Early-stage endometrial cancer can be cured surgically using three techniques: laparotomy, laparoscopic surgery, and robotic-assisted surgery. In this retrospective study, we used 4 years of data from the Japanese Diagnosis Procedure Combination to analyze perioperative sequelae for each surgical procedure. Patients with early-stage endometrial cancer were classified into three groups: laparotomy, laparoscopic surgery, and robotic-assisted surgery. The number of robotic-assisted surgeries is increasing, but hospitals with low surgical volumes performed laparotomy at a rate of 57.6%, while hospitals with higher surgical volume tended to perform fewer laparotomies and more laparoscopic and robotic-assisted surgeries. Compared with the laparotomy group, the in-hospital risk ratios and 95% confidence intervals for postoperative sequelae were 0.28 (0.18–0.44) and 0.39 (0.23–0.69) in the laparoscopic and robotic-assisted surgery groups, respectively (P = 0.001 for all). Regarding blood transfusion treatment, the incidence rate ratios were lower for laparoscopic and robotic-assisted surgery than for laparotomy (P < 0.001 for all), even after multivariable analysis. In conclusion, in early-stage endometrial cancer, laparotomy had the highest incidence of perioperative sequelae compared to laparoscopic surgery and robotic-assisted surgery.
Endometrial cancer (EC) originates from the endometrium and is a common cancer of the female reproductive organs. EC is more likely to occur in postmenopausal women with a median age at diagnosis of 65 years, approximately 4% of whom are younger than 40 years, and approximately 70% of whom have never given birth [1]. Based on Japan’s national cancer statistics, 68.2% of endometrial cancer cases are diagnosed at an early stage [2]; therefore, the 5-year survival rate is 80–85% [3]. According to the National Comprehensive Cancer Network guidelines, the primary treatment for early-stage EC (FIGO stage 1–2) is total hysterectomy with salpingo-oophorectomy [4].
The morbidity and mortality rates of EC have increased in recent years, particularly in developed countries. The number of laparoscopic and robotic-assisted surgeries, known as minimally invasive surgeries, especially robotic-assisted surgeries, is expected to continue to increase. Recently, approximately 80% of hysterectomies in patients with cancer have been performed using robotic-assisted surgery in the United States [5]. However, robotic-assisted surgery for EC in Japan has a short history, with insufficient research conducted on its surgical outcomes and safety. On the other hand, laparoscopic surgery for early-stage EC has been covered by insurance since 2014, although its adoption as an advanced medical treatment had already been expanding even prior to that. Minimally invasive surgery, such as laparoscopic and robotic-assisted surgery, is preferred when technically feasible.
Compared to laparotomy, laparoscopic surgery offers the advantages of reduced blood loss and quicker postoperative recovery. Robotic-assisted surgery shows better perioperative outcomes in terms of reduced pain and complications than conventional laparoscopic surgery and laparotomy [6, 7]. Because early-stage EC can be cured by surgery, it is essential to understand the perioperative risks associated with the three surgical procedures and the trends in their occurrence.
It has been reported that the outcome of cancer surgery varies depending on the number of cases a hospital handles [8]. Because of the large number of hospitals, case numbers are widely dispersed across Japan. The electronic medical record system is not shared, making it impossible to grasp the number of surgeries performed at each hospital. Hence, observational studies using DPC data provide important information.
Randomized controlled trials (RCTs) have already demonstrated that laparoscopic and robotic-assisted surgery have a lower risk of perioperative complications for early-stage endometrial cancer than laparotomy [9, 10]. However, because such RCTs are reports from specific institutions, it would be useful to examine whether the benefits of laparoscopy and robotic-assisted surgery are evident in clinical practice, regardless of hospital case volume or region. There are no nationwide comparative studies that include small-scale hospitals, which account for a substantial portion of actual clinical practice.
This study examined the perioperative sequelae of laparotomy, laparoscopic surgery, and robotic-assisted surgery for new-onset and early-stage EC in Japan, utilizing data from the Diagnosis Procedure Combination (DPC) system.
