Introduction
In 2021, the American Society for Reproductive Medicine (ASRM) updated its guidance on the limits to the number of embryos to transfer, recommending single embryo transfer (SET) as the preferred strategy to promote singleton gestation and reduce multiple pregnancies in assisted reproductive technology (ART).1 Although this consensus provided evidence-based guidance for the clinical application of SET in selected patients, several barriers to its widespread implementation remain. Minimizing multiple-pregnancy-related obstetric complications while safeguarding maternal and neonatal health—without compromising in vitro fertilization (IVF) pregnancy rates—remains a major challenge in contemporary ART practice.
Maternal age is a well-established independent risk factor affecting fertility and pregnancy outcomes. Chinese clinical practice guidelines for ART define advanced maternal age (AMA) as ≥35 years.2 However, many women maintain a relatively high ovarian reserve beyond age 35, with a marked decline often not occurring until approximately age 38.3 Consequently, some investigators define AMA in infertile women as ≥38 years, although the optimal threshold remains a subject of debate. According to the age-stratified guidelines of the ASRM regarding limits on the number of embryos to transfer, elective SET is recommended for patients with a favorable prognosis aged <35 years as well as those aged 35–37 years, with the latter group representing a critical clinical decision point for SET implementation.1 In 2024, our team reported that among women aged 35–37 years undergoing high-quality single blastocyst transfer (SBT) in frozen-thawed embryo transfer (FET) cycles, pregnancy outcomes were comparable to those of younger patients.4
Nevertheless, several uncertainties persist. Previous studies evaluating age-related outcomes following SBT have been predominantly based on FET cycles; these findings cannot be directly extrapolated to fresh transfer cycles due to fundamental differences in the endometrial environment, embryo–endometrial synchrony, and endocrine physiology between fresh and frozen cycles. Moreover, women aged 35–37 years constitute a clinically important transitional group, yet their pregnancy and neonatal outcomes following fresh high-quality SBT remain poorly characterized.
Therefore, it remains unclear whether clinical outcomes after fresh high-quality SBT differ substantially across the narrow maternal age strata of <30, 30–34, and 35–37 years. To address this knowledge gap, we conducted a large retrospective cohort study comparing pregnancy and neonatal outcomes among 4352 women who underwent fresh high-quality SBT at a single tertiary reproductive medicine center.
Materials and Methods
Study Population
This retrospective cohort study was conducted at the Reproductive Center of the Second Affiliated Hospital of Wenzhou Medical University. The study population comprised 4352 women who underwent fresh embryo transfer between January 2016 and June 2023. Women were included if they met all of the following criteria: (1) age ≤37 years; (2) endometrial thickness ≥7 mm on the day of embryo transfer; (3) no more than two previous embryo transfer cycles; (4) transfer of a single high-quality blastocyst on Day 5; and (5) controlled ovarian hyperstimulation (COH) using an early follicular phase long-acting gonadotropin-releasing hormone agonist (GnRH-a) long (EFLL) protocol.
Exclusion criteria were as follows: (1) uterine developmental abnormalities, endometrial polyps, intrauterine adhesions, adenomyosis, or submucosal fibroids resulting in endometrial compression; (2) a history of genetic disease, malignancy, autoimmune disease, or hematological disorder in the woman or her partner; and (3) a history of recurrent miscarriage.
The EFLL protocol was selected because it is the standard COH regimen used in our center and provides consistent follicular development and endometrial preparation. We restricted the study to women who received a single high-quality Day-5 blastocyst in a fresh transfer cycle to minimize confounding by embryo quality, transfer strategy, and stimulation protocol, thereby improving the internal validity of the age-group comparisons. However, this restriction may limit the generalizability of the findings to other stimulation protocols or transfer strategies, which is acknowledged as a limitation.
Eligible patients were stratified into three age groups: <30 years (Group A, n = 1939), 30–34 years (Group B, n = 1801), and 35–37 years (Group C, n = 612) (Figure 1). The grouping was based on the 35–37-year cutoff recommended in the ASRM guidelines; younger patients were further divided into <30 years and 30–34 years to test for prognostic gradients within the younger population, and women aged ≥38 years were excluded to reduce age-related heterogeneity in outcomes. The study was conducted in strict compliance with the ethical principles of the Declaration of Helsinki of the World Medical Association.
Figure 1Flowchart of patient selection and grouping. A total of 4352 eligible patients who received assisted reproductive technology treatment at the reproductive center of the Second Affiliated Hospital of Wenzhou Medical University between January 2016 and June 2023 were enrolled and stratified into three groups according to maternal age: <30 years (group A, n = 1939), 30–34 years (group B, n = 1801), and 35–37 years (group C, n = 612). Patient data were compared among the three groups.Flowchart sequence: Fresh single blastocyst transfer cycles (n equals 4,352), less than 30 years (group A, n equals 1,939), 30-34 years (group B, n equals 1,801), 35-37 years (group C, n equals 612), One-way ANOVA analysis, Chi-squared test, Compare baseline characteristics pregnancy outcomes and neonatal outcomes, Logistic regression analysis, Excluding cycles of beta-hCG less than 15mIU/mL (n equals 1,091), Adverse pregnancy outcomes (n equals 3,261), Live birth (n equals 4,352), Excluding non-live birth (n equals 1,997), Obstetric and neonatal complications (n equals 2,355).Flowchart sequence: Fresh single blastocyst transfer cycles grouped by age, then analyses and outcomes.
COH Protocol
All patients underwent the EFLL protocol. A single 3.75 mg dose of a GnRH-a (Decapeptyl, Ferring GmbH, Germany) was administered on days 3–5 of the menstrual cycle. Pituitary desensitization was assessed approximately 30–32 days later.
Ovarian stimulation with gonadotropins (Gn) was initiated once downregulation was confirmed, defined as serum follicle-stimulating hormone (FSH) <5 mIU/mL, luteinizing hormone (LH) <5 mIU/mL, and estradiol (E2) <50 pg/mL, together with transvaginal ultrasound findings of no follicles >10 mm in diameter and endometrial thickness <5 mm. The type and dosage of Gn were individualized according to patient characteristics, including age, anti-Müllerian hormone (AMH) level, body mass index (BMI), antral follicle count (AFC), and baseline FSH level.
