Preimplantation genetic testing for aneuploidy in poor ovarian responders with four or fewer oocytes retrieved

Abstract

Purpose

To assess whether preimplantation genetic testing for aneuploidies (PGT-A) at the blastocyst stage improves clinical outcomes compared with transfer of embryos without PGT-A in poor ovarian response (POR) patients.

Methods

Retrospective cohort study of IVF cycles from 2016 to 2019 at a single academic fertility center. IVF cycles with POR and four or fewer oocytes retrieved were stratified into PGT-A (n = 241) and non-PGT (n = 112) groups. In PGT-A cycles, trophectoderm biopsy, next-generation sequencing with 24-chromosome screening, and single euploid frozen embryo transfer were performed. In non-PGT cycles, fresh or frozen transfer of untested embryos on day 3 or 5 was performed. Main outcomes included live birth rate and miscarriage rate per retrieval.

Result(s)

Patients who underwent PGT-A cycles were significantly less likely to reach embryo transfer compared with those who underwent non-PGT cycles (13.7% vs 70.6%). The live birth rate per retrieval did not differ between the PGT-A and non-PGT groups (6.6% vs 5.4%). Overall, the miscarriage rate was low. The PGT-A group demonstrated a significantly lower miscarriage rate per retrieval (0.4% vs 3.6%) as well as per pregnancy (5.9% vs 40.0%) compared with the non-PGT group. The number needed to treat to avoid one clinical miscarriage was 31 PGT-A cycles.

Conclusion(s)

PGT-A did not improve live birth rate per retrieval in POR patients with four or fewer oocytes retrieved. PGT-A was associated with a lower miscarriage rate; however, a fairly large number of PGT-A cycles were needed to prevent one miscarriage.

Electronic supplementary material

The online version of this article (10.1007/s10815-020-01765-y) contains supplementary material, which is available to authorized users.

Introduction

Diminished ovarian reserve (DOR) and poor ovarian response (POR) are frequently encountered as barriers to infertility treatment [1], leading to persistently low success rates with IVF [2, 3]. Compounding the problem of low oocyte quantity, patients with DOR and POR are often of advanced maternal age and must also contend with poor oocyte quality and increased embryo aneuploidy [4–6]. In the last decade, whole genome amplification techniques, including comparative genomic hybridization arrays (CGH), single nucleotide polymorphism (SNP) microarrays, and next generation sequencing (NGS), have been applied to preimplantation genetic testing for aneuploidy (PGT-A) and 24-chromosome screening [7]. Although a recent randomized controlled trial suggested PGT-A did not improve overall pregnancy outcomes [8], most randomized controlled trials in the past decade examining genetic testing in trophectoderm biopsies have suggested that PGT-A enhances embryo selection, improves implantation rate, and decreases miscarriage rate per embryo transfer. However, the majority of these studies only included subjects with an adequate ovarian response, resulting in at least a modest number of blastocysts available for PGT-A [8–11]. Furthermore, in many studies, pregnancy rates were calculated in reference to only those patients who proceeded to embryo transfer on days 5/6. Data are rarely presented as success rate per oocyte retrieval, which is particularly relevant in patients with a limited number of oocytes and embryos. Due to the high likelihood of failing to obtain embryos for transfer, couples with severe DOR or POR should be counseled on outcomes per retrieval rather than per embryo transfer.

There is a lack of data regarding effectiveness of PGT-A in the severe DOR or POR patient population. Failure to reach embryo transfer because of poor or arrested embryo development is more likely in extended culture compared with cleavage-stage embryo transfer [12], especially in cycles with few day 3 cleavage-stage embryos. While PGT-A can select euploid embryos and avoid futile transfer, the PGT-A procedure is costly and time-consuming, and potentially decreases the success rate per cycle by loss of viable embryos through extended culture, biopsy, freezing, and misdiagnosis. For patients with low oocyte yield, the impact of PGT-A, in terms of live birth and miscarriage per retrieval, has not been well studied.

A significant proportion of patients at our center are poor ovarian responders who are of advanced maternal age and/or have low anti-Mullerian hormone (AMH). The decision to proceed with PGT-A versus non-PGT is a difficult choice for our patients and physicians. Some of these patients elect to undergo PGT-A despite the retrieval of four or fewer oocytes—patients we will refer as “extremely poor responders.” Counseling such patients regarding cycle outcomes with and without PGT-A is challenging in the absence of robust clinical data. To that end, we reviewed all charts of poor responder patients undergoing IVF cycles where four or fewer oocytes were retrieved and compared outcomes in patients who underwent PGT-A with those choosing to transfer unscreened embryos. The intention-to-treat analysis at the point of retrieval was performed with respect to the clinical pregnancy rate, miscarriage rate, and live birth rate.

