Hysteroscopic suture of levonorgestrel-releasing intrauterine system: learning curve for in vitro simulation training

Abstract

Objective

To simulate hysteroscopic suturing in vitro and analyze the learning curve of gynecologists with different experience levels.

Methods

Three gynecologists were trained on uterine models in a circulating water box. The posterior uterine wall was sutured 10 times under hysteroscopy for 5 consecutive days, and the time of each suture procedure was recorded.

Results

Doctors A, B, and C completed 50 posterior uterine sutures. After Dr. C completed 50 sutures on the posterior wall, he added 50 sutures on the anterior wall (Group D). The mean suturing time was 71.54 ± 68.158 s in Group A, 50.10 ± 28.060 s in Group B, 34.04 ± 10.457 s in Group C, and 30.38 ± 8.734 s in Group D. The difference between Groups C and B and between Groups B and A was statistically significant. There was no statistically significant difference between Groups C and D. Simulation curves were created using the number of features as the abscissa and cumulative sum as the coordinate, with peak curves of 19, 27, and 18 cases for Group A, B, and C, respectively.

Conclusion

Doctors with experience in single-hole laparoscopic surgery or hysteroscopic suture surgery can significantly shorten the hysteroscopic suturing time.

Introduction

The levonorgestrel-releasing intrauterine system (LNG-IUS) (Mirena; Bayer Pharmaceuticals, Berlin, Germany) has a T-shaped appearance with 32-mm longitudinal and transverse arms, and there are no other specifications regarding this product in China. The LNG-IUS not only provides contraception but also treats multiple diseases such as menorrhagia, dysmenorrhea, endometrial hyperplasia.^1,2 However, some patients repeatedly lose the LNG-IUS device. Patients with adenomyosis have a significantly higher risk of losing the LNG-IUS because of expansion of the uterine cavity and heavy menstruation, and in one study, the cumulative rate of losing the IUS was 11% in 1 year. ³ This raises the question of whether there is a reliable way to guarantee that the LNG-IUS will stay in the uterine cavity. Zhu et al. ⁴ were the first to report the procedure of suturing the LNG-IUS under hysteroscopy. We have also demonstrated the feasibility of suturing the LNG-IUS under hysteroscopy. ⁵ However, this technique is difficult to perform. To shorten the operation time and reduce the occurrence of complications, gynecologists must undergo in vitro simulation training to shorten the learning curve and improve their operational skills. Throughout the process of hysteroscopic fixation of the LNG-IUS, suturing under hysteroscopy is the key point of the procedure as well as the most important factor affecting the operation time. Because this is a new technology, evaluating its operability and practicality is crucial in determining whether it can be promoted. Studying the learning curve not only helps to demonstrate stable technical proficiency levels but also helps to establish a training plan. A large amount of surgical and gynecological literature supports the performance of objective evaluation of the reliability and effectiveness of surgical skills through simulation training.^6–10 However, there are no relevant literature reports to date, either internationally or domestically. We therefore performed an experiment to preliminarily explore the practicality of suturing under in vitro simulated hysteroscopy.

Methods

This study involved the use of a hysteroscopy training box (KMS Medical Technology Co., Ltd., Changsha City, Hunan Province, China), endoscope camera system (KMS Medical Technology Co., Ltd.), hysteroscopic cold knife surgery system (KMS-II; KMS Medical Technology Co., Ltd.), silicone uterine model (X519H; Johnson & Johnson, New Brunswick, NJ, USA), and PROLENE monofilament polypropylene suture (W8977; Johnson & Johnson). The uterine model was placed in the water box, the slot of the uterus was fixed, the water box was filled with tap water, and the uterine model was submerged to the highest water level line. The hysteroscopy training box (Figure 1) automatically circulated the liquid and could provide a pressure of 100 to 150 mmHg of intrauterine pressure; this pressure was uniformly set at 100 mmHg.

Figure 1.

The hysteroscopy training box and silicone uterine model.

