Why not vaginal?-Nationwide trends and surgical outcomes in low-risk hysterectomies: A retrospective cohort study

Johanna Wagenius and Andrea Stuart – design of study, applying for ethical approval, data preparation, statistical analysis, clinical interpretation, writing and amending manuscript, submitting for publication. Karin Källén – data preparation, statistical analysis, amending manuscript. Sophia Ehrström and Jan Baekelandt – design of study, clinical interpretation, amending manuscript. All authors contributed to the interpretation of the data and read and approved the final version of the article for publication.

The study was approved on September 21, 2022, by the Swedish Ethical Review Authority (project number 2022–04358‐01).

The study was supported by grants from the Stig and Ragna Gorthon foundation.

Table  1 shows patient demographics depending on the mode of hysterectomy. Compared to vaginal hysterectomy, there were significant differences in patient demographics in the AH‐, LH‐, and RH‐groups. Patients in the VH‐group had generally lower BMI, more vaginal deliveries, and had less previous abdominal surgery (including caesarean section). To minimize bias due to the chosen surgical techniques, the hysterectomies were divided into a low‐risk group including patients with no previous CS, no previous abdominal surgery, BMI 30 and/or nullipara. Characteristics by surgical techniques. The rate of abdominal hysterectomies and vaginal hysterectomies decreased between 2014 and 2023. The distribution of surgical techniques is shown in Figure  2 . In the study, 3427 patients were operated on vaginally, of which 15.9% were vNOTES hysterectomies ( n  = 545). RH was the most common surgical technique 2021–2023 ( n  = 1816) followed by LH. In the low‐risk group, 25.2% of the patients were operated on vaginally. Additionally, 30.4% were operated on laparoscopically, 23.0% robotically, and 21.4% abdominally. In the standard group, 14.9% were operated on vaginally. Also, 32.7% were operated on laparoscopically, 25.5% robotically, and 26.8% abdominally. Distribution of surgical techniques 2014–2023. The mean surgical time has not changed over time. The mean surgical time was 106 min for AH, 118 min for LH, and 91 min for RH. VH had the shortest mean surgical time, 67 min. VH and vNOTES hysterectomies had about the same surgical time. Table  2 shows surgical outcomes and complications in the low‐risk group, and Table  3 surgical outcomes and complications in the standard group depending on hysterectomy technique. Complications and surgical outcomes in the low‐risk group. Abbreviations: AH, Abdominal hysterectomy; LH, Laparoscopic hysterectomy; RH, Robot assisted hysterectomy; VH, Vaginal hysterectomy. As compared to 45–90 min. Complications and surgical outcomes in standard group. Abbreviations: AH, Abdominal hysterectomy; LH, Laparoscopic hysterectomy; RH, Robot assisted hysterectomy; VH, Vaginal hysterectomy. As compared to 45–90 min. Comparing VH and AH, vaginal hysterectomies had better surgical outcomes in both the low‐risk group and the standard group, except for intraoperative complications where there were no differences between VH and AH. Abdominal hysterectomies had longer surgical times and longer hospital admissions. In the low‐risk group, laparoscopic hysterectomy had less severe intraoperative complications, ARR = 0.38 (95% CI 0.17–0.86), but more mild postoperative complications, ARR = 1.24 (95% CI 1.05–1.46), compared to vaginal hysterectomy. There were no differences in severe postoperative complications, bleeding >500 mL, and reoperations in LH versus VH. LH had more conversions, ARR = 1.46 (95% CI 1.00–2.12), longer surgical time, ARR = 2.73 (95% CI 2.46–3.00) and longer hospital admission, ARR = 1.26 (95% CI 1.12–1.43) compared to VH. In the standard group, there were no significant differences concerning intra/postoperative complications, reoperations, and bleeding >500 mL LH versus VH. LH had more conversions, ARR = 1.62 (95% CI 1.23–2.12); longer surgical time, ARR = 2.22 (95% CI 2.04–2.43); and longer hospital admission, ARR = 1.31 (95% CI 1.17–1.47), compared to VH. Mild (ARR = 0.33 (95% CI 0.16–0.66)) and severe intraoperative complications (ARR = 0.17 (95% CI 0.05–0.58)) and bleeding >500 mL (ARR = 0.12 (95% CI 0.04–0.34)) were less common comparing RH versus VH in the low‐risk group. The absolute rate was very low (severe intraoperative complications 1.0% and bleeding >500 mL 2.0% in the VH‐group). RH also had fewer conversions compared to VH, ARR = 0.40 (95% CI 0.22–0.72). There were no differences between RH and VH regarding postoperative complications and reoperations. Surgical time <45 min was less common in the RH group, ARR = 0.47 (95% CI 0.42–0.54) and RH had a significantly longer postoperative hospital stay, ARR = 1.16 (95% CI 1.02–1.33). Comparing RH versus VH in the standard group shows the same significant results as in the low‐risk group; besides that, there were no differences concerning severe intraoperative complications between the two methods. Table  4 shows estimated costs for each surgical technique. AH had lower perioperative costs but was the most expensive method due to a longer hospital stay. RH was the most expensive minimally invasive method, and VH was the cheapest method. vNOTES was cost neutral to vaginal hysterectomy if a sealing device was used for VH. The costs were lower in the low‐risk group due to shorter surgical time and hospitalization, but the distribution did not differ compared to the whole cohort. Estimated costs by surgical techniques. Abbreviations: AH, Abdominal hysterectomy; LH, Laparoscopic hysterectomy; RH, Robot assisted hysterectomy; VH, Vaginal hysterectomy.