Follicular growth was monitored by ultrasound beginning on day 5 or 6 of Gn stimulation, and the Gn regimen was subsequently adjusted based on ultrasound findings and serial serum hormone measurements. Final oocyte maturation was triggered with either 4000–10000 IU of human chorionic gonadotropin (hCG) or 250 μg of recombinant hCG when at least two to three follicles reached ≥18 mm in diameter. A pregnancy test was performed before Gn initiation if the post-downregulation endometrial thickness exceeded 5 mm (Figure 2).
Figure 2Controlled ovarian hyperstimulation (COH) protocol and follow-up. A single 3.75 mg dose of a gonadotropin-releasing hormone agonist (GnRH-a) was administered on days 3–5 of the menstrual cycle. Approximately 30–32 days later, once pituitary desensitization was confirmed, ovarian stimulation was initiated with individually selected gonadotropins (gn).
Ultrasound monitoring was started on day 5 or 6 of Gn stimulation, and the type and dose of Gn were adjusted according to ultrasound findings and serum hormone levels. The oocyte maturation trigger was administered when at least two to three follicles reached ≥18 mm in diameter. Ultrasound-guided transvaginal oocyte retrieval was performed 36–38 hours after trigger administration.
Luteal phase support was initiated on the day of oocyte retrieval. on day 5 after oocyte retrieval, a single high-quality blastocyst was selected for embryo transfer under abdominal ultrasound guidance. Luteal phase support medications were continued unchanged after embryo transfer. Pregnancy was determined by measuring serum β-human chorionic gonadotropin (β-hCG) levels on days 10–12 after embryo transfer.
The first ultrasound examination was performed 14 days after the β-hCG test, and the second ultrasound examination was performed 14 days after the first.A timeline diagram of controlled ovarian hyperstimulation steps from period day 3 to follow up ultrasounds.
Oocyte Retrieval and Embryo Transfer
Transvaginal ultrasound-guided oocyte retrieval was performed 36–38 hours after the hCG trigger. Retrieved oocytes were fertilized via conventional IVF or intracytoplasmic sperm injection (ICSI). Normal fertilization was confirmed 16–18 hours later by the presence of two pronuclei (2PN).
Embryo development was assessed on Day 3 (cleavage stage) and Day 5/6 (blastocyst stage). Blastocyst quality was evaluated using the Gardner grading system.5 Blastocysts were assigned a numerical score (1–6) based on the degree of expansion and hatching status, along with letter grades for the inner cell mass (ICM) and trophectoderm (TE) (A, B, or C). For this study, a high-quality blastocyst was defined as one with a stage ≥3 and ICM and TE grades of BB or better.6 A single high-quality blastocyst was selected for fresh transfer on Day 5 under abdominal ultrasound guidance.
Luteal phase support was initiated on the day of oocyte retrieval and consisted of oral dydrogesterone (20 mg twice daily) and micronized progesterone (200 mg administered orally or vaginally twice daily); this regimen was continued unchanged after embryo transfer (Figure 2).
Assessment of Pregnancy Outcomes
Follow-up was conducted via telephone to record pregnancy outcomes, delivery details, and neonatal status. Serum β-human chorionic gonadotropin (β-hCG) levels were measured 10–12 days after embryo transfer for pregnancy determination. Ultrasonography was performed 14 days after a positive β-hCG test to confirm the initial pregnancy, followed by a second scan 14 days later to assess ongoing viability (Figure 2).
According to the ninth edition of the textbook Obstetrics and Gynecology, published by the People’s Medical Publishing House in China,7 the outcome definitions were as follows:
Biochemical pregnancy: serum β-hCG level ≥15 mIU/mL within one month of transfer, without subsequent ultrasound evidence of a gestational sac.Clinical pregnancy: ultrasound evidence of a gestational sac with cardiac activity in the uterine cavity.Ectopic pregnancy: implantation of a gestational sac outside the uterine cavity.Miscarriage: pregnancy loss before 28 weeks of gestation with fetal weight <1000 g.Preterm birth: delivery between 28 and 37 weeks of gestation.Low birth weight: newborn weight <2500 g at birth.Macrosomia: newborn weight ≥4000 g at birth.
The three study outcomes were classified a priori as follows: (1) the primary outcome was non-live birth, defined as failure to achieve a live birth after embryo transfer; (2) the secondary outcome was adverse pregnancy outcomes, evaluated among women with a positive β-hCG result and defined as biochemical pregnancy, miscarriage, ectopic pregnancy, or other pregnancy-related complications resulting in pregnancy failure. Because these events represent a continuum of early pregnancy failure, they were combined into a composite secondary outcome to reduce the number of statistical tests; and (3) exploratory outcomes were obstetric and neonatal complications, evaluated among women who achieved a live birth and defined as the occurrence of gestational diabetes, gestational hypertension, preterm birth, low birth weight, birth defects, or other major obstetric/neonatal adverse events.
Statistical Methods
Statistical analyses were performed using SPSS software version 26.0 (IBM Corp., Armonk, NY, USA). Normally distributed continuous variables are presented as mean ± standard deviation and were compared among the three age groups using one-way analysis of variance (ANOVA). Non-normally distributed continuous variables are expressed as median with interquartile range (IQR), and comparisons were performed using the Kruskal–Wallis test.
Categorical variables are presented as frequencies and percentages and were compared using the chi-square test or Fisher’s exact test, as appropriate. All tests were two-tailed, and a P value <0.05 was considered statistically significant. For multi-group comparisons of pregnancy and neonatal outcomes, post hoc pairwise comparisons were conducted with Bonferroni correction.
Because 18 outcomes were compared among the three age groups, the Bonferroni-corrected threshold for pairwise comparisons was set at P < 0.001 (ie, 0.05 divided by 54 pairwise comparisons). The primary outcome was tested without adjustment; secondary and exploratory comparisons were interpreted against the Bonferroni-corrected threshold. Pairwise comparisons of baseline and laboratory characteristics in Tables 1–2 are reported with unadjusted P values for descriptive purposes.
