An Emergency Robotic Programme for Acute Complicated Diverticulitis: A Tertiary Centre Experience

ABSTRACT

Introduction

Robotic‐assisted surgery is now the standard for elective colorectal procedures, but its use in emergencies remains limited. This study describes the feasibility of a dedicated emergency robotic programme for acute complicated diverticulitis.

Methods

We retrospectively analysed a prospectively maintained database (2020–2023). Of 31 patients undergoing robotic surgery for suspected complicated diverticulitis, 19 met inclusion criteria and underwent robotic left colectomy or Hartmann's procedure.

Results

Median total procedure time was 171 minutes (IQR 42) and median total OR time was 235 minutes (IQR 55), with no conversions to open surgery. Overall 30‐day morbidity was 47.4% (9/19) and mortality was 10.5% (2/19), broadly consistent with published laparoscopic/open outcomes, though no concurrent comparator group was analysed.

Conclusion

This initial, uncontrolled experience suggests that a dedicated emergency robotic programme for acute complicated diverticulitis is technically feasible. These hypothesis‐generating results warrant confirmation through prospective, comparative studies before conclusions about safety or superiority can be drawn.

1. Introduction

Robotic‐assisted surgery has seen remarkable growth in colorectal surgery and is projected to surpass laparoscopy in the United States [1]. This rise is attributed to its superior ergonomics, enhanced precision, 3‐D visualisation, and evidence of improved outcomes compared with traditional laparoscopy [2, 3, 4]. However, despite these advantages, robotic surgery has been slower to gain traction in emergency procedures due to higher costs, limited access, longer preparation times, and increased operative times compared with traditional laparoscopy. Limited platform availability and a shortage of trained personnel further restrict its use in emergency settings.

While emerging evidence supports the feasibility and safety of robotic surgery in abdominal emergencies, particularly in hemodynamically stable patients treated at well‐trained centers, its application remains limited [5]. Studies have shown potential benefits in procedures such as cholecystectomy, colectomy, and hernia repair, including lower conversion rates to open surgery and reduced complication rates [6, 7, 8].

To address the gap in robotic surgery for emergency cases and contribute to the growing body of evidence, the colorectal unit at UZ Brussel launched a dedicated program in December 2019 for left colectomy or Hartmann's procedure in patients with complicated diverticulitis [9].

Led by two robotic‐trained colorectal surgeons, this program sought to extend the benefits of robotic surgery to patients undergoing emergency procedures, who are often affected by multiple comorbidities or are suboptimally prepared for surgery.

This study presents the initial, preliminary results of this emergency surgery program and describes its feasibility in the emergency setting. Given the retrospective, uncontrolled, single‐center design and small sample size, the findings presented here should be considered hypothesis‐generating rather than definitive evidence of safety or comparative benefit.

2. Materials and Methods

This retrospective study, conducted in accordance with the STROBE guidelines for observational research [10], analysed data from a prospectively maintained database (January 2020 to December 2023). The study was approved by the ethical committee of Universitair Ziekenhuis Brussel (BUN B1432023000299).

Included were adult patients (≥ 18 years) who underwent emergent or urgent robotic left colectomy or Hartmann's procedure for perforated diverticulitis, where emergent surgery was defined as unplanned surgery within 24 hours of diagnosis and urgent surgery as surgery within 72 hours of diagnosis. Patients under 18 years of age were excluded. Patients with a known preoperative or intraoperative diagnosis of malignancy, an operative indication other than diverticulitis (e.g., ischaemic colitis, diverticular bleeding, or vascular perforation), or a delay from diagnosis to surgery exceeding 72 hours were excluded. Two patients with an incidental postoperative or unrelated concurrent malignancy diagnosis were retained under an intention‐to‐treat framework, as detailed in the Results and Discussion.

2.1. Patient Triage and Selection

All haemodynamically stable patients presenting with suspected complicated diverticulitis who tolerated pneumoperitoneum at the start of surgery were eligible for the robotic approach, regardless of age, body mass index, or comorbidities. However, assignment to the robotic pathway was contingent on the availability of one of the two robotically trained colorectal surgeons at the time of presentation. When neither surgeon was on call, the on‐call surgical team selected a laparoscopic or open approach instead. This on‐call, surgeon‐availability‐dependent assignment introduces an inherent selection bias, as discussed further below.

