Impact of different endometrial preparation protocols on pregnancy outcomes in patients at high risk for ovarian hyperstimulation syndrome: a propensity score matched retrospective cohort study.

The flowchart for the enrollment of the study participants is depicted in Fig.  1 . A total of 1783 patients were enrolled ( n  = 1431 for HRT protocol; n  = 212 for NC protocol; n  = 140 for OS protocol). Fig. 1 Flow diagram showing patient selection. Note: FET: frozen-thawed embryo transfer; HRT: hormone replacement therapy; NC: natural cycle; OS: ovarian stimulation Flow diagram showing patient selection. Note: FET: frozen-thawed embryo transfer; HRT: hormone replacement therapy; NC: natural cycle; OS: ovarian stimulation Table  1 presents the baseline characteristics and information on FET cycles before matching across the three groups. Compared with the NC group, the HRT group was younger, had higher baseline FSH levels, and exhibited a higher prevalence of female-factor infertility. In the fresh cycle, HRT patients more frequently used the follicular-phase depot GnRH agonist protocol, while NC patients favored the mid-luteal GnRH agonist protocol. The NC group had the lowest BMI, the highest proportion of regular menstrual patterns, the lowest AFC, and the thinnest pre-transplantation endometrial thickness. Additionally, the HRT group had a lower incidence of triple-line endometrial patterns compared to the OS group. No significant differences were observed in other baseline characteristics among the three groups. To further compare pregnancy outcomes, PSM was employed to balance baseline characteristics and sample sizes across the groups. The matched dataset demonstrated balance in these factors, with detailed results presented in Table  2 . Table 1 Comparison of baseline characteristics and frozen-thawed embryo transfer cycle information of three groups before matching Characteristics HRT ( n  = 1431) NC ( n  = 212) OS ( n  = 140) P value Age of the woman (years, mean ± SD) 29.64 ± 3.41 30.37 ± 3.33 29.86 ± 3.04 0.011 Age of the man (years, mean ± SD) 31.32 ± 3.81 31.78 ± 3.47 31.20 ± 3.71 0.217 BMI (kg/m 2 ,mean ± SD) 22.94 ± 3.51 21.27 ± 2.61 22.65 ± 3.29 < 0.001 Regular pattern of menstruation, n(%) 804 (56.18) 194 (91.51) 92 (65.71) < 0.001 Type of infertility, n(%) 0.703  Primary 528 (36.90) 82 (38.68) 56 (40.00)  Secondary 903 (63.10) 130 (61.32) 84 (60.00) Infertility factors, n(%) < 0.001  Female factor 954 (66.67) 99 (46.70) 87 (62.14)  Male factor 227 (15.86) 69 (32.55) 30 (21.43)  Bilateral factor 176 (12.30) 19 (8.96) 15 (10.71)  Unexplained 74 (5.17) 25 (11.79) 8 (5.71) Years of infertility (years, mean ± SD) 3.37 ± 2.10 3.46 ± 2.30 3.29 ± 2.19 0.732 Basal FSH(U/L, mean ± SD) 6.21 ± 1.67 6.63 ± 2.08 6.23 ± 1.75 0.012 Basal LH (U/L, mean ± SD) 7.09 ± 6.96 6.88 ± 11.38 5.77 ± 4.07 0.110 AFC (n, mean ± SD) 19.14 ± 6.18 16.20 ± 4.84 18.35 ± 6.37 < 0.001 Protocol in fresh cycle, n(%) < 0.001  Follicular-phase depot GnRH agonist 611 (42.70) 65 (30.66) 52 (37.14)  Mid-luteal GnRH agonist 349 (24.39) 93 (43.87) 48 (34.29)  Antagonist 471 (32.91) 54 (25.47) 40 (28.57) Fertilization type, n(%) 0.798  IVF 1187 (82.95) 159 (75.00) 113 (80.71)  ICSI 244 (17.05) 53 (25.00) 27 (19.29) Gn time of use (d, mean ± SD) 10.62 ± 2.31 10.28 ± 1.91 10.50 ± 2.14 0.226 Gn dosage (U, mean ± SD) 1935.89 ± 771.14 1854.54 ± 665.56 1919.45 ± 785.78 0.460 Trigger day E2 value (ng/L, mean ± SD) 5631.25 ± 2542.10 6063.53 ± 2648.25 5444.94 ± 2688.26 0.320 No. of oocyte retrieved(n, mean ± SD) 19.91 ± 6.16 19.40 ± 5.65 19.76 ± 6.03 0.428 Pre-transplantation endometrial thickness mm, n(%) < 0.001  14 22 (1.54) 8 (3.77) 9 (6.43) Triple-line endometrial pattern, n(%) 0.001  A 237 (16.56) 52 (24.53) 39 (27.86)  B 1054 (73.65) 141 (66.51) 84 (60.00)  C 140 (9.78) 19 (8.96) 17 (12.14) Type of embryo transferred, n(%) 0.734 Cleavage stage 170 (11.88) 27 (12.74) 14 (10.00) Blastocyst stage 1261 (88.12) 185 (87.26) 126 (90.00) No. of embryos transferred, n(%) 0.471  1 1280 (89.45) 185 (87.26) 127 (90.71)  2 151 (10.55) 27 (12.74) 13 (9.29) Good quality embryo transferred, n(%) 1240 (86.65) 179 (84.43) 122 (87.14) 0.657 Note: HRT: hormone replacement therapy; NC: natural cycle; OS: ovarian stimulation; SD: standard deviation; BMI: body mass index; FSH: follicle stimulating hormone; LH: luteinizing hormone; AFC: antral follicle count; IVF: in vitro fertilization; ICSI: intracytoplasmic sperm injection; Gn: gonadotropin; E2:estradiol; Triple-line endometrial pattern: pattern A (a triple-line pattern consisting of a central hyperechoic line surrounded by two hypoechoic layers), pattern B (an intermediate isoechogenic pattern with the same reflectivity as the surrounding myometrium and a poorly defined central echogenic line), and pattern C (homogenous, hyperechogenic endometrium).Good quality embryo: good quality cleavage-stage embryos were defined as those derived from normally zygote with an embryonic cell number of 7–9 and a fragmentation level of less than 10 on day 3 after fertilisation; high-quality blastocysts were defined as those with a score of better than 4CC Comparison of baseline characteristics and frozen-thawed embryo transfer cycle information of three groups before matching Note: HRT: hormone replacement therapy; NC: natural cycle; OS: ovarian stimulation; SD: standard deviation; BMI: body mass index; FSH: follicle stimulating hormone; LH: luteinizing hormone; AFC: antral follicle count; IVF: in vitro fertilization; ICSI: intracytoplasmic sperm injection; Gn: gonadotropin; E2:estradiol; Triple-line endometrial pattern: pattern A (a triple-line pattern consisting of a central hyperechoic line surrounded by two hypoechoic layers), pattern B (an intermediate isoechogenic pattern with the same reflectivity as the surrounding myometrium and a poorly defined central echogenic line), and pattern C (homogenous, hyperechogenic endometrium).Good quality embryo: good quality cleavage-stage embryos were defined as those derived from normally zygote with an embryonic cell number of 7–9 and a fragmentation level of less than 10 on day 3 after fertilisation; high-quality blastocysts were defined as those with a score of better than 4CC Table 2 Comparison of baseline characteristics and frozen-thawed embryo transfer cycle information of three groups after matching Characteristics Matched I Matched II Matched III HRT ( n  = 394) NC ( n  = 194) P value HRT ( n  = 272) OS( n  = 136) P-value NC ( n  = 109) OS ( n  = 109) P value Age of the woman (years, mean ± SD) 30.09 ± 3.30 30.33 ± 3.37 0.411 29.94 ± 2.97 29.92 ± 3.02 0.953 29.85 ± 3.32 29.84 ± 3.09 0.983 Age of the man (years, mean ± SD) 31.74 ± 3.98 31.85 ± 3.54 0.745 31.29 ± 3.53 31.34 ± 3.66 0.907 31.18 ± 3.43 31.33 ± 3.96 0.770 BMI (kg/m 2 ,mean ± SD) 21.56 ± 2.99 21.45 ± 2.59 0.665 22.50 ± 3.61 22.72 ± 3.29 0.550 22.12 ± 2.80 22.05 ± 2.90 0.856 Regular pattern of menstruation, n(%) 352 (89.34) 179 (90.86) 0.665 171 (62.87) 88 (64.71) 0.799 92 (84.40) 88 (80.73) 0.592 Type of infertility, n(%) 0.929 0.803 0.780  Primary 159 (40.36) 78 (39.59) 111 (40.81) 53 (38.97) 43 (39.45) 40 (36.70)  Secondary 235 (59.64) 119 (60.41) 161 (59.19) 83 (61.03) 66 (60.55) 69 (63.30) Infertility factors, n(%) 0.196 0.667 0.644  Female factor 223 (56.60) 99 (50.25) 175 (64.34) 84 (61.76) 54 (49.54) 59 (54.13)  Male factor 95 (24.11) 64 (32.49) 45 (16.54) 29 (21.32) 33 (30.28) 29 (26.61)  Bilateral factor 38 (9.64) 17 (8.63) 32 (11.76) 15 (11.03) 10 (9.17) 13 (11.93)  Unexplained 38 (9.64) 17 (8.63) 20 (7.35) 8 (5.88) 12 (11.01) 8 (7.34) Years of infertility (years, mean ± SD) 3.41 ± 2.18 3.37 ± 2.28 0.834 3.26 ± 2.08 3.35 ± 2.19 0.663 3.05 ± 2.08 3.38 ± 2.31 0.278 Basal FSH(U/L, mean ± SD) 6.52 ± 1.84 6.49 ± 2.09 0.885 6.17 ± 1.78 6.19 ± 1.81 0.929 6.24 ± 2.03 6.20 ± 1.64 0.875 Basal LH (U/L, mean ± SD) 6.59 ± 8.53 6.39 ± 10.17 0.798 5.94 ± 3.47 5.78 ± 4.07 0.685 5.51 ± 4.91 5.44 ± 3.98 0.913 AFC (n, mean ± SD) 16.90 ± 6.42 16.43 ± 4.88 0.364 19.00 ± 7.10 18.43 ± 6.33 0.435 16.81 ± 4.88 17.09 ± 6.03 0.702 Protocol in fresh cycle, n(%) 0.736 0.292 0.594  Follicular-phase depot GnRH agonist 127 (32.23) 59 (29.95) 118 (43.38) 52 (38.24) 33 (30.28) 36 (33.03)  Mid-luteal GnRH agonist 159 (40.36) 86 (43.65) 70 (25.74) 45 (33.09) 38 (34.86) 42 (38.53)  Antagonist 108 (27.41) 52 (26.40) 84 (30.88) 39 (28.68) 38 (34.86) 31 (28.44) Fertilization type, n(%) 1.000 1.000 0.874  IVF 321 (81.47) 161 (81.73) 