How to balance the live birth rate and the multiple pregnancy rate by selecting the cleavage-stage embryo number and quality for POSEIDON Group 1 and Group 2? A retrospective study.

For unexpected low-prognosis patients from POSEIDON Group 1, transferring a single good-quality cleavage-stage embryo could have acceptable live birth rates without much concern about multiple births. For patients from POSEIDON Group 2, transferring mixed-quality cleavage-stage embryos was an option to improve the live birth rate and have reasonable multiple pregnancy risks.

Of the total population, 672 patients with frozen cleavage-stage embryo transfer cycles were classified into POSEIDON Group 1 and 503 were classified into Group 2. The number of patients transferring one embryo and two embryos was 83 versus 589 and 96 versus 407 in Poseidon Group 1 and Group 2, respectively. The baseline characteristics of subgroups are presented in Table  1 . There were significant differences in LH (3.69 (2.67, 5.08) versus 4.30 (3.27, 5.68), P  = 0.015), AMH (2.58 (1.73, 3.87) versus 2.86 (1.99, 4.66), P  = 0.026), and number of oocytes retrieved (6.00 (4.00, 8.00) versus 7.00 (5.00, 8.00), P  = 0.017) between SET and DET in Group 1. All indicators were comparable between SET and DET within Group 2, except for FSH (8.80 (7.12, 9.70) versus 7.90 (6.78, 9.70), P  = 0.024), and AMH (2.08 (1.60, 2.74) versus 2.37 (1.72, 3.52), P  = 0.006). Table 1 Baseline clinical characteristics of patients subgrouped by transferred embryo number in the first FET cycle Group 1 ( n  = 672) Group 2 ( n  = 503) SET ( n  = 83) DET ( n  = 589) p value SET ( n  = 96) DET ( n  = 407) p value Age (years) 32.00 (30.00, 33.00) 31.00 (29.00, 33.00) 0.302 38.00 (36.25, 41.00) 38.00 (36.00, 40.00) 0.057 Body mass index (kg/m 2 ) 20.81 (18.91, 23.53) 20.32 (19.00, 22.43) 0.410 22.03 (19.86, 23.96) 21.48 (19.79, 23.44) 0.397 Duration of infertility (years) 3.00 (2.00, 4.00) 3.00 (2.00, 5.00) 0.235 3.00 (2.00, 7.75) 3.00 (2.00, 5.00) 0.217 Indications for IVF  Tubal factor 39.76% 43.80% 0.707 50.00% 25.80% 0.500  Endometriosis 8.43% 5.09% 5.21% 4.67%  Male factor 21.69% 18.34% 13.54% 15.72%  Combined factors 10.84% 10.87% 10.42% 24.32%  Unknown factors 16.25% 13.58% 12.50% 12.78% AFC 9.50 (6.00, 14.00) 10.00 (8.00, 14.00) 0.062 8.00 (6.00, 10.00) 8.00 (6.00, 11.00) 0.293 FSH (IU/L) 7.79 (6.76, 8.60) 7.90 (6.70, 9.50) 0.573 8.80 (7.12, 9.70) 7.90 (6.78, 9.70) 0.024 LH (IU/L) 3.69 (2.67, 5.08) 4.30 (3.27, 5.68) 0.015 4.01 (3.12, 5.53) 4.08 (3.03, 5.49) 0.815 E2 (pg/mL) 47.00 (36.00, 58.00) 46.00 (35.00, 58.00) 0.470 48.00 (35.00, 60.00) 48.00 (36.00, 60.00) 0.926 P (ng/mL) 0.55 (0.37, 0.86) 0.57 (0.38, 0.83) 0.835 0.56 (0.37, 0.74) 0.54 (0.35, 0.80) 0.940 PRL (ng/mL) 12.51 (9.42, 16.05) 13.27 (9.91, 18.19) 0.103 13.36 (9.19, 16.61) 12.26 (9.36, 16.86) 0.975 T (ng/mL) 0.48 (0.32, 0.63) 0.45 (0.34, 0.60) 0.536 0.40 (0.29, 0.54) 0.40 (0.29, 0.54) 0.876 AMH (ng/mL) 2.58 (1.73, 3.87) 2.86 (1.99, 4.66) 0.026 2.08 (1.60, 2.74) 2.37 (1.72, 3.52) 0.006 COH protocol  GnRH-Agonist 44.58% 44.48% 0.987 40.63% 48.65% 0.157  GnRH-Antagonist 55.42% 55.52% 59.38% 51.35% Total Gn dose (IU) 2700.00 (2025.00, 35, 250.00) 2550.00 (1912.50, 3337.50) 0.278 3075.00 (2400.00, 3900.00) 3075.00 (2400.00, 3975.00) 0.808 Percentage of ICSIs (%) 31.33% 27.84% 0.510 30.21% 32.43% 0.674 No. oocyte retrieved 6.00 (4.00, 8.00) 7.00 (5.00, 8.00) 0.017 6.00 (3.00, 8.00) 6.00 (4.00, 8.00) 0.170 BMI body mass index, IVF in vitro fertilization, AFC antral follicle count, FSH follicle-stimulating hormone, LH luteinizing hormone, E2 Estradiol, P Progesterone, PRL prolactin, T testosterone, AMH anti-mullerian hormone COH controlled ovarian hyperstimulation, Gn gonadotropin, GnRH- gonadotropin-releasing hormone- Baseline clinical characteristics of patients subgrouped by transferred embryo