Cumulative pregnancy and cumulative live birth rates and perinatal and obstetric outcomes in POSEIDON patients: insights from 19,371 IVF/ICSI cycles

This study presents an analysis of cumulative pregnancy rates, live birth rates, and perinatal and obstetric outcomes among 19,371 infertile patients categorized under the POSEIDON classification system. Our findings demonstrate that maternal age exerts a stronger influence than ovarian reserve status on ART success rates. Crucially, by applying the POSEIDON criteria to predict pregnancy and perinatal outcomes, clinicians can identify high-risk groups early and implement timely, targeted interventions to improve reproductive prognosis.

This study was an institutional review board-approved study based on retrospective data collected from patients with infertility who had their first IVF/ICSI cycle at the reproductive medicine center of Tongji hospital from January 2016 to December 2021. This study adhered strictly to the standards for reporting observational studies (STROBE) [ 17 ]. The study population comprised individuals who had undergone standard ovarian stimulation and IVF/ICSI. Only the first cycle per patient was analysed [ 18 ]. All the retrospective data were obtained from the center’s electronic medical record (EMR) system, which provides standardized and routinely audited documentation for all IVF/ICSI cycles. The exclusion criterion was failure to meet any one of the inclusion criteria: (1) Patients who underwent preimplantation genetic testing and those who had undergone IVF/ICSI for purposes other than infertility were excluded. (2) Patients who had polycystic ovary or polycystic ovary syndrome (PCOS), or chronic diseases, such as hypertension, cardiovascular disease, autoimmune disease, diabetes, abnormal hepatic and renal function, hypothalamic amenorrhea, pelvic tuberculosis, uterine malformation, intrauterine adhesions, or defibrillators. (3) In addition, patients who received mild or minimal stimulation protocols were excluded, including those treated with clomiphene citrate, letrozole, or other oral ovulation induction agents, as the POSEIDON criteria require standard ovarian stimulation with gonadotropin doses ≥ 150 IU [ 19 ]. (4) Patients who were lost to follow-up [ 18 ]. Patients with male factor infertility, including those with severe oligozoospermia or azoospermia requiring testicular sperm extraction (TESE), were not excluded from this study. These cases were included as part of real-world ART populations. An assessment of the ovarian reserve was performed 1–3 months prior to starting stimulation [ 18 ]. The antral follicle count was measured using 2-dimensional transvaginal ultrasonography during the early follicular phase as recommended in practical guidelines for standardized use of the AFC. AFC was defined as the number of follicles 2–9 mm in diameter in both ovaries [ 20 ]. All patients underwent either conventional in vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI) procedures. The fertilization method was determined according to standard clinical protocols. Conventional IVF was performed when the total motile sperm count (TMSC) was ≥ 10 million and no significant male factor was identified. Intracytoplasmic sperm injection (ICSI) was indicated in cases of severe oligoasthenozoospermia, obstructive azoospermia, previous fertilization failure, use of testicular sperm, or when the number of TMSCs was < 10 million. These criteria were applied to ensure optimal fertilization outcomes and minimize the risk of total fertilization failure [ 21 – 23 ]. The controlled ovarian hyperstimulation (COH) protocol for each patient was customized based on the ovarian reserve assessment and combined with either a GnRH agonist or antagonist regimen [ 24 ]. On the day of hCG administration, serum levels of estradiol (E2) and progesterone (P) were measured [ 11 ]. Oocyte retrieval was performed 34–36 h after hCG injection via transvaginal ultrasound guidance [ 25 , 26 ]. In IVF cycles, each oocyte was inseminated with 10,000 motile sperm after 4 h [ 27 , 28 ]. Mature metaphase II (MII) oocytes were selected under an inverted microscope, and successful fertilization was confirmed by observing 2 pronuclei (2PN) in zygotes [ 25 ]. In both the IVF cycle and the ICSI cycle, all the oocytes were routinely denuded to allow assessment of nuclear maturity. The presence of the first polar body was used to identify metaphase II (MII) oocytes, which were recorded prior to insemination. On Day 3 post-retrieval, the embryos were morphologically assessed on the basis of blastomere number, symmetry, and degree of fragmentation. Embryos with fewer than 4 blastomeres or > 50% fragmentation were considered to have low developmental potential, as per established grading systems (ASEBIR) [ 29 , 30 ]. Embryos with ≥ 4 blastomeres and < 50% fragmentation are eligible for extended culture. Cryopreservation was performed by vitrification at either the cleavage or blastocyst stage for embryos that met the laboratory criteria for freezing. The subjects were allocated into 5 groups according to the Poseidon criteria [ 13 , 18 ]. The criteria were as follows: 1. POSEIDON Group 1 (PG1): Age < 35 years, AFC ≥ 5, AMH ≥ 1.2 ng/ml, and a standard ovarian stimulation with < 10 oocytes retrieved. 2. POSEIDON Group 2 (PG 2): Age ≥ 35 years, AFC ≥ 5, AMH ≥ 1.2 ng/ml, and a standard ovarian stimulation with < 10 oocytes retrieved. 3. POSEIDON Group 3 (PG 3): Age < 35 years, AFC < 5, and AMH < 1.2 ng/ml. 4. POSEIDON Group 4 (PG 4): Age ≥ 35 years, AFC < 5, and AMH < 1.2 ng/ml. 5. Non-POSEIDON (non-PG): Patients with an AFC ≥ 5, AMH ≥ 1.2 ng/ml, and a standard ovarian stimulation with ≥ 10 oocytes retrieved. 1. POSEIDON Group 1 (PG1): Age < 35 years, AFC ≥ 5, AMH ≥ 1.2 ng/ml, and a standard ovarian stimulation with < 10 oocytes retrieved. 2. POSEIDON Group 2 (PG 2): Age ≥ 35 years, AFC ≥ 5, AMH ≥ 1.2 ng/ml, and a standard ovarian stimulation with < 10 oocytes retrieved. 3. POSEIDON Group 3 (PG 3): Age < 35 years, AFC < 5, and AMH < 1.2 ng/ml. 4. POSEIDON Group 4 (PG 4): Age ≥ 35 years, AFC < 5, and AMH < 1.2 ng/ml. 5. Non-POSEIDON (non-PG): Patients with an AFC ≥ 5, AMH ≥ 1.2 ng/ml, and a standard ovarian stimulation with ≥ 10 oocytes retrieved. Fertilization outcomes, including the number of fertilized oocytes, normal fertilization rate, available embryos, frozen embryo on day 3, extended to blastocyst culture and transferable embryos, were recorded for both the IVF and ICSI cycles and analysed collectively to represent the total embryo development performance across the cohort. Normal fertilization rate was defined as the proportion of oocytes exhibiting 2 pronuclei (2PN) approximately 16–18 h after insemination or ICSI [ 31 ] . Clinical pregnancy was defined as the presence of an intrauterine gestational sac observed via ultrasound. The live birth rate per first cycle was calculated by dividing the number of women in the first cycle who had a live birth by the total number of women who attempted that complete cycle. The cumulative live birth rate (CLBR) was defined as the proportion of women who achieved at least one live birth over the course of all embryo transfer cycles resulting from a single ovarian stimulation, divided by the total number of women who initiated IVF treatment. For the CLBR, all complete cycles were included until either no transferable embryos remained or the first live birth occurred, whichever occurred first [ 32 ]. Patients with no available embryos included those with no oocytes retrieved or total fertilization failure. These patients were included in the denominators for cumulative outcomes (CPR and CLBR), but were not incorporated into the denominators of fresh embryo transfer cycles outcome such as clinical pregnancy rate (PR) and live birth rate (LBR), which were calculated per fresh embryo transfer cycle. This approach reflected real-world clinical effectiveness, particularly for patients who have characteristics associated with a poor prognosis, for whom failure to obtain transferable embryos was itself a critical outcome. Statistical analyses were performed via SPSS 20 software. Continuous data were summarized as the means with standard deviations (SDs), whereas categorical data were expressed as frequencies and percentages. The strength of the associations was examined through univariate analysis, and odds ratios (ORs) with corresponding 95% confidence intervals (CIs) were calculated. To control for confounding factors, adjusted odds ratios (AORs) were obtained via multiple logistic regression. Continuous variables were first assessed for normality using the Shapiro–Wilk test. For comparisons between 2 independent groups, continuous variables with a normal distribution were analyzed using Student’s t test, whereas non-normally distributed variables were compared using the Mann–Whitney U test. For comparisons across more than 2 groups (non-PG, PG1–PG4), one-way analysis of variance (ANOVA) was applied for normally distributed variables, and the Kruskal–Wallis H test was used for non-normally distributed variables. Differences in categorical variables were assessed via chi-square tests or Fisher’s exact tests. All p values were calculated as 2-tailed values, with a significance threshold set at p < 0.05. Information regarding all maternal and neonatal outcomes was obtained via telephone interviews after delivery [ 25 ]. All information was collected with the participant's consent.

