Joint associations of endometrial thickness and pattern with live birth after fresh and frozen-thawed embryo transfer: a retrospective cohort study of 45,729 cycles.

OBJECTIVE

To investigate how endometrial thickness and pattern jointly relate to live birth in fresh and frozen-thawed embryo transfer cycles.

DESIGN

Retrospective cohort study.

SUBJECTS: A total of 45,729 embryo transfer cycles performed between January 2018 and August 2022, including 16,566 fresh and 29,163 frozen-thawed cycles.

EXPOSURE: Endometrial thickness and pattern were assessed by transvaginal ultrasound on the trigger day in fresh cycles and on the day of progesterone initiation in frozen-thawed cycles. Fresh and frozen-thawed cycles were analyzed separately.

MAIN OUTCOME MEASURES: Live birth per embryo transfer cycle.

RESULTS

In both fresh and frozen-thawed cycles, restricted cubic spline analyses demonstrated significant overall and nonlinear associations between endometrial thickness and live birth (all P<0.001), while non-A morphology was associated with a lower adjusted live birth risk than type A morphology in no-interaction modified Poisson models (fresh: aRR, 0.880; 95% CI, 0.832-0.930; frozen-thawed: aRR, 0.892; 95% CI, 0.863-0.923). In fresh cycles, type B/C was associated with lower live birth risk than type A within the 7.0-7.9 mm and 8.0-8.9 mm strata. In frozen-thawed cycles, live birth rates increased from 16.7% in the <7 mm B/C group to 44.7% in the ≥14 mm A group, with a significant type A advantage within strata below 10 mm. In some frozen-cycle comparisons, thinner type A endometrium significantly outperformed the adjacent thicker B/C stratum, indicating that type A morphology was associated with higher live birth rates within thinner endometrial strata. Adjusted spline curves showed that live birth risk increased most rapidly at lower thicknesses, with the slope beginning to attenuate at approximately 11-12 mm in fresh cycles and 9-10 mm in frozen-thawed cycles. Sensitivity analyses reproduced the same overall pattern. Additionally, endometrial thickness and morphology varied with ovarian stimulation protocol and gonadotropin duration in fresh cycles, and with endometrial preparation regimen in frozen-thawed cycles.

CONCLUSION

Endometrial thickness showed a nonlinear association with live birth, and trilaminar morphology was associated with higher adjusted live birth rates in both fresh and frozen-thawed embryo transfer cycles. The prognostic contribution of endometrial pattern was most evident in thin-to-intermediate endometrium.