Variables with P < 0.05 in univariable analysis, together with female age groups, were entered into multivariable logistic regression models using the forced-entry method. Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated from these models, and the Wald test was used for statistical inference. Model fit was assessed using the Hosmer–Lemeshow goodness-of-fit test, which indicated no significant lack of fit (P > 0.05).Table 1Baseline CharacteristicsGroup A (n=1939)Group B (n=1801)Group C (n=612)P valueFemale age, median (IQR)(years)U27.6(25.9,28.8)a,b32.2(31.0,33.5)c36.2(35.6,37.0)<0.001*Male age, mean (SD) (years)U30.02±3.44a,b34.06±3.42c37.67±3.67<0.001*Duration of infertility, median (IQR)(years)U3.0(1.0,4.0)a,b3.0(2.0,5.0)3.0(2.0,5.0)<0.001*AMH, median (IQR)(ng/ml)U3.56(2.56,5.26)a,b3.19(2.16,4.60)c2.58(1.80,3.67)<0.001*BMI, mean (SD)(kg/m2)21.91±3.29b22.06±3.1922.28±3.120.041*Infertility type<0.001* Primary infertility %(n)61.17(1186/1939)a,b40.92(737/1801)c21.57(132/612) Secondary infertility %(n)38.83(753/1939)a,b59.08(1064/1801)c78.43(480/612)Fertilization method<0.001* IVF %(n)81.07(1572/1939)a85.73(1544/1801)84.31(516/612) ICSI %(n)18.93(367/1939)a14.27(257/1801)15.69(96/612)Testicular sperm extraction<0.001* Yes %(n)4.38(85/1939)a,b1.67(30/1801)1.80(11/612) No %(n)95.62(1854/1939)a,b98.33(1771/1801)98.20(601/612)Transfer cycle<0.001* First cycle %(n)97.52(1891/1939)a,b95.34(1717/1801)94.60(579/612) Second cycle %(n)2.48(48/1939)a,b4.66(84/1801)5.40(33/612)Infertility causes Female factor %(n)64.88(1258/1939)b62.19(1120/1801)58.33(357/612)0.011* Male factor %(n)10.93(212/1939)9.55(172/1801)9.97(61/612)0.368 Combined-factor%(n)^10.57(205/1939)8.77(158/1801)9.80(60/612)0.178 Unexplained factor %(n)13.62(264/1939)a,b19.49(351/1801)21.90(134/612)<0.001*Basal hormone levels LH, mean (SD) (IU/L)5.16±2.895.07±3.115.04±2.710.545 FSH, mean (SD)(IU/L)6.88±2.00a,b7.08±1.987.22±1.96<0.001* E2, median (IQR)(pg/mL)U40.58(32.15,51.53)41.87(32.56,52.69)42.29(32.29,54.16)0.058 P, mean (SD)(ng/mL)0.52±0.21a,b0.50±0.210.50±0.200.010* PRL, mean (SD)(mIU/L)13.13±6.3212.69±6.6212.89±6.720.692Notes: ^Combined-factor was defined as having more than one cause of infertility. *P < 0.05 was considered statistically significant.
“a” represents P value less than 0.05 between groups A and B, “b” represents P value less than 0.05 between groups A and C, “c” represents P value less than 0.05 between groups B and C. U: Evaluated across groups using the Kruskal–Wallis H-test, with pairwise comparisons performed using the Mann–Whitney U-test.Abbreviation: SD, Standard deviation; IQR, Interquartile range; AMH, Anti-Müllerian hormone; BMI, Body mass index; IVF, In vitro fertilization; ICSI, Intracytoplasmic sperm injection; LH, Luteinizing hormone; FSH, Follicle-stimulating hormone; E2, Estradiol; P, Progesterone; PRL, Prolactin.Table 2Ovarian Stimulation Characteristics and Laboratory OutcomesGroup A (n=1939)Group B (n=1801)Group C (n=612)P valueDownregulated hormone levels LH, mean (SD)(IU/L)0.45±0.290.44±0.280.43±0.380.130 E2, mean (SD)(pg/mL)24.26±11.0623.94±10.6623.93±11.040.629 P, median (IQR)(ng/mL)U0.43(0.30,0.59)a,b0.40(0.28,0.55)c0.37(0.27,0.50)<0.001*Hormone levels on hCG day LH, mean (SD) (IU/L)0.71±0.550.71±0.530.67±0.530.270 E2, mean (SD)(pg/mL)2526.52±1210.89b2475.65±1170.032380.74±1195.920.029* P, mean (SD) (ng/mL)0.67±0.290.68±0.300.69±0.300.132Endometrial thickness on hCG day, median(IQR)(mm)U11.10(9.80,12.60)11.00(9.70,12.70)10.80(9.40,12.70)0.096Incidence of OHSS %(n)1.75(34/1939)1.44(26/1801)0.65(4/612)0.142Gn dosage, median (IQR) (IU)U2025(1575,2625)a,b2400(1800,2987.5)c2800(2250,3300)<0.001*Duration of Gn treatment, median (IQR) (day)U11.0(10.0,13.0)11.0(10.0,13.0)11.0(10.0,13.0)0.065No. of oocytes retrieved, mean (SD)14.82±5.30a,b14.12±5.31c13.28±5.32<0.001*No. of mature oocytes, mean (SD)13.12±4.86a,b12.55±4.91c11.84±4.81<0.001*No. of normally fertilized oocytes, mean (SD)10.62±4.29a,b10.23±4.36c9.72±4.22<0.001*No. of cleavage embryos, mean (SD)10.36±4.22a,b10.02±4.30c9.52±4.18<0.001*No. of frozen embryos, median (IQR)U4.0(3.0,7.0)a,b4.0(2.0,6.0)c4.0(2.0,5.0)<0.001*Transferred blastocyst grade AA %(n)2.01(39/1939)a1.00(18/1801)0.98(6/612)0.020* AB %(n)26.35(511/1939)a,b22.32(402/1801)c20.26(124/612)0.001* BA %(n)26.00(504/1939)27.04(487/1801)27.45(168/612)0.684 BB %(n)45.64(885/1939)a,b49.64(894/1801)51.31(314/612)0.012*Notes: *P < 0.05 was considered statistically significant. “a” represents P value less than 0.05 between groups A and B, “b” represents P value less than 0.05 between groups A and C, “c” represents P value less than 0.05 between groups B and C.
