Comparison of live-birth defects after

luteal-phase ovarian stimulation vs.

conventional ovarian stimulation for
in vitro fertilization and vitrified
embryo transfer cycles
Hong Chen, M.D.,a Yun Wang, M.D.,a Qifeng Lyu, Ph.D.,a Ai Ai, M.D.,a Yonglun Fu, M.D.,a Hui Tian, M.D.,a
Renfei Cai, M.D.,a Qingqing Hong, M.D.,a Qiuju Chen, Ph.D.,a Zeev Shoham, M.D.,b and Yanping Kuang, M.D.a
a
Department of Assisted Reproduction, Shanghai Ninth People's Hospital, Shanghai Jiaotong University School of
Medicine, Shanghai, People's Republic of China; and b Department of Obstetrics and Gynecology, Kaplan Medical
Center, Rehovot, Israel

Objective: To assess live-birth defects after a luteal-phase ovarian-stimulation regimen (LPS) for in vitro fertilization (IVF) and vitrified
embryo transfer (ET) cycles.
Design: Retrospective cohort study.
Setting: Tertiary-care academic medical center.
Patient(s): Infants who were born between January 1, 2013 and May 1, 2014 from IVF with intracytoplasmic sperm injection (ICSI)
treatments (n ¼ 2,060) after either LPS (n ¼ 587), the standard gonadotropin-releasing hormone–agonist (GnRH-a) short protocol
(n ¼ 1,257), or mild ovarian stimulation (n ¼ 216).
Intervention(s): The three ovarian-stimulation protocols described and assisted reproductive technology (ART) treatment (IVF or ICSI,
and vitrified ET) in ordinary practice.
Main Outcome Measure(s): The main measures were: gestational age, birth weight and length, multiple delivery, early neonatal mor-
tality, and birth defects. Associations were assessed using logistic regression by adjusting for confounding factors.
Result(s): The final sample included 2,060 live-born infants, corresponding to 1,622 frozen–thawed (FET) cycles, which led to: 587
live-born infants from LPS (458 FET cycles); 1,257 live-born infants from the short protocol (984 FET cycles); and 216 live-born
infants from mild ovarian stimulation (180 FET cycles). Birth characteristics regarding gestational age, birth weight and length,
multiple delivery, and early neonatal death were comparable in all groups. The incidence of live-birth defects among the LPS group
(1.02%) and the short GnRH-a protocol group (0.64%) was slightly higher than in the mild ovarian-stimulation group (0.46%).
However, none of these differences reached statistical significance. For congenital malformations, the risk significantly increased
for the infertility-duration factor and multiple births; the adjusted odds ratios were 1.161 (95% confidence interval [CI]: 1.009–
1.335) and 3.899 (95% CI: 1.179–12.896), respectively. No associations were found between congenital birth defects and various
ovarian-stimulation regimens, maternal age, body mass index, parity, insemination method, or infant gender.
Conclusion(s): To date, the data do not indicate an elevated rate of abnormality at birth after LPS, but further study with larger pop-
ulations is needed to confirm these results. However, infertility itself poses a risk factor for congenital malformation. A higher likelihood
of birth defects in multiple births may lead couples to favor elective, single ET; couples under-
taking ART should be made aware of the known increased birth defects associated with a twin birth.
(Fertil SterilÒ 2015;103:1194–201. Ó2015 by American Society for Reproductive Medicine.) Use your smartphone
Key Words: Luteal phase ovarian stimulation, in vitro fertilization, intracytoplasmic sperm to scan this QR code
injection, live birth, frozen embryo transfer, congenital malformation and connect to the
discussion forum for
this article now.*
Discuss: You can discuss this article with its authors and with other ASRM members at http://
fertstertforum.com/chenh-live-birth-defects-luteal-phase-ovarian-stimulation/ * Download a free QR code scanner by searching for “QR
scanner” in your smartphone’s app store or app marketplace.

Received December 27, 2014; revised January 28, 2015; accepted February 16, 2015; published online March 23, 2015.
H.C. has nothing to disclose. Y.W. has nothing to disclose. Q.L. has nothing to disclose. A.A. has nothing to disclose. Y.F. has nothing to disclose. H.T. has nothing
to disclose. R.C. has nothing to disclose. Q.H. has nothing to disclose. Q.C. has nothing to disclose. Z.S. has nothing to disclose. Y.K. has nothing to disclose.
H.C. and Y.W. should be considered similar in author order.
Supported in part by grants from the National Natural Foundation of People's Republic of China (81270749 to Y.K., 31071275 to Q.L. and 81470064 to Y.W.)
and the Natural Science Foundation of Shanghai, People's Republic of China (11411950105 to Y.K. and 14ZR1423900 to Y.W.).
Reprint requests: Yanping Kuang, M.D., Department of Assisted Reproduction, Shanghai Ninth People's Hospital, Zhizaoju Road No 639, Huangpu District,
Shanghai, People's Republic of China (E-mail: Kuangyanp@126.com).

Fertility and Sterility® Vol. 103, No. 5, May 2015 0015-0282/$36.00

Copyright ©2015 American Society for Reproductive Medicine, Published by Elsevier Inc.
http://dx.doi.org/10.1016/j.fertnstert.2015.02.020