Data Sources
This study was conducted as a retrospective observational cohort study using administrative data known as DPC data. Japan introduced the DPC system in 2003 following advances in information technology. The system was developed as a measuring tool to enhance transparency in acute inpatient care, to standardize Japanese medical practices, and to evaluate and improve the quality of medical services. The DPC is both a unique reimbursement system that records the medical procedures performed during hospitalization at the voucher level and a comprehensive payment system. The data for each inpatient includes patient information such as the patient’s date of birth and admission and discharge dates, as well as the principal diagnosis, sequelae, and comorbidities. This data is recorded in an administrative database linked to the DPC system, and then a discharge summary known as Form 1 is created. The DPC system contains medical performance data known as EF file, which records the performance of each medical procedure, such as anesthesia, surgery, prescriptions, and blood purification, using Japan’s Healthcare Payment System codes. The dates of each medical procedure and medication are also recorded using the corresponding codes. Approximately 90% of all acute care hospitals and all university hospitals in Japan are registered as eligible hospitals in the DPC database. Several papers have reported that DPC data are highly valid for cancer diagnosis [11, 12]. In 2022, of the 1,746 hospitals participating in the DPC system, 1,377 (79%) agreed to provide data for this study.
Participant Selection
Participants who were treated from April 1, 2018, to March 31, 2022, were enrolled. First, we extracted data on inpatients admitted with a diagnosis of EC from the diseases that received the most medical resources. EC was diagnosed based on the International Classification of Diseases for Oncology, third edition, site code C54 [13]. New-onset and surgical cases were extracted from that data. We excluded cases with missing variables, such as body mass index (BMI). In addition, we excluded patients who underwent chemotherapy during hospitalization in order to examine only perioperative complications.
We identified each surgery by the code the health care provider used for billing. Using the data in Form 1 of the DPC, for laparotomy, cases described as stage I or stage IA were identified as early-stage EC. Under the Japanese insurance system, laparoscopic and robotic-assisted hysterectomy is reimbursed only for early-stage endometrial cancer. Therefore, patients recorded as having undergone these procedures in the DPC database were regarded as early-stage cases. Patients with early-stage EC who had undergone either laparotomy, laparoscopic, or robotic-assisted surgery were included in this study (n = 36,291). The study selection process is illustrated in Figure 1.
Comorbidities of Inpatients
Hypertension, thrombosis, and diabetes were identified as comorbidities among patients using the International Classification of Diseases for Oncology, third edition. The codes are as follows: I802, I822, I10, E10, and E11.
Hospital Surgery Volume
The number of surgeries for early EC performed at each hospital during the 4-year observation period was calculated, and hospitals were divided into quartiles based on the number of surgeries performed. Quartile 1 represents hospitals with the lowest number of surgeries (<43), quartile 2 represents 43–70, quartile 3 represents 71–106, and quartile 4 represents the highest number of surgeries (>107).
Identifying Perioperative Sequelae and Medical Procedures
Postoperative complications were identified using the International Classification of Diseases for Oncology, third edition, codes if the sequelae included thrombosis, compartment syndrome, glaucoma code, or postoperative ileus. The codes are as follows: I802, I822, M6226, M6229, T796, K56, K913, and H40.
Using medical performance data from the DPC system, we identified ureterorrhaphy for surgical ureteral trauma and blood transfusion by the medical billing code and the code assigned to each drug.
Statistical Analysis
We performed an analysis using a Poisson regression with robust variance to investigate the relationship between surgical outcomes and surgical procedures. In this model, patients were nested within 376 facilities to control for the influence of each facility. All facilities included in the analysis were hospitals that performed any type of endometrial cancer surgery during the study period.
Surgical outcomes are not influenced solely by each surgical procedure; therefore, it was necessary to adjust for individual factors that are constitutive of the patient. Overall perioperative sequelae included ileus, thrombosis, surgical ureteral trauma, compartment syndrome, and glaucoma. Ileus and thrombosis were analyzed separately as sequelae outcomes. Surgical ureteral trauma, compartment syndrome, and glaucoma were not analyzed separately due to the very small number of cases. Patient-level factors were selected based on their potential association with outcomes and analyzed using a multivariate model incorporating patient-level factors.
The following potential confounding factors were selected for this model: age, BMI, number of surgeries per hospital, and comorbidities, including thrombosis, hypertension, and diabetes.
Data were analyzed using STATA/MP 17 (StataCorp LLC, College Station, Texas, USA).
Changes in the Number of Surgeries
Figure 2 shows trends in the number of surgeries for early-stage EC by surgical procedure each year. Robotic-assisted surgeries have been increasing since they were covered by insurance in FY2018. The number of robotic-assisted surgeries exceeded that of laparotomy surgeries in FY2019 and has continued to increase.