Our study shows a decline in AH and VH over the last decade in Sweden among benign hysterectomies with prerequisites for vaginal surgery. During the same period, there was an increase in LH and RH. VH should be the preferred method in this cohort but was the least chosen technique. VH had the shortest surgical time and was the cheapest method. VH had few postoperative complications and short hospitalization. Robotic hysterectomies had a 71% higher cost than VH without considering the investment cost and yearly maintenance. According to the Cochrane review from 2023 regarding the recommended technique for benign hysterectomy, vaginal hysterectomy is the method of choice. When VH is not feasible, LH has multiple advantages over AH, but a slightly increased risk of ureteric injury. Evidence is limited for RH and vNOTES. vNOTES is a new technique that shows promising results. 20 , 21 It was first introduced in Sweden in 2021. The technique is cost‐neutral to VH and allows the removal of adnexa if needed. Further studies will show if vNOTES is as effective as VH, LH, and RH. Recently, a study protocol for a large multicenter RCT comparing VH, vNOTES, and LH has been published. 22 Hysterectomies due to the indications of prolapse and endometriosis were excluded from the study. Prolapse surgery is almost exclusively performed vaginally, is considered to be surgically easy, and would give a false benefit to vaginal hysterectomy if included in the cohort. On the other hand, hysterectomies due to endometriosis are almost never performed vaginally, are considered technically challenging, and will give a misleading increased complication rate for non‐vaginal hysterectomy. Our results suggest that surgeons tend to choose patients with lower BMI, higher parity, and no previous CS for a vaginal hysterectomy. Therefore, in a second step, we selected a group with BMI <30, no nulliparous women, no previous CS, and no previous abdominal surgery. These parameters are commonly used when the surgeon decides on a hysterectomy technique. In this group with optimal conditions for vaginal hysterectomy, only 25.2% of the hysterectomies were performed as VH, and in 2021–2023, only 20% of the hysterectomies were performed vaginally. There was a shift over time from the most minimally invasive method (VH) to less minimally invasive methods (LH and RH). The reduction in AH over the time period was equivalent to the reduction in VH. During the last decade, a regression to less minimally invasive methods has been seen among low‐risk hysterectomies in this cohort. The reason for this is unclear, but obviously, the surgical training for vaginal hysterectomies is declining to the benefit of laparoscopic and robotic training. VH had better surgical outcomes compared to AH and LH in the low‐risk group. Compared to RH, there was an increased risk of intraoperative complications and bleeding >500 mL in the VH group, but the absolute rate was very low. There was no difference in postoperative complications, and VH had quicker discharge than RH. We are inclined to believe that intraoperative complications of significance usually lead to postoperative complications. The significantly shorter surgical time and hospital stay in VH compared to RH also indicate that the difference in intraoperative complications registered in the GynOp registry is of small clinical relevance. Another risk factor for VH is a history of multiple conizations due to cervical dysplasia, which presents a high risk for bladder injury. No analysis concerning previous conizations is done in this study since the data is not available in the registry. A recent large study including 400 000 hysterectomies in the USA 7 showed that VH had the lowest inpatient total cost and RH had the highest total cost. They concluded that complication rates were low in all minimally invasive hysterectomies, suggesting that it was unlikely that the robotic approach provided an appreciable improvement in perioperative outcomes. In our study, 23.0% of the patients were operated on robotically and 30.4% laparoscopically in the low‐risk group. Besides being more costly, it also gave these patients an unnecessary abdominal wall incision. All procedures with abdominal entrance versus vaginal entrance will give some defect to the fascia and abdominal wall structures. Complications due to abdominal wall defects, such as hernia, neuralgia, adhesions, and strangulation of intestines, can present themselves years after surgery. 23 Future studies are needed to address and analyze this problem after gynecological surgery. Lonnerfors et al. performed a RCT in Sweden, comparing the cost between VH or LH with RH. 24 In the study, they initially excluded the purchase price of the da Vinci Si system robot, as it was considered to be a pre‐existing investment. When the purchase price of the robot was included, the cost of a robotic hysterectomy was 2600 USD more expensive than VH and/or LH. Swenson et al. showed that non‐robotic (laparoscopic or vaginal) routes had a 24% lower cost compared to the robotic‐assisted route and had no increase in clinical complications. 17 Bijen et al. compared the costs between LH and AH and concluded that the shorter hospital stay in the LH group and lower morbidity compensate for the increased procedure costs. 9 Furthermore, Lonnerfors et al. showed that the determining factor regarding hospital cost was surgical time; therefore, we chose to mainly focus on costs due to surgical time in our cost overview. In Sweden, approximately 30 gynecological clinics perform robotic surgery. 1 The initial investment cost of the robot is usually, but not always, shared between other specialties, and therefore not included in this study. Surgical time is commonly calculated as the time from incision to the last stitch. In our study, VH had the shortest surgical time. An alternative way to calculate operation time and cost is how many hysterectomies of a certain type can be performed per day in an operating room. In the GynOp registry, we have access to surgical time, but not the total time spent in the OR. A better calculation could have been performed regarding the cost of a hysterectomy per route if we had information on the total OR time per procedure. This study is strengthened by being a large cohort study (almost 18 000 hysterectomies) with prospectively collected data from the nationwide GynOp registry. Most variables are compulsory in the registry, and we excluded hysterectomies that had missing data. A weakness of our study is that it is a cohort study rather than an RCT, with inherent bias to the chosen surgical technique. As we have selected a low‐risk hysterectomy group, many of the problems with selection bias should be dealt with. Given the high incidence of hysterectomies performed in the world and the numbers of patients that could possibly be included in RCTs, there are, in fact, not many large RCTs comparing techniques for hysterectomy. The latest Cochrane review is based on 63 RCTs within a total of only 6811 patients. 5 Another weakness is that the experience of the surgeon has not been taken into account. In Sweden, RH is usually performed by experienced surgeons with a higher yearly volume compared to LH and VH. 1 That could be an explanation for the differences seen in intraoperative complications mainly between LH and RH. Robotic surgeons have a compulsory surgical education before they are allowed to do surgery with the robot. Doctors in training usually do not perform robotic surgery, but they can perform hysterectomies with the other techniques in Sweden.

