As a complication of assisted reproductive technology (ART), ectopic pregnancy (EP) is a serious consequence on women’s health. In the United States [1], from 2001 to 2011, the incidence of EP after ART ranged from 2.0% to 1.6%. It has also been reported that the incidence of EP after in vitro fertilization-embryo transfer (IVF-ET) is between 2.1 and 8.6% [2]. The widely accepted risk factors for EP include tubal factor infertility, the number of embryos transferred, embryo transfer in a fresh cycle, and day 3 embryo transfer [3, 4, 5]. Tubal factor infertility significantly increases the risk of EP by 25% [1].

In addition, the incidence of some rare EPs, such as interstitial pregnancies (IPs) and heterotopic pregnancies (HPs) (both an ectopic and intrauterine pregnancy at the same time [6]), is also increasing. These conditions are difficult to diagnose and treat after ART. IP refers to an EP in which implantation occurs in the interstitial duct, which is the fallopian tubal segment that traverses through the myometrium. Some IPs may last up to 7–16 weeks of gestation, at which time rupture can result in massive bleeding and high mortality [7]. Some recent studies have indicated that the incidence of IP is 0.8% per pregnancy and 35.5% in all EP cases after ART [8]. The mortality rate of IP is 2.5%, which is seven times greater than that of EP [9, 10, 11].

The incidence of EP from IVF-ET is increased 2.5- to 5.0-fold compared with that of ectopic spontaneous pregnancy [3]. Similarly, the incidence of IP is lower in natural pregnancy but more common after ART [8]. The reasons for this are unclear. In our original research, we analyzed the clinical data of 21 patients with IP after ART. The results showed that 90.48% of the IP patients had a history of tubal-related surgery, among whom 78.95% had unilateral or bilateral tubal resection or proximal tubal occlusion. Approximately 85.71% of the IP patients had undergone double embryo transfer. Therefore, we hypothesized that there was an increased incidence of EP and IP after ART in patients with a history of tubal surgery, especially in those who had undergone salpingectomy or proximal tubal occlusion. The number of embryos transferred may be a related risk factor.

The purpose of our study was to investigate the effect of different numbers of transferred embryos on the EP outcome after ET in patients who had undergone salpingectomy or proximal tubal occlusion, especially regarding the related factors of IP.

The clinical outcomes of 11,609 patients undergoing IVF-ET in the Centre for Clinical Reproductive Medicine, The First Affiliated Hospital of Nanjing Medical University, from January 2005 to November 2017 were retrospectively analyzed. Among them, 5388 women who achieved a clinical pregnancy were included in the study. The inclusion criteria were as follows: Patients who had a history of at least one unilateral/bilateral salpingectomy or proximal tubal occlusion and patients with a clinical pregnancy that was obtained after ART. The exclusion criterion was as follows: Patients who had not undergone salpingectomy or proximal tubal occlusion.

Once IVF treatment was initiated in our clinical center, the patients’ baseline characteristics, medical history, IVF process data, and follow-up pregnancy outcomes were routinely captured in the reproductive medicine clinical center database. Trained nurses continued to collect information about the patient’s treatment during the entire course of treatment until the end of each treatment with a clinical outcome.

The selection process of the patients is shown in Fig. 1. A clinical pregnancy was defined as the ultrasound observation of one or more pregnancy sacs after ET, including intrauterine and extrauterine pregnancies. A clinical intrauterine pregnancy was diagnosed as one or more pregnancy sacs in the uterus. EP was defined as one or more pregnancy sacs outside the uterus, including a clinical HP. Ipsilateral EP on a tubal remnant, contralateral EP and bilateral IP can occur after salpingectomy.

Fig. 1.

Patient selection flowchart.

The diagnosis criteria of IP under ultrasound were as follows: no pregnancy sac in the uterus, a gestational sac located laterally in the interstitial (intramural) part of the tube and surrounded by thin (5 mm) myometrial tissue, and the presence of the “interstitial line sign” [12, 13]. EP and IP were diagnosed by laparoscopic surgery or ultrasound.

