How endometrial thickness affects pregnancy and perinatal outcomes: a retrospective study of 16,566 frozen-thawed single blastocyst transfer cycles.

A total of 16,566 single blastocyst FET cycles were analyzed. The EMT in the entire cohort ranged from 2 mm to over 20 mm, with the majority of patients falling within the 7–10 mm range (Fig.  1 A). Notably, four LBs occurred in women with an EMT  20 mm. Further analysis of the EMT distribution stratified by pregnancy outcome showed a greater density of women with an EMT < 7 mm among those who did not achieve a LB than among those who did (Fig.  1 B). Similarly, women who experienced miscarriage showed a higher density of EMT < 7 mm than those who did not (Fig.  1 C). Fig. 1 Primary outcomes by endometrial thickness groups. A Overall distribution of endometrial thickness; B Density of endometrial thickness by live birth status; C Density of endometrial thickness by miscarriage status; D Adjusted partial effect (log odds) of endometrial thickness on live birth estimated by a generalized additive model. The partial effect of endometrial thickness on live birth was expressed as the estimated adjusted log odds. The black curve represents the estimated adjusted log odds of experiencing live birth, and the shaded gray area around the black curve indicates the corresponding 95% CI; E Adjusted partial effect of endometrial thickness on gestational age estimated by a generalized additive model; F Distribution of stabilized IPTW across endometrial thickness groups, indicating adequate weight overlap and absence of extreme weights. IPTW: inverse probability of treatment weights, EMT: endometrial thickness Primary outcomes by endometrial thickness groups. A Overall distribution of endometrial thickness; B Density of endometrial thickness by live birth status; C Density of endometrial thickness by miscarriage status; D Adjusted partial effect (log odds) of endometrial thickness on live birth estimated by a generalized additive model. The partial effect of endometrial thickness on live birth was expressed as the estimated adjusted log odds. The black curve represents the estimated adjusted log odds of experiencing live birth, and the shaded gray area around the black curve indicates the corresponding 95% CI; E Adjusted partial effect of endometrial thickness on gestational age estimated by a generalized additive model; F Distribution of stabilized IPTW across endometrial thickness groups, indicating adequate weight overlap and absence of extreme weights. IPTW: inverse probability of treatment weights, EMT: endometrial thickness Participants were grouped according to EMT as follows: ≤7 mm ( N  = 751, 4.5%), (7,8] mm ( N  = 1,492, 9.0%), (8,9] mm ( N  = 5,771, 34.8%), (9,10] mm ( N  = 4,563, 27.5%), and > 10 mm ( N  = 3,989, 24.1%) (Table  1 ). Given that embryo transfers may be cancelled on an individualized basis when the EMT is < 8 mm, it should be noted that the ≤ 7 mm and (7,8] mm categories represent clinically selected subgroups. A positive trend was observed in the proportions of women achieving LB and CP with increasing EMT. Notably, women with an EMT of ≤ 7 mm had lower LB (24.6%) and CP (35.2%) rates compared to other groups (43.4–50.6% for LB; 53.4–61.0% for CP). The proportions of other pregnancy outcomes remained relatively stable across the EMT groups. Table 1 Characteristics and pregnancy outcomes of 16,566 FET cycles stratified by endometrial thickness Variable Endometrial Thickness Group (mm) ≤ 7 (7,8] (8,9] (9,10] > 10 N  = 751 N  = 1,492 N  = 5,771 N  = 4,563 N  = 3,989 Maternal age at oocyte retrieval, yrs 32.0 (29.0, 36.0) 32.0 (29.0, 35.0) 31.0 (28.0, 34.0) 31.0 (28.0, 34.0) 31.0 (28.0, 34.0) Category by maternal age at oocyte retrieval, yrs  < 35 506 (67.4%) 1,093 (73.3%) 4,452 (77.1%) 3,667 (80.4%) 3,233 (81.5%)  ≥ 35 245 (32.6%) 399 (26.7%) 1,319 (22.9%) 896 (19.6%) 756 (19.0%) BMI, kg/m² 21.5 (20.0, 23.4) 21.2 (19.6, 23.4) 21.3 (19.6, 23.4) 21.3 (19.6, 23.5) 21.8 (20.0, 24.0) BMI category, kg/m²  Underweight (< 18.5) 72 (9.6%) 172 (11.5%) 689 (11.9%) 508 (11.1%) 371 (9.3%)  Normal weight (18.5–23.9) 526 (70.0%) 1,027 (68.8%) 3,904 (67.6%) 3,047 (66.8%) 2,574 (64.5%)  Overweight/Obesity (≥ 24.0) 153 (20.4%) 293 (19.6%) 1,178 (20.4%) 1,008 (22.1%) 1,044 (26.2%) Baseline FSH 1 , mIU/mL 7.3 (6.2, 8.5) 7.2 (6.1, 8.5) 7.2 (6.1, 8.4) 7.1 (6.1, 8.4) 7.2 (6.1, 8.4)  Missing (N) 0 1 6 3 3 AMH level, ng/ml 3.6 (2.1, 6.3) 3.7 (2.2, 6.4) 4.0 (2.3, 6.9) 4.1 (2.3, 7.1) 4.0 (2.2, 6.8) Duration of infertility, yrs 2.0 (1.0, 4.0) 3.0 (1.0, 4.0) 3.0 (2.0, 4.0) 3.0 (2.0, 4.0) 3.0 (2.0, 5.0) Infertility diagnosis  Primary infertility 340 (45.3%) 807 (54.1%) 3,716 (64.4%) 3,137 (68.7%) 2,915 (73.1%)  Secondary infertility 411 (54.7%) 685 (45.9%) 2,055 (35.6%) 1,426 (31.3%) 1,074 (26.9%) Infertility etiology  Male factor 198 (26.4%) 468 (31.4%) 2,107 (36.5%) 1,804 (39.5%) 1,625 (40.7%)  Tubal 355 (47.3%) 633 (42.4%) 2,276 (39.4%) 1,667 (36.5%) 1,401 (35.1%)  Ovulatory 115 (15.3%) 157 (10.5%) 412 (7.1%) 300 (6.6%) 301 (7.5%)  DOR 44 (5.9%) 107 (7.2%) 413 (7.2%) 324 (7.1%) 274 (6.9%)  Endometriosis 6 (0.8%) 20 (1.3%) 66 (1.1%) 39 (0.9%) 43 (1.1%)  Others 33 (4.4%) 107 (7.2%) 497 (8.6%) 429 (9.4%) 345 (8.6%) Ovarian stimulation protocols  GnRH antagonist 334 (44.5%) 591 (39.6%) 2,239 (38.8%) 1,728 (37.9%) 1,456 (36.5%)  GnRHa ultra-long 251 (33.4%) 518 (34.7%) 2,188 (37.9%) 1,879 (41.2%) 1,651 (41.4%)  Long GnRHa 87 (11.6%) 237 (15.9%) 859 (14.9%) 632 (13.9%) 557 (14.0%)  PPOS 66 (8.8%) 107 (7.2%) 362 (6.3%) 237 (5.2%) 252 (6.3%)  Other protocols 13 (1.7%) 39 (2.6%) 123 (2.1%) 87 (1.9%) 73 (1.8%) Endometrial preparation protocols  GnRHa + HRT 131 (17.4%) 197 (13.2%) 538 (9.3%) 517 (11.3%) 758 (19.0%)  HRT 563 (75.0%) 1,214 (81.4%) 5,009 (86.8%) 3,819 (83.7%) 2,896 (72.6%)  NC 57 (7.6%) 81 (5.4%) 224 (3.9%) 227 (5.0%) 335 (8.4%) No. of oocytes retrieved 13.0 (9.0, 18.0) 13.0 (9.0, 19.0) 14.0 (10.0, 20.0) 14.0 (10.0, 20.0) 13.0 (9.0, 19.0) Oocyte maturation rate 0.9 (0.8, 1.0) 0.9 (0.8, 1.0) 0.9 (0.8, 1.0) 0.9 (0.8, 1.0) 0.9 (0.8, 1.0) Fertilization  ICSI 158 (21.0%) 360 (24.1%) 1,620 (28.1%) 1,406 (30.8%) 1,182 (29.6%)  IVF 572 (76.2%) 1,070 (71.7%) 3,888 (67.4%) 2,935 (64.3%) 2,626 (65.8%)  Rescue ICSI 21 (2.8%) 62 (4.2%) 263 (4.6%) 222 (4.9%) 181 (4.5%) Normal fertilization rate 0.7 (0.6, 0.8) 0.7 (0.5, 0.8) 0.7 (0.5, 0.8) 0.6 (0.5, 0.8) 0.6 (0.5, 0.8) Maternal age at FET, yrs 33.0 (30.0, 37.0) 32.0 (29.0, 35.0) 31.0 (29.0, 35.0) 31.0 (29.0, 34.0) 31.0 (29.0, 34.0) Interval between FET and IVF/ICSI, days 151.0 (73.0, 329.0) 105.0 (59.0, 241.0) 83.0 (39.0, 184.0) 80.0 (38.0, 181.0) 89.0 (56.0, 187.0) Embryo of good quality 532 (70.8%) 1061 (71.1%) 4163 (72.1%) 3365 (73.7%) 2796 (70.1%) LB 185/751 (24.6%) 648/1,492 (43.4%) 2,630/5,771 (45.5%) 2,173/4,563 (47.6%) 2,021/3,989 (50.6%) CP 264/751 (35.2%) 796/1,492 (53.4%) 3,244/5,771 (56.2%) 2,693/4,563 (59.0%) 2,432/3,989 (61.0%) Miscarriage 79/264 (29.9%) 148/796 (18.6%) 614/3,244 (18.9%) 520/2,693 (19.3%) 411/2,432 (16.9%) Biochemical pregnancy 65/751 (8.7%) 126/1,492 (8.4%) 481/5,771 (8.3%) 360/4,563 (7.9%) 308/3,989 (7.7%) Ectopic pregnancy 1/751 (0.1%) 6/1,492 (0.4%) 23/5,771 (0.4%) 14/4,563 (0.3%) 5/3,989 (0.1%) Multiple pregnancy 14/751 (1.9%) 21/1,492 (1.4%) 83/5,771 (1.4%) 70/4,563 (1.5%) 66/3,989 (1.7%) All categorical variables are presented as N (%); all continuous variables are presented as median (Q1,Q3) 1 baseline FSH levels summarized using complete data BMI Body mass index, FSH Follicle stimulating hormone, AMH Anti-Müllerian hormone, DOR Diminished ovarian reserve, GnRH Gonadotropin-releasing hormone, GnRHa Gonadotropin-releasing hormone agonist, NC Natural cycle, HRT Hormone replacement treatment, PPOS Progestin-primed ovarian stimulation, ICSI Intracytoplasmic sperm injection, IVF In vitro fertilization, FET Frozen embryo transfer, LB Live birth, CP Clinical pregnancy Characteristics and pregnancy outcomes of 16,566 FET cycles stratified by endometrial thickness 185/751 (24.6%) 648/1,492 (43.4%) 264/751 (35.2%) 796/1,492 (53.4%) 79/264 (29.9%) 148/796 (18.6%) All categorical variables are presented as N (%); all continuous variables are presented as median (Q1,Q3) 1 baseline FSH levels summarized using complete data BMI Body mass index, FSH Follicle stimulating hormone, AMH Anti-Müllerian hormone, DOR Diminished ovarian reserve, GnRH Gonadotropin-releasing hormone, GnRHa Gonadotropin-releasing hormone agonist, NC Natural cycle, HRT Hormone replacement treatment, PPOS Progestin-primed ovarian stimulation, ICSI Intracytoplasmic sperm injection, IVF In vitro fertilization, FET Frozen embryo transfer, LB Live birth, CP Clinical pregnancy Figure 1 D illustrates a non-monotonic relationship regarding the partial effect of EMT on LB. The partial effect of EMT on LB increases steeply at < 8 mm, translating minor EMT increments into significant outcome improvements. This effect plateaus between 8 and 10 mm, and exhibits a slower rate of improvement from 10 to 12 mm. Conversely, an EMT exceeding 12 mm yields minimal benefit and is associated with a slight decline in LB likelihood. The clinical characteristics of 7,366 single blastocyst FET cycles evaluated for singleton perinatal outcomes are summarized in Table  2 . Median GA increased from 38.4 weeks (IQR: 37.6–39.0) in the ≤ 7 mm cohort to 39.0 weeks in thicker EMT groups. Median birth weight rose from 3.2 kg (IQR: 3.0–3.5) in the ≤ 7 mm group to 3.4 kg (IQR: 3.1–3.6) in the > 10 mm group. Correspondingly, pregnancies with an EMT ≤ 7 mm had the highest rates of LBW (9.7%) and PTB (15.3%), but the lowest rate of macrosomia (3.4%) compared with the other EMT groups. After adjusting for confounders, Fig.  1 E illustrates a non-linear relationship between EMT and GA; the partial effect curve rises steeply for EMT values < 8 mm, further indicating that a lower EMT is associated with shorter GA. Table 2 Clinical characteristics and perinatal outcomes of 7,366 singleton live births by endometrial thickness Variable Endometrial Thickness Group (mm) ≤ 7 (7,8] (8,9] (9,10] > 10 N  = 176 N  = 615 N  = 2,534 N  = 2,084 N  = 1,957 Maternal age at oocyte retrieval, yrs 31.0 (28.0, 34.0) 31.0 (28.0, 34.0) 30.0 (28.0, 33.0) 30.0 (28.0, 33.0) 30.0 (28.0, 33.0) Category by maternal age at oocyte retrieval, yrs  < 35 133 (75.6%) 500 (81.3%) 2,121 (83.7%) 1,814 (87.0%) 1,696 (86.7%)  ≥ 35 43 (24.4%) 115 (18.7%) 413 (16.3%) 270 (13.0%) 261 (13.3%) BMI, kg/m² 21.2 (19.9, 23.2) 21.2 (19.6, 23.2) 21.3 (19.6, 23.3) 21.3 (19.6, 23.4) 21.9 (19.9, 24.0) BMI category, kg/m²  Underweight (< 18.5) 23 (13.1%) 68 (11.1%) 304 (12.0%) 230 (11.0%) 191 (9.8%)  Normal weight (18.5–23.9) 123 (69.9%) 436 (70.9%) 1,735 (68.5%) 1,405 (67.4%) 1,262 (64.5%)  Overweight/Obesity (≥ 24.0) 30 (17.0%) 111 (18.0%) 495 (19.5%) 449 (21.5%) 504 (25.8%) Baseline FSH 1 , mIU/mL 7.3 (6.1, 8.5) 7.1 (6.0, 8.2) 7.0 (6.0, 8.3) 7.1 (6.0, 8.2) 7.1 (6.0, 8.3)  Missing (N) 0 1 2 1 3 AMH level, ng/ml 3.6 (1.9, 6.8) 4.0 (2.5, 6.6) 4.4 (2.7, 7.5) 4.5 (2.7, 7.4) 4.3 (2.5, 7.4) Duration of infertility, yrs 2.0 (1.0, 4.0) 2.5 (1.0, 4.0) 3.0 (2.0, 4.0) 3.0 (2.0, 4.0) 3.0 (2.0, 4.0) Infertility diagnosis  Primary infertility 88 (50.0%) 355 (57.7%) 1,710 (67.5%) 1,468 (70.4%) 1,452 (74.2%)  Secondary infertility 88 (50.0%) 260 (42.3%) 824 (32.5%) 616 (29.6%) 505 (25.8%) Infertility etiology  Male factor 49 (27.8%) 216 (35.1%) 946 (37.3%) 827 (39.7%) 795 (40.6%)  Tubal 86 (48.9%) 264 (42.9%) 1,002 (39.5%) 796 (38.2%) 691 (35.3%)  Ovulatory 18 (10.2%) 52 (8.5%) 194 (7.7%) 139 (6.7%) 154 (7.9%)  DOR 12 (6.8%) 30 (4.9%) 133 (5.2%) 104 (5.0%) 114 (5.8%)  Endometriosis 1 (0.6%) 5 (0.8%) 25 (1.0%) 15 (0.7%) 20 (1.0%)  Others 10 (5.7%) 48 (7.8%) 234 (9.2%) 203 (9.7%) 183 (9.4%) Ovarian stimulation protocols  GnRH antagonist 84 (47.7%) 247 (40.2%) 977 (38.6%) 757 (36.3%) 677 (34.6%)  GnRH-a ultra-long 58 (33.0%) 228 (37.1%) 1,009 (39.8%) 933 (44.8%) 905 (46.2%)  Long GnRH-a 16 (9.1%) 105 (17.1%) 392 (15.5%) 294 (14.1%) 266 (13.6%)  PPOS 17 (9.7%) 25 (4.1%) 107 (4.2%) 67 (3.2%) 75 (3.8%)  Other protocols 1 (0.6%) 10 (1.6%) 49 (1.9%) 33 (1.6%) 34 (1.7%) Endometrial preparation protocols  GnRHa + HRT 30 (17.0%) 64 (10.4%) 210 (8.3%) 243 (11.7%) 382 (19.5%)  HRT 126 (71.6%) 520 (84.6%) 2,239 (88.4%) 1,761 (84.5%) 1,425 (72.8%)  NC 20 (11.4%) 31 (5.0%) 85 (3.4%) 80 (3.8%) 150 (7.7%) No. of oocytes retrieved 13.0 (9.0, 19.0) 15.0 (10.0, 20.0) 15.0 (10.0, 21.0) 15.0 (10.5, 21.0) 14.0 (10.0, 20.0) Oocyte maturation rate 0.9 (0.8, 1.0) 0.9 (0.8, 1.0) 0.9 (0.8, 1.0) 0.9 (0.8, 1.0) 0.9 (0.8, 1.0) Fertilization  ICSI 41 (23.3%) 167 (27.2%) 729 (28.8%) 642 (30.8%) 574 (29.3%)  IVF 132 (75.0%) 422 (68.6%) 