Training method

The reporting of this study conforms to the STROBE guidelines. ¹¹ Three gynecologists with >5 years of experience in performing hysteroscopic surgery were trained on uterine models in a circulating water box. Before the practice, the gynecologists received an explanation of the details and key points of the hysteroscopic suture technique and watched teaching videos and previous surgical videos. During the training, the trainees were guided and timed by doctors with surgical experience. The suturing time was defined as follows. After the suture was placed in the uterine cavity, the hysteroscope was placed at the external cervical os and the timer was started. The suture was placed in the posterior wall of the uterus (Figure 2(b–g)), and the procedure was complete when the suture was pulled to the cervical os. The trainees performed the suturing procedure 10 times a day under hysteroscopy for 5 consecutive days, and the time of each suture procedure was recorded. During this training, we followed the technique of suturing the posterior wall of the uterus at the 5-o’clock position for needle insertion and the 7-o’clock position for needle extraction. Suturing of the anterior wall involved inserting the needle at the 11-o’clock position and exiting the needle at the 1-o’clock position. If the gynecologists needed to practice tying knots using a knotter, the knot was tied four to six times in vitro. Then, under hysteroscopy, they snipped the suture 1 cm below the knot.^5,12

Figure 2.

Simulated hysteroscopic suture procedure. (a) Entry and exit points of the needle in the uterine cavity. (b)–(g) Suture process of the posterior wall of the uterus: 5 o’clock in and 7 o’clock out with the needle (denoted as 5–7). (h) Suture process of the anterior wall of the uterus: 11 o’clock in and 1 o’clock out with the needle (denoted as 11–1) and (i) The suture is too deep to see the needle tip.

Statistical analysis

The cumulative sum (CUSUM) was used to evaluate the learning curve. ¹³ The CUSUM was calculated as follows: CUSUM $={\sum }_{i=1}^n(Ti-Tm)$ , where n represents the number of sutures, $\mathit{Ti}$ represents the time of each suture, and Tm represents the average suture time. Statistical analysis was performed with SPSS Version 23.0 software (IBM Corp., Armonk, NY, USA). Continuous variables are expressed as mean ± standard deviation. Differences between subsets were evaluated with the independent-samples t-test, and differences between proportions were compared with the χ² test or Fisher’s exact test. Regression analysis was used to evaluate the relationship between variables and consecutive cases. Analysis of variance was used to compare the mean of each value among the groups. A P-value of <0.05 was considered statistically significant.

Results

Three doctors participated in the training. Doctor A had never performed either hysteroscopic suture surgery or single-port laparoscopic surgery. Doctor B had not performed hysteroscopic suture surgery but had independently performed single-hole laparoscopic surgery. Doctor C had performed 20 cases of hysteroscopic suture surgery and had independently carried out single-port laparoscopy. All three doctors completed 50 posterior uterine sutures. After Dr. C had completed 50 sutures on the posterior wall, he added 50 sutures on the anterior wall (Group D). The means suture time in each group was 71.54 ± 68.158 seconds in Group A, 50.10 ± 28.060 s in Group B, 34.04 ± 10.457 s in Group C, and 30.38 ± 8.734 s in Group D. The difference between Groups C and B was statistically significant (P = 0.0003). The difference between Groups B and A was also statistically significant (P = 0.0424) (Figure 3). There was no statistically significant difference between Groups C and D. The number of sutures was used as the abscissa, and the suture time was used as the ordinate to draw scatter plots. Simulation curves were created using the number of sutures as the abscissa and CUSUM as the ordinate (Figure 4). The best fitting models of the CUSUM curves were as follows:

Figure 3.

Distribution of hysteroscopic suture times among the three doctors. Groups A, B, and C sutured the posterior wall of the uterus, while Group D sutured the anterior wall of the uterus.

Figure 4.

Scatter plot and CUSUM learning curve. CUSUM, cumulative sum.

A (R² = 0.971, P = 0.000): y = −326.0471 + 172.7683x − 6.3303x² + 0.0606x³

B (R² = 0.916, P = 0.000): y = 176.2668 − 1.9843x + 0.3265x² − 0.0072x³

C (R² = 0.805, P = 0.000): y= −21.6256 +14.2525x − 0.5271x² + 0.0050x³

Each curve showed two parts: the rising part at the front and the falling part at the back, with peaks located in the 19th case (A), 27th case (B), and 18th case (C).

Discussion

The CUSUM learning curve has been used to evaluate the learning difficulty of many types of surgical procedures.^14–18 A systematic review of six articles involving 250 gynecologists (32.4% experts and 67.6% doctors) participating in prospective studies concluded that virtual gynecological simulators improved the surgical skill of both novice and experienced surgeons, potentially reducing the complication rates of gynecological procedures. ¹⁹