Materials and methods

Study population

We initially identified all autologous IVF cycles with 1–4 oocytes retrieved at a single academic fertility center between June 2016 and August 2019. These identified patients were further evaluated for poor ovarian response. Only patients who met the definition of POR and had four or fewer oocytes retrieved were included in the analysis. Poor ovarian responders were defined according to Bologna ESHRE consensus [13]: two of the following three features present (i) advanced maternal age ≥ 40 years or any other risk factor for POR; (ii) a previous POR; (iii) an abnormal ovarian reserve test (ORT) with AMH < 1.1 ng/ml or AFC < 5–7 follicles; two episodes of POR after maximal stimulation or advanced maternal age and an abnormal ORT. Patients who used PGT for the indication of a known history of chromosomal translocation or single-gene disorders, for sex selection, and for embryo banking were excluded from the analysis. Patients who underwent testicular sperm extraction (TESE) for non-obstructive azoospermia were excluded from the study to avoid the confounding factor of severe male factor. The study cohort was stratified into PGT with 24-chromosome aneuploidy screening (n = 241 retrieval cycles) and non-PGT (n = 112 retrieval cycles) groups. The decision of whether to use PGT-A was made by the patient after a counseling session with a genetic counselor and reproductive endocrinology physician. Treatment outcomes were collected through December 2019, or to the birth of a live-born baby. This retrospective cohort study was performed with institutional review board approval.

IVF treatment

The ovarian stimulation protocol was chosen by each patient’s primary physician and was based on patient’s ovarian reserve (AMH and AFC). Protocols used included GnRH antagonist with or without birth control pills or estradiol-priming, GnRH agonist microflare, and Clomid or letrozole with or without combined gonadotropin. Stimulation medications included FSH (recombinant), HMG (human menopausal gonadotropin), Clomid, or letrozole. Oocyte maturation was triggered by hCG and/or GnRH agonist. Choice of medication and dose of medication were adjusted according to each patient’s response to stimulation as determined by ultrasound or serum estradiol measurement. The majority of patients who underwent non-PGT treatment received fresh embryo transfer on day 3, with only a small number of patients receiving frozen embryo transfer. Fresh embryo transfer was conducted only if the patient’s serum progesterone level was less than 1.5 ng/ml and the endometrial thickness was at least 7 mm on the day of hCG trigger. Depending on patient age and embryo quality and number, one to four embryos were transferred on day 3, and the remaining embryos were cultured until day 5–7. Blastocysts with a grade of 3CC and above subsequently underwent cryopreservation. Vaginal progesterone (100 mg Endometrin, Ferring Pharma, Inc.) was administered three times daily beginning 1 day after oocyte retrieval for luteal support in fresh transfer. In each frozen embryo transfer cycle, one to four day 3 embryos, one euploid embryo, or one unknown ploidy blastocyst were transferred. Endometrial preparation and transfer procedures were performed per standard protocols at our center. Both natural and programmed protocols were utilized in frozen embryo transfer cycle, with the aim to achieve an endometrial thickness of 7 mm by ultrasound. In the natural FET cycles, luteal support was provided with vaginal progesterone (100 mg Endometrin, Ferring Pharma, Inc.) two times per day. In programmed FET cycles, luteal support was provided with estradiol 2 mg orally three times daily, intramuscular progesterone in oil at a dose of 50 mg per night, and 100 mg vaginal progesterone twice daily, starting in the evening 5 days prior to transfer. Regardless of fresh or cryopreserved transfer, luteal phase support was continued through 10 weeks of gestation if pregnancy was achieved.