222 (81.62) 111 (81.62) 82 (75.23) 84 (77.06)  ICSI 73 (18.53) 36 (18.27) 50 (18.38) 25 (18.38) 27 (24.77) 25 (22.94) Gn time of use (d, mean ± SD) 10.38 ± 2.01 10.30 ± 1.96 0.683 10.61 ± 2.12 10.54 ± 2.15 0.755 10.41 ± 2.06 10.27 ± 1.68 0.565 Gn dosage (U, mean ± SD) 1858.66 ± 673.85 1863.97 ± 665.70 0.928 1946.19 ± 750.44 1941.27 ± 827.32 0.952 1950.94 ± 699.34 1900.11 ± 731.45 0.601 Trigger day E2 value (ng/L, mean ± SD) 5934.80 ± 2390.08 5986.06 ± 2572.71 0.811 5499.41 ± 2382.22 5436.10 ± 2656.40 0.808 5660.43 ± 2497.75 5514.87 ± 2668.19 0.678 No. of oocyte retrieved(n, mean ± SD) 19.64 ± 6.14 19.50 ± 5.86 0.792 20.08 ± 6.36 20.10 ± 6.45 0.974 19.39 ± 5.64 19.88 ± 5.85 0.533 Note: HRT: hormone replacement therapy; NC: natural cycle; OS: ovarian stimulation; Matched I: The HRT and NC groups were matched at a 2:1 ratio using a propensity score matching tolerance of 0.02. Matched II: The HRT and OS groups were matched at a 2:1 ratio using a propensity score matching tolerance of 0.02. Matched III: The NC and OS groups were matched at a 1:1 ratio using a propensity score matching tolerance of 0.02.SD: standard deviation; BMI: body mass index; FSH: follicle stimulating hormone; LH: luteinizing hormone; AFC: antral follicle count; IVF: in vitro fertilization; ICSI: intracytoplasmic sperm injection; Gn: gonadotropin; E2:estradiol; Comparison of baseline characteristics and frozen-thawed embryo transfer cycle information of three groups after matching Note: HRT: hormone replacement therapy; NC: natural cycle; OS: ovarian stimulation; Matched I: The HRT and NC groups were matched at a 2:1 ratio using a propensity score matching tolerance of 0.02. Matched II: The HRT and OS groups were matched at a 2:1 ratio using a propensity score matching tolerance of 0.02. Matched III: The NC and OS groups were matched at a 1:1 ratio using a propensity score matching tolerance of 0.02.SD: standard deviation; BMI: body mass index; FSH: follicle stimulating hormone; LH: luteinizing hormone; AFC: antral follicle count; IVF: in vitro fertilization; ICSI: intracytoplasmic sperm injection; Gn: gonadotropin; E2:estradiol; Based on the initial data, the HRT group exhibited the lowest live birth rate (54.79%), along with the highest rate of miscarriage (18.43%) and preterm birth (< 37 weeks: 11.61%). In contrast, the NC group had the lowest cesarean delivery rate(63.16%) and the highest proportion of SGA(7.52%). The OS group showed the highest live birth rate (70.71%) and lowest miscarriage rate (6.48%). Following PSM, the NC group demonstrated significantly higher rate of live birth (63.96% vs. 53.81%), biochemical pregnancy (77.16% vs. 68.27%), and clinical pregnancy (73.60% vs. 64.97%), alongside a significantly lower cesarean delivery rate (63.49% vs. 78.77%) when compared to the HRT group. Similarly, the OS group showed significantly higher live birth rate (70.59% vs. 51.84%), biochemical pregnancy rate (81.62% vs. 69.49%), and clinical pregnancy rate (77.21% vs. 66.91%), as well as a significantly lower miscarriage rate (6.67% vs. 20.33%) compared to the HRT group. However, no statistically significant differences were observed in pregnancy outcomes between the NC and OS groups following matching (Table  3 ). Table 3 Comparison of pregnancy outcomes before and after matching Pre-matching Matched I Matched II Matched III HRT ( n  = 1431) NC ( n  = 212) OS ( n  = 140) P value HRT ( n  = 394) NC ( n  = 194) P value HRT ( n  = 272) OS( n  = 136) P-value NC ( n  = 109) OS ( n  = 109) P value Live birth rate, n(%) 784 (54.79) 133 (62.74) 99 (70.71) < 0.001 212 (53.81) 126 (63.96) 0.024 141 (51.84) 96 (70.59) < 0.001 72 (66.06) 74 (67.89) 0.884 Biochemical pregnancy rate, n(%) 1044 (72.96) 161 (75.94) 114 (81.43) 0.510 269 (68.27) 152 (77.16) 0.031 189 (69.49) 111 (81.62) 0.012 85 (77.98) 85 (77.98) 1.000 Clinical pregnancy rate, n(%) 982 (68.62) 154 (72.64) 108 (77.14) 0.070 256 (64.97) 145 (73.60) 0.043 182 (66.91) 105 (77.21) 0.042 80 (73.39) 79 (72.5) 1.000 Miscarriage rate, n(%) 181 (18.43) 19 (12.34) 7 (6.48) 0.002 40 (15.62) 18 (12.41) 0.465 37 (20.33) 7 (6.67) 0.004 8 (10.00) 4 (5.06) 0.380 Gestational age (week), mean ± SD 38.83 ± 1.93 38.91 ± 1.85 38.93 ± 1.66 0.804 38.97 ± 1.82 38.91 ± 1.89 0.780 38.95 ± 1.90 38.88 ± 1.66 0.777 38.87 ± 2.23 39.01 ± 1.66 0.662 < 37 weeks, n (%) 91 (11.61) 6 (4.51) 10 (10.10) 0.047 23 (10.85) 6 (4.76) 0.053 12 (8.51) 10 (10.42) 0.620 3 (4.17) 7 (9.46) 0.348 < 34 weeks, n (%) 25 (3.19) 2 (1.50) 2 (2.02) 0.487 6 (2.83) 2 (1.59) 0.467 135 (95.74) 94 (97.92) 0.587 2 (2.78) 2 (2.70) 0.978 Twin birth rate, n(%) 30 (3.83) 2 (1.50) 4 (4.04) 0.392 10 (4.72) 2 (1.59) 0.230 4 (2.84) 4 (4.17) 0.849 1 (1.39) 2 (2.70) 1.000 Cesarean delivery, n(%) 589 (75.13) 84 (63.16) 76 (76.77) 0.011 167 (78.77) 80 (63.49) 0.003 102 (72.34) 75 (78.12) 0.394 47 (65.28) 55 (74.32) 0.312 Birth weight (g, mean ± SD) 3309.09 ± 533.72 3288.01 ± 560.07 3285.05 ± 476.49 0.853 3299 0.01 ± 526.32 3297.26 ± 568.34 0.977 3345.67.± 528.62 3280.42 ± 482.87 0.335 3343.96 ± 643.56 3308.04 ± 512.43 0.709 LBW, n (%) 47 (5.99) 7 (5.26) 5 (5.05) 0.893 16 (7.55) 7 (5.56) 0.482 10 (7.09) 5 (5.21) 0.559 4 (5.56) 5 (6.76) 0.763 HBW, n (%) 72 (9.18) 9 (6.77) 8 (8.08) 0.640 20 (9.43) 9 (7.14) 0.467 8 (5.67) 8 (8.33) 0.423 8 (11.11) 8 (10.81) 0.953 SGA, n (%) 32 (4.08) 10 (7.52) 0 (0.00) 0.017 9 (4.25) 9 (7.14) 0.251 4 (2.84) 0 (0.00) 0.250 2 (2.78) 0 (0.00) 0.464 LGA, n (%) 123 (15.69) 20 (15.04) 9 (9.09) 0.225 31 (14.62) 20 (15.87) 0.756 20 (14.18) 9 (9.38) 0.364 13 (18.06) 8 (10.81) 0.311 Note: HRT: hormone replacement therapy; NC: natural cycle; OS: ovarian stimulation; Matched I: The HRT and NC groups were matched at a 2:1 ratio using a propensity score matching tolerance of 0.02. Matched II: The HRT and OS groups were matched at a 2:1 ratio using a propensity score matching tolerance of 0.02. Matched III: The NC and OS groups were matched at a 1:1 ratio using a propensity score matching tolerance of 0.02.SD: standard deviation; LBW: low birth weight; HBW: high birth weight; SGA: small for gestational age; LGA: large for gestational age Comparison of pregnancy outcomes before and after matching Note: HRT: hormone replacement therapy; NC: natural cycle; OS: ovarian stimulation; Matched I: The HRT and NC groups were matched at a 2:1 ratio using a propensity score matching tolerance of 0.02. Matched II: The HRT and OS groups were matched at a 2:1 ratio using a propensity score matching tolerance of 0.02. Matched III: The NC and OS groups were matched at a 1:1 ratio using a propensity score matching tolerance of 0.02.SD: standard deviation; LBW: low birth weight; HBW: high birth weight; SGA: small for gestational age; LGA: large for gestational age Based on the results of multivariate regression analysis, we explored the relationship between endometrial preparation protocols and pregnancy outcomes in the matched dataset.After adjusting for confounding factors, the NC group had higher rates of live birth (adjusted OR:1.50,95% CI:1.03–2.19) and biochemical pregnancy (adjusted OR: 1.57, 95% CI: 1.03–2.39), along with a lower cesarean delivery rate (adjusted OR: 0.44, 95% CI: 0.26–0.74) compared to the HRT group. Similarly, the OS group demonstrated higher rate of live birth (adjusted OR: 2.53, 95% CI:1.55–4.14), biochemical pregnancy (adjusted OR:2.14,95% CI: 1.22–3.75), and clinical pregnancy (adjusted OR: 1.86, 95% CI: 1.10–3.15), alongside a lower miscarriage rate (adjusted OR: 0.29, 95% CI: 0.12–0.71) compared to the HRT group(Table  4 ; Fig.  2 ). Fig. 2 Forest plot of endometrial preparation protocols and pregnancy outcomes. Note: NC: natural cycle; OS: ovarian stimulation; *reference HRT; **:reference NC; CI: confidence interval; Adjusted for pre-transplantation endometrial thickness, triple-line endometrial pattern, type, number and quality of embryos transferred Forest plot of endometrial preparation protocols and pregnancy outcomes. Note: NC: natural cycle; OS: ovarian stimulation; *reference HRT; **:reference NC; CI: confidence interval; Adjusted for pre-transplantation endometrial thickness, triple-line endometrial pattern, type, number and quality of embryos transferred Table 4 Relationship between endometrial Preparation and pregnancy outcomes Pregnancy outcomes Endometrial preparation Adjusted model OR/β(95% CI) P value Live birth rate HRT Reference NC 1.50 (1.03, 2.19) 0.033 OS 2.53 (1.55, 4.14) < 0.001 NC Reference OS 