Short‐term outcomes between robot‐assisted and open pancreaticoduodenectomy in patients with high body mass index: A propensity score matched study

Abstract Background High body mass index was considered as a risk factor for minimally invasive surgery. The short‐term outcomes of robot‐assisted pancreaticoduodenectomy (RPD) remain controversial. This study aims to investigate the feasibility and advantage of RPD in patients with high body mass index compared to open pancreaticoduodenectomy (OPD). Methods Clinical data of 304 patients who underwent pancreaticoduodenectomy from January 2016 to December 2019 in Ruijin Hospital, Shanghai Jiao Tong University School of Medicine was collected. Patients with BMI >25 kg/m2 were included and divided into RPD and OPD group. After PSM at a 1:1 ratio, 75 patients of OPD and 75 patients of RPD were recorded and analyzed. Results The RPD group showed advantages in the estimated blood loss (EBL) (323.3 mL vs. 480.7 mL, p = 0.010), the postoperative abdominal infection rate (24% vs. 44%, p = 0.010), the incidence of Clavien‐Dindo III‐V complications (14.7% vs. 28.0%, p = 0.042) over OPD group. Conclusion RPD shows advantages in less EBL, lower incidence rate of Clavien‐Dindo III‐V complications over OPD in overweight and obese patients. RPD was confirmed as a safe and feasible surgical approach for overweight or obsess patients.


| INTRODUCTION
Overweight and obesity are now considered nonnegligible health problems worldwide. By 2015, obesity had been diagnosed in over 107 million children and 603 million adults, 1,2 and approximately 58% of people are expected to be overweight or obese by 2030. 3 Overweight and obesity are defined as abnormal or excessive fat accumulation that may impair health. Body mass index (BMI) is a simple index of weight-for-height that is commonly used to classify overweight and obesity in adults. It is defined as a person's weight in kilograms divided by the square of his height in meters (kg/m 2 ). World Health Organization (WHO) defines overweight and obesity as follows: overweight is a BMI ≥25 and obesity is a BMI ≥30. 4 The definition of obesity was BMI > 28 kg/m 2 in China relatively. Obesity is regarded as a risk factor for some chronic diseases, such as Type 2 diabetes, hypertension, dyslipidemia, and coronary heart disease. 5 In recent years, studies have found that obesity is an independent risk factor for the development of pancreatic and periampullary tumors. [6][7][8][9][10] As reported in previous studies, an increase of 5 kg/m 2 in BMI is associated with a 10% or even higher risk of developing pancreatic neoplasms. [11][12][13] Based on the data of our center, the percentages of overweight and obese patients with pancreatic neoplasms increased continuously from 2016 to 2019 from 16.5% to 24.7%. More proportion of overweight and obese patients would suffer from pancreatic neoplasms due to metabolic disorders.
Pancreaticoduodenectomy is still the only radical cure for pancreatic head neoplasms and periampullary tumors. 14 With the advancement of technology, minimally invasive pancreatic resection, especially robot-assisted pancreaticoduodenectomy (RPD), has gained popularity. Meanwhile, it has been shown to be safe and feasible not only for benign pancreatic tumors but also for malignant tumors. [15][16][17] RPD was proved with less intraoperative estimated blood loss (EBL) and fewer postoperative complications resulting in rapid recovery compared to traditional open pancreaticoduodenectomy (OPD). 17,18 However, overweight and obesity were considered to be associated with higher perioperative complications and inappropriate for pancreatic surgery. The high BMI was proved as an independent risk factor to postoperative pancreatic fistula (POPF), wound infection, and delayed gastric emptying (DGE). [19][20][21] Several reasons included excessive adipose tissue, decreased operative space, and omental or mesenteric thickening [19][20][21] which were supposed to explain the different short-term outcomes between normal and overweight or obese patients. Compared to OPD, it was still unclear whether the effect of high BMI on the perioperative outcomes of RPD. This article aims to confirm the feasibility of RPD and to investigate the advantages of RPD compared to OPD for overweight and obese patients.

