The practice of bariatric surgery has risen dramatically with the worsening epidemic of obesity. Laparoscopic sleeve gastrectomy (LSG) and laparoscopic Roux-en-Y gastric bypass (LRYGB) are the two most commonly performed bariatric procedures [1]. Experience in these two procedures has been extensive and exceeds 10 years in LSG and 25 years in LRYGB [1]. During these years, surgeons have been studying several modifications to the surgical technique in both procedures; one of which is the utility of intraoperative endoscopy (IOE). Several publications have recommended the routine use of IOE to decrease complications such as leak, stenosis, and bleeding following either LSG or LRYGB [2,3,4,5,6,7]. Haddad et al. [6] reported the utility of IOE in detecting and curing leaks during LRYGB in 96% of patients who had a sustained positive intraoperative leak test. Nimeri et al. [4] concluded that IOE could aid in the prevention of LSG postoperative stenosis through endoscopic intraoperative detection and removal of over-sewing sutures. However, most of the publications on IOE in bariatric surgery were limited to reporting the experience of individual surgeons and bariatric centers which may not represent the population of bariatric surgeons [4,5,6,7,8,9,10,11,12,13,14,15,16]. Thus, many of the assumptions regarding the use of IOE might not be firmly evidence-based. This was reflected in a statement by The American Society for Metabolic and Bariatric Surgery (ASMBS) in 2015 on the prevention of gastrointestinal leaks after LSG and LRYGB where it has not found any evidence to support the use of IOE for the reduction of leaks [17]. Considering the limitation of previous studies, we aim to compare outcomes of bariatric surgeries with IOE to surgeries without IOE using a large sample of patients that would be representative of the national outcomes.

Methods

Study design

Our study was based on the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) prospective database. ACS-NSQIP includes data on over 130 preoperative, intraoperative, and postoperative variables that are prospectively collected by academic and community hospitals (680 participating sites in 2016 and 315 sites in 2011). To ensure high quality of collected data, ACS-NSQIP has implemented several mechanisms to confirm reliability and consistency of the data [18, 19]. Due to the de-identified nature of the NSQIP database, no institutional review board (IRB) authorization is required.

Inclusion/exclusion criteria

Using Current Procedural Terminology (CPT), LSG patients (CPT code 43775) and LRYGB (CPT codes 43644 and 43645) were identified in ACS-NSQIP database from years 2011 till 2016. Patients with body mass index (BMI) ≥ 35 kg/m2 were included. We excluded emergency cases, patients with disseminated cancer, preoperative sepsis, and concomitant procedures, namely cholecystectomy and removal of gastric band as these might increase the risk of postoperative complications [20, 21].

Exposure

Patients were then categorized based on IOE, which was identified using CPT codes 43234, 43235, 43236, and 43239 described in Table 1. Baseline characteristics, preoperative laboratory values, intraoperative parameters, and postoperative complications in patients who have been endoscoped were compared to patients who have not been endoscoped. The analysis was separately done for LSG and LRYGB.

Table 1 Description of intraoperative endoscopy CPT codes
Full size table

Outcomes

Thirty-day postoperative outcomes that were studied include sepsis (sepsis and septic shock), organ space infection, bleeding, unplanned reoperation, unplanned readmission, mortality, and prolonged length of stay (PLOS). PLOS was defined as length of hospital stay greater than or equal to 3 days (90th percentile). A composite outcome was also studied and included complications such as sepsis, organ space infection, bleeding, unplanned reoperation, and mortality.

Statistical analysis

Continuous variables were presented by means and standard deviations, while categorical variables were presented as frequencies and percentages. Student t test was used for the comparison of continuous variables, and Pearson χ2 test or Fisher’s exact test was used for categorical variables. Binary multiple logistic regression of outcomes, using forward stepwise method, was used to adjust for age, gender, BMI, ASA ≥ 3, diabetes, hypertension, and operative time. Statistical significance was set at the 5% level. Statistical analysis was performed using SAS, version 9.2 (SAS Institute, Cary, NC, USA).