1194 VOL. 103 NO. 5 / MAY 2015


ince the first live birth resulting from in vitro fertiliza-
S
tion (IVF) in the United Kingdom in 1978 (1), IVF has
become a major alternative in the treatment of infertile
couples. The number of infants born as a result of this tech-
nology has risen rapidly. According to a report from a 2009
survey on assisted reproductive technology (ART) in Europe,
a total of 399,020 ART cycles were performed in a population
of 373.8 million, corresponding to 1,067 cycles per million in-
habitants (2), and this proportion may be increasing every
year.
Numerous studies have explored the type and incidence
of ART-related side effects in women who use these fertility
services, and the potential impact on their offspring.
Although IVF complication rates seem to be low (3), concern
is still increasing about risks regarding this treatment proce-
dure, e.g., suboptimal oocyte or embryo quality induced by
premature luteinization (4), ovarian hyperstimulation
syndrome (OHSS), and thromboembolism after hormonal
stimulation, and infection and abdominal bleeding caused
by egg collection (3). Therefore, influential innovations in
stimulation regimens have been an integral part of IVF treat-
ment for the past 30 years. However, these innovations have
always limited ovarian stimulation initiation to the early
follicular phase of the menstrual cycle, which may have
elicited a premature luteinizing hormone (LH) surge, with
multiple follicular development stimulated by exogenous
gonadotropin. Consequently, the question of how to effec-
tively suppress premature LH surge has traditionally been a
researcher obsession.
To circumvent the premature LH surge obstacle, researchers
have designed numerous stimulation regimens, ranging from
pituitary down-regulation technology (beginning in the early
1980s) (5) to aggressive stimulation in combination with a
gonadotropin-releasing hormone (GnRH) antagonist (during
the past decade) (6). These solutions make stimulation complex
and make complications caused by OHSS, which vary from mild
to extremely severe illness, into rare but potentially life-
threatening conditions.
In 2009, a 40-year-old woman with a 10-year history of
primary infertility became pregnant with a twin pregnancy af-
ter accidental luteal-phase ovarian stimulation (LPS) at our
center, and had a favorable delivery (7). This case paved the
way to the possibility of successful outcomes with LPS. We
investigated 242 women who received LPS (Supplemental
Fig. 1, available online).
A satisfactory clinical pregnancy rate of 55.46% was ob-
tained, resulting in 68 live births and 44 ongoing pregnancies
at that time. Moreover, no premature LH surges occurred, nor
did moderate-to-severe OHSS during the stimulation. The
study proved that LPS is feasible for producing competent oo-
cytes and/or embryos with optimal pregnancy outcomes (8).
To date, this unique approach has resulted in hundreds of in-
fants born after embryo vitrification.
Accompanying the desire to continuously optimize
ovarian stimulation is a concern about whether LPS is safe
as a baby grows into childhood and adulthood. Today,
GnRH-a regimens using exogenous gonadotropins, which
were found to cause no greater incidence of bodily malforma-
tions than that expected in the general population (9), are still
VOL. 103 NO. 5 / MAY 2015
Fertility and Sterility®
the most frequently applied ovarian-stimulation approaches
worldwide. They appear to be safe for the offspring who orig-
inate from this regimen.
The mild ovarian-stimulation approach administers either
lower doses (over fewer days) of exogenous gonadotropins, or
other compounds (such as antiestrogens, aromatase inhibitors,
or GnRH-antagonist), to limit the number of oocytes obtained
for IVF to <7 (10). This regimen can reduce patient discomfort
and risk, and lower the duration and intensity of the pharma-
cologic interventions (11). Thus, mild stimulation may inter-
fere less with normal ovarian physiology and result in better
oocyte and/or embryo quality with reduced aneuploidy in
the human preimplantation embryo (12). In consideration of
their safety statistics for offspring, we selected the standard
GnRH-a short protocol, and the mild ovarian-stimulation
approach, to be used in the control groups, to evaluate the
safety for offspring born from IVF and vitrified embryo trans-
fer (ET) after LPS.
MATERIALS AND METHODS
Study Population and Design
A retrospective cohort study was conducted at the Depart-
ment of Assisted Reproduction of the Ninth People's Hospital
of Shanghai Jiao Tong University School of Medicine
(Shanghai, People's Republic of China). The study was
approved by the hospital's ethics committee. Infertile couples,
who underwent IVF or intracytoplasmic sperm injection (ICSI)
treatment with frozen and then thawed embryo transfer (FET)
using LPS, the standard GnRH-a short protocol, or the mild
ovarian-stimulation regimen, were recruited at our center.
Informed written consent was obtained from patients in
accordance with the ethics committee protocol.
These patients underwent the procedures from March 1,
2012 to July 1, 2013, leading to births between January 1,
2013 and May 1, 2014. Infants born to mothers with reported
maternal diseases, such as gestational diabetes mellitus (13,
14), hypertension (15), and thyroid disorders (16, 17), or
adverse environmental exposure (18–22) during pregnancy,
were excluded from this analysis because of the possible
association of these factors with birth defects. The final
data, involving 2,060 live-born infants, delivered after IVF
with ICSI and FET, were stratified into groups according to
the method of ovarian stimulation: 587 births after LPS,
1,257 births after the standard GnRH-a short protocol, and
216 births after mild ovarian stimulation. The study design
and participant selection procedure are presented in
Supplemental Figure 2 (available online).
Treatment
The details of the ovarian-stimulation regimen for LPS are
extensively described elsewhere (Supplemental Fig. 1) (8).
Briefly, ovarian stimulation was conducted for patients with
antral follicles <8 mm on days 1–3 after ovulation, by inject-
ing 225 IU of hMG (Anhui Fengyuan Pharmaceutical Co. Ltd)
and simultaneously administering 2.5 mg of letrozole
(Jiangsu Hengrui Medicine Co. Ltd) every day. Daily adminis-
tration of 10 mg of medroxyprogesterone acetate was added
1195
ORIGINAL ARTICLE: ASSISTED REPRODUCTION
beginning on stimulation day 12, for cases in which postovu-
lation follicle size was <14 mm in diameter and stimulation
needed to continue for several more days to avoid oocyte
retrieval during menstruation.
In the short GnRH-a group, ovarian stimulation was car-
ried out using a daily dose of 0.1 mg of triptorelin (Decapeptyl,
Ipsen) injected subcutaneously, starting on day 2 or 3 of the
natural cycle, and continuing until the trigger day. This treat-
ment was accompanied by intramuscular injection of 150 IU
or 225 IU of hMG, beginning on the same day as the first trip-
torelin adminisration. In the mild ovarian-stimulation group,
25 mg per day of clomiphene citrate (CC, Fertilan, Codal-
Synto Ltd) and 2.5 mg per day of letrozole (LE, Jiangsu
Hengrui Medicine Co.) were coadministered from cycle-day
3 onward, but letrozole was given for only 4 days, and CC
was continuously used until the trigger day. Patients started
to receive injections of 150 IU of hMG every other day, begin-
ning on cycle-day 6. The doses of hMG were adjusted after
7–8 days for LPS, 5–7 days for the short GnRH-a protocol,
and 4 days for the mild ovarian-stimulation protocol, accord-
ing to the patient's ovarian response.
Monitoring was performed using transvaginal ultrasound
scanning of the ovaries and serum estradiol measurements.
When 1–3 dominant follicles reached 18–20 mm in diameter,
the final stage of oocyte maturation was induced with 0.1 mg
of triptorelin for LPS and the mild ovarian-stimulation proto-
col, and with 5,000 IU of hCG (Zhuhai Livzon Pharmaceutical
Group Inc.) for the short GnRH-a protocol. A 0.6-g dose of
ibuprofen was used on the trigger day and the next day, for
LPS and the mild ovarian-stimulation protocol.
Oocyte retrieval was carried out 32–36 hours after matu-
ration induction or ovarian stimulation (based on each group
protocol). Conventional IVF was performed for women with
tubal or idiopathic infertility indications, and ICSI was per-
formed for male-factor indications. Embryos were cultured
and scored (23) as previously described. The freezing and
thawing procedure for embryos (24); the method of embryo
and endometrium synchronization during a natural cycle,
an induced ovulation cycle, or an artificial cycle; and the
timing of ET are described elsewhere (8). Once pregnancy
was achieved, the progesterone (P) supplement was continued
until 8 weeks of gestation.
Outcome Assessment
The couples completed a telephone interview during each
stage of pregnancy, up to 1 week after delivery. The interview
questionnaire included information about: a wide range of
preconception and pregnancy exposures; gestational weeks;
pregnancy complications: mode of delivery; birth date and lo-
cality; birth weight and length; infant gender; and neonatal
diseases, if any. Those whose response(s) were outside the
norm were questioned further about specific neonatal disease
indicators, including diagnosis, exact point in time of the
diagnosis, treatment, and prognosis.
For neonates born in our university hospitals, a detailed
physical examination and a routine ultrasound examination
of the brain, kidneys, and heart were done at birth. For neo-
nates born elsewhere, written reports were obtained from
1196
the pediatrician in charge. Pertinent information on the ques-
tionnaire was researched through family planning service
agencies when attempts to contact couples failed. For live-
born children with birth defects, case information was re-
viewed by a specially designated nurse, to ensure that infants
met the case definition of the Chinese Birth Defects Moni-
toring Program.
The outcomes were defined based on the International
Committee for Monitoring Assisted Reproductive Technology
and the World Health Organization revised glossary of ART
terminology 2009 (10). For example, ‘‘delivery’’ was defined
as the expulsion or extraction of R1 fetuses from the mother
after 20 completed weeks of gestational age. ‘‘Live birth’’ was
defined as the complete expulsion or extraction from its
mother of a product of fertilization, irrespective of the dura-
tion of the pregnancy, which, after such separation, breathes
or shows any other evidence of life such as heart beat, umbil-
ical cord pulsation, or definite movement of voluntary mus-
cles, irrespective of whether the umbilical cord has been cut
or the placenta is attached.
‘‘Gestational age’’ was defined as the age of an embryo or
fetus calculated by adding 2 weeks (14 days) to the number of
completed weeks since fertilization. (For FETs, the estimated
fertilization date was computed by subtracting the embryo
age on the date of cryopreservation from the FET-cycle trans-
fer date). ‘‘Preterm birth’’ was defined as a live birth or still-
birth that takes place after R20 but <37 completed weeks
of gestational age. ‘‘Postterm birth’’ was defined as a live birth
or stillbirth that takes place after 42 completed weeks of
gestational age. ‘‘Low birth weight’’ was defined as birth
weight <2,500 grams.
‘‘Multiple birth’’ was defined as a pregnancy and/or deliv-
ery with >1 fetus or neonate. ‘‘Early neonatal death’’ was
defined as the death of a live-born baby within 7 days of birth.
‘‘Congenital anomalies’’ were defined as all structural, func-
tional, and genetic anomalies diagnosed in aborted fetuses,
at birth or in the neonatal period. Congenital malformations
were classified using the International Classification of Dis-
eases Q codes (Q00–Q99) as conditions registered in the Inter-
national Statistical Classification of Diseases and Related
Health Problems (25).
Statistical Analysis
Statistical analyses were carried out using SPSS 17.0 soft-
ware (SPSS, Inc.). Data were presented as means  SD if
they demonstrated normal distributions, or as medians
(ranges) for non-normal distributions; qualitative data were
presented as percentages. Differences in the means of contin-
uous parametric data were analyzed using one-way ANOVA
or Kruskal-Wallis tests, and comparisons of rates among 3
groups were performed by the c2 test or Fisher's exact test,
as appropriate. Multiple pregnancy data for analysis were
entered as either individual deliveries or a multiple delivery.
With regard to birth weight and malformations, ‘‘infant’’
was the unit of analysis.
Binary logistic regression was performed to quantify the
effect of risk factors on congenital malformations between
groups. The significance of the model was calculated using
VOL. 103 NO. 5 / MAY 2015
 Fertility and Sterility®