1,687 (66.6%) 1,341 (64.3%) 1,298 (66.3%)  Rescue ICSI 3 (1.7%) 26 (4.2%) 118 (4.7%) 101 (4.8%) 85 (4.3%) Normal fertilization rate 0.7 (0.6, 0.8) 0.7 (0.5, 0.8) 0.6 (0.5, 0.8) 0.6 (0.5, 0.8) 0.6 (0.5, 0.8) Maternal age at FET, yrs 32.0 (29.0, 35.0) 31.0 (28.0, 34.0) 31.0 (28.0, 34.0) 31.0 (29.0, 33.0) 31.0 (28.0, 33.0) Interval between FET and IVF/ICSI, days 141.5 (68.5, 305.5) 91.0 (41.0, 201.0) 78.0 (39.0, 170.0) 75.0 (36.0, 175.5) 87.0 (56.0, 179.0) Embryo of good quality 140 (79.5%) 497 (80.8%) 2047 (80.8%) 1709 (82.0%) 1521 (77.7%) Delivery mode  Cesarean delivery 159 (90.3%) 526 (85.5%) 2,176 (85.9%) 1,760 (84.5%) 1,635 (83.5%)  Natural labor 17 (9.7%) 89 (14.5%) 358 (14.1%) 324 (15.5%) 322 (16.5%) Fetal Sex  Female 104 (59.1%) 333 (54.1%) 1,421 (56.1%) 1,158 (55.6%) 1,075 (54.9%)  Male 72 (40.9%) 282 (45.9%) 1,113 (43.9%) 926 (44.4%) 882 (45.1%) GA, wk 38.4 (37.6, 39.0) 39.0 (38.0, 39.4) 39.0 (38.1, 39.6) 39.0 (38.0, 39.6) 39.0 (38.1, 39.6) Birth weight, kg 3.2 (3.0, 3.5) 3.3 (3.0, 3.6) 3.4 (3.1, 3.6) 3.4 (3.1, 3.7) 3.4 (3.1, 3.6) LBW ( 4000 g) 6 (3.4%) 31 (5.0%) 151 (6.0%) 124 (6.0%) 120 (6.1%) PTB (< 37wk) 27 (15.3%) 60 (9.8%) 197 (7.8%) 164 (7.9%) 158 (8.1%) GDM 6 (3.4%) 22 (3.6%) 90 (3.6%) 77 (3.7%) 82 (4.2%) HDP 9 (5.1%) 20 (3.3%) 83 (3.3%) 64 (3.1%) 67 (3.4%) Placenta previa 5 (2.8%) 13 (2.1%) 52 (2.1%) 49 (2.4%) 35 (1.8%) Fetal malformation 1 (0.6%) 2 (0.3%) 6 (0.2%) 6 (0.3%) 3 (0.2%) All categorical variables are presented as N (%); all continuous variables are presented as median (Q1,Q3) 1 baseline FSH levels summarized using complete data BMI Body mass index, FSH Follicle stimulating hormone, AMH Anti-Müllerian hormone, DOR Diminished ovarian reserve, GnRH Gonadotropin-releasing hormone, GnRHa Gonadotropin-releasing hormone agonist, HRT Hormone replacement treatment, NC Natural cycle, PPOS Progestin-primed ovarian stimulation, ICSI Intracytoplasmic sperm injection, IVF  In vitro fertilization, FET Frozen embryo transfer, GA Gestational age, GDM Gestational diabetes mellitus, HDP Hypertensive disorders of pregnancy, LBW Low birth weight, PTB Preterm birth Clinical characteristics and perinatal outcomes of 7,366 singleton live births by endometrial thickness All categorical variables are presented as N (%); all continuous variables are presented as median (Q1,Q3) 1 baseline FSH levels summarized using complete data BMI Body mass index, FSH Follicle stimulating hormone, AMH Anti-Müllerian hormone, DOR Diminished ovarian reserve, GnRH Gonadotropin-releasing hormone, GnRHa Gonadotropin-releasing hormone agonist, HRT Hormone replacement treatment, NC Natural cycle, PPOS Progestin-primed ovarian stimulation, ICSI Intracytoplasmic sperm injection, IVF  In vitro fertilization, FET Frozen embryo transfer, GA Gestational age, GDM Gestational diabetes mellitus, HDP Hypertensive disorders of pregnancy, LBW Low birth weight, PTB Preterm birth Regarding obstetric complications (Table  2 ), the proportion of GDM appeared similar across all EMT groups (ranging from 3.4% to 4.2%). Conversely, HDP peaked at 5.1% in the ≤ 7 mm group, declining modestly as EMT increased. The rates of other perinatal outcomes, including placenta previa and fetal malformations, appeared comparable across all EMT groups. To further address the potential bias introduced by repeated embryo transfer cycles from the same individual, we fitted GLMMs with a random intercept for each patient. Both weighted and unweighted GLMMs were constructed. Before outcome evaluation, covariate balance was assessed. As detailed in Table S1, adequate balance was achieved across all EMT groups following IPTW, with all maximum absolute SMDs falling below 0.1. Adequate overlap of the stabilized IPTW weights was demonstrated, as shown in Fig.  1 F. Results from the IPTW-weighted GLMMs are presented in Table  3 . Table 3 Adjusted associations of endometrial thickness with reproductive outcomes from stabilized IPTW-weighted generalized linear mixed-effects models Endometrial Thickness Group (mm) Outcome ≤ 7 (7,8] (9,10] > 10 LB 0.428*** [0.353, 0.519] 1.011 1.063 1.254*** [1.108, 1.420] [0.890, 1.148] [0.975, 1.159] Miscarriage 1.660*** 0.924 1.026 0.875 [1.231, 2.238] [0.750, 1.138] [0.896, 1.174] [0.758, 1.009] Birth weight -0.036 -0.008 0.020 0.007 [-0.094, 0.022] [-0.043, 0.027] [-0.003, 0.043] [-0.016, 0.031] GDM 0.841 1.000 1.047 1.156 [0.386, 1.834] [0.619, 1.616] [0.761, 1.442] [0.844, 1.585] HDP 1.457 1.073 0.959 1.036 [0.725, 2.928] [0.651, 1.769] [0.680, 1.354] [0.736, 1.458] Macrosomia 0.616 0.874 1.005 0.965 [0.286, 1.326] [0.579, 1.318] [0.775, 1.303] [0.742, 1.254] LBW 1.334 1.391 0.712 0.977 [0.639, 2.784] [0.798, 2.424] [0.454, 1.118] [0.633, 1.509] PTB 1.919* 1.196 1.085 0.941 [1.075, 3.426] [0.767, 1.866] [0.795, 1.479] [0.683, 1.297] GA -0.505*** -0.120 -0.042 0.022 [-0.743, -0.266] [-0.260, 0.020] [-0.134, 0.050] [-0.073, 0.117] Values represent adjusted odds ratios (aORs) for binary outcomes and regression coefficients (β) for continuous outcomes Reference group: Endometrial Thickness of (8,9] mm Models were adjusted for the following covariates – LB and miscarriage: AMH level, infertility etiology, infertility diagnosis, endometrial preparation protocols, maternal age at oocyte retrieval, BMI, interval between oocyte retrieval and embryo transfer, and embryo of good quality – GA and PTB: Cesarean section (C-section) status, embryo of good quality, maternal age at oocyte retrieval, BMI, interval between oocyte retrieval and embryo transfer, and endometrial preparation protocols – HDP and GDM: Maternal age at oocyte retrieval, BMI, infertility etiology, endometrial preparation protocols, ovarian stimulation protocols, and the interval between FET and IVF/ICSI – Birth weight, LBW, and macrosomia: C-section status, embryo of good quality, maternal age at oocyte retrieval, BMI, gestational age at delivery, interval between oocyte retrieval and embryo transfer, and endometrial preparation protocols IPTW Inverse probability treatment weighting, LB Live birth, GDM Gestational diabetes mellitus, HDP Hypertensive disorders of pregnancy, LBW Low birth weight, PTB Preterm birth, GA Gestational age * p < 0.05, ** p < 0.01, *** p < 0.001 Adjusted associations of endometrial thickness with reproductive outcomes from stabilized IPTW-weighted generalized linear mixed-effects