Robotic-Assisted versus Laparoscopic Proctectomy with Ileal Pouch-Anal Anastomosis for Ulcerative Colitis: An Analysis of Clinical and Financial Outcomes from a Tertiary Referral Center

Background: Robotic-assisted colorectal surgery is gaining popularity, but limited data are available on the safety, efficacy, and cost of robotic-assisted restorative proctectomy with the construction of an ileal pouch and ileal pouch-anal anastomosis (IPAA) for ulcerative colitis (UC). Methods: A retrospective study was conducted comparing consecutively performed robotic-assisted and laparoscopic proctectomy with IPAA between 1 January 2016 and 31 September 2021. In total, 67 adult patients with medically refractory UC without proven dysplasia or carcinoma underwent surgery: 29 operated robotically and 38 laparoscopically. Results: There were no differences between both groups regarding postoperative complications within 30 days according to Clavien-Dindo classification’ grades 1–5 (51.7% vs. 42.1%, p = 0.468) and severe grades 3b–5 (17.2% vs. 10.5%, p = 0.485). Robotic-assisted surgery was associated with an increased urinary tract infection rate (n = 7, 24.1% vs. n = 1, 2.6%; p = 0.010) and longer operative time (346 ± 65 min vs. 281 ± 66 min; p < 0.0001). Surgery costs were higher when operated robotically (median EUR 10.377 [IQR EUR 4.727] vs. median EUR 6.689 [IQR EUR 3.170]; p < 0.0001), resulting in reduced total inpatient profits (median EUR 110 [IQR EUR 4.971] vs. median EUR 2.853 [IQR EUR 5.386]; p = 0.001). Conclusion: Robotic-assisted proctectomy with IPAA can be performed with comparable short-term clinical outcomes to laparoscopy but is associated with a longer duration of surgery and higher surgery costs. As experience increases, some advantages may become evident regarding operative time, postoperative recovery, and length of stay. The robotic procedure might then become cost-efficient.


Introduction
Almost 20% of patients with ulcerative colitis (UC) require surgery during their lifetime [1]. The procedure of choice is a restorative proctocolectomy with the construction of an ileal pouch with ileal pouch-anal anastomosis (IPAA), offering an effective cure when medical treatments have been insufficient while avoiding a permanent stoma. The surgical approach has evolved from an open technique to a variety of minimally invasive procedures, including the use of a robotic platform [2][3][4]. The evolution toward a less invasive approach in this young patient population was advocated early on by the clear benefits of better cosmesis [4], shorter convalescence [5], less pronounced intra-abdominal adhesions [6], and lower rates of abdominal wall herniation [7].
Despite these advantages of laparoscopy, there are also limitations, including a limited range of motion within the narrow pelvis and confined visibility when performing distal rectal dissection and subsequent anastomosis of the ileal pouch. Ongoing innovation in minimally invasive surgery has led to advances in robotic surgical technology [8,9]. The improved ergonomics, dexterity, stable visualization of the robotic platform, and equivalent safety and efficacy outcomes have contributed to its increasing use in rectal cancer surgery [10,11].
However, are these advantages relevant in benign disease when performing roboticassisted proctectomy with IPAA for UC? The literature comparing robotic and laparoscopic proctectomy with IPAA is scarce. Most studies include only a few patients with heterogenous patient outcomes and limited data on in-hospital costs of the two approaches [12,13].
Therefore, this study aimed to present our first series of consecutive patients who underwent robotic-assisted proctectomy with IPAA for UC. We hypothesized that the robotic approach is associated with a lower 30-day postoperative complication rate, a reduced length of hospital stay (LOS), and a lower conversion rate to open surgery. We performed an in-hospital cost analysis to compare both techniques' costs, revenues, and profits in the operating room and on the ward.