a likelihood ratio; uncertainty was evaluated and explained
with Nagelkerke's R2. The effect of risk factors on congenital
malformation was expressed through an adjusted odds ratio
(OR) and 95% confidence interval (CI). In the regression anal-
ysis, the following possible factors were considered: maternal
age, infertility duration, parity, ovarian stimulation method,
method of insemination, multiple birth, and infant gender.
All of these factors were introduced into the regression equa-
tion using the forward conditional method.
RESULTS
The sample of 3,637 pregnancy cycles resulting from the IVF or
ICSI and FET procedures, from March 1, 2012 to July 1, 2013,
consisted of 3,090 ongoing pregnancies, 78 ectopic pregnan-
cies, 433 spontaneous abortions, 13 induced abortions, 1 still-
birth and 22 that were lost to follow-up, resulting in a total of
2,060 live-born infants (1,622 live-birth cycles). Among them,
458 pregnancies originating from 885 ETs led to the birth of
587 live-born neonates after treatment with LPS; 984 pregnan-
cies originating from 1,929 ETs led to the birth of 1,257 live-
born babies after treatment with the short GnRH-a protocol;
and 180 pregnancies originating from 334 ETs led to the birth
of 216 live-born neonates after treatment with mild ovarian
stimulation. These results mean that comparable percentages
of pregnancies led to the live births of infants for each of these
3 ovarian stimulation protocols (details of group distribution
profiles are presented in the flow chart in Supplemental
Figure 2 (available online) and Table 1).
The average age of women included was notably higher
in the mild ovarian-stimulation group (32.52  4.12 years)
than that in the other 2 groups (31.35  4.14 years in the
LPS group, and 31.49  4.05 years in the short GnRH-a pro-
tocol group; P< .004 and P< .006, respectively). Other
maternal characteristics, such as body mass index, infertility
duration, and proportion of nullipara, were comparable in the
3 groups. Although preterm delivery rate, postterm pregnancy
rate, and low birth weight in the LPS group (21.18%, 0.22%,
and 22.66%, respectively) were higher than those in the short
GnRH-a protocol group (16.47%, 0.10%, and 17.98%, respec-
tively) and the mild ovarian-stimulation group (18.89%, 0%,
and 19.91%, respectively), these differences were not statisti-
cally significant (P>.05).
Birth length was comparable in all groups (48.93 
2.29 cm vs. 49.11  2.14 cm vs. 48.94  2.18 cm, P>.05). Simi-
larly, a nonsignificant difference was found among the 3
groups for the incidence of multiple delivery, despite the fact
that pregnancies of patients who conceived after LPS and
the short GnRH-a protocol were more likely to result in multi-
ple pregnancies and, correspondingly, less likely to result in a
singleton. On the other hand, early neonatal death was higher
in the short GnRH-a protocol group (0.56%) than that in the
LPS group (0.17%) and the mild ovarian-stimulation group
(0%); however, these values were all very low, and and none
of the differences were statistically significant (P>.05).
A total of 15 cases (0.73%) qualified as having congenital
defects in all live-born infants, according to the definition in
the International Classification of Diseases. Defects were
observed in 6 of 587 (1.02%) in the LPS group, 8 of 1,257
(0.64%) in the short GnRH-a protocol group, and 1 of 216
(0.46%) in the mild ovarian-stimulation group; the differ-
ences were not significant. Comparisons between groups of
birth defects according to neonatal gender, singletons, and
multiples were carried out and reached the undifferentiated