Data collected included patient demographics (age, sex, body mass index [BMI]), American Society of Anesthesiologists (ASA) classification, World Society of Emergency Surgery (WSES) classification for diverticular disease, presence of generalized purulent or fecal peritonitis, time from diagnosis to surgical incision, surgical history (prior abdominal surgery), and operative details (procedure type, total OR time, total procedure [skin‐to‐skin] time, splenic flexure mobilization, intraoperative complications, conversion to open surgery, and estimated blood loss). Postoperative variables included intensive care unit admission, length of hospital stay, 30‐day morbidity, time to recovery of gastrointestinal function (first flatus, first bowel movement, full mobility, and resumption of a soft diet), 30‐day readmission, 30‐day reoperation, and 30‐day mortality.

2.2. Outcome Definitions

Overall 30‐day morbidity was defined as the occurrence of any complication of any severity within 30 days of the index surgery, with each patient classified according to the single highest Clavien‐Dindo grade reached during that period (i.e., a patient who experienced both a minor and a major complication was classified only under the major/highest grade, to avoid double‐counting). Surgery‐related morbidity was defined as complications directly attributable to the operative procedure itself (e.g., anastomotic leak, surgical site infection, intra‐abdominal collection), as distinct from medical complications unrelated to the surgical technique. Thirty‐day readmission was defined as any unplanned hospital readmission within 30 days of the index procedure, and 30‐day reoperation as any unplanned return to the operating room within the same period.

2.3. Emergency Robotic Colorectal Surgery Programme

This emergency programme, developed and implemented by two experienced (each with 200+ robotic colorectal resections performed prior to programme implementation) robotic‐trained colorectal surgeons (Authors E.V.E. and J.S.), aimed to provide rapid robotic surgical intervention for patients with suspected complicated diverticulitis.

Running such a highly specialised programme, with surgery often performed outside normal working hours, demands institutional support and a team of robotically trained OR staff, including scrub nurses, first assistants, and anaesthesiologists experienced with robotic surgery. Our goal was to offer this service 24/7 to ensure timely treatment. However, due to logistical limitations in this initial implementation phase, the two dedicated colorectal surgeons could not provide 24/7 coverage. On the occasions when only non‐robotically trained colleagues were on call, immediate robotic surgery was not possible, and the on‐call surgeon determined the most appropriate alternative approach (laparoscopic or open). A true 24/7 robotic emergency service was therefore not achieved during this pilot phase; a subset of the 49 contemporaneous non‐robotic (laparoscopic or open) emergency cases performed during the same period occurred when the robotic team was not available on call. An overview of the emergency robotic colorectal surgery team is represented in Figure 1.

FIGURE 1.

Overview of the team offering emergency robotic colorectal surgery to the patient (p).

All robotic procedures were performed using the da Vinci Xi surgical system (Intuitive Surgical Inc., Sunnyvale, CA, USA). A standardized four‐port robotic approach with a 12‐mm assistant port was employed. The operating surgeon determined whether to proceed with resection and primary anastomosis or a Hartmann's procedure based on clinical judgement.

2.4. Statistical Methods

Categorical variables were reported as frequencies with percentages, and quantitative variables were expressed as medians with interquartile ranges (Median [IQR]) throughout the manuscript and tables. Given the limited size of the two procedure‐based subgroups (Hartmann's procedure, n = 10; resection with primary anastomosis, n = 9), between‐group comparisons are presented descriptively; Fisher's exact test and the Mann‐Whitney test were used only for the two prespecified overall‐cohort outcomes discussed in the Results (length of hospital stay and 30‐day morbidity), and results for these should be interpreted with appropriate caution given the small subgroup sizes. All statistical calculations were performed using Stata 18 SE.

3. Results

Between 2020 and 2023, 84 patients underwent emergency Hartmann's procedure or left hemicolectomy at our institution, of which 31 were approached robotically and 49 were performed using a laparoscopic or open technique. Of the 31 robotic cases, 12 were excluded from analysis: 3 for a preoperative or intraoperative diagnosis of malignancy, 5 for an operative indication other than diverticulitis (ischaemic colitis, diverticular bleeding, or vascular perforation), and 4 for a delay from diagnosis to surgery exceeding 72 hours. This left 19 patients in the final analytic cohort (Figure 2).

FIGURE 2.

Patient inclusion and exclusion flow diagram, from all emergency/urgent robotic cases (N = 31) to the final analytic cohort (N = 19).