A total of 19,371 patients were included in this study, 7009 of whom met the POSEIDON criteria and were classified into one of the ‘low-prognosis’ groups according to their age and ovarian functions; these patients were significantly older, and/or had decreased AMH levels, AFCs and numbers of oocytes retrieved if they previously underwent standard ovarian stimulation in the POSEIDON groups (PGs) (Table  1 , all p < 0.001). According to the POSEIDON criteria, PG1 comprised 4003 individuals, PG2 included 1581 individuals, PG3 consisted of 638 individuals, PG4 involved 787 individuals, and non-PG contained 12,362 individuals (Fig.  1 ). Table 1 Obstetrical history with demographic data from the first IVF/ICSI cycle Non-POSEIDON (n = 12,362) POSEIDON 1 (n = 4003) POSEIDON 2 (n = 1581) POSEIDON 3 (n = 638) POSEDON 4 (n = 787) POSEDON (n = 7009) p value † Age (years) # 30.17 ± 3.80 29.88 ± 2.83 37.92 ± 2.59* 30.39 ± 2.55 40.09 ± 3.56* 32.88 ± 5.04*  < 0.001 BMI (kg/m 2 ) # 21.67 ± 2.87 21.78 ± 3.04 22.45 ± 2.87* 21.86 ± 3.10 22.49 ± 2.82* 22.02 ± 3.00*  < 0.001 Baseline FSH (mIU/mL) # 7.19 ± 1.69 7.94 ± 2.27* 8.19 ± 2.40* 11.36 ± 6.67* 12.08 ± 6.24* 8.78 ± 3.87*  < 0.001 Infertility, n (%)  < 0.001 Primary infertility 8576 (69.4) 2928 (73.1)* 748 (47.3)* 454 (71.2) 329 (41.8)* 4459 (63.6)* Secondary infertility 3786 (30.6) 1075 (26.9)* 833 (52.7)* 184 (28.8) 458 (58.2)* 2550 (36.4)* Duration of infertility (years) 3.20 ± 2.25 3.25 ± 2.18 4.14 ± 3.72* 3.27 ± 2.36 3.78 ± 4.01 3.52 ± 2.88*  < 0.001 AMH (ng/mL) # 5.51 ± 3.40 3.67 ± 2.67* 2.76 ± 1.66* 0.63 ± 0.31* 0.55 ± 0.31* 2.84 ± 2.48*  < 0.001 AFC& # 14.99 ± 5.54 11.34 ± 5.04* 9.06 ± 3.52* 2.97 ± 1.05* 2.88 ± 1.04* 9.12 ± 5.31*  < 0.001 INHB # 98.62 ± 57.30 86.99 ± 40.10* 87.55 ± 47.79* 68.43 ± 106.18* 59.38 ± 39.49* 83.24 ± 50.17*  < 0.001 Indication, n (%) Male factor infertility 4605 (37.3) 1340 (33.5)* 490 (31.0)* 178 (27.9)* 180 (22.9)* 2188 (31.2)*  < 0.001 Tubal 6724 (54.4) 2186 (54.6) 809 (51.2) 279 (43.7)* 327 (41.6)* 3601 (51.4)*  < 0.001 Endometriosis 1005 (8.1) 486 (12.1) 133 (8.4) 98 (15.4) 75 (9.5) 792 (11.3)*  < 0.001 Unexplained 964 (7.8) 282 (7.0) 57 (3.6)* 2 (0.3)* 0 (0)* 341 (4.9)*  < 0.001 Total gonadotropin dose (IU) # 2228.42 ± 767.41 2590.29 ± 870.15* 2899.94 ± 793.02* 2962.88 ± 1074.91* 2897.60 ± 1,030.68* 2728.56 ± 907.80*  < 0.001 Duration of stimulation (days) # 10.37 ± 1.69 10.13 ± 1.96* 9.88 ± 1.80* 9.56 ± 2.28* 9.31 ± 2.23* 9.93 ± 2.01*  < 0.001 Protocol, n (%)  < 0.001 GnRH agonist 9831 (79.5) 2482 (62.0)* 596 (37.7)* 30 (4.7)* 4 (0.5)* 3112 (44.4)* GnRH antagonist 2439 (19.8) 1391 (34.7)* 785 (49.7)* 370 (58.0)* 293 (37.2)* 2837 (40.5)* luteal phase stimulation 92 (0.7) 127 (3.2)* 200 (12.7)* 238 (37.3)* 490 (62.3)* 1055 (15.1)* Fertilization method, n (%)  < 0.001 IVF cycles 8024 (64.9) 2678 (66.9) 1115 (70.5)* 475 (74.5)* 571 (72.6)* 4839 (69.0)* Intracytoplasmic sperm injection (ICSI) cycles 3629 (29.4) 1060 (26.5)* 390 (24.7)* 154(24.2)* 178 (22.6)* 1717 (24.5)* E2 on hCG (pg/mL) # 3347.98 ± 1827.18 1899.26 ± 1018.68* 1762.08 ± 939.44* 1100.21 ± 707.19* 884.05 ± 595.38* 1681.20 ± 1003.71*  < 0.001 LH on hCG (IU/L) # 2.62 ± 2.27 3.28 ± 2.57* 3.92 ± 2.72* 4.84 ± 5.40* 5.78 ± 5.70* 4.14 ± 3.98*  < 0.001 Endometrial thickness (mm) # 12.03 ± 2.53 11.80 ± 2.85* 10.84 ± 2.77* 10.06 ± 2.59* 8.66 ± 2.73* 11.08 ± 2.98*  < 0.001 P on hCG day (ng/mL) # 1.00 ± 0.58 0.78 ± 0.75* 0.80 ± 1.00* 0.78 ± 1.11* 0.91 ± 2.04* 0.80 ± 1.03*  < 0.001 Oocytes retrieved 16.17 ± 5.33 6.68 ± 2.03* 5.95 ± 2.20* 4.03 ± 2.79* 3.12 ± 2.36* 6.45 ± 3.26*  < 0.001 Number of patients with no oocytes retrieved, n (%) 0 (0) 17 (0.42)* 3 (0.19)* 30 (4.70)* 36 (4.57)* 86 (1.22)*  < 0.001 Number of metaphase II oocytes # 13.88 ± 5.02 5.79 ± 2.04* 5.25 ± 2.11* 3.47 ± 2.48* 2.76 ± 2.11* 5.12 ± 2.37*  < 0.001 Number of patients with no metaphase II oocytes, n (%) 9 (0.07) 34 (0.85)* 7 (0.44)* 47 (7.3)* 54 (6.86)* 142 (2.02)*  < 0.001 Number of fertilized oocytes (2PN) # 9.70 ± 4.36 4.01 ± 2.00* 3.56 ± 1.98* 2.38 ± 2.09* 1.84 ± 1.76* 3.52 ± 2.12*  < 0.001 Normal fertilization rate (%) # 60.01 ± 18.46 60.09 ± 24.56 60.01 ± 26.24 58.32 ± 33.42 58.77 ± 35.79 59.77 ± 27.23* 0.279 Number of available embryos # 9.60 ± 4.45 3.97 ± 2.02* 3.56 ± 2.00* 2.36 ± 2.06* 1.85 ± 1.77* 3.49 ± 2.13*  < 0.001 Day 3 frozen embryo 0.22 ± 0.55 0.26 ± 0.64 0.44 ± 0.75* 0.56 ± 0.79* 0.72 ± 0.86* 0.38 ± 0.72*  < 0.001 Number of embryos extended to blastocyst culture 8.47 ± 4.72 2.63 ± 2.06* 2.21 ± 1.97* 1.28 ± 1.89* 0.85 ± 1.52* 2.21 ± 2.06*  < 0.001 Number of blastocyst a ,# 5.83 ± 4.16 1.66 ± 1.74* 1.25 ± 1.52* 0.78 ± 1.47* 0.44 ± 1.02* 1.35 ± 1.66*  < 0.001 Blastocyst formation rate (%) b ,# 65.56 ± 26.16 58.67 ± 35.76* 53.02 ± 36.77* 53.87 ± 40.64* 46.15 ± 41.27* 56.32 ± 36.82*  < 0.001 No transferred embryo, n (%) c 72 (0.6) 168 (4.2)* 84 (5.3)* 118 (18.5)* 171 (21.7)* 541 (7.7)*  < 0.001 Cumulative clinical pregnancy, n (%) d 10,459 (84.6) 2902 (72.5)* 865 (54.7)* 330 (51.7)* 242 (30.7)* 4339 (61.9)*  < 0.001 Cumulative live birth, n (%) e 10,013 (81.0) 2682 (67.0)* 714 (45.2)* 289 (45.3)* 180 (22.9)* 3865 (55.1)*  < 0.001 † p values represent comparisons between each POSEIDON group and the Non-POSEIDON group. Continuous variables with a normal distribution were analyzed using Student’s t test, whereas non-normally distributed variables were compared using the Mann–Whitney U test. Differences in categorical variables were assessed via chi-square tests or Fisher’s exact tests *indicate significant differences between each POSEIDON group (1–4) and the Non-POSEIDON group (p < 0.05), calculated using the Bonferroni-corrected multiple comparison test. One-way analysis of variance (ANOVA) was applied for normally distributed variables, and the Kruskal–Wallis H test was used for non-normally distributed variables # Median (interquartile range) for continuous variables a Blastocyst number refers to the number of embryos that were embryotransferd on Day5 or successfully developed to the blastocyst stage (Day 5–7) following extended culture of Day 3 embryos b Blastocyst rate refers to the proportion of embryos that developed to the blastocyst stage among embryos that were actually cultured for Day 5–7, rather than all fertilized oocytes c No transferred embryo refers to cycles with no mature oocyte or developed embryos obtained d