TABLE 1

Characteristics and neonatal outcome by regimen group.

Short protocol Mild ovarian
Characteristics and outcomes LPS (n [ 458) (n [ 984) stimulation (n [ 180) F/c2 P value
Maternal age (y) 31.35  4.14 31.49  4.05 32.52  4.12 5.701 < .003
Body mass index (kg/m2) 21.41  3.16 21.57  6.09 21.39  2.68 0.192 .826
Infertility duration (y; median [range]) 3.0 (0–20) 3.0 (0–17) 3.0 (1–20) 0.131 .937
Nullipara 428 (93.44) 872 (88.62) 170 (94.00) 3.103 .212
Total ETs (n) 885 1,929 334 – –
Embryos from IVF cycles 583 (65.88) 1,333 (69.10) 238 (71.26) 4.313 .116
Embryos from ICSI cycles 302 (34.12) 596 (30.90) 96 (28.74) 4.313 .116
Cleavage-stage ETs 802 (90.62) 1,740 (90.20) 308 (92.22) 1.358 .507
Blastocyst ETs 83 (9.38) 189 (9.80) 26 (7.78) 1.358 .507
Live born infants (n) 587 1,257 216 – –
Gestational age (wk) 37.91  1.98 38.02  2.07 38.20  1.88 1.416 .243
28%age<37 96 (20.96) 157 (15.96) 34 (18.89) 5.575 .062
<28 1 (0.22) 5 (0.51) 0 (0) 0.656 .838
R42 1 (0.22) 1 (0.10) 0 (0) 1.044 .632
Birth weight (g) 2,965.49  639.41 3,020.97  635.51 3,037.79  624.69 1.786 .168
Birth weight <2,500 g 133 (22.66) 226 (17.98) 43 (19.91) 5.599 .061
Birth length (cm) 48.93  2.29 49.11  2.14 48.94  2.18 0.993 .371
Multiple delivery cycles (n) 158 339 50 – –
Multiple delivery rate (%) 34.49 34.45 27.78 3.203 .202
Early neonatal death 1 (0.17) 7 (0.56) 0 (0) 1.452 .470
Note: Values are n (%), or mean (SD), unless otherwise indicated. For normal-distribution data, presented as mean  SD, comparisons were performed using 1-way ANOVA. For non-normal-
distribution data, presented as median (range), comparisons were performed using the Kruskal-Wallis test. For qualitative data, presented as percentage, comparisons were made using Pearson's c2
analysis; exceptions are early neonatal death and gestational age <28 wk and R42 wk, for which Fisher's exact test was used. For maternal age, a significant difference was detected among groups,
P< .003; between LPS and mild ovarian stimulation using least significant digit, P< .004; and between the short protocol and mild ovarian stimulation using least-significant difference (LSD)
P< .006.
Chen. Birth defects after ovarian stimulation. Fertil Steril 2015.

