Patients who had undergone an embryo transfer cycle using frozen thawed embryos from May 2016 to October 2020 at a university-based hospital were the subject of this retrospective analysis. The hospital’s database was used to collect all the information anonymously. The patients were separated into the following four groups based on the body mass index (BMI): Group 1: BMI 18.5 kg/m${}^2$; Group 2: 18.5 kg/m${}^2$ $\le$ BMI 24 kg/m${}^2$; Group 3: 24 kg/m${}^2$ $\le$BMI 28 kg/m${}^2$; Group 4: BMI $\ge$28 kg/m${}^2$ [14].

For these women, there were no definite restrictions on endometrial preparation procedures.

Natural cycle: Patients with regular menstrual cycles underwent a natural cycle. Ultrasonic monitoring was conducted on 10th of the menstruation, luteal support of progesterone capsules (Xianju Pharmaceutical, batch number: H20041902, Taizhou, Zhejiang, China) 150 mg po bid, Progynova (Bayer, batch number: H20160679, Guangzhou, Guangdong, China) 2 mg po tid, Chorionic Gonadotrophin (HCG) (Lizhu, batch number: H44020673, Zhuhai, Guangdong, China) 2000 IU im qod was added when ovulation occurred, and the timing of embryo transfer was planned based on the stage of embryonic developmental on the day of vitrification.

Hormonal replacement cycle: Patients with irregular menstruation underwent hormonal replacement on Days 2–3 of menstruation. Finmatone (Abbott Biologicals B.V., batch number: H20150345, CP Weesp, Holland, Netherlands) 2 mg, tid. Fourteen days later, ultrasound was performed and the hormone levels were checked. When the thickness of the endometrium was more than 8 mm and the E2 level was $>$200 pg/mL, yellow tablets containing 2 mg of estrogen and 10 mg of progesterone and progesterone injection (Xianju Pharmaceutical, batch number: H33020828, Taizhou, Zhejiang, China) were added to initiate transformation and provide luteal support. For cleavage, the transfer took place after 4 days and for blastocysts, it happened after 6 days.

Stimulation cycle: Patients with irregular menstruation could also undergo a stimulation cycle. Letrozole was provided orally 5 mg per day for 5 days on Days 2–3 of menstruation. When a dominant follicle has reached or over 10 mm after 5’days of letrozole (LE) administration, ultrasound monitoring as a natural cycle was conducted. In the case that has no dominant follicle growing, 37.5 IU of HMG (Lizhu, batch number: H10940097, Zhuhai, Guangdong, China) was added to increase follicle growth, with additional doses of 37.5 IU when needed. For embryos in the cleavage stage, the transfer took place after 4 days and for blastocysts, it happened after 6 days. Luteal support matched the natural cycle in many ways.

Down regulation cycle: Patients with endometriosis or adenomyosis underwent a pituitary down regulation protocol. Gonadotropin-releasing hormone agonist (GnRHa) administrated on Day 2–3 of the cycle. Fourteen to 28 days later, ultrasonic monitoring was performed and the hormonal levels were tested; if down regulated successfully, hormone replacement and luteal support were similar to the hormone replacement cycle.

The main measure was the live birth rate, which is defined as the ratio of live births delivered after 28 weeks of gestation to the number of embryo transfer cycles. The rates of chemical pregnancy, clinical pregnancy, implantation and miscarriage were decided to be the secondary outcomes indicator. Serum $\beta$-hCG levels $>$50 U/L on Day 12 following blastocyst transplantation or Day 14 following cleavage-stage embryo transplantation were defined as biochemical pregnancy. A gestational sac founded by transvaginal ultrasonography 4 weeks after embryo transfer was considered a clinical pregnancy. The number of gestational sacs divided by the number of embryos transferred equals the implantation rate. Miscarriage was the unintentional loss of a pregnancy before 28 weeks of gestation.

Means $\pm$ SD (standard deviation) was the form of data presenting. SPSS Version 23.0 (SPSS Inc., Chicago, IL, USA) was used for data analyze. The Kolmogorov–Smirnov test was used to determine whether continuous variables had a normal distribution, and ANOVA was used to compare normally distributed count data with non-normally distributed count data. $\chi$${}^2$ test were used to compare Categorical variables where appropriate. Logistic regression analysis was applied to examine the impact of BMI on the clinical result after confounding factors were adjusted. The relationship between the trend of clinical outcomes based on the normal weight group and increasing BMI was analyzed by Chi Square for Trend test.