U: Evaluated across groups using the Kruskal–Wallis H-test, with pairwise comparisons performed using the Mann–Whitney U-test.Abbreviation: SD, Standard deviation; IQR, Interquartile range; LH, luteinizing hormone; E2, Estradiol; P, Progesterone; hCG, Human chorionic gonadotropin; OHSS, Ovarian hyperstimulation syndrome; Gn, Gonadotropin.
Results
Baseline Characteristics
Baseline characteristics are summarized in Table 1. Group A comprised the youngest women and had the highest AMH levels, the highest proportion of primary infertility (61.17%), the highest proportion of cycles involving testicular sperm extraction (4.38%), the highest proportion of first transfer cycles (97.52%), and the lowest proportion of unexplained infertility (13.62%). With increasing age, AMH levels decreased, while basal FSH levels increased.
Group C had a higher BMI than Group A. No significant differences were observed among the groups regarding male-factor or combined-factor infertility. The remaining baseline variables are presented in Table 1.
Ovarian Stimulation Characteristics and Laboratory Outcomes
Stimulation and laboratory outcomes are shown in Table 2. With increasing age, total Gn dose increased (median 2025 IU vs 2400 IU vs 2800 IU), whereas the numbers of oocytes retrieved, mature oocytes, fertilized oocytes, cleavage-stage embryos, and frozen embryos decreased progressively. The proportion of top-quality blastocysts (AA and AB) was highest in Group A and declined with age.
Endometrial thickness on the day of hCG trigger, OHSS incidence, and duration of Gn treatment were comparable among the three groups. Additional comparisons are presented in Table 2.
Pregnancy and Neonatal Outcomes
Pregnancy and neonatal outcomes are summarized in Table 3, with unadjusted P values shown for between-group comparisons. After Bonferroni correction (P < 0.001), the following differences remained statistically significant: Group C had a lower positive hCG rate (67.65%) and clinical pregnancy rate (55.39%) than Groups A (76.84% and 64.26%) and B (75.35% and 63.58%) (all P < 0.001). The miscarriage rate in Group C (20.94%) remained significantly higher than that in Group A (11.48%, P < 0.001), but the difference from Group B (14.06%, unadjusted P = 0.002) did not remain significant after correction.
The live birth rate, the primary outcome, was significantly lower in Group C (43.79%) than in Groups A (56.88%) and B (54.64%) (all P < 0.001). The spontaneous delivery rate was higher in Group A (52.77%) than in Groups B (42.48%) and C (39.55%) (all P < 0.001).Table 3Pregnancy and Neonatal OutcomesGroup A (n=1939)Group B (n=1801)Group C (n=612)P valuePositive rate of hCG test %(n)76.84(1490/1939)b75.35(1357/1801)c67.65(414/612)<0.001**Clinical pregnancy rate %(n)64.26(1246/1939)b63.58(1145/1801)c55.39(339/612)<0.001**Biochemical pregnancy rate %(n)12.17(236/1939)11.27(203/1801)11.93(73/612)0.689Miscarriage rate %(n)11.48(143/1246)b14.06(161/1145)c20.94(71/339)<0.001**Ectopic pregnancy rate %(n)0.41(8/1939)0.50(9/1801)0.33(2/612)0.904Twin pregnancy rate %(n)2.73(34/1246)3.14(36/1145)2.06(7/339)0.554Preterm birth rate%(n)5.80(64/1103)b6.81(67/984)10.07(27/268)0.043*Live birth rate %(n)56.88(1103/1939)b54.64(984/1801)c43.79(268/612)<0.001**Neonatal birth weight, mean (SD)(g)3239.79±533.643226.06±539.643236.77±545.970.180Gestational age at birth, mean (SD)(weeks)39.09±1.67b39.02±1.6738.80±1.620.045*Incidence of macrosomia %(n)6.80(77/1133)6.69(68/1016)7.69(21/273)0.840Incidence of low birth weight infants %(n)8.21(93/1133)8.37(85/1016)8.79(24/273)0.952Birth defect rate %(n)1.50(17/1133)a0.49(5/1016)0(0/273)0.014*Neonatal sex ratio %(n)0.732 Male56.75(643/1133)58.37(593/1016)56.78(155/273) Female43.25(490/1133)41.63(423/1016)43.22(118/273)Delivery mode<0.001** Spontaneous delivery %(n)52.77(582/1103)a,b42.48(418/984)39.55(106/268) Cesarean section %(n)47.23(521/1103)a,b57.52(566/984)60.45(162/268)Obstetric complications GDM %(n)2.33(29/1246)b3.58(41/1145)5.31(18/339)0.015* ICP %(n)0(0/1246)0.09(1/1145)0(0/339)0.554 Gestational hypertension %(n)0.80(10/1246)0.52(6/1145)0.59(2/339)0.717Notes: Outcomes are classified as primary (non-live birth), secondary (adverse pregnancy outcomes), or exploratory (preterm birth, GDM, birth defects, macrosomia, low birth weight, and other perinatal indicators). **After Bonferroni correction for 54 pairwise comparisons, the significance threshold was P <0.001. *P < 0.05 in unadjusted between-group comparison; these exploratory comparisons did not remain statistically significant after Bonferroni correction (threshold P < 0.001). “a” represents P value less than 0.05 between groups A and B, “b” represents P value less than 0.05 between groups A and C, “c” represents P value less than 0.05 between groups B and C.Abbrevations: SD, Standard deviation; hCG, Human chorionic gonadotropin; GDM, Gestational diabetes mellitus; ICP, Intrahepatic cholestasis of pregnancy.
