Beyond the Scale: Effects of Maternal Obesity on Embryo Morphokinetics and IVF Outcomes.

Obesity affects over 40% of reproductive-aged women worldwide [ 1 ]. It is consistently linked to poorer reproductive outcomes, including ovulatory dysfunction and elevated risks of gestational diabetes (GDM), hypertensive disorders, and cesarean delivery [ 2 ]. In IVF, concerns persist regarding how maternal body mass index (BMI) impacts both ovarian response and embryo development. Findings regarding BMI and IVF success remain inconsistent. Some studies report lower oocyte yield and reduced live birth rates with elevated BMI [ 3 ], citing altered steroidogenesis and inflammation [ 4 ]. Conversely, others suggest success rates are comparable when high-quality embryos are transferred, implying age or endometrial factors may be more decisive [ 5 ]. Most studies rely on static embryo assessment, providing limited insight into early embryogenesis. Time-lapse incubator (TLI) systems allow for continuous monitoring of cleavage timing and blastocyst formation [ 6 ]. However, few studies have utilized TLI to examine how maternal BMI affects specific morphokinetics [ 7 ]. This study aimed to assess the association between maternal BMI and embryo morphokinetics and determine their correlation with clinical and obstetric outcomes. Primary outcomes included cleavage timing and embryo quality (KIDscore), while secondary outcomes focused on pregnancy, delivery, and complication rates.

This retrospective cohort study at Hillel Yaffe Medical Center (January 2018–December 2024) included fresh IVF cycles using autologous oocytes, TLI culture, and fresh embryo transfer (ET). We excluded severe male factor (few motile cells) and uterine abnormalities (septate, T-shaped, unicornuate, or markedly myomatous uteri), though mild-to-moderate male factor was included. To maximize power, fertility preservation cycles were included for oocyte yield analysis; however, pregnancy rates were calculated strictly per fresh ET to avoid bias from elective cryopreservation. The Institutional Review Board approved the study (0026-20-HYMC), and informed consent was waived. Patients were categorized by BMI: underweight (<18.5 kg/m 2 ), normal weight (18.5–24.99 kg/m 2 ), overweight (25–29.99 kg/m 2 ), and obese (≥30 kg/m 2 ) per CDC criteria [ 8 ]. All patients used tailored GnRH antagonist protocols [ 9 ]. Gonadotropin dosing (recombinant FSH and/or purified HMG) was individualized based on age, BMI, and ovarian reserve. Serum E2, progesterone, and LH were measured at every follow-up alongside transvaginal ultrasound (TVUS). Antagonists (0.25 mg Cetrorelix (Merck Serono, Darmstadt)/Ganirelix (Organon, Oss, The Netherlands)) were initiated when follicles reached ≥12 mm. Once leading follicles were ≥17 mm, ovulation was triggered with recombinant hCG and/or 0.2 mg GnRH agonist, followed by retrieval 36 h later and fertilization via IVF or ICSI [ 10 ]. ET occurred on Day 3 or 5. Luteal support included oral dydrogesterone combined with vaginal progesterone gel or tablets. ET decisions were based on oocyte yield, E2 levels (to mitigate OHSS risk), and embryo quality. Single ET was standard; two embryos were transferred only after >3 unsuccessful cycles. All embryos were cryopreserved if endometrial abnormalities were detected or if progesterone rose prematurely. Embryos were assessed via TLI (EmbryoViewer, Unisense FertiliTech, Aarhus, Denmark), which provided continuous image acquisition without disturbing culture conditions. Only 2PN embryos were included; for Day 3 transfers, data were censored at incubator removal, while later milestones were derived from embryos cultured until Day 5/6. Arrested embryos were analyzed until the point of arrest. Recorded timings included tPB2, tPNa, tPNf, and cleavage stages t2, t3, t4, t5, and t8. Cell cycle intervals (Cc2, Cc3), synchrony (s2, s3), and KIDscore, integrating morphokinetic and morphological data for implantation potential, were assessed. Morphokinetic analysis was performed at the individual oocyte level. Biochemical pregnancy was confirmed via serum β-hCG 12 days post-transfer, and clinical pregnancy via TVUS at 6 weeks. All pregnancies were followed until delivery to record maternal/fetal complications and mode of delivery. Clinical outcome analyses were performed at the cycle level, with one observation per fresh embryo transfer. Analysis was performed using SPSS version 29. Normality was assessed using the Shapiro–Wilk test. Continuous variables were compared using one-way ANOVA or the Kruskal–Wallis test, as appropriate, with post hoc pairwise comparisons adjusted by Bonferroni correction. Independent samples t -test and Mann–Whitney U test were used for comparisons between two groups. Categorical variables were analyzed using Pearson’s chi-square test or Fisher’s exact test. To account for the non-independence of multiple embryos originating from the same patient, morphokinetic parameters were analyzed using a linear mixed-effects model (LMM) with patient as a random effect. The fixed effects included BMI category, developmental timepoint, and their interaction. Pairwise comparisons between BMI groups at each timepoint were performed with Bonferroni correction. Given the large number of comparisons across BMI categories and multiple outcome variables, the Benjamini–Hochberg false discovery rate (FDR) procedure was applied to adjust for multiple testing. All reported p -values reflect FDR-adjusted values, with significance set at q < 0.05. Multivariable logistic regression was used to identify independent predictors of clinical pregnancy, adjusting for BMI category, maternal age, male factor infertility, endometrial thickness, and LH and progesterone levels at trigger. The authors used Gemini 3 (Google, California, USA, January 2026, as available online) to perform spellchecking and to improve the language, flow, and structural clarity of the manuscript. Following this, all authors manually reviewed, edited, and verified the generated text to ensure it accurately reflects the study’s data and findings. The authors take full responsibility for the final content and the integrity of the work.