Laboratory methods

Oocyte retrieval occurred approximately 35 h after administration of hCG. Oocytes were then fertilized using either conventional insemination or ICSI. Approximately 18 h after insemination or ICSI, oocytes were examined for the presence of pronuclei. Zygotes displaying two pronuclei were group cultured at 37 °C in separate 80 ul single-microdrop Sage Single-step media in an atmosphere containing 5% O2 and 6% CO2. In PGT-A cycles, embryos were cultured to the blastocyst stage (days 5–7). Embryos that reached blastocysts were graded according to their morphological quality based on the Gardner criteria [14]. Blastocysts with a grade of 3CC and above at days 5–7 underwent trophectoderm biopsy. All blastocysts were frozen by vitrification after biopsy. Comprehensive chromosome testing was conducted by Ion GeneStudio™ S5 with the Ion Chef™ system (Thermo Fisher Scientific) at Igenomix testing center. Vitrification and thawing of embryos were performed using the Cryotech Vitrification Kit. Only one euploid blastocyst was transferred in each frozen embryo transfer cycle. In fresh embryo transfer cycles without PGT, embryos were graded on day 3 of culture by morphological criteria on the basis of the number and size of blastomeres and the percentage of fragmentation [15]. After selection of embryos for fresh transfer, the remaining embryos were either frozen on day 3, or continued to be cultured in the same media until the blastocyst stage. All viable blastocysts from PGT-A cycles and day 3 embryos from non-PGT cycles with a grade of 3CC and above at days 5–7 were frozen by vitrification as above.

The fertilization rate was calculated by the number of oocytes with two pronuclei (2PN) per number of inseminated cumulus-oocyte complexes (conventional IVF) or per the number of mature oocytes injected (ICSI). The blastocyst formation rate was analyzed based on the number of usable blastocysts that were available for biopsy or freezing per normally fertilized MII egg (2PN). A clinical pregnancy was defined as a serum quantitative hCG level > 5 mIU/ml and the presence of a gestational sac on transvaginal ultrasound at 6–7 weeks of gestation. A clinical miscarriage was defined as a loss of an intrauterine pregnancy after a gestational sac had been identified on ultrasound and between 6 and 20 weeks gestational age. Live births were defined as birth of a neonate at or beyond 24 weeks gestation and were documented by patient report. Clinical pregnancy rates and live birth rates were calculated per retrieval and per embryo transfer. Miscarriage rates were calculated per retrieval and per pregnancy.

Statistical analysis

Independent t test, Pearson’s chi-square test, or Fisher’s exact test were used for continuous or categorical variables, respectively. Results are presented as mean ± standard deviation unless otherwise stated. P < 0.05 was considered statistically significant.

Results

One hundred and sixty-five patients received PGT-A treatment and 54 of them underwent more than one PGT-A cycle. Ninety-six patients had non-PGT treatment and 16 of them had more than one non-PGT cycle. In total, 241 PGT-A cycles and 112 non-PGT cycles were included in this study. The patients’ baseline characteristics are reported in Table 1. All patients had ovarian reserve testing (anti-Mullerian hormone test) within 8 months prior to IVF treatment. Mean level of AMH were 0.7 ng/ml in the PGT-A group and 0.63 ng/ml in the non-PGT group. Patients in PGT-A groups were slightly older than patients in non-PGT group (40.8 vs 39.4, P < 0.001). The patient groups did not differ in BMI, parity, average number of previous miscarriages, and average number of previous failed IVF cycles. The most commonly selected protocol was GnRH antagonist (72.3% in PGT-A group and 71.8% in non-PGT group). There were more cycles using the microdose flare protocol with GnRH agonist in the PGT-A group compared with the non-PGT group (26.1% vs 16.1%). Clomid and letrozole with or without gonadotropins were used more in non-PGT group (8.9%) than PGT-A group (1.6%). GnRH agonist long protocol was not used in either group. The total dose of gonadotropins was higher in PGT-A groups while the serum peak estradiol level, the average number of eggs retrieved, and the average number of 2PN zygotes were similar in both groups (Table 2).

Table 1.

Clinical characteristics of patients

	PGT-A	Non PGT-A	P value
Number of patients	165	96
Age	40.8 ± 3.4	39.4 ± 4.4	< 0.001
BMI (kg/m2)	24.6 ± 5.0	24.8 ± 4.5	0.677
AMH level (ng/ml)	0.70 ± 0.74	0.63 ± 0.60	0.347
Gravida (n)	1.1 ± 1.4	1.1 ± 1.5	0.900
Parity (n)	0.4 ± 0.7	0.4 ± 0.7	0.672
Prior spontaneous miscarriage (n)	0.5 ± 0.9	0.5 ± 1.0	0.861
Prior failed IVF cycles (n)	1.4 ± 1.6	1.3 ± 1.1	0.090
Stimulation protocol			0.293
Antagonist	183/253 (72.3%)	84/112 (71.8%)
Microdose Lupron flare	66/253 (26.1%)	18/112 (16.1%)
Other	4/253 (1.6%)	10/112 (8.9%)
Gonadotropin dosage (iu)	4678 ± 1404	4184 ± 1753	0.007
Peak E2 level (pg/ml)	1090.4 ± 690.8	996.0 ± 643.6	0.191

Table 2.