Some exploratory differences did not remain significant after Bonferroni correction: preterm birth (Group C vs A: 10.07% vs 5.80%, unadjusted P = 0.012), gestational age at birth (Group A vs C: 39.09 ± 1.67 vs 38.80 ± 1.62 weeks, unadjusted P = 0.039), birth defects (Group A vs B: 1.50% vs 0.49%, unadjusted P = 0.020), and gestational diabetes mellitus (Group C vs A: 5.31% vs 2.33%, unadjusted P = 0.004). These exploratory findings were not significant after correction and should be interpreted cautiously.
Factors Associated with Non-Live Birth
As the prespecified primary outcome, non-live birth was analyzed in all 4352 patients. Univariable logistic regression showed that male age, female age, AMH, transferred blastocyst grade, polycystic ovary syndrome (PCOS), Gn dose, endometrial thickness, and E2 level on the day of hCG trigger were significantly associated with non-live birth. These variables were entered into a multivariable logistic regression model, with female age group forced into the model.
After multivariable adjustment, the 30–34-year group did not have a significantly increased risk of non-live birth compared with the <30-year group (OR = 1.046, 95% CI 0.905–1.208, P = 0.545), whereas the 35–37-year group had significantly higher odds (OR = 1.539, 95% CI 1.220–1.942, P < 0.001). An E2 level >4000 pg/mL on the day of hCG trigger remained an independent risk factor for non-live birth (OR = 1.413, 95% CI 1.140–1.751, P = 0.002). An endometrial thickness of 8–12 mm on the day of hCG trigger remained a protective factor (OR = 0.844, 95% CI 0.737–0.967, P = 0.014).
Male age, AMH, PCOS, transferred blastocyst grade, and Gn dose were retained for adjustment but did not reach statistical significance in the multivariable model (all P > 0.05) (Table 4).Table 4Factors Associated with Non-Live BirthUnivariable OR (95%CI)P valueMultivariable OR (95%CI)P valueEntered into Multivariable Model?Infertility duration (years) <3RefNo ≥31.020 (0.905–1.150)0.742NoBMI (kg/m2) <18.5RefNo 18.5≤BMI<240.974 (0.807–1.175)0.783No 24≤BMI<281.010 (0.811–1.256)0.932No ≥280.987 (0.725–1.345)0.936NoMale age (years) <35RefYes 35–401.251 (1.093–1.431)0.001*1.068 (0.906–1.259)0.435Yes >401.268 (0.947–1.698)0.1111.013 (0.728–1.410)0.938YesFemale age (years) <30RefYes 30–341.095 (0.963–1.246)0.1671.046 (0.905–1.208)0.545Yes 35–371.694 (1.410–2.034)<0.001*1.539 (1.220–1.942)<0.001*YesAMH (ng/ml) <2RefYes 2–50.817 (0.692–0.965)0.017*0.874 (0.732–1.044)0.137Yes >50.733 (0.604–0.890)0.002*0.848 (0.672–1.071)0.166YesInfertility type Primary infertilityRefNo Secondary infertility1.034 (0.917–1.165)0.584NoTransfer cycle First cycleRefNo Second cycle1.087 (0.796–1.484)0.601NoTransferred blastocyst grade AARefYes AB1.569 (0.922–2.672)0.0971.369 (0.722–2.430)0.283Yes BA1.351 (0.794–2.297)0.2671.192 (0.672–2.114)0.548Yes BB1.751 (1.036–2.961)0.036*1.523 (0.864–2.686)0.146YesTESE NoRefNo Yes1.185 (0.831–1.689)0.348NoICSI NoRefNo Yes1.088 (0.927–1.277)0.302NoPCOS NoRefYes Yes0.821 (0.710–0.950)0.008*0.901 (0.754–1.076)0.249YesEndometriosis NoRefNo Yes1.047 (0.809–1.354)0.727NoMale factor NoRefNo Yes1.127 (0.971–1.308)0.115NoOHSS NoRefNo Yes1.041 (0.635–1.707)0.874NoNo. of oocytes retrieved ≤10RefNo 11–140.994 (0.853–1.159)0.940No ≥151.082 (0.928–1.262)0.314NoGn dosage (IU) ≤2000RefYes 2001–40001.115 (0.983–1.264)0.0900.998 (0.863–1.154)0.978Yes >40001.438 (1.046–1.978)0.025*1.140 (0.808–1.608)0.457YesDuration of Gn treatment (day) ≤10RefNo 11–141.001 (0.877–1.141)0.993No ≥150.970 (0.768–1.224)0.796NoEndometrial thickness on hCG day (mm) <8RefYes 8–120.816 (0.716–0.929)0.002*0.844 (0.737–0.967)0.014*Yes >120.736 (0.558–0.971)0.030*0.758 (0.569–1.009)0.057YesE2 level on hCG day (pg/mL) ≤2000RefYes 2001–40001.064 (0.935–1.210)0.3481.118 (0.976–1.280)0.108Yes >40001.315 (1.072–1.612)0.008*1.413 (1.140–1.751)0.002*YesNotes: Model fit for Table 4: Hosmer–Lemeshow χ2 = 5.300, P = 0.725. The multivariable model was adjusted for male age, female age, AMH, transferred blastocyst grade, PCOS, Gn dosage, endometrial thickness, and E2 level on hCG day. *P < 0.05 was considered statistically significant.Abbrevation: BMI, Body mass index; AMH, Anti-Müllerian hormone; TESE, Testicular sperm extraction; ICSI, Intracytoplasmic sperm injection; PCOS, Polycystic ovary syndrome; OHSS, Ovarian hyperstimulation syndrome; Gn, Gonadotropin; E2, Estradiol; hCG, Human chorionic gonadotropin; CI, Confidence interval; OR, Odds ratio; Ref, Reference category.
