation phase length (linear relationship)
SPL2 = Stimulation phase length (parabolic relationship)
There was a greater incidence of non-smokers versus smokers within all 3 groups (p<0.05). The incidence of smokers was greater in women with a moderate versus a short or long SPL (p=0.04).
Negative parabolic associations were detected
between SPL and the number of follicles that developed to ≥ 6, 10 and 15 mm, number of oocytes
collected, and serum estradiol concentrations on the
day of hCG (
After adjusting for all potentially confounding
variables in our multivariate regression models, we
found that the number of follicles that developed to
≥ 6, 10 and 15 mm and number of oocytes collected
was negatively confounded by the FSH stimulation
start dose (p<0.05,
Associations between SPL and serum estradiol concentrations (D), stratified for GnRH agonist (A) versus antagonist (B) cycles. The solid line on each graph represents the best fit regression line.
It is not practical to monitor the growth of individual follicles that develop during ovarian stimulation, using current clinical assisted reproductive practices. We have conducted the present study to determine whether the duration of ovarian stimulation could be used as a reliable marker of mean ovarian follicular growth to predict clinical outcomes in couples undergoing IVF. The results of the present study have indicated that the length of the ovarian stimulation phase can be used to predict the number of estrogen-producing follicles that develop and oocytes retrieved prior to IVF. However, endometrial, fertilization, embryo and pregnancy outcomes were not influenced by the SPL. Thus, our hypotheses were only partially supported.
Increasing the FSH dose in patients whose follicles are growing slowly or decreasing the FSH dose in patients whose follicles are developing very quickly to achieve a SPL of approximately 11 days should therefore optimize follicular and oocyte outcomes, but not increase the probability of fertilization, high quality embryos or pregnancy. Collectively, we interpret these data to mean the SPL is not a reliable indicator of IVF success.
The effects of age on menstrual cycle length, conception and delivery rates are well-documented; however, potential age-related changes in the length of the SPL are not fully understood. We did not find that age had an effect on the SPL nor did it appear to confound the association between SPL and study outcomes. Thus, our hypothesis that women of advanced age would exhibit shorter SPLs and poorer clinical outcomes was not supported. Data on follicular phase length during previous natural cycles were not available in the present study. It is plausible that an age-related shortening of the follicular phase during the natural menstrual cycle does not necessarily correspond to a decreased SPL during ovarian stimulation. Continued research should be conducted to test this hypothesis.
In addition to age, we studied the potentially confounding effects of the ovarian stimulation hormonal treatment regimen on the association between SPL and IVF outcomes. A high FSH start dose was associated with an overall reduction in the number of follicles that developed. However, the pattern of change in follicle number in relation to SPL did not differ with different FSH start doses. The negative association between the FSH start dose and the number of follicles that developed was attributed to a higher dose of FSH that may have been administered to women expected to have a poor response to stimulatory therapy (even though women with a known history of a poor response were excluded from our analyses).
In addition, we found that the use of GnRH antagonist
was strongly associated with lower overall
estradiol concentrations throughout the stimulation
phase compared to the use of agonists. The pattern of
change in estradiol concentrations in relation to SPL,
however, was not different in agonist versus antagonist
cycles. Our findings were consistent with previous
research that demonstrated lower serum estradiol
concentrations in women using GnRH antagonists
versus agonists during ovarian stimulation (
The ability to obtain more mature follicles and
oocytes by optimizing the SPL, but not more embryos
and pregnancies, supports the notion that follicle
quality is more important than follicle quantity
in predicting the probability of pregnancy following
IVF. These results are consistent with those from
a previous study in which the number of dominant
follicles during ovarian stimulation was not associated
with pregnancy success (
A short or long ovarian stimulation phase length is associated with a suboptimal number of follicles that develop, serum estradiol concentrations and number of oocytes retrieved in couples undergoing assisted reproduction. However, the length of the stimulation phase does not predict embryo development or pregnancy outcomes.