In the present study, a striking finding was that weight increased significantly with age. In line with the finding of Barbara Luke et al.’s [15] study that older women were more likely to be obese before having children. Our clinical outcomes were similar in all groups, but showed an inverse tendency with increasing BMI. Interestingly, our results agreed with those of another study conducted in a Chinese group. They found that patients with obesity negatively affected the clinical outcomes compared with the overweight group, likely because they did not use the China classification: obesity was defined as a BMI $>$27 kg/m${}^2$, overweight was defined as 23–27.4 kg/m${}^2$and a normal BMI range was 18.5–22.9 kg/m${}^2$ [16]; additionally, the sample size was large. By contrast, these results may be inconsistent with the limited study conducted by Insogna et al. [17] reveal that the BMI in overweight individuals was not associated with a normal BMI range, likely because of racial differences. The proposal of Asians at high risk for chronic diseases such as type 2 diabetes and cardiovascular disease is considerable at lower BMI cutoffs for overweight individuals [15]. Considering that Asians have a higher percentage of body fat than Caucasians, the Working Group on Obesity in China recommends lower cutoffs for BMI to define overweight and obesity than those used for Caucasians [1, 18]. We should not neglect the influence of overweight in female individuals.

Female obesity is associated with a higher proportion of PCOS patients. This finding agrees with the theory that obesity in women contributes to the development of PCOS [8]. A majority of studies have revealed that the oocyte quality, maturity, and size decreases with increasing BMI. Also, it has been observed that female obesity is correlated with a reduced number of embryos [5]. Additionally, studies on mice have provided laboratory evidence for embryo development. In a prospective study of mouse embryos, Wenhong Ma et al. [19] found that the lipid content and apoptosis rate of fresh embryos from obese mice were significantly higher and the development of embryos was notably delayed, likely caused by endoplasmic reticulum stress, mitochondrial dysfunction, and apoptosis resulting from lipid accumulation. Another study from mice provided laboratory evidence that obesity might impair oocyte maturation and development [20].

Patients who are overweight or obese have reduced implantation rates, which also lead to fewer biochemical pregnancies, clinical pregnancies and live births, suggesting a suboptimal endometrial receptivity in these individuals. According to a recent study, gene expression within the window of implantation in obese women, demonstrated that genes mainly regulating development and biological function such as transcription and biosynthesis were deregulated, particularly in women who are obese with infertility or polycystic ovary syndrome [21]. Other studies on ovum donation have revealed a decrease in endometrium receptivity in obese women [22, 23, 24].

Although our study did not explore the mechanism or discuss treatments for patients who were overweight or obese to improve pregnancy outcomes, research has shown that these patients appear to improve pregnancy outcomes significantly from the administration of myo-inositol, vitamin D and alpha-lipoic acid in order to improve insulin sensitivity [25, 26]. Administration of myo-inositol can lower the hyper-insulinemic state in obese women with and without PCOS [25, 27]. A pilot study has found that adding myo-inositol, alpha-lipoic acid and folic acid to IVF treatment for overweight/obese non-PCOS women improved the quality of their oocytes [27]. Nevertheless, future investigation into the mechanism underlying the improvement in patients’ metabolic and psychological status following adjuvant treatments is necessary.

The shortcomings of our study were its retrospective nature and small sample size, which reduces its strength of evidence. Prospective studies in population with large sample sizes would strengthen the conclusion that overweight may reduce the rate of live birth in FET cycles. Additionally, too few patients were in the obese cohort (6%), possibly weakening the power to detect differences.

However, we adjusted for confounding factors by logistic analysis to balance the data and make convincing conclusions. We observed a decrease in endometrium receptivity and resulted in a decrease in the implantation rate in patients who were overweight or obese. Large sample studies are needed in the future to make robust conclusions. Furthermore, this study is the first to focus on overweight in Chinese women based on the reasonable cutoff BMI of a Chinese cohort (underweight: BMI 18.5 kg/m${}^2$; normal weight: 18.5 kg/m${}^2$ $\le$ BMI 24 kg/m${}^2$; overweight: 24 kg/m${}^2$ $\le$ BMI 28 kg/m${}^2$; obese: BMI $\ge$28 kg/m${}^2$) [14]. Further trend analysis revealed that a decreasing tendency of the rates of clinical pregnancy and live delivery with BMI change. Overweight and obese women’s endometrial receptivity should not be neglected.