Factors Associated with Adverse Pregnancy Outcomes
Among the 3261 women with a positive β-hCG result, participants were divided into two groups according to the occurrence of adverse pregnancy outcomes. Univariable analysis identified male age, female age, AMH, and male-factor infertility as factors significantly associated with adverse pregnancy outcomes. These variables were entered into a multivariable logistic regression model, with female age group forced into the model.
After multivariable adjustment, the 30–34-year group showed no significant difference in the risk of adverse pregnancy outcomes compared with the <30-year group (OR = 1.048, 95% CI 0.871–1.261, P = 0.621). Women aged 35–37 years had significantly higher odds (OR = 1.515, 95% CI 1.135–2.024, P = 0.005). An AMH level of 2–5 ng/mL remained a protective factor (OR = 0.782, 95% CI 0.631–0.969, P = 0.025).
Male age and male-factor infertility were retained for adjustment but were not independently significant in the multivariable model (Table 5). Among the individual components of the composite outcome, miscarriage was the main driver of adverse pregnancy outcomes; the miscarriage rate in Group C was markedly higher than that in Groups A and B (remaining statistically significant vs Group A after Bonferroni correction), whereas biochemical pregnancy and ectopic pregnancy rates did not differ significantly among the three groups.Table 5Factors Associated with Adverse Pregnancy OutcomesUnivariable OR (95%CI)P valueMultivariable OR (95%CI)P valueEntered into Multivariable Model?Infertility duration (years) <3RefNo ≥31.005 (0.862–1.172)0.950NoBMI (kg/m2) <18.5RefNo 18.5≤BMI<240.950 (0.746–1.211)0.680No 24≤BMI<281.134 (0.859–1.497)0.375No ≥281.127 (0.766–1.658)0.545NoMale age (years) <35RefYes 35–401.227 (1.033–1.457)0.020*1.088 (0.882–1.342)0.433Yes >401.153 (0.788–1.686)0.4640.974 (0.637–1.488)0.903YesFemale age (years) <30RefYes 30–341.080 (0.915–1.276)0.3621.048 (0.871–1.261)0.621Yes 35–371.553 (1.231–1.959)<0.001*1.515 (1.135–2.024)0.005*YesAMH (ng/ml) <2RefYes 2–50.747 (0.605–0.924)0.007*0.782 (0.631–0.969)0.025*Yes >50.863 (0.679–1.098)0.2310.936 (0.732–1.197)0.597YesInfertility type Primary infertilityRefNo Secondary infertility0.997 (0.855–1.162)0.967NoTransfer cycle First cycleRefNo Second cycle1.283 (0.880–1.870)0.196NoTransferred blastocyst grade AARefNo AB1.515 (0.745–3.079)0.251No BA1.458 (0.719–2.957)0.296No BB1.706 (0.847–3.436)0.135NoTESE NoRefNo Yes1.419 (0.928–2.168)0.106NoICSI NoRefNo Yes1.134 (0.926–1.389)0.225NoPCOS NoRefNo Yes1.008 (0.841–1.209)0.929NoEndometriosis NoRefNo Yes1.076 (0.776–1.492)0.606NoMale factor NoRefYes Yes1.226 (1.017–1.478)0.033*1.210 (0.996–1.469)0.054YesOHSS NoRefNo Yes0.994 (0.522–1.892)0.985NoNo. of oocytes retrieved ≤10RefNo 11–141.065 (0.873–1.300)0.535No ≥151.156 (0.947–1.410)0.154NoGn dosage (IU) ≤2000RefNo 2001–40001.035 (0.881–1.216)0.673No >40001.359 (0.911–2.030)0.133NoDuration of Gn treatment (day) ≤10RefNo 11–140.988 (0.834–1.171)0.891No ≥151.013 (0.753–1.361)0.933NoEndometrial thickness on hCG day (mm) <8RefNo 8–120.847 (0.717–1.001)0.051No >120.933 (0.665–1.308)0.687NoE2 level on hCG day (pg/mL) ≤2000RefNo 2001–40001.130 (0.957–1.335)0.149No >40001.296 (0.997–1.685)0.053NoNotes: Model fit for Table 5: Hosmer–Lemeshow χ2 = 3.988, P = 0.781. The multivariable model was adjusted for male age, female age, AMH, and male factor. *P < 0.05 was considered statistically significant.Abbrevation: BMI, Body mass index; AMH, Anti-Müllerian hormone; TESE, Testicular sperm extraction; ICSI, Intracytoplasmic sperm injection; PCOS, Polycystic ovary syndrome; OHSS, Ovarian hyperstimulation syndrome; Gn, Gonadotropin; E2, Estradiol; hCG, Human chorionic gonadotropin; CI, Confidence interval; OR, Odds ratio; Ref, Reference category.
Factors Associated with Obstetric and Neonatal Complications
Patients who achieved a live birth (n = 2355) were stratified into two groups based on the occurrence of obstetric and neonatal complications. Univariable analysis revealed that maternal BMI, male age, duration of Gn stimulation, endometrial thickness, and serum E2 level on the day of hCG trigger were significantly associated with these complications. Although female age did not show a significant association in the univariable analysis, it was forced into the multivariable model.