1.04(0.58, 1.87) 0.906 Biochemical pregnancy rate HRT Reference NC 1.57 (1.03, 2.39) 0.035 OS 2.14 (1.22, 3.75) 0.008 NC Reference OS 0.87 (0.44, 1.70) 0.678 Clinical pregnancy rate HRT Reference NC 1.46 (0.98, 2.18) 0.065 OS 1.86 (1.10, 3.15) 0.021 NC Reference OS 0.88 (0.47, 1.64) 0.679 Miscarriage rate HRT Reference NC 0.74 (0.39, 1.40) 0.355 OS 0.29 (0.12, 0.71) 0.007 NC Reference OS 0.46 (0.12, 1.75) 0.256 Gestational age HRT Reference NC -0.18 (-0.61, 0.26) 0.430 OS 0.08 (-0.43, 0.58) 0.766 NC Reference OS 0.30 (-0.38, 0.97) 0.386 < 37 weeks HRT Reference NC 0.41 (0.15, 1.12) 0.082 OS 0.83 (0.30, 2.27) 0.719 NC Reference OS 2.44 (0.57, 10.47) 0.229 < 34weeks HRT Reference NC 0.84 (0.16, 4.48) 0.834 OS 0.19 (0.03, 1.20) 0.077 NC Reference OS 0.69 (0.09, 5.54) 0.724 Twin birth rate HRT Reference NC 0.43 (0.08, 2.38) 0.331 OS 1.42 (0.20, 10.12) 0.725 NC Reference OS 0.74(0.02, 27.72) 0.868 Cesarean delivery HRT Reference NC 0.44 (0.26, 0.74) 0.002 OS 1.65 (0.84, 3.25) 0.144 NC Reference OS 1.45 (0.68, 3.10) 0.338 Birth weight HRT Reference NC -42.86 (-170.72, 85.01) 0.512 OS -49.81 (-190.17, 90.54) 0.487 NC Reference OS -46.99 (-244.49, 150.51) 0.642 LBW HRT Reference NC 0.82 (0.30, 2.21) 0.696 OS 0.44 (0.13, 1.52) 0.193 NC Reference OS 1.20 (0.29, 4.97) 0.802 HBW HRT Reference NC 0.60 (0.25, 1.45) 0.258 OS 1.45 (0.45, 4.70) 0.540 NC Reference OS 0.96 (0.31, 2.944) 0.945 SGA HRT Reference NC 2.28 (0.82, 6.34) 0.116 OS / / NC Reference OS / / LGA HRT Reference NC 0.85 (0.43, 1.66) 0.626 OS 0.64 (0.25, 1.59) 0.334 NC Reference OS 0.47 (0.17, 1.30) 0.147 Note: Univariate logistic regression and linear regression were used to calculate OR (95% CI) and β (95% CI), respectively OR: Odds ratio; CI: confidence interval; HRT: hormone replacement therapy; NC: natural cycle; OS: ovarian stimulation; LBW: low birth weight; HBW: high birth weight; SGA: small for gestational age; LGA: large for gestational age; Adjusted for pre-transplantation endometrial thickness, triple-line endometrial pattern, type, number and quality of embryos transferred Relationship between endometrial Preparation and pregnancy outcomes Note: Univariate logistic regression and linear regression were used to calculate OR (95% CI) and β (95% CI), respectively OR: Odds ratio; CI: confidence interval; HRT: hormone replacement therapy; NC: natural cycle; OS: ovarian stimulation; LBW: low birth weight; HBW: high birth weight; SGA: small for gestational age; LGA: large for gestational age; Adjusted for pre-transplantation endometrial thickness, triple-line endometrial pattern, type, number and quality of embryos transferred A stratified analysis was performed based on menstrual regularity using matched data to evaluate the impact of various endometrial preparation protocols on pregnancy outcomes.In both the regular and irregular menstrual subgroups, the OS protocol was associated with a significantly higher live birth rate compared to the HRT group. In patients with regular menstrual cycles, the NC group demonstrated a significantly lower cesarean delivery rate, but a higher proportion of SGA infants, compared to the HRT group. In contrast, in patients with irregular menstrual cycles, the OS group exhibited biochemical pregnancy rate, and clinical pregnancy rate, as well as a lower miscarriage rate and birth weight compared to the HRT group (Supplemental Table S1 ). After adjusting for potential confounding factors, we conducted a multivariate regression analysis using unmatched data. The results showed that, compared to the HRT group, the NC group had significantly higher rate of live birth, biochemical pregnancy, and clinical pregnancy, as well as lower rate of preterm birth (< 37 weeks) and cesarean delivery rate. The OS group demonstrated superior outcomes in terms of live birth rate and biochemical pregnancy rate, with significantly lower miscarriage rates and a lower proportion of LGA infants compared to the HRT group. Additionally, the OS group exhibited a higher cesarean delivery rate compared to the NC group (Supplemental Table S2 ).These findings align with those of the primary analysis, further supporting the impact of different endometrial preparation protocols on pregnancy outcomes.