Cumulative clinical pregnancy rate (CPR) was defined as the proportion of women who achieved at least one clinical pregnancy from their first complete IVF/ICSI treatment, divided by the total number of women with that treatment, irrespective of whether oocytes were retrieved or an embryo transfer was performed e Cumulative live birth rate (CLBR) was defined as the proportion of women who achieved their first treatment-dependent live birth from the first oocyte retrieval cycle, divided by the total number of women who received their first IVF/ICSI. All fresh and subsequent frozen-thawed embryo transfers originating from the first oocyte retrieval were included until either a live birth occurred or no transferable embryos remained, whichever came first Fig. 1 Flowchart of the participants. In the figure, n represents the number of subjects, with values in parentheses corresponding to fresh embryo transfer cycles and frozen embryo transfer cycles Obstetrical history with demographic data from the first IVF/ICSI cycle † p values represent comparisons between each POSEIDON group and the Non-POSEIDON group. Continuous variables with a normal distribution were analyzed using Student’s t test, whereas non-normally distributed variables were compared using the Mann–Whitney U test. Differences in categorical variables were assessed via chi-square tests or Fisher’s exact tests *indicate significant differences between each POSEIDON group (1–4) and the Non-POSEIDON group (p < 0.05), calculated using the Bonferroni-corrected multiple comparison test. One-way analysis of variance (ANOVA) was applied for normally distributed variables, and the Kruskal–Wallis H test was used for non-normally distributed variables # Median (interquartile range) for continuous variables a Blastocyst number refers to the number of embryos that were embryotransferd on Day5 or successfully developed to the blastocyst stage (Day 5–7) following extended culture of Day 3 embryos b Blastocyst rate refers to the proportion of embryos that developed to the blastocyst stage among embryos that were actually cultured for Day 5–7, rather than all fertilized oocytes c No transferred embryo refers to cycles with no mature oocyte or developed embryos obtained d Cumulative clinical pregnancy rate (CPR) was defined as the proportion of women who achieved at least one clinical pregnancy from their first complete IVF/ICSI treatment, divided by the total number of women with that treatment, irrespective of whether oocytes were retrieved or an embryo transfer was performed e Cumulative live birth rate (CLBR) was defined as the proportion of women who achieved their first treatment-dependent live birth from the first oocyte retrieval cycle, divided by the total number of women who received their first IVF/ICSI. All fresh and subsequent frozen-thawed embryo transfers originating from the first oocyte retrieval were included until either a live birth occurred or no transferable embryos remained, whichever came first Flowchart of the participants. In the figure, n represents the number of subjects, with values in parentheses corresponding to fresh embryo transfer cycles and frozen embryo transfer cycles Statistical analysis revealed that patients in PGs presented increased BMIs and baseline FSH levels. Infertility and tubal factors were predominant in the non-PG patients who underwent ART, whereas the prevalence of endometriosis was significantly greater in the patients in the PGs (Table  1 , p < 0.001). After stimulation, the number of oocytes retrieved, the number of metaphase II (MII) oocytes, the number of available embryos, the number of 2PN embryos, and the number of blastocysts gradually decreased from the non-PG group to the PG 1–4 subgroups, which corresponded with the observed decline in ovarian function (Table  1 , all p < 0.001). However, the normal fertilization rates did not differ between PG patients and their counterparts (p = 0.210). When the number of embryos transferred was assessed, PG4 patients had the highest rates of embryos that were not transferred, indicating potential treatment failure (p < 0.001). Most importantly, the CPR gradually decreased from non-PG patients to PG 1–4 patients, with rates of 84.6%, 72.5%, 54.7%, 51.7%, and 30.7%, respectively. The CLBR was highest in non-PG patients, followed by PG1 patients, PG3 patients, PG2 patients, and PG4 patients, with rates of 81.0%, 67.0%, 45.3%, 45.2%, and 22.9%, respectively (p < 0.001). Given the significant impact of age on the CLBR, we accordingly restricted our analysis to different age populations within specific age cohorts. The trends in differences between the non-PG (< 35 years/ ≥ 35 years) and PG 1/3 subgroups were consistent with those observed when the non-PG was compared with the PG group (Table S1 ). The CPR and CLBR showed a clear trend, with non-PG patients having significantly better outcomes than PG patients. We then questioned whether patients across different PGs exhibited variations in the CPR and the CLBR. We first conducted pairwise univariable logistic regression to compare rates between PG 1/2/3/4 patients and non-PG patients (Table  2 ). Patients categorized within the PGs had significantly lower CPRs and CLBRs, regardless of grade (Table  2 , all p < 0.001). Specifically, the odds ratios (ORs) of PG1 vs. non-PG were 0.480 (95% confidence interval (CI) 0.441–0.522) and 0.476 (95% CI 0.440–0.516) for the CPR and CLBR, respectively. These numbers continuously decreased in PG4 (OR = 0.081, 95% CI 0.069–0.095 and OR = 0.070, 95% CI 0.059–0.083 for the CPR and CLBR, respectively). We performed multivariable logistic regression analyses adjusting for clinically significant covariates (Table  2 ). The above trends still exist. Specifically, compared with non-PG patients, PG4 patients had AORs of 0.259 (95% CI 0.207–0.324) and 0.259 (95% CI 0.206–0.326) for CPR and the CLBR, respectively. Interestingly, the adjusted odds ratios (AORs) for PG3 were 0.404 (95% CIs 0.328–0.497) and 0.404 (95% CI 0.330–0.495), which were slightly higher than those for PG 2, which were 0.364 (95% CI 0.315–0.420) and 0.357 (95% CI 0.311–0.409) for the CPR and CLBR, respectively. Table 2 Odds ratios and adjusted odds ratios of pregnancy outcomes for the first IVF/ICSI cycle OR (95% CI) p value AOR (95% CI) p value Cumulative pregnancy