Following multivariable adjustment, a BMI of ≥28 kg/m2 remained an independent risk factor for obstetric and neonatal complications (OR = 2.414, 95% CI: 1.405–4.149, P = 0.001). Conversely, an endometrial thickness of >12 mm on the day of hCG trigger (OR = 0.361, 95% CI: 0.183–0.712, P = 0.003) and a Gn stimulation duration of 11–14 days (OR = 0.680, 95% CI: 0.530–0.872, P = 0.002) were identified as independent protective factors. Female age, male age, and serum E2 level on the day of hCG trigger were retained in the model for adjustment but did not reach statistical significance (Table 6).Table 6Factors Associated with Obstetric and Neonatal ComplicationsUnivariable OR (95%CI)P valueMultivariable OR (95%CI)P valueEntered into Multivariable Model?Infertility duration (years) <3RefNo ≥31.141 (0.906–1.437)0.261NoBMI (kg/m2) <18.5RefYes 18.5≤BMI<241.027 (0.702–1.503)0.8910.982 (0.668–1.443)0.925Yes 24≤BMI<281.441 (0.942–2.207)0.0921.336 (0.864–2.066)0.193Yes ≥282.589 (1.534–4.370)<0.001*2.414 (1.405–4.149)0.001*YesMale age (years) <35RefYes 35–401.329 (1.031–1.713)0.028*1.214(0.899–1.638)0.206Yes >401.370 (0.796–2.356)0.2561.313 (0.727–2.372)0.367YesFemale age (years) <30RefYes 30–341.098 (0.859–1.404)0.4541.061(0.811–1.389)0.664Yes 35–371.411 (0.990–2.010)0.0571.256 (0.818–1.929)0.298YesAMH (ng/ml) <2RefNo 2–51.148 (0.812–1.622)0.435No >51.368 (0.934–2.005)0.108NoInfertility type Primary infertilityRefNo Secondary infertility1.141 (0.907–1.435)0.259NoTransfer cycle First cycleRefNo Second cycle1.684 (0.999–2.841)0.051NoTransferred blastocyst grade AARefNo AB1.141 (0.434–2.997)0.789No BA1.156 (0.442–3.022)0.767No BB1.356 (0.525–3.505)0.530NoTESE NoRefNo Yes0.604 (0.259–1.413)0.245NoICSI NoRefNo Yes0.965 (0.705–1.321)0.825NoPCOS NoRefNo Yes1.179 (0.907–1.534)0.218NoEndometriosis NoRefNo Yes1.253 (0.787–1.996)0.341NoMale factor NoRefNo Yes0.957 (0.714–1.284)0.771NoOHSS NoRefNo Yes0.994 (0.382–2.586)0.990NoNo. of oocytes retrieved ≤10RefNo 11–140.881(0.660–1.176)0.389No ≥150.928 (0.695–1.240)0.613NoGn dosage (IU) ≤2000RefNo 2001–40001.034 (0.814–1.313)0.785No >40001.287 (0.701–2.363)0.415NoDuration of Gn treatment (day) ≤10RefYes 11–140.688 (0.538–0.879)0.003*0.680 (0.530–0.872)0.002*Yes ≥150.887 (0.581–1.354)0.5780.822 (0.476–1.147)0.178YesEndometrial thickness on hCG day (mm) <8RefYes 8–120.832 (0.651–1.062)0.1400.873 (0.680–1.122)0.289Yes >120.380 (0.194–0.745)0.005*0.361 (0.183–0.712)0.003*YesE2 level on hCG day (pg/mL) ≤2000RefYes 2001–40000.764 (0.600–0.971)0.028*0.837 (0.653–1.072)0.158Yes >40000.748 (0.493–1.134)0.1720.836 (0.547–1.280)0.410YesNotes: Model fit for Table 6: Hosmer–Lemeshow χ2 = 8.031, P = 0.430.
The multivariable model was adjusted for BMI, female age, male age, duration of Gn treatment, endometrial thickness, and E2 level on hCG day. *P < 0.05 was considered statistically significant.Abbrevations: BMI, Body mass index; AMH, Anti-Müllerian hormone; TESE, Testicular sperm extraction; ICSI, Intracytoplasmic sperm injection; PCOS, Polycystic ovary syndrome; OHSS, Ovarian hyperstimulation syndrome; Gn, Gonadotropin; E2, Estradiol; hCG, Human chorionic gonadotropin; CI, Confidence interval; OR, Odds ratio; Ref, Reference category.
Discussion
The present study shows that, among women undergoing fresh high-quality SBT, fertility outcomes declined with increasing maternal age; however, pregnancy outcomes in women aged 30–34 years were not statistically different from those in women younger than 30 years. In contrast, women aged 35–37 years did not achieve comparable outcomes. Because all participants received SBT, these findings reflect age-related differences within the SBT population rather than a comparison of SBT with other transfer strategies.
To our knowledge, this is the first study to use a female age cutoff of 37 years to evaluate the feasibility of fresh high-quality SBT in women aged 37 years or younger.
In 2019, a multicenter, open-label, randomized controlled trial conducted at 21 reproductive centers in China showed that although frozen-thawed SBT resulted in a higher singleton live birth rate than fresh SBT in ovulatory women with a good prognosis, the increased risk of preeclampsia after FET warrants further investigation.8 In 2020, another multicenter randomized controlled study found that a freeze-all strategy was not superior to fresh blastocyst transfer in women with regular menstrual cycles, offering no significant therapeutic advantage9 and underscoring the importance of avoiding unnecessary delays to pregnancy. Wong et al reported similar findings.10 The present study focused on fresh SBT cycles; by stratifying infertile women aged ≤37 years, we further examined the impact of age on pregnancy outcomes, thereby enabling a more individualized application of SBT strategies.
Previous studies using predictive reference models found that the cumulative live birth rate (CLBR) was highly similar among women aged 20–28 years, with differences beginning after age 29 and becoming more pronounced after age 35.11 Our findings suggest that, although ovarian stimulation and laboratory outcomes were less favorable in women aged 30–34 years than in those aged <30 years, pregnancy outcomes were comparable between the two groups. This indicates that, for women younger than 35 years, pregnancy outcomes after fresh high-quality SBT do not vary significantly with age. Unfortunately, women aged 35–37 years did not achieve outcomes as favorable as those aged 30–34 years following this transfer strategy; their ovarian stimulation, laboratory, and pregnancy outcomes were markedly inferior.
Multivariable logistic regression further confirmed that women aged 35–37 years had 1.539-fold higher odds of non-live birth, whereas women aged 30–34 years and younger were not at increased risk, supporting the feasibility of fresh high-quality SBT in women younger than 35 years. Although the proportion of top-quality blastocysts (AA and AB) decreased progressively with maternal age, blastocyst quality was entered as a covariate in the multivariable analysis, and the independent association between maternal age 35–37 years and non-live birth remained unchanged. These findings suggest that differences in blastocyst quality among the three age groups are unlikely to account for the age-related decline in pregnancy outcomes.