number in the first FET cycle BMI body mass index, IVF in vitro fertilization, AFC antral follicle count, FSH follicle-stimulating hormone, LH luteinizing hormone, E2 Estradiol, P Progesterone, PRL prolactin, T testosterone, AMH anti-mullerian hormone COH controlled ovarian hyperstimulation, Gn gonadotropin, GnRH- gonadotropin-releasing hormone- To explore the influence factors on the live birth rate, potentially associated variables including age, BMI, FSH, LH, AMH, the number of oocytes retrieved, and the number of transferred embryos, were tested in the univariate and multivariate analyses (Table  2 ). In Poseidon Group 1, the number of transferred embryos was found to be significantly associated with the live birth rate (OR 1.756 95% CI 1.057–2.919, P  = 0.030) before adjustment for confounders, but the difference was not significant after adjustment for confounders (aOR 1.647 95% CI 0.985–2.755, P  = 0.057). This association in Group 2 was significant (OR 4.320, 95% CI 1.830–10.194, P  = 0.001 and aOR 3.707, 95% CI 1.547–8.884, P  = 0.003) even after adjusting for potential confounders. Additionally, age was shown to have a significant association with the live birth rate in both Group 1 (OR 0.903, 95% CI 0.846–0.963, P  = 0.002 and aOR 0.910, 95% CI 0.852–0.971, P  = 0.004) and 2 (OR 0.764, 95% CI 0.689–0.847, P  = 0.000 and aOR 0.782, 95% CI 0.703–0.870, P  = 0.000). And AMH was found to have a significant association with the live birth rate only in Group 2 before adjustment for confounders (OR 1.227 95% CI 1.087–1.385, P = 0.001). We did not observe other variables affecting in terms of live birth rate. Table 2 Logistics regression analysis on live birth in patients with unexpected poor prognosis Group 1 Group 2 Univariate analysis Multivariate analysis Univariate analysis Multivariate analysis OR (95% CI) p value OR (95% CI) p value OR (95% CI) p value OR (95% CI) p value Age 0.903 (0.846–0.963) 0.002 0.910 (0.852–0.971) 0.004 0.764 (0.689–0.847) 0.000 0.782 (0.703–0.870) 0.000 Body mass index/BMI 1.012 (0.963–1.063) 0.646 1.011 (0.960–1.065) 0.671 0.975 (0.904–1.051) 0.506 0.974 (0.902–1.052) 0.506 FSH 0.974 (0.914–1.038) 0.422 0.979 (0.913–1.049) 0.540 0.958 (0.886–1.036) 0.279 0.987 (0.929–1.049) 0.673 LH 1.039 (0.973–1.109) 0.252 1.025 (0.956–1.100) 0.482 0.977 (0.906–1.053) 0.539 0.939 (0.852–1.035) 0.206 AMH 1.016 (0.977–1.056) 0.417 1.005 (0.966–1.045) 0.808 1.227 (1.087–1.385) 0.001 1.139 (0.991–1.310) 0.068 No. oocyte retrieved 1.048 (0.977–1.124) 0.187 1.029 (0.958–1.106) 0.431 1.099 (0.995–1.214) 0.063 1.036 (0.928–1.157) 0.529 No. transferred number 1.756 (1.057–2.919) 0.030 1.647 (0.985–2.755) 0.057 4.320 (1.830–10.194) 0.001 3.707 (1.547–8.884) 0.003 Logistics regression analysis on live birth in patients with unexpected poor prognosis The outcomes of the SET group and the DET group stratified by Group 1 and Group 2 were displayed in Table  3 . For Poseidon Group 1, live birth rate (40.2% vs 27.7%, OR 1.756 95% CI 1.057–2.919, P  = 0.030; aOR 1.647 95% CI 0.985–2.755, P  = 0.057), conception rate (51.1% vs 33.7%, OR 2.053 95% CI 1.267–3.327, P  = 0.004; aOR 1.895 95% CI 1.162–3.092, P  = 0.010), clinical pregnancy rate (46.9% vs 30.1%, OR 2.046 95% CI 1.246–3.360, P  = 0.005; aOR 1.884 95% CI 1.140–3.114, P  = 0.014), and ongoing pregnancy rate ( 40.7% vs 28.9%, OR 1.691 95% CI 1.023–2.793, P  = 0.040; aOR 1.577 95% CI 0.948–2.623, P  = 0.080) were higher in the DET group than the SET group. However, compared with transferring a single embryo, the multiple birth rate (29.1% vs 0%, P  = 0.003) was found significantly higher in transferring double embryos. For the live birth rate in Poseidon Group 2, we found that the DET was significantly higher than that of the SET (22.4% vs 6.3%, OR 4.320 95% CI 1.830–10.194, P  = 0.001; aOR 3.707 95% CI 1.547–8.884, P  = 0.003), and