The study population included women undergoing ART treatment at the Center for Assisted Reproductive Technology of Northwest Women’s and Children’s Hospital in the People’s Republic of China between January 2021 and December 2022. All patients were followed up for more than one year. Inclusion criteria were: (1) patients at high risk of OHSS; (2) patients undergoing their first FET cycle after the cryopreservation of all embryos; (3) patients with ≥ 1 cryopreserved cleavage-stage embryo or blastocyst; and (4) patients aged 20–40 years. Exclusion criteria included: (1) patients with uterine malformations, hydrosalpinx, endometriosis, adenomyosis, uterine fibroids, endometrial abnormalities, or other conditions that could affect ET outcomes; (2) patients with genetic abnormalities or hereditary diseases; (3) patients with a history of recurrent implantation failure or recurrent spontaneous abortion; and (4) patients whose embryo transfer was canceled, were lost to follow-up, or had missing data. Criteria for high risk of OHSS leading to cancellation of the fresh ET cycle included: retrieval of ≥ 15 oocytes in the fresh cycle with an estradiol level ≥ 18,350 pmol/L on the day of oocyte maturation induction; or the presence of clinical symptoms such as abdominal pain, bloating, nausea, and vomiting after oocyte retrieval; or a maximum ovarian diameter > 8 cm three days after oocyte retrieval, with or without significant pelvic effusion. This study was approved by the ethics committee of Northwest Women’s and Children’s Hospital (Grant number: 2022007) and has been performed in accordance with the principles of the Declaration of Helsinki.The ethics committee that approved this study waived the need to obtain informed consent. Endometrial preparation protocols were customized based on both the patient’s specific condition and the clinician’s expertise. Patients in the HRT group involved oral administration of estradiol valerate at 6 mg per day on the fifth day of their menstrual cycle.Serum progesterone levels were measured and vaginal ultrasonography was performed 10–12 days later to assess endometrial thickness.When endometrial thickness reached ≥ 7 mm and serum progesterone levels were < 1.5 ng/mL, endometrial transformation was initiated. This transformation was achieved through the intramuscular injection of progesterone at 60 mg daily.Embryo transfer occurred on the fourth day for cleavage-stage embryos and on the sixth day for blastocyst-stage embryos following endometrial transformation. Patients in the NC group, vaginal ultrasound was performed to monitor follicular growth on days 8–10 of the menstrual cycle. When the diameter of the dominant follicle reached ≥ 17 mm and the thickness of the endometrium was ≥ 7 mm, 10,000 International Units (IU) of human chorionic gonadotropin (HCG) were administered to trigger ovulation. Endometrial transformation (intramuscular injection of progesterone at 60 mg daily) was performed on the day of ovulation. Embryo transfer occurred on the third day for cleavage-stage embryos and on the fifth day for blastocyst-stage embryos following endometrial transformation. Patients in the OS group, letrozole (2.5-5.0 mg per day) was administered orally from days 3–5 of the menstrual cycle. The dose of human menopausal gonadotropin (HMG) was adjusted based on the diameter of the dominant follicle after 5 days of treatment. The protocols for triggering ovulation, endometrial transformation, and ET selection followed the same procedures as those used in the NC group. Luteal phase support started from the day of beginning the progesterone administration for HRT cycles and from the day of ovulation for NC and OS cycles. For patients on the HRT protocol, the regimen consisted of a progesterone injection at 60 mg per day, administered intramuscularly, along with oral progesterone at 30 mg per day and estradiol valerate at 6 mg per day. For patients on the NC and OS protocols, the regimen consisted of a progesterone injection at 60 mg per day, administered intramuscularly, along with oral progesterone at 30 mg per day. Serum beta-human chorionic gonadotropin (β-hCG) level was measured 12 or 14 days after ET. If the serum β-hCG was ≥ 50 IU/L, luteal support continued until 10 weeks of gestation, and a vaginal ultrasound was performed 5 weeks after ET to confirm the presence of the gestational sac and fetal heartbeat. If the β-hCG was < 50 IU/L, luteal support was discontinued. According to American Society for Reproductive Medicine (ASRM) 2017 consensus definitions [ 15 ], the definition of clinical outcomes were as follows: Live birth was defined as the birth of at least one live-born infant at ≥ 22 weeks of gestation. Biochemical pregnancy was determined by serum β-hCG levels ≥ 50 IU/L measured 14 days after ET. Clinical pregnancy was defined as the presence of one or more gestational sacs observed by ultrasound at 6–8 weeks after ET. Miscarriage was characterized by the clinical loss of pregnancy before the 22nd week of gestation. Gestational age (GA) was determined by adding the days from ET to birth, with Day 17 for cleavage-stage ET and Day 19 for blastocyst transfer [ 16 ].Preterm birth (PTB) was defined as GA of less than 37 weeks; Moderate preterm birth (MPTB) was defined as GA less than 34 weeks [ 17 ]. Low birth weight (LBW) was defined as a birth weight  4000 g. Small for gestational age (SGA) was defined as a birth weight below the 10th percentile, and large for gestational age (LGA) was defined as above the 90th percentile, based on Chinese population birthweight percentiles [ 18 ]. Based on our retrospective