This study was reviewed and approved by the Human Ethical Committee of the University of Teikyo Hospital (trial registration number: 20-094). The deidentified medical records of 150 female patients who underwent laparoscopic surgeries, performed after obtaining informed consent, from June 1, 2015 to December 31, 2021, were reviewed retrospectively. In total, 22 cases were excluded for the following reasons: peritoneal ectopic pregnancy (9 cases), miscarriage from the fallopian tube (8 cases), no detection of pregnancy tissue (3 cases), ovarian ectopic pregnancy (1 case) and uterine horn pregnancy (1 case). Since there were relatively few patients in our hospital whose first choice was treatment with intramuscular methotrexate injection (IMI), we did not analyse these cases. We extracted data on representative patient characteristics, such as age, delivery history, presenting symptoms, and physical data, from medical records. In particular, in all patients, serum hCG levels were examined several times, including before and after surgery, and we extracted these data. To evaluate the failure rates of laparoscopic surgeries, we extracted the data on PT (13 cases), in which patients were diagnosed per rising or plateauing serum hCG levels postoperatively and needed to be treated by IMI.

The basic procedures of laparoscopic salpingostomy were performed according to the following steps: (1) an umbilical trocar (12 mm) and three lower abdominal trocars (5 mm) were placed in a diamond pattern trocar conformation to access the abdominal and pelvic cavities; (2) diluted vasopressin was injected, and a horizontal incision was made over the ectopic gestation using an ultrasonic scalpel; (3) the products of conception were picked up into a 200-mL MemoBag™ and removed from the umbilical trocar; (4) trophoblastic tissue was verified macroscopically; and (5) the abdominal cavity was washed out with saline.

First, to compare the rates of PT between laparoscopic salpingostomy and salpingectomy, we divided the patients into two groups according to these treatment methods. In our hospital, the choice between these two methods depended largely on the patient’s situation, and the numbers were approximately equal. In only 6 cases, tubal pregnancy rupture was detected during surgery. In these two groups, we compared the 19 indexes shown in Table 1 by using Student’s t-test, Pearson’s chi-square test and the Mann–Whitney U test. By referring to a previous study evaluating the influence of the ratio of serum hCG levels before and after laparoscopic salpingostomy on the possibility of PT [12], we also compared this index (“hCG ratio” in Table 1). Second, we tried to evaluate the risk factors for PT, especially to detect the predictive values of preoperative serum hCG levels. Then, four approximate cut-off values were used for serum hCG levels: 2000, 4000, 6000 and 8000 mIU/mL. For each of these four parameters, we performed similar analyses. We first analysed the 62 patients who underwent laparoscopic salpingostomy and then analysed all 128 patients. To control for confounding factors, we divided the patients into two groups according to the presence or absence of each factor and performed multivariate logistic regression analysis. In this analysis, we assessed the influence of the following 14 factors, which were seemed to be related to the conditions of patients and the difficulty of operations: (1) Advanced age, defined as an age $\ge$38 years; (2) High body mass index (BMI), defined as a BMI $\ge$25 (kg/m${}^2$); (3) Nulliparity, defined as no previous delivery; (4) Embryo transfer, defined as cases in which the patients became pregnant after embryo transfer; (5) Abnormal bleeding, defined as patients with abnormal vaginal bleeding caused by TEP; (6) Abdominal pain, defined as patients with abdominal pain caused by TEP; (7) Pelvic haematoma, defined as cases in which pelvic haematoma was detected by TVUS before surgery or during surgery; (8) Foetal heartbeat, defined as cases in which foetal heartbeat was detected by TVUS; (9) Isthmic tubal pregnancy, defined as cases in which isthmic tubal pregnancy was diagnosed during surgery; (10) Abdominal adhesion, defined as abdominal adhesion detected by laparoscopic inspection immediately after the start of surgery; (11) Salpingostomy; (12) High hCG, defined as a serum hCG level $\ge$2000, 4000, 6000, or 8000 mIU/mL; (13) Coexistent leiomyoma, defined as leiomyoma detected by laparoscopic inspection; and (14) Coexistent endometriosis, defined as endometriosis detected by laparoscopic inspection. Since the average age of the patients was 33.4 $\pm$ 4.8 years and since over 40% of the included patients became pregnant after infertility treatments, “Advanced age” was defined as patients aged 38 years or older with reference to a previous report [16]. Statistical analyses were performed using Microsoft Excel (Microsoft Corporation, Redmond, WA, USA) and JMP version 12 for Windows (SAS Institute, Inc., Tokyo, Japan) to determine the correlations between patient characteristics and the failure of laparoscopic surgery. The odds ratios (ORs) and 95% confidence intervals (CIs) were estimated to determine the strengths of the correlations. p 0.05 was considered statistically significant.

To estimate the appropriate postoperative follow-up period, we created linear regression lines based on the scatter plot of the association between reduction rates of serum hCG levels and elapsed time after surgery (laparoscopic salpingostomy: number of spots = 236; laparoscopic salpingectomy: number of spots = 289) by using the ordinary least-squares method with Microsoft Excel. In this analysis, 13 patients with PT were excluded. These plots did not include data below the detection sensitivity (laparoscopic salpingostomy: number of spots = 7; laparoscopic salpingectomy: number of spots = 4). We calculated the indexes as follows: (1) we divided the serum hCG levels postoperatively by these levels immediately before the operation and (2) calculated the Log${}_{10}$ values. The formula used was Log${}_{10}$ (serum hCG level after/before operation (mIU/mL)). The relationship between this logarithmic index and elapsed time after surgery (days) was plotted. Then, by referring to the slopes of these regression lines, we estimated each required follow-up period for patients with laparoscopic salpingostomy or salpingectomy.