A total of 2238 fresh IVF cycles were stratified by BMI: underweight ( n = 102), normal weight (n = 1001), overweight (n = 560), and obese (n = 575). Maternal age was comparable across groups ( p = 0.054). Infertility etiologies differed significantly: male factor ( p = 0.009), tubal/uterine factor ( p = 0.008), and anovulation ( p < 0.001) were significantly different. Notably, anovulation was absent in underweight women but reached 15.7% in the obese group. Endometrial thickness increased progressively with BMI, from 9.34 ± 2.43 mm in underweight to 10.05 ± 2.44 mm in obese patients ( p < 0.001) ( Table 1 ). A sub-analysis of fertility preservation cycles is presented in Supplementary Table S1 . These cycles showed similar trends in ovarian response relative to BMI. Importantly, inclusion of these patients primarily contributed to the analysis of oocyte yield. Oocyte yield ( p = 0.98), the number of fertilized oocytes ( p = 0.43), and fertilization rates (59–60%; p = 0.23) were similar across all BMI categories. In the fertility preservation subgroup, oocyte yield showed a trend toward higher retrieval in the normal weight group compared with the obese group, though this did not reach significance after FDR correction (q = 0.064). Peak estradiol levels on trigger day decreased significantly with increasing BMI, with median levels ranging from 1505 pg/mL in underweight to 627 pg/mL in obese patients (normal weight vs. obese, q < 0.001). The estradiol-to-oocyte ratio followed a similar significant decline (normal weight vs. obese q = 0.006; normal weight vs. overweight q = 0.037). KIDScore values were numerically highest in the underweight group (4.62 ± 0.79) and lowest in the normal weight group (4.29 ± 1.1), but the difference did not remain significant after FDR correction (q = 0.064). The distribution of cleavage-stage (Day 3) versus blastocyst-stage (Day 5) transfers was comparable across all study groups ( p = 0.13) ( Table 2 ). A linear mixed-effects model with patient as a random effect confirmed a significant BMI × developmental timepoint interaction ( p < 0.001), indicating that embryo developmental trajectories differ across BMI categories. All morphokinetic comparisons that were significant in the original analysis remained significant after Benjamini–Hochberg FDR correction. Early cleavage occurred sooner in underweight patients. tPNf was significantly shorter in the underweight group (23.8 ± 3.0 h) compared to the normal weight (P1 = 0.028) and obese groups (P3 = 0.027). Similarly, t2 was reached earlier in the underweight group (26.68 ± 3.52 h) than in the normal weight (P1 = 0.005) and obese (P3 = 0.035) cohorts. Normal weight embryos also reached t2 later than overweight embryos (P4 = 0.039). Acceleration in the underweight group persisted through t3 (37.1 ± 5.8 h vs. 38.4–38.6 h; P1 = 0.008, P3 = 0.022), where overweight embryos were also significantly faster than normal weight (P4 = 0.005). This trend reversed at the 5-cell stage (t5), with overweight patients reaching cleavage earlier (P4 = 0.019). By t8, embryos from overweight women were the fastest (58.8 ± 10.3 h), significantly outpacing the normal weight (P4 = 0.019) and obese (P6 = 0.036) cohorts. At the blastocyst stage (tB), development remained faster in overweight women (107.6 ± 8.89 h) compared to underweight patients (P2 = 0.019), indicating faster late-stage development in higher BMI groups ( Table 3 ). Figure 1 illustrates these deviations from the normal weight mean; negative Δ values indicate earlier development, while positive values indicate delay ( Figure 1 ). Rates of positive β-hCG, clinical, ongoing, and live birth were not significantly different across groups (clinical pregnancy: 33.3% underweight vs. 29.9% obese; p = 0.93) ( Table 2 ). Cesarean section (CS) rates differed significantly ( p < 0.001), with obese patients reaching 51.9% compared to 27.3% in underweight and 25.0% in normal weight women. Gestational diabetes (GDM) was more frequent in overweight (19.1%) and obese (19.4%) patients than normal weight (10.3%) and underweight (0%) women ( p = 0.060). Mean birthweight showed a non-significant increasing trend from 2734 ± 542 g in underweight to 3156 ± 664 g in obese patients ( Table 4 ) ( Figure 2 ). In order to predict clinical pregnancy, we compared baseline characteristics. We found that clinical pregnancy was associated with younger age (32.7 ± 