cohort study, the live birth rate in the HRT group at our center was 55.1%. According to previous studies, we hypothesized that the live birth rate for the NC and OS group would be 65.1%, representing a 10% increase compared to the HRT group. We assumed a 10% minimal clinically important difference and set the effect size to medium (Cohen’s f = 0.25). Using a two-sided test with a significance level of 5.0% and 80% statistical power, sample size estimation via G*Power indicated that approximately 120–140 participants would be needed per group. Considering a 5% drop-out rate, we adjusted the sample size to 140–150 participants per group, resulting in a total required sample size of 420–450 participants. However, the actual sample sizes in our study were 1431 participants for the HRT group, 212 participants for the NC group and 140 participants for the OS group. These actual sample sizes far exceed the estimated requirements, ensuring that the study has sufficient power to detect clinically significant differences in live birth rates among the three groups. Continuous variables were presented as means ± standard deviations (SD), while categorical variables were summarized as frequencies and percentages. For group comparisons, the Student’s t-test or analysis of variance (ANOVA) was used for normally distributed continuous variables, while the Kruskal-Wallis H test or Mann-Whitney U test was applied to non-normally distributed continuous variables. Categorical variables were analyzed using the Chi-square test, with the exact Pearson Chi-square test applied when the assumptions of the Chi-square test were not met. To address differences in baseline characteristics and account for sample size discrepancies across the three groups, propensity score matching (PSM) was employed. Matching was performed based on baseline characteristics prior to endometrial preparation, including the woman’s age, the man’s age, body mass index (BMI), menstrual regularity, type, factors and years of infertility, basal follicle-stimulating hormone (FSH) and luteinizing hormone (LH), antral follicle count (AFC), protocol in fresh cycle, fertilization type, timing and dosage of gonadotropin use, estradiol levels on the trigger day, and number of oocytes retrieved. Matching was conducted using the HRT group as the control in a 2:1 ratio with the NC and OS groups, and with the NC group as the control in a 1:1 ratio with the OS group, using a matching tolerance of 0.02. After adjusting for potential confounders, multiple regression analyses were performed on the matched datasets to assess the associations between endometrial preparation protocols and pregnancy outcomes. Sensitivity analyses were conducted to ensure the robustness of the results, including (i) regression analyses on unmatched datasets and (ii) stratified analyses based on menstrual regularity. All statistical analyses were performed using EmpowerStats ( www.empowerstats.com , X&Y Solutions, Inc., Boston, MA, USA) and R software ( http://www.r-project.org ). A P-value of < 0.05 was considered statistically significant.

In recent years, frozen embryo transfer (FET) has become increasingly widespread in assisted reproductive technology (ART). This growing trend is primarily attributed to advancements in embryo cryopreservation techniques, particularly the development of vitrification (ultra-rapid freezing), which significantly improves the survival rates of thawed embryos [ 1 ] and enhances live birth outcomes [ 2 ] compared to traditional slow freezing methods.Moreover, FET has been associated with higher live birth rates [ 3 ]and lower preterm birth rates [ 4 ]compared to fresh embryo transfer (ET), while also significantly reducing the risk of ovarian hyperstimulation syndrome (OHSS) [ 5 ]. OHSS is an iatrogenic complication of ovarian stimulation, characterized by ovarian enlargement, ascites, and pleural effusion, which in severe cases can lead to multi-organ failure or death. OHSS typically occurs during the luteal phase or early pregnancy and can be triggered by almost all ovulation-stimulating medications [ 6 ]. Several studies have demonstrated that selective cryopreservation of whole embryos is an effective strategy to mitigate the risk of OHSS [ 7 – 9 ]. The “implantation window” refers to the period when the embryo first embeds into the endometrium, and successful ET hinges on synchronizing early embryo development with endometrial readiness [ 10 ].Different endometrial preparation protocols aim to optimize this synchronization to improve implantation and pregnancy outcomes. Common protocols include hormone replacement therapy (HRT), natural cycle (NC), and ovarian stimulation (OS). However, existing studies comparing these protocols have reported inconsistent results regarding their effects on pregnancy outcomes [ 11 – 14 ]. With the increasing prevalence of ART procedures, OHSS has garnered significant attention. Many patients opt for whole embryo cryopreservation over fresh embryo transfer to reduce the risk of OHSS. However, there is limited research specifically addressing the optimal endometrial preparation protocols for patients at high risk of OHSS.Clinicians typically select among available protocols based on their experience and the patient’s condition, potentially overlooking the most suitable option for some patients, which could impact pregnancy outcomes adversely. Therefore, this retrospective cohort study aimed to compare the pregnancy outcomes of three different protocols (HRT, NC, and OS) and identify the optimal endometrial preparation protocol for patients at high risk of OHSS.