rate POSEIDON 1 0.480 (0.441–0.522)  < 0.001 0.501 (0.455–0.551)  < 0.001 POSEIDON 2 0.220 (0.197–0.245)  < 0.001 0.364 (0.315–0.420)  < 0.001 POSEIDON 3 0.195 (0.166–0.229)  < 0.001 0.404 (0.328–0.497)  < 0.001 POSEIDON 4 0.081 (0.069–0.095)  < 0.001 0.259 (0.207–0.324)  < 0.001 Cumulative live birth rate POSEIDON 1 0.476 (0.440–0.516)  < 0.001 0.506 (0.463–0.553)  < 0.001 POSEIDON 2 0.193 (0.173–0.215)  < 0.001 0.357 (0.311–0.409)  < 0.001 POSEIDON 3 0.194 (0.165–0.228)  < 0.001 0.404 (0.330–0.495)  < 0.001 POSEIDON 4 0.070 (0.059–0.083)  < 0.001 0.259 (0.206–0.326)  < 0.001 1. Odds ratios (ORs) and 95% confidence intervals (CIs) are based on the univariate analysis. Adjusted ORs (AORs), 95% CIs, and adjusted P values are based on the multiple logistic regression model 2. AORs and P values for age, BMI, type of infertility, infertility duration, tubal factors, endometriosis, male infertility factors, unexplained factors, gonadotropin dose, duration of stimulation, type of regimen, mode of fertilization (IVF or ICSI), type of embryo transferred (cleavage embryo or blastocyst), number of embryos transferred, AMH, FSH and AFC 3. The reference group is non-PG patients Odds ratios and adjusted odds ratios of pregnancy outcomes for the first IVF/ICSI cycle 1. Odds ratios (ORs) and 95% confidence intervals (CIs) are based on the univariate analysis. Adjusted ORs (AORs), 95% CIs, and adjusted P values are based on the multiple logistic regression model 2. AORs and P values for age, BMI, type of infertility, infertility duration, tubal factors, endometriosis, male infertility factors, unexplained factors, gonadotropin dose, duration of stimulation, type of regimen, mode of fertilization (IVF or ICSI), type of embryo transferred (cleavage embryo or blastocyst), number of embryos transferred, AMH, FSH and AFC 3. The reference group is non-PG patients We subsequently screened fresh embryo transfer cyclses and monitored oocyte and embryo quality, as well as pregnancy outcomes. Patient demographics are shown in Table  3 . Similar to the entire population, patients classified as PG patients were older; had increased BMI, baseline FSH, and secondary infertility rates; and had decreased AMH and AFC levels (Table  3 , all p < 0.005). The usage of the GnRH antagonist was more pronounced in the PGs (p < 0.001). The number of oocytes retrieved and the number of MII oocytes also decreased from the non-PG to the PG 1–4 subgroups (Table  3 , all p < 0.001). The normal fertilization rate in non-PG patients remained at approximately 60%, increasing to 68.39 and 67.96% in POSEIDON 3/4 patients, respectively. Despite the higher rate of double embryo transfer in PGs, both the PR and LBR, including singleton and twin births, were lower than those in non-PG patients. This decline was particularly evident in older age groups, where younger subgroups (PG1, PG3) demonstrated superior PR and LBR outcomes compared to older subgroups (PG2, PG4). The LBRs for patients in PG1 and PG3 were 48.5 and 42.8%, respectively, whereas those in PG2 and PG4 were 32.2 and 23.7%, respectively (p = 0.001). For non-PG patients, the LBR was 52.2%. Additionally, the PR of young patients in the PG1 subgroup was 55.9%, which was slightly lower than that of normal individuals (59.2%), whereas PG3 could still achieve a PR, reaching 47.3% at one point, which was higher than that of PG2 (40.3%) and approximately 1.63 times greater than that of PG4. Table 3 Comparison of IVF/ICSI-technique parameters between POSEIDON and non-POSEIDON in fresh embryo transfer cycles Non-POSEIDON (n = 8356) POSEIDON 1 (n = 3137) POSEIDON 2 (n = 1011) POSEIDON 3 (n = 243) POSEDON 4 (n = 152) POSEIDON (n = 4543) p value † Age (years) # 30.15 ± 3.71 29.79 ± 2.85 37.33 ± 2.17* 30.46 ± 2.55 38.26 ± 2.98* 31.79 ± 4.29*  < 0.001 BMI (kg/m 2 ) # 21.79 ± 2.89 21.81 ± 3.03 22.40 ± 2.82* 21.81 ± 3.13 22.09 ± 2.98 21.95 ± 2.99* 0.001 Baseline FSH (mIU/mL) # 7.24 ± 1.70 7.87 ± 2.10* 8.12 ± 2.22* 9.91 ± 4.47* 11.26 ± 5.30* 8.15 ± 2.59*  < 0.001 Infertility, n (%) 0.001 Primary infertility 5758 (68.9) 2293 (73.1)* 500 (49.5)* 166 (68.3) 72 (47.4)* 3031 (66.7) * Secondary infertility 2598 (31.1) 844 (26.9)* 511 (50.5)* 77 (31.7) 80 (52.6)* 1512 (33.3) * Duration of infertility (years) # 3.21 ± 2.24 3.22 ± 2.16 4.14 ± 3.60* 3.08 ± 2.21 3.09 ± 3.35* 3.41 ± 2.62* 0.001 AMH (ng/mL) # 5.26 ± 3.26 3.80 ± 2.75* 2.94 ± 1.74* 0.71 ± 0.28* 0.65 ± 0.30* 3.34 ± 2.59*  < 0.001 AFC # 14.65 ± 5.31 11.70 ± 5.12* 9.43 ± 3.62* 3.25 ± 0.94* 3.09 ± 0.97* 10.45 ± 5.19*  < 0.001 Indication, n (%) INHB # 97.31 ± 53.70 87.72 ± 41.27* 87.67 ± 50.08* 70.13 ± 48.09* 66.67 ± 34.52* 86.39 ± 44.00*  < 0.001 Male factor infertility 3115 (37.3) 1071 (34.1)* 306 (30.3)* 79 (32.5) 38 (25.0)* 1494 (32.9)* 0.001 Tubal 4555 (54.5) 1705 (54.4) 533 (52.7) 106 (43.6)* 72 (47.4) 2416 (53.2)* 0.005 Endometriosis 650 (7.8) 355 (11.3)* 72 (7.1) 37 (15.2)* 10 (6.6) 474 (10.4)* 0.001 Unexplained 691 (8.3) 241 (7.7) 46 (4.5)* 0 (0)* 0 (0)* 287 (6.3)* 0.001 Total gonadotropin dose (IU) # 2256.12 ± 758.93 2557.50 ± 855.98* 2892.60 ± 755.20* 2912.65 ± 851.76* 2903.29 ± 780.13* 2662.64 ± 850.35*  < 0.001 Duration of stimulation # 10.36 ± 1.69 10.16 ± 1.94* 10.03 ± 1.74* 9.45 ± 1.90* 9.27 ± 1.73* 10.06 ± 1.90*  < 0.001 Protocol, n (%)  < 0.001 GnRH agonist 6870 (82.2) 2085 (66.5)* 479 (47.4)* 21 (8.6)* 3 (2.0)* 2588 (57.0) * GnRH antagonist 1486 (17.8) 1052 (33.5)* 532 (52.6)* 222 (91.4)* 149 (98.0)* 1955 (43.0) * Fertilization method 0.001 IVF cycles 5454 (65.3) 2121 (67.6) 712 (70.4)* 189 (77.8)* 113 (74.3) 3135 (69.0)* Intracytoplasmic sperm injection (ICSI) cycles 2902 (34.7) 1016 (32.4) 299 (29.6)* 54 (22.2)* 39 (25.7) 1168 (25.7)* E2 on hCG (pg/mL) # 2963.73 ± 1348.96 1880.84 ± 941.40* 1831.16 ± 918.74* 1337.14 ± 650.34* 1242.72 ± 658.39* 1819.18 ± 928.81*  < 0.001 LH on hCG (IU/L) # 2.69 ± 2.23 3.23 ± 2.48* 3.77 ± 2.47* 4.28 ± 4.14* 5.31 ± 4.55* 3.63 ± 2.93*  < 0.001 Endometrial thickness (mm) # 12.06 ± 2.49 11.87 ± 2.39* 11.30 ± 2.49* 11.20 ± 2.29* 10.46 ± 2.31* 11.66 ± 2.43*  < 0.001 P on hCG day (ng/mL) 0.86 ± 0.36 0.71 ± 0.35* 0.70 ± 0.31* 0.67 ± 0.38* 0.66 ± 0.35* 0.70 ± 0.34*  < 0.001 Oocytes retrieved # 14.52 ± 3.46 6.87 ± 1.87* 6.27 ± 2.03* 5.06 ± 2.52* 4.52 ± 2.47* 6.56 ± 2.05*  < 0.001 Number of metaphase II oocytes # 12.50 ± 3.49 6.02 ± 1.87* 5.59 ± 1.95* 4.52 ± 