Embryologic and clinical outcome of the patients

	PGT-A	Non PGT-A	P value
Number of retrieval cycles	241	112
No. of oocytes retrieved	2.7 ± 1.0	2.5 ± 1.1	0.062
Fertilization rate	66.2%	63.1%	0.331
No. of 2PNs	1.5 ± 1.1	1.4 ± 1.0	0.056
Rate of cycles with embryo available for transfer	13.7% (33/241)	70.6% (79/112)	< 0.001
No. of embryo transfer cycles	34	79
Clinical pregnancy rate per retrieval	7.1% (17/241)	8.9% (10/112)	0.526
Miscarriage rate per retrieval	0.4% (1/241)	3.6% (4/112)	0.036
Miscarriage rate per pregnancy	5.9% (1/17)	40% (4/10)	0.047
Live birth rate per retrieval	6.6% (16/241)	5.4% (6/112)	0.814

Patients who underwent PGT-A treatment had a significantly lower chance to reach embryo transfer compared with those in the non-PGT group. In the PGT-A group, 51.5% of PGT-A cycles (n = 124) had blastocysts available for biopsy on culture day 5–6. A total of 169 blastocysts were biopsied; 97.0% were informative and 21.9% were euploid (n = 37). Ultimately, 13.7% of PGT-A cycles (n = 33) had euploid embryos available for FET. In the non-PGT group, 70.6% of cycles (n = 79) had at least one embryo available for transfer. Of the 79 non-PGT cycles with at least one embryo available for transfer, 74 cycles had fresh embryo transfer on day 3 or 5, three cycles had all embryos frozen on D3, and two cycles had embryos frozen on day 5 followed by frozen embryo transfers.

The clinical pregnancy rate and live birth rate per retrieval cycle did not differ between the PGT-A (7.1% and 6.6%) and non-PGT groups (8.9% and 5.4%, P value = 0.526 and 0.814 respectively) (Table 2). Thirty-three patients in PGT-A group had 34 frozen single euploid blastocyst transfer cycles and 79 patients in non-PGT group had either fresh (n = 74) or frozen transfer (n = 5). Of the 74 fresh transfers, only 2 patients received day 5 blastocyst transfer. The remaining 72 patients received day 3 fresh embryo transfer. The mean number of transferred embryos per cycle in the non-PGT group was 1.6 (SD = 0.6). Per embryo transfer, the PGT-A group had a higher clinical pregnancy rate (50.0% vs 12.7%, P value < 0.001) and live birth rate (47.1% vs 7.6%, P value < 0.001) than the non-PGT group. The non-PGT group had a significantly higher clinical miscarriage rate per retrieval (3.6%) as well as per pregnancy (40%) compared with PGT-A group (0.4% and 5.6%, P value = 0.036 and 0.047, respectively). The number needed to treat to avoid one clinical miscarriage was 31 PGT-A cycles.

Four patients in the PGT-A group had two euploid embryos, of which one patient underwent the second embryo transfer and had no pregnancy. All patients in the non-PGT group had transfer of all available embryos at one time and none had surplus embryo for 2nd transfer. There was one twin pregnancy in the non-PGT group. There was no multi-pregnancy in the PGT-A group. Because almost all cycles had available embryos transferred at one time, the pregnancy rate per embryo transfer was almost equal to the pregnancy rate per retrieval. Therefore, the pregnancy rate per transfer almost represented the total cycle potential in both groups.

Supplemental Figure 1 is the dispersion graph that displays the distribution of embryo transfer and live birth with respect to both the woman’s age at oocyte retrieval and the number of oocytes retrieved. It clearly shows that in the non-PGT group, even though pregnancy can be obtained beyond age 38, all live birth clustered in women at age 38 or younger. In contrast, in the PGT-A group, once an euploid embryo was obtained, maternal age did not impact the chance for live birth. The difference of live birth rates in women younger than age 38 and women age ≥ 38 at egg retrieval is shown in Fig. 1. Although the sample size in age-stratified groups was not large enough to achieve statistical significance, the data suggests that PGT-A did not demonstrate superior live birth rate per retrieval compared with non-PGT in women younger than age 38; however, PGT-A demonstrated a higher live birth rate in women at 38 years old or beyond. The euploidy rates were 32.4% in women younger than age 38 and 19.2% in women with age ≥ 38.

Fig. 1.