VOL. 103 NO. 5 / MAY 2015 1197

ORIGINAL ARTICLE: ASSISTED REPRODUCTION
outcomes presented in Table 2. The detailed breakdown of de-
tected malformations according to the various organ systems
is presented in Table 3. As the results show, higher percent-
ages of cardiac defects (0.24%) and digestive problems
(0.19%) were observed, compared with other malformations,
and in total, accounted for 0.43 percentage points of the
0.73% malformations in the 2,060 live-born infants, meaning
that the remaining 0.30 percentage points accounted for other
malformations.
Table 4 presents the results of unadjusted and adjusted
analyses for factors that may have influenced congenital mal-
formations in IVF or ICSI, FET cycles. Binary logistic regres-
sion analyses were conducted of treatment and patient
characteristics associated with the adverse outcomes of each
risk factor hypothesized or proven to be associated with
adverse outcomes. These characteristics included maternal
age, body mass index, parity, infertility duration, ovarian-
stimulation method, method of insemination, multiple births,
and infant gender, which were included in the analysis using
the forward conditional method. Infertility duration (x1),
body mass index (x2), and multiple births (x3) were included
in the logistic regression equation as: logit P¼7.139 þ
0.149x1 þ 0.051x2 þ 1.361x3 (P< .05, with the 2 log likeli-
hood ¼ 119.85, and the Nagelkerke R2 ¼ 0.135).
The results show a statistically significant increase in the
probability of an adverse outcome for infertility duration, for
multiple births, in both the unadjusted and adjusted models.
Increasing maternal infertility duration was associated with
an increase in the likelihood of congenital malformation (OR:
1.144, 95% CI: 1.000–1.309; P< .05 for the unadjusted model;
and OR: 1.161, 95% CI: 1.009–1.335; P< .037 for the adjusted
model). Compared with singletons, multiples were 3.58 times
more likely to experience an adverse outcome (OR: 3.58, 95%
CI: 1.194–10.734; P< .023), in analyses using the unadjusted
model; the likelihood was 3.899 times greater in analyses using
the adjusted model (OR: 3.899, 95% CI: 1.179–12.896;
P< .026). No association was found between various ovarian-
stimulation methods and congenital malformations.
DISCUSSION
Our previous studies have demonstrated that LPS is feasible
for producing competent oocytes and embryos with optimal
pregnancy outcomes (7, 8). Nevertheless, this issue is
TABLE 2
Incidence of birth defects in live-born infants.
Characteristic LPS (n [ 587)
Number of birth defects 6 (1.02)
Number of deliveries
Singletons 2 (0.34)
Multiples 4 (0.68)
Birth defects, by gender
Male 2 (0.34)
Female 4 (0.68)
Note: Data are n (%) and were compared using Fisher's exact test.
Chen. Birth defects after ovarian stimulation. Fertil Steril 2015.
1198
complicated by the suspected, but unconfirmed, detrimental
effect of LPS on oocyte and/or embryo quality when the P
level is high. The currently available literature does not
provide clear evidence on this point, but rather has been
varied and contradictory.
Sohn et al. (26) found that exposure of the oocyte to P
in vivo had a significant adverse effect on implantation
rate. Others found that the addition of P to cumulus oocyte
complexes reduced, by approximately 40%, the proportion
of both total oocytes and cleaved oocytes that reached the
blastocyst stage. This effect was partially reversed by the
administration of mifepristone, an antiprogestin, also called
RU486 (27). In contrast, some believe that exposure to P
in vitro did not affect the proportion of embryos that devel-
oped to the blastocyst stage, the number of blastocyst cells,
or the relative abundance of selected transcripts in the blasto-
cyst (28).
Furthermore, a significant positive correlation between
the level of P and the proportion of oocytes undergoing
germinal vesicle breakdown was observed when cumulus
oocyte complexes were cultured for 20 hours (29). However,
unlike these studies that directly exposed the oocyte to P,
the intrafollicular hormonal milieu of early antral follicles
during the luteal phase might be similar throughout the
various phases of the menstrual cycle, and these antral folli-
cles do not respond to rising LH and P concentrations, owing
to a lack of LH receptors (30). As a result, the rising LH and P
levels may not adversely affect oocyte quality.
Powerful evidence has shown that oocytes that are aspi-
rated during the luteal phase are capable of maturing in vitro
and being fertilized (8, 30–33). Even so, no tangible evidence
has proven the safety of LPS for offspring that originate from
this ovarian-stimulation regimen. Therefore, we followed up
on the live-birth defects of infants born from LPS, in this
study, as data on the occurrence of malformations in
offspring are an important tool for assessing potential adverse
effects of an ovarian-stimulation regimen.
Use of a mild ovarian-stimulation approach has been pro-
posed as a method that can yield good embryologic results
(fertilization rate, clinical pregnancy rate, and live-birth
rate) in poor responders; this effect was particularly evidenced
in the subgroup of women age >37 years, who had slightly
higher pregnancy rates than those using conventional
ovarian stimulation. Thus, we tend to select the mild
Short protocol Mild ovarian
(n [ 1,257) stimulation (n [ 216) c2 P value
8 (0.64) 1 (0.46) 0.970 .624
3 (0.24) 0 (0) 0.494 .802
5 (0.40) 1 (0.46) 0.942 .710
3 (0.24) 0 (0) 0.494 .802
5 (0.40) 1 (0.46) 0.942 .710
VOL. 103 NO. 5 / MAY 2015
 Fertility and Sterility®

TABLE 3

Types of malformations among 2,060 live-born infants.

Malformation type ICD-10 code and diagnosis (n) n (prob./thousand)
Circulatory system Q20.0: persistent truncus arteriosus (1); Q21.0: ventricular septal defect (1); Q21.1: atrial septal defect 5 (2.4)
(2); Q25.6: stenosis of pulmonary artery (1)
Digestive system Q39.0: atresia of esophagus (1); Q41.2: congenital atresia of ileum (1); Q43.1: congenital megacolon (2) 4 (1.9)
Urinary system Q62.0: congenital hydronephrosis (1) 1 (0.5)
Musculoskeletal Q68.0: congenital sternomastoid torticollis (1) 1 (0.5)
system
Nervous system Q00.0: anencephaly (1); Q01.9: encephalocele (1) 2 (1.0)
Other congenital Q82.5: vascular nevus (1); Q28.8: congenital aneurysm (1) 2 (1.0)
malformation
Birth defects total 15 (7.3)
Note: ICD-10 ¼ International Classification of Diseases, 10th edition. Prob. ¼ probability.
Chen. Birth defects after ovarian stimulation. Fertil Steril 2015.

ovarian-stimulation approach for routine infertility treatment
for women who have advanced age and/or decreased ovarian
function. This selection approach may explain why women in
the mild ovarian-stimulation group were significantly older
(32.52  4.12 years) than those in the LPS group (31.35 
4.14 years, P< .004) and than those in the short GnRH-a pro-
tocol group (31.49  4.05 years, P< .006), as shown in
Table 1.
For various newborn parameters, including weeks of
gestation, preterm birth rate, postterm delivery rate, birth
weight and length, odds of low birth weight, multiple delivery
rate, early neonatal mortality, and incidence of live-birth de-
fects, none of the differences among groups was statistically
significant (Tables 1 and 2). This finding indicated that
neonatal outcomes in the LPS vs. the classic ovarian-
stimulation protocol did not have obvious differences, regard-
less of normality or adverse consequences.
Most previous studies have documented that oocyte qual-
ity is closely related to offspring health, as oocyte quality af-
fects early embryonic survival, the establishment and
maintenance of pregnancy, and even growth abnormalities
(34). The most-common research studies on this topic
analyzed the relationship between metabolic abnormalities
of oocytes (35), or age-related infertility (36), and abnormal-
ities in offspring. These findings suggest that birth outcomes
were worse when more-immature oocytes were obtained, as
an oocyte that had not completed cytoplasmic maturation
was of poor quality, and thus, unable to successfully complete
normal developmental processes (34).
Our research confirmed this possibility to some extent, as
higher oocyte maturation rates were observed when ovarian
stimulation occurred during the luteal phase (37), and the
incidence of live-birth defects was not higher, compared
with the rate in those using the traditional ovarian-
stimulation regimen. In fact, the study indicated that LPS is
safe for live-born infants at birth. In line with previous
studies, organ systems most frequently affected by congenital
malformations in our study cohort were the cardiovascular
system and the digestive system. The total rate of congenital
malformation in live-born infants was 0.73% in our study

