The n e w e ng l a n d j o u r na l of m e dic i n e

Original Article

Fresh versus Frozen Embryos for Infertility

in the Polycystic Ovary Syndrome
Z.-J. Chen, Y. Shi, Y. Sun, B. Zhang, X. Liang, Y. Cao, J. Yang, J. Liu, D. Wei,
N. Weng, L. Tian, C. Hao, D. Yang, F. Zhou, J. Shi, Y. Xu, J. Li, J. Yan, Y. Qin,
H. Zhao, H. Zhang, and R.S. Legro

A BS T R AC T

BACKGROUND
The transfer of fresh embryos is generally preferred over the transfer of frozen The authors’ full names, academic de-
embryos for in vitro fertilization (IVF), but some evidence suggests that frozen- grees, and affiliations are listed in the
Appendix. Address reprint requests to
embryo transfer may improve the live-birth rate and lower the rates of the ovarian Dr. Chen at the Center for Reproductive
hyperstimulation syndrome and pregnancy complications in women with the Medicine, Shandong Provincial Hospital,
polycystic ovary syndrome. Shandong University, 324 Jingwu Rd.,
Jinan 250021, China, or at chenzijiang@
hotmail.com.
METHODS
In this multicenter trial, we randomly assigned 1508 infertile women with the N Engl J Med 2016;375:523-33.
DOI: 10.1056/NEJMoa1513873
polycystic ovary syndrome who were undergoing their first IVF cycle to undergo Copyright © 2016 Massachusetts Medical Society.
either fresh-embryo transfer or embryo cryopreservation followed by frozen-
embryo transfer. After 3 days of embryo development, women underwent the
transfer of up to two fresh or frozen embryos. The primary outcome was a live
birth after the first embryo transfer.
RESULTS
Frozen-embryo transfer resulted in a higher frequency of live birth after the first
transfer than did fresh-embryo transfer (49.3% vs. 42.0%), for a rate ratio of 1.17
(95% confidence interval [CI], 1.05 to 1.31; P = 0.004). Women who underwent
frozen-embryo transfer also had a lower frequency of pregnancy loss (22.0% vs.
32.7%), for a rate ratio of 0.67 (95% CI, 0.54 to 0.83; P<0.001), and of the ovarian
hyperstimulation syndrome (1.3% vs. 7.1%), for a rate ratio of 0.19 (95% CI, 0.10
to 0.37; P<0.001), but a higher frequency of preeclampsia (4.4% vs. 1.4%), for a rate
ratio of 3.12 (95% CI, 1.26 to 7.73; P = 0.009). There were no significant between-
group differences in rates of other pregnancy and neonatal complications. There
were five neonatal deaths in the frozen-embryo group and none in the fresh-
embryo group (P = 0.06).
CONCLUSIONS
Among infertile women with the polycystic ovary syndrome, frozen-embryo transfer
was associated with a higher rate of live birth, a lower risk of the ovarian hyper-
stimulation syndrome, and a higher risk of preeclampsia after the first transfer
than was fresh-embryo transfer. (Funded by the National Basic Research Program
of China and others; ClinicalTrials.gov number, NCT01841528.)