Patient demographics, disease acuity, and intraoperative details are shown in Table 1. The majority of patients were men (68.4%), and the median age was 67 years (IQR 27). The median BMI was 25.1 kg/m² (IQR 6.2), and most patients had an ASA score of 3 or higher (52.6%). Previous abdominal surgery had been performed in 52.6% of patients. By WSES classification, 4 patients (21.1%) were stage 2a, 7 (36.8%) were stage 2b, 5 (26.3%) were stage 3, and 3 (15.8%) were stage 4; generalised purulent or faecal peritonitis was present in 7 patients (36.8%) and absent in 12 (63.2%). Time from diagnosis to surgical incision was within 24 hours in 14 patients (73.7%), 24–48 hours in 2 patients (10.5%), and 48–72 hours in 3 patients (15.8%).

TABLE 1.

Patient demographics, disease acuity, and intra‐operative details.

Parameter	Value
Sex:
Male	13 (68.4%)
Female	6 (31.6%)
Age, median (IQR) (years)	67 (27)
Body mass index, median (IQR) (kg/m2)	25.1 (6.2)
ASA score:
1	1 (5.3%)
2	8 (42.1%)
3	5 (26.3%)
4	5 (26.3%)
Previous abdominal surgery	10 (52.6%)
Disease acuity:
WSES classification 2a	4 (21.1%)
WSES classification 2b	7 (36.8%)
WSES classification 3	5 (26.3%)
WSES classification 4	3 (15.8%)
Generalized purulent/faecal peritonitis	7 (36.8%)
Time from diagnosis to incision ≤ 24h	14 (73.7%)
Time from diagnosis to incision 24–48h	2 (10.5%)
Time from diagnosis to incision 48–72h	3 (15.8%)
Procedure:
Hartmann's procedure	10 (52.6%)
Resection with primary anastomosis	9 (47.4%)
Total OR time, median (IQR) (minutes)	235 (55)
Total procedure time (skin‐to‐skin), median (IQR) (minutes)	171 (42)
Splenic flexure mobilisation	10 (52.6%)
Estimated blood loss, median (IQR) (mL)	50 (140)
Intraoperative complication	0 (0%)
Conversion to open surgery	0 (0%)

Note: Values are n (%) unless otherwise indicated.

Abbreviations: ASA, American Society of Anaesthesiologists; IQR, interquartile range; WSES, World Society of Emergency Surgery.

Hartmann's procedure was performed in 10 of 19 patients (52.6%), and resection with primary anastomosis in 9 patients (47.4%). The median total procedure (skin‐to‐skin) time was 171 minutes (IQR 42), and the median total OR time (including anaesthesia induction, positioning, and docking) was 235 minutes (IQR 55). Splenic flexure mobilisation was performed in 52.6% of cases, and the specimen was most often extracted through a suprapubic Pfannenstiel incision (89.4%). No intraoperative complications occurred, and no conversions to open surgery were necessary. The majority of bedside assistants were fellows or surgeons in training (84.2%). The median estimated blood loss was 50 mL (IQR 140).

Postoperative outcomes are shown in Table 2. Thirty‐day overall morbidity occurred in 9 patients (47.4%), classifying each patient by their single highest Clavien‐Dindo grade, minor complications (Clavien‐Dindo I‐II) occurred in 3 patients (15.8%) and major complications (Clavien‐Dindo ≥ III) in 6 patients (31.6%). Surgery‐related morbidity, as defined above, occurred in 4 patients (21.1%). No anastomotic leak was reported among patients who underwent primary anastomosis. Postoperative wound infection occurred in 1 patient (5.3%).

TABLE 2.

Postoperative outcomes.

Post‐operative outcome	Value
30‐day overall morbidity	9 (47.4%)
Clavien‐Dindo I‐II (minor)	3 (15.8%)
Clavien‐Dindo ≥ III (major)	6 (31.6%)
Surgery‐related morbidity	4 (21.1%)
Anastomotic leak	0 (0%)
Wound infection	1 (5.3%)
ICU admission after surgery	15 (78.9%)
Days to flatus, median (IQR)	2 (1)
Days to first bowel movement, median (IQR)	3 (1)
Days to full mobility, median (IQR)	4 (5)
Days to soft diet, median (IQR)	2 (2)
Length of hospital stay (days), median (IQR)	11 (15)
30‐day readmission	3 (15.8%)
30‐day reoperation	1 (5.3%)
30‐day mortality	2 (10.5%)

Note: Values are n (%) unless otherwise indicated. ICU: intensive care unit. Each patient is classified under the single highest Clavien‐Dindo grade reached; overall and grade‐specific morbidity therefore sum correctly to 9 patients (47.4%).

