Faculty Bibliography

Introduction: Rapid weight loss after bariatric surgery (BS) has been associated with the formation of gallstones, and subsequent acute cholecystitis and pancreatitis (AP). However, the complex post-surgical anatomy limits the possibility of performing an ERCP as part of AP treatment. Therefore, the aim of this study was to assess the impact of bariatric surgery on mortality, morbidity and resource utilization among patients with AP using a national database. Aims & Methods: This was a case-control study using the National Inpatient Sample 2013, the largest publically available inpatient database in the United States. All patients with an ICD-9 CM code for a principal diagnosis of AP were included. There were no exclusion criteria. Patients with a past history of BS were identified using the appropriate ICD-9CM codes. The primary outcome was all cause mortality. The secondary outcomes were: 1-Morbidity: Intensive care unit (ICU) admission, shock and multi-organ failure. 2- Resource utilization: use of ERCP, cholecystectomy, total parenteral nutrition (TPN), length of hospital stay (LOS), total hospitalization charges and costs. Multivariate regression analyses were used to adjust for the following confounders: Age, sex, race, income in patients' zip code, Charlson Comorbidity Index, hospital region, location, size and teaching status. Results: A total of 274,775 patients with AP were included in the study, of which 4240 (1.7%) had undergone bariatric surgery. The mean patient age was 51 years and 48% were female. After adjusting for confounders, patients with and without history of bariatric surgery had similar adjusted odds of mortality (adjusted Odds Ratio (aOR): 1.37, 95% CI: 0.51-3.65, p=0.52). Looking at morbidity, patients with and without a history of bariatric surgery had similar adjusted odds of shock (aOR: 0.71, 95% CI: 0.23 - 2.24, p=0.57), multiple organ failure (aOR: 0.61, 95% CI: 0.08-4.37, p=0.62), and ICU admission (aOR: 0.34, 95% CI: 0.11 - 1.06, p=0.62). As far as resource utilization, patients with bariatric surgery had lower adjusted odds of ERCP (aOR: 0.55, 95% CI: 0.38 - 0.79, p<0.01), but both patient groups had similar adjusted odds of cholecystectomy (aOR: 1.21, 95% CI: 0.98 - 1.49, p=0.07) and TPN use (aOR: 1.21, 95% CI: 0.77 - 1.90, p=0.41). Interestingly, although patients with bariatric surgery had shorter adjusted length of stay (adjusted mean difference: -0.48 days, 95% CI: -0.74 - -0.23, p<0.01) and total hospitalization costs (adjusted mean difference: - $755, 95% CI: -$1295 to -$215, p<0.01), both patient groups had similar total hospitalization charges (adjusted mean difference: -$1949, 95% CI: -$4366 - $467, p=0.11) Conclusion: Bariatric surgery has no impact on inpatient all-cause mortality among patients who develop acute pancreatitis, despite its association gallstone acute pancreatitis and limited ERCP performance. In addition, bariatric surgery does not affect morbidity in this patient population as measured by rate of shock, multiple organ failure and ICU admission. However, a history of bariatric surgery is associated with decreased rates of ERCP, shorter length of stay and lower hospitalization costs

Introduction: Helicobacter Pylori (H. Pylori) has been implicated in worsening outcomes in patients with hepatic encephalopathy. This is believed to be the result of its urease enzyme that increases the production of ammonia. Small studies so far have yielded contradictory results on whether the presence of H. pylori worsens treatment outcomes in hepatic encephalopathy. Therefore, the aim of this study was to assess the impact of H. pylori on mortality, morbidity and resource utilization among patients with hepatic encephalopathy using a national database. Aims & Methods: This was a case-control study using the National Inpatient Sample 2013, the largest publically available inpatient database in the United States. All patients with an ICD-9 CM code for a principal diagnosis of hepatic encephalopathy were included. There were no exclusion criteria. Patients positive for H. pylori were identified using the appropriate ICD-9CM codes. The primary outcome was all cause mortality. The secondary outcome was resource utilization as measure by use of abdominal imaging (CT scan and ultrasound of the abdomen), length of hospital stay (LOS), total hospitalization charges and costs. Multivariate regression analyses were used to adjust for the following confounders: Age, sex, race, income in patients' zip code, Charlson Comorbidity Index, hospital region, location, size and teaching status. Results: A total of 55,360 patients with hepatic encephalopathy were included in the study, of which 20 had H. pylori infection. The mean patient age was 60 years and 42% were female. After adjusting for confounders using multivariate analysis, patients with and without H. pylori had similar adjusted odds of mortality (adjusted Odds Ratio (aOR): 1.71, 95% CI: 0.62-4.74, p=0.30). As far as resource utilization, patients with and without H. pylori had similar adjusted odds of abdominal imaging (aOR: 3.02, 95% CI: 0.88-10.40, p=0.08), LOS (adjusted mean difference: 1.7 days, 95% CI: -0.02-3.42, p=0.52), and total hospitalization charges (adjusted mean difference: $16588, 95% CI: -$4499 - $37675, p=0.12). However, patients with H. pylori had higher adjusted total hospitalization costs compared with patients without H. pylori (adjusted mean difference: $6128, 95% CI: $1141 - $11115, p=0.01 Conclusion: Presence of Helicobacter Pylori has no impact on inpatient mortality among patients with liver cirrhosis and hepatic encephalopathy. In addition, the presence of Helicobacter Pylori is not associated with any increase in resource utilization among this patient population, with the exception of total hospitalization costs. It is surprising to note that, although total hospitalization costs differed between the two groups, they received the same total hospitalization charges from admitting hospitals