n engl j med 375;6 nejm.org August 11, 2016 523

The New England Journal of Medicine
Downloaded from nejm.org by ade tonara on March 27, 2019. For personal use only. No other uses without permission.
Copyright © 2016 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l
I
n vitro fertilization (IVF) is widely
performed as an infertility treatment and has
resulted in the births of more than 5 million
infants worldwide.1 However, there are concerns
about the safety of the procedures for women
and for their infants.1,2 The ovarian hyperstimu-
lation syndrome (which is caused by ovarian
enlargement, an increase in vascular permea-
bility and abdominal ascites, and intravascular
hemoconcentration) is a potentially life-threat-
ening complication of ovarian stimulation.3 Preg-
nancies conceived by means of IVF are associ-
ated with greater risks of maternal and neonatal
complications, including preeclampsia, preterm
delivery, low birth weight, and congenital anom-
alies, than are spontaneous pregnancies.4-7
Women with the polycystic ovary syndrome who
undergo IVF are at increased risk for the ovarian
hyperstimulation syndrome and complications
later in the pregnancy.3,8
The transfer of fresh embryos is the usual
practice in most IVF units when such embryos
are available. In some cases, all embryos may be
cryopreserved without a fresh-embryo transfer,
most commonly to prevent the ovarian hyper-
stimulation syndrome when excess follicle devel-
opment has occurred. After ovarian recovery,
embryos are thawed and transferred to the
uterus after a programmed physiologic cycle of
hormone replacement to prepare the endome-
trium. With fresh-embryo transfer, the medica-
tions given for ovarian stimulation or the result-
ing supraphysiologic sex steroids may alter the
endometrial receptivity 9-11 and adversely affect
trophoblastic invasion or placentation.12
Observational studies have shown that rates
of singleton pregnancy are higher after frozen-
embryo transfer than after fresh-embryo trans-
fer.13 A meta-analysis of 11 observational studies
showed that pregnancies from frozen-embryo
transfer were associated with lower risks of pre-
term birth, low birth weight, and perinatal
death than were pregnancies from fresh-embryo
transfer.14 Subsequently, small, single-center,
randomized studies comparing the two proce-
dures have shown better ongoing pregnancy
rates with frozen-embryo transfer.15 However,
data are lacking from randomized trials to de-
termine any differences between the two proce-
dures in live-birth rates and pregnancy compli-
cations.
524 n engl j med 375;6
The New England Journal of Medicine
Downloaded from nejm.org by ade tonara on March 27, 2019. For personal use only. No other uses without permission.
Copyright © 2016 Massachusetts Medical Society. All rights reserved.
of m e dic i n e
We designed a multicenter, randomized, con-
trolled trial to assess whether initial frozen-
embryo transfer would result in a higher fre-
quency of a live birth and fewer treatment-related
and subsequent pregnancy complications than
fresh-embryo transfer in infertile women with
the polycystic ovary syndrome.
Me thods
Study Design and Oversight
The study rationale and a detailed study protocol
were published previously.16 The study protocol
and statistical analysis plan are available with
the full text of this article at NEJM.org. The
study was approved by the ethics committees at
the Reproductive Medical Center of Shandong
University and other clinical sites. A data and
safety monitoring board was established to over-
see the study. All the patients provided written
informed consent before participation.
Enrollment began in June 2013 and was com-
pleted in May 2014 at 14 reproductive medical
centers throughout China. Follow-up of preg-
nancies from the initial embryo transfer was
completed in July 2015. All data entry, data
management, and analyses were coordinated or
performed at Shandong University, which was
the data-coordinating center for this study. The
first two authors assume responsibility for the
accuracy and completeness of the data, and all
the authors vouch for the fidelity of the report to
the study protocol. There was no commercial
support for this study.
Study Population
A total of 1508 infertile women with the polycys-
tic ovary syndrome who were undergoing their
first IVF cycle were randomly assigned to undergo
either fresh-embryo transfer or embryo cryo-
preservation followed by frozen-embryo trans-
fer. All the women were between the ages of 20
and 34 years and weighed at least 40 kg.
To diagnose the polycystic ovary syndrome,
we used modified Rotterdam criteria,17,18 which
included menstrual abnormalities (irregular
uterine bleeding, oligomenorrhea, or amenor-
rhea) combined with either hyperandrogenism
or polycystic ovaries, as validated in our Chinese
population.19 Hyperandrogenism was diagnosed
on the basis of either hirsutism or hyperandro-
nejm.org August 11, 2016
 Fresh vs. Frozen Embryos in Polycystic Ovary Syndrome
genemia. Hirsutism was determined by means
of a modified Ferriman–Gallwey score of more
than 6 (on a scale of 0 to 36, with higher scores
indicating increased hair growth) at the screen-
ing examination.20 Hyperandrogenemia was de-
fined as an elevated total testosterone level ac-
cording to local laboratory criteria.18,21 All serum
laboratory values were obtained in local labora-
tories as part of the clinical evaluation and treat-
ment of the patients. Polycystic ovaries were
defined as the presence of either an ovary con-
taining 12 or more antral follicles measuring
2 to 9 mm in diameter or an increased ovarian
volume (>10 cm3).22 Other causes of hyperan-
drogenism and ovulation dysfunction — includ-
ing tumors, congenital adrenal hyperplasia,
hyperprolactinemia, and thyroid dysfunction —
were ruled out.
Patients were also excluded from the study if
they had a history of unilateral oophorectomy,
recurrent spontaneous abortion (defined as three
or more previous spontaneous pregnancy losses),
congenital or acquired uterine malformations,
abnormal results on parental karyotyping, or
medical conditions that contraindicated assisted
reproductive technology or pregnancy. Well-
controlled diabetes and hypertension were not
exclusion criteria. There were no male-factor
exclusions, and the use of donor semen was al-
lowed. All patients had tried other infertility
therapies without success.
Study Procedures
All the patients received a standardized ovarian
stimulation regimen, oocyte retrieval, and fertil-
ization, followed by a planned transfer of up to
two day-3 embryos, as recommended for this
age group by the American Society of Reproduc-
tive Medicine and Chinese guidelines.16,23 In
brief, recombinant follicle-stimulating hor-
mone (Gonal-f, Merck Serono) at a daily dose of
112.5 IU for patients weighing 60 kg or less and
150 IU for those weighing more than 60 kg was
started on day 2 or 3 of the menstrual cycle.
These doses were adjusted according to the ovar-
ian response, as monitored on ultrasonography
and the measurement of serum sex steroids. Hu-
man menopausal gonadotropin (Menopur, Fer-
ring) could be added at the discretion of local
investigators. Gonadotropin-releasing–hormone
antagonist (Cetrorelix, Merck Serono) at a daily
n engl j med 375;6
The New England Journal of Medicine
Downloaded from nejm.org by ade tonara on March 27, 2019. For personal use only. No other uses without permission.
Copyright © 2016 Massachusetts Medical Society. All rights reserved.
dose of 250 µg was started when the largest fol-
licle exceeded 12 mm. Urinary human chorionic
gonadotropin at a dose of 4000 to 8000 IU was
administered to induce oocyte maturation when
two or more follicles measured 18 mm or more.
Oocyte retrieval was performed 34 to 36 hours
later. On the day of oocyte retrieval, if an ade-
quate number of oocytes (>3 and <30) were re-
trieved and the patient was at low risk for the
ovarian hyperstimulation syndrome (as deter-
mined by local investigators), patients were ran-
domly assigned to one of the two study groups
in a 1:1 ratio, stratified according to study site.
An online central randomization system (www
.medresman.org) was used to automatically gen-
erate the assignment sequence, which was un-
known to the clinical investigators.
For patients who were assigned to the fresh-
embryo group, intramuscular progesterone at a
daily dose of 80 mg was administered for luteal-
phase support, beginning on the day of oocyte
retrieval until 10 weeks after conception. For pa-
tients who were assigned to the frozen-embryo
group, no luteal-phase support was administered
after oocyte retrieval, and day-3 embryos were
cryopreserved for later transfer. Oral estradiol
valerate (Progynova, Delpharm Lille) was admin-
istered for endometrial preparation on day 2 or 3
of the second menstrual cycle after oocyte re-
trieval. Intramuscular progesterone at a dose of
80 mg per day was added when the endometrial
thickness reached 8 mm or more or at the physi-
cian’s discretion if the thickness of the endome-
trium was less than 8 mm (with the latter occur-
ring in 63 of 746 patients). On day 4 of the
progesterone regimen, two day-3 frozen embryos
were thawed and transferred. The luteal-phase
support with estradiol valerate and intramuscu-
lar progesterone for endometrium preparation
continued until 10 weeks after conception.
Embryo Culture, Evaluation, and Selection
for Transfer
The oocytes were inseminated approximately 4 to
6 hours after follicular aspiration by a conven-
tional method or intracytoplasmic sperm injec-
tion, according to the sperm quality. Morpho-
logic criteria were used for embryo scoring.24 On
day 3, two high-quality embryos were picked out
for fresh transfer23 or cryopreserved by means
of vitrification in the group undergoing frozen-
nejm.org August 11, 2016 525
 The n e w e ng l a n d j o u r na l
embryo transfer. Local investigators had the
option to transfer day-2 embryos if there were
fewer than three embryos on day 2. Biochemical
pregnancy was defined as a human chorionic
gonadotropin level of more than 10 mIU per
milliliter, as measured at 14 days after embryo
transfer. Clinical pregnancy was defined as the
presence of a gestational sac in the uterine cav-
ity at 35 days after embryo transfer, as detected
on ultrasonography. Ongoing pregnancy was
defined as the presence of a fetus with heart
motion at 11 to 12 weeks of gestation. All preg-
nancy and neonatal outcomes were obtained
through review of medical records. Moderate
and severe ovarian hyperstimulation syndromes
were defined according to accepted criteria.25
Study Outcomes
The primary outcome was a live birth, which
was defined as delivery of any viable infant at
28 weeks or more of gestation during the first
embryo transfer. Prespecified secondary out-
comes included biochemical pregnancy, clinical
pregnancy, ongoing pregnancy, singleton live
birth, cumulative live birth (including subse-
quent frozen-embryo transfer), pregnancy loss,
moderate or severe ovarian hyperstimulation
syndrome, ectopic pregnancy, pregnancy and
neonatal complications, and congenital anom-
alies. (Criteria for the secondary outcomes are
provided in Table S1 in the Supplementary Ap-
pendix, available at NEJM.org.)
Statistical Analysis
The study was designed to have a power of 80%
at a two-sided significance level of 0.01 to de-
tect an absolute difference of 10 percentage points
in the live-birth rate between the two study groups
(on the basis of anticipated rates of 30% after
fresh-embryo transfer vs. 40% after frozen-
embryo transfer) by means of Pearson’s chi-
square test.16 We calculated that at least 530 pa-
tients per study group were required, a number
that we increased to 590 to allow for a dropout
rate of 10%. (Details are provided in the Meth-
ods section in the Supplementary Appendix.)
We performed intention-to-treat analyses to
compare the frequency of live births in the two
study groups. Categorical data were represented
as frequency and percentage; differences in these
measures between the study groups were as-
526 n engl j med 375;6
The New England Journal of Medicine
Downloaded from nejm.org by ade tonara on March 27, 2019. For personal use only. No other uses without permission.
Copyright © 2016 Massachusetts Medical Society. All rights reserved.
of m e dic i n e
sessed by means of chi-square analysis, with the
use of Fisher’s exact test for expected frequencies
of less than 5. Continuous data were expressed
as means (±SD), with a Wilcoxon rank-sum test
for between-group differences. We used multi-
variate logistic regression to adjust for the effect
of study site and baseline characteristics. We
also performed post hoc analyses of data from
patients who received treatment and from those
who complied with the study protocol (Tables
S3, S4, and S5 in the Supplementary Appendix).
All analyses were performed with the use of SAS
software, version 9.2 (SAS Institute).
R e sult s
Study Patients
The baseline characteristics of the patients were
similar in the two study groups (Table 1), as
were the characteristics of the IVF procedures
(Table 2, and Table S2 in the Supplementary Ap-
pendix). A total of 170 patients dropped out of
the study or deviated from the protocol, includ-
ing 82 of 762 (10.8%) in the fresh-embryo group
and 88 of 746 (11.8%) in the frozen-embryo
group (P = 0.53) (Fig. 1).
Live Birth and Secondary Outcomes
The frequency of live births after frozen-embryo
transfer was significantly higher than that after
fresh-embryo transfer (49.3% vs. 42.0%), for a
rate ratio of 1.17 (95% confidence interval [CI],
1.05 to 1.31; P = 0.004). Frozen-embryo transfer
also resulted in a higher frequency of singleton
live births than fresh-embryo transfer (33.5% vs.
27.8%), for a rate ratio of 1.20 (95% CI, 1.03 to
1.40; P = 0.02), and a lower frequency of preg-
nancy loss (22.0% vs. 32.7%), for a rate ratio of
0.67 (95% CI, 0.54 to 0.83; P<0.001) (Table 3).
The incidence of moderate or severe ovarian
hyperstimulation syndrome was significantly
lower in the frozen-embryo group than in the
fresh-embryo group (1.3% vs. 7.1%), for a rate
ratio of 0.19 (95% CI, 0.10 to 0.37; P<0.001)
(Table 4). However, the incidence of preeclamp-
sia was significantly higher with frozen-embryo
transfer (4.4% vs. 1.4%), for a rate ratio of 3.12
(95% CI, 1.26 to 7.73; P = 0.009), although no
patient in either group had severe preeclampsia.
There were two stillbirths and five neonatal
deaths in the frozen-embryo group and none in
nejm.org August 11, 2016
 Fresh vs. Frozen Embryos in Polycystic Ovary Syndrome