5.8 vs. 35.4 ± 6.5 years; p < 0.001), anovulatory ( p = 0.013) or male factor infertility ( p = 0.038), and higher oocyte yield (10.0 ± 6.3 vs. 8.7 ± 6.0; p < 0.001). For live birth, BMI was similar between groups (26.7 ± 6.12 vs. 26.5 ± 6.22; p = 0.75), as was BMI distribution ( p = 0.62). Women with live birth were younger (32.3 ± 5.6 vs. 34.5 ± 6.32; p = 0.002), had slightly lower basal FSH ( p = 0.02), and fewer previous transfers (2.49 ± 2.45 vs. 3.22 ± 2.94; p = 0.028) ( Table 5 ). Based on the univariate, we conducted a multivariate logistic regression to predict clinical pregnancy. Independent predictors of clinical pregnancy were endometrial thickness (aOR 1.118, 95% CI: 1.018–1.229; p = 0.02) and serum progesterone at trigger (aOR 1.057, 95% CI: 1.038–1.077; p < 0.001). Adjusted for confounders, BMI was not significantly associated with pregnancy, and neither overweight (aOR 1.015; p = 0.962) nor obesity (aOR 1.432; p = 0.201) demonstrated significant impacts. For live birth, maternal age remained a significant independent predictor (aOR 0.925, 95% CI: 0.868–0.986; p = 0.016). BMI (aOR 1.024; p = 0.353) and transfer type (single/double) were not significant. Oocyte yield was positively associated with live birth (aOR 1.079, 95% CI: 1.013–1.150; p = 0.019) ( Table 6 ).

In this cohort of over 2200 fresh ET cycles, maternal BMI influenced hormonal profiles and embryo developmental kinetics but did not compromise pregnancy outcomes. Live birth and clinical pregnancy rates were comparable across BMI groups, suggesting preserved implantation potential in the fresh transfer setting. However, obese women had markedly higher cesarean rates, highlighting BMI as a determinant of perinatal course rather than conception success. The higher prevalence of anovulatory infertility with increasing BMI is consistent with obesity-mediated ovulatory dysfunction [ 11 ]. The thicker endometrium observed in higher BMI groups may reflect peripheral aromatization of androgens to estrone in adipose tissue [ 12 ], though it is worth noting that endometrial receptivity could be impaired by altered inflammatory and cytokine profiles, potentially offsetting this anatomical advantage [ 5 , 13 ]. Although oocyte yield was preserved across BMI groups, peak estradiol and estradiol-to-oocyte ratios decreased with increasing BMI, which may reflect reduced granulosa cell responsiveness and altered steroidogenesis [ 14 , 15 , 16 ]. Despite these hormonal disparities, pregnancy rates remained comparable, suggesting that competent embryos can be obtained even with suboptimal hormonal markers [ 17 ]. The relatively young and uniform age of our cohort likely contributed to this finding, as age remains the dominant determinant of IVF success [ 18 ]. A linear mixed-effects model confirmed a significant BMI interaction with developmental timepoint ( p < 0.001), and all individual pairwise comparisons remained significant after Benjamini–Hochberg FDR correction, supporting the robustness of the observed morphokinetic patterns. Morphokinetic analysis revealed a biphasic pattern ( Figure 1 ): embryos from underweight women reached early cleavage milestones faster (shorter tPNf, t2, t3), but this advantage reversed around the 4-cell stage. From t5 onward, overweight women’s embryos progressed through later stages (tM, tSB, tB) faster than those of normal weight and obese groups. It is well established that embryonic genome activation (EGA) occurs between the 4- and 8-cell stages, and that cleavage before this transition depends largely on oocyte quality [ 19 ]. One possible explanation for the early acceleration in underweight women is that their oocytes may be less susceptible to lipotoxic damage; studies in animal models have shown that oocytes from obese individuals display mitochondrial abnormalities and reduced membrane potential [ 20 ], and human data suggest lower mitochondrial DNA expression in the cumulus cells of overweight women [ 21 , 22 ]. However, we did not directly measure lipotoxicity markers or mitochondrial function, so this interpretation remains speculative. The post-EGA catch-up in overweight embryos is also intriguing. At the blastocyst stage, metabolic demand rises and fatty acid β-oxidation becomes an important ATP source [ 23 , 24 ]. Follicular fluid triglyceride levels have been reported to be highest in overweight women [ 25 ], which could theoretically provide