This study shows a decline of vaginal hysterectomies in Sweden 2014–2023 among patients with prerequisites for vaginal surgery. VH is a cheap and safe method with few postoperative complications and short hospitalization. Our results support the vaginal route for low‐risk hysterectomies, both considering surgical outcomes and costs.

Hysterectomy is the most common surgical procedure performed in women next to caesarean section, and in Sweden about 4000 hysterectomies are performed annually. 1 The primary indications are abnormal uterine bleeding, symptomatic uterine fibroids, or cervical dysplasia. 2 Hysterectomies can be performed as an abdominal hysterectomy (AH) or minimally invasively as a vaginal hysterectomy (VH), laparoscopic hysterectomy (LH), or robot‐assisted hysterectomy (RH). In recent years, a new technique has been introduced that combines vaginal hysterectomy and endoscopy via the vagina: vNOTES (vaginal Natural Orifice Transluminal Endoscopic Surgery). Globally, the incidences of both abdominal and vaginal hysterectomies are decreasing to the benefit of laparoscopic and robot‐assisted hysterectomies. 3 , 4 The medical gain of a reduction in AH is clear, but the shift from performing vaginal surgery to robotic surgery is not. 5 , 6 Furthermore, there are distinct differences in financial costs depending on the route of hysterectomy. 7 , 8 , 9 , 10 , 11 , 12 , 13 , 14 Selection bias complicates the analysis of outcomes depending on the technique of hysterectomy; only the easiest (and therefore probably cheapest) hysterectomies are performed vaginally, and the more complicated hysterectomies are performed robotically, laparoscopically, or abdominally. 15 If feasible, the recommended route of hysterectomy is vaginal due to quicker recovery and lower costs. 5 , 16 Previous studies have shown that hysterectomies that have prerequisites of being performed vaginally instead are being performed robotically or laparoscopically. 17 , 18 , 19 The primary objective of this study was to describe the surgical techniques used for benign hysterectomies with prerequisites of being performed vaginally: uterus weight under 300 grams, no endometriosis, and no prolapse. Our secondary objective was to identify a low‐risk group of patients with optimal conditions for a vaginal hysterectomy and present the surgical outcomes and costs depending on the route of hysterectomy used.