Patient Cohort
The study was a retrospective, comparative study performed at the Department of General and Visceral Surgery, Charité Berlin-Campus Benjamin Franklin. The department acts as a referral center for inflammatory bowel disease (IBD) and is a highly specialized center for laparoscopic colorectal surgery. Robotic-assisted colorectal resections have been regularly performed since 2016. Using our hospital's patient database (SAP SE, Walldorf, Baden-Württemberg, Germany), we identified all adult patients (age ≥ 18 years) with UC without proven carcinoma or dysplasia who had undergone elective robotic-assisted or laparoscopic restorative proctectomy with IPAA for UC between 1 January 2016 and 31 September 2021. The 67 consecutive patients suffered from medically refractory UC and had previously undergone laparoscopic colectomy with rectal stump closure and creation of an end ileostomy as a first step according to a three-step procedure. Patients with Crohn's disease, indeterminate colitis, or familial adenomatous polyposis were excluded. No patient was excluded beyond these criteria. The choice of the surgical procedure depended on the surgeon's preference and the availability of the robotic platform. Patient characteristics did not influence the choice of approach. All robotic-assisted proctectomies were performed using the Da Vinci X Surgical System (Intuitive Surgical, Mountain View, Sunnyvale, CA, USA). The study was approved by the local ethical review committee (EA4/049/21) and was in accordance with the Helsinki Declaration of 1975 and its later amendments or comparable ethical standards.
Inclusion criteria:

Surgical Interventions
All five operating surgeons were experienced laparoscopic colorectal surgeons, whereas they were in their learning curve for robotic-assisted proctectomy. The surgeons completed their basic training on the robotic console in 2015, 2017, 2019, 2020, and 2021.
In robotic-assisted proctectomy, the takedown of the terminal ileostomy was conducted first, followed by placing four 8 mm robotic trocars (Intuitive Surgical, Sunnyvale, CA, USA) in a horizontal line, a few centimeters cephalad of the umbilicus. A 12 mm auxiliary trocar was placed on the right flank, and another 12 mm auxiliary trocar was inserted via a Pfannenstiel incision as an optical trocar for the conventional laparoscopic part of the operation (Applied Medical, Santa Clara, CA, USA) ( Figure 1). Laparoscopic lysis of adhesions and mobilization of the mesenteric root was then performed, followed by a robotic-assisted proctectomy in the total mesorectal excision (TME) plane with the preservation of pelvic nerves. Robotic-assisted proctectomy was conducted with monopolar curved scissors (arm 1), fenestrated bipolar forceps (arm 3), and cadiere forceps (arm 4) (Intuitive Surgical, Mountain View, Sunnyvale, CA, USA;). The robotic camera was placed in arm 2 (Intuitive Surgical, Mountain View, Sunnyvale, CA, USA). The rectum was transected at the level of the pelvic floor using either a linear cutter via the Pfannenstiel incision (TA 30; Medtronic, Dublin, Ireland) or a robotic linear stapler (SureForm 45 mm green cartridge; Intuitive Surgical, Mountain View, Sunnyvale, CA, USA). If a robotic stapler was used, a 12 mm robotic arm was placed as arm 1, and stapling was conducted via this arm. USA) in a horizontal line, a few centimeters cephalad of the umbilicus. A 12 mm auxiliary 100 trocar was placed on the right flank, and another 12 mm auxiliary trocar was inserted via 101 a Pfannenstiel incision as an optical trocar for the conventional laparoscopic part of the 102 operation (Applied Medical, Santa Clara, CA, USA) (Figure 1). Laparoscopic lysis of ad-103 hesions and mobilization of the mesenteric root was then performed, followed by a ro-104 botic-assisted proctectomy in the total mesorectal excision (TME) plane with the preser-105 vation of pelvic nerves. Robotic-assisted proctectomy was done with monopolar curved 106 scissors (arm 1), fenestrated bipolar forceps The ileal pouch was constructed extracorporeally with three staple loads (GIA 75 120 mm blue cartridge; Medtronic, Dublin, Ireland) with a pouch length of 12 -14 centimeters. 121 The type of anastomosis -hand-sewn or stapled -was at the discretion of the operating 122 surgeon. In this cohort, without evidence of dysplasia or carcinoma, our standard anasto-123 mosis was created with circular stapling (CEEA 29 mm; Johnson & Johnson; New Bruns-124 The ileal pouch was constructed extracorporeally with three staple loads (GIA 75 mm blue cartridge; Medtronic, Dublin, Ireland) with a pouch length of 12-14 cm. The type of anastomosis-hand-sewn or stapled-was at the discretion of the operating surgeon. In this cohort, without evidence of dysplasia or carcinoma, our standard anastomosis was created with circular stapling (CEEA 29 mm; Johnson & Johnson; New Brunswick, NJ, USA). Hand-sewn anastomoses were performed on rare occasions, for example, in case of technical difficulties with the stapled anastomosis. Rigid pouchoscopy with air leak test was performed regularly. In case of leakage, an endosponge was placed at the level of the anastomosis. A loop ileostomy was created in all patients.
In the laparoscopic group, proctectomy was performed laparoscopically with mesorectal excision as described previously [14,15]. The remaining steps of the operation were performed equally in both groups.

