Professional Documents
Culture Documents
Dor 1
Dor 1
Dor 1
1489–1498, 2018
Advanced Access publication on July 10, 2018 doi:10.1093/humrep/dey238
*Correspondence address. Reproductive Medicine Associates of New Jersey, 140 Allen Rd. Basking Ridge, NJ 07920, USA.
E-mail: smorin@rmanj.com
Submitted on January 8, 2018; resubmitted on May 4, 2018; accepted on June 15, 2018
STUDY QUESTION: Do infertile women aged <38 years with quantitative evidence of diminished ovarian reserve and/or poor response
to stimulation also exhibit poor oocyte quality as measured by blastulation rates, aneuploidy rates, and live birth rates?
SUMMARY ANSWER: Young women with evidence of accelerated follicular depletion, either by precycle ovarian reserve testing or post-
cycle evidence of low oocyte yield, exhibit equivalent blastulation rates, aneuploidy rates and live birth rates per euploid embryo transfer as
age-matched controls with normal precycle and postcycle parameters.
WHAT IS KNOWN ALREADY: Previous studies are conflicted as to whether women with evidence of diminished ovarian reserve and/
or poor ovarian response are also at increased risk of exhibiting evidence of poor oocyte quality. Most prior studies have failed to adequately
control for the confounding effect of female age on typical markers of oocyte quality in poor responders. The rate of follicular depletion
occurs at around 38 years on average; thus, evidence of quantitative depletion before this would indicate a premature diminution of ovarian
reserve and allow evaluation of whether markers of oocyte quality are tied to quantitative markers.
STUDY DESIGN, SIZE, DURATION: This was a retrospective cohort study at a single center between 2012 and 2016. This time frame was
specifically chosen as all embryos were cultured to the blastocyst stage at this center during the study period (no cleavage stage transfers were per-
formed). Two comparisons were made: precycle assessment of ovarian reserve (based on anti-mullerian hormone (AMH) level) and postcycle
oocyte yield results. For each comparison, patients in <10th percentile were compared to patients in the interquartile range (IQR) with respect to
blastulation rate, aneuploidy rate and live birth rate. A mixed effects model was created to control for female age (in the <38 year old range) and cor-
relation among oocytes from a given cohort.
PARTICIPANTS/MATERIALS, SETTING, METHODS: For the precycle blastulation analysis, only patients with AMH data available
were included (345 patients with AMH in the <10th percentile versus 1758 patients with AMH in the 25th to 75th percentile (IQR)). To
compare aneuploidy rates, the subset of these patients who pursued preimplantation genetic testing for aneuploidy (PGT-A) was then ana-
lyzed (124 patients in the <10th percentile versus 782 patients in the IQR). For the postcycle blastulation analysis, all patients who proceeded
to retrieval (whether or not they also had AMH data available) were included (535 patients with oocyte yield in the <10th percentile versus
2675 patients in the IQR). To compare aneuploidy rates, the subset of these patients who pursued PGT-A was then analyzed (156 patients in
the <10th percentile versus 1100 patients in the IQR).
MAIN RESULTS AND THE ROLE OF CHANCE: The adjusted odds of a given fertilized oocyte developing to a blastocyst, being aneu-
ploid or leading to a live birth after euploid transfer were no different if the oocyte was retrieved from a cycle with ovarian reserve parameters
© The Author(s) 2018. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved.
For Permissions, please e-mail: journals.permissions@oup.com
1490 Morin et al.
or oocyte yield in the <10th percentile compared to an oocyte retrieved in a cycle with those parameters in the 25–75th percentile.
An AMH level in the <10th percentile did more commonly result in cycle cancellation prior to retrieval and after retrieval prior to transfer
due to global arrest of embryos.
LIMITATIONS, REASONS FOR CAUTION: The timing of retrieval in patients with fewer oocytes may be more optimal given the great-
er ability to discern the overall maturity of the cohort, thus enhancing performance per retrieved oocyte. Analyses included only first cycles.
Subsequent adjustment of protocol due to prior performance may mean that some patients in the <10th percentile for oocyte yield are actu-
ally better prognosis patients than their first cycle indicates. Data on whether or not patients were on oral contraceptives at time that AMH
level drawn was not available. Other unknown biases are also likely to be present given retrospective nature of the study.