TABLE 4

Unadjusted and adjusted probability ratios of the influencing factors for congenital malformation.
Variables Coefficient (B) OR (Exp[B]) (95% CI) Wald (c2) P value
Unadjusted model
Maternal age 0.048 1.049 (0.919–1.196) 0.503 .478
BMI 0.064 1.066 (0.940–1.208) 0.979 .323
Infertility duration 0.135 1.144 (1.000–1.309) 3.858 < .050
Parity –16.421 0.000 0.000 .996
Method of insemination 0.419 1.521 (0.525–4.407) 0.598 .440
Multiple births 1.275 3.580 (1.194–10.734) 5.182 < .023
Gender of infants 0.733 0.480 (0.144–1.603) 1.422 .233
LPS 0.681 1.976 (0.229–17.029) 0.384 .536
Short protocol 0.383 1.467 (0.182–11.803) 0.130 .719
Adjusted model
Infertility duration 0.149 1.161 (1.009–1.335) 4.355 < .037
Multiple births 1.361 3.899 (1.179–12.896) 4.973 < .026
BMI 0.051 1.052 (0.962–1.151) 1.241 .265
Note: The reference for LPS and the short protocol is the mild ovarian-stimulation protocol.
The variables in the adjusted model are adjusted for factors in the model, plus maternal age, parity, different ovarian-stimulation methods, method of insemination, and gender of infants, using the
forward conditional method. These confounding factors were excluded after adjustment. Infertility duration, BMI, and multiple births are included in the logistic regression equation as:
logit P¼7.139 þ 0.149x1 þ 0.051x2 þ 1.361x3. The –2 log likelihood ¼ 119.85; the Nagelkerke R2 ¼ 0.135. BMI ¼ body mass index.
Chen. Birth defects after ovarian stimulation. Fertil Steril 2015.

VOL. 103 NO. 5 / MAY 2015 1199

ORIGINAL ARTICLE: ASSISTED REPRODUCTION
(Table 3), slightly lower than the 1.09% rate for abnormalities
occurring in infants within 7 days after delivery, in People's
Republic of China, as reported in another study (38).
Possible explanations for the lower rate of birth defects in
our study are: [1] Unlike the other study, our purposeful re-
striction of participants, to those women with no reported
maternal diseases or adverse environmental exposure during
pregnancy, allowed a more precise evaluation of the isolated
effect of various ovarian-stimulation regimens on subsequent
neonatal outcomes. [2] In the present study, we focused on the
incidence of congenital abnormalities in live-born infants;
therefore, the findings may not represent the overall birth-
defect rates, owing to spontaneous abortion, stillbirth, or
induced abortion, which may have occurred precisely because
of detectable malformations by antenatal examination, all of
which were excluded from the study. Furthermore, if subjects
were lost to follow-up, they were excluded from this analysis,
and some may have had birth defects. [3] Infants in this study
were all born from FET cycles, which have been reported to
have a lower risk of birth defects, compared with fresh ET
cycles (39).
Embryos that survive the freeze–thaw process, as in an
early natural-selection process, are less likely to be aneuploi-
dal, and therefore are of superior quality, which may have a
positive influence on birth outcome. In addition, unlike fresh
ET, developing embryos in FET cycles may be shielded from
exposure to the excessively hormonal ovarian environment,
which occurs immediately prior to fresh ET, but not before
thawed ET, a difference that could have various adverse ef-
fects on very early pregnancy (39).
Not surprisingly, and in line with previous studies, infer-
tility duration as well as multiple pregnancy, was found to
be a significant risk factor for congenital malformation
(Table 3). On one hand, higher rates of congenital abnormality
were considered to be associated with infertility duration, as a
long waiting time to pregnancy is a presentation condition in
infertile couples, and this and others of these conditions may
share R1 causal path with adverse birth outcomes (40). Zhu
et al. (41) found that the overall prevalence of congenital mal-
formations rose with increasing time to pregnancy. Singletons
who were born to subfertile couples who either conceived
spontaneously or received fertility treatments had an even
higher risk of congenital anomalies than singletons born to
fertile couples with a time to pregnancy of <1 year (41).
Thus, increased risk is assumed to be mainly a result of subfer-
tility or intrinsic parental characteristics, defined as years of
known childlessness, and played a significant role in the asso-
ciation between ART and congenital anomalies, rather than of
any aspect of the ART procedure itself.
On the other hand, multiple births, which are associated
not only with a higher risk of maternal morbidity, but also
a higher likelihood of perinatal morbidity and mortality,
have been proven to be more likely than singletons to be
affected by congenital malformations (42). The significantly
increased risk of birth defects with multiple births might be
explained in several ways. First, nutritional supply is as an
important factor in normal fetal development; insufficient
nutrition from the mother may cause fetal retardation or
1200
make the fetus susceptible and less resistant to birth defect–
related factors.
Second, common genetic backgrounds and crowded uter-
ine conditions were both found to occur in high proportions in
association with some inherent and mechanical deformations
of fetuses in multiple births. Finally, the fact that pregnancy
achieved via ART is so precious for a woman who has had a
long period of infertility means that intensive prenatal diag-
nosis is carried out. For this reason, fetal deformities can be
diagnosed early, and timely termination of an abnormal preg-
nancy may occur more often with singletons than twins
because mothers may be reluctant to interrupt such a preg-
nancy if a normal sibling is present from the same multifeta-
tion. As a result, this situation may selectively affect the
probability of abortion in singletons and twins. Overall, a
higher incidence of birth defects in children from multiple
births may result from a combination of factors, and their
interaction may play an important role in the development
of these birth defects.
The study does have limitations. Our data were limited to
neonatal information extracted from parent questionnaires,
rather than from direct access to medical records. Therefore,
minor problems likely escaped detection, and birth defect
rates may have been underestimated, although these issues
are unlikely to have altered infant birth characteristics.
In conclusion, our large, retrospective study found that
the data to date do not indicate an elevated rate of birth ab-
normalities after luteal-phase ovarian stimulation, but further
study with larger populations is needed to confirm this
finding. In addition, the risk of congenital anomalies was
found to be significantly correlated with infertility duration
and multiple pregnancy. This finding suggests that the
increased incidence of congenital abnormality may result
from intrinsic parental characteristics, rather than from the
particular fertility-treatment protocol or techniques used. A
higher risk of deformity in multiple births may favor the
recommendation to use elective, single ET, and couples un-
dergoing ART should be made aware of the known increased
risk of birth defects associated with a twin birth. Long-term
follow-up, throughout childhood and into adolescence and
adulthood, is needed to evaluate the health of children born
after LPS.
Acknowledgments: The authors thank Meiping Sheng for
patients' follow-up and Suqun Zhang for data auditing.
REFERENCES
1. Steptoe PC, Edwards RG. Birth after the reimplantation of a human embryo.
Lancet 1978;2:366.
2. Ferraretti AP, Goossens V, Kupka M, Bhattacharya S, de Mouzon J,
Castilla JA, et al. Assisted reproductive technology in Europe, 2009: results
generated from European registers by ESHRE. Hum Reprod 2013;28:
2318–31.
3. Serour GI, Aboulghar M, Mansour R, Sattar MA, Amin Y, Aboulghar H.
Complications of medically assisted conception in 3,500 cycles. Fertil Steril
1998;70:638–42.
4. Macklon NS, Stouffer RL, Giudice LC, Fauser BC. The science behind 25
years of ovarian stimulation for in vitro fertilization. Endocr Rev 2006;27:
170–207.
VOL. 103 NO. 5 / MAY 2015

