L26/P-518 The Endometrial Paradox: Distinct Obstetric Risk Profiles of Endometrial Thickness in Fresh, Artificial Frozen, and Natural Frozen Cycles

<jats:title>Abstract</jats:title> <jats:sec> <jats:title>Study question</jats:title> <jats:p>Does endometrial thickness impact on adverse obstetric outcomes (hypertension, low birth weight, prematurity) differ between fresh IVF, artificial FET (E+P), and natural cycle FET protocols?</jats:p> </jats:sec> <jats:sec> <jats:title>Summary answer</jats:title> <jats:p>Thin endometrium (&amp;lt;8mm) drives adverse outcomes in fresh cycles; thick (&amp;gt;12mm) predicts complications in artificial FETs. Natural FETs showed best safety despite thinnest linings.</jats:p> </jats:sec> <jats:sec> <jats:title>What is known already</jats:title> <jats:p>Endometrial thickness is a key predictor of implantation, yet its influence on downstream obstetric complications remains controversial. While thin endometrium is linked to placental dysfunction in fresh cycles, the “upper limit” of safety in frozen cycles is debated. Furthermore, the potential protective role of the corpus luteum (present in natural cycles/fresh, absent in artificial FETs) in mitigating these risks requires clarification.</jats:p> </jats:sec> <jats:sec> <jats:title>Study design, size, duration</jats:title> <jats:p>This retrospective cohort study analyzed 481 live births from a single center. The cohort included 255 Fresh cycles, 180 Frozen Artificial (E+P) cycles, and 46 Frozen Natural/Modified Natural (MNC) cycles. Cycles were stratified by EMT groups (&amp;lt;8mm, 8-12mm, &amp;gt;12mm) and outcomes were compared across protocols.</jats:p> </jats:sec> <jats:sec> <jats:title>Participants/materials, setting, methods</jats:title> <jats:p>Women undergoing IVF/ICSI who achieved a live birth were included. Outcomes measured included Hypertensive Disorders of Pregnancy (HDP/Preeclampsia), Low Birth Weight (LBW &amp;lt;2500g), and Prematurity (&amp;lt;37 weeks). Comparisons used t-tests and Fisher’s exact tests.</jats:p> </jats:sec> <jats:sec> <jats:title>Main results and the role of chance</jats:title> <jats:p>Risk profiles differed fundamentally by protocol:</jats:p> <jats:p>• Fresh Cycles: “Thin is Bad.” An EMT &amp;lt;8mm was the primary driver of complications, significantly increasing the risk of HDP (13.6%), LBW (45.5%), and Prematurity (50.0%) compared to thicker linings (p &amp;lt; 0.05).</jats:p> <jats:p>• Frozen Artificial (E+P): “Thick is Bad.” Excessive thickness (&amp;gt;12mm) was associated with drastically higher rates of LBW (55.0% vs 14.9%) and Prematurity (60.0% vs 17.2%) compared to normal linings (p &amp;lt; 0.001), though notably protective against HDP (0%).</jats:p> <jats:p>• Frozen Natural/MNC: “The Natural Protective Effect.” Despite having significantly thinner endometrium (Mean: 9.6mm) compared to Fresh (10.4mm, p = 0.04) and Artificial cycles (10.1mm), Natural Frozen cycles achieved the best obstetric outcomes of all groups. They demonstrated the lowest rates of Prematurity (8.7% vs 23.3% in Artificial, p = 0.038; vs 28.2% in Fresh, p = 0.005) and LBW (10.9%).</jats:p> </jats:sec> <jats:sec> <jats:title>Limitations, reasons for caution</jats:title> <jats:p>Retrospective analysis with smaller Natural Frozen cohort (n = 46). Infertility etiology balanced (p = 0.11), but Artificial group BMI higher (27.5 vs 25.6 Natural, p = 0.05). Artificial cycles’ distinct profile (high LBW, 0% HDP) suggests protocol—not BMI—primarily drives obstetric outcomes.</jats:p> </jats:sec> <jats:sec> <jats:title>Wider implications of the findings</jats:title> <jats:p>Hormonal environment dictates optimal endometrial window. Fresh cycles: thin (&amp;lt;8mm) linings predict insufficiency. Artificial FETs: thick (&amp;gt;12mm) linings signal exogenous dysregulation. Natural FETs achieve superior safety despite thinnest endometrium, suggesting corpus luteum/physiological ovulation provides unique placental protection beyond absolute thickness.</jats:p> </jats:sec> <jats:sec> <jats:title>Trial registration number</jats:title> <jats:p>Yes</jats:p> </jats:sec>