WIDER IMPLICATIONS OF THE FINDINGS: While young women with evidence of quantitative depletion of ovarian reserve have lower
live birth rates per stimulation cycle, this not attributable to poor oocyte quality because the blastulation rate per fertilized oocyte and live birth
rate per embryo transfer are equivalent to that in women with normal quantitative markers of ovarian reserve. Thus, the pathophysiology mediat-
Key words: ovarian reserve / oocyte quality / aneuploidy / ovarian stimulation / assisted reproduction
This time frame was chosen as all embryos were cultured to blastocyst at In order to compare aneuploidy rates between the groups, a subgroup
this clinic during the study period. No cleavage stage transfers were per- analysis was performed among only patients who underwent PGT-A. In
formed, regardless of the number of embryos present on Day 3. Thus, order to be included in this group, at least one blastocyst must have been
performance through blastulation could be observed of all embryos in the available for biopsy. For aneuploidy screening, diagnostic categories during
laboratory. Only the first cycle was included for each patient to eliminate the study period were only euploid or aneuploid (there were no segmental
previous failure bias. Patients with a history of ovarian surgery, chemother- imbalance or mosaic aneuploidy diagnoses made). Finally, IRs, LBRs and
apy or radiation exposure, fragile X premutation or abnormal karyotype pregnancy loss rates after euploid embryo transfers were compared
were excluded from the analysis as our intention was to characterize between the groups.
patients with idiopathic reduction in ovarian reserve. Cases involving surgi-
cal sperm retrieval or chromosomal translocations were excluded given Post-cycle diagnosis of PR
their propensity for poor blastocyst formation (Balaban et al., 2001;
Munne et al., 1998). This study was performed under institutional review In order to assess ovarian reserve based on post-treatment parameters,
board approval. the number of oocytes collected for each patient who proceeded to
The stimulation approach was selected by each patient’s primary retrieval was recorded. Again, the number of oocytes retrieved and the
correlation between oocytes and embryos derived from the same patient. percentile was ≤5 oocytes retrieved. A total of 535 patients had
Using this model, adjusted odds ratios for blastulation, aneuploidy and live oocyte yields in this range. The IQR for yield was 10 to 21 oocytes. A
birth were calculated. total of 2675 patients had oocyte yields in this range. Patients with
A subgroup analysis of patients who met the Bologna Criteria for PR to oocyte yield in the <10th percentile were slightly older and had higher
ovarian stimulation was also performed. As per the Bologna criteria, at
basal FSH levels and lower AMH levels than patients in the IQR. They
least two of the following risk factors are required: (i) female age ≥40 years
also had lower estradiol levels on the day of LH surge and their cycles
(or other risk factor for PR, such as genetic causes known to reduce follicle
produced fewer oocytes and embryos. The cycles were more likely to
number), (ii) previous PR (≤3 oocytes) and (iii) abnormal ovarian reserve
testing (AMH < 0.5 or AFC < 5–7). Given that patients over 38 and be canceled prior to embryo transfer due to global arrest of embryos.
patients with genetic risk factors for DOR were excluded, the oocyte yield As a result, the overall LBR per retrieval was lower for patients in the
in the studied cycle could only be used to qualify patients for the Bologna <10th percentile. However, the overall blastulation rate was equiva-
criteria (as opposed to a prior history of PR to stimulation). As a result, lent between the two groups. Furthermore, the LBR once a blastocyst
patients in this analysis had ≤3 oocytes retrieved and had an AMH of <0.5. was available for transfer was no different (Table II).
Blastulation rates, aneuploidy rates for PGT-A patients and LBRs per Among the subset of patients utilizing PGT-A, there were similar dif-
Table I Baseline characteristics and IVF cycle outcomes for patients with AMH levels <10th percentile compared to
patients with AMH values in the 25–75th percentile.
Continued
1494 Morin et al.
Table I Continued
Figure 1 Patients with precycle AMH levels <10th percentile have lower oocyte yields than those in the 25–75th percentile. However, on mixed
effects model, a successfully fertilized oocyte derived from a patient with AMH in the <10th percentile has the same odds of forming a quality blasto-
cyst, being euploid and producing a live birth after transfer of an euploid blastocyst.
addressing this issue have failed to account for the confounding impact The data presented here suggest that in patients <38 years old with
of age on factors such as blastulation, aneuploidy and clinical either precycle evidence of DOR or PR to stimulation, there is no
outcomes. associated qualitative decline in oocyte performance. Thus, compared
We sought to separate age-related diminution in oocyte quality by to normal responders, a fertlised oocyte retrieved from a young
studying only patients <38 years old. This age was carefully selected patient with DOR or PR is no less likely to form a quality blastocyst, be
given evidence from multiple studies that (i) average rate of follicular euploid or produce a live birth.