Results

Out of 62,805 cases of LSG and 50,047 cases of LRYGB, 11,254 cases (17.9%) and 9,854 cases (19.7%) had IOE, respectively. Mean age in endoscoped LSG patients was slightly lower compared to non-endoscoped (43.7 vs. 44.9 years, p < 0.003). Mean age in endoscoped LRYGB was similar to non-endoscoped (44.7 vs. 44.6 years, p = 0.16). Mean BMI in IOE patients was equivalent to non-IOE (rounded to 46 kg/m2 in LSG and 47 kg/m2 in LRYGB). The prevalence of American Society of Anesthesiologists (ASA) classification more than III was slightly higher in IOE group (LSG: 72.9% vs. 70.1%, p < 0.001; LRYGB: 78.9% vs. 73.8%, p < 0.001). Operative time of LSG increased on average by 9.4 min when IOE was performed (93.6 min vs. 84.2 min, p < 0.001). In LRYGB, operative time was slightly greater by 2 min with IOE (130.3 min vs. 132.3 min, p = 0.002). All other preoperative comorbidities and laboratory values in LSG and LRYGB were clinically comparable between IOE and non-IOE group (Table 2).

Table 2 Baseline characteristics and preoperative laboratory values by IOE status for LSG and LRYGB patients
Full size table

Postoperative outcomes of LSG are presented in Table 3. Endoscoped LSG patients had lower sepsis rate (0.21% vs. 0.37%, p = 0.009), prolonged hospital stay (14.0% vs. 14.9%, p = 0.014), unplanned reoperation (0.38% vs. 0.58%, p = 0.006), and composite complications (1.17% vs. 1.43%, p = 0.02). Multivariate logistic regression was performed to adjust for confounding factors such as age, BMI, gender, ASA, diabetes, hypertension, and operative time. IOE in LSG independently decreased sepsis [adjusted odds ratio (AOR) = 0.55 (0.36, 0.84)], unplanned reoperations [AOR = 0.61 (0.44, 0.85)], prolonged length of stay [AOR = 0.87 (0.82, 0.92)] and composite complications [AOR = 0.78 (0.65, 0.94)]. Other LSG complications such as organ space infection, bleeding, and mortality showed no significant difference between the two groups at the crude and adjusted analysis.

Table 3 Postoperative complications for IOE vs. non-IOE among LSG patients
Full size table

All postoperative complications of LRYGB, except for prolonged length of stay, were comparable between IOE and non-IOE groups of patients (Table 4). Prolonged length of hospital stay in LRYGB was lower in endoscoped patients (20.6% vs. 22.4%, p < 0.001), and similarly lower after multivariate analysis [AOR = 0.89 (0.84, 0.94)].

Table 4 Postoperative complications for IOE vs. non-IOE among LRYGB patients
Full size table

Discussion

Routine use of IOE has been supported by some bariatric surgeons to decrease the incidence of gastrointestinal leakage, bleeding, and stenosis [2,3,4,5,6,7]. Most of the available literature on the utility of IOE in bariatric surgery represents a small patient population and a single center experience [4,5,6,7,8,9,10,11,12,13,14,15,16]. Our study aims to estimate the utilization of IOE in the practice of bariatric surgery and test its effectiveness in decreasing postoperative complications using a large representative sample of patients.

Our results report that 17.9% of LSG and 19.7% of LRYGB cases during the period of 2011 till 2016 were assisted with endoscopy. This implies that IOE, while utilized by a considerable portion of bariatric surgeons across the US, is still underutilized by the majority. The outcomes also show a statistically significant decrease in complications after endoscopy-assisted LSG. IOE was independently associated with decrease in sepsis [0.37% vs. 0.21%, (AOR) = 0.55 (0.36, 0.84)], unplanned reoperations [0.58% vs. 0.38%, AOR = 0.61 (0.44, 0.85)], prolonged length of stay [14.9% vs. 14.0%, AOR = 0.87 (0.82, 0.92)], and composite complications [1.43% vs. 1.17%, AOR = 0.78 (0.65, 0.94)].