models 0.428*** [0.353, 0.519] 1.254*** [1.108, 1.420] Values represent adjusted odds ratios (aORs) for binary outcomes and regression coefficients (β) for continuous outcomes Reference group: Endometrial Thickness of (8,9] mm Models were adjusted for the following covariates – LB and miscarriage: AMH level, infertility etiology, infertility diagnosis, endometrial preparation protocols, maternal age at oocyte retrieval, BMI, interval between oocyte retrieval and embryo transfer, and embryo of good quality – GA and PTB: Cesarean section (C-section) status, embryo of good quality, maternal age at oocyte retrieval, BMI, interval between oocyte retrieval and embryo transfer, and endometrial preparation protocols – HDP and GDM: Maternal age at oocyte retrieval, BMI, infertility etiology, endometrial preparation protocols, ovarian stimulation protocols, and the interval between FET and IVF/ICSI – Birth weight, LBW, and macrosomia: C-section status, embryo of good quality, maternal age at oocyte retrieval, BMI, gestational age at delivery, interval between oocyte retrieval and embryo transfer, and endometrial preparation protocols IPTW Inverse probability treatment weighting, LB Live birth, GDM Gestational diabetes mellitus, HDP Hypertensive disorders of pregnancy, LBW Low birth weight, PTB Preterm birth, GA Gestational age * p < 0.05, ** p < 0.01, *** p < 0.001 Compared to the (8,9] mm group, women with an EMT ≤ 7 mm had significantly lower odds of LB (aOR: 0.428, 95% CI: 0.353–0.519, p  < 0.001), higher odds of miscarriage (aOR: 1.660, 95% CI: 1.231–2.238, p  < 0.001), higher odds of PTB (aOR: 1.919, 95% CI: 1.075–3.426, p  < 0.05), and reduced GA (adjusted β: −0.505 weeks, 95% CI: −0.743 to − 0.266, p   10 mm, the odds of LB were significantly higher than the reference group (aOR: 1.254, 95% CI: 1.108–1.420, p  < 0.001), with no significant increase in any adverse perinatal outcomes. In the unweighted GLMMs (Table S2), similar patterns were observed. An EMT ≤ 7 mm was associated with lower odds of LB, higher odds of miscarriage and PTB, and shorter GA, whereas an EMT > 10 mm was associated with higher odds of LB. Women with an EMT > 10 mm also had a modestly lower risk of miscarriage (OR: 0.859, 95% CI: 0.744–0.991, p  < 0.05). Associations for other perinatal outcomes were largely unchanged across EMT groups, consistent with the weighted models. Significant interaction terms were identified for miscarriage by maternal age at oocyte retrieval (Table S3) and for birth weight by maternal BMI (Table S4), whereas no significant interaction was detected for other outcomes. For miscarriage, a statistically significant interaction between an EMT of (9,10] mm and maternal age was detected (interaction p  < 0.01). Compared with the reference category of EMT (8,9] mm, an EMT of (9,10] mm was not associated with miscarriage among women aged  0.05), whereas higher odds of miscarriage were observed among women aged ≥ 35 years (aOR = 1.47, 95% CI 1.12–1.94, p   10 mm among women with overweight/obesity (interaction p  < 0.05) (Table S4). Specifically, among women with normal weight, EMT showed no association with birth weight, with effect sizes close to null for both the (9,10] mm (adjusted β = 0.005 kg, 95% CI: -0.023 to 0.032, p  > 0.05) and > 10 mm (adjusted β = -0.016 kg, 95% CI: -0.045 to 0.012, p  > 0.05) categories. However, among women with underweight, an EMT of (9,10] mm was associated with a higher birth weight relative to the reference EMT category of (8,9] mm (adjusted β = 0.087 kg, 95% CI: 0.019 to 0.154, p   10 mm among women with underweight (adjusted β = 0.051 kg, 95% CI: -0.019 to 0.120, p  < 0.1), although the corresponding interaction term was not statistically significant. Among women with overweight/obesity, there was no evidence of a difference in birth weight between EMT (9,10] mm and the reference category of EMT (8,9] mm (adjusted β = 0.027 kg, 95% CI: -0.023 to 0.076, p  > 0.05). In contrast, an EMT > 10 mm was associated with higher birth weight (adjusted β = 0.062 kg, 95% CI: 0.012 to 0.111, p  < 0.01).

This retrospective cohort study analyzed 16,566 single blastocyst FET cycles from 12,800 couples (Figure S1) conducted at the Reproductive Medical Center of Tongji Hospital. Oocyte aspiration was performed from 2015 to September 2022, and FET was performed from 2016 to September 2022. Cycles utilizing three common endometrial preparation protocols were included: hormone replacement treatment (HRT), natural cycles (NC), and HRT with gonadotropin-releasing hormone agonist (GnRHa) pretreatment (GnRHa + HRT). Inclusion required fertilization via IVF, ICSI, or rescue ICSI. Patients were excluded for missing EMT or any other variables used in the analyses, untreated uterine factors (uterine malformations, intrauterine adhesions, uterine fibroids, endometrial tuberculosis, and endometrial polyps), PGT use, or canceled transfers. Perinatal outcome analyses were restricted to 7,366 cycles yielding singleton LBs. Approval for this study was obtained from the Institutional Review Board of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology (No. TJ-IRB202404020). Briefly, NC were monitored via transvaginal ultrasound and serum hormones, with luteal support initiated one day post-ovulation. In HRT cycles, oral estradiol was administered from menstrual day 2 and adjusted according to endometrial growth, followed by progesterone supplementation once adequate proliferation was achieved. The GnRHa + HRT approach included a prior depot injection of leuprorelin 28 days before commencing the standard HRT regimen [ 21 ]. In instances where the EMT was less than 8 mm and failed to reach the measurement recorded on the hCG day of the corresponding fresh cycle, cancellation of the embryo transfer may be considered based on the patient’s condition and the physician’s recommendation [ 22 ]. Embryo culture, vitrification, and warming procedures were performed as previously published [ 23 , 24 ]. Morphological evaluation of embryos was conducted preceding vitrification. Good quality status was attributed to blastocysts that achieved grading profiles of AA, AB, BA, or BB, as determined by the Gardner scoring system [ 25 ]. EMT was defined by the final ultrasound measurement obtained before embryo transfer. Specifically, the EMT recorded on