Table 1. Characteristics of the Patients at Baseline.*

Frozen-Embryo Transfer Fresh-Embryo Transfer

Characteristic (N = 746) (N = 762)

No. of No. of
Patients† Value Patients† Value
Age — yr 28.1±3.0 28.2±3.1
Body-mass index‡ 23.8±3.8 23.9±3.6
Blood pressure — mm Hg
Systolic 119±12 119±12
Diastolic 75±9 76±8
Fertility history
Duration of attempt to conceive — yr 3.8±2.2 3.8±2.2
Previous conception — no. (%) 233 (31.2) 212 (27.8)
Concomitant infertility factors —
no. (%)
Tubal factors 421 (56.4) 403 (52.9)
Male factors 152 (20.4) 163 (21.4)
Ultrasonographic findings
Antral follicle count in both ovaries 668 29.2±7.7 690 29.2±8.6
Polycystic ovaries, according to 686/746 (92.0) 678/761 (89.1)
modified Rotterdam criteria
— no./total no. (%)§
Laboratory tests
Total testosterone — ng/dl 714 44.4±26.3 726 41.1±18.5
Ratio of luteinizing hormone to 1.6±1.0 760 1.5±0.9
follicle-stimulating hor-
mone¶
Estradiol — pg/ml 746 43.7±24.5 758 42.4±26.8
Progesterone — ng/ml 562 0.6±0.3 580 0.6±0.3
Glucose — mg/dl
Fasting 744 94.9±12.2 758 94.2±11.6
At 2 hr∥ 722 120.6±32.9 743 122.4±32.8