that the conception rate (30.5% vs 14.6%, OR 2.566 95% CI 1.402–4.699, P  = 0.002; aOR 2.240 95% CI 1.205–4.166, P  = 0.011), clinical pregnancy rate (27.0% vs 12.5%, OR 2.593 95% CI 1.363–4.933, P  = 0.004; aOR 2.271 95% CI 1.175–4.389, P  = 0.015) and ongoing pregnancy rate (22.6% vs 7.3%, OR 3.713 95% CI 1.662–8.295, P  = 0.001; aOR 3.200 95% CI 1.412–7.252), P  = 0.005) were significantly higher than those of the SET group. An increase of 18.70% ( P  = 0.586) in the multiple birth rate without significant difference was shown between the groups. Table 3 Pregnancy outcomes of first FET cycle (SET versus DET) Group 1 Group 2 SET DET Crude OR (95% CI) p value Adjusted OR (95% CI) p value SET DET Crude OR (95% CI) p value Adjusted OR (95% CI) p value Number of patients 83 589 96 407 Endometrial thickness (mm) 11.03 ± 2.88 11.52 ± 2.35 / 0.141 / / 10.95 ± 2.44 11.05 ± 2.33 / 0.714 / / Conception (%) 28(33.7%) 301(51.1%) 2.053 (1.267–3.327) 0.004 1.895 (1.162–3.092) 0.010 14(14.6%) 124(30.5%) 2.566 (1.402–4.699) 0.002 2.240 (1.205–4.166) 0.011 Clinical pregnancy rate (%) 25(30.1%) 276(46.9%) 2.046 (1.246–3.360) 0.005 1.884 (1.140–3.114) 0.014 12/(12.5%) 110(27.0%) 2.593 (1.363–4.933) 0.004 2.271 (1.175–4.389) 0.015 Ongoing pregnancy rate (%) 24(28.9%) 240(40.7%) 1.691 (1.023–2.793) 0.040 1.577 (0.948–2.623) 0.080 7(7.3%) 92(22.6%) 3.713 (1.662–8.295) 0.001 3.200 (1.412–7.252) 0.005 Live birth rate (%) 23(27.7%) 237(40.2%) 1.756 (1.057–2.919) 0.030 1.647 (0.985–2.755) 0.057 6(6.3%) 91(22.4%) 4.320 (1.830–10.194) 0.001 3.707 (1.547–8.884) 0.003 Multiple birth rate (%) 0 69/237(29.1%) / 0.003 / / 0 17/91(18.7%) / 0.586 / / * Adjusted for age, BMI, FSH, LH, AMH, No. oocytes retrieved Pregnancy outcomes of first FET cycle (SET versus DET) * Adjusted for age, BMI, FSH, LH, AMH, No. oocytes retrieved The outcomes of subgroups stratified by embryo quality in Poseidon Group 1 and 2 were displayed in Table  4 . For patients in Group 1, we found that group GQE + GQE achieved a significantly higher clinical pregnancy rate (58.3% vs 36.2%, OR 2.466 95% CI 1.272–4.782, P  = 0.008; aOR 2.229 95% CI 1.138–4.366, P  = 0.019), ongoing pregnancy rate (54% vs 34%, OR 2.275 95% CI 1.166–4.439, P  = 0.016; aOR 2.122 95% CI 1.077–4.182, P  = 0.030) and live birth rate (53.5% vs 31.9%, OR 2.452 95% CI 1.246–4.827, P  = 0.009; aOR 2.302 95% CI 1.158–4.574, P  = 0.017) than the group GQE, and the differences were significant even after adjusting for potential confounders. Additionally, group GQE + PQE had also a higher clinical pregnancy rate (47.7% vs 36.2%, OR 1.608 95% CI 0.842–3.072, P  = 0.150; aOR 1.475 95% CI 0.764–2.845, P  = 0.247), ongoing pregnancy rate (40.9% vs 34%, OR 1.342 95% CI 0.696–2.588, P  = 0.379; aOR 1.272 95% CI 0.653–2.476, P  = 0.479) and live birth rate (40.5% vs 31.9%, OR 1.452 95% CI 0.746–2.827, P  = 0.272; aOR 1.382 95% CI 0.704–2.714, P = 0.347) than group GQE, all with the non-significant differences. Although improvement in outcome indicators could be seen in the group GQE + GQE and group GQE + PQE than the group GQE or PQE, the multiple birth rates increased to 31.0% ( P  = 0.027) and 28.1% ( P  = 0.042), respectively. And there were no significant differences in clinical pregnancy rate (32.7%vs 22.2%, OR 1.703 95% CI 0.727–3.985, P  = 0.220; aOR 1.645 95% CI 0.693–3.905, P  = 0.260), ongoing pregnancy rate (25.5% vs 22.2%, OR 1.195 95% CI 0.506–2.825, P  = 0.685; aOR 1.126 95% CI 0.463–2.737, P  = 0.794), and live birth rate between group PQE + PQE and group PQE (24.9% vs 22.2%, OR 1.157 95% CI 0.489–2.739, P  = 0.740; aOR 1.089 95% CI 0.447–2.655, P  = 0.851), and the addition of a PQE increased multiple birth rate (26.8%, P = 0.063). In Group 2, the live birth rate increased from 8.5% to 28.2% (OR 4.231, 95% CI 1.419–12.622, P  = 0.010; aOR 4.044, 95% CI 1.318–12.408 