2.31* 4.08 ± 2.26* 5.78 ± 1.99*  < 0.001 Number of fertilized oocytes (2PN) # 8.78 ± 3.34 4.26 ± 1.83* 3.93 ± 1.78* 3.35 ± 1.93* 2.94 ± 1.91* 4.09 ± 1.85*  < 0.001 Normal fertilization rate (%) # 60.74 ± 18.15 62.77 ± 21.68* 63.79 ± 21.98* 68.39 ± 23.42* 67.96 ± 24.92* 63.43 ± 22.00* 0.001 Number of available embryos # 8.75 ± 3.50 4.25 ± 1.84* 3.96 ± 1.83* 3.33 ± 1.98* 2.96 ± 1.83* 4.09 ± 1.87*  < 0.001 Number of available blastocyst # 5.00 ± 3.28 1.76 ± 1.71* 1.41 ± 1.54* 1.07 ± 1.62* 0.77 ± 1.34* 1.61 ± 1.67*  < 0.001 Blastocyst formation rate (%) # 65.43 ± 26.45 60.94 ± 34.62* 56.68 ± 35.46* 55.57 ± 38.73 55.36 ± 40.35 59.69 ± 35.14* 0.001 Type of embryos transferred, n (%) 0.001 Cleavage, n (%) 7153 (85.6)* 3042 (97.0)* 969 (95.8)* 235 (96.7)* 150 (98.7)* 4396 (96.8)* Blastocyst, n (%) 1201 (14.4)* 95 (3.0)* 42 (4.2)* 8 (3.3)* 2 (1.3)* 147 (3.2)* Number of embryos transferred 0.001 1, n (%) 5552 (66.4)* 1972 (62.9)* 600 (59.3)* 158 (65.0) 90 (59.2) 2820 (62.1)* 2, n (%) 2804 (33.6)* 1165 (37.1)* 411 (40.7)* 85 (35.0) 62 (40.8) 1723 (37.9)* Biochemical pregnancy, n (%) 425 (5.1) 178 (5.7) 54 (5.3) 15 (6.2) 10 (6.6) 257 (5.7)* 0.644 Clinical pregnancy, n (%) 4948 (59.2) 1755 (55.9)* 407 (40.3)* 115 (47.3)* 44 (28.9)* 2321 (51.1)* 0.001 Pregnancy loss rate, n (%) 586 (7.0) 234 (7.5) 81 (8.0) 11 (4.5) 8 (5.3) 334 (7.35)* 0.001 Live birth, n (%) 4362 (52.2) 1521 (48.5)* 326 (32.2)* 104 (42.8)* 36 (23.7)* 1987 (43.7)* 0.001 Singleton live birth, n (%) 3678 (44.0) 1282 (40.9)* 286 (28.3)* 86 (35.4) 34 (22.4)* 1688 (37.2)* 0.001 Twins live birth, n (%) 684 (8.2) 238 (7.6) 40 (4.0)* 18 (7.4) 2 (1.3)* 298 (6.6)* 0.001 † p values represent comparisons between each POSEIDON group and the Non-POSEIDON group. Continuous variables with a normal distribution were analyzed using Student’s t test, whereas non-normally distributed variables were compared using the Mann–Whitney U test. Differences in categorical variables were assessed via chi-square tests or Fisher’s exact tests *indicate significant differences between each POSEIDON group (1–4) and the Non-POSEIDON group (p < 0.05), calculated using the Bonferroni-corrected multiple comparison test. One-way analysis of variance (ANOVA) was applied for normally distributed variables, and the Kruskal–Wallis H test was used for non-normally distributed variables # Median (interquartile range) for continuous variables Comparison of IVF/ICSI-technique parameters between POSEIDON and non-POSEIDON in fresh embryo transfer cycles † p values represent comparisons between each POSEIDON group and the Non-POSEIDON group. Continuous variables with a normal distribution were analyzed using Student’s t test, whereas non-normally distributed variables were compared using the Mann–Whitney U test. Differences in categorical variables were assessed via chi-square tests or Fisher’s exact tests *indicate significant differences between each POSEIDON group (1–4) and the Non-POSEIDON group (p < 0.05), calculated using the Bonferroni-corrected multiple comparison test. One-way analysis of variance (ANOVA) was applied for normally distributed variables, and the Kruskal–Wallis H test was used for non-normally distributed variables # Median (interquartile range) for continuous variables Using logistic regression, we found that all PGs had decreased PRs and LBRs (Table  4 , all p < 0.005). After adjusting for confounding factors (Table  4 ), we noticed that the PR and LBR did not differ between PG1/3 patients and non-PG patients. However, PG2 patients had a decreased PR (OR = 0.667, 95% CI 0.567–0.783, p < 0.001) and LBR (OR = 0.670, 95% CI 0.568–0.792, p < 0.001). These effects were more pronounced in PG4 patients (OR = 0.513, 95% CI 0.348–0.755, p < 0.001 for pregnancy; OR = 0.590, 95% CI 0.394–0.885, p = 0.011 for live birth). These results suggested that fresh embryo transfer might be more effective in patients under 35 years of age than in those over 35 years of age in PGs. Table 4 Odds ratios and adjusted odds ratios of pregnancy outcomes for the fresh embryo transfer cycle OR (95% CI) p value AOR (95% CI) p value Pregnancy rate PG1 0.875 (0.805–0.950) 0.002 0.943 (0.860–1.034) 0.211 PG2 0.464 (0.406–0.530)  < 0.001 0.667 (0.567–0.783)  < 0.001 PG3 0.619 (0.479–0.799)  < 0.001 0.956 (0.722–1.266) 0.752 PG4 0.281 (0.197–0.399)  < 0.001 0.513 (0.348–0.755)  < 0.001 Live birth rate PG1 0.862 (0.794–0.936)  < 0.001 0.928 (0.847–1.016) 0.107 PG2 0.436 (0.379–0.501)  < 0.001 0.670 (0.568–0.792)  < 0.001 PG3 0.685 (0.529–0.886) 0.004 1.064 (0.802–1.412) 0.666 PG4 0.284 (0.195–0.414)  < 0.001 0.590 (0.394–0.885) 0.011 1. Odds ratios (ORs) and 95% confidence intervals (CIs) are based on the univariate analysis. Adjusted ORs (AORs), 95% CIs, and adjusted P values are based on the multiple logistic regression model 2. AORs and P values for age, BMI, type of infertility, infertility duration, tubal factors, endometriosis, male infertility factors, unexplained factors, gonadotropin dose, duration of stimulation, type of regimen, mode of fertilization (IVF or ICSI), type of embryo transferred (cleavage embryo or blastocyst), number of embryos transferred, AMH, FSH, AFC, P, and intimal thickness 3. The reference group is the non-PG Odds ratios and adjusted odds ratios of pregnancy outcomes for the fresh embryo transfer cycle 1. Odds ratios (ORs) and 95% confidence intervals (CIs) are based on the univariate analysis. Adjusted ORs (AORs), 95% CIs, and adjusted P values are based on the multiple logistic regression model 2. AORs and P values for age, BMI, type of infertility, infertility duration, tubal factors, endometriosis, male infertility factors, unexplained factors, gonadotropin dose, duration of stimulation, type of regimen, mode of fertilization (IVF or ICSI), type of embryo transferred (cleavage embryo or blastocyst), number of embryos transferred, AMH, FSH, AFC, P, and intimal thickness 3. The reference group is the non-PG Therefore, we next sought to identify differences between non-PG patients under 35 years of age and those in PG 1/3 (Table S3). The infertility rates did not differ between the groups (p = 0.162), nor did BMI (p = 0.180). Other indicators were consistent with the differences between the non-PG and PG subgroups when fresh embryos were transferred. The differences in demographics between the non-PG (≥ 35 years) and