Johanna Wagenius, Andrea Stuart, and Jan Baekelandt declare consultancy for Applied Medical.

Data were extracted from the Swedish quality register GynOp between the years 2014 to 2023. Inclusion criteria were elective benign, non‐prolapse, non‐endometriosis hysterectomies with uterus weight less than 300 grams. In the study, 17 804 hysterectomies were included; see the flowchart presented in Figure  1 . The hysterectomies were divided into a low‐risk group including patients with no previous caesarean section (CS), no previous abdominal surgery, BMI 30 and/or nulliparity. Flowchart. The GynOp Registry contains pre‐, intra‐, and post‐operative information regarding gynecological surgery since 1997. Data are collected from both the surgeon and the patient. Data entry is compulsory for surgeons, and the dataset is a complete set for the country with 90%–93% coverage during the period studied. The patient fills in a preoperative questionnaire in order to collect baseline demographic data and also fills in a postoperative questionnaire 8 weeks and 12 months after surgery regarding patient‐reported complications. The surgeon confirms or rejects the patient‐reported complications in the registry. Demographic variables are filled in at the time of surgery (age, BMI, tobacco use, previous abdominal surgery, previous CS), type of surgery (hysterectomy and/or adnexal surgery), primary incision (abdominal, laparoscopic, vaginal, and conversion to laparoscopy/laparotomy), and indication for surgery. Postoperatively, the surgeon registers data regarding uterus weight, blood loss, surgical time, and complications according to Clavien Dindo. Clavien Dindo I–II are considered mild complications, and Clavien Dindo III–IV are considered severe complications. Hysterectomies were divided into abdominal, vaginal, laparoscopic (including laparoscopically assisted vaginal hysterectomy, LAVH, and total laparoscopic hysterectomy, TLH), and robot‐assisted hysterectomy. Vaginal hysterectomy and vNOTES hysterectomy were included in the group vaginal hysterectomies. The cost of the equipment, instruments (ports, optics, disposables), and sterilization were calculated for every type of hysterectomy. The cost of a sealing device was included in the cost for VH and RH but not for AH and LH, as this is commonly used in VH and RH but not always in AH and LH. Data on costs were extracted from the financial department at Helsingborg Hospital, Sweden. Helsingborg hospital is a tertiary teaching hospital in southern Sweden, with approximately 3400 deliveries and 250 in‐patient beds. The Department of Gynecology performs only benign surgery and is among the five clinics in Sweden with the highest volume of benign gynecological surgery. The Swedish health‐care system is government‐funded and universal for all citizens, and the vast majority are regionally owned public hospitals. For all procedures, the cost of personnel, anesthesia, draping, and antibiotics was the same. The per‐minute charge for use of the operating room was based on the previous year's mean institutional cost and includes the operating room, OR staff, one gynecological surgeon, cleaning, and basic expendables such as gowns and gloves and was 17.4 Euro per minute. The cost for one overnight stay at the gynecological ward was 1224 Euro (year 2021). The initial investment cost of the robot and yearly maintenance were not included in the cost calculation. Chi‐square tests were used to evaluate heterogeneity between groups displayed in descriptive tables. The risk of miscellaneous complications and other evaluated binary outcomes after vaginal hysterectomies was compared to the corresponding outcomes after abdominal, laparoscopic, or robot‐assisted hysterectomies, respectively. The results were stratified by subpopulation (low‐risk or standard group). Crude and adjusted Risk Ratios (RR, ARR) were obtained by modified Poisson‐regression analyses. Adjustments were made for possible confounders with a pre‐assumed association with both operation method and the miscellaneous outcome measures. The variables included in the multiple models were: maternal age (years, continuous), parity (0/1+), BMI (kg/m 2 , continuous), smoking (yes/no), and uterus weight (g, continuous). Missing data for smoking were rare (0.4%), and women without information about smoking were assumed to be non‐smokers. A p ‐value <0.05 was considered significant. Data were analyzed with the statistical package SPSS version 27, IBM Corp., NY, USA.