We evaluated the learning curve of hysteroscopic sutures using the CUSUM method. We evaluated the effectiveness of hysteroscopic suture surgery techniques in helping gynecologists successfully complete the operation. Through training and practice, the time required for the three doctors to suture under hysteroscopy was reduced. The rising part of the learning curve is considered the initial learning stage, during which the surgical time is longer and there is an increased risk of serious complications such as fluid overload.^20,21 In particular, patients who undergo hysteroscopic suturing often have uterine enlargement, deformation caused by adenomyosis, or endometrial polyps, and the corresponding surgical time will be further extended. The downward part of the curve is considered a stage of ability consolidation, during which the suture time is shortened. Doctor A’s ability tended to stabilize after hysteroscopic suturing in the 19th case. It is worth noting that Doctor B, who had experience in single-port laparoscopic surgery, needed 27 cases to achieve stability after hysteroscopic suturing. However, the average suturing time of Doctor B was significantly shorter than that of Doctor A, who had no experience in single-hole laparoscopic surgery. This may have been related to the need for one-handed needle clamping operations during single-hole laparoscopic suturing, and skilled one-handed operations may reduce the time required for needle clamping and adjustment. Among the 50 simulated hysteroscopic sutures performed by Doctor C, the time taken (mean of 34.04 ± 10.457 s) was significantly shorter than that of Doctor B. This may have been related to Doctor C’s previous experience in single-port laparoscopic surgery and having successfully performed 20 cases of hysteroscopic sutures in the clinical setting. He therefore may have mastered more uterine surgical techniques. The curve slowly decreased after reaching the 18th case, indicating that after 18 attempts, it was difficult to acquire more surgical skills and that the surgical abilities of the gynecologists had reached their limits. These three learning curves indicate that different doctors have varying numbers of cases and times required to achieve stable suturing under hysteroscopy. Doctors with experience in single-port laparoscopic surgery or hysteroscopic suture surgery can achieve a significantly shorter surgical time. The learning curves of surgeons with different backgrounds of experience in hysteroscopic surgery vary, and experienced surgeons perform better using simulators. ¹⁴ In this training, Doctor C completed 50 sutures of the posterior wall of the uterus and then performed another 50 sutures of the anterior wall of the uterus (Group D). The results showed that although the mean time for Group D was shorter, there was no statistically significant difference compared with Group C.

Three factors that affect the suture time of the simulated hysteroscopy suture procedures can be summarized as follows. First, the orientation of the needle tip may affect the speed of needle adjustment in the needle holder. If the posterior wall of the uterus is sutured and the needle tip is facing toward the right side of the patient, it is beneficial for intraoperative needle holding. The simulation of hysteroscopic suturing showed that the shortest suturing time for each of the three doctors was 16 s for Doctor A, 15 s for Doctor B, and 19 s for Doctor C. This may have been because the direction of the suture needle in the uterine cavity was conducive to needle holding and that no unnecessary suturing action occurred during the operation. Second, the posterior wall of the uterus may be easier to suture than the anterior wall. As shown in Figure 2(a), the posterior wall of the uterus can be sutured from the 5-o’clock direction, exiting the needle from the 7-o’clock direction. This suture method is in line with the suturing habits of most doctors and may become faster with additional practice time. When suturing the anterior wall of the uterus (Figure 2(h)), if the needle holding method is the same as that of the posterior wall, the needle holder should simply be rotated 180 degrees. After completing 50 sutures of the posterior wall of the uterus, Doctor C performed another 50 sutures of the anterior wall of the uterus (Group D). The results showed that the mean time in Group D was shorter, but the difference was not statistically significant. This indicates that the doctor has fully grasped the suture technique under hysteroscopy. Third, the depth and frequency of sutures also affect the surgical time. If the suture is placed too deeply during the surgery, as shown in Figure 2(i), it will not be possible to smoothly remove the needle. The needle needs to be returned and sutured again a second time, and the surgeon may be concerned about the risk of uterine perforation. If the suture is too shallow, it will need to be sutured again because it cannot be fixed to the myometrium.

This emerging minimally invasive procedure is still in the early stages of research and learning, and additional basic research is required before it can become a mature clinical technique. This experiment was a simulated uterine suture, which has some differences from the actual intrauterine suture procedure in humans.

Conclusion

We used the CUSUM method to analyze a simulated learning curve of hysteroscopic suturing and found that gynecologists who had not performed this surgery have achieved proficiency and stability after 19 training cases. Doctors with experience in single-hole laparoscopic surgery or hysteroscopic suture surgery may significantly shorten the surgical time of hysteroscopic suturing.

Data availability statement

The data analyzed in this study are available from the corresponding author upon reasonable request.

Declaration of conflicting interests

The authors declare that there is no conflict of interest.

Ethics and consent

The requirement for ethics approval was waived because this study protocol involved only in vitro model training. All participants provided written informed consent.