Outcome Measures
Preoperative, perioperative, and 30-day postoperative outcomes were retrospectively analyzed. Preoperative data included American Society of Anesthesiologists (ASA) classification, anemia, diabetes mellitus, immunosuppressive medication (corticosteroids and/ or immunomodulators) within 30 days before surgery, number of previous abdominal surgeries, and smoking status. Perioperative characteristics included the anastomotic technique (hand-sewn or stapled anastomosis), frequency of intraoperative adverse events such as damage to organ structures (lesion of the vagina, bladder, or the ureter) or serosal tears, the operative time and the conversion rate to open surgery.
The primary endpoint was the rate of 30-day postoperative complications, including any deviation of the postoperative course defined as grades 1-5 of the Clavien-Dindo classification of surgical complications [16]. Postoperative complications included anastomotic leakage (seen on abdominopelvic imaging or endoscopy), surgical site infection (SSI [defined according to Centers for Disease Control and Prevention Definition [17]), hemorrhage (defined by blood count and/or abdominopelvic imaging, leading to transfusion and/or intervention), small bowel obstruction (SBO [defined by insertion of a nasogastric tube]), urinary tract infection (UTI [defined by a positive urinary culture and symptoms]), pneumonia (defined by imaging and need of antibiotic therapy), and pulmonary artery embolism (defined by CT imaging).
Secondary endpoints were operative time, rate of conversion to open surgery, LOS, and detailed in-hospital cost analysis. Our study population's calculated costs, revenues, and profits refer to the Department of General and Visceral Surgery, Charité Berlin-Campus Benjamin Franklin, where the study was conducted. The calculation is based on the official guidelines of the German Institute for remuneration in hospitals (InEK) for the calculation of German Diagnosis Related Groups (G-DRGs). The tool of our corporate controlling department is a graphic display that outlines the calculation of costs and the comparison of proceeds and expenses. This tool calculated the exact inpatient costs, revenues, and profits based on the case number of every single patient, including surgery costs, costs on surgical ward, medication costs, laboratory costs, and costs of diagnostic procedures. Surgery costs were mainly referred to cutting-suture time and time of anesthesia. Surgery costs also involved costs for revision surgery. Revenues were based on the G-DRGs, in which there is a fixed case-related sum regardless of the treatment of an individual patient [18]. No additional reimbursement is granted for robotic-assisted procedures in Germany. A detailed description of the composition of the underlying inpatient cost analysis can be found in previously published data of our working group [19].

Statistical Analysis
Parameters were depicted according to their scale and distribution with absolute and relative frequencies for categorical parameters and mean, standard deviation (SD), median, and interquartile range (IQR) for quantitative parameters. For quantitative outcomes, statistical group comparisons were performed using the t-test for independent samples. Because of the skewed distribution of some of the variables, group differences for quantitative variables were analyzed by the Mann-Whitney U test. For categorical outcomes, group comparisons were performed using cross-tabulation and the chi-square test.
Moreover, a multiple logistic regression was conducted to identify potential UTI risk factors. The analyzed independent variables were included due to a p value of 0.05 or less on simple regression analysis with one independent variable. Subsequently, we performed stepwise backward variable elimination with a threshold p > 0.1. Values of the multivariate analysis were expressed as odds ratio (OR), 95% confidence interval (CI), and p value. Furthermore, a multiple linear regression analysis was performed for potential risk factors of longer operative time. Again, stepwise backward variable elimination was achieved with a threshold of p > 0.1. Values from this second multivariate analysis were reported as regression coefficient, 95% CI, standardized coefficient, and p value.
p values of 0.05 or less were considered statistically significant. Due to the exploratory character of the analyses, no Bonferroni correction has been performed. Data analysis was performed with IBM SPSS Statistics 27 (IBM, Armonk, NY, USA).