This study was approved by the Ethics Committee of the First Affiliated Hospital of Nanjing Medical University (2018-SR-063). This study was retrospective, so informed consent was not needed.

Descriptive statistics (frequencies and means) were initially performed. Before analysis, continuous variables (maternal age, infertility duration, body mass index (BMI), antral follicle count, follicle-stimulating hormone (FSH) level) were examined to determine compliance with normal distribution. Since these variables were not normally distributed, we used the Mann–Whitney U test to examine the association between these variables in participants. Categorical variables were analyzed with chi-square tests. If the expected frequencies were less than 5, Fisher’s exact test was used for the analysis. Variables with a p 0.05 were included in multivariate logistic regression analysis. Statistical Package for Social Science (SPSS) 25.0 (IBM, Armonk, New York, USA) was used for statistical analysis. A value of p 0.05 was considered statistically significant.

Among the 11,609 patients undergoing IVF-ET, 5388 women achieved clinical pregnancy in our study. The overall clinical pregnancy rate was 46.4%. Among these 5388 women, 131 women (2.4%) had an EP, and 26 women (0.5%) had an IP. The characteristics of the patients are shown in Table 1.

To explore the potential risk factors for EP and IP, the baseline and clinical characteristics were analyzed. The results are shown in Tables 2,3. Maternal age, infertility duration, BMI, antral follicle count, FSH level, type of infertility and type of embryo transfer cycle were not significantly different between the two groups (p $\mathrm{>}$ 0.05). The day of embryo transfer and number of embryos transferred were significantly different between the nonectopic pregnancy and ectopic pregnancy groups (p 0.05). However, only the number of embryos transferred was significantly different between the noninterstitial and interstitial pregnancy groups (p 0.05).

During ART treatment, fallopian tube factors remain a major risk factor for EP, especially in women with a prior history of fallopian tube surgery [14]. Patients who undergo a conservative operation or tubal microsurgery have a higher incidence of EP after assisted pregnancy [15]. Some researchers also supposed that salpingectomy and tubal infertility were the most important variables affecting the risk of IP [13]. Therefore, it is of great significance to find effective intervention measures to reduce the risk of EP for this population experiencing infertility. In our original research, a history of salpingectomy or proximal tubal occlusion and the number of embryos transferred were suggested to be risk factors for IP. Therefore, we specifically analyzed the clinical outcome during IVF treatment for patients who had undergone salpingectomy or proximal tubal occlusion. We found that the incidence of EP was 2.4% (131/5388) in the patients who obtained a clinical pregnancy in our study, which is higher than the average incidence of EP after ART (range between 1.38 and 2.10% of clinical pregnancies) [3]. IP is a rare condition, accounting for 2–4% of all tubal pregnancies after ART [11]. However, IPs accounted for 19.85% of the EPs in our study. The reason may be that the electrocoagulation-damaged fallopian tube stump produces chemokines that may lead to embryonic migration to and implantation in the uterine horn [16].

We found that the incidence of EP and IP in patients with single embryo transfer was significantly lower than that in patients with double embryo transfer during all ET cycles, especially frozen-thaw ET cycles. Perhaps we can select single embryo transfer for IVF-ET treatment in frozen-thawed ET cycles in patients who have undergone salpingectomy or proximal tubal occlusion to reduce the risk of EP, especially that of IP. In recent years, elective single embryo transfer (eSET) has been supported. Some researchers have suggested that eSET is associated with a better perinatal outcome [17]. eSET has been shown to reduce the number of pregnancy complications and neonatal morbidity and mortality with no impact on cumulative live-birth rates [18].

