Nationwide Introduction of Minimally Invasive Robotic Surgery for Early-Stage Endometrial Cancer and Its Association With Severe Complications

Key Points

Question  Is the nationwide introduction of minimally invasive robotic surgery associated with a decreased risk of severe complications in patients with early-stage endometrial cancer?

Findings  A Danish nationwide cohort of 5654 women with early-stage endometrial cancer was divided into 2 groups based on the time of the introduction of minimally invasive robotic surgery in their region. The risk of severe complications was significantly reduced in the group undergoing surgery after the introduction of minimally invasive robotic surgery.

Meaning  The national implementation of minimally invasive robotic surgery was associated with an increased proportion of minimally invasive surgical procedures, which translated into a reduced risk of severe complications in women with early-stage endometrial cancer.

Importance  Minimally invasive laparoscopic surgery (MILS) for endometrial cancer reduces surgical morbidity compared with a total abdominal hysterectomy. However, only a minority of women with early-stage endometrial cancer undergo MILS.

Objective  To evaluate the association between the Danish nationwide introduction of minimally invasive robotic surgery (MIRS) and severe complications in patients with early-stage endometrial cancer.

Design, Setting, and Participants  In this nationwide prospective cohort study of 5654 women with early-stage endometrial cancer who had undergone surgery during the period from January 1, 2005, to June 30, 2015, data from the Danish Gynecological Cancer Database were linked with national registers on socioeconomic status, deaths, hospital diagnoses, and hospital treatments. The women were divided into 2 groups; group 1 underwent surgery before the introduction of MIRS in their region, and group 2 underwent surgery after the introduction of MIRS. Women with an unknown disease stage, an unknown association with MIRS implementation, unknown histologic findings, sarcoma, or synchronous cancer were excluded, as were women who underwent vaginal or an unknown surgical type of hysterectomy. Statistical analysis was conducted from February 2, 2017, to May 4, 2018.

Exposure  Minimally invasive robotic surgery, MILS, or total abdominal hysterectomy.

Main Outcomes and Measures  Severe complications were dichotomized and encompassed death within 30 days after surgery and intraoperative and postoperative complications diagnosed within 90 days after surgery.

Results  A total of 3091 women (mean [SD] age, 67 [10] years) were allocated to group 1, and a total of 2563 women (mean [SD] age, 68 [10] years) were allocated to group 2. In multivariate logistic regression analyses, the odds of severe complications were significantly higher in group 1 than in group 2 (odds ratio [OR], 1.39; 95% CI, 1.11-1.74). The proportion of women undergoing MILS was 14.1% (n = 436) in group 1 and 22.2% in group 2 (n = 569). The proportion of women undergoing MIRS in group 2 was 50.0% (n = 1282). In group 2, multivariate logistic regression analyses demonstrated that a total abdominal hysterectomy was associated with increased odds of severe complications compared with MILS (OR, 2.58; 95% CI, 1.80-3.70) and MIRS (OR, 3.87; 95% CI, 2.52-5.93). No difference was found for MILS compared with MIRS (OR, 1.50; 95% CI, 0.99-2.27).

Conclusions and Relevance  The national introduction of MIRS changed the surgical approach for early-stage endometrial cancer from open surgery to minimally invasive surgery. This change in surgical approach was associated with a significantly reduced risk of severe complications.

Several randomized clinical trials have confirmed that minimally invasive laparoscopic surgery (MILS) for early-stage endometrial cancer is associated with reduced morbidity compared with a total abdominal hysterectomy (TAH).^1-9 Globally, the MILS rate for endometrial cancer is low, which is most likely related to the steep learning curve in complex oncologic surgical procedures.^10-15 Minimally invasive robotic surgery (MIRS) was approved for treatment of gynecologic conditions in 2005.¹⁶ The technique offers enhanced visualization, movements, and ergonomics and was rapidly accepted by surgeons worldwide. Data from highly specialized institutions suggest that the implementation of MIRS increases the use of minimally invasive surgery (MIS).^17-21

In Denmark, the treatment of early-stage endometrial cancer was gradually centralized from 28 departments in 2005 to 6 national cancer centers in 2012. During the first years of the centralization process, only patients with high-risk histologic characteristics were referred, followed by a transition to the referral of the entire group as requested by the National Board of Health. Each center implemented MIRS between 2008 and 2013.

In the present population-based study, we aimed to evaluate whether the nationwide introduction of MIRS was associated with a decreased risk of severe complications in women with early-stage endometrial cancer.