depletion, (ii) aneuploidy rate and (iii) and embryo arrest rate all These data contrast with prior observations that concluded that
increase significantly after age 38 (Faddy et al., 1992; Franasiak et al., aneuploid ongoing gestations and aneuploid miscarriages were more
2014; Thomas et al., 2010). Thus, if an accelerated deterioration in common in poor responders (Haadsma et al., 2010). Furthermore, the
aneuploidy, blastulation and pregnancy outcomes occurred in parallel only study to utilize trophectoderm biopsy and 24-chromosome
with DOR/PR in young patients, a unifying mechanism responsible for screening platforms to address this issue at the embryonic level sug-
both qualitative and quantitative decline would be the likely culprit. gested that aneuploidy was increased in blastocysts of patients with
However, if the blastulation, aneuploidy and pregnancy outcomes DOR (Katz-Jaffe et al., 2013). However, patients in the DOR arm in
remained consistent with age-related controls with normal response, this study were older than patients with normal ovarian reserve para-
then the mechanisms governing follicular depletion and quality para- meters, thus confounding the aneuploidy results. We also specifically
meters would appear to be divergent. included blastulation rates in the analysis, given evidence that
Young patients with DOR do not have higher aneuploidy rates 1495
Table II Baseline characteristics and IVF cycle outcomes for patients with oocyte yields in <10th percentile compared to
patients with oocyte yields in the 25–75th percentile
Characteristic <10th percentile (≤5 oocytes retrieved) 25–75th percentile (10–21 oocytes retrieved) P-value
.............................................................................................................................................................................................
All patients (including PGT-A and non-PGT-A patients)
n 535 2675
Age (y) 34.4 ± 2.9 32.9 ± 3.1 <0.001a
Duration of infertility (months) 10 ± 3.1 10 ± 3.0 0.45
Primary diagnosis DOR: 73.9% (395/535) DOR: 22.4% (599/2675) <0.001c
Male factor: 13.3% (71/535) Male factor: 48.0% (1284/2675)
Tubal factor: 3.8% (20/535) Tubal factor: 8.9% (238/2675)
Unexplained: 9.2% (49/535) Unexplained: 20.7% (554/2675)
aneuploid embryos are more likely to arrest in extended culture (Vega size. While prior literature has implied that more embryos were
et al., 2014). Thus, while ploidy was not measured in the arrested required for each live birth in these patients, our data suggest that
embryos, equivalent blastulation in both groups is indicative of similar ovarian reserve and response do not impact the anticipated compe-
aneuploidy rates in arrested embryos as well. Including blastulation in tence of a given embryo. This is a valuable information for pretreat-
the analysis also addressed nongenetic causes of embryo developmen- ment counseling.
tal competence that may be impacted by mechanisms of ovarian aging. While these data provide the largest experience comparing
However, no association was observed as blastulation rate was not extended culture, aneuploidy screening and pregnancy outcomes
different between the groups. between DOR/PR patients and normal responders, there are limita-
The equivalent aneuploidy rate in DOR/PR patients to age-related tions given the retrospective nature of the analysis. Some patients in
controls also provides some insight into the biological processes medi- the PR group may have done better in subsequent stimulations with
ating the age-related increase in meiotic errors in the oocyte. Indeed, different protocols and thus may have no longer met criteria for PR.
there is still disagreement regarding whether segregation errors in Indeed, some of these patients may have been included in the ‘poor
oocytes are a reflection of the size of the remaining follicular pool or a responder’ group but in fact behaved more similarly to their age-
function of cumulative, temporal exposure to oxidative damage and matched controls in subsequent cycles. Furthermore, given the limited
other stressors that predispose to aneuploidy (Nagaoka et al., 2012). number of growing follicles in DOR/PR patients, retrieval timing in
Our data do not support the ‘first in, last out’ hypothesis, whereby the patients with fewer follicles may have produced better average oocyte
last oocytes ovulated (irrespective of age) have the fewest recombin- developmental competence per follicle.