Seventy‐eight point 9% of patients were admitted to the ICU postoperatively. The median length of hospital stay was 11 days (IQR 15). The median time to first flatus was 2 days (IQR 1), to first bowel movement 3 days (IQR 1), to full mobility 4 days (IQR 5), and to resumption of a soft diet 2 days (IQR 2). Thirty‐day readmission occurred in 3 patients (15.8%), for reasons including an abdominal collection requiring CT‐guided drain placement, an episode of ileus treated with nasogastric tube insertion, and one case of delayed perforation with septic shock; this last patient was the only one requiring reoperation within 30 days of the index procedure.

Two patients died within 30 days of surgery. Both were operated on under the intention‐to‐treat working diagnosis of acute perforated diverticulitis, consistent with our exclusion criterion of known preoperative colorectal malignancy rather than any malignancy whatsoever. The first patient died of a pulmonary embolism and was subsequently found, on postoperative histopathology, to have a stage T3N1 perforated sigmoid adenocarcinoma that had not been suspected preoperatively; because the clinical and radiological presentation was indistinguishable from complicated diverticulitis, a standard diverticular resection was performed rather than a formal oncologic resection with high vascular tie and complete mesocolic excision, and this patient's death was attributed to pulmonary embolism rather than to the malignancy itself. The second patient died of a pelvic abscess in the setting of known stage IV lung cancer under palliative, concurrent chemotherapy—a pre‐existing comorbidity that was not the indication for abdominal surgery.

Descriptive comparisons according to procedure type (resection with primary anastomosis vs. Hartmann's procedure) are presented in Table 3; given the small subgroup sizes (n = 9 and n = 10), p‐values for most comparisons have been omitted, and these findings should be interpreted descriptively rather than as evidence of a statistical difference between groups. For the two prespecified overall outcomes assessed statistically, patients undergoing Hartmann's procedure had a longer median length of hospital stay (19 days [IQR 20] vs. 10 days [IQR 5], p = 0.03) and a higher rate of 30‐day morbidity (70.0% vs. 22.2%, p = 0.04) compared with those undergoing resection and primary anastomosis; these comparisons should be interpreted with caution given the limited sample size.

TABLE 3.

Descriptive comparison of outcomes by procedure type: resection with primary anastomosis versus Hartmann's procedure.

Variable	Primary anastomosis (N = 9)	Hartmann's procedure (N = 10)
Male, n (%)	5 (55.6)	8 (80.0)
Female, n (%)	4 (44.4)	2 (20.0)
Age, median (IQR), years	65 (13)	68 (27)
Body mass index, median (IQR), kg/m2	26.6 (6.7)	23.9 (6.4)
Previous abdominal surgery, n (%)	6 (66.7)	4 (40.0)
Total OR time, median (IQR), minutes	270 (70)	230.5 (49)
Total procedure time (skin‐to‐skin), median (IQR), minutes	171 (30)	171 (40)
Splenic flexure mobilisation, n (%)	7 (77.8)	3 (30.0)
Estimated blood loss, median (IQR), ml	40 (110)	50 (100)
Length of hospital stay (days), median (IQR)	10 (5)	19 (20)
30‐day overall morbidity, n (%)	2 (22.2)	7 (70.0)
Surgery‐related morbidity, n (%)	1 (11.1)	3 (30.0)
Complication with Clavien‐Dindo ≥ 3, n (%)	2 (22.2)	3 (30.0)
Days to flatus, median (IQR)	2 (1)	2 (1)
30‐day readmission, n (%)	1 (11.1)	2 (20.0)
30‐day mortality, n (%)	1 (11.1)	1 (10.0)

Note: Values are n (%) unless otherwise indicated. p‐values have been omitted due to the small subgroup sample sizes, which preclude robust statistical comparison; these data should be interpreted descriptively.

Abbreviations: IQR, interquartile range; OR, operating room.

4. Discussion

Robotic surgery in the emergency setting remains in an exploratory phase, with most published literature limited to case reports and small case series [5, 11]. Our initial experience with 19 patients contributes descriptive, hypothesis‐generating data towards defining the feasibility and preliminary safety profile of an emergency robotic programme, though it cannot establish safety or superiority given its retrospective, uncontrolled, single‐centre design. In this cohort, the programme demonstrated a 0% conversion rate and morbidity, reoperation, and mortality rates broadly in the range reported for other emergency approaches to complicated diverticular disease, suggesting that the robotic platform may be a technically feasible option for selected urgent and emergency cases; confirmation in prospective, comparative studies is required before broader conclusions can be drawn.