Because different IVF treatments would bias the results, we assessed fresh and frozen-thaw ET cycles respectively. In our study, the incidence of EP per clinical pregnancy following frozen-thaw ET cycles (2.2%) appeared to be slightly lower than that following fresh ET cycles (2.7%). The EP rate in our study was similar to the rate reported previously, which ranged from 1.28% to 2.22% for frozen–thawed ET cycles and from 1.92% to 4.62% for fresh ET cycles [5, 19, 20]. We hypothesized that the reason for the low EP rate of frozen–thawed ET cycles compared to fresh ET cycles was that frozen-thaw ET cycles have better endometrial receptivity and synchronization than stimulated fresh ET cycles, and the uterine environment is better for frozen-thaw embryo implantation. Unexpectedly, the incidence of IP per clinical pregnancy following frozen-thaw ET cycles (0.5%) appeared to be slightly higher than that following fresh ET cycles (0.4%). This observation is in contrast with the report that frozen-thaw ET cycles have a lower incidence of EP but is similar to the findings of Wang et al. [8]. The exact reason is unclear, and further studies are required to determine the reason.

Our study also found that the day of embryo transfer were associated with the incidence of EP in women undergoing frozen-thaw ET cycles. Transplanting blastocysts (day 5–6 embryos) in patients undergoing frozen-thaw ET cycles significantly reduces the risk of EP. However, multivariate analysis showed that the incidence of EP was only related to the number of embryos transferred. This may be related to the small number of EP patients. Cong Fang et al. [21] showed that fresh transfers and frozen-thawed day 3 transfers were associated with a higher EP rate than frozen-thawed day 5 blastocyst transfers in patients undergoing ART. Bingqian Zhang et al. [22] indicated that day 5 blastocyst transfer reduces the risk for EP in cycles compared with day 3 blastocyst transfer. It was also reported that [23] the EP rate was 3.4% in day 3 embryo transfer cycles and 2.1% in day 5 blastocyst transfer cycles. The reason may be that higher levels of progesterone and 17 $\beta$-estradiol can allow transferred day 3 embryos to readily enter the fallopian tube and have a negative effect on intrauterine embryo implantation. Single blastocyst transplantation is recommended to increase the implantation rate and reduce the EP rate and IP rate. We hope to expand the sample size in subsequent studies to determine the relationship between the day of embryo transfer and EP.

We recognize that this study has some limitations. First, although our study enrolled 11,609 subjects, it was limited by the retrospective nature of the analysis. It is possible that not all confounders were accounted for, and a prospective cohort study is needed. More confounders, such as controlled ovarian hyperstimulation protocols and doses, need to be included. Second, the number of women with ectopic pregnancy and interstitial pregnancy whose data were analyzed was relatively small. The limited sample size precluded us from drawing solid conclusions. Finally, the perinatal outcome results were lacking, which biased our conclusions.

In summary, the findings of our study showed that the embryo transfer number is a risk factor for EP, especially IP, in patients undergoing in frozen-thaw ET cycles who have undergone salpingectomy or proximal tubal occlusion. Therefore, it is recommended to select single embryo transfer for IVF-ET treatment in patients undergoing frozen-thawed ET who have undergone salpingectomy or proximal tubal occlusion to reduce the risk of EP, especially IP. Furthermore, we considered how to manage the interstitial part of the fallopian tube in salpingectomy or proximal tubal occlusion operations to reduce the risk of IP. Can we perform blastocyst transplantation to reduce the incidence of ectopic pregnancy in women with tubal infertility?

Conception and design of the study, acquisition of data, or analysis and interpretation of data—CW, YZ, XX, XM, YM. Drafting the article or revising it critically for important intellectual content—CW, YZ, DWW. Final approval of the version to be submitted—DWW and JL. All authors read and approved the final manuscript.

The Ethics Committee of the First Affiliated Hospital of Nanjing Medical University approved the study (2018-SR-063). This study was retrospective, so informed consent was not needed. This manuscript conforms to the Enhancing the QUAlity and Transparency Of health Research (EQUATOR) network guidelines.

We thank all the peer reviewers for their opinions and suggestions.

This work was supported by the National Nature and Science Foundation of China (81730041); the National Key R&D Program of China (2017YFC1001604; 2016YFC1000207); the collaborative research program of Southeast University and Nanjing Medical University (2017DN26); and the Women and Children Youth Talent Project of Jiangsu Province (FRC201793[EKRC]).

The authors declare no conflict of interest.