| Patient selection and surgical procedure
A total of 304 patients who underwent OPD and RPD from January 2016 to December 2019 in the Department of General Surgery, Pancreatic Disease Center, Ruijin Hospital Shanghai Jiao Tong University School of Medicine, Shanghai, China, were included in this study. This study was approved by the Ruijin Hospital Ethics Committee. Every patient was asked to sign the informed consent to make sure that they agreed with the operation and the use of data we collected before and after surgery. The study was undertaken according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines. 22 The requirement for informed consent was waived due to our retrospective study. Preoperative computed tomography (CT) and magnetic resonance imaging were routinely performed for all patients. Endoscopic ultrasonography would be applied if the diagnosis of tumor remained unclear. The diagnosis of the tumor and surgical protocol was determined by our multidisciplinary team. All these surgeries were performed by the same group of surgeons in our center which had previous experience (>300 cases). All RPD were carried out by the da Vinci Si Surgical System, Intuitive TM, five-port technique 23 and 10-step operative technique 24 with experienced professors who had passed the learning curve. The inclusion criteria and exclusion criteria are as follows. 304 patients were then enrolled. There were three cases of conversion in the RPD group.

| Inclusion criteria
1. Patients with BMI ≥25 kg/m 2 . 2. Patients underwent OPD or RPD and were pathologically diagnosed with pancreatic neoplasm located in the head of the pancreas or periampullary area. 3. The neoplasms were resectable conforming to the consensus proposed by National Comprehensive Cancer Network.

| Exclusion criteria
1. Preoperative suspicious metastasis or unexpected involvement of major vessels were proved intraoperatively 2. History of upper-abdomen surgery 3. Robotic-assisted surgery converted to open laparotomy intraoperatively The patient characteristics including age, sex, BMI, hemoglobin level, diabetes, bilirubin level, albumin level, American Society of Anesthesiologists score (ASA), alcohol/smoking history and pathological results were recorded. We used propensity score matching (PSM) to minimize the selection bias caused by different characteristics of the patients. After PSM at a 1:1 ratio, 75 pairs of patients were finally included in this study. The study flowchart is shown in Figure 1.

| Matching and data acquisition
When comparing two groups, inevitable bias may occur due to treatment selection. Propensity score matching (PSM) is a good method to minimize the bias. As a statistical method used to process data from observational studies, it is designed to reduce the effects of selection bias and confounding variables in order to make more reasonable comparisons between experimental and control groups. We considered there was no significant difference between two groups when the value was less than 0.1. We collected patient data from our own database. The propensity score was calculated based on the covariates age, sex, BMI, ASA anesthesia score, and tumor type, and then paired one to one into two groups based on surgical type.
All outcomes were defined according to the International Study Group on Pancreatic Surgery (ISGPS). 25 The perioperative outcomes included POPF, post-pancreatectomy hemorrhage (PPH), bile leakage, infection, DGE, severe complications, reoperation, operation time, estimated blood loss, postoperative length of stay (LOS), and 90-day mortality. The "Infection" is defined as postoperative signs of peritonitis, manifestations of incisional infection, the presence of positive pathogenic bacteria on culture, and imaging or surgical confirmation of the presence of infectious lesions in the abdominal cavity or under the incision, such as purulent exudate and localized abscess necrosis. The "severe complications" are defined due to Clavien-Dindo criterion, including complications of Grade III and above, such as reoperation, abdominal gastrointestinal fistulas, effusions, and infections requiring interventional procedures, DGE requiring endoscopic treatment, and other serious lifethreatening postoperative complications and death.
Long-term outcomes were not taken into account.

| Statistical analysis
We used the computer software SPSS 26.0 for Windows (IBM Corp.) to do all statistical analyses. We use medians and interquartile ranges or means and standard deviations to describe continuous data, and use numbers and percentages to describe categorical data. Propensity score was calculated for each patient by using logistic regression modeling. The variables used for matching including: age, sex, BMI, Hb level, Alb level, bilirubin level, ASA score, diabetes history, smoking history, alcohol history, tumor size and pathology. Then we set a 0.01 SD as the caliper width and patients were matched