* Plus–minus values are means ±SD. There were no significant differences between groups in any of the baseline char-
acteristics except for total testosterone (P = 0.005). To convert the values for total testosterone to nanomoles per liter,
multiply by 0.03467. To convert the values for estradiol to picomoles per liter, multiply by 3.671. To convert the values
for progesterone to nanomoles per liter, multiply by 3.180. To convert the values for glucose to millimoles per liter,
multiply by 0.05551.
† The number of patients who were included in each analysis is provided if it differs from the total number in the study
group.
‡ The body-mass index is the weight in kilograms divided by the square of the height in meters.
§ Polycystic ovaries were defined as the presence of an antral follicle count of 12 or more or a volume of more than 10 cm3
in at least one ovary.
¶ Luteinizing hormone and follicle-stimulating hormone were measured in units per liter.
∥ The glucose level was measured 2 hours after the oral administration of 75 g of glucose.

the fresh-embryo group (P = 0.50 and P = 0.06, clinical pregnancy, ongoing pregnancy, ectopic
respectively). pregnancy, other pregnancy complications, neo-
There were no significant between-group dif- natal complications, or congenital anomalies
ferences in the rates of biochemical pregnancy, (Tables 3 and 4). The congenital anomalies were

n engl j med 375;6 nejm.org August 11, 2016 527

The New England Journal of Medicine
Downloaded from nejm.org by ade tonara on March 27, 2019. For personal use only. No other uses without permission.
Copyright © 2016 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

Table 2. Outcomes of Controlled Ovarian Hyperstimulation.*

Frozen-Embryo Transfer Fresh-Embryo Transfer

Characteristic (N = 746) (N = 762)

No. of No. of
Patients Value Patients Value
No. of days of ovarian stimulation 10.3±2.1 10.3±2.2
Gonadotropin dose — IU 715 1592±686 732 1571±667
Estradiol level on hCG trigger day — pg/ml 728 4288±2210 746 4141±2159
Progesterone level on hCG trigger day — ng/ml 745 1.0±0.5 1.0±0.6
Endometrial thickness — mm
On hCG trigger day 10.5±2.1 10.4±2.1
Before frozen-embryo transfer 740 9.2±1.4 NA
No. of oocytes retrieved 14.4±6.0 14.2±5.8
No. of good-quality embryos on day 3 726 6.3±4.0 733 6.3±4.0
Patients undergoing embryo transfer —
no./total no. (%)†
Day 2 10/726 (1.4) 24/748 (3.2)
Day 3 689/726 (94.9) 682/748 (91.2)
Day 5 27/726 (3.7) 42/748 (5.6)
No. of embryos transferred 726 748
Mean 1.95±0.23 1.96±0.19
1 — no./total no. (%) 39/726 (5.4) 28/748 (3.7)
2 — no./total no. (%) 687/726 (94.6) 720/748 (96.3)

* Plus–minus values are means ±SD. There was no significant difference between groups in any of the outcomes of ovarian
stimulation except for the day of embryo transfer (P = 0.01). The abbreviation hCG denotes human chorionic gonadotropin,
and NA not applicable.
† Local investigators had the option of transferring the day-2 embryos if there were fewer than three embryos on day 2.
Day-5 embryo transfers were performed in case of poor embryo quality or at the request of a patient.

varied and were not specific to any organ system
(Table S6 in the Supplementary Appendix). The
incidences of adverse events and serious adverse
events were similar in the two groups (Table 4).
Logistic-regression analyses showed that the
frequency of a live birth was still higher in the
frozen-embryo group than in the fresh-embryo
group after adjustment for study site and base-
line characteristics (Table S10 in the Supple-
mentary Appendix). There was no significant
between-group difference in the cumulative live-
birth rate, including all frozen-embryo transfers
performed within 12 months after the initial
transfer (P = 0.30) (Table S9 in the Supplemen-
tary Appendix).
The results of post hoc as-treated and per-
protocol analyses were consistent with the results
of the intention-to-treat analysis. An additional
post hoc analysis showed that the singleton
birth weight in the frozen-embryo group was
higher than that in the fresh-embryo group
(P = 0.005) (Table 3).
Discussion
In this large study involving infertile women
with the polycystic ovary syndrome, the frequen-
cies of a live birth and a singleton live birth after
frozen-embryo transfer were significantly higher
than the frequencies after fresh-embryo transfer.
The between-group difference was mediated
through a significantly lower rate of pregnancy
loss in the frozen-embryo group. The incidence
of the ovarian hyperstimulation syndrome was
significantly lower with frozen-embryo transfer,
although the rate of preeclampsia was higher.

528 n engl j med 375;6 nejm.org August 11, 2016

The New England Journal of Medicine

Downloaded from nejm.org by ade tonara on March 27, 2019. For personal use only. No other uses without permission.
Copyright © 2016 Massachusetts Medical Society. All rights reserved.
 Fresh vs. Frozen Embryos in Polycystic Ovary Syndrome

1900 Women provided consent

392 Were excluded

1508 Underwent randomization

762 Were assigned to undergo 746 Were assigned to undergo

fresh-embryo transfer frozen-embryo transfer

82 Discontinued participation 88 Discontinued participation

680 Complied with protocol 658 Complied with protocol
or had protocol violation or had protocol violation
14 Did not complete 20 Did not complete
embryo transfer embryo transfer
8 Received fresh-blasto- 23 Received frozen-
cyst transfer blastocyst transfer
57 Received frozen-embryo 44 Received fresh-embryo
transfer transfer
34 Received frozen- 4 Received fresh-
blastocyst transfer 1 Was lost blastocyst transfer 4 Were lost
23 Received day 3 frozen- to follow-up 40 Received day 3 fresh- to follow-up
embryo transfer embryo transfer
3 Did not fulfill eligibility 1 Did not fulfill eligibility
criteria criteria

320 Delivered live-born infants 368 Delivered live-born infants

Figure 1. Study Enrollment and Outcomes.