This is a study of nationwide registers based on prospectively obtained data. In the design, we take advantage of the “natural experiment” that took place when MIRS was introduced nationwide in Denmark. The introduction took place gradually across the 6 national cancer centers, which, to some extent, mimics a multicenter stepped-wedge design. The main division of the data corresponds to a multicenter before-after design (eFigure in the Supplement), followed by an adjusted comparison of the 3 concurrent surgical modalities. The study was approved by the Data Protection Agency. The Danish Gynecological Cancer Database (DGCD) is a scientific quality assurance register with mandatory entry of the data (commanded by the Danish government). Danish researchers may request access to the data in the DGCD, which requires approval from the Danish Gynecological Cancer Group and the Danish Data Protection Agency.

All consecutive women who underwent surgical treatment for International Federation of Gynecology and Obstetrics (FIGO) stage I or II endometrial cancer from January 1, 2005, to June 30, 2015, were identified.

Surgical modality was categorized as TAH, MILS, or MIRS. Minimally invasive laparoscopic surgery encompassed total laparoscopic hysterectomy and laparoscopically assisted vaginal hysterectomy. The women were divided into 2 groups: group 1 underwent surgery before MIRS was introduced, and group 2 underwent surgery after MIRS was introduced. The date dividing the groups was set as the date of the first such procedure performed at each cancer center for early-stage endometrial cancer. If a woman underwent TAH or MILS outside a cancer center during the MIRS introductory period and the information on region was missing, she was excluded from the study.

The primary outcome was severe complications, which were dichotomized for each woman into absent or present. Prior to data collection, severe complications were defined to include 30-day mortality and severe intraoperative and postoperative complications. Severe intraoperative complications were defined as unintended vascular, urinary tract, bowel, or nerve damage. Severe postoperative complications encompassed 90 postoperative days and included acute renal failure, paralytic ileus, deep venous thrombosis, pulmonary embolism, acute myocardial infarction, sepsis, fistula, postoperative deep or intra-abdominal hematoma, surgical evacuation of cavities, and the need for gynecologic reoperation.

All Danish citizens are at birth or immigration provided with a unique civil registration number that is used for all contacts with the municipality and the health care system (the Danish Civil Registration System).²² More than 100 nationwide registers and clinical databases using the civil registration number exist and provide researchers with the possibility to unambiguously link the information between registers. All linkage of data with information from nationwide registers was performed using the Danish civil registration number.²² Linkage of the data was performed by Statistics Denmark. The research group performing the data entry and data merging were blinded.

The cohort was extracted from the DGCD, which is a nationwide and validated cancer register with mandatory registration of all women with newly diagnosed gynecologic cancer.²³ The register holds prospectively obtained information on demographic, clinical, surgical, and pathologic data related to the index operation.

The Danish National Patient Register (NPR) holds information on all hospital contacts with the corresponding diagnoses and treatments coded.²⁴ The NPR holds high validity on the registration of severe conditions and was used to extract data on comorbidities and severe intraoperative and postoperative complications. Minor complications (eg, cystitis) were excluded because they are often treated by the general practitioner, whose diagnoses and treatments are not registered in the NPR. The Charlson Comorbidity Index was estimated based on diagnoses recorded in the NPR up to 10 years before surgery and categorized into none (0), mild (1), and moderate to severe (≥2).^25-27

The Danish Register of Causes of Death is based on the mandatory death certificate that is completed by physicians during medical inquest. The 30-day mortality used in the registration of severe complications is derived from this register.²⁸

Information on the educational level of each patient was obtained from the Population Education Register, and information on the disposable income of each patient the year before surgery was obtained from the Income Statistics Register.^29,30 The socioeconomic status of each patient was ranked on a 5-point scale based on their educational level and their disposable income the year before surgery.^31,32