Live birth rates in two age groups

Discussion

In the present study, we set out to report cycle outcomes in women with low egg numbers undergoing IVF with and without PGT-A, in order to provide data that can be used in counseling patients in the poor prognostic group. PGT-A was not associated with superior pregnancy rates or live birth rate per retrieval compared with non-PGT treatment within this extremely poor responder population. PGT-A was associated with a lower chance of miscarriage when compared with the IVF alone group; however, the absolute numbers of miscarriages were low in both groups, and patients would need to undergo 31 PGT-A cycles to prevent one miscarriage with no impact on live birth.

Studies that investigated PGT-A in extremely poor responders are limited. Studies that investigated PGT-A in advanced maternal age or DOR/POR have either been limited to day 3 biopsy, or been limited to normal response population with a number of oocytes retrieved that would make it reasonable to proceed with PGT-A treatment [16–18]. As such, it is challenging to extrapolate the data to a population with extremely poor response and a limited number of eggs who face the difficult decision of whether to proceed with PGT-A. Although the Bologna criteria defined poor ovarian response (POR) during IVF treatment [13], they still allow for clinical heterogeneity within the defined population [19]. Cutoff points used for AMH to predict POR have ranged from 0.10–1.66 ng/mL, with reported sensitivities of 44–97% and specificities of 41–100% [20]. In addition, despite the presence of abnormal ovarian reserve tests, a significant percentage (10–20%) of women will eventually exhibit normal ovarian response and subsequently afford better prognosis [20]. In this study, examining outcomes within the population of women with four or fewer oocytes retrieved allows us to provide a clearer framework from which to counsel such patients.

In this study, a large majority of (86.3%) patients who underwent intended PGT-A treatment did not obtain euploid blastocyst for transfer. If one euploid embryo was obtained and transferred, then the live-birth rate per embryo transfer was 50.0%, similar to the rate reported in the literature for the female patient population across the board regardless of age or AMA, good or poor prognosis after PGT-A cycles [21]. However, because of the low chance to reach the stage of euploid embryo transfer, the overall live birth rate per retrieval was only 6.6%. Therefore, PGT-A did not improve the live birth rate per retrieval compared with the non-PGT group. Notably, although the miscarriage rate per pregnancy was higher in non-PGT group, the overall miscarriage rates per retrieval in both PGT-A and non-PGT groups were very low (0.4% and 3.6%), which is associated with the low odds of pregnancy in this specific population. Furthermore, PGT-A is costly, adds time to diagnosis, and requires additional resources. With no improvement in live birth rate and a small difference in miscarriage, the number needed to treat to prevent one miscarriage was 31 PGT-A cycles. Therefore, adding PGT-A to cycles with four or fewer eggs may not to be a cost-effective strategy.

When the analysis was performed by stratifying patients by age at retrieval, the results show that in women with the same low ovarian reserve, younger age is associated with lower rate of aneuploidy in blastocysts and is a significant protective factor in achieving live birth. Aneuploidy rate in the group of women younger than age 38 was 32.4%, which was similar to aneuploidy rates found in previous studies where younger women with DOR and POR exhibited equivalent blastulation rates, aneuploidy rates, and live birth rates compared with age-matched controls with normal ovarian reserve [22]. In this study, PGT-A did not significantly improve live birth rate compared with non-PGT treatment in patients younger than age 38. This is consistent with prior studies that have showed PGT-A blastocyst selection does not result in an enhanced live birth rate in young infertile women [18]. There was one twin pregnancy in non-PGT group and no multi-pregnancy in PGT-A group. Due to the limited number of cases, we are unable to make conclusions comparing multiple pregnancy rates between the two groups.

In the non-PGT group, no live birth occurred in women with age ≥ 38 after 49 embryo transfer cycles. There were two clinical pregnancies that ended with spontaneous miscarriage in the non-PGT group. Notably, due to the limited number of embryos available for transfer, all patients in non-PGT group transferred their all available embryos at one time. In contrast, 11 live births after 24 FET cycles were achieved in women with age ≥ 38 in the PGT-A group. Because PGT-A allows for selection of the most viable embryos but does not increase the number of euploid embryos present, PGT-A is not expected to improve the outcome when the total cycle potential or cumulative delivery rate is considered. Although the sample size in each age group was not large enough to achieve statistical significance (requiring at least 135 cycles in each treatment group), the data still suggested that patients with age ≥ 38 in the non-PGT group had a much lower pregnancy rate and live birth rate than those in PGT-A group.

