Human Reproduction, Vol.28, No.11 pp. 3000–3006, 2013
Advanced Access publication on September 5, 2013 doi:10.1093/humrep/det344
ORIGINAL ARTICLE Infertility
Uterine length and fertility outcomes: a
cohort study in the IVF population
L.K. Hawkins 1,2,*, K.F. Correia2, S.S. Srouji 2, M.D. Hornstein 2,
and S.A. Missmer 2
1
Department of Obstetrics and Gynecology, Northwestern University, Chicago, IL, USA 2Department of Obstetrics, Gynecology and
Reproductive Biology, Brigham and Women’s Hospital and Harvard Medical School, Boston, MA, USA
*Correspondence address. E-mail: [email protected]
Submitted on May 13, 2013; resubmitted on June 26, 2013; accepted on July 18, 2013
study question: What is the relationship between pre-cycle uterine length and IVF outcome (chemical pregnancy, clinical pregnancy,
spontaneous abortion and live birth)?
summary answer: Women at extremes of uterine length (,7.0 or .9.0 cm) were less likely to achieve live birth and women with
uterine lengths ,6.0 cm were also more likely to experience spontaneous abortion.
what is known already: A prospective study of 807 women published in 2000 found that implantation and clinical pregnancy rates
were highest in women with uterine lengths between 7.0 and 9.0 cm, though the difference was not significant. The relationship between pre-cycle
uterine length and live birth has not been evaluated.
study design, size, duration: A retrospective cohort study of all cycles performed after uterine length measurement at an academic
hospital IVF clinic from 2001 to 2012.
participants/materials, setting, methods: A total of 8981 fresh cycles were performed in 5120 adult women with normal
uterine anatomy. Women with uterine anomalies (unicornuate, bicornuate, septate or uterus exposed to diethylstilbestrol) were excluded and
women with fibroids were identified for subanalysis. Uterine length was measured by uterine sounding. Cycles were divided by uterine length into
groups: ,6.0 cm (very short, n ¼ 76), 6.0–6.9 cm (short, n ¼ 2014), 7.0–7.9 cm (referent, n ¼ 4984), 8.0 –8.9 cm (long, n ¼ 1664) and ≥9 cm
(very long, n ¼ 243). Multivariate logistic regression (first-cycle analyses) and generalized estimating equations (all-cycle analyses) were adjusted
for age, fibroids and ART treatment (assisted hatching, intracytoplasmic sperm injection) to generate relative risk (RR) of cycle outcomes by
uterine length.
main results and the role of chance: Median uterine length in the IVF population was 7.0 cm (interquartile range 7.0 –7.8) and
was positively associated with BMI (P , 0.001) and fibroids (P ¼ 0.02). Compared with the referent group, women with uterine lengths ,6.0 cm
were half as likely to achieve live birth (RR: 0.53; 95% confidence interval (CI): 0.35– 0.81) and women with lengths of 6.0–6.9 cm were also less
likely (RR: 0.91; CI: 0.85 –0.98). Cubic regression spline identified a significant inverse U-shaped association whereby women with uterine lengths
,7.0 or .9.0 cm were less likely to achieve live birth. Women with lengths ,6.0 cm were also more likely to experience spontaneous abortion
(RR: 2.16; CI: 1.23– 3.78). Results remained consistent when excluding women with a uterine factor diagnosis (n ¼ 8823), when limiting to the
first cycle at our institution (n ¼ 5120) and when further restricting to first-ever cycles (n ¼ 3941).
limitations, reasons for caution: Optimal assessment of uterine length by ultrasound was not feasible due to time and
cost limitations, though uterine sounding is a clinically relevant measurement allowing for results with practical implications. Findings from our
predominantly Caucasian clinic population may not be generalizable to infertile populations with different ethnic compositions.
wider implications of the findings: Reproducibility of results would solidify findings and inform patient counseling in women
undergoing IVF.
study funding/competing interest(s): No funding was sought for this investigation. MD declares relationships with
UpToDate (royalties) and WINFertlity (consultant).
Key words: IVF/ICSI outcome / embryo transfer / uterus / cohort study / infertility
& The Author 2013. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.
For Permissions, please email: [email protected]
3001
Uterine length and IVF outcomes
Introduction
Anatomic and functional abnormalities of the uterus are associated with
adverse fertility outcomes. These include uterine anomalies (Ozgur et al.,
2007; Mollo et al., 2009), cavity distortions affected by submucosal or
large (.4 cm) intramural fibroids (Farhi et al., 1995; Jun et al., 2001;
Somigliana et al., 2011) and defects of the endometrium (Revel, 2012).