Surgical modality and the dates derived from the DGCD were used for grouping of the exposures as previously described. Age was reported in years and grouped into quartiles. Body mass index (BMI; calculated as weight in kilograms divided by height in meters squared) was grouped into 4 categories according to cutoff values in the World Health Organization classification³³ (<25.0, 25.0-29.9, 30.0-34.9, and ≥35.0). A BMI less than 13.0 and a BMI greater than 98.0 were considered implausible and were recorded as missing. Smoking status was grouped into 3 categories in the DGCD: nonsmoker, smoker, and previous smoker. Physical status was reported according to the American Society of Anesthesiologists (ASA) score and categorized into no systemic disease (I), mild systemic disease (II), and moderate to severe systemic disease (≥III).³⁴ Information on intra-abdominal adhesions was dichotomized as yes or no.³⁵ Definitions of histopathologic risk groups (low risk, intermediate risk, and high risk) were based on the depth of myometrial invasion and histologic grade according to the FIGO 2009 guidelines.^36,37 Furthermore, women with stage II disease or nonendometrioid adenocarcinoma (carcinosarcoma included) were considered to be at high risk, in line with the new ESMO-ESGO-ESTRO (European Society for Medical Oncology–European Society of Gynaecological Oncology–European Society for Radiotherapy and Oncology) guideline.³⁸ Thus, histopathologic risk was defined as follows: low, endometrioid adenocarcinoma stage IA with grade I or II; intermediate, endometrioid adenocarcinoma stage IB with grade I or II or endometrioid adenocarcinoma stage IA with grade III; high, endometrioid adenocarcinoma stage IB with grade III, endometrioid adenocarcinoma stage II, or nonendometrioid carcinoma; and unknown, endometrioid adenocarcinoma stage I with unknown grade. Lymph node dissection was reported as a dichotomized variable (yes or no). Information on the number of lymph nodes removed was not available. Treatment at a cancer center vs a nonspecialized gynecologic department was noted as a dichotomized variable.

Statistical analysis was conducted from February 2, 2017, to May 4, 2018. The distribution for demographic and tumor characteristics between group 1 and group 2 and among the 3 surgical modalities (TAH, MILS, and MIRS) was compared by use of χ² tests.

A logistic regression model was used to estimate the odds ratio (OR) with 95% CI for severe complications. We compared severe complications in group 1 and group 2 and stratified them using the 2 surgical modalities (TAH and MIS). Furthermore, restricted analyses in group 2 were performed comparing severe complications among the 3 surgical modalities (TAH, MILS, and MIRS).

Univariate analyses and adjusted ORs for prespecified potential confounders were performed and included age, BMI, smoking status, ASA score, intra-abdominal adhesions, socioeconomic status, histologic risk groups, lymphadenectomy, and surgical year. The models were initially adjusted for confounders that were considered the most important clinically: age, ASA score, lymphadenectomy, and BMI. Furthermore, the model was tested for other potential confounders, and those that had a significant association were included in the model. Likelihood ratio tests were used to evaluate whether potential interactions between the key confounding factors were statistically significant. The Hosmer-Lemeshow test was used to evaluate the goodness of fit of the models. Furthermore, the model that compared the 3 concomitant surgical modalities was adjusted for site-level clustering after the centralization was accomplished (ie, to account for the correlation between the sites and the risk of severe complications). Women with missing data were excluded from the adjusted analyses. In addition, an explorative adjusted model was generated to see whether an MIRS learning curve could be identified within group 2. In this model, the surgical year was replaced by a variable of time since MIRS introduction alongside an interaction term between the new time variable and surgical modalities.

Sensitivity analyses were performed to evaluate the potential effect of missing data. The OR of severe complications for the entire population and the OR of severe complications for the women included in the fully adjusted analyses were compared.

All P values were from 2-sided tests, and results were deemed statistically significant at P < .05. All analyses were conducted using STATA, version 15.0 (StataCorp).

In total, 5654 women with stage I or II endometrial cancer underwent surgery during the study period (Figure 1). Of these women, 3091 (mean [SD] age, 67 [10] years) underwent surgery before the implementation of MIRS (group 1) and 2563 (mean [SD] age, 68 [10] years) underwent surgery after the implementation of MIRS (group 2). The 2 groups differed significantly regarding several clinical, sociodemographic, and histopathologic characteristics (Table 1).

A total of 226 women in group 1 experienced severe complications (7.3%; 95% CI, 6.4%-8.3%), and a total of 160 women in group 2 experienced severe complications (6.2%; 95% CI, 5.3%-7.2%) (Table 2). We also observed that 7 of 5654 women (0.12%) had acute renal failure and that 15 of 5654 (0.27%) had sepsis. A multiple logistic regression analysis demonstrated that the odds of severe complications were associated with lymphadenectomy, histopathologic tumor risk group, ASA score, intra-abdominal adhesions, and BMI. Age, surgical year, smoking status, and socioeconomic status were not significantly associated with the odds of severe complications. In the adjusted analyses, the odds of severe complications were significantly higher among women in group 1 compared with women in group 2 (OR, 1.39; 95% CI, 1.11-1.74). In the overall comparison of groups, the surgical approach was not incorporated in the analysis. The model was repeated within stratified analyses among women who underwent minimally invasive techniques and open procedures. The stratified and adjusted analyses did not demonstrate any difference in odds of severe complications (Table 3).