| Patient characteristics
In the initial enrollment session, we enrolled a total of 228 patients undergoing OPD and 76 patients undergoing RPD. In comparing the baseline data between the two groups, we observed a significant difference in ASA scores (p < 0.05) between the two groups. And because of the large difference in population size between the two groups, we therefore used PSM to complete 1:1 matching. After matching, each group had 75 patients. In terms of age, sex, BMI, hemoglobin level, albumin level, bilirubin level, ASA score, tumor size, pathology results, diabetes, smoking and alcohol history, there were no significant differences between 2 groups after PSM. In the OPD group, 45 patients (60.0%) were men, the mean age was 61.5 ± 9.6 years, and the mean BMI was 27.1 ± 2.01 kg/m 2 . In the RPD group, 45 patients (60.0%) were male, the mean age was 61.8 ± 10.3 years, and the mean BMI was 27.1 ± 2.9 kg/m 2 . The patient characteristics are displayed in Table 1.

| Perioperative outcomes
There were three cases of conversion in the RPD group. One patient with BMI at 25.5 kg/m 2 and one patient with BMI at 31.1 kg/m 2 underwent conversion of RPD because of unexpected intense contact between the tumor margin and superior mesenteric artery (SMA)/superior mesenteric vein (SMV). One patient with BMI at 29.3 suffered intraoperative bleeding of during the resection of uncinate process of the pancreas and received OPD considering the perioperative safety. The perioperative outcomes of patients in RPD group were compared to those in OPD group and are shown in Table 2. Tche EBL was significantly lower in the RPD group than in the OPD group (323.3 vs. 480.7 mL, p = 0.010). The operation time and postoperative complications, including POPF, bile leakage, PPH, DGE, and reoperation, showed no differences between the 2 groups ( Table 2). The occurrences of infection and severe complications (Clavien-Dindo Grade ≥ III) were significantly higher in the OPD group than in the RPD group (44.0% vs. 24.0%, p = 0.010; 28.0% vs. 14.7%, p = 0.046). The occurrence of 90-day mortality was lower in the RPD group (1.3% vs. 8.0%, p = 0.042). Six patients died within 90 days after surgery in the OPD group. The reasons included hemorrhage in two patients, severe abdominal infection in one patient, liver metastasis in one patient, and dyscrasia in the other two patients. Only one patient, who was in the RPD group, died on the postoperative day fifth because of an unexpected cardiopulmonary failure.

| DISCUSSION
Obesity and overweight are thought to be associated with the development of pancreatic neoplasms and periampullary tumors. [6][7][8] According to previous reports, obesity also has an impact on patients undergoing PD surgery. 19,26,27 Shengliang He proposed that RPD was associated with decreased blood loss and shorter hospital stays in obese patients. 18 Peng L indicated that the wound infection rate was lower in RPD than OPD. 17 However, their conclusion was limited by their sample size and there were few reports or studies that compared the effects of obesity on perioperative outcomes in RPD versus OPD. Our propensity score matched study showed that the RPD group had advantages compared to the OPD group in several terms, illustrating the considerable advantage of RPD in reducing surgical trauma in obese and overweight patients. Tjeertes EK found that obesity is a significant risk factor for postoperative abdominal infection, greater T A B L E 2 Intraoperative and perioperative outcomes before and after propensity score matching.