Results
Sixty-seven patients underwent minimally invasive restorative proctectomy with IPAA, of which 29 were performed robotically and 38 laparoscopically. Slightly more patients were female (n = 34, 50.7%), with a mean age of 38 years (SD = 13 years) ( Table 1). There was no difference in ASA classification among patients undergoing robotic-assisted or laparoscopic surgery, with the majority classified as ASA 2 (72.4% vs. 78.9%; p = 0.454) ( Table 1). More patients in the robotic-assisted group had more than one previous abdominal surgery than in the laparoscopic group (48.3% vs. 21.1%; p = 0.028) ( Table 1). Both groups showed no differences concerning age, sex, BMI, smoking status, diabetes, anemia, and preoperative steroid or immunomodulatory medication (Table 1).
In the robotic-assisted group, one patient was converted to open surgery versus four patients in the laparoscopic group (3.4% vs. 10.5%; p = 0.379) ( Table 4). Compared to four patients in the laparoscopic group, one patient in the robotic-assisted group showed a positive air leak test when IPAA was performed (3.4% vs. 10.5%; p = 0.379) ( Table 4). Regarding the LOS, both groups showed no difference, with a median of 9 days (IQR 4 days) for the robotic-assisted group and a median of 8.5 days (IQR 7 days) for the laparoscopic group (p = 0.877) ( Table 4).      The total inpatient costs did not differ between both groups (median 15.051 EUR/16.275 USD vs. median 13.243 EUR/14.320 USD; p = 0.118) ( Table 6). The robotic-assisted approach was associated with higher surgery costs compared to the laparoscopic approach (median 10.377 EUR/11.221 USD vs. median 6.689 EUR/7.233 USD; p < 0.0001); total inpatient revenues did not differ (median 15.524 EUR/16.766 USD vs. median 15.524 EUR/16.766 USD; p = 0.256) ( Table 6), resulting in reduced total inpatient profits for robotic surgery (median 110 EUR/119 USD vs. median 2.853 EUR/3.081 USD; p = 0.001) ( Table 6). A detailed in-hospital cost analysis can be found in Table 6.