ation events and are thus at higher risk for segregation errors It is also important to note that the average duration of infertility in
(Henderson and Edwards, 1968). Instead, the normal age-related the patients in this study was less than 12 months (the duration
aneuploidy rate and clinical performance of oocytes in these DOR/PR required to meet the ICMART/WHO definition of infertility) (Zegers-
patients suggest that errors are more a function of time-related dam- Hochschild et al., 2009). Thus, it is possible that the favorable blastula-
age or deterioration in mechanisms responsible for maintaining cohe- tion rate, euploidy rate and LBRs described in these patients may be at
sion between sister chromatids (such as cohesins) (Liu and Keefe, least partially reflect the fact that some better prognosis patients were
2008; Chiang et al., 2011; Jeffreys et al., 2003). included in the DOR/PR groups. This is especially important in the
The ultimate test of oocyte and embryo quality is establishment of a precycle assessment, given evidence that AMH alone is poorly predict-
healthy pregnancy that progresses to delivery. While prior studies ive of fecundity (Steiner et al., 2017). However, a subset analysis of
have reported an age-independent association between DOR/PR and patients with more than 12 months of infertility demonstrated similar
decreased IRs (El Toukhy et al., 2002) and increased miscarriage rates outcomes to the entire population analyzed here, so any confounding
(Levi et al., 2001), we found no such effect. Our observation that an influence of including women with less than 12 months of infertility
oocyte retrieved from a DOR patient performs similarly to that from appears relatively limited.
age-matched controls has important implications for cycle planning. Another limitation of this study is the inability to decipher whether
Many patients bank embryos according to clinic-based projections or not patients were on oral contraceptive pills (OCPs) at the time
regarding how many embryos are needed for a patient’s ideal family their AMH level was drawn. We included all patients who had an
Young patients with DOR do not have higher aneuploidy rates 1497
AMH drawn within 6 months of their cycle starts, but we did not have Funding
any information regarding whether or not they were on OCPs at that
time. It is possible that the AMH of some patients in the <10th per- No specific research funding was used for this study.
centile group may have been influenced by suppressive effect of
OCPs. This may also partially explain why some patients in the low
AMH group had more oocytes retrieved than expected. However, Conflict of interest
this may also simply reflect the limitations of AMH to prognosticate
None.
stimulation outcomes in all patients. These are only some examples of
potential confounders that may influence our retrospective data.
There are almost certainly others that are present but difficult to con-
trol for in this retrospective study. Prospective studies will be required References
to better elucidate the issues addressed here. Abdalla H, Thum MY. An elevated basal FSH reflects a quantitative rather
Another caveat to mention is the fact that results from the blastula-
Gardner DK, Lane M, Stevens J, Schlenker T, Schoolcraft WB. Blastocyst the time of HCG administration in IVF cycles predict both ovarian
score affects implantation and pregnancy outcome: towards a single reserve and embryo morphology. Hum Reprod 2006;21:159–163.
blastocyst transfer. Fertil Steril 2000;73:1155–1158. Smeenk JM, Sweep FC, Zielhuis GA, Kremer JA, Thomas CM, Braat DD.
Gougeon A, Ecochard R, Thalabard JC. Age-related changes of the popula- Antimullerian hormone predicts ovarian responsiveness, but not
tion of human ovarian follicles: increase in the disappearance rate of embryo quality or pregnancy after in vitro fertilization or intracytoplas-
non-growing and early-growing follicles in aging women. Biol Reprod mic sperm injection. Fertil Steril 2007;87:223–226.
1994;50:653–663. Steiner AZ, Pritchard D, Stanczyk FZ, Kesner JS, Meadows JW, Herring
Haadsma ML, Mooji TM, Groen H, Burger CW, Lambalk CB, Brokemans AH, Baird DD. Association between biomarkers of ovarian reserve and
FJM, van Leeuwen FE, Bouman K, Hoek A. A reduced size of the ovarian infertility among older women of reproductive age. JAMA 2017;318:
follicle pool is associated with an increased risk of a trisomic pregnancy 1367–1376.
in IVF-treated women. Hum Reprod 2010;25:552–558. Sun W, Stegmann BJ, Henne M, Catherino MH, Segars JH. A new
Hehenkamp W, Looman C, Themmen A, de Jong F, te Velde E, approach to ovarian reserve testing. Fertil Steil 2008;90:196–202.
Broekmans F. Anti-mullerian hormone levels in the spontaneous men- Tal R, Seifer DB, Khanimov M, Malter HE, Grazi RV, Leader B.
strual cycle do not show substantial fluctuation. J Clin Endocrinol Metab Characterization of women with elevated antimullerian hormone levels