There were no significant between-group differ-
ences in the rates of other pregnancy-related or
neonatal complications.
Previous randomized trials comparing the
transfer of fresh embryos versus frozen embryos
have not reported live-birth rates, although this
outcome is the recommended end point for in-
fertility trials.26 Our results in women with the
polycystic ovary syndrome showed similar clini-
cal pregnancy rates in the two groups but a
higher frequency of live births with frozen-
embryo transfer, which underscores the impor-
tance of considering live birth as the primary
outcome.26 In a previous randomized trial in-
volving women with an expected high response
to ovarian stimulation (defined as >15 antral
follicles observed on baseline ultrasonography),
no significant difference was found in the rate
of clinical pregnancy between the two embryo-
transfer procedures,27 whereas another study
from the same group reported a higher rate of
clinical pregnancy after frozen-embryo transfer
than after fresh-embryo transfer among women
with an expected normal response to ovarian
stimulation (defined as 8 to 15 antral follicles
observed on baseline ultrasonography).15
The higher rate of preeclampsia that we ob-
served in the frozen-embryo group is consistent
with some previous reports that have assessed
this outcome.28-30 An observational study com-
paring outcomes in the same mother similarly
showed that frozen-embryo transfer was associ-
ated with a higher risk of hypertensive disorders
than fresh-embryo transfer.29 In our study, the

n engl j med 375;6 nejm.org August 11, 2016 529

The New England Journal of Medicine
Downloaded from nejm.org by ade tonara on March 27, 2019. For personal use only. No other uses without permission.
Copyright © 2016 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

Table 3. Live Birth, Pregnancy, and Pregnancy Loss.*

Frozen-Embryo Fresh-Embryo Absolute Difference Rate Ratio in

Transfer Transfer between Groups Frozen-Embryo Group P
Outcome (N = 746) (N = 762) (95% CI) (95% CI) Value
Primary outcome: live birth — no. (%)† 368 (49.3) 320 (42.0) 7.3 (2.3 to 12.4) 1.17 (1.05 to 1.31) 0.004
Singleton 250 (33.5) 212 (27.8) 5.7 (1.0 to 10.3) 1.20 (1.03 to 1.40) 0.02
Twin 117 (15.7) 108 (14.2) 1.5 (–2.1 to 5.1) 1.11 (0.87 to 1.41) 0.41
Triplet‡ 1 (0.1) 0
Birth weight — g
Singleton 3511.2±593.6 3349.4±553.2 161.8 (56.2 to 267.3) 0.005
Twin 2479.7±503.2§ 2481.7±496.0 –2.0 (–94.9 to 90.8) 0.97
Secondary outcomes — no. (%)
Biochemical pregnancy¶ 492 (66.0) 492 (64.6) 1.4 (–3.4 to 6.2) 1.02 (0.95 to 1.10) 0.57
Clinical pregnancy∥ 438 (58.7) 428 (56.2) 2.5 (–2.4 to 7.5) 1.05 (0.96 to 1.14) 0.32
Ongoing pregnancy** 393 (52.7) 372 (48.8) 3.9 (–1.1 to 8.9) 1.08 (0.98 to 1.19) 0.13
Pregnancy loss — no./total no. (%)
Among biochemical pregnancies 108/492 (22.0) 161/492 (32.7) –10.8 (–16.3 to –5.2) 0.67 (0.54 to 0.83) <0.001
Among clinical pregnancies 64/438 (14.6) 107/428 (25.0) –10.4 (–15.7 to –5.1) 0.58 (0.44 to 0.77) <0.001
First trimester 42/438 (9.6) 56/428 (13.1) –3.5 (–7.7 to 0.7) 0.73 (0.50 to 1.07) 0.10
Second trimester 22/438 (5.0) 51/428 (11.9) –6.9 (–10.6 to –3.2) 0.42 (0.26 to 0.68) <0.001

* Plus–minus values are means ±SD.

† Live birth was defined as the delivery of a live-born infant after 28 weeks of gestation or more. However, two deliveries — one at 27 weeks
3 days and one at 25 weeks 4 days — were counted as live births.
‡ Two embryos were transferred, but one embryo developed into a monozygotic twin. Most of the patients with triplet pregnancy underwent
fetus reduction, but one patient declined to undergo the procedure and delivered three live-born infants.
§ Birth weights were missing for one pair of twins.
¶ Biochemical pregnancy was defined as a serum level of human chorionic gonadotropin of more than 10 mIU per milliliter.
∥ Clinical pregnancy was defined as the observation of a gestational sac on ultrasonography.
** Ongoing pregnancy was defined as the presence of a fetal heartbeat on ultrasonography at 12 weeks of gestation. Four patients in the frozen-
embryo group and one in the fresh-embryo group were lost to follow-up after 7 to 12 weeks of gestation.

rates of congenital anomalies were not signifi-
cantly different between the two groups and
were consistent with rates previously reported
after IVF.31
The potential excess of neonatal death, owing
primarily to prematurity, in the frozen-embryo
group warrants attention. The P value for the
between-group difference was of borderline sig-
nificance; especially given the multiple compari-
sons performed, chance cannot be ruled out. In
a retrospective Nordic population-based cohort
study, singleton infants who were born after
frozen-embryo transfer had a higher risk of peri-
natal death than singletons born after fresh-
embryo transfer (odds ratio, 1.49; 95% CI, 1.07
to 2.07).32 However, another study in a Danish
national cohort did not show a significant be-
tween-group difference in the rate of perinatal
death among singletons,33 and a meta-analysis
of observational studies showed that frozen-
embryo transfer was associated with a lower risk
of perinatal death than fresh-embryo transfer
(relative risk, 0.68; 95% CI, 0.48 to 0.96).14 Our
finding in a post hoc analysis that the birth
weight of singleton newborns in the frozen-
embryo group was higher than that in the fresh-
embryo group is consistent with the findings in
previous observational studies.32-34
Our study was not designed to identify the
mechanisms underlying our results. However,
frozen-embryo transfer allows the ovary to re-
cover from the ovarian stimulation and the ex-
posed endometrial lining to shed, providing a
fresh start for both. Increased risks of the ovar-
ian hyperstimulation syndrome35 and pregnancy
loss36 have been associated with higher estradiol

530 n engl j med 375;6 nejm.org August 11, 2016

The New England Journal of Medicine

Downloaded from nejm.org by ade tonara on March 27, 2019. For personal use only. No other uses without permission.
Copyright © 2016 Massachusetts Medical Society. All rights reserved.
 Fresh vs. Frozen Embryos in Polycystic Ovary Syndrome

levels after IVF, and the estradiol levels in fresh

fant, congenital esophageal atresia in one, and prematurity in three (at 28 weeks 6 days in one infant and 29 weeks 5 days in twins, one of whom also had the respiratory distress syn-
P Value
cycles are much higher than in frozen cycles.