Among women in group 2, severe complications occurred in 81 of 712 who underwent TAH (11.4%; 95% CI, 9.1%-13.9%), 29 of 569 who underwent MILS (5.1%; 95% CI, 3.4%-7.2%), and 59 of 1282 who underwent MIRS (4.6%; 95% CI, 2.9%-5.1%). A multiple logistic regression model for severe complications in group 2 demonstrated that the odds of severe complications were associated with lymphadenectomy, ASA score, BMI, socioeconomic status, and surgical year but were not associated with age, smoking status, tumor risk group, socioeconomic group, and intra-abdominal adhesions. In the adjusted analyses, TAH was associated with increased odds of complications compared with MILS (OR, 2.58; 95% CI, 1.80-3.70) and MIRS (OR, 3.87; 95% CI, 2.52-5.93), whereas no significant difference in the odds of complications was found when women who had MILS were compared with those who had MIRS (OR, 1.50; 95% CI, 0.99-2.27). No significant differences were found in sensitivity analyses.

In group 1, 2655 women (85.9%) underwent TAH, and 436 women (14.1%) underwent MILS. In group 2, 712 women (27.8%) underwent TAH, 569 women (22.2%) underwent MILS, and 1282 women (50.0%) underwent MIRS. The nationwide use of MIS increased from 3% in 2005 to 95% in 2015 (Figure 2). Minimally invasive laparoscopic surgery was accessible throughout the study period, whereas MIRS was gradually implemented between 2008 and 2013 in all cancer centers. In 2005, 202 of 488 women (41.4%) had surgical procedures at gynecologic cancer centers; this number gradually increased to 535 of 554 (96.6%) in 2012. In 2005, 49 of 488 women (10.0%) underwent staging lymph node dissections, gradually increasing to 201 of 525 (38.3%) in 2010. The proportion of staging lymph node dissections remained stable, between 34.9% (197 of 565) and 39.5% (209 of 529), from 2010 to 2015. Trends of changes in the OR of severe complications after MIRS along a potential learning curve after the introduction of MIRS were not observed.

To our knowledge, this study describes the first analyses of how a nationwide implementation of MIRS affected the risk of surgical complications in an unselected national cohort of women with early-stage endometrial cancer. The implementation of MIRS led to an increase in the proportion of women who underwent MIS, and it reduced the number of severe complications. The reduction in the number of severe complications was observed despite a higher proportion of women with an older age, a high ASA score, high-risk histopathologic characteristics, and intra-abdominal adhesions being offered MIS and a higher proportion of women undergoing staging lymphadenectomy. Randomized trials on TAH vs MIRS have reported a decreased risk of complications after laparoscopy.^3-5,8,39-42 Thus, in line with earlier findings from single-institution trials,^17,19,43-52 multi-institution trials,^10,11,53-57 and randomized clinical trials in selected cohorts,^3-5,8,39-42 our nationwide study confirms that MIS for women with early-stage endometrial cancer decreases the risk of severe complications compared with open surgery. Furthermore, the present study supports earlier reports from specialized single-institution trials that using MIRS for early-stage endometrial cancer decreases the risk of surgical complications.^{17,20,21,58,59}

Studies in selected cohorts have compared complications between minimally invasive approaches and reported equal or improved results after MIRS.^{17,45,46,48,51,52,54-56} In contrast, Wright et al¹⁰ demonstrated an increased complication rate after MIRS (23.7%) compared with MILS (19.5%) for endometrial cancer. The study reported comparatively high, severe complication rates of 39.7% after TAH and 22.7% after MIS. The authors ascribe the higher risk of complications after MIRS to an increased proportion of medical complications, in particular, bacteremia and respiratory and renal failures in the MIRS group. The data in the study by Wright et al¹⁰ were derived from the Surveillance, Epidemiology, and End Results (SEER) Medicare database, which combines data from the National Cancer Database and claims from patients covered by Medicare.⁶⁰ Medicare is a US government insurance system that aims to support persons 65 years of age or older, persons receiving chemotherapy, persons with end-stage renal disease, persons with amyotrophic lateral sclerosis, and younger persons with disabilities.⁵⁷ Wright et al¹⁰ adjusted for the complexity within this cohort by using propensity score matching weighted by the inverse probability of treatment with additional stabilizing techniques. However, we believe that the increased odds of complications after MIRS compared with MILS is a result of selection bias related to the allocation of women with a higher risk of complications to undergo MIRS rather than to the higher risk of complications after MIRS. In our study, we observed that 0.12% of women had acute renal failure and that 0.27% had sepsis. The proportions of acute renal failure and of sepsis found in our study correspond with previous reports on women with endometrial cancer.^45,53,55,61

Bergstrom et al¹¹ have recently proposed a minimally invasive hysterectomy benchmark of more than 80% when performed at high-volume institutions. The study included 1621 women who underwent surgery for endometrial cancer during the period from 2013 to 2014 at 4 national cancer centers. Our study suggests that a nationwide, minimally invasive hysterectomy benchmark of 95% is feasible for early-stage endometrial cancer if treatment is centralized to high-volume cancer centers with MIRS adopted.