P  = 0.015) comparing the group GQE to the group GQE + GQE. Interestingly, we found that the live birth rate for group GQE + PQE had also a higher live birth rate than group GQE (26.1% vs 8.5%, OR 3.794 95% CI 1.272–11.316, P  = 0.017; aOR 3.219 95% CI 1.041–9.961, P  = 0.042). However, a significant difference of 24.10% was shown in the multiple birth rates between the group GQE + PQE (8.30%) and the group GQE + GQE (32.40%) (OR 0.220 95% CI 0.061–0.800, P  = 0.021; aOR 0.146 95% CI 0.035–0.605, P  = 0.008). The outcome in clinical pregnancy rate (8.2%vs 14.5%, OR 1.907 95% CI 0.618–5.886, P  = 0.262; aOR 1.591 95% CI 0.496–5.106, P  = 0.435), ongoing pregnancy rate (4.1% vs 13.0%, OR 3.525 95% CI 0.787–15.787, P  = 0.100; aOR 3.126 95% CI 0.677–14.430, P  = 0.144), and live birth rate (4.1% vs13.0%, OR 3.525 95% CI 0.787–15.787, P  = 0.100; aOR 3.126 95% CI 0.677–14.430, P  = 0.144) tends to be unsatisfactory for both transferring only one PQE or two PQEs. Table 4 Comparison of pregnancy outcomes of patients with different embryo quality in the first FET cycle Group 1 Group 2 GQE GQE + GQE GQE + PQE PQE + PQE PQE p value GQE GQE + GQE GQE + PQE PQE + PQE PQE p value Number of patients 47 187 237 165 36 47 131 138 138 49 Conception (%) 19 (40.4%) a,a’ 118 (63.1%) e,e’,f,f’ 123 (51.9%) g,g’ 60 (36.4%) 9 (25.0%) 0.000* 9 (19.1%) a,a’ 53 (40.5%) f,f’ 47 (34.1%) g,g’ 24 (17.4%) 5 (10.2%) 0.000* Clinical pregnancy rate (%) 17 (36.2%) a,a’ 109 (58.3%) e,e’,f,f’ 113 (47.7%) g,g’ 54 (32.7%) 8 (22.2%) 0.000* 8 (17.0%) a,a’ 49 (37.4%) e,f,f’ 41 (29.7%) g,g’ 20 (14.5%) 4 (8.2%) 0.000* Ongoing pregnancy rate (%) 16 (34.0%) a,a’ 101 (54.0%) e,e’,f,f’ 97 (40.9%) g,g’ 42( 25.5%) 8 (22.2%) 0.000* 5 (10.6%) a,a’,b 38 (29.0%) f,f’ 36 (26.1%) g,g’ 18 (13.0%) 2 (4.1%) 0.000* Live birth rate (%) 15 (31.9%) a,a’ 100 (53.5%) e,e’,f,f’ 96 (40.5%) g,g’ 41 (24.9%) 8 (22.2%) 0.000* 4 (8.5%) a,a’,b,b’ 37 (28.2%) f,f’ 36 (26.1%) g,g’ 18 (13.0%) 2 (4.1%) 0.000* Multiple birth rate (%) 0 a,b 31/100 (31.0%) f,f’ 27/96 (28.1%) 11/41 (26.8%) 0 0.092 0 12/37 (32.4%) e,e’,f,f’ 3/36 (8.3%) 2/18 (11.1%) 0 0.045* ** p  < 0.05 was statistical significance. Crude: “a” represents p value less than 0.05 between groups GQE and GQE + GQE, “b” represents p value less than 0.05 between groups GQE and GQE + PQE, “c” represents p value less than 0.05 between groups GQE and PQE + PQE, “d” represents p value less than 0.05 between groups GQE and PQE, “e” represents p value less than 0.05 between groups GQE + GQE and GQE + PQE, “f” represents p value less than 0.05 between groups GQE + GQE and PQE + PQE, “g” represents p value less than 0.05 between groups GQE + PQE and PQE + PQE, “h” represents p value less than 0.05 between groups PQE and PQE + PQE; Adjusted: “a ’ ” represents p value less than 0.05 between groups GQE and GQE + GQE, “b ’ ” represents p value less than 0.05 between groups GQE and GQE + PQE, “c’” represents p value less than 0.05 between groups GQE and PQE + PQE, “d ’ ” represents p value less than 0.05 between groups GQE and PQE, “e ’ ” represents p value less than 0.05 between groups GQE + GQE and GQE + PQE, “f ’ ” represents p value less than 0.05 between groups GQE + GQE and PQE + PQE, “g ’ ” represents p value less than 0.05 between groups GQE + PQE and PQE + PQE, “h’” represents p value less than 0.05 between groups PQE and PQE + PQE Comparison of pregnancy outcomes of patients with different embryo quality in the first FET cycle ** p  < 0.05 was statistical significance. Crude: “a” represents p value less than 0.05 between groups GQE and GQE + GQE, “b” represents p value less than 0.05 between groups GQE and GQE + PQE, “c” represents p value less than 0.05 between groups GQE and PQE + PQE, “d” represents p value less than 0.05 between groups GQE and PQE, “e” represents p value less than 0.05 between groups GQE + GQE and GQE + PQE, “f” represents p value less than 0.05 between groups GQE + GQE and PQE + PQE, “g” represents p value less than 0.05 between groups GQE + PQE and PQE + PQE, “h” represents p value less than 0.05 between groups PQE and PQE + PQE; Adjusted: “a ’ ” represents p value less than 0.05 between groups GQE and GQE + GQE, “b ’ ” represents p value less than 0.05 between groups GQE and GQE + PQE, “c’” represents p value less than 0.05 between groups GQE and PQE + PQE, “d ’ ” represents p value less than 0.05 between groups GQE and PQE, “e ’ ” represents p value less than 0.05 between groups GQE + GQE and GQE + PQE, “f ’ ” represents p value less than 0.05 between groups GQE + GQE and PQE + PQE, “g ’ ” represents p value less than 0.05 between groups GQE + PQE and PQE + PQE, “h’” represents p value less than 0.05 between groups PQE and PQE + PQE

This was a retrospective study performed at Peking University Shenzhen Hospital Reproductive Medicine Center. A total of 2970 patients were retrospectively classified according to the POSEIDON criteria into four groups based on the POSEIDON criteria as described below [ 2 ] in the period between January 2018 and December 2021. The POSEIDON criteria divided patients with a reduced ovarian response to exogenous gonadotropins into two main categories: the unexpected low prognosis (Groups 1 and 2) and the expected low prognosis (Groups 3 and 4). Group 1: patients < 35 years with AFC ≥ 5 and/or AMH ≥ 1.2 ng/mL with an unexpected poor or suboptimal ovarian response (retrieved ≤ 9 oocytes). Group 2: patients ≥ 35 years with AFC ≥ 5 and/or AMH ≥ 1.2 ng/mL and with an unexpectedly poor or suboptimal ovarian response. Group 3: patients < 35 years with AFC < 5 and/or AMH < 1.2 ng/ml. Group 4: patients ≥ 35 years with AFC < 5 and/or AMH < 1.2 ng/ml. AMH and AFC biomarkers were demonstrated to have equivalent values in predicting ovarian response [ 14 ]. Here AMH was used for ovarian reserve stratification to avoid subjective bias in ultrasound assessment for AFC. Our study only included patients who had been from Group 1 and Group 2 with cleavage-stage embryo(s) transferred in their first FET cycle. Patients in POSEIDON Groups 3 and 4, as well as those who underwent blastocyst transfer and those with failed follow-ups, were excluded. Patients were grouped into SET and DET based on transferred embryo number. According to embryo quality, patients were further classified into SET with one good-quality embryo (GQE), SET with one poor-quality embryo (PQE), DET with two GQE, DET with one GQE plus one PQE, and DET with two PQE (Fig.  1 ). Fig. 1 Flowchart Flowchart Ovarian response assessment was based on the standard ovarian stimulation procedures including gonadotrophin-releasing hormone (GnRH) antagonist and GnRH agonist (ultra-long/long) protocols. Gonadotropin (recombinant FSH (GONAL-f, Merck Serono, Germany) /purified hMG (Urofollit ALPHA ropin, Livzon Pharmaceuticals, China) /recombinant FSH plus hMG or recombinant LH (Luveris, Merck Serono, Germany), 2:1 ratio) was used initially with a dosage of 150-450 IU according to growing follicles monitored by ultrasound. When two follicles with diameters more than 18 mm were observed, HCG (Choriogonadotropin alfa, Merck Serono, Germany) or HCG combined with GnRH agonist was triggered to reach final maturation. Oocyte retrieval was implemented under transvaginal ultrasound after 36–38 h. The cryopreservation and thawing procedures were performed according to standard protocols of vitrification freezing and warming kits (Kitazato, Shizuoka, Japan). For vitrification, the embryos were first placed in an equilibrium solution for 9 min, then transferred to a vitrification-freezing solution for 1 min, and subsequently loaded onto the cryotron and immediately immersed in the liquid nitrogen within the same device