<0.001

§ Neonatal death was defined as the death of a newborn within 28 days after delivery. The causes of the five neonatal deaths were the neonatal respiratory distress syndrome in one in-
0.009
1.00

0.20
0.29
0.50
0.52
0.06
Higher live-birth rates have been reported with
the use of letrozole as an ovulation-induction
agent than with clomiphene; letrozole results

Rate Ratio in Frozen-

in lower midluteal estrogen levels than clomi-

0.19 (0.10 to 0.37)

1.00 (0.42 to 2.38)

3.12 (1.26 to 7.73)
1.97 (0.68 to 5.71)
1.17 (0.87 to 1.57)

1.24 (0.68 to 2.28)

Embryo Group
phene.18

(95% CI)
Our results — along with the increased use
of frozen-embryo transfer — have implications
for the public reporting and interpretation of
IVF outcomes in national registries (e.g., the

‡ The denominator for congenital anomalies includes the number of live newborns plus the number of fetuses that were therapeutically terminated.
registries maintained by the Centers for Disease
Control and Prevention and the Society for As-

Absolute Difference
sisted Reproductive Technology in the United

–5.7 (–7.7 to –3.7)

between Groups

0.0 (–1.8 to 1.8)

1.1 (–0.6 to 2.9)

2.7 (–2.3 to 7.8)

0.9 (–1.7 to 3.6)

3.0 (0.8 to 5.2)
States). Until recently, the statistic that has typi-

(95% CI)

† Four patients in the frozen-embryo group and one in the fresh-embryo group were lost to follow-up after 7 to 12 weeks of gestation.
cally been used to reflect the success rate of an
individual clinic has been the clinical pregnancy
rate per initiated cycle with fresh-embryo trans-
fer, although this outcome has been criticized
for potential bias.37,38 Our observation that there
was no significant between-group difference in Fresh-Embryo

68/427 (15.9)
54/762 (7.1)

10/492 (2.0)

17/432 (3.9)
6/427 (1.4)
5/427 (1.2)
cumulative live-birth rates underscores the ef- (N = 762)
Transfer

ficacy of subsequent frozen-embryo transfer

0
among women in whom fresh-embryo transfer no./total no. (%)
was not successful. The cumulative live-birth
rate (including transfers of both fresh and fro-
zen embryos) may provide a better indication of
Frozen-Embryo

81/434 (18.7)
ultimate success of a single IVF cycle but may
10/746 (1.3)

10/492 (2.0)
19/434 (4.4)
10/434 (2.3)

24/491 (4.9)
2/370 (0.5)

5/487 (1.0)

* A full listing of adverse events is provided in Table S7 in the Supplementary Appendix.

(N = 746)
Transfer

not fully account for between-group differences

in maternal and perinatal outcomes.
Our study has some limitations. First, about
10% of the patients in each group deviated from
the protocol, which in most cases involved not
receiving the assigned treatment, even though we
Moderate or severe ovarian hyperstimulation syndrome before bio-

delayed randomization until the day of oocyte

retrieval to minimize the likelihood of nonad-
herence. Such deviations would be expected to
Table 4. Maternal, Fetal, and Neonatal Adverse Events.*

Gestational hypertension among clinical pregnancies

attenuate the between-group difference in the

Ectopic pregnancy among biochemical pregnancies

intention-to-treat analysis. In addition, results of

the per-protocol analysis were consistent with
Preterm delivery among clinical pregnancies
Preeclampsia among clinical pregnancies†

those of the intention-to-treat analysis. Second,

Neonatal death among live newborns§

our study specifically involved women with the

polycystic ovary syndrome, a group that is at
increased risk for the ovarian hyperstimulation
Stillbirth among all deliveries

syndrome and pregnancy complications, so our

chemical pregnancy

findings may not be applicable to other women

Congenital anomalies‡

undergoing IVF.
In conclusion, in a randomized trial involving
infertile women with the polycystic ovary syn-
Outcome

drome, we found that frozen-embryo transfer

drome).

resulted in a higher frequency of live births than

fresh-embryo transfer, a difference that was at-

n engl j med 375;6 nejm.org August 11, 2016 531

The New England Journal of Medicine
Downloaded from nejm.org by ade tonara on March 27, 2019. For personal use only. No other uses without permission.
Copyright © 2016 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

tributed to a lower rate of pregnancy loss. Women
in the frozen-embryo group also had a lower
frequency of the ovarian hyperstimulation syn-
drome but a higher frequency of preeclampsia.
Supported by a grant from the National Basic Research Pro-
gram of China (973 Program) (2012CB944700), by grants from
the National Natural Science Foundation of China (81430029
and 81471428), and by grants from the Thousand Talents Pro-
gram (to Drs. Legro and H. Zhang).
Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
We thank the Reproductive Medicine Network Steering Com-
mittee of the National Institutes of Health for sharing the pro-
tocol and case-report forms from the Pregnancy in Polycystic
Ovary Syndrome II study.

