In this retrospective study with a large sample size, no significant association was found between female BMI and the risk of EP in fresh ET cycles after controlled ovarian hyperstimulation (COH) after adjusting for potential confounding factors. In addition, the EP rates among three BMI groups were also comparable in FET cycles.

To date, there has been numerous data identifying risk factors for EP during ART treatment [4, 7, 19]. Whether a correlation exists between female BMI and the risk of EP has been previously discussed. Several Chinese studies of natural conceptions and artificial insemination cycles with donor sperm observed higher EP rates in obese women than those with underweight and normal weight [20, 21], suggesting potential obesity-related tubal dysfunction. In the study including 136,605 clinical pregnancies about the effect of ovarian hyperstimulation protocols on the EP in the United States, Londra et al. also observed that female obesity was associated with an increased risk of EP (BMI ≥ 30 kg/m2 vs. < 30 kg/m2: adjusted OR 1.33; 95% CI, 1.19–1.47; P < 0.001) in fresh autologous ET cycles [8]However, limited by the relatively low incidence of EP and the small proportion of underweight women in the infertile population, the effect of low BMI per se on EP in women was ignored. To our knowledge, Cai et al. firstly proposed a clear association between low BMI and EP during IVF treatment [15]. In the cohort of 16,378 pregnancies derived from fresh and frozen-thawed ET cycles including 2155 underweight women, low BMI (< 18.5 kg/m2) was associated with the increased odds of EP (2.92% vs. 2.02%, adjusted OR 1.61; 95% CI, 1.19–2.16; P = 0.002, compared with the normal BMI group) but not high BMI (> 24.9 kg/m2). They conjectured that underweight women might have a nutrition-related unfavorable uterine environment for embryos, which possibly involved in the underlying mechanism of the higher EP risk. However, Bellver J made a dissent that a suboptimal endometrial milieu would be more likely to hamper embryo implantation or ongoing pregnancy in uterus, resulting in lower implantation rates or higher miscarriage rates rather than implanting in fallopian tube, a tissue non-physiologically suitable for conception [11]. And in women undergoing either oocyte donation or autologous euploid embryo transfer, no significant differences in pregnancy outcomes have been previously reported between underweight and normal weight groups [22,23,24,25], which did not support the premise that being underweight might have an adverse effect on the uterine environment. In our study, neither an independent effect of female BMI on EP in fresh cycles after adjusting for potential confounders nor a significant difference in EP rates among the three BMI groups in FET cycles was demonstrated, which differs from the findings of Cai et al. This difference may be attributed to discordant BMI classification criteria and variations in the proportion of the subjects among three groups: 6.2% (1703/27,600) of underweight women, 61.0% (16,840/27,600) of women with normal weight, and 32.8% (9057/27,600) women with overweight/obesity were included in the present study according to the standards of WGOC and ILSI, whereas the proportions were 12.1% (1324/10,930), 82.8% (9047/10,930) and 5.1% (559/10,930) respectively in the study by Cai et al. based on the WHO criteria.

In this study, the EP rates in fresh ET and FET cycles were 2.43% and 2.82% respectively. Our data also identified some risk factors for EP related to IVF treatment, such as tubal factor, elevated peak estrogen levels after COH, endometrial thickness, and stage of embryo for transfer in fresh ET cycles, which concurred with those previously reported [7, 10, 19]. For fresh ET cycles, the results of multivariate regression analysis suggested that the uneven distribution of these factors among the three groups potentially contributed to the higher EP rates in underweight women rather than the independent effect of low female BMI.

Supraphysiologic hormonal milieu after COH in fresh ET cycles is known to increase the risk of EP during IVF treatment, and high estradiol levels may interfere the regulation of tubal physiologic process, which possibly plays a role in the pathophysiological mechanism of tubal EP development after embryos transfer [26]. Wang et al. reported that in fresh ET cycles, high estradiol levels [estradiol on hCG day > 4085 pg/mL] were associated with the increased EP risk (3.4% vs. 2.0%, adjusted OR, 1.99; 95% CI, 1.19–3.35; P = 0.009) in women without polycystic ovary syndrome [27]. In our study, lean women accompanied with higher estradiol levels on average yielded higher EP rates in fresh ET cycles, and high peak estradiol levels were related to an increased risk of EP after adjusting potential confounding factors, which was consistent with our previous findings in both tubal infertile women and non-tubal infertile women undergoing fresh ET cycles [7]. For FET cycles without exogenetic high-dose gonadotropin supplementation, hormone levels could be considered close to natural condition, and EP rates were similar among three BMI groups. Taken together, our findings supported the conclusions of Wang et al., and we noticed that lean women appeared to have a sensitivity to ovarian stimulation and develop higher estradiol levels after COH [28]. Thus, when exploring the independent impact of female BMI on EP risk, attention should be paid to eliminating the interference of estradiol levels.

Our findings also confirmed that thin endometrium thickness (EMT) contributed to an increased risk of EP in both fresh cycles and FET cycles, which is in line with multiple existing reports [1, 5]. One potential explanation for the relationship between thin endometrium and higher EP rates is the difference in oxygen tension between the thin endometrium and the fallopian tube. In the thin endometrium with a thin or absent functional layer, the implanting embryos would be closer to the spiral arteries in the basal endometrium layer where the embryos are exposed to higher oxygen concentrations, which may be detrimental to their growth. By contrast, the oxygen tension is relatively lower in the fallopian tube [29]. Consistent with many past studies, this study also presented similar higher EP rates for cleavage-stage embryos compared to blastocyst transfer [7, 9, 30]. Several speculations have been proposed: transferring blastocysts into the uterine cavity seems closer to the physiological state than that of cleavage-stage embryos, requiring a shorter interval for further development before implantation; the size of a blastocyst is larger than that of a cleavage-stage embryo, all of which may reduce the chance of embryos migrating to the fallopian tube [9, 31].

The strength of the current study is the large sample size with the adequate number of underweight women in a single center. Given that the characteristics and distribution of BMI vary by race and region, BMI categories were based on the recommendations on cut-off points of BMI in Chinese adults by WGOC. Therefore, the results were more suitable to provide counseling and guidance for Chinese women undergoing IVF treatment. And the independent effect of female BMI on EP development was discussed in fresh ET and FET cycles respectively. It must be acknowledged that this study has some following limitations. First, as a retrospective study, some potential confounders, such as the volume of transfer fluid, transfer depth and smoking habits, were not available in our database. Second, when investigating the association between development stages of embryos transferred and the occurrence of EP, our study mainly focused on the comparison between cleavage-stage embryos and blastocysts without differentiating the specific stage of a certain blastocyst. In addition, this study spans a long duration over ten years when IVF technology has progressed rapidly, and thus there might be potential biases related to the evolution of IVF behind the results. More well-designed prospective studies with large cohort are needed to evaluate the effect of female underweight or excess weight on EP development after IVF treatment, and possible underlying mechanisms.