In conclusion, the results of our study demonstrated that in patients at high risk for OHSS, the OS protocol and the NC protocol for endometrial preparation resulted in better pregnancy outcomes compared to the HRT protocol.

Our primary objective of this study was to evaluate the association between different endometrial preparation protocols and pregnancy outcomes among patients at high risk for OHSS.After adjusting for potential confounders, our results indicated that both the NC and OS protocols were associated with significantly higher live birth rate, biochemical pregnancy rate, and clinical pregnancy rate compared to the HRT protocol.Additionally, the NC protocol showed a lower cesarean delivery rate, while the OS protocol had a significantly lower miscarriage rate compared to the HRT protocol. Several studies have examined the clinical outcomes of different endometrial preparation protocols, but the results remain inconsistent. Two previous studies from our center found that, among patients with regular menstrual cycles and those experiencing recurrent miscarriage, the NC protocol was associated with a higher live birth rate and a lower miscarriage rate compared to the HRT protocol [ 19 , 20 ].Furthermore, a meta-analysis comparing reproductive outcomes between the HRT and OS protocols in women with PCOS, which included eight retrospective studies and one randomized controlled trial (RCT), indicated that the OS protocol resulted in higher live birth and clinical pregnancy rate, along with lower miscarriage rate [ 21 ]. Additionally, a large retrospective cohort study involving 12,950 FET cycles reported that live birth rate in HRT cycles was slightly lower than that in NC cycles. Meanwhile, miscarriage rate in both HRT and OS cycles were significantly higher compared to NC cycles [ 22 ].However, other studies have reported no significant differences in pregnancy outcomes among the HRT, NC, and OS protocols [ 23 – 26 ]. The inconsistencies in these findings may be attributed to differences in patient populations, sample sizes, and study designs. To the best of our knowledge, this is the first study to specifically evaluate endometrial preparation protocols in patients at high risk for OHSS. Successful embryo implantation depends on the synchronization between the embryo and the maternal endometrium, a process regulated by hormonal influences [ 27 ].In the NC protocol, the natural hormonal fluctuations during the ovulatory cycle closely resemble physiological conditions, creating an optimal environment for embryo implantation. This natural hormonal environment improves maternal circulatory function and avoids the potential adverse effects of exogenous hormones [ 28 – 30 ].In the OS protocol, the use of letrozole may enhance endometrial receptivity [ 31 ]. Recently, significant increases in endometrial receptivity markers and ultrasound indicators for evaluating endometrial receptivity have been reported with the use of letrozole [ 31 – 33 ].Additionally, letrozole improves the hormonal environment for implantation by reducing luteal phase estradiol levels and increasing progesterone levels [ 34 ]. In contrast, the HRT protocol relies on exogenous hormonal supplementation, with non-physiological hormone levels potentially impairing endometrial receptivity and placental formation [ 29 , 31 ].Furthermore, the absence of a corpus luteum in HRT cycles leads to insufficient endogenous progesterone production. Progesterone is essential for endometrial preparation and successful embryo implantation, and low serum progesterone levels combined with corpus luteum deficiency have been shown to reduce live birth rates and increase miscarriage rates [ 35 – 37 ]. This study had several strengths. Firstly, the relatively large sample size, the consideration of more confounding factors, and the use of PSM contributed to the accuracy of the outcomes. Secondly, our data of study came from the actual clinical data, minimizing observational bias.Thirdly, this is the first study to specifically evaluate endometrial preparation protocols in high-risk OHSS patients, providing valuable insights for improving the management of these patients in clinical settings. This study also had some limitations. First, as a single-center retrospective study, our results may be subject to selection bias, as patients were assigned to different protocols based on clinician preferences and patient choices. Second, the OS group had a small sample size, which may affect the precision of the results and limit our ability to draw robust conclusions about this protocol. Third, although we adjusted for several confounders, there may still be unmeasured factors, such as lifestyle choices, psychological factors, or variations in clinical management, that could influence pregnancy outcomes.Future research should prioritize larger, multi-center, prospective studies to validate the findings of this study and improve the generalizability of the results. Additionally, studies should explore potential effect modifiers, such as age and BMI, to better understand how these factors interact with endometrial preparation protocols. This study underscores the potential advantages of the NC and OS protocols relative to the HRT protocol for patients at high risk for OHSS. Nevertheless, the clinical implementation of these protocols may encounter several challenges—most notably for the NC protocol, which necessitates frequent monitoring and may impose considerable financial and logistical burdens. It is imperative that patient preferences—such as the inclination for fewer clinic visits or a readiness to adhere to medication regimens—be judiciously considered when determining the most appropriate protocol. Ultimately, addressing these practical issues is vital for the customization of endometrial preparation protocols to suit individual patients, thereby enhancing pregnancy outcomes following embryo transfer.

Below is the link to the electronic supplementary material. Supplementary Material 1 Supplementary Material 1