PG 2/4 subgroups were similar to those between the non-PG and PG subgroups when fresh embryos were transferred, except for fertilization methods (p = 0.588). Singleton live births were evaluated for obstetric complications and adverse birth outcomes, which were reported in Table  5 . There were 8795 singleton live births in the non-POSEIDON group, whereas 3423 singleton live births occurred in PGs. Similar to the entire population, patients categorized within the PGs were older, had a higher BMI, a decreased prevalence of male factor infertility, and had a higher incidence of endometriosis (11.2% vs. 8.1%, p < 0.001). Caesarean section rates were also slightly higher in the PGs compared to non-POSEIDON patients (80.3% vs. 78.0%, p < 0.001). Compared with non-PG patients, a greater proportion of patients in the PGs preferred IVF over ICSI (p < 0.001). The usage of GnRH antagonists was more pronounced, and higher dosages were used in the PGs (p < 0.001). For singleton live births, gestational age (p < 0.001) and birth weight (p = 0.003) decreased gradually from non-PG to PG 1–4. The preterm birth rate progressively increased from the non-PG to the PG 1–4 (p = 0.008), with the PG4 subgroup showing a particularly high rate of 13.0%. Overall, the POSEIDON group had a significantly higher preterm birth rate than the non-POSEIDON group (8.5% vs. 7.0%, p < 0.05). The prevalence of gestational diabetes was higher in the PG1 (5.4%) and PG2 (5.9%) subgroups than in the non-PG (3.9%) (p < 0.05), with an overall rate of 5.2% in the PGs (vs. 3.9% in non-PGs, p < 0.05). Additionally, the rate of low birth weight was likewise increased in the PG group (4.8% vs. 3.8%, p < 0.05). With respect to obstetric complications, the incidence of gestational hypertension, placenta previa, fetal malformation, neonatal death, premature rupture of membranes, neonatal pneumonia, and postpartum hemorrhage did not differ significantly between the non-PG and PG 1–4 (p > 0.05). Table 5 Perinatal and obstetric outcomes in comparison of singleton pregnancies Non-POSEIDON (N = 8795) POSEIDON 1 (N = 2393) POSEIDON 2 (N = 624) POSEIDON 3 (N = 244) POSEDON 4 (N = 162) POSEDON (N = 3423) p value † Maternal age (years) +  30.07 ± 3.55 29.79 ± 2.80 37.12 ± 2.12* 30.40 ± 2.56 37.88 ± 2.31* 31.55 ± 4.09  < 0.001 BMI (kg/m 2 ) # 21.65 ± 2.86 21.83 ± 3.07 22.41 ± 2.76* 21.81 ± 3.06 22.09 ± 2.54 21.95 ± 3.00  < 0.001 Type of transfer embryos, n (%)  < 0.001 Fresh 3678 (41.8) 1282 (53.6)* 286 (45.8) 86 (35.2) 34 (21.0)* 1688 (49.3)* Frozen 5117 (58.2) 1111 (46.4)* 338 (54.2) 158 (64.8) 128 (79.0)* 1735 (50.7)* Fertilization method, n (%)  < 0.001 IVF cycles 5713 (65.0) 1638 (68.4)* 464 (74.4)* 194 (79.5)* 123 (75.9)* 2419 (70.67) Intracytoplasmic sperm injection (ICSI) cycles 3082 (35.0) 755 (31.6)* 160 (25.6)* 50 (20.5)* 39 (24.1)* 1004 (29.33) Protocol, n (%)  < 0.001 GnRH agonist 6982 (79.4) 1491 (62.3)* 260 (41.7)* 12 (4.9)* 3 (1.9)* 1766 (51.6) GnRH antagonist 1765 (20.1) 852 (35.6)* 319 (51.1)* 161 (66.0)* 80 (49.4)* 1412 (41.3) luteal phase stimulation 48 (0.5) 50 (2.1)* 45 (7.2)* 71 (29.1)* 79 (48.8)* 245 (7.2) Total gonadotropin dose (IU) # 2200.88 ± 757.67 2591.33 ± 884.77* 2872.81 ± 827.83* 2916.02 ± 868.39* 2929.03 ± 772.54* 2681.79 ± 879.08  < 0.001 Duration of stimulation # 10.36 ± 1.67 10.19 ± 1.93* 9.97 ± 1.80* 9.42 ± 1.89* 9.60 ± 1.94* 10.06 ± 1.92  < 0.001 Indication, n (%) Male factor infertility 3291 (37.4) 774 (32.3)* 181 (29.0)* 72 (29.5) 39 (24.1)* 1066 (31.1)*  < 0.001 Tubal 4750 (54.0) 1338 (55.9) 334 (53.3) 112 (45.9) 75 (46.3) 1859 (54.3) 0.008 Endometriosis 709 (8.1) 276 (11.5)* 49 (7.9) 42 (17.2)* 17 (10.5) 384 (11.2)*  < 0.001 Unexplained 707 (8.0) 173 (7.2) 26 (4.2)* 1 (0.4)* 0 (0)* 200 (5.8)  < 0.001 Gestational age (days) # 271.45 ± 29.61 270.90 ± 11.17 269.36 ± 10.45* 270.26 ± 12.43 268.72 ± 11.19* 270.47 ± 11.16*  < 0.001 Birth weight (g) # 3306.96 ± 472.10 3277.50 ± 487.37 3281.47 ± 498.65 3220.74 ± 529.66 3258.77 ± 446.77 3273.29 ± 490.76 0.003 Delivery mode, n (%)  < 0.001 Cesarean delivery 6862 (78.0) 1877 (78.4) 538 (86.2)* 184 (75.4) 148 (91.4)* 2747 (80.3)* Natural labor 1933 (22.0) 516 (21.6) 86 (13.8)* 60 (24.6) 14 (8.6)* 676 (19.7)* Gender, n (%) 0.156 Male 4665 (53.0) 1206 (50.4) 340 (54.5) 132 (54.1) 88 (54.3) 1766 (51.6) Female 4130 (47.0) 1187 (49.6) 284 (45.5) 112 (45.9) 74 (45.7) 1657 (48.4) Abnormal perinatal outcomes, n (%) Preterm delivery, < 37 weeks 619 (7.0) 194 (8.1) 53 (8.5) 24 (9.8) 21 (13.0)* 292 (8.5)* 0.008 Gestational hypertension 199 (2.3) 63 (2.6) 23 (3.7) 5 (2.0) 5 (3.1) 91 (2.7) 0.192 Gestational diabetes mellitus 343 (3.9) 130 (5.4)* 37 (5.9) 8 (3.3) 4 (2.5) 177 (5.2)* 0.027 Placenta previa 214 (2.4) 69 (2.9) 18 (2.9) 9 (3.7) 4 (2.5) 100 (2.9) 0.546 Fetal malformation 122 (1.4) 36 (1.5) 10 (1.6) 3 (1.2) 2 (1.2) 51 (1.5) 0.910 Neonatal death 13 (0.1) 4 (0.2) 1 (0.2) 0 (0) 0 (0) 5 (0.1) 0.955 Premature rupture of membranes 104 (1.2) 40 (1.7) 5 (0.8) 3 (1.2) 2 (1.2) 50 (1.5) 0.315 Neonatal pneumonia 55 (0.6) 12 (0.5) 2 (0.3) 0 (0) 2 (1.2) 16 (0.5) 0.417 Postpartum hemorrhage 34 (0.4) 12 (0.5) 2 (0.3) 2 (0.8) 1 (0.6) 17 (0.5) 0.763 Low birth weight,  4000 g 439 (5.0) 107 (4.5) 32 (5.1) 12 (4.9) 9 (5.6) 160 (4.6) 0.974 † p values represent comparisons between each POSEIDON group and the Non-POSEIDON group. Continuous variables with a normal distribution were analyzed using Student’s t test, whereas non-normally distributed variables were compared using the Mann–Whitney U test. Differences in categorical variables were assessed via chi-square tests or Fisher’s exact tests *indicate significant differences between each POSEIDON group (1–4) and the Non-POSEIDON group (p < 0.05), calculated using the Bonferroni-corrected multiple comparison test. One-way analysis of variance (ANOVA) was applied for normally distributed variables, and the Kruskal–Wallis H test was used for non-normally distributed variables Perinatal and obstetric outcomes in comparison of singleton pregnancies † p values represent comparisons between each POSEIDON group and the Non-POSEIDON group. Continuous variables with a normal distribution were analyzed using Student’s t test, whereas non-normally distributed variables were compared using the Mann–Whitney U test. Differences in categorical variables were assessed via chi-square tests or Fisher’s exact tests *indicate significant differences between each POSEIDON group (1–4) and the Non-POSEIDON group (p < 0.05), calculated using the Bonferroni-corrected multiple comparison test. One-way analysis of variance (ANOVA) was applied for normally distributed variables, and the Kruskal–Wallis H test was used for non-normally distributed variables