the day of ovulation was utilized for patients undergoing NC, whereas the EMT documented upon progesterone initiation was applied for those receiving HRT or GnRHa + HRT regimens [ 21 ]. Baseline endocrine profiles, including FSH and AMH levels, were measured on day 3 of the natural menstrual cycle using enzyme-linked immunosorbent assay kits (before October 2019) or magnetic particle-based chemiluminescence immunoassays (after October 2019) in our hospital’s central laboratory [ 26 ]. Regarding embryological parameters, the maturation rate was defined as the fraction of metaphase II (MII) oocytes out of the total oocyte yield. The normal fertilization (defined as 2PN) rate was determined relative to the total number of oocytes retrieved in IVF, or relative to the number of mature MII oocytes in ICSI. The definitions of pregnancy and perinatal outcomes were adopted from a previously established high-quality clinical trial in reproductive medicine [ 27 ]. A live birth (LB) was defined as the delivery of an infant at ≥ 28 gestational weeks exhibiting vital signs. CP was determined by the presence of at least one intrauterine gestational sac with detectable fetal heart activity during ultrasound examination. Biochemical pregnancy loss was identified when an initially positive hCG pregnancy test was followed by a subsequent decline, without any ultrasonographic evidence of a gestational sac. Miscarriage was defined as the spontaneous loss of an established CP occurring before 28 weeks of gestation. Ectopic pregnancy was defined as a pregnancy that developed outside the uterine cavity. Records of all perinatal outcomes were compiled through post-delivery telephone consultations by trained follow-up staff. The variables documented included basic neonatal indices, namely fetal sex, GA, and birth weight, as well as the occurrence of abnormal perinatal outcomes such as hypertensive disorders of pregnancy (HDP), gestational diabetes mellitus (GDM), placenta previa, and fetal malformation. Low birth weight (LBW) was defined as a birth weight of less than 2500 g, while macrosomia was defined as a birth weight exceeding 4000 g. Preterm birth (PTB) was defined as an LB occurring before 37 weeks of gestation. The primary outcomes were the LB, birth weight, GA, LBW, macrosomia, and PTB. The secondary outcomes included miscarriage, HDP, and GDM. EMT was categorized into five groups: ≤7 mm, (7,8] mm, (8,9] mm, (9,10] mm, and > 10 mm, based on both clinical relevance and the observed distribution in our study population. For the descriptive analysis, baseline characteristics and outcome variables of the study population were summarized across these EMT groups. Continuous variables were presented as medians and interquartile ranges (IQRs), while categorical variables were summarized as counts and percentages. To evaluate the potentially non-linear associations of EMT with LB and GA, generalized additive models (GAMs) incorporating thin plate regression splines were fitted after adjusting for potential confounders. Furthermore, we employed generalized linear mixed-effects models (GLMMs) with inverse probability of treatment weighting (IPTW) [ 28 ]. A random intercept was included to account for multiple embryo transfer cycles per patient. A logistic link function was specified for binary outcomes, and an identity link function was used for continuous outcomes (birth weight and GA). Propensity scores for the multiple EMT categories were estimated using a multinomial logistic regression model specifically implemented with the multinom() function from the nnet package in R. The EMT group was treated as the dependent variable, and clinically relevant baseline characteristics, including maternal age at oocyte retrieval, BMI, infertility etiology, endometrial preparation protocol, AMH level, fertilization method, and embryo of good quality, were included as predictors. IPTW weights were then calculated as the inverse of the estimated propensity score. Stabilized weights were derived by multiplying the IPTW weights by the marginal probability of each EMT group to improve the efficiency and robustness of the estimates. To limit the influence of extreme values, predicted probabilities were truncated at the 1st and 99th percentiles prior to weight calculation. Standardized mean differences (SMDs) were calculated for each covariate across all pairwise EMT group comparisons before and after applying weighting. The maximum absolute SMD was used to summarize overall balance. SMDs were calculated using the cobalt package [ 29 ] in R, with an absolute SMD < 0.1 considered indicative of adequate balance. Propensity score distributions and weight diagnostics were also evaluated graphically using density plots to assess the overlap between EMT groups. For each GLMM, the EMT category was included as the main exposure with the EMT group of (8,9] mm serving as the reference category. Stabilized IPTW weights were applied to account for confounding. Additional covariates were included as fixed effects to account for residual confounding, implementing a doubly robust estimation approach. Covariates included in GLMMs were selected a priori based on clinical relevance and prior literature and included maternal age at oocyte retrieval, BMI, AMH level, infertility diagnosis, infertility etiology, embryo of good quality, endometrial preparation protocol, and the interval between oocyte retrieval and embryo transfer. Specific covariates were tailored for each outcome depending on the clinical context. A sensitivity analysis using an unweighted GLMM (without IPTW) was conducted to examine the robustness of the primary results. We further assessed potential effect modification of maternal age at oocyte retrieval and maternal BMI on the associations between EMT and pregnancy or perinatal outcomes. Maternal age was dichotomized as < 35 years and ≥ 35 years [ 30 ]. Maternal BMI was classified as underweight (< 18.5 kg/m²), normal weight (18.5–23.9 kg/m², reference), or overweight/obesity (≥ 24.0 kg/m²) [ 31 ]. Interaction terms were tested for all prespecified pregnancy and perinatal outcomes. For each outcome, two IPTW‑weighted GLMMs were fit separately: one including an interaction term between EMT categories and maternal age categories, and the other including an interaction term between EMT categories and maternal BMI categories. Effect estimates were derived from these models, with reference groups defined as women aged < 35 years with an EMT of (8,9] mm and normal-weight women with an EMT of (8,9] mm, respectively. In our primary analysis, adjusted odds ratios (aORs) or adjusted β coefficients, along with 95% confidence intervals (CIs), were reported as appropriate for binary or continuous outcomes. All analyses were conducted using R (version 4.3.1), and statistical significance was defined as a two-sided p-value < 0.05.