Appendix
The authors’ full names and academic degrees are as follows: Zi-Jiang Chen, M.D., Ph.D., Yuhua Shi, M.D., Ph.D., Yun Sun, M.D.,
Ph.D., Bo Zhang, M.D., Xiaoyan Liang, M.D., Ph.D., Yunxia Cao, M.D., Ph.D., Jing Yang, M.D., Ph.D., Jiayin Liu, M.D., Ph.D., Daimin
Wei, M.D., Ph.D., Ning Weng, M.D., Lifeng Tian, M.D., Ph.D., Cuifang Hao, M.D., Ph.D., Dongzi Yang, M.D., Ph.D., Feng Zhou, M.S.,
Juanzi Shi, M.D., Ph.D., Yongle Xu, M.D., Jing Li, M.D., Junhao Yan, M.D., Ph.D., Yingying Qin, M.D., Ph.D., Han Zhao, M.D., Ph.D.,
Heping Zhang, Ph.D., and Richard S. Legro, M.D.
The authors’ affiliations are as follows: Center for Reproductive Medicine, Shandong Provincial Hospital Affiliated to Shandong
University, Key Laboratory of Reproductive Endocrinology, Shandong University, Ministry of Education, and National Research Center
for Assisted Reproductive Technology and Reproductive Genetics, Jinan (Z.-J.C., Y. Shi, D.W., J. Li, J. Yan, Y.Q., H. Zhao), Center for
Reproductive Medicine, Ren Ji Hospital, School of Medicine, Shanghai Jiao Tong University (Z.-J.C., Y. Sun), and Shanghai Key Labora-
tory for Assisted Reproduction and Reproductive Genetics (Z.-J.C., Y. Sun), Shanghai, Center for Reproductive Medicine, Maternal and
Child Health Hospital in Guangxi, Guangxi (B.Z.), Reproductive Medicine Center, the Sixth Affiliated Hospital of Sun Yat-sen Univer-
sity (X.L.) and Center for Reproductive Medicine, Sun Yat-sen Memorial Hospital, Sun Yat-sen University (D.Y.), Guangzhou, Center for
Reproductive Medicine, the First Affiliated Hospital of Anhui Medical University, Hefei (Y.C.), Center for Reproductive Medicine, Wuhan
University, Wuhan (J. Yang), Department of Obstetrics and Gynecology, First Affiliated Hospital of Nanjing Medical University, Nanjing
(J. Liu), Reproductive Medicine Center of Jinghua Hospital, Shenyang (N.W.), Center for Reproductive Medicine, Jiangxi Provincial
Maternal and Child Health Hospital, Nanchang (L.T.), Center for Reproductive Medicine of Yantai Yuhuangding Hospital, Yantai (C.H.),
Center for Reproductive Medicine, Department of Obstetrics and Gynecology, Sir Run Run Shaw Hospital Affiliated to Zhejiang Univer-
sity School of Medicine, Hangzhou (F.Z.), Assisted Reproduction Center, Maternal and Child Health Care Hospital of Shanxi Province,
Xi’an (J.S.), and Center for Reproduction and Genetics, Suzhou Municipal Hospital, Suzhou (Y.X.) — all in China; Department of Bio-
statistics, Yale University School of Public Health, New Haven, CT (H. Zhang); and Department of Obstetrics and Gynecology, Penn
State College of Medicine, Hershey (R.S.L.).

References
1. Kissin DM, Jamieson DJ, Barfield WD.
Monitoring health outcomes of assisted
reproductive technology. N Engl J Med
2014;371:91-3.
2. Kawwass JF, Kissin DM, Kulkarni AD,
et al. Safety of assisted reproductive tech-
nology in the United States, 2000-2011.
JAMA 2015;313:88-90.
3. Delvigne A, Rozenberg S. Epidemiol-
ogy and prevention of ovarian hyperstim-
ulation syndrome (OHSS): a review. Hum
Reprod Update 2002;8:559-77.
4. Hansen M, Kurinczuk JJ, Bower C,
Webb S. The risk of major birth defects
after intracytoplasmic sperm injection
and in vitro fertilization. N Engl J Med
2002;346:725-30.
5. Helmerhorst FM, Perquin DA, Donker
D, Keirse MJ. Perinatal outcome of single-
tons and twins after assisted conception:
a systematic review of controlled studies.
BMJ 2004;328:261.
6. Jackson RA, Gibson KA, Wu YW,
Croughan MS. Perinatal outcomes in sin-
gletons following in vitro fertilization:
a meta-analysis. Obstet Gynecol 2004;103:
551-63.
7. Pandey S, Shetty A, Hamilton M,
Bhattacharya S, Maheshwari A. Obstetric
and perinatal outcomes in singleton preg-
nancies resulting from IVF/ICSI: a sys-
tematic review and meta-analysis. Hum
Reprod Update 2012;18:485-503.
8. Tang T, Glanville J, Hayden CJ, White
D, Barth JH, Balen AH. Combined life-
style modification and metformin in obese
patients with polycystic ovary syndrome:
a randomized, placebo-controlled, double-
blind multicentre study. Hum Reprod
2006;21:80-9.
9. Devroey P, Bourgain C, Macklon NS,
Fauser BC. Reproductive biology and IVF:
ovarian stimulation and endometrial re-
ceptivity. Trends Endocrinol Metab 2004;
15:84-90.
10. Bourgain CDP, Devroey P. The endo-
metrium in stimulated cycles for IVF.
Hum Reprod Update 2003;9:515-22.
11. Horcajadas JA, Riesewijk A, Polman J,
et al. Effect of controlled ovarian hyper-
stimulation in IVF on endometrial gene
expression profiles. Mol Hum Reprod
2005;11:195-205.
12. Choux C, Carmignac V, Bruno C,
Sagot P, Vaiman D, Fauque P. The placenta:
phenotypic and epigenetic modifications
induced by assisted reproductive tech-
nologies throughout pregnancy. Clin Epi-
genetics 2015;7:87.
13. Kansal Kalra S, Ratcliffe SJ, Milman
L, Gracia CR, Coutifaris C, Barnhart KT.
Perinatal morbidity after in vitro fertiliza-
tion is lower with frozen embryo transfer.
Fertil Steril 2011;95:548-53.
14. Maheshwari A, Pandey S, Shetty A,
Hamilton M, Bhattacharya S. Obstetric
and perinatal outcomes in singleton preg-
nancies resulting from the transfer of fro-
zen thawed versus fresh embryos generat-
ed through in vitro fertilization treatment:
a systematic review and meta-analysis.
Fertil Steril 2012;98:368-77.
15. Shapiro BS, Daneshmand ST, Garner
FC, Aguirre M, Hudson C, Thomas S. Evi-
dence of impaired endometrial receptivity
after ovarian stimulation for in vitro fer-
tilization: a prospective randomized trial
comparing fresh and frozen-thawed em-
bryo transfer in normal responders. Fertil
Steril 2011;96:344-8.
16. Shi Y, Wei D, Liang X, et al. Live birth
after fresh embryo transfer vs elective em-
bryo cryopreservation/frozen embryo trans-
fer in women with polycystic ovary syn-
drome undergoing IVF (FreFro-PCOS):
study protocol for a multicenter, prospec-
tive, randomized controlled clinical trial.
Trials 2014;15:154.
17. Rotterdam ESHRE/ASRM-Sponsored
PCOS Consensus Workshop Group. Re-
vised 2003 consensus on diagnostic crite-
ria and long-term health risks related to
polycystic ovary syndrome. Fertil Steril
2004;81:19-25.
18. Legro RS, Brzyski RG, Diamond MP,
et al. Letrozole versus clomiphene for in-
fertility in the polycystic ovary syndrome.
N Engl J Med 2014;371:119-29.
19. Chen ZJ, Zhang YW, Liu JY, et al. China