Before propensity score matching
After propensity score matching intraoperative blood loss, and a longer operation time. 27 The possible reasons were as follows: (1) Excessive subcutaneous fat tissue predisposes these patients to impaired healing due to low regional perfusion and oxygen tension. 28 (2) A longer operation time has been described as a significant predictor of postoperative wound infections. 29,30 (3) Impaired immunity, elevated blood glucose levels, and too much tension in the surgical incision are also factors that contribute to impaired wound healing 31,32 (4) Anatomical difficulties caused by excessive fat distribution. Meanwhile, the smaller incisions and more delicate manipulations of robot-assisted surgery, as well as faster recovery, can bring significant advantages. 33,34 In our study, the EBL was lower in the RPD group than in the OPD group. For overweight and obese patients, the level of surgical challenge increases due to their complicated regional anatomy of the duodenum, SMA and SMV. More precise manipulation should be made because of the vascular fragility and unexpected bleeding during stretch of the mesentery. On the contrary, robot-assisted surgical systems can filter hand tremor; thus, for obese patients with brittle tissues, a robotic approach may be a better choice than open surgery to ensure that accidental injury is avoided. Especially when dealing with vessels, robotassisted surgery may help prevent damaging the vascular wall and reduce the incidence of bleeding. Better field exposure and higher magnification during robot-assisted surgery can also effectively reduce estimated blood loss. Furthermore, increased EBL was well-documented to be associated with higher morbidity in terms of postoperative complications which could explain the lower incidence of Complications of Grade (≥ III) and abdominal infection. 35,36 We proposed that the robot-assisted approaches can benefit the patients with less EBL and lower postoperative severe complications by more flexible and precise manipulation in overweight and obese patients.
In our study, the operation time was not different between the two groups. In previous studies and reports, robot-assisted surgery was reported with much longer time than traditional open surgery. Reasons may include the longer preparation times needed for installation and because the procedures and processes are not as skilled as those performed during traditional surgery. Based on the previous experimental results obtained at our center with the standardized surgical procedure, ineffective intraoperative manipulations can be significantly reduced, thus achieving a stable operation time and a low level of surgical difficulty. 37 This modular surgical concept and approach has nearly eliminated time-consuming missteps during RPD operations, and the surgical procedure can be modified by adjusting the surgical path and sequence for overweight and obese patients individually and feasibly.
POPF represents one of the most common and serious complications. Buchs et al showed that more accurate surgical process and less manipulation around the residual pancreas in RPD could decrease the incidence of POPF. 38 Patricio M. Polanco et al 39 and Bing-Yang Hu et al 40 pointed out that higher BMI was relevant with the POPF which required more precise surgical procedure. There was no significant difference of the incidence of POPF between the two groups in our study. We proposed that with the current surgical technique, RPD was safe for overweight and obese patients.
The advantages of RPD versus OPD in overweight and obese patients may be concluded as following. The operators of robot-assisted surgery have a more comfortable operating space. During the operation, there was no need to open and close the large abdominal incision, thereby simplifying the operation procedure, reducing the level of difficulty of the exposure process, and decreasing morbidity due to infection. However, there are inevitable challenges in performing robot-assisted surgery in overweight and obese patient population. First, thicker subcutaneous fat requires precise establishment of the Trocar orifice which is prone to slip out and affects the size of the intra-abdominal space after pneumoperitoneum establishment. Besides, overweight and obese patients have a thicker omentum and mesentery, making them more prone to bleeding and damage to the intestinal segment during surgery when separating the transverse colon and duodenum. In addition, fatty pancreas leads to increased difficulty in suturing the pancreatic-enteric anastomosis, which has a significant impact on the occurrence of POPF. The learning curve for RPD, especially with those overweight and obese patients requires further prospective studies to verify based on our conclusion and recent study.
There are several limitations to our study. First, there was still bias associated with patient selection and the retrospective nature of the research although PSM method was made. Second, BMI is widely used to distinguish overweight and obese people, visceral fat which was more complex to calculate by CT scan was proposed as a prior measure than BMI for clinical researches. In addition, PSM is statistically inferior to randomized studies. Further prospective randomized clinical trials should be designed and carried out to better understand the benefits of robot-assisted surgery in overweight and obese populations.

| CONCLUSION
In this study, we compared RPD surgery with OPD surgery in a population of overweight and obese patients. We found that RPD has advantages in less EBL, lower incidence rate of Clavien-Dindo III-V complications over OPD. RPD was confirmed as a safe and feasible surgical approach for overweight or obsess patients.

| FUTURE OUTLOOK
We plan to conduct a randomized clinical trial in the future on the treatment of obese patients with robot-assisted pancreatic surgery. The definition of obesity will not be limited to BMI, but will be defined by assessing the patient's visceral fat, subcutaneous fat, pancreatic steatosis, and other aspects. We also hope to conduct long-term prognostic follow-up and to provide evidence for clinical management through survival analysis.

FUNDING INFORMATION
This study was funded by the National Natural Science Foundation of China (81672325).

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon reasonable request.

ETHICS STATEMENT
This study was approved by the institutional review board of Shanghai Ruijin Hospital.