The risk of leakage, based on large published series after LSG and LRYGB, has been estimated to be 0.7% and 0.8%, respectively [22]. IOE in LSG has been studied and recommended to be routinely used for air leak test mainly based on expert opinion [3, 4, 23]. In LRYGB, IOE has frequently been studied as well and recommended to test the integrity of the gastric pouch and gastrojejunostomy anastomosis [2, 5,6,7,8,9,10,11,12,13,14]. A study by Alaedeen et al. [2] compared outcomes of IOE in LRYGB to orogastric tube methylene blue test and concluded that IOE is superior in decreasing postoperative leaks. A recently published randomized control trial tested IOE in LRYGB [24]. The study recruited 50 patients in the endoscopy arm and 50 patients in the non-endoscopy arm. The trial reported a significant decrease in postoperative leaks (0% vs. 8%), reoperations (0% vs. 8%), and length of stay with IOE. On the other hand, it also reported more than 30 min increase in operative time with the use of IOE. Despite the significance of the results, outcomes of the trial cannot be generalized due to the variability in surgical technique and consequently in outcomes and complications. For example, the trial reported a leak rate of 8% in the non-endoscopy group which is relatively very high compared to LRYGB literature-reported leak rate at 0.8% [22]. In our study, we did not specifically measure leakage rates as it is not reported in ACS-NSQIP database. Nonetheless, we measured outcomes such as sepsis, mortality, organ space infections, bleeding, unplanned reoperations, and readmissions, which can be indicators of gastrointestinal leakage and stenosis. Our results showed no major difference in outcomes between IOE and non-IOE patients except for a slightly decreased prolonged length of stay. An explanation for the discrepancy between our results and the literature could be that intraoperative endoscopies were not always routinely performed. Some surgeons selectively perform intraoperative endoscopies based on the difficulty and complexity of cases, which could increase complications in IOE group. Therefore, the effect of routine IOE in LRYGB and LSG is expected to be greater as selective IOE might be masking the decrease in complications.

IOE was also reported to have benefits other than prevention of leak. Some surgeons support the use of endoscopy during LSG for the calibration of the gastric tube size instead of bougie and have reported good weight loss outcomes [15, 16, 23, 25, 26]. Other studies have also proposed IOE for the prevention of postoperative stenosis [4, 5]. With the privilege of lumen visualization, IOE also might help in the detection of intraluminal bleeding in LSG or LRYGB, which could be controlled endoscopically or laparoscopically [7, 8, 23]. Endoscopy may also assist in guiding the correct placement of stapling instruments, as the endoscopy light illuminates through the gastric wall [3]. Despite the safety of IOE in bariatric surgery, few complications as iatrogenic injury and lacerations have been reported [6, 16]. Ruiz-Tovar et al. [23] also reported a decrease in major postoperative complications with the use of IOE in comparison with bougie but his study was limited with a small sample size of 50 LSGs.

Our study is the first to report large dataset outcomes of endoscopy-assisted bariatric procedures. It is also the first to estimate the prevalence of IOE utilization in bariatric surgery. A relatively low percentage ranging around 18–20% of bariatric surgeries are assisted with IOE. The lack of strong evidence supporting the utility of IOE might have contributed to its underutilization in bariatric surgery. Our study results powered by a significantly large number of patients show a decrease in complications associated with the utilization of IOE, specifically in LSG. Those results may be generalizable as well to other foregut procedures. Moreover, our study shows that the increase in intraoperative time with IOE is reasonable and lower than previous studies [24].

This study also has some limitations. Our paper focuses on the impact of IOE on postoperative complications but does not study other potential uses of IOE. In addition, the NSQIP database was not designed to be a bariatric specific database, thus lacking parameters including weight loss outcomes and complications such as leak or stenosis. NSQIP also inherently report data up to 30-day postoperatively; therefore, we could not assess late complications. Although our results do not explicitly measure the risk of leak, all our studied outcomes can be direct complications of leakage especially sepsis which is a major and potentially life-threatening morbidity. Large multicenter prospective studies are needed to explore the benefits of IOE in bariatric surgery, particularly the intermediate or long-term benefits beyond 30 days.

Conclusion

IOE is generally underutilized in bariatric procedures. IOE is associated with decreased risk of postoperative complications particularly sepsis, unplanned reoperations, prolonged hospital stay, and composite complications after LSG; and hospital stay after LRYGB. Large multicenter prospective studies are needed to explore the benefits of IOE in bariatric surgery, particularly the intermediate or long-term benefits.