In this large-scale retrospective study of 16,566 FET cycles and 7,366 singleton live births, an EMT of 7 mm emerged as a potential common reference point for evaluating both pregnancy and perinatal outcomes. Specifically, an EMT below this threshold was associated with lower odds of LB, increased odds of miscarriage, shorter GA, and higher odds of PTB. Conversely, an EMT approaching 10 mm was associated with higher odds of LB. Furthermore, maternal age and BMI modified the associations between EMT and miscarriage as well as birth weight, respectively. In concordance with two prior studies [ 7 , 11 ], we found that an EMT < 7 mm was associated with significantly poorer pregnancy outcomes in FET cycles. Specifically, an EMT below this cutoff was associated with 57% lower odds of LB and 66% higher odds of miscarriage. Nevertheless, our data indicate that patients with an EMT > 7 mm still achieved an LB rate of 24.6%, and LBs remained attainable even among those with an EMT < 4 mm. This aligns with research by Ata et al. [ 32 ], which demonstrated that LB could be obtained even with an EMT as low as 3 mm. Consequently, our findings suggested that a thin endometrium alone should not necessarily be regarded as an indication for cancellation of an embryo transfer. Research exploring obstetric and neonatal outcomes remains relatively limited [ 16 – 20 , 33 – 35 ]. Regarding FET cycles in particular, few studies have evaluated the effects of an EMT below 8 mm, attributed to restricted sample sizes. Hu et al. [ 19 ] analyzed 5,220 cleavage-stage FET cycles resulting in singleton LBs and found that an EMT < 8 mm was associated with an elevated risk of PTB. In the present study, an EMT ≤ 7 mm was associated with shorter GA and higher odds of PTB, while no significant associations with adverse perinatal outcomes were observed for EMT values of 7–8 mm. Although previous research has connected an EMT < 8 mm with an increased risk of LBW or reduced birth weight, the findings may be partially confounded by unadjusted GA [ 19 ] or limited sample sizes within thin EMT cohorts [ 33 ]. After controlling for GA, our study found no significant association between EMT and the risk of LBW. Notably, while several studies [ 17 , 34 , 35 ] reported that an EMT of either 12 mm was associated with increased odds of HDP, our findings demonstrated a similar but non-significant pattern. This lack of significance may stem from the risk of recall bias inherent in self-reported HDP data; therefore, these results should be interpreted with caution. The biological mechanisms through which a thin endometrium leads to adverse clinical outcomes remain a subject of ongoing debate. EMT is a clinically accessible phenotypic marker of uterine status; its thinness potentially mirrors compromised deep biological processes, including implantation, receptivity, decidualization, angiogenesis, and early placentation [ 36 – 39 ]. During the early stages of implantation, thin endometrium exhibits significant transcriptomic aberrations characterized by the abnormal activation of pro-inflammatory microenvironments and imbalances in oxidative stress [ 36 ], both of which could potentially interfere with the implantation process. Furthermore, disruptions in internal miRNA-mRNA regulatory networks, such as the aberrant expression of the miR-200 family, alongside the dysregulation of core signaling pathways including FoxO, PI3K-Akt, p53, and Wnt, have been observed, which potentially exert an inhibitory effect on cellular proliferation and angiogenesis [ 36 , 40 ]. These factors may lead to a compromised state of endometrial receptivity, possibly contributing to the observed decline in LB rates and a concomitant increase in miscarriage risk. Furthermore, shortened GA and the increased risk of PTB may be attributable to placental and fetal maldevelopment. It has been proposed that embryos implanting in a thin endometrium may reside in closer proximity to the basal layer and spiral arteries, where locally elevated oxygen tension may be detrimental to embryonic development [ 41 – 43 ]. An alternative mechanism may involve disturbances in the uteroplacental circulation. The remodeling of maternal spiral arteries to establish a low-resistance uteroplacental circulation is essential for normal fetal development [ 44 ]. A thin endometrium is characterized by vascular dysplasia and high uterine blood flow impedance [ 45 ]. It is hypothesized that such vascular anomalies might impede the crucial spiral artery remodeling process, thereby potentially contributing to adverse perinatal outcomes [ 16 , 46 ]. Notably, Hadas et al. [ 38 ] demonstrated that anatomical changes and histological malperfusion lesions were present within a thin endometrium. Moreover, the cesarean section rate in our cohort was relatively high, reflecting the overall high prevalence of cesarean delivery in China [ 47 – 49 ], which is further elevated among ART-conceived pregnancies [ 50 , 51 ]. While cesarean section does not alter the clinical classification of preterm birth, a portion of the observed variance in GA may be attributable to iatrogenic preterm deliveries. The impact of an excessively thick endometrium on clinical outcomes was not fully explored in this study. In our GAM model, the adjusted partial effect of EMT on LB appeared to rise up to 12 mm, followed by a decline; however, the GLMM analyses were confined to an EMT > 10 mm due to sample size limitations. Although the clinical significance of a thick endometrium was not the primary focus of this study, it is important to highlight its potential pathological implications. A limited number of previous reports suggested that an excessively thick EMT might be associated with adverse clinical outcomes. For instance, two