for 60 s. When thawing embryos, the embryos were taken out of the liquid nitrogen and immediately placed into a thawing solution for 1 min. Next, the embryos were exposed to a dilution solution at room temperature for 3 min, followed by washing solution 1 for 5 min, and then washing solution 2 for 1 min, before finally being transferred into the culture medium. Endometrial preparation protocols included natural cycles (NC), hormone replacement therapy cycles (HRT) with or without gonadotropin-releasing hormone agonist (GnRH-a) therapy and Ovulation induction cycles. (i) NC: no exogenous drugs were used to promote follicle growth or induce ovulation, but rather the treatment relied on the natural menstrual cycle of the woman. (ii) HRT or HRT with GnRH-a: oral estradiol valerate (Progynova, Bayer-Schering Pharma AG, Berlin, Germany) was given to the patients. While the estradiol and endometrial thickness were suitable, progesterone supplementation was started. Or before hormone replacement therapy, GnRH-a (3.75 mg triptorelin acetate, Dophereline) was administered for down-regulation. (iii) ovulation induction cycles: medications such as human menopausal gonadotropin, clomiphene or letrozole were used to promote follicle growth and ovulation in the cycle. All the above protocols involved monitoring the size of follicles, endometrial thickness, and serum hormone levels with vaginal ultrasound to determine the timing of embryo transfer, which was performed when the endometrial thickness was ≥ 7mm. In our study, these methods were found to be comparable across the different groups. Embryo transfer decisions were opted for the specific clinical situation with fully informed consent considering factors including age, medical history, personal willingness and quantity and quality of the available embryos on the day of transfer. The quality of cleavage-stage embryos was commonly graded as I to IV on day 3 according to the 2023 Chinese expert consensus referring to blastomere counts, blastomere size, and degree of fragmentation [ 15 ]. Grade I referred to 8 embryo cells of even cell sizes and fragmentation < 10%. Grade II referred to 6–8 or ≥ 8 embryo cells of even or mostly even cell sizes and fragmentation < 10% or between 10 and 25%. Grade III referred to 4–5 embryo cells with varying in size and fragmentation  50%. We defined good-quality embryos as those with 7 to 9 blastomeres on the third day and classified as Grade I or II and otherwise PQE. The primary outcome was the live birth rate, and the secondary outcomes included the conception rate, clinical pregnancy rate, ongoing pregnancy rate, and multiple birth rate. The live birth rate was defined as a survived fetus born beyond 28 weeks or with a birth weight ≥ 1000 g. The conception rate was defined as the rate of positive detection of the urine or serum beta-HCG test. The clinical pregnancy rate was defined as the rate of ultrasound-confirmed intrauterine gestational sac 28 days after embryo transfer. The ongoing pregnancy rate was defined as the rate of ultrasound-confirmed intrauterine fetal heartbeat after 12 weeks of pregnancy. The multiple birth rate was defined as the rate of giving birth to more than one infant. SPSS 25.0 (IBM, Armonk, NY, USA) was used for statistics. The baseline characteristics for subgroups were reported as quartiles for continuous variables and analyzed by the Mann–Whitney U-Test. Categorical variables were expressed as percentages and the chi-square test or Fisher’s exact test was performed for analysis. Logistic regression was performed to explore the potential confounders for the live birth rate among age, BMI, FSH, LH, AMH, the number of oocytes retrieved, and the number of transferred embryos, and the results were expressed as OR with 95% CI. P  < 0.05 were considered to be statistically significant.