This study revealed that among 19,371 initial IVF/ICSI cycles, 36.7% (7109 cases) met the POSEIDON criteria for low ovarian response. These patients exhibited a dual decline in both the quantity and quality of oocytes/embryos, with significant deterioration in fertility outcomes such as the CPR and the CLBR. The perinatal data revealed elevated rates of gestational diabetes, preterm delivery, and low birth weight in the POSEIDON group, whereas metrics such as gestational hypertension showed no intergroup differences. For ART outcomes [ 33 ], we calculated the CPRs and CLBRs for the first IVF/ICSI cycle, as well as the PRs and LBRs for fresh embryo transfers. Interestingly, PG3 resulted in a slightly higher CLBR than PG2. This finding implied that maternal age (and therefore oocyte quality) rather than the quantitative ovarian reserve was the primary factor determining the CLBR [ 34 ]. This observation is consistent with the OPTIMIST study group analysis by Leijdekkers et al. [ 35 ], which showed that differences in cumulative live birth rates across POSEIDON groups are largely driven by female age rather than ovarian reserve parameters. Van Rooij et al. [ 36 ] reported that some young patients with DOR, despite having a low AFC, were still able to produce high-quality embryos. In contrast, for older women, the percent decline in blastocyst euploid probability increased progressively with advancing age [ 37 ]. Overall, ART was less likely to be successful for patients who met the POSEIDON criteria because of a decreased number of oocytes, which limited the number of embryos that could be produced. The condition might be exacerbated by advanced female age, affecting the availability of genetically normal embryos for transfer, and thereby impacting the CLBR per cycle started [ 38 – 40 ]. Regarding perinatal outcomes, the incidence of low birth weight increased in the PGs. This may be due to female age and could also be associated with a greater incidence of endometriosis, which increases the risk of preterm delivery, caesarean delivery, and delivery of small-for-gestational-age (SGA) infants [ 41 , 42 ]. Furthermore, no statistically significant differences were detected in key perinatal indicators (e.g., gestational hypertension, placenta previa, and fetal malformation) between PGs and the non-PGs. The lack of significant differences may be attributed to limitations in sample size or to the minimal impact of female age and low ovarian reserve on perinatal outcomes. Once a morphologically high-quality embryo is achieved and successfully establishes early pregnancy, the subsequent risk of obstetric and perinatal complications is relatively lower than that of embryos of poorer quality [ 43 , 44 ]. This may be attributed to better developmental competence, euploidy potential, and synchronized interactions between the endometrium and embryo. Notably, P3 and P4 patients received more GnRH antagonists and luteal phase stimulation, whereas P1 and P2 patients received more GnRH agonists. This distribution of stimulation protocols aligns with internationally recognized clinical practice patterns. For patients in PG1 and PG2 with normal or age-appropriate ovarian reserves, GnRH agonist or antagonist protocols are commonly preferred, offering reliable follicular synchronization and a satisfactory oocyte yield. In contrast, patients in PG3 and PG4, characterized by a diminished ovarian reserve (DOR), more frequently undergo GnRH antagonist protocols or luteal phase stimulation (LPS), which have been shown to be more flexible and effective in maximizing oocyte yield within a limited follicular pool. This protocol selection strategy is consistent with recommendations from the European Society of Human Reproduction and Embryology (ESHRE) and other expert consensus statements [ 45 , 46 ]. To further reduce potential confounding factors, the stimulation protocol type was included as a covariate in the multivariable regression analyses. Our findings indicated that patients in PG1, PG2, and PG3 exhibited acceptable pregnancy outcomes compared with non-PG patients. This finding suggested that a greater number of embryos may be necessary for live birth in these individuals, yet their ovarian reserve and response do not appear to affect the predicted embryo competency. Notably, the perinatal outcomes were generally similar to those of non-PG patients, suggesting that a fertilized oocyte from a young patient in PGs is no less likely to develop into a euploid blastocyst and result in a live birth [ 47 ]. The significantly lower pregnancy and live birth rates observed in PG4 are expected, as this group comprises older women with diminished ovarian reserve—a combination strongly associated with impaired oocyte competence and increased aneuploidy rates. Prior studies have consistently demonstrated that PG4 patients exhibit the lowest cumulative live birth rates among all POSEIDON groups, highlighting the need for early fertility counselling, adjunctive therapies, or consideration of donor oocytes in selected cases [ 12 , 48 ]. However, a relatively higher blastocyst formation rate was observed in the PG4 subgroup than in previously published studies of similarly defined poor ovarian responders [ 18 ]. This may be partially explained by the strict embryo selection criteria for extended culture, which are limited to Day 3 embryos with ≥ 4 blastomeres, ≤ 50% fragmentation, and regular morphology, as well as optimized laboratory conditions, including time-lapse incubation, sequential media, and individualized culture systems. Notably, Day 3 embryos that failed to meet the above criteria were excluded from the analysis of blastocyst formation rates. Therefore, the reported blastocyst formation rates reflect only the developmental competence of selected embryos that entered extended culture, rather than the entire cohort of fertilized oocytes. However, it is important to note that the blastocyst formation rate in the PG4 subgroup remained lower