In women without uterine abnormalities, little is known about what
uterine factors, including uterine length, are predictive of fertility outcomes. Defining the relationship between fertility outcomes and
pre-cycle uterine length would assist in counseling and management
of women undergoing assisted reproduction technology (ART)
treatments.
Studies of the relationship between uterine length and fertility outcomes in both the ART and general population are few. In the ART population, a prospective study by Egbase and colleagues found that women
with a short (,7.0 cm) uterine length experienced a higher rate of
ectopic pregnancy (Egbase et al., 2000). In the general population, crosssectional studies paint an incomplete picture: while parous women have
a longer average uterine length than nulliparous women (Platt et al., 1990;
Esmaelzadeh et al., 2004; Canteiro et al., 2010; Langer et al., 2012),
uterine involution studies in parous women fail to demonstrate a paritydependent difference outside the puerperal period (de Castro, 1988;
Negishi et al., 1999; Olayemi et al., 2002). The question remains
whether uterine length is associated with fertility potential. Routine
pre-IVF cycle uterine length measurements allow for the use of prospectively gathered data to assess uterine length and IVF outcomes.
We hypothesize that uterine length is not associated with the incidence of chemical pregnancy, clinical pregnancy, spontaneous abortion
and live birth in women with normal uterine anatomy undergoing IVF.
Methods
Study population
All women presenting to the Center for Infertility and Reproductive Surgery
at Brigham and Women’s Hospital, Boston, MA, for pre-IVF uterine length
measurement between 2001 and 2011 were screened for inclusion in this
retrospective cohort study. All women ages 21 – 45 years old presenting
for a uterine length measurement prior to fresh embryo transfer were eligible
for inclusion. Exclusion criteria included women with uterine anomalies (unicornuate, bicornuate, septate uterus) and those with potentially altered
anatomy due to history of diethylstilbestrol exposure, peritoneal chemotherapy or radiation and uterine perforation or rupture. Gestational carrier cycles
and egg donor cycles were excluded and cycles not resulting in fresh embryo
transfer were excluded to standardize time between uterine length measurement and IVF outcome. Women with balanced translocations and Turner’s
mosaics were excluded due to concern for a higher spontaneous abortion
rate in this population. The protocol was approved by the Partners Human
Research Committees. HIPAA waiver of authorization was accepted in
place of written informed consent.
Out of concern for influence on cycle outcome, women with abnormal
uterine physiology and those at higher risk for recurrent pregnancy loss
were identified and only included after demonstrating homogeneity.
Women considered as potentially having abnormal uterine physiology
included women with any fibroids, those with clinically significant (defined
as submucosal or large intramural) fibroids and women with adenomyosis
as documented in the cycle-specific fertility treatment record. A ‘concern
for recurrent spontaneous abortion (RSAB) category’ was developed to
represent women in whom a preexisting diagnosis may systematically
impact cycle outcomes and included women with a positive RSAB workup
(prothrombin, antithrombin, homocysteine or Factor VLeiden mutations,
Protein C or S deficiencies, anti-cardiolipin antibodies, lupus anticoagulant),
women with a history of hypothyroidism and women with Asherman’s
syndrome (Ford and Schust, 2009).
Data collection
Information on infertility diagnoses, cycle-specific uterine length and cycle
outcomes was abstracted from two clinical databases. Missing information
was obtained by individual query of online medical records. A woman was
considered as having a specific infertility diagnosis if it was one of the cyclespecific, provider-recorded infertility diagnoses. Providers included attending
physicians, fellows and nurses.
Uterine Length measurement
Uterine sounding was performed in the proliferative phase (Days 5 – 12) of
the cycle antecedent to IVF treatment. It is practice in our clinic for
embryo transfer to be performed within 1 month of uterine assessment.
Uterine length measurements were performed according to standard procedure at our clinic (Schoolcraft et al., 2001; Coroleu et al., 2002). After
placement in the dorsal lithotomy position, speculum examination permitted
visualization of the cervix. Blood or mucous surrounding the os was removed
with cotton swabs. The outer sheath of the Rocket Duo catheter (Rocket
Medical plc, Hingham, USA) was inserted through the cervix for measurement of uterine length, defined as the distance from the external cervical
os to the uterine fundus (Rocket Medical plc 09/09/2011). For statistical analysis, cycles were grouped by initial uterine length: ,6.0 cm (very short, n ¼
76), 6.0– 6.9 cm (short, n ¼ 2014), 7.0 – 7.9 cm (referent, n ¼ 4984), 8.0 –
8.9 cm (long, n ¼ 1664), and ≥9 cm (very long, n ¼ 243).