Basic Characteristics of the Study Participants
The basic characteristics of the study participants are shown in Table 1. We classified them into four groups based on their surgical volume for early-stage EC in the observation period. Approximately 58% of patients who had undergone laparotomy had been operated on at hospitals with a surgical volume <43. Approximately 74% of patients who had undergone robotic-assisted surgery had been operated on at hospitals with a volume of ≥71 surgeries. During the observation period, hospitals with higher surgical volume tended to perform fewer laparotomies and more minimally invasive procedures.
| Variables |
Laparotomy (n = 2,344) |
Laparoscopic surgery (n = 8,064) |
Robotic-assisted surgery (n = 3,096) |
|---|---|---|---|
| n (%) | n (%) | n (%) | |
| Age, mean (SD) | 59.7 (12.0) | 58.5 (12.0) | 56.7 (11.0) |
| 20–29 | 8 (0.3%) | 24 (0.3%) | 9 (0.3%) |
| 30–39 | 86 (3.7%) | 334 (4.1%) | 153 (4.9%) |
| 40–49 | 370 (15.8%) | 1,440 (17.9%) | 582 (18.8%) |
| 50–59 | 770 (32.9%) | 2,963 (36.7%) | 1,276 (41.2%) |
| 60–69 | 564 (24.1%) | 1,685 (20.9%) | 648 (20.9%) |
| 70–79 | 416 (17.8%) | 1,166 (14.5%) | 353 (11.4%) |
| >80 | 130 (5.6%) | 452 (5.6%) | 75 (2.4%) |
| BMI (kg/m2), mean (SD) | 25.2 (5.9) | 25.4 (9.7) | 26.1 (6.6) |
| >18.5 | 171 (7.3%) | 527 (6.5%) | 184 (5.9%) |
| 18.5–25 | 1,147 (48.9%) | 3.961 (49.1%) | 1,413 (45.6%) |
| >25 | 1,026 (43.8%) | 3,576 (44.4%) | 1,499 (48.4%) |
| Comorbidities | |||
| Thrombosis | 85 (3.6%) | 292 (3.6%) | 104 (3.3%) |
| High blood pressure | 384 (16.4%) | 915 (11.4%) | 304 (9.8%) |
| Diabetes | 306 (13.0%) | 907 (11.3%) | 383 (12.4%) |
| The number of surgeries per hospital in the observation period | |||
| <43 | 1,349 (57.6%) | 1,892 (23.5%) | 320 (10.4%) |
| 43–70 | 576 (24.6%) | 2,153 (26.7%) | 479 (15.5%) |
| 71–106 | 246 (10.5%) | 1,920 (23.8%) | 1,202 (38.8%) |
| >106 | 173 (7.4%) | 2,099 (26.0%) | 1,095 (35.4%) |
Length of hospitalization
Among the patients who underwent laparotomy, the mean length of hospitalization was 11.2 days (SD 4.6). In contrast, those who underwent laparoscopic surgery had a shorter mean hospital stay of 7.8 days (SD 4.0), and patients treated with robotic-assisted surgery showed a similar mean stay of 7.6 days (SD 4.0).
Outcome of Perioperative Sequela and Blood Transfusion
Table 2 presents the postoperative sequelae of patients who underwent each surgery. Ureterorrhaphy for surgical ureteral trauma occurred only in the laparotomy group (0.1%), while no cases were reported in the laparoscopic or robotic-assisted groups. Blood transfusion was significantly more common in laparotomy patients (6.1%) compared to laparoscopic (1.1%) and robotic-assisted surgery (1.1%). Immediate postoperative sequelae, such as glaucoma and compartment syndrome, were rare, occurring only in the laparoscopic (0.04%) and robotic-assisted groups (0.1%), with no cases in the laparotomy group. Postoperative complications, including thrombosis and ileus, were more frequent in the laparotomy group (1.3% and 2.7%, respectively) than in the laparoscopic (0.4% and 0.7%) and robotic-assisted groups (0.6% and 0.9%).