Hitherto, diminished ovarian reserve (DOR) or poor ovarian response (POR) remains a thorny issue during infertility treatment [ 1 ]. To manage patients with POR more efficiently, in 2016, the Patient-Oriented Strategies Encompassing IndividualizeD Oocyte Number (POSEIDON) criteria were established to develop a more nuanced stratification of the “low prognosis patient”. The Poseidon criteria were divided into four subgroups based on age (cut-off by 35 years), ovarian reserve parameters (AFC, AMH), and ovarian response [ 2 ]. Poor ovarian responders probably fail to obtain oocytes of sufficient quantity, resulting in unsatisfying pregnancy rates than normal responders [ 3 ]. Various interventions, including the pretreatment and ovarian stimulation protocols, have been intensely discussed to improve reproductive outcomes of POR [ 1 ]. However, in addition to the low number of eggs retrieved, there may be other underlying factors that are detrimental to the successful implantation of embryos due to low ovarian response. In general, there was a physiological decline in oocyte quality and an increase in the incidence of aneuploidy with increasing maternal age [ 4 ]. However, the question of whether the early ovarian hyporesponse in young patients is accompanied by decreased oocyte formation or embryo quality remains controversial. A recent study has demonstrated an endogenous steroid dysregulation of luteal granulosa cells from POR patients [ 5 ]. These might reveal that POR is not only a transient manifestation during ovarian stimulation but potentially impacts embryo implantation by oocyte or embryo quality and endometrium receptivity. Embryo transfer strategy for this patient population that involves embryo development stage, embryo quantity, and embryo quality is expected to be explored. With the advance of vitrification and culture technology, frozen blastocyst transfer has been widely applied in assisted reproductive technology (ART). However, extended culture to the blastocyst stage might cause the loss of embryos available for transfer [ 6 ] and lead to a higher failure rate than cleavage-stage embryo transfer for the arrested development during cell culture [ 7 ]. Prolonged embryo culture might also be associated with some adverse obstetric and perinatal outcomes, such as preterm birth, large for gestational age and monozygotic twin, and might also change the epigenetic mark of embryonic development [ 8 , 9 ]. Considering the potential limitation of extended culture, cleavage-stage embryo transfer might remain practical in some clinical situations especially when the number of embryos was in shortage. For POR patients who have a limited number of embryos, it is a challenge for physicians to tailor the most suited embryo transfer protocol by deciding the numbers and quality of transferred embryos. For decades, the development of assisted reproduction technology (ART) has not only improved the live birth rate but also increased the risk of multiple pregnancies, which is complicated with adverse obstetric and neonatal outcomes [ 10 ]. Single embryo transfer (SET) was recommended to reduce the complications associated with multiple pregnancies [ 11 ]. Holschbach V et al. pointed out that the number of embryo transfer was an important and independent factor affecting the outcomes of frozen embryo transfer [ 12 ]. The 2021 ASRM /SART guideline suggested that transferring more than one embryo could be an option in patients with unfavorable prognoses to improve the prognosis [ 13 ]. At present, there are few studies discussing the above issues among low prognosis patients stratified by the POSEIDON criteria. This study focused on the association of the live birth rates and multiple pregnancy rates with a certain quantity or quality of cleavage-stage embryos transferred for POSEIDON Group 1 and 2, trying to explore an embryo transfer strategy that balanced the efficiency and safety and improved the clinical outcomes for this population.

For Poseidon Group 1, considering the markedly increased risk of twin pregnancy resulting from double cleavage-embryo transfer, a single good-quality cleavage-stage embryo transfer could be attempted. For Poseidon Group 2, the combination of good- and poor-quality cleavage-stage embryos is superior to one or two good-quality embryos when trying to balance efficacy and safety.

To the best of our knowledge, there is insufficient evidence to support a rational embryo transfer strategy for POSEIDON low prognosis patients specific to the decisions on embryo number and quality. Our study showed that although DET including two GQEs or a GQE plus a PQE helped improve live birth rates, they led to a significantly increased risk of multiple births for patients in Poseidon Group 1. For Poseidon Group 2, DET including a GQE plus a PQE was found to have a benefit in live birth rates, with a modest increase in the multiple birth rate. Therefore, this study balanced the live birth rate and multiple birth rates by tailoring the number and quality of embryos, and tried to make optimal cleavage-embryo transfer decisions for low prognosis patients, the results of which could enrich current research in this field. According to the 2020 International Committee for Monitoring Assisted Reproductive Technologies (ICMART) preliminary world report on ART, the world average for delivery rate of frozen-thawed embryo cycle was 31.3% per transfer and the multiple rate was 16.5% in 2020 [ 16 ]. There is no denying that single blastocyst transfer is often recommended in clinical practice, as it tends to provide satisfying live birth rates per cycle and decrease multiple birth rates [ 17 ]. It was reported in a previous study that single cleavage-embryo transfer showed a poorer live birth rate (11.7% versus 38.1%, P  < 0.001) in a fresh cycle compared to single blastocyst transfer[ 18 ]. In contrast, in our study, the live birth rate of single cleavage-stage embryo transfer (27.7%) fell into a more acceptable outcome, especially single good-quality embryo (31.9%) in Poseidon Group 1. Meanwhile, based