Previous evidence indicates that, in fresh embryo transfer cycles, male partner age was significantly associated with live birth rate when female partners were aged 35–39 years, but this association was not observed when female partners were aged <35 years.12 In light of this, this interaction may partially explain the poorer pregnancy outcomes observed in the 35–37-year group; alongside maternal reproductive aging, the more advanced paternal age in this group represents an additional contributing factor that should not be overlooked.
Adverse pregnancy outcomes are an important indicator for evaluating the safety and reliability of transfer strategies. Previous studies have shown that women aged ≥35 years have a higher risk of adverse obstetric outcomes, such as gestational diabetes mellitus (GDM), gestational hypertension, and preterm birth, than younger women.13 As an exploratory finding, we observed a higher incidence of GDM in the 35–37-year group (5.31%) than in the <30-year group (2.33%); however, this difference should be interpreted with caution given the exploratory nature of the analysis. Consistent with previous reports, the incidence of miscarriage was highest in the 35–37-year group (20.94%) and increased with maternal age.13–15 Multivariable logistic regression further showed that the 35–37-year group had 1.515-fold higher odds of adverse pregnancy outcomes compared with the <30-year group.
Among the individual components, miscarriage was the main contributor to this association, consistent with previous evidence that the risk of clinical pregnancy loss increases after 35 years of age. Nevertheless, a Finnish study offers an alternative perspective on the relationship between maternal age and birth outcomes by accounting for unobserved maternal characteristics shared among siblings.16 That study found that advanced maternal age was not associated with an increased probability of low birth weight or preterm birth when sibling-shared factors—such as parental social background, healthy behaviors during pregnancy, maternal height, genetic factors related to infertility, and health status—were held constant.
Most studies suggest that the likelihood of delivery by cesarean section increases with maternal age.13,17,18 Similarly, we found a significant association between age and cesarean delivery: the lowest rate was observed in the <30-year group (47.23%) and the highest in the 35–37-year group (60.45%). Slow labor progression, impaired uterine contractility, and uterine changes with aging may be common underlying reasons for the higher cesarean section rates among women of advanced maternal age. In addition, the level of the delivery hospital, the socioeconomic status of the pregnant woman,19 and the presence of metabolic abnormalities in the third trimester18 may also influence the mode of delivery.
It is also plausible that the increased rate of cesarean section in older mothers reflects shared decision-making between patients and obstetricians to minimize the risk of adverse pregnancy and delivery outcomes. However, we were unable to test this hypothesis in the present study because detailed information on the indications for cesarean section was not recorded during follow-up.
The higher birth defect rate in the <30-year group was an unexpected exploratory finding. In this group, 17 birth defects (1.50%) were recorded, including six cardiac malformations, five limb malformations, two external auditory canal malformations, and one case each of glucose-6-phosphate dehydrogenase deficiency (favism), solitary lung, omphalocele, and ectopic kidney. In the 30–34-year group, five birth defects were identified: one case each of digital deformity, congenital esophageal atresia, unilateral renal agenesis, glucose-6-phosphate dehydrogenase deficiency (favism), and biliary atresia.
No congenital anomaly was observed in Group C (35–37 years). Because the number of events was small and multiple outcomes were compared, the between-group difference did not remain statistically significant after Bonferroni correction and may represent a chance finding. A pure maternal age effect is unlikely to explain this pattern.
Published systematic reviews consistently report that the risk of congenital anomalies increases with advanced maternal age, whereas evidence for an increased risk in younger mothers is weak and limited to certain defect categories.20 The highest defect rate in our youngest group is therefore opposite to the expected age-related trend, arguing against age per se as the main driver. Lifestyle and environmental factors, such as periconceptional smoking, alcohol use, and inadequate folic acid intake, are associated with specific congenital anomalies;21 however, these variables were not collected retrospectively, so their potential contribution cannot be evaluated. Differences in infertility etiology and fertilization method may be more relevant.
Indeed, the <30-year group had the highest proportions of primary infertility (61.17%), testicular sperm extraction (4.38%), and ICSI use (18.93% vs 14.27% in the 30–34-year group). Sperm source, including TESE, was not significantly associated with any outcome and is therefore unlikely to explain the age-related differences observed. Studies on ART have shown that birth defect risk is influenced by the procedure itself, the underlying infertility causes,22 male factor,22 and ICSI use.23,24 Nevertheless, the small number of events and residual confounding preclude causal inference.
This observation should be regarded as hypothesis-generating and warrants confirmation in prospective studies.
This study has several limitations that should be considered when interpreting the findings. First, although we endeavored to ensure data integrity, the retrospective nature of historical medical records may introduce missing values or recording inaccuracies. Second, follow-up was restricted to post-transfer pregnancy and neonatal outcomes at delivery and did not include long-term physical growth or neurodevelopmental assessments during childhood.
Third, given the retrospective, single-center design of this study, large-scale, prospective, multicenter randomized controlled trials are warranted to validate the clinical feasibility and safety of our findings. Fourth, although we prespecified the primary outcome and classified all other outcomes into secondary or exploratory tiers, the possibility of type I errors due to multiple testing cannot be fully excluded, particularly for the exploratory findings. For instance, the exploratory finding of a higher birth defect rate in the youngest group was based on a small number of events and should therefore be interpreted with caution.
Finally, because all participants were treated with the EFLL protocol and underwent fresh transfer of a single high-quality Day-5 blastocyst, the findings may not be fully generalizable to patients receiving other ovarian stimulation protocols, cleavage-stage embryo transfer, or embryos of lower quality.
Conclusion
Among women undergoing fresh high-quality single blastocyst transfer, those aged 30–34 years achieved pregnancy outcomes comparable to those of women younger than 30 years, whereas women aged 35–37 years had significantly lower live birth rates and higher miscarriage rates. These findings suggest that fresh high-quality SBT is an effective strategy for women younger than 35 years, while more individualized transfer strategies may be required for women aged 35–37 years. Because all participants underwent SBT, the results reflect age-related differences within the SBT population rather than a comparison between transfer strategies.
Prospective multicenter randomized trials are warranted to confirm these findings.