5. Porter RN, Smith W, Craft IL, Abdulwahid NA, Jacobs HS. Induction of ovula-
tion for in-vitro fertilisation using buserelin and gonadotropins. Lancet 1984;
2:1284–5.
6. Tarlatzis BC, Fauser BC, Kolibianakis EM, Diedrich K, Rombauts L, Devroey P.
GnRH antagonists in ovarian stimulation for IVF. Hum Reprod Update 2006;
12:333–40.
7. Kuang Y, Chen Q, Hong Q, Lyu Q, Fu Y, Ai A, et al. Luteal-phase ovarian
stimulation case report: three-year follow-up of a twin birth. J IVF Reprod
Med Genet 2013;1:106. http://dx.doi.org/10.4172/jfiv.1000106.
8. Kuang Y, Hong Q, Chen Q, Lyu Q, Ai A, Fu Y, et al. Luteal-phase ovarian
stimulation is feasible for producing competent oocytes in women undergo-
ing in vitro fertilization/intracytoplasmic sperm injection treatment, with
optimal pregnancy outcomes in frozen-thawed embryo transfer cycles. Fertil
Steril 2014;101:105–11.
9. Tarlatzis BC, Bili H. Safety of GnRH agonists and antagonists. Expert Opin
Drug Saf 2004;3:39–46.
10. Zegers-Hochschild F, Adamson GD, de Mouzon J, Ishihara O, Mansour R,
Nygren K, et al. The International Committee for Monitoring Assisted Repro-
ductive Technology (ICMART) and the World Health Organization (WHO)
revised glossary on ART terminology, 2009. Hum Reprod 2009;24:2683–7.
11. Fauser BC, Nargund G, Andersen AN, Norman R, Tarlatzis B, Boivin J, et al.
Mild ovarian stimulation for IVF: 10 years later. Hum Reprod 2010;25:
2678–84.
12. Baart EB, Martini E, Eijkemans MJ, Van Opstal D, Beckers NG, Verhoeff A,
et al. Milder ovarian stimulation for in-vitro fertilization reduces aneuploidy
in the human preimplantation embryo: a randomized controlled trial. Hum
Reprod 2007;22:980–8.
13. Keshavarz M, Cheung NW, Babaee GR, Moghadam HK, Ajami ME,
Shariati M. Gestational diabetes in Iran: incidence, risk factors and preg-
nancy outcomes. Diabetes Res Clin Pract 2005;69:279–86.
14. Chung CS, Myrianthopoulos NC. Factors affecting risks of congenital mal-
formations. II. Effect of maternal diabetes on congenital malformations.
Birth Defects Orig Artic Ser 1975;11:23–38.
15. Van Zutphen AR, Werler MM, Browne MM, Romitti PA, Bell EM, McNutt LA,
et al. Maternal hypertension, medication use, and hypospadias in the Na-
tional Birth Defects Prevention Study. Obstet Gynecol 2014;123:309–17.
16. Khoury MJ, Becerra JE, D'Almada PJ. Maternal thyroid disease and risk of
birth defects in offspring: a population-based case-control study. Paediatr
Perinat Epidemiol 1989;3:402–20.
17. A search for the possible molecular mechanisms of thyroid dysgenesis: sex
ratios and associated malformations. J Clin Endocrinol Metab 1999;84:
2502–6.
18. Alverson CJ, Strickland MJ, Gilboa SM, Correa A. Maternal smoking and
congenital heart defects in the Baltimore-Washington Infant Study. Pediat-
rics 2011;127:e647–53.
19. Gaspari L, Paris F, Jandel C, Kalfa N, Orsini M, Daures JP, et al. Prenatal envi-
ronmental risk factors for genital malformations in a population of 1442
French male newborns: a nested case-control study. Hum Reprod 2011;
26:3155–62.
20. Hackshaw A, Rodeck C, Boniface S. Maternal smoking in pregnancy and
birth defects: a systematic review based on 173,687 malformed cases and
11.7 million controls. Hum Reprod Update 2011;17:589–604.
21. Larkby CA, Goldschmidt L, Hanusa BH, Day NL. Prenatal alcohol exposure is
associated with conduct disorder in adolescence: findings from a birth
cohort. J Am Acad Child Adolesc Psychiatry 2011;50:262–71.
22. Richardson S, Browne ML, Rasmussen SA, Druschel CM, Sun L, Jabs EW,
et al. Associations between periconceptional alcohol consumption and cra-
niosynostosis, omphalocele, and gastroschisis. Birth Defects Res A Clin Mol
Teratol 2011;91:623–30.
VOL. 103 NO. 5 / MAY 2015
Fertility and Sterility®
23. Cummins JM, Breen TM, Harrison KL, Shaw JM, Wilson LM, Hennessey JF. A
formula for scoring human embryo growth rates in in vitro fertilization: its
value in predicting pregnancy and in comparison with visual estimates of
embryo quality. J Vitro Fertil Embryo Transf 1986;3:284–95.
24. Br€amer GR. International statistical classification of diseases and related
health problems. Tenth revision. World Health Stat 1988;41:32–6.
25. World Health Organization. International statistical classification of diseases
and related health problems. Available at: http://apps.who.int/iris/bitstream/
10665/61362/1/WHO_MNH_MEP_87.1_REV.2.pdf. Last accessed March
17, 2015.
26. Sohn SH, Penzias AS, Emmi AM, Dubey AK, Layman LC, Reindollar RH, et al.
Administration of progesterone before oocyte retrieval negatively affects
the implantation rate. Fertil Steril 1999;71:11–4.
27. Silva CC, Knight PG. Effects of androgens, progesterone and their antago-
nists on the developmental competence of in vitro matured bovine oocytes.
J Reprod Fertil 2000;119:261–9.
28. Lonergan P. Influence of progesterone on oocyte quality and embryo devel-
opment in cows. Theriogenology 2011;76:1594–601.
29. Yamashita Y, Shimada M, Okazaki T, Maeda T, Terada T. Production of pro-
gesterone from de novo-synthesized cholesterol in cumulus cells and its
physiological role during meiotic resumption of porcine oocytes. Biol Reprod
2003;68:1193–8.
30. Demirtas E, Elizur SE, Holzer H, Gidoni Y, Son WY, Chian RC, et al. Immature
oocyte retrieval in the luteal phase to preserve fertility in cancer patients. Re-
prod Biomed Online 2008;17:520–3.
31. von Wolff M, Thaler CJ, Frambach T, Zeeb C, Lawrenz B, Popovici RM, et al.
Ovarian stimulation to cryopreserve fertilized oocytes in cancer patients can
be started in the luteal phase. Fertil Steril 2009;92:1360–5.
32. Bedoschi GM, de Albuquerque FO, Ferriani RA, Navarro PA. Ovarian stimula-
tion during the luteal phase for fertility preservation of cancer patients: case
reports and review of the literature. J Assist Reprod Genet 2010;27:491–4.
33. Maman E, Meirow D, Brengauz M, Raanani H, Dor J, Hourvitz A. Luteal
phase oocyte retrieval and in vitro maturation is an optional procedure for
urgent fertility preservation. Fertil Steril 2011;95:64–7.
34. Krisher RL. The effect of oocyte quality on development. J Anim Sci 2004;82
E-Suppl:E14–23.
35. Jungheim ES, Schoeller EL, Marquard KL, Louden ED, Schaffer JE, Moley KH.
Diet-induced obesity model: abnormal oocytes and persistent growth ab-
normalities in the offspring. Endocrinology 2010;151:4039–46.
36. Baird DT, Collins J, Egozcue J, Evers LH, Gianaroli L, Leridon H, et al. Fertility
and ageing. Hum Reprod Update 2005;11:261–76.
37. Cha K, Do B, Chi H, Yoon T, Choi D, Koo J, et al. Viability of human follicular
oocytes collected from unstimulated ovaries and matured and fertilized
in vitro. Reprod Fertil Dev 1992;4:695–701.
38. Yan J, Huang G, Sun Y, Zhao X, Chen S, Zou S, et al. Birth defects after as-
sisted reproductive technologies in China: analysis of 15,405 offspring in
seven centers (2004 to 2008). Fertil Steril 2011;95:458–60.
39. Halliday JL, Ukoumunne OC, Baker HW, Breheny S, Jaques AM, Garrett C,
et al. Increased risk of blastogenesis birth defects, arising in the first 4 weeks
of pregnancy, after assisted reproductive technologies. Hum Reprod 2010;
25:59–65.
40. Basso O, Baird DD. Infertility and preterm delivery, birthweight, and
Caesarean section: a study within the Danish National Birth Cohort. Hum
Reprod 2003;18:2478–84.
41. Zhu JL, Basso O, Obel C, Bille C, Olsen J. Infertility, infertility treatment, and
congenital malformations: Danish national birth cohort. BMJ 2006;333:679.
42. Mastroiacovo P, Castilla EE, Arpino C, Botting B, Cocchi G, Goujard J, et al.
Congenital malformations in twins: an international study. Am J Med Genet
1999;83:117–24.
1201
ORIGINAL ARTICLE: ASSISTED REPRODUCTION