Outcome measures
Chemical pregnancy was defined as cycles resulting in serum hCG ≥1 but
without subsequent development of a gestational sac. Clinical pregnancy
denoted cycles resulting in ultrasound confirmation of an intrauterine gestational sac but not fetal cardiac activity and/or followed by the passage of
pathology-confirmed fetal tissue prior to the development of fetal cardiac activity. Spontaneous abortion was defined as cycles ending in the passage of
pathology-confirmed fetal tissue or loss after documented fetal cardiac activity. Live birth denoted cycles resulting in live birth.
Statistical analyses
Analyses were performed using Statistical Analysis Software 9.2 (SAS Institute Inc., Cary, NC, USA). Continuous variables were presented as mean
(+SD) or median (quartile 1 – quartile 3). Spearmen rank correlations and
simple linear regression models were used to identify potential confounders,
which were further evaluated in multivariate-adjusted linear regression
models. Multivariate log binomial regression was used to evaluate cycle outcomes in the first cycle analysis. Generalized estimating equations, applying a
binomial distribution, log link and compound symmetry covariance structure,
were used to compare cycle outcomes in the instance of multiple cycles per
woman. All models are adjusted for age, fibroids, assisted hatching (AH) and
intracytoplasmic sperm injection (ICSI). Graphs were produced using cubic
regression splines with knots specified at uterine length cut points of 6.0,
7.0, 8.0 and 9.0 cm. Statistical significance was assumed for P , 0.05.
To neither bias toward nor against those who achieved success in early
cycles, all cycles were included in the primary analysis, which specifically
accounted for the likelihood of treatment cessation based on outcome and
the correlation between cycles in one woman (Missmer et al., 2011). To
address potential heterogeneity between subjects whereby women with
3002
Hawkins et al.
more than one cycle may experience suboptimal outcomes, additional analyses restricting to one cycle per woman and further restricting to first-ever
cycles were performed.
while mean BMI, mean height and prevalence of fibroids and AfricanAmerican ethnicity increased across exposure groups (Table I).
Distribution of uterine length in the IVF
population
Results
The final cohort included 8981 cycles from 5120 adult women. The majority of patients were Caucasian (85.4%) with a mean age of 36.7 + 4.0
years and mean BMI of 25.7 + 6.4 kg/m2. Age, infertility diagnoses and
stimulation protocols were similarly distributed across exposure groups
Median uterine length at the time of pre-cycle uterine measurement was
7.00 cm (7.00–7.75) with a mean of 7.24 cm (+0.71) and range of 5.00–
12.00 cm. The majority (55.5%) of patients had uterine lengths between
7.00 and 7.99 cm, while about a fourth (23.3%) had uterine lengths
,7.00 cm and about a fifth (21.2%) had lengths ≥8.00 cm (Fig. 1).
Table I Demographic and cycle characteristics for 8981 embryo transfers at a university-affiliated hospital.
Uterine length (cm)
......................................................................................................................................................
Characteristic
<6.00, n 5 76
(0.8%)
6.00– 6.99, n 5 2014
(22.4%)
7.00–7.99, n 5 4984
(55.5%)
8.00–8.99, n 5 1664
(18.5%)
9.001, n 5 243
(2.7%)
.............................................................................................................................................................................................
Age at embryo transfer
(years)
35.8 + 3.4
36.3 + 4.1
36.6 + 4.1
37.1 + 3.9
37.7 + 3.6
BMI (kg/m2)
22.7 + 2.9
24.5 + 4.8
25.6 + 6.5
27.4 + 7.2
29.6 + 7.5
Height (in.)