| Variables |
Laparotomy (n = 2,344) |
Laparoscopic surgery (n = 8,064) |
Robotic-assisted surgery (n = 3,096) |
|---|---|---|---|
| n (%) | n (%) | n (%) | |
| Intraoperative medical procedures | |||
| ureterorrhaphy for surgical ureteral trauma | 2 (0.1%) | 0 (0%) | 0 (0%) |
| blood transfusion treatment | 142 (6.1%) | 87 (1.1%) | 35 (1.1%) |
| Immediate postoperative sequelae | |||
| Glaucoma | 0 (0%) | 3 (0.04%) | 3 (0.1%) |
| compartment syndrome | 0 (0%) | 3 (0.04%) | 3 (0.1%) |
| Postoperative sequelae | |||
| Thrombosis | 31 (1.3%) | 31 (0.4%) | 17 (0.6%) |
| Ileus | 63 (2.7%) | 53 (0.7%) | 29 (0.9%) |
Table 3 presents the results of univariable and multivariable analyses examining the associations between postoperative sequela and surgical procedure, as well as between blood transfusion and surgical procedure. Compared to laparotomy, both laparoscopic and robotic-assisted surgeries were associated with significantly lower incidence rate ratios of overall perioperative sequelae. In the multivariable analysis, laparoscopic surgery showed an adjusted risk ratio (RR) of 0.28 (95% confidence interval [CI]: 0.18–0.44, P = 0.001), and robotic-assisted surgery had an adjusted RR of 0.39 (95% CI: 0.23–0.69, P = 0.001) after adjusting for age, BMI, number of surgeries per hospital, and comorbidities, including thrombosis, hypertension, and diabetes.
| Outcome | Surgery | Number of events (%) | Univariable | Multivariable* | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| RR | 95% CI | P | Adjusted RR | 95% CI | P | |||||
| Overall perioperative sequelae† | ||||||||||
| laparotomy | 93 (4.0%) | Reference | Reference | |||||||
| laparoscopic surgery | 88 (1.1%) | 0.27 | 0.18 | 0.42 | <0.001 | 0.28 | 0.18 | 0.44 | 0.001 | |
| robotic assisted surgery | 51 (1.7%) | 0.38 | 0.22 | 0.63 | <0.001 | 0.39 | 0.23 | 0.69 | 0.001 | |
| Postoperative thrombosis | ||||||||||
| laparotomy | 31 (1.3%) | Reference | Reference | |||||||
| laparoscopic surgery | 31 (0.4%) | 0.33 | 0.1 | 0.59 | <0.001 | 0.32 | 0.16 | 0.63 | 0.001 | |
| robotic assisted surgery | 17 (0.6%) | 0.47 | 0.22 | 1.03 | 0.06 | 0.46 | 0.19 | 1.09 | 0.078 | |
| Postoperative ileus | ||||||||||
| laparotomy | 63 (2.7%) | Reference | Reference | |||||||
| laparoscopic surgery | 53 (0.7%) | 0.22 | 0.12 | 0.40 | <0.001 | 0.22 | 0.12 | 0.41 | <0.001 | |
| robotic assisted surgery | 29 (0.9%) | 0.27 | 0.12 | 0.57 | 0.001 | 0.28 | 0.13 | 0.62 | 0.002 | |
| Blood transfusion treatment | ||||||||||
| laparotomy | 142 (6.1%) | Reference | Reference | |||||||
| laparoscopic surgery | 87 (1.1%) | 0.17 | 0.13 | 0.23 | <0.001 | 0.16 | 0.11 | 0.23 | <0.001 | |
| robotic assisted surgery | 35 (1.1%) | 0.17 | 0.1 | 0.26 | <0.001 | 0.15 | 0.09 | 0.26 | <0.001 | |
RR: risk ratio, CI: confidence interval, BMI: body mass index.
*Estimated by the multilevel logistic regression model after adjusting for age, BMI, number of surgeries per hospital, comorbidities including thrombosis, hypertension and diabetes
†Overall perioperative sequelae include surgical ureteral trauma, glaucoma, compartment syndrome, thrombosis, and ileus
For postoperative thrombosis, laparoscopic surgery significantly reduced the risk (adjusted RR: 0.32, 95% CI: 0.16–0.63, P = 0.001), while robotic-assisted surgery showed a non-significant trend toward reduction (adjusted RR: 0.46, 95% CI: 0.19–1.09, P = 0.078).
In addition, postoperative ileus was less frequent in minimally invasive approaches. Laparoscopic surgery had an adjusted RR of 0.16 (95% CI: 0.09–0.28, P < 0.001), and robotic-assisted surgery had an adjusted RR of 0.22 (95% CI: 0.11–0.44, P < 0.001).
The need for blood transfusion was markedly lower in both the laparoscopic (adjusted RR: 0.22, 95% CI: 0.12–0.41, P < 0.001) and robotic-assisted surgery groups (adjusted RR: 0.28, 95% CI: 0.13–0.62, P = 0.002) compared to laparotomy.
An increasing number of hospitals are introducing robotic-assisted surgery, which is expected to continue to replace laparotomy and laparoscopic surgery. However, few studies have compared patient background and perioperative complications in these three types of surgery for EC using a large dataset. In the present study, we used DPC data to investigate the perioperative sequelae of three surgical procedures.