greater lipid substrate for later-stage development. However, this hypothesis requires direct validation with paired metabolomic and morphokinetic data. Importantly, these phase-specific differences appear to converge by transfer time: underweight embryos’ early advantage is offset by post-EGA deceleration, while overweight embryos catch up. Laboratory selection of the best-performing embryo further compresses these kinetic differences. In our cohort, endometrial thickness—not BMI—was the independent predictor of clinical pregnancy. These findings are consistent with Weinerman et al. [ 26 ] and Rubio et al. [ 27 ], and align with the recent work of Younes et al. [ 28 ], who similarly reported that BMI-related differences in embryo development did not translate to impaired clinical outcomes. Taken together, these data suggest that once embryos reach blastocyst competence, the uterine environment may matter more than subtle developmental timing differences. The sharp rise in cesarean delivery among obese women is consistent with the literature linking a higher BMI to longer labor, macrosomia, and clinician preference for surgical delivery [ 29 ]. The trend toward higher GDM rates in overweight and obese groups (19.1% and 19.4% vs. 10.3% in normal weight; p = 0.06) aligns with established pathophysiology—including insulin resistance, chronic low-grade inflammation, and altered adipokine profiles—though our study may have been underpowered to detect a statistically significant difference given the relatively small number of deliveries per group. Similarly, the preeclampsia trend ( p = 0.08) is consistent with shared pathophysiological mechanisms between obesity and preeclampsia, including endothelial dysfunction and oxidative stress [ 30 ]. It is possible that obesity-related metabolic dysfunction compounds the already elevated metabolic demands of IVF pregnancies. Other perinatal outcomes did not differ significantly, which may reflect both sample size limitations and the effectiveness of antenatal surveillance in this cohort [ 31 , 32 ]. In our multivariable model, endometrial thickness and progesterone at trigger were independent predictors of clinical pregnancy, in line with previous studies [ 33 , 34 ], while BMI category was not significant after adjustment. These findings suggest that BMI influences reproductive outcomes primarily through hormonal and endometrial modulation—including impaired granulosa responsiveness [ 14 , 15 , 16 ] and increased adipose-driven estrogen metabolism [ 12 ]—rather than through direct impairment of implantation potential. This mechanistic pattern aligns with our morphokinetic data: although embryo development differed across BMI groups, these differences did not translate into reduced pregnancy rates once transfer occurred. Nonetheless, the association between elevated BMI and increased cesarean delivery rates, along with a trend toward greater metabolic complications, underscores the importance of preconception optimization and targeted antenatal care. It is worth noting that this cohort was uniformly young, which may limit the ability to detect age–BMI interactions and partially explain why BMI did not independently predict clinical pregnancy. Strengths include the large single-center cohort with standardized TLI culture, broad BMI spectrum, adjustment for confounders, and follow-up through delivery. Focusing on fresh transfers avoids cryopreservation confounding. Limitations include the retrospective, single-center design with inherent selection biases. Embryo ploidy was not assessed by preimplantation genetic testing (PGT-A), and the potential effect of BMI on aneuploidy rates cannot be excluded. Additionally, residual confounding from unmeasured variables such as diet, physical activity, and environmental exposures cannot be ruled out [ 35 , 36 ].

In conclusion, in this large cohort of fresh IVF cycles, maternal BMI influenced baseline characteristics, hormonal profiles, and delivery mode but did not reduce pregnancy rates. The most pronounced risk associated with obesity was a higher cesarean section rate, with a trend toward increased metabolic complications. Effective embryo selection and preserved implantation potential may explain the similar pregnancy outcomes, but the elevated obstetric risk profile underscores the need for individualized counseling and proactive perinatal management in this population.