Discussion
As the safety and feasibility of robotic-assisted low anterior resection have been demonstrated, the technique has gained momentum in rectal cancer surgery [10,20]. However, robotic-assisted proctectomy with IPAA for benign disease is still rarely performed. We maintained the well-established laparoscopic steps when working outside the pelvis and used the robotic platform specifically for proctectomy, where robotic surgery is probably most beneficial. Our technique of robotic-assisted proctectomy in UC involves a TME to the level of the pelvic floor. Although not essential in benign disease, TME provides an almost bloodless plane of dissection, allowing us to identify and maintain the surgical planes during pelvic dissection.
To our knowledge, we studied the largest cohort of patients who received roboticassisted or laparoscopic proctectomy with IPAA for UC. Our data demonstrated the safety and feasibility of robotic-assisted proctectomy with IPAA for UC. We showed equivalent 30-day postoperative complication rates for robotic-assisted and laparoscopic proctectomy with IPAA, as indicated by comparable overall (Clavien-Dindo grade 1-5) and severe complications (Clavien-Dindo grade 3b-5) (51.7% vs. 42.1% and 17.2% vs. 10.5%, respectively).
A systemic review and meta-analysis by Flynn et al. also showed no differences in 30-day postoperative complication rates between robotic-assisted and laparoscopic proctectomy with IPAA. The systematic review revealed no randomized trials, the number of patients in most studies was small, and patient cohorts were heterogeneous [12]. In contrast, our study cohort was homogeneous: all patients suffered from ulcerative colitis without evidence of dysplasia or carcinoma, were status post colectomy, and underwent proctectomy with the construction of an ileal pouch with IPAA.
In our cohort, deaths were reported in neither group. Interestingly, UTIs occurred more often in patients undergoing robotic-assisted versus laparoscopic proctectomy: 24.1% vs. 2.6%. At the same time, urinary retention was detected in neither group (defined as failure to wean off the urinary catheter after two attempts during the hospital stay with a residual urine volume ≥ 50 mL). In our study, only patients undergoing roboticassisted proctectomy received a TME to the level of the pelvic floor, potentially injuring the hypogastric plexus and/or the pelvic splanchnic nerves/pelvic plexus, possibly resulting in higher rates of UTI and long-term urinary dysfunction. However, the study is limited by the lack of long-term data on urinary function and the short follow-up. In addition, no patient in the robotic group developed urinary retention, making pelvic nerve damage with consecutive UTI rather unlikely. In total, 17.2% of patients in the robotic group versus 7.9% of patients in the laparoscopic group had to return to the operating room postoperatively before discharge. These patients required reinsertion of a Foley catheter, which may have contributed to the higher rate of UTIs in the robotic group. Furthermore, although not significant, there were slightly more women in the robotic group than in the laparoscopic group. Since women are at higher risk for UTIs, this may have contributed to the higher UTI rate. The proportion of patients with previous abdominal surgeries was significantly higher in the robotic group. The probably higher degree of adhesions may have contributed to a more complex pelvic dissection, a longer operative time, and a consecutively higher UTI rate.
Consistent with other studies [3,21], robotic-assisted surgery took longer than laparoscopic surgery (346 min. vs. 281 min. on average). Furthermore, multiple linear regression analyses associated hand-sewn anastomosis and higher BMI with longer operative time.
Interestingly, compared to others [3,8], we could not demonstrate a progressive reduction in operative time with increasing surgeon experience (data not shown) as would be expected in any new technique. The number of robotic-assisted surgeries performed by individual surgeons may have been too small to detect such a reduction.
Based on the hypothesis that the technical advantages of the robotic platform should facilitate proctectomy and avoid the need for conversion to open surgery, we compared conversion rates between both groups. Although patients in the robotic-assisted group underwent previous abdominal surgeries more often, the conversion rate was low at 3.4%, whereas the conversion rate in the laparoscopic group was 10.5%. Remarkably, the low conversion rate was achieved in the robotic group, although the number of previous abdominal surgeries was higher in the robotic group than in the laparoscopic group. However, the difference did not reach a significance level; our study cohort may have been too small to detect such a difference. In the study of Rencuzogullari, robotic-assisted proctectomy with IPAA was also not associated with a reduced conversion rate to open surgery [22]. The prospective randomized ROLARR trial revealed a decreased conversion rate for robotic-assisted versus laparoscopic low anterior resection for rectal cancer in the subgroup of male patients [20].
Consistent with other studies, we could not demonstrate shorter LOS for patients receiving robotic-assisted proctectomy with IPAA [3,21,22]. Regarding the LOS, both groups showed no difference, with a median of 9 days for the robotic-assisted group and a median of 8.5 days for the laparoscopic group.
To our knowledge, this analysis is the first detailed economic evaluation of inpatient healthcare costs comparing robotic-assisted and laparoscopic proctectomy with IPAA. Our corporate controlling team assessed healthcare costs based on the G-DRG system. The exact costs, revenues, and profits based on the case numbers of each individual patient were calculated, enabling accurate estimates. Our results suggest that robotic-assisted proctectomy with IPAA is unlikely to be cost-saving. Regarding surgery costs, the roboticassisted approach was more expensive than the laparoscopic approach (median 10 When considering robotic-assisted surgery, the acquisition and maintenance costs, the operational life, and the total utilization of the robotic platform per year must be considered. A detailed health economic analysis by Jayne et al. [20], based on data provided by Intuitive Surgical in 2017 (Mountain View, Sunnyvale, CA, USA), stated that "the net benefits (excluding fixed costs) of each robotic-assisted operation included in a set of cost-effective procedures must be positive, and the entire set of cost-effective procedures must have an average net benefit of at least 1.491 € (1.611 USD). On average, all robotic-assisted operations combined must exceed this figure, with each operation making at least some positive contribution." Based on the data of our retrospective study, robotic-assisted proctectomy was associated with median higher total inpatient costs of 1.808 EUR (1.955 USD) per case. It must be considered that all surgeons were experienced laparoscopic colorectal surgeons. In contrast, they were in their learning curve for robotic-assisted proctectomy. This might have biased our results. A lower 30-day postoperative complication rate leads to reduced costs, and LOS might occur with more advanced experience regarding the robotic-assisted technique.
It seems reasonable to further develop the technology in the rational belief that the platform's significant potential for improvement can lead to clinical benefits in the future. Robotic technology has far greater potential for further advancement than laparoscopy [23,24]. Some potential advantages of robotic proctectomy include nerve-sparing surgery with improved long-term functional outcomes and enhanced preparation to the pelvic floor with a shorter rectal stump and lower risk of cuffitis. As new competitors and further developed robotic platforms enter the market, acquisition, maintenance, and instrument costs will decrease [25]. Under these circumstances, the robotic platform will likely become a routine part of colorectal surgery, especially for complex operations such as proctectomy with IPAA.