Table 3 shows that, compared with the laparotomy group, the laparoscopic surgery group and the robotic-assisted surgery group had lower rates of thrombosis and postoperative ileus. Reports indicate that minimally invasive surgery reduces the risk of thrombosis [14]. Multivariate analysis showed that laparoscopic surgery, in particular, significantly reduced the risk of thrombosis. Furthermore, more invasive surgeries using laparotomy are associated with a more extensive ileus [15]. Laparotomy can be associated with a high risk of perioperative complications and excessive bleeding during surgery. There have been several reports that laparotomy is more prone to cause bleeding than minimally invasive surgery, and indeed, our data also showed that laparotomy showed the highest transfusion rate [16, 17].
Robotic-assisted surgery is associated with fewer complications and reduced intraoperative blood loss compared to laparotomy and laparoscopic surgery [18]. However, in the current study, rather than comparing laparoscopy and robotic-assisted surgery, we compared these two minimally invasive surgeries separately with laparotomy. The volume of robotic-assisted surgery is expected to grow faster than that of laparoscopic surgery because surgical proficiency in robotic-assisted surgery is generally faster than in laparoscopy [19, 20].
Our multivariate analysis suggests that minimally invasive surgery has a lower risk of complications than laparotomy. Because robotic-assisted surgery requires steep Trendelenburg positioning and long surgery times, there are concerns about the effects on the respiratory and cardiovascular systems in elderly patients [21], although many reports have demonstrated the safety of robotic-assisted surgery in older patients [21, 22].
Robotic surgery will become more widespread among the older population in the future, and the number of minimally invasive surgeries will likely increase. Currently, the proportion of minimally invasive surgeries for early-stage endometrial cancer is higher in high-volume hospitals.
Several studies have shown that hospitals with high surgical volumes have lower mortality and postoperative complication rates [23, 24]. One approach to increasing the number of minimally invasive surgeries performed in hospitals is to centralize cancer surgery within hospitals with multidisciplinary teams [25, 26]. Our data, showing that minimally invasive surgery affects short-term postoperative outcomes, can contribute to the discussion of the centralization of cancer surgery.
This study has some limitations. First, the data does not cover all surgical cases of early-stage endometrial cancer in Japan. For example, regarding early-stage endometrial cancer in 2021, the DPC data surveyed in this study accounted for approximately 30.5% of the national cancer statistics. This is due to the exclusion of cases with missing variables such as stage and BMI. Another reason is that some hospitals that are not DPC-eligible perform surgery for early-stage endometrial cancer, though these facilities are typically low-volume. Furthermore, it was difficult to accurately identify cases in which laparotomy was chosen because minimally invasive surgery was predicted to be impossible preoperatively, or cases in which minimally invasive surgery was converted to laparotomy due to excessive bleeding or sequelae during surgery. Second, although comorbidities and sequelae may include suspected disease names, there are no studies or literature showing differences between suspected and definitive diagnoses. Third, factors related to postoperative complications include comorbidities such as dyslipidemia, smoking history, and Activities of Daily Living (ADL), but these data were not collected and could not be analyzed as confounding factors. Although indices such as the Charlson Comorbidity Index and the Elixhauser Index allow for more detailed adjustment, confounding by comorbidities was less applicable in our study because we focused on short-term outcomes closely related to the surgical approach. This study lacked detailed person-time information; consequently, incidence rates could not be estimated, and the results should be interpreted with caution when considering time-at-risk or incidence-based outcomes. A future task will be to verify the differences between gynecological oncology data held by the Japan Society of Obstetrics and Gynecology and DPC data.
Regarding the perioperative sequelae for early-stage EC, laparotomy had the highest incidence of sequelae compared to laparoscopic surgery and robotic-assisted surgery. As robotic-assisted surgery is currently on the rise, a comparison between laparoscopy and robotic-assisted surgery is a future challenge.
This study was funded by the Ministry of Health, Labour, and Welfare, Japan (grant number:24AA2006).
None.
The datasets used and/or analyses conducted during the current study are not publicly available for ethical reasons. The datasets are available from the corresponding author upon reasonable request, provided that ethical approval has been obtained.
This study was approved by the Institutional Review Board of the University of Occupational and Environmental Health, Japan (approval number R4-045), which waived the requirement for obtaining informed consent from the participants. All experiments were performed in accordance with relevant guidelines and regulations.