studies reported a decrease in CP rates when EMT exceeded 14 mm [ 52 , 53 ]. Furthermore, Rombauts et al. observed that patients with an EMT > 12 mm had nearly a threefold increased risk of placenta previa compared to those with EMT < 9 mm [ 54 ]. Given the paucity of relevant research, future studies with larger sample sizes in cohorts with thick endometrium are warranted to further elucidate these potential impacts. The modifying effects of age and BMI on the relationship between EMT and reproductive outcomes in FET remain underexplored. Interestingly, we observed that an EMT of (9,10] mm was associated with increased odds of miscarriage in women aged ≥ 35 years, while no significant decrease in LB was observed. We hypothesize that this phenomenon arises from a synergistic interaction between age-related endometrial dysfunction and embryonic abnormalities. With advancing maternal age, the incidence of embryonic chromosomal anomalies driven by oocyte meiotic errors rises significantly [ 55 ]. A physiologically normal endometrium can screen and prevent the implantation of abnormal embryos during the implantation phase via functional decidualization and implantation checkpoints [ 55 ]. However, evidence indicates that the human endometrium undergoes significant transcriptomic alterations starting at age 35, characterized by the abnormal upregulation of immune-inflammatory responses and the accumulation of senescent cells [ 56 , 57 ]. Consequently, we hypothesize that age-related deficiencies may impair embryonic screening in the endometrium of women aged ≥ 35 years, potentially creating a more permissive environment for the implantation of aneuploid embryos. In this context, a relatively thick endometrium, which is potentially more receptive, might facilitate such abnormal implantation events, thereby contributing to an increased incidence of clinically recognized miscarriages. Simultaneously, the overly permissive endometrial environment appears not to compromise the implantation or developmental potential of high-quality euploid embryos, which may explain why overall LB rates are maintained despite the increased risk of miscarriage. Furthermore, the distribution of miscarriage events within our study cohort might be confounded by other occult genetic factors. The proportion of male factor infertility rose with increasing EMT (ranging from approximately 26% to 41%), and such paternal abnormalities might further predispose to embryonic genetic instability. Moreover, given the retrospective nature of this study, we could not exclude potential parental chromosomal structural abnormalities, which serve as a critical risk factor for early pregnancy loss. Future investigations incorporating parental karyotyping and PGT data are required to definitively decouple the respective roles of endometrial function and embryonic genetic status in pregnancy maintenance among women of advanced reproductive age. We also observed a statistically significant, albeit clinically modest, positive interaction between a thick endometrium and BMI categories. After adjusting for confounding factors, including GA, an EMT > 10 mm in women with overweight/obesity was associated with an approximate 62 g increase in fetal birth weight when compared to an EMT of (8,9] mm. Mechanistically, this may be attributed to the influence of maternal-fetal exchange. In women with overweight/obesity, whose intrauterine environment is often marked by hyperglycemia, hyperlipidemia, or hyperinsulinemia, a thick EMT may enhance placental vascularization and the transplacental nutrient flow. Based on our data, however, this magnitude of change might not manifest as pathological macrosomia at the individual level. Clinically, the findings of this study may serve as a valuable reference for individualized decision-making during FET cycles. Although an EMT ≤ 7 mm is not an absolute barrier to achieving pregnancy, the decreased odds of LB and the elevated risk of PTB indicate that optimizing EMT remains a rational initial therapeutic objective. Nevertheless, in the absence of other definitive contraindications, it may be inappropriate to routinely cancel embryo transfer based solely on the isolated morphological parameter of a thin endometrium. For patients whose endometrium fails to thicken further despite clinical intervention, proceeding with the transfer may still be considered following a comprehensive evaluation of clinical indices, thorough risk counseling, and careful alignment with patient preferences. However, it must be emphasized that if a pregnancy is successfully established, these patients will likely require more rigorous prenatal surveillance and prophylactic interventions against PTB. For women aged ≥ 35 years, adopting a more cautious, stratified management strategy may yield clinical benefits. In women of advanced reproductive age with a thick endometrium, the overall prospects for LB do not seem to be markedly compromised; however, given a possible predisposition to miscarriage, enhanced first-trimester monitoring, alongside careful expectation management, may be clinically beneficial. Because embryo ploidy was not assessed in our dataset, the exact mechanisms underlying the observed miscarriage patterns remain to be elucidated. Consequently, investigating whether preimplantation genetic testing for aneuploidy (PGT-A) could offer clinical utility for this specific demographic represents an interesting avenue for future research. The primary strengths of this study lie in its utilization of a large-scale, single-center cohort, which ensures standardized clinical protocols and comprehensive follow-up data. Furthermore, we employed rigorous statistical approaches, including GLMMs and IPTW, to effectively control for clustering and confounding. Unlike prior research that primarily focused on pregnancy outcomes, our study concurrently analyzed obstetric and neonatal outcomes to comprehensively elucidate the clinical value of EMT. Additionally, we investigated the modifying effects of maternal age and BMI on the association between EMT and clinical outcomes, thereby providing evidence for the individualized treatment strategies in clinical practice. Several limitations should be acknowledged. First, given its retrospective design, we cannot establish a definitive causal