than that observed in the non-POSEIDON population and the PG1–PG3 subgroups. This reflects the inherent limitations associated with advanced maternal age and reduced ovarian reserve, such as lower oocyte quality, reduced embryo viability, and increased chromosomal abnormalities. These biological constraints continue to influence the developmental potential of embryos, despite improvements in stimulation and culture protocols. However, it is important to acknowledge the limitations of our study. A limitation of this study is the unequal sample distribution across POSEIDON subgroups, particularly in PG3 and PG4. Although this distribution reflects the underlying structure of the real-world ART population, in which normal responders account for the majority of IVF/ICSI cases, whereas poor responders—particularly younger patients with diminished ovarian reserve—are relatively less common in routine clinical practice. This disparity may introduce bias and reduce the statistical robustness of subgroup analyses for these subgroups. The potential impact was addressed by reporting both crude and adjusted ORs with 95% confidence intervals and by interpreting subgroup-level findings conservatively. Future studies incorporating more comprehensive data from multiple IVF/ICSI cycles are used to further validate the subgroup-specific pregnancy and perinatal outcomes within the POSEIDON classification. Furthermore, although POSEIDON classification permits finer stratification (e.g., PG1a vs. PG1b, PG2a vs. PG2b), we did not adopt this approach because the number of patients—particularly those with pregnancy outcomes—was insufficient to allow meaningful statistical analysis. This limitation underscores the need for larger, multicenter datasets in future research. This study has several notable strengths. It includes a large real-world cohort encompassing 19,371 IVF/ICSI cycles, enabling robust statistical comparisons across the 4 POSEIDON subgroups and with non-POSEIDON patients. The incorporation of both cumulative pregnancy and live birth outcomes, alongside detailed obstetric and perinatal data, provides a comprehensive view of reproductive prognosis. To avoid the confounding effects of repeated ovarian stimulation cycles and accumulated embryos [ 42 , 43 ], only the first oocyte retrieval cycle was included, with follow-up extended to live birth or depletion of all embryos. This design aligns with the clinical need to predict reproductive outcomes from the first cycle, thereby reducing the financial and physical burden of repeated treatments. The study population reflects the diversity of patients encountered in routine clinical practice, including those with cancelled cycles or no oocyte retrieval, thereby enhancing generalizability and real-world applicability. Unlike prior studies that have focused primarily on ovarian response or oocyte yield, this study emphasized treatment-dependent live birth and cumulative outcomes, which are more meaningful for clinical counselling and prognostication. This nuanced understanding of treatment efficacy may inform individualized strategies for patients who have characteristics associated with a poor prognosis. Moreover, this study provided additional insights into the rationale behind the POSEIDON classification. The higher CLBR observed in PG3 compared to PG2 suggests that maternal age may have a greater impact on reproductive outcomes than ovarian reserve. This finding reinforces the need for age-focused clinical strategies when managing low responders.

The proportion of infertility attributed to elderly individuals and/or those with a low ovarian reserve is increasing globally [ 1 – 3 ]. According to various definitions, the estimated prevalence rates of poor ovarian response (POR) and diminished ovarian reserve (DOR) ranged from 6 to 35% [ 4 – 6 ]. DOR patients exhibit diminished fertility and elevated miscarriage risks, creating significant time, financial strain, and psychological stress for their families [ 7 , 8 ]. Nevertheless, owing to the lack of precise diagnostic criteria and comprehensive management guidelines for DOR, research on the management and prognostic evaluation of different patient groups is scarce [ 3 , 9 , 10 ]. The POSEIDON group (patient-oriented strategies encompassing the individualized oocyte number) has established criteria to enhance the management of patients receiving ART who have characteristics associated with a poor prognosis [ 11 – 14 ]. The POSEIDON criteria establish a comprehensive management framework for advanced maternal age (AMA) and DOR patients through multidimensional stratification, encompassing the selection of ovarian stimulation protocols to embryo screening. However, achieving effective personalized guidance and preemptive interventions under the POSEIDON criteria necessitates multidimensional assessment of subgroup-specific pregnancy outcomes (e.g., live birth rate, cumulative clinical pregnancy) and composite clinical indicators (e.g., oocytes retrieved, normal fertilization rate). Previous studies have revealed no significant difference in the live birth rate (LBR) between POSEIDON Groups 1 and 2 compared with the non–POSEIDON group of women, whereas groups 3 and 4 had significantly lower LBRs [ 15 ]. However, few relevant studies exist. Additionally, there is an absence of analysis between POSEIDON subgroups, which is necessary to address clinical queries on whether patients with young-aged poor ovarian reserves and older-aged normal ovarian reserves can be similarly managed. Moreover, previous studies have focused solely on pregnancy outcomes while neglecting perinatal outcomes. Long-term perinatal follow-up is essential, as interventions during this critical window not only target immediate morbidity prevention but also optimize neurodevelopmental trajectories. This evidence-based approach may necessitate reevaluation of optimal diagnostic and therapeutic strategies for pregnant women classified within specific POSEIDON subgroups [ 16 ]. Thus, we compared pregnancy and perinatal outcomes among the POSEIDON subgroups to infer the significance of age, ovarian reserve, and ovarian response parameters in population stratification.

Below is the link to the electronic supplementary material. Supplementary file1 (XLSX 30 KB) Supplementary file1 (XLSX 30 KB)