The average age, BMI, operation time, blood loss volume and treatment period of the included patients were 33.4 $\pm$ 4.8 (20–44) years, 21.3 $\pm$ 2.8 (17.1–29.9) kg/m${}^2$, 54.1 $\pm$ 20.1 (25–144) min, 126.4 $\pm$ 237.5 (0–1334) mL and 32.4 $\pm$ 17.9 (9–110) days, respectively. The average and median serum hCG levels just before surgery were 6528.1 $\pm$ 11513.8 (2.9–91673.2) and 2879.5 mIU/mL, respectively. Among these cases, we detected one rare case in which the hCG level of intraperitoneal haematoma was 606.0 mIU/mL and laparoscopic salpingostomy was successfully performed. The average ratio of serum hCG levels before and after surgery was 0.38 $\pm$ 0.16 (0.10–1.18). When the characteristics of the patients with laparoscopic salpingostomy and salpingectomy were compared, significant differences in patient age, operation time, treatment period, the number of patients with embryo transfer, PT and coexistent leiomyoma were detected (Table 1). In the analysis of serum hCG levels, we detected a trend wherein laparoscopic salpingectomy was chosen to treat patients with high serum hCG levels, and the average and median hCG levels were higher in patients with laparoscopic salpingectomy than in those with salpingostomy (9393.8 $\pm$ 15177.3 vs. 3477.4 $\pm$ 3493.0 and 4093.4 vs. 2117.0 mIU/mL). Although laparoscopic salpingostomy tended to be performed for patients with relatively low serum hCG levels, 12 out of the 13 cases with PT were detected among the 62 cases with laparoscopic salpingostomy. Among these 12 cases, PT was detected approximately two weeks after surgery (15.0 $\pm$ 8.8, 3–36 days) by the rebound of decreased serum hCG levels (1304.6 $\pm$ 1177.5, 101.2–3465.3 mIU/mL). In 11 out of 12 cases, we performed IMI once, but in one case, we needed to perform IMI twice. Among patients undergoing laparoscopic salpingectomy, only one was diagnosed with PT and needed IMI 4 days after surgery (serum hCG levels: 456.6 mIU/mL). Next, we compared the serum hCG levels of these 12 PT cases with those of the 50 cases with successive laparoscopic salpingostomy. The average serum hCG level tended to be high among the 12 cases with PT (4635.4 $\pm$ 2401.1 vs. 3199.5 $\pm$ 3672.6 mIU/mL, p = 0.20), and the median was more than twice that of the non-PT cases (4627.6 vs. 1899.1 mIU/mL, p 0.05). On the other hand, the ratios of the serum hCG levels before and after laparoscopic salpingostomy were similar (0.45 $\pm$ 0.23 vs. 0.40 $\pm$ 0.17, p = 0.43).

To evaluate the association between PT and serum hCG levels, we referred to the aforementioned difference in serum hCG levels between successful and failed cases after laparoscopic salpingostomy. Specifically, we set the cut-off values based on serum hCG levels of 2000, 4000, 6000 and 8000 mIU/mL and performed a multivariate analysis that included an additional 13 factors. First, in this analysis of 62 cases with laparoscopic salpingostomy (Table 2), “high hCG”, including serum hCG levels of 2000 and 4000 mIU/mL, showed a significant difference (OR = 14.0, p 0.01 and OR = 6.3, p 0.01), but significance was not apparent for other factors. When focusing on patients whose serum hCG levels were lower than 2000 mIU/mL, we detected few cases with PT after laparoscopic salpingostomy (OR = 0.071, 95% CI: 0.0086–0.60, n = 1/29). Second, as expected, the analysis of all 128 cases showed the following significant factors: “hCG $\ge$2000” (OR = 8.6, p 0.01), “hCG $\ge$4000” (OR = 3.5, p 0.01) and “salpingostomy” (OR = 15.6, p 0.01) (Table 3). In both analyses, two values for serum hCG levels, 2000 and 4000 mIU/mL, were considered to be important. However, other significant predictors could not be detected. From these results, we can provide the following three classifications: (1) for patients whose serum hCG levels are under 2000 mIU/mL, we can safely perform laparoscopic salpingostomy; (2) for patients whose serum hCG levels are over 4000 mIU/mL, we should select laparoscopic salpingectomy to avoid PT; and (3) for patients whose serum hCG levels are between 2000 and 4000 mIU/mL, we should determine the method while considering other patient factors.

To estimate the required follow-up period after laparoscopic salpingostomy and salpingectomy, we calculated the rate of decrease in the serum hCG levels by using the logarithmic index (Fig. 1). In this analysis, the slopes of the regression lines in cases with laparoscopic salpingostomy and salpingectomy were –0.092 and –0.11 (R${}^2$ = 0.74 and 0.87), respectively. By using these results, to pursue the decline in serum hCG levels until reaching approximately one-thousandth to one ten-thousandth, we needed follow-up periods of 27 to 36 days after laparoscopic salpingectomy and 33 to 43 days after laparoscopic salpingostomy.

Fig. 1.

Reduction rate of serum hCG levels after surgery. The reduction rates of serum hCG levels after laparoscopic salpingostomy (black triangles) and salpingectomy (white circles) are shown with regression lines and R${}^2$ values. X-axis: elapsed time after surgery (days). Y-axis: Log${}_{10}$ (serum hCG level after/before operation (mIU/mL)). Abbreviations: Log, logarithm; hCG, human chorionic gonadotropin.