Limitations
This study has several limitations. First, it is a monocentric retrospective comparative study of a prospectively maintained database, limiting the analysis of postoperative outcomes. Second, we act as a referral center for IBD and are a highly specialized center for laparoscopic colorectal surgery. Thus, our patient population, perioperative settings, and operative steps may differ from outside institutions, limiting the transferability of our findings. Third, many of our patients travel long distances. Therefore, we do not have routine postoperative visits; instead, we often rely on patient phone calls and nursing communication. Thus, our 30-day postoperative complication rate may be underreported. Still, we think that the vast majority of complications were detected in our cohort due to the relatively lengthy hospital stay in Germany, our policy to initiate diagnostics early in vague suspicion of a postoperative complication, and the encouragement of patients to contact our department in case of postoperative discomfort.
Fourth, the number of patients is small but will continue to increase as our experience with robotic-assisted proctectomy grows. Due to the small number, the study might have been underpowered, thus showing no significant differences. Considering these limitations, our data suggest that robotic-assisted proctectomy is feasible and has comparable shortterm outcomes to the laparoscopic approach. Fifth, patients in the robotic group had more previous abdominal surgeries. This may have biased the results regarding postoperative complications, conversion rate, hospital stay, and costs.
The study's strength is the detailed assessment of specific clinical and financial outcomes for either robotic-assisted or laparoscopic proctectomy with IPAA. To our knowledge, this analysis is the first thorough economic evaluation of inpatient healthcare costs comparing the largest cohort of patients who received robotic-assisted or laparoscopic proctectomy with IPAA for UC. The 67 consecutive patients suffered from medically refractory UC without dysplasia or carcinoma and had previously undergone laparoscopic total abdominal colectomy as a first step according to a three-step procedure. Therefore, our cohort was homogenous and comparable; both groups were well-balanced. All robotic-assisted proctectomies were performed by surgeons already considered experts in laparoscopic colorectal surgery. They followed well-established surgical training, proctoring, and protocol using the Da Vinci X Surgical System.

Conclusions
Robotic-assisted proctectomy with IPAA is a safe and effective technique for patients with UC, showing equivalent postoperative morbidity and conversion rates compared to laparoscopic surgery. With more experience, some benefits in operative time, postoperative recovery, and long-term functional outcomes may become evident. Expenses might decrease as new competitors and further developed robotic platforms enter the market.  Informed Consent Statement: Patient consent was waived, as this was a non-interventional retrospective study.

Data Availability Statement:
Publicly available datasets were analyzed in this study. These data can be found here: https://www.synapse.org/#!Synapse:syn38741915; accessed on 2 October 2022.
Conflicts of Interest: J.C.L. received remuneration for travel, accommodation, and fees for robotic courses from Intuitive Surgical; otherwise, the authors declare no conflict of interest.