We compared serum progesterone (P4) level on the day of trigger in PGT-A vs non-PGT patients in this age group. The mean progesterone level was lower in non-PGT group (0.5 ng/ml) than in the PGT-A group (0.63 ng/ml), although progesterone level does not appear to have clinical significance in terms of affecting embryo implantation and pregnancy rates. There was a higher gonadotropin dosage administered in PGT-A group, suggesting intentional protocol-driven effects by the treating clinician to attempt to achieve a sufficient number of embryos for testing. The rest of the clinical characteristics, including age, AMH level, mean 2PN number, and prior history of failed IVF cycles, were not significantly different between the two groups, and did not suggest patient selection bias. However, we cannot exclude the possibility that patients with age ≥ 38 with a poorer prognosis were disproportionately diverted to non-PGT group. Another possible explanation of the lower pregnancy rate in non-PGT in the older age group could be the less natural endometrial environment that was associated with fresh embryo transfer. Numerous studies have suggested that frozen embryo transfer might improve pregnancy rates compared with fresh transfer [23], although there is insufficient data to explain the lack of this effect in younger patients in this study. We did not have enough frozen embryo transfers after non-PGT to assess the isolated effect of PGT-A. In the recent randomized controlled trial published by Munné, S et al. [8], compared with control, use of PGT-A was not associated with improvement in ongoing pregnancy rate with the intention to treat in women aged 35–40 years. Of note, patients in that study had at least two blastocysts of sufficient quality for biopsy and vitrification by day 6. In addition, both PGT-A and control groups underwent frozen-thaw embryo transfers. In the real world, poor ovarian response patients and physicians tend to select fresh embryo transfer on day 3 to save time and cost and avoid extended embryo culture and potential loss of embryos. Given the suggested findings of lower pregnancy rate in older patients in non-PGT group in our study, further research to compare fresh embryo transfer with frozen transfer without PGT-A in this specific age group might provide a clearer recommendation regarding use of PGT-A in older patients.

Our study was limited by its retrospective nature. Although our study is one of the few studies to address the controversy of whether PGT-A should be pursued in the context of an extremely low number of oocytes, the sample size was still suboptimal. We were not able to exclude patient selection bias despite no significant clinical characteristics being observed between the two groups. We excluded patients who underwent embryo banking from the study, which potentially biased the results. Recently reported studies demonstrated that older and poorer prognosis IVF patients are disproportionately diverted to embryo banking [24, 25]. Another limitation was that we were not able to assess the time to live birth in this research due to inadequate data. Many argue that time to a successful pregnancy is shorter with PGT-A with respect to fewer embryo transfers and less time lost with a miscarriage. However, in a population in which most patients will require multiple IVF-PGT attempts to establish a successful pregnancy, the amount of time spent on PGT-A treatment may exceed the time that is saved by avoiding a miscarriage. Additional research requiring repeated cycles to be performed within a certain time frame or a set number of cycles to allow for cost-effectiveness analysis is needed to answer this important question. Despite the above limitations, in the absence of a large prospective trial, our study provided data that can be used to guide our consultation in our day-to-day clinical practice.

In conclusion, proceeding with PGT-A in extremely poor ovarian responders with four or fewer oocytes retrieved did not provide better live birth rate per intent-to-treat cycle compared with untested embryo transfers on day 3 or day 5. Overall, live birth rate and miscarriage rate per retrieval were low in both PGT-A and non-PGT groups. Although PGT-A was associated with reduced odds of miscarriage compared with non-PGT, a fairly large number of PGT-A cycles would be needed to prevent one miscarriage in this specific patient population. The decision to proceed with PGT-A in the setting of low oocyte numbers should take into account patient preferences and values as well as expected outcomes of each decision. We hope that the information provided here will aide in these discussions. Future research regarding patient experience, time to pregnancy, and cost effectiveness are needed.

Authors’ contribution

J.D. performed study design, execution, statistical analysis, and wrote the manuscript. H.H., Q.Z., A.N., and S.A. coordinated data collection and statistical analysis. B.B. and R.L. contributed interpretation of the data and edit of the manuscript. We thank Stephanie A. Leonard for assistance in statistics. All authors reviewed the manuscript and provided critical feedback and discussion.

Data availability

The material contained in the manuscript has not been published, has not been submitted, or is not being submitted elsewhere for publication.

Code availability

N/A.

Compliance with ethical standards

Conflict of interest

The authors declare that they do not have conflicts of interest.

Ethics approval

This research is approved by IRB at Stanford University.

Consent to participate

Waivered.