SUPPLEMENTAL FIGURE 1

Timeline of LPS stimulation, showing the LPS protocol. A total of 225 international units (IU) of human menopausal gonadotropin (hMG), and 2.5
mg of letrozole were simultaneously administered daily, starting the day after ovulation. The dosage and duration of hMG and letrozole were
adjusted after 7–8 days, according to follicular development and sex hormones. Daily administration of 10 mg of medroxyprogesterone acetate
was added, according to the stimulus duration, to avoid oocyte retrieval during menstruation. Ovulation was induced with 0.1 mg of a GnRH-
agonist (GnRH-a) whenR3 follicles reached diameters of 18 mm, or 1 dominant follicle reached 20 mm. Ibuprofen, in a dose of 0.6 g, was used
on the trigger day and the day after. GnRH-a ¼ gonadotropin-releasing hormone–agonist; HMG ¼ human menopausal gonadotropin; IU ¼
international unit; LPS ¼ luteal-phase stimulation; MPA, medroxyprogesterone acetate.
Chen. Birth defects after ovarian stimulation. Fertil Steril 2015.

1201.e1 VOL. 103 NO. 5 / MAY 2015

 Fertility and Sterility®

SUPPLEMENTAL FIGURE 2

Flow chart of the study. Ongoing pregnancy was defined as the presence of a gestational sac and fetal heart beat detection at 12 weeks. Rate of
spontaneous abortion was significantly different from the groups (P<.016). Differences with P value (2-tailed) <.016 were considered significant
after Bonferroni adjustments when pairwise comparisons were performed. A significant difference was found between the short-protocol group
and the mild ovarian-stimulation group (P<.005). No obvious difference was found between the luteal-phase stimulation group and the mild
ovarian-stimulation group (P¼.021). There were a total of 8 induced abortions caused by fetus malformations in all groups: 2 in the luteal-
phase stimulation group, 5 in the short-protocol group, and 1 in the mild ovarian-stimulation group. The differences were not significant
(P>.05). The malformations were classified as congenital heart disease (3 cases), Down syndrome (2 cases), 18-trisomy (1 case),
cheilopalatognathus (1 case), and cleft vertebra (1 case).
Chen. Birth defects after ovarian stimulation. Fertil Steril 2015.

VOL. 103 NO. 5 / MAY 2015 1201.e2