In 2002 and 2005, the Danish government launched a nationwide Cancer Action Plan I and II, respectively, in which endometrial cancer treatment was centralized to specialized centers.^62,63 The centralization of endometrial cancer treatment was completed in 2012 after gradual implementation. A nationwide centralization of complex surgical procedures induces a high flow of patients and gives the optimal conditions for improving surgical expertise in complex surgical procedures for the treatment of cancer. A high level of surgical expertise may be associated with a decreased number of complications, but centralization may also facilitate more comprehensive surgery and may offer surgery for patients with cancer and comorbid conditions. Only 101 women in group 2 received treatment outside the cancer centers; these women had few severe complications. Therefore, adjustment or stratification for centralization in the adjusted regression analyses was not possible, and the degree and direction of a potential association of centralization on our outcome remains unknown.

Our study was strengthened by the inclusion of prospectively entered national data during a 10-year period with the gradual introduction of MIRS. By dividing the women into groups according to the availability of MIRS, we avoided a comparison across surgical modalities that were not available during the time frame.

Our study included 97% of all women who underwent surgery for early-stage endometrial cancer in Denmark, thus minimizing selection. The DGCD consecutively performs external validation and data quality improvement through updated national registers generating deficiency lists that are completed by the departments responsible for reporting. However, all register-based studies are at risk of information bias. The registration of minor complications may be underreported in the NPR; to minimize such information bias, we included only major complications. Furthermore, severe complications were dichotomized to decrease the risk of multiple registrations of complications with the same origin, increasing the robustness of the analyses.

The MIRS learning curve has been suggested to encompass up to 30 operations,^14,61,64,65 and as the nationwide introduction of MIRS was effectuated during the present study, our estimates may be considered conservative.^10,13,14 Information on the MILS and MIRS experiences of the surgeons was not available; accordingly, adjustment for the surgeon-specific learning curve was not possible. The absence of an overall learning curve may be due to several factors. The MIRS learning curve is reported to be less steep than that for conventional laparoscopy and therefore was less of a factor. Furthermore, the first surgical procedures performed at each center were performed by skilled cancer surgeons, which means that the learning curve for the Danish surgeons performing MIRS is likely to be scattered over time.

The national introduction of MIRS changed the surgical approach for early-stage endometrial cancer from open surgery to MIS. This change in surgical approach was associated with a significantly reduced risk of severe complications.

Accepted for Publication: November 23, 2018.

Corresponding Author: Siv Lykke Jørgensen, MD, Department of Gynecology and Obstetrics, Odense University Hospital, Kloevervaenget 10, Tenth Floor, Odense 5000, Denmark (siv.lykke.joergensen@rsyd.dk).

Published Online: February 27, 2019. doi:10.1001/jamasurg.2018.5840

Author Contributions: Drs Jørgensen and Wu had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Jørgensen, Mogensen, Lund, Jensen.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Jørgensen, Mogensen, Lund, Jensen.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Jørgensen, Wu, Iachina, Jensen.

Obtained funding: Jørgensen, Mogensen, Jensen.

Administrative, technical, or material support: Jørgensen, Korsholm, Jensen.

Supervision: Mogensen, Jensen.

Conflict of Interest Disclosures: None reported.

Funding/Support: The project was funded by the region of Southern Denmark, the University of Southern Denmark, the Danish Cancer Society, the Department of Gynecology and Obstetrics Odense University Hospital Denmark, the Senior Consultant Research Council Odense University Hospital Denmark, the A.P. Moeller Foundation, and Carpenter Axel Kastrup-Nielsen’s memorial fund. The study was granted epidemiological support by Program for Clinical Research Infrastructure, which was established by the Lundbeck Foundation and the Novo Nordisk Foundation.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Additional Contributions: Claus Høgdall, Department of Gynaecology Rigshospitalet Denmark, assisted with merging the Danish Gynecological Cancer Database data set. He was not compensated for his contribution.