relationship between EMT and clinical outcomes. The cohort remains highly heterogeneous, and some patients contributed repeated cycles spanning across EMT groups. While IPTW-weighted GLMMs with random intercepts were used to account for these complexities, unmeasured residual confounding persists. For example, embryo ploidy serves as a central factor influencing implantation, miscarriage, LB, and distal perinatal outcomes; however, relevant data were not available in this study. Consequently, the observed associations between EMT and clinical outcomes may partially reflect unmeasured embryonic chromosomal abnormalities, including aneuploidy, mosaicism, and possible polyploidy. Additionally, our cohort did not comprehensively capture confounders such as patient lifestyle factors (e.g., smoking, alcohol intake) [ 58 ], underlying metabolic disorders [ 59 ], use of supplements [ 60 ], ongoing treatment adjustments, and individual clinical decision-making. Consequently, we cannot rule out the alternative explanation that the observed adverse obstetric outcomes were, to some extent, driven by occult maternal comorbidities or non-EMT-mediated primary placental dysfunction. Future prospective multicenter cohort studies incorporating comprehensive multidimensional covariates alongside PGT-A data are required to corroborate our findings. Additionally, specific clinical decision-making processes might have introduced selection bias within the groups with an EMT < 8 mm. At our center, when the endometrium is deemed excessively thin, interventions such as prolonging the cycle or administering supplemental estrogen are often implemented, enabling some patients to ultimately achieve a thicker endometrium. For patients whose endometrium remains refractory to further proliferation, clinicians determine whether to proceed with or cancel the transfer based on the overall assessment and patient preference. This practice pattern may inadvertently select a specific subpopulation with a persistently thin endometrium, concurrently resulting in a smaller sample size for this group. Such selection bias is a common challenge in research regarding EMT [ 61 ] and must be carefully weighed when interpreting our findings. Moreover, data on perinatal outcomes were derived from self-reported telephone follow-ups. This reliance introduces a risk of recall bias, warranting caution when interpreting these outcomes. Finally, although GAMs were used for visualization, our primary analysis relied on artificially defined EMT categories. This categorization may reduce statistical power and obscure underlying continuous or non-linear associations.

Although the precise risk of thin endometrium remains to be carefully evaluated due to the truncation effect in clinical data, the existing evidence supports that an EMT ≤ 7 mm is associated with lower odds of LB and increased odds of PTB, while an EMT reaching 10 mm is associated with higher odds of LB in FET cycles. The clinical impact of EMT seems to be modulated by maternal age and BMI, indicating the need for personalized treatment protocols. In clinical practice, a thin endometrium alone may not be considered an absolute indication for cancelling an embryo transfer; however, the associated risks warrant careful consideration, and enhanced perinatal monitoring may be beneficial. The findings, particularly those regarding obstetric outcomes and subgroup interactions, should be validated through future large-scale prospective studies.

Successful embryo implantation relies on intricate synchrony and cross-talk between a developing embryo and a receptive endometrium during a critical and brief period known as the window of implantation [ 1 ]. In assisted reproductive technology (ART), optimal endometrial receptivity remains one of the primary factors for achieving a clinical pregnancy (CP). Endometrial thickness (EMT) currently serves as a widely used clinical indicator for assessing endometrial receptivity before embryo transfer [ 2 ]. Despite extensive research on this parameter, its association with clinical outcomes remains controversial [ 3 ], especially in frozen embryo transfer (FET) cycles. Numerous studies have indicated that an endometrium thinner than 7 mm [ 4 – 7 ] or 8 mm [ 8 – 10 ] compromises pregnancy outcomes. An extensive Canadian cohort involving 96,000 fresh and frozen cycles reported a substantial decline in LB rate when the EMT fell below 6 mm [ 11 ]. Beyond merely defining the “thin” endometrium, identifying the optimal EMT to maximize reproductive success has yielded heterogeneous findings. Some studies suggested a lower optimal threshold of 8 mm [ 12 , 13 ], while the upper limits ranged from 13 [ 13 ] to 14 mm [ 14 ]. Data from the aforementioned large-scale Canadian study demonstrated that pregnancy rates in FET cycles plateaued after 7–10 mm [ 11 ]. Moreover, a comprehensive study comprising 244,001 cycles demonstrated that the LB rate plateaued at an EMT of 9 mm [ 15 ]. Conversely, a large-scale study by Julian A et al. involving preimplantation genetic testing (PGT) cycles reported that the LB rate continued to increase up to an EMT of 12 mm [ 9 ]. With improvements in ART, reproductive medicine has shifted focus from implantation alone to broader maternal and neonatal outcomes. However, most prior EMT studies focus on pregnancy endpoints, with limited exploration of obstetric and perinatal risks. Emerging evidence suggests that EMT may be associated with obstetric complications [ 16 , 17 ] and adverse fetal outcomes [ 18 – 20 ], such as small-for-gestational-age [ 18 , 20 ] and hypertensive disorders [ 17 , 20 ]. Despite this, no consensus exists on the optimal EMT range for guiding clinical decisions. Further investigation is warranted to clarify the role of EMT in both pregnancy and perinatal outcomes. In this study, we examined 16,566 single blastocyst FET cycles to evaluate the associations between EMT and a range of clinical outcomes, including LB, gestational age (GA), birth weight, and perinatal outcomes. We also explored potential effect modification by maternal age and BMI. Through these analyses, we aim to provide comprehensive and clinically relevant evidence to inform embryo transfer decisions and improve maternal-fetal health.

Supplementary Material 1. Supplementary Material 1.