Two multicenter randomized controlled trials (Binda et al. and Oberkofler et al.) compared resection with primary anastomosis (with or without ileostomy) versus Hartmann's procedure in patients with Hinchey III‐IV diverticulitis. These trials, using laparoscopic and open approaches, reported similar morbidity rates between the two procedures (35%–80% for primary anastomosis and 46%–84% for Hartmann's), with no statistical difference [12, 13]. Our robotic cohort's overall morbidity of 47.4% (reduced to 21.1% when considering only surgery‐related complications) falls within this previously reported range. This finding is of interest given the complexity of our patient population, in which over half of patients had an ASA score ≥ 3, indicating significant comorbidity burden and potentially increased surgical risk; however, the absence of a concurrent comparator cohort in our own study means this comparison to historical literature cannot substitute for a direct, adequately powered comparative analysis.

Another notable observation is the 0% conversion rate in our robotic series. Reported conversion rates for laparoscopic emergency colorectal surgery range from 5.3% to 12.2% in the literature [14, 15, 16]. While unplanned conversion to open surgery has been associated with worse patient outcomes, and robotic approaches have shown lower conversion rates than laparoscopy across several abdominal procedures [17, 18, 19, 20, 21, 22], our small, single‐arm sample size precludes any definitive statement that the robotic platform reduces conversion or postoperative morbidity in this population; this remains a hypothesis for future comparative studies.

A recent systematic review by Shaban et al. [23] analysed mortality after emergency surgery for complicated acute diverticulitis across 14 studies, reporting mortality ranging from 2.9% to 66.7% for resection with primary anastomosis and from 8.7% to 60% after Hartmann's procedure. Our cohort's mortality rate of 10.5% falls towards the lower end of this range; however, given the small sample size and absence of a concurrent comparator, this observation should be considered exploratory and hypothesis‐generating rather than evidence of a mortality benefit associated with the robotic approach.

Several important limitations must be acknowledged. First, this is a retrospective, single‐centre study without a concurrent comparator group; although 49 non‐robotic (laparoscopic or open) cases were performed during the same period, they were not included in this descriptive analysis of the robotic pathway, and we have not attempted a formal comparison here. This omission limits the interpretability of our findings and introduces the risk that observed differences reflect patient or surgeon selection rather than a true treatment effect. Second, assignment to the robotic approach was not randomized or protocol‐driven, but depended on the on‐call availability of the two robotically trained surgeons; this surgeon‐availability‐dependent selection introduces an inherent selection bias that cannot be fully accounted for in a retrospective design. Third, our granular data did not consistently capture separate docking and console times, precluding a more detailed breakdown of operative time beyond total OR time and total skin‐to‐skin procedure time; future prospective work from this programme should capture these intervals separately. Fourth, the small subgroup sizes in our procedure‐based comparison (n = 9 and n = 10) make between‐group statistical testing unstable, and most p‐values have therefore been omitted from Table 3. Fifth, this study involved two highly experienced robotic surgeons, a level of expertise not universally available, which may limit generalisability to other centres. Finally, scaling a true 24/7 emergency robotic service requires substantial institutional investment: continuous platform availability, a dedicated pool of robotically trained on‐call OR staff (scrub nurses, first assists, and anaesthesiologists), and the opportunity cost of reserving robotic theatre time for unscheduled cases, all factors that may limit the scalability of this model to centres without a critical mass of trained personnel and institutional support.

This study describes the preliminary feasibility of a dedicated robotic surgery program for treating acute complicated diverticulitis in a small, uncontrolled, single‐center cohort. The 0% conversion rate and morbidity/mortality rates broadly consistent with the published range for other approaches are encouraging but hypothesis‐generating only. Larger, prospective studies with concurrent comparator arms are needed to confirm these findings and to define the role, safety, and cost‐effectiveness of robotic surgery in managing acute complicated diverticulitis before it can be recommended as a preferred approach.

Author Contributions

E.V.E. and T.V. contributed equally to this work. T.V. and E.V.E. drafted the manuscript. S.V., L.V.P., J.V.Z., M.V., and M.R. created tables and figures and collected data. M.R. and J.S. supervised the project. All authors reviewed and approved the final manuscript.

Funding

The authors have nothing to report.

Ethics Statement

This study was approved by the Institutional Review Board of Universitair Ziekenhuis Brussel (BUN B1432023000299).

Consent

All patients signed an authorisation form.

Conflicts of Interest

The authors declare no conflicts of interest.

Permission to Reproduce Material From Other Sources

The authors have nothing to report.

Data Availability Statement

Data are available from the corresponding author upon reasonable request.