532 n engl j med 375;6 nejm.org August 11, 2016

The New England Journal of Medicine

Downloaded from nejm.org by ade tonara on March 27, 2019. For personal use only. No other uses without permission.
Copyright © 2016 Massachusetts Medical Society. All rights reserved.
 Fresh vs. Frozen Embryos in Polycystic Ovary Syndrome

diagnosis criteria of polycystic ovarian
syndrome diagnosis. Zhonghua Fu Chan
Ke Za Zhi 2012;47:74-5.
20. Hatch R, Rosenfield RL, Kim MH,
Tredway D. Hirsutism: implications, eti-
ology, and management. Am J Obstet Gy-
necol 1981;140:815-30.
21. Legro RS, Barnhart HX, Schlaff WD,
et al. Clomiphene, metformin, or both for
infertility in the polycystic ovary syn-
drome. N Engl J Med 2007;356:551-66.
22. Balen AH, Laven JS, Tan SL, Dewailly D.
Ultrasound assessment of the polycystic
ovary: international consensus definitions.
Hum Reprod Update 2003;9:505-14.
23. Practice Committee of American Soci-
ety for Reproductive Medicine; Practice
Committee of Society for Assisted Repro-
ductive Technology. Criteria for number
of embryos to transfer: a committee opin-
ion. Fertil Steril 2013;99:44-6.
24. Puissant F, Van Rysselberge M, Bar-
low P, Deweze J, Leroy F. Embryo scoring
as a prognostic tool in IVF treatment.
Hum Reprod 1987;2:705-8.
25. Golan A, Ron-el R, Herman A, Soffer
Y, Weinraub Z, Caspi E. Ovarian hyper-
stimulation syndrome: an update review.
Obstet Gynecol Surv 1989;44:430-40.
26. Harbin Consensus Conference Work-
shop Group. Improving the reporting of
clinical trials of infertility treatments
(IMPRINT): modifying the CONSORT
statement. Hum Reprod 2014;29:2075-82.
27. Shapiro BS, Daneshmand ST, Garner
FC, Aguirre M, Hudson C, Thomas S. Evi-
dence of impaired endometrial receptivity
after ovarian stimulation for in vitro fer-
tilization: a prospective randomized trial
comparing fresh and frozen-thawed em-
bryo transfers in high responders. Fertil
Steril 2011;96:516-8.
28. Sazonova A, Källen K, Thurin-Kjell-
berg A, Wennerholm UB, Bergh C. Ob-
stetric outcome in singletons after in vitro
fertilization with cryopreserved/thawed
embryos. Hum Reprod 2012;27:1343-50.
29. Opdahl S, Henningsen AA, Tiitinen
A, et al. Risk of hypertensive disorders in
pregnancies following assisted reproduc-
tive technology: a cohort study from the
CoNARTaS group. Hum Reprod 2015;30:
1724-31.
30. Ishihara O, Araki R, Kuwahara A, Ita-
kura A, Saito H, Adamson GD. Impact of
frozen-thawed single-blastocyst transfer
on maternal and neonatal outcome: an
analysis of 277,042 single-embryo trans-
fer cycles from 2008 to 2010 in Japan.
Fertil Steril 2014;101:128-33.
31. Davies MJ, Moore VM, Willson KJ, et al.
Reproductive technologies and the risk of
birth defects. N Engl J Med 2012;366:
1803-13.
32. Wennerholm UB, Henningsen AK,
Romundstad LB, et al. Perinatal outcomes
of children born after frozen-thawed em-
bryo transfer: a Nordic cohort study from
the CoNARTaS group. Hum Reprod 2013;
28:2545-53.
33. Pinborg A, Loft A, Aaris Henningsen
AK, Rasmussen S, Andersen AN. Infant
outcome of 957 singletons born after fro-
zen embryo replacement: the Danish Na-
tional Cohort Study 1995-2006. Fertil
Steril 2010;94:1320-7.
34. Wennerholm UB, Söderström-Anttila V,
Bergh C, et al. Children born after cryo-
preservation of embryos or oocytes: a sys-
tematic review of outcome data. Hum
Reprod 2009;24:2158-72.
35. Haning RV Jr, Austin CW, Carlson IH,
Kuzma DL, Shapiro SS, Zweibel WJ. Plasma
estradiol is superior to ultrasound and
urinary estriol glucuronide as a predictor
of ovarian hyperstimulation during induc-
tion of ovulation with menotropins. Fertil
Steril 1983;40:31-6.
36. Wang JX, Norman RJ, Wilcox AJ. Inci-
dence of spontaneous abortion among
pregnancies produced by assisted repro-
ductive technology. Hum Reprod 2004;19:
272-7.
37. Abdalla HI, Bhattacharya S, Khalaf Y.
Is meaningful reporting of national IVF
outcome data possible? Hum Reprod 2010;
25:9-13.
38. Maheshwari A, Bhattacharya S. Elec-
tive frozen replacement cycles for all:
ready for prime time? Hum Reprod 2013;
28:6-9.
Copyright © 2016 Massachusetts Medical Society.

ARTICLE METRICS NOW AVAILABLE

Visit the article page at NEJM.org and click on the Metrics tab to view
comprehensive and cumulative article metrics compiled from multiple sources,
including Altmetrics. Learn more at www.nejm.org/page/article-metrics-faq.

n engl j med 375;6 nejm.org August 11, 2016 533

The New England Journal of Medicine
Downloaded from nejm.org by ade tonara on March 27, 2019. For personal use only. No other uses without permission.
Copyright © 2016 Massachusetts Medical Society. All rights reserved.