on our previous research concerning blastocyst transfer, younger patients have demonstrated superior clinical outcomes compared to older patients [ 19 ]. Consistently, our current study has revealed comparable outcomes (with live birth rates of 27.7% and 40.2% for SET and 6.3% and 22.4% for DET in Groups 1 and 2, respectively). It could be explained by a more recent report that euploidy rates for unexpected low-prognosis patients in Poseidon Group 1 were not significantly different from patients without a low prognosis and were not influenced by the ovarian reserve [ 20 ]. Although the DET group resulted in a higher success rate than the SET group for unexpected low-prognosis patients in the study, it obviously increased the multiple pregnancy rate. As recommended by the ASRM Committee, thorough counseling on trade-offs between risks and benefits is needed when offering an embryo transfer protocol to patients with very poor prognosis (greater than 1% and lesser than 5% success rate) or futility (lesser than 1% success rate) [ 21 ]. In light of the aforementioned considerations, it is posited that within the context of patients presenting with a dearth of blastocysts suitable for transfer or harboring apprehensions regarding the viability of blastocyst culture, the contemplation of a single, high-quality cleavage-stage embryo transfer emerged as a potential alternative strategy for select individuals within Poseidon Group 1. On the contrary, a meta-analysis study published in 2022 indicated that transferring one embryo alone could result in a decrease in live births in women at advanced reproductive age [ 8 ]. In 2015, Gleicher et al. showed that three or more embryos seemed to improve live birth rates in poor responders aged 35–40 by Bologna criteria [ 22 ]. However, an early study indicated that the rate of multiple pregnancies might increase in women with poor response who adopted double embryo transfer [ 23 ]. As described in the latest 2024 ESHRE guideline, there was insufficient definitive evidence regarding the comparison between single embryo transfer and double embryo transfer. The guideline also indicated that sometimes this patient population might choose to undergo double embryo transfer after careful consideration including advanced age, poor-quality embryos, and failure to achieve a live birth in previous ART cycles [ 24 ]. For Poseidon Group 2, it seemed that double embryo transfer may be preferable when merely focusing on the live birth rate. However, the effect of adding the extra number of cleavage-embryo is needed to try to guarantee safety as well as efficacy. When more than one embryo is adopted to transfer, the potential interaction among embryos may influence the growth and the implantation of each other. One hypothesis for this phenomenon was that the paracrine factors produced by the pre-implanted embryos might have detrimental effects on the development of the other [ 25 , 26 ]. If impaired, the embryo might signal to the endometrium and alter the receptivity to inhibit its implantation [ 27 ]. In addition to age and embryo number, embryo quality would be another consideration for maximizing the utilization of embryos in poor ovarian responders with limited embryos available for transfer when embryo accumulation procedures were not conducted. A retrospective multicenter study reported that both the pregnancy rate and implantation rate were lowered in DET with a GQE plus a PQE than in the good-quality embryo alone, indicating that the transfer of abnormal embryos could cause damage to the implantation rate of a good-quality embryo [ 28 ]. However, our study observed a beneficial effect of a poor-quality embryo to a good-quality embryo due to the increased live birth rate when transferred together. Another study found that transferring an impaired-quality embryo along with a good-quality embryo resulted in a lower implantation rate, than transferring two good-quality embryos [ 29 ] which was in line with our research results. We also found that embryo transfers with two good-quality embryos yielded higher multiple pregnancy rates than transfers with one good-quality and one poor-quality embryo. This could be attributed to the higher implantation chance of good-quality embryos than poor-quality embryos [ 30 , 31 ]. For unexpected low prognosis patients ≥ 35 years, we found that transferring the DET with one PQE plus one GQE as an alternative transfer scheme controlled the accepted multiple pregnancies (8.3%) below the world average whilst maintaining reasonable live birth rates (26.1%). To reduce the risk of a multiple pregnancy, the GQE + PQE transfer could be chosen over two GQEs transfer. In addition, we also noticed that either transferring a GQE or one or two PQEs was associated with a low live birth rate for patients aged ≥ 35 years with an unexpected low prognosis. A recent report suggested that the live birth rate of a GQE for blastocyst transfer was 42.99% in contrast, double cleavage-stage embryo transfer for hybrid-quality embryos could also achieve a relatively satisfactory live birth rate (37.25%) in the first frozen embryo transfer [ 32 ]. Though it did not target low responders and not stratify patients by age, this may indicate that without the availability of good-quality blastocyst, DET with GQE + PQE for cleavage-stage embryo transfer could be an alternative, as concluded by our study. However, several limitations of this study need to be addressed and further discussed. First, it was a retrospective design with inherent bias and unknown confounders. The disparity of the sample size for some subgroup counterparts may limit the statistical power, which could be attributed to the non-routine use of single-cleavage-stage embryo transfer at our center. Moreover, the Chinese standard for embryo quality assessment was used in this study while other criteria may be adopted in current research, which should be noticed when interpreting, comparing and analyzing. A large-scale multi-center prospective analysis is expected to further confirm these findings.