Introduction: Bariatric surgery provides a durable method of weight loss but is associated with serious adverse events. Some studies report an increase in druginduced acute liver injury following bariatric surgery. Aims & Methods: We aimed to assess if bariatric procedures increase the risk of acute liver failure in a large inpatient cohort. We retrospectively analyzed discharge data on patients who developed acute liver injury (ALI) using the Nationwide Inpatient Sample (NIS) database from 2010-2013. Discharges with an ICD-9 code indicating ALI were included. The primary outcome was ALI in patients with a history of bariatric surgery compared to all other patients with an inpatient diagnosis of ALI. Secondary outcomes were mortality in the two cohorts and independent socio-demographic and medical risk factors for mortality in each cohort. Variables tested include age, gender, race, income, Charlson criteria, hospital factors and medical comorbidities including Malnutrition, HTN, Anemia, CKD, Diabetes, CHF, Coagulopathy, Alcoholism, HBV and HCV. Univariate and multivariate logistic regression analyses were performed to identify independent predictors. Results: During the study period, a total of 437,390 patients were diagnosed with acute liver injury and were included in the study, of which 3,799 had previously undergone bariatric surgery. In the post-bariatric cohort, mean age was 58.7 years and 77% were women. The prevalence of acute liver injury in all inpatient admissions for that time period was higher in patients with history of bariatric surgery (0.85%) than in non-bariatric patients (0.75%), p<0.01. Patients with history of bariatric surgery displayed odds ratio of 1.52 of developing ALI when compared to patients with no history of bariatric surgery (95%CI: 1.43-1.61, p<0.01). The rate of overall inpatient mortality was higher in non-bariatric cohort (15.9% versus 9.3%). Post-bariatric patients admitted for ALI were more likely to be younger, female, Caucasian and residing in more affluent areas. Post-bariatric patients were also more likely to have higher rates of malnutrition, anemia, alcoholism, and significantly lower prevalence of hepatitis B and C, CHF, diabetes and kidney disease. In a multivariate regression model, the presence of CHF and coagulopathy increased mortality risk, and diagnosis of alcoholism was associated with lower mortality risk from ALI in patients with prior history of bariatric surgery (Table). Conclusion: Bariatric surgery increases the risk of subsequent acute liver injury. Post-bariatric surgery patients admitted for ALI are more likely to have anemia, malnutrition, and alcoholism, supporting the hypothesis that baseline nutritional status may predispose to drug-induced ALI. Addressing these potentially modifiable risk factors may decrease the significant morbidity and mortality of ALI

Introduction: IGastric stenosis is a rare but potentially serious complication following sleeve gastrectomy, occurring in 0.1%-4% of cases. Surgical revision is invasive and technically challenging. Endoscopic balloon dilatation (EBD) has emerged as a potential treatment option for post-sleeve gastrectomy stenosis (SGS). Aim: To assess the overall event success rate of endoscopic balloon dilation for sleeve gastrectomy stenosis. Methods: Searches of MEDLINE and Embase databases through May 2017 were performed. Examination of titles and abstracts, full review of potentially relevant studies, and data extraction were performed independently by 2 authors. Studies that reported EBD success rates for SGS and had sample size greater than 5 adults were selected. The primary outcome was overall success rate with EBD, expressed as pooled event rate and 95% confidence interval (95% CI). The secondary outcomes were EBD success rates in early (up to 3 months after surgery) and late (3 months or longer post-surgery) SGS; SGS in distal(antrum/body ) or proximal(cardia)location , as well as percent of cases requiring salvage surgery. Random effects metaanalysis was performed for all outcomes. Results: Of the 184 citations identified, 16 studies including 301 patients were eligible for one or more analysis. Among the study participants, the average age was 41.7 years (range 34.6-45.6), average BMI was 40.1 (range 32.9-46.5), with 70.7% female patients. The average number of dilations per person was 1.7. The overall success rate of endoscopic dilation was 78.2% (95% CI 67.7%-85.9%); salvage surgery was required in 14.6% of patients (95