63.2 + 2.5
64.2 + 2.7
64.7 + 2.6
64.9 + 2.5
65.7 + 2.9
Race
White
32 (78.0%)
1057 (85.6%)
2614 (85.4%)
848 (84.5%)
99 (76.2%)
African American
1 (2.4%)
32 (2.6%)
108 (3.5%)
57 (5.7%)
15 (11.5%)
Asian
5 (12.2%)
94 (7.6%)
239 (7.8%)
67 (6.7%)
13 (10.0%)
Other
3 (7.3%)
52 (4.2%)
99 (3.2%)
31 (3.1%)
3 (2.3%)
13 (18.3%)
670 (34.8%)
1579 (33.3%)
497 (31.6%)
59 (26.1%)
Infertility diagnosisa
Male factor
Ovulation dysfunction
10 (14.1%)
208 (10.8%)
456 (9.6%)
175 (11.1%)
35 (15.5%)
Decreased ovarian reserve
10 (14.1%)
219 (11.4%)
484 (10.2%)
129 (8.2%)
19 (8.4%)
Endometriosis
10 (14.1%)
228 (11.8%)
533 (11.2%)
172 (10.9%)
27 (11.9%)
Tubal factor
11 (15.5%)
284 (14.7%)
752 (15.8%)
253 (16.1%)
43 (19.0%)
Uterine factor
2 (2.8%)
31 (1.6%)
86 (1.8%)
39 (2.5%)
10 (4.4%)
Unexplained
18 (25.4%)
480 (24.9%)
1288 (27.1%)
422 (26.8%)
39 (17.3%)
Fibroids
5 (6.6%)
106 (5.3%)
358 (7.2%)
191 (11.5%)
40 (16.5%)
Clinically relevant fibroidsb
0 (0%)
12 (0.6%)
41 (0.8%)
33 (2.0%)
5 (2.1%)
Adenomyosis
0 (0%)
12 (0.6%)
13 (0.3%)
4 (0.2%)
6 (2.5%)
RSAB
7 (9.2%)
184 (9.1%)
403 (8.1%)
146 (8.8%)
29 (11.9%)
Pre-cycle uterine length
5.5 + 0.3
6.4 + 0.3
7.2 + 0.3
8.1 + 0.2
9.3 + 0.6
Down-regulation
50 (65.8%)
1215 (60.3%)
3007 (60.3%)
1025 (61.6%)
127 (52.3%)
Antagonist
11 (14.5%)
334 (16.6%)
855 (17.2%)
271 (16.3%)
53 (21.8%)
Poor responder
15 (19.7%)
453 (22.5%)
1103 (22.1%)
361 (21.7%)
62 (25.5%)
Other/missing
0 (0%)
12 (0.6%)
19 (0.4%)
7 (0.4%)
1 (0.4%)
Stimulation typec
Total FSH dose (IU)
3948 + 2252
3890 + 2186
4004 + 2135
4048 + 2058
407 + 2015
ICSI
20 (26.3%)
804 (39.9%)
2058 (41.3%)
647 (38.9%)
80 (32.9%)
Assisted hatching
34 (44.7%)
1130 (56.1%)
2790 (56.0%)
971 (58.4%)
163 (67.1%)
Embryos transferred (#)
2.5 + 1.1
2.8 + 1.5
2.9 + 1.5
3.0 + 1.6
3.0 + 1.7
Data are presented as mean + SD or n (%).
RSAB, recurrent spontaneous abortion.
a
Column total may exceed 100% due to patients with multiple diagnoses.
b
Clinically relevant fibroids ¼ submucosal or large (≥4 cm) intramural fibroids.
c
Down-regulation ¼ long luteal GnRH agonist protocols; antagonist ¼ antagonist protocols, poor responder ¼ flare or estradiol patch priming protocols.
3003
Uterine length and IVF outcomes
and height were not significantly associated with uterine length after
adjusting for fibroids and weight.
Discussion
Figure 1. Distribution of pre-cycle uterine length measured by
uterine sounding in 8981 IVF cycles at a university-affiliated hospital.
Comparison of IVF outcomes
When compared with women with uterine lengths of 7.00 –7.99 cm
(referent), women with very short (,6.00 cm) uterine lengths were
half as likely to experience live birth (RR: 0.53, 95% CI: 0.35 –0.81)
and women with short (6.00–6.99 cm) uterine lengths were also less
likely (RR: 0.91, 95% CI: 0.85 –0.98). Women with very long
(.9.0 cm) uterine lengths were less likely to experience live birth,
though not significantly less likely when compared with the referent
group (Table II). Women with short uterine lengths (,6.00 cm) were
also more likely to experience spontaneous abortion compared with
the referent group (RR: 2.16, 95% CI: 1.23 –3.78). When assessing firstever cycles only, women with uterine lengths ,6.00 cm experienced a
higher likelihood of chemical pregnancy (Table II). Results remained consistent when assessing one cycle per woman (n ¼ 5120) and when excluding women with a uterine factor diagnosis (n ¼ 8823).
Cubic regression spline models a significant inverse U-shaped association between uterine length and predicted probability of live birth
whereby women with uterine lengths ,7.00 cm or .9.00 cm had a
lower predicted probability of live birth while women with a uterine
length of 8.00 cm experienced the highest predicted probability of
live birth (Fig. 2).
Assessment of covariates: BMI, fibroids
and ethnicity
In a multivariate model adjusting for age, ethnicity and fibroids, obese
(BMI ≥30 kg/m2) women had a longer average uterine length than
normal weight (BMI 20–24.9 kg/m2) women (7.65 cm + 0.06 and
7.29 cm + 0.05, respectively, P , 0.001). When including adjustment
for height, uterine length increased across increasing quartiles of
weight (Q1, ≤56.8 kg: 7.19 cm (referent); Q2, 57.3–63.6 kg:
7.32 cm, P ¼ 0.008; Q3, 64.1– 75.5 kg: 7.36 cm, P , 0.001; Q4,
.75.5 kg: 7.62 cm, P , 0.001) and increasing percentiles of weight
(10th percentile, ,52.3 kg: 7.13 cm, P , 0.001; 10–90th percentile,
52.3 –93.6 kg: 7.38 cm (referent); 90th percentile, .93.6 kg: 7.77 cm,
P , 0.001). In all multivariate-adjusted models, women with fibroids
had a longer average uterine length than women without fibroids
(7.45 cm + 0.07 and 7.28 cm + 0.04, respectively, P ¼ 0.02). Ethnicity
The major findings of this investigation demonstrate that women at
extremes of uterine length (,7.00 cm and .9.00 cm) were less likely
to achieve live birth after IVF. Compared with the referent group (precycle uterine lengths of 7.00–7.99 cm), women with short lengths
(,7.00 cm) were less likely to achieve live birth and those with very
short lengths (,6.00 cm) were more likely to experience spontaneous
abortion. BMI, weight and fibroids were positively associated with uterine
length.
Although significant cubic regression spline demonstrated that women
with very long (.9.00 cm) uterine length experienced a lower likelihood
of live birth, this difference was not significant when compared with the
referent group. This could be related to our designation of women with
the most common pre-cycle uterine length (7.00–7.99 cm) as referent.
The regression spline demonstrates that the uterine length conferring the
highest probability of live birth is closer to 8.0 cm (7.75–8.25 cm) which
might have been a more appropriate referent group. Our designation of a
referent group with a suboptimal likelihood of live birth might have biased
results toward the null.
Potential underlying mechanisms for a lower likelihood of live birth in
women at extremes of uterine length are multiple and include anatomical
and hormonal explanations. Though women with perceptible uterine
anomalies were excluded, extremes of uterine length may signify anatomical variations that contribute a decreased likelihood of implantation.
An anatomical argument is supported by prospective data showing
women with uterine lengths ,7.0 cm experience higher rates of
ectopic pregnancy (Egbase et al., 2000). Research documenting a positive
association between uterine size and estrone concentration (Chua et al.,
1991; Ciobanu et al., 2003; Dandolu et al., 2010) implies extremes of
uterine length might represent estrone deficiency or excess that adversely impacts cycle success (e.g. local or systemic aberrations in hormonal
signaling might contribute to dysynchronous maturation of the endometrium). In women with very long uterine lengths, it is also possible a large
uterus serves as an intermediate in the pathway between obesity and
reduced cycle success (Shah et al., 2011).
The finding that women with uterine lengths ,6.00 cm experienced a
higher likelihood of chemical pregnancy when assessing initial but not all
cycles may denote inadequate uterine receptivity in women with a small
uterus that is overcome with different, prolonged or simply repeated
uterine preparation, though a small sample size in this exposure group
obscures substantive discussion.
The association of BMI and weight with uterine length is consistent
with the hyperestrogenic environment of obesity in which estrone acts
as a growth factor (Chua et al., 1991; Ciobanu et al., 2003; Dandolu
et al., 2010). In the present work, incremental increase in average
uterine length across increasing percentiles of weight suggests that
weight is associated with uterine length in a dose-dependent fashion
(7.13 cm (10th percentile of BMI), 7.19 cm (25th percentile), 7.62 cm
(75th percentile) and 7.77 cm (90th percentile)).
The association of fibroids and uterine length is consistent with cavity
distortion from submucosal or large (≥4 cm) intramural fibroids (Jun
et al., 2001; Oliveira et al., 2004). Although infertile women with cavitydistorting fibroids are typically treated prior to IVF, it is possible our
3004
Hawkins et al.
Table II Pre-cycle uterine length and IVF outcomes following embryo transfers at a university-affiliated hospital.
Uterine length (cm)
...............................................................................................................................................
<6.00
6.00– 6.99
7.00– 7.99
8.00–8.99
9.001
.............................................................................................................................................................................................
Chemical pregnancy only
First cycle, n ¼ 3941
RR (95% CI)a
All cycles, n ¼ 8981
RR (95% CI)b
7 (18.9%)
2.28 (1.14– 4.57)
8 (10.5%)
1.32 (0.68– 2.56)
82 (9.0%)
1.06 (0.83– 1.36)
183 (8.5%)
1.00 (referent)
54 (7.6%)
0.88 (0.66– 1.18)
175 (8.7%)
412 (8.3%)
129 (7.8%)
1.05 (0.88– 1.25)
1.00 (referent)
0.92 (0.76– 1.12)
7 (6.3%)
0.70 (0.34–1.46)
16 (6.6%)
0.77 (0.48–1.24)
Clinical pregnancy only
First cycle, n ¼ 3941
RR (95% CI)a
All cycles, n ¼ 8981
b
RR (95% CI)
17 (45.9%)
0.93 (0.65– 1.33)
29 (38.2%)
0.79 (0.58– 1.06)
397 (43.6%)
996 (46.0%)
331 (46.5%)
0.94 (0.86– 1.02)
1.00 (referent)
1.03 (0.94– 1.13)
868 (43.1%)
2224 (44.6%)
760 (45.7%)
0.95 (0.90– 1.01)
1.00 (referent)
1.04 (0.98– 1.11)
47 (42.0%)
0.98 (0.79–1.22)
94 (38.7%)
0.92 (0.79–1.08)
Spontaneous abortion
First cycle, n ¼ 3941
RR (95% CI)a
All cycles, n ¼ 8981
RR (95% CI)b
8 (21.6%)
3.16 (1.68– 5.95)
12 (15.8%)
2.16 (1.23– 3.78)
76 (8.4%)
1.13 (0.87– 1.46)
162 (7.5%)
1.00 (referent)
53 (7.5%)
0.97 (0.72– 1.30)
187 (9.3%)
399 (8.0%)
131 (7.9%)
1.18 (0.99– 1.39)
1.00 (referent)
0.97 (0.79– 1.18)
316 (34.8%)
820 (38.0%)
273 (38.4%)
0.91 (0.83– 1.01)
1.00 (referent)
1.04 (0.94– 1.16)
668 (33.3%)
1776 (35.8%)
613 (37.0%)
0.91 (0.85– 0.98)
1.00 (referent)
1.06 (0.99– 1.14)
9 (8.0%)
1.00 (0.52–1.90)
21 (8.6%)
1.01 (0.65–1.58)
Live birthc
First cycles only, n ¼ 3941
a
RR (95% CI)
All cycles, n ¼ 8981
RR (95% CI)b
8 (21.6%)
0.51 (0.28– 0.96)
16 (21.1%)
0.53 (0.35– 0.81)
38 (33.9%)
0.99 (0.77–1.28)
72 (29.6%)
0.91 (0.75–1.11)
a
First-ever cycles only. RR calculated from log binomial model adjusted for age, fibroids, AH (assisted hatching) and ICSI.
All cycles. RR calculated from generalized estimating equations to account for multiple cycles per woman of 8981 embryo transfers adjusted for age, fibroids, AH and ICSI.
Due to missing follow-up information, 36 cycles were not included in the birth analysis.
b
c
cohort includes women who did not receive pre-IVF myomectomy for
reasons such as advanced maternal age. Notwithstanding, the statistical
difference between a mean uterine length of 7.45 cm in women with
fibroids and 7.28 cm in women without fibroids is unlikely to be clinically
significant. Further, effect estimates were appreciably unchanged by adjustment for fibroids such that fibroids did not confound the association
of uterine length and fertility outcomes in our models.
This study is the first to evaluate the association of pre-IVF cycle
uterine length with live birth. In 2000, Egbase and colleagues conducted
a prospective study comparing rates of implantation, clinical pregnancy
and ectopic pregnancy by pre-IVF uterine length among 807 women
and found that women with uterine lengths ,7 cm were more likely
to experience ectopic pregnancy compared with women with uterine
lengths of 7–9 cm (14.9 and 1.8%, respectively, P,0.001) and, though
not significantly different, rates of clinical pregnancy were highest
(46.5%) in women with uterine lengths of 7–9 cm compared with
those in women with lengths ,7 cm (36.7%) or .9 cm (32.9%)
(Egbase et al., 2000). Their findings are in keeping with ours, though
we compared more narrow exposure categories, assessed incidence
of live birth and did not specifically evaluate ectopic pregnancy incidence.
Strengths of this investigation include a clinic population-based study
population, isolation of women with anatomically normal uteri and assessment of live birth incidence. A sample size of 8981 cycles powered
comparison of multiple IVF outcomes across five categories of uterine
length and permitted more resolute assessment than prior studies.
Access to complete records allowed for assessment of the impact of
fibroids, ethnicity and BMI on the relationship between uterine length
and IVF outcomes. Variability in measurement was minimized by use of
an identical catheter for all measurements and a standardized measurement technique.
The crude nature of uterine sounding limits our results. Though
uterine sounding is a clinically relevant measurement that allows for
results with practical implications, 3D saline infusion sonohysterography would provide optimal assessment of uterine measurements
(Wachsberg et al., 1994). Inter-observer variability from the recording
of measurements by multiple providers may have introduced bias.
Although the measurement technique is standardized, provider-specific
differences in technique may impact our findings, as they have been
shown to impact outcomes after embryo transfer (Hearns-Stokes
et al., 2000). The catheter used for uterine sounding is low resolution
with 1 cm markings that require estimation of in-between lengths. Additionally, uterine length (from external os to uterine fundus) is a nonspecific measure composed of endometrial length (from internal os to
uterine fundus) and cervical length (from internal to external os). If the
ratio of these components varies as a function of uterine length, this
could conceivably cause confounding, particularly given the well-studied
association between a short cervix and adverse pregnancy outcomes
(Berghella et al., 2011; Owen and Mancuso, 2012). Information on
BMI was only available for 25.6% of the cohort such that we chose not
to sacrifice power for inclusion of BMI in our final model despite its
3005
Uterine length and IVF outcomes
insight into the biological plausibility of study findings and contributed
critical revisions. M.D.H. contributed to study conception, data interpretation and provided critical revisions. S.A.M. led data analysis and interpretation and provided critical revisions and guidance.
Funding
No external funding was sought for this study.
Conflict of interest
M.D.H. discloses relationships with UpToDate (royalties) and WINFertilility (paid consultant). All other authors have no conflicts of interest
associated with this study.
References
Figure 2. Cubic regression spline of predicted probabilities of live
birth by pre-cycle uterine length in women less than 35 years old
without fibroids undergoing autologous IVF cycle at a university-affiliated
hospital. Dashed lines represent 95% confidence intervals.
potential as a confounder of uterine length and fertility outcomes (Shah
et al., 2011). Results are limited by incomplete information on gravidity
and parity at the time of pre-cycle uterine length measurement. Across
a 12-year study period, gravidity and parity were not reliably recorded
in the query-able clinical databases, and individual medical record
review for some 8000 cycles was impractical. The study population,
while representative of the demographic composition at our clinic, was
predominantly Caucasian such that results may not be generalizable to
IVF clinic populations with other ethnic compositions.
In summary, this study demonstrates that women with short uterine
lengths have lower live birth rates and higher rates of spontaneous abortion. Median uterine length in our IVF population is 7.00 cm (7.00–7.75)
and is positively associated with BMI and fibroids. Women at extremes of
uterine length experience lower likelihood of live birth after IVF such that
women with the highest probability of live birth have a pre-cycle uterine
length between 7.00 and 9.00 cm. Reproducibility of results would solidify findings and inform patient counseling in women undergoing IVF.
Authors’ roles
L.K.H. and S.A.M. conceived this research question and designed the investigation. L.K.H. initiated this project, led data collection, contributed
to data analysis and interpretation and wrote this manuscript. K.F.C. contributed to data collection, performed data analysis and contributed substantially to data interpretation and manuscript revisions. S.S.S. provided
Berghella V, Rafael TJ, Szychowski JM, Rust OA, Owen J. Cerclage for short
cervix on ultrasonography in women with singleton gestations and
previous preterm birth: a meta-analysis. Obstet Gynecol 2011;117:663–671.
Canteiro R, Bahamondes MV, dos Santos Fernandes A, Espejo-Arce X,
Marchi NM, Bahamondes L. Length of the endometrial cavity as measured
by uterine sounding and ultrasonography in women of different parities.
Contraception 2010;81:515–519.
Chua BH, Chua CC, Zhao ZY, Krebs CJ. Estrone modulates the EGF
receptor in the liver of db/db mouse. J Recept Res 1991;11:941– 957.
Ciobanu LC, Luu-The V, Martel C, Labrie F, Poirier D. Inhibition of estrone
sulfate-induced uterine growth by potent nonestrogenic steroidal
inhibitors of steroid sulfatase. Cancer Res 2003;63:6442 – 6446.
Coroleu B, Barri PN, Carreras O, Martinez F, Parriego M, Hereter L,
Parera N, Veiga A, Balasch J. The influence of the depth of embryo
replacement into the uterine cavity on implantation rates after IVF: a
controlled, ultrasound-guided study. Hum Reprod 2002;17:341 – 346.
Dandolu V, Singh R, Lidicker J, Harmanli O. BMI and uterine size: is there any
relationship? Int J Gynecol Pathol 2010;29:568 – 571.
de Castro A. Assessment of uterine shape and size using Kurz’s Cavimeter.
Adv Contracept 1988;4:115 –123.
Egbase PE, Al-Sharhan M, Grudzinskas JG. Influence of position and length of
uterus on implantation and clinical pregnancy rates in IVF and embryo
transfer treatment cycles. Hum Reprod 2000;15:1943 – 1946.
Esmaelzadeh S, Rezaei N, HajiAhmadi M. Normal uterine size in women of
reproductive age in northern Islamic Republic of Iran. East Mediterr
Health J 2004;10:437– 441.
Farhi J, Ashkenazi J, Feldberg D, Dicker D, Orvieto R, Ben Rafael Z. Effect of
uterine leiomyomata on the results of in-vitro fertilization treatment. Hum
Reprod 1995;10:2576 – 2578.
Ford HB, Schust DJ. Recurrent pregnancy loss: etiology, diagnosis, and
therapy. Rev Obstet Gynecol 2009;2:76 – 83.
Hearns-Stokes RM, Miller BT, Scott L, Creuss D, Chakraborty PK, Segars JH.
Pregnancy rates after embryo transfer depend on the provider at embryo
transfer. Fertil Steril 2000;74:80– 86.
Jun SH, Ginsburg ES, Racowsky C, Wise LA, Hornstein MD. Uterine leiomyomas
and their effect on in vitro fertilization outcome: a retrospective study. J Assist
Reprod Genet 2001;18:139–143.
Langer JE, Oliver ER, Lev-Toaff AS, Coleman BG. Imaging of the female pelvis
through the life cycle. Radiographics 2012;32:1575 – 1597.
Missmer SA, Pearson KR, Ryan LM, Meeker JD, Cramer DW, Hauser R.
Analysis of multiple-cycle data from couples undergoing in vitro
fertilization: methodologic issues and statistical approaches. Epidemiology
2011;22:497 – 504.
3006
Mollo A, De Franciscis P, Colacurci N, Cobellis L, Perino A, Venezia R, Alviggi C,
De Placido G. Hysteroscopic resection of the septum improves the pregnancy
rate of women with unexplained infertility: a prospective controlled trial. Fertil
Steril 2009;91:2628–2631.
Negishi H, Kishida T, Yamada H, Hirayama E, Mikuni M, Fujimoto S. Changes
in uterine size after vaginal delivery and cesarean section determined by
vaginal sonography in the puerperium. Arch Gynecol Obstet 1999;263:
13 – 16.
Olayemi O, Omigbodun AA, Obajimi MO, Odukogbe AA, Agunloye AM,
Aimakhu CO, Okunlola MA. Ultrasound assessment of the effect of
parity on postpartum uterine involution. J Obstet Gynaecol 2002;22:
381 – 384.
Oliveira FG, Abdelmassih VG, Diamond MP, Dozortsev D, Melo NR,
Abdelmassih R. Impact of subserosal and intramural uterine fibroids that
do not distort the endometrial cavity on the outcome of in vitro
fertilization-intracytoplasmic sperm injection. Fertil Steril 2004;81:582–587.
Owen J, Mancuso M. Cervical cerclage for the prevention of preterm birth.
Obstet Gynecol Clin North Am 2012;39:25– 33.
Hawkins et al.
Ozgur K, Isikoglu M, Donmez L, Oehninger S. Is hysteroscopic correction of
an incomplete uterine septum justified prior to IVF? Reprod Biomed Online
2007;14:335 – 340.
Platt JF, Bree RL, Davidson D. Ultrasound of the normal nongravid uterus:
correlation with gross and histopathology. J Clin Ultrasound 1990;18:15–19.
Revel A. Defective endometrial receptivity. Fertil Steril 2012;97:1028– 1032.
Rocket Medical plc. Rocket DUO IUI Catheters. 09/09/2011.
Schoolcraft WB, Surrey ES, Gardner DK. Embryo transfer: techniques and
variables affecting success. Fertil Steril 2001;76:863 – 870.
Shah DK, Missmer SA, Berry KF, Racowsky C, Ginsburg ES. Effect of obesity
on oocyte and embryo quality in women undergoing in vitro fertilization.
Obstet Gynecol 2011;118:63 – 70.
Somigliana E, De Benedictis S, Vercellini P, Nicolosi AE, Benaglia L, Scarduelli C,
Ragni G, Fedele L. Fibroids not encroaching the endometrial cavity and IVF
success rate: a prospective study. Hum Reprod 2011;26:834–839.
Wachsberg RH, Kurtz AB, Levine CD, Solomon P, Wapner RJ. Real-time
ultrasonographic analysis of the normal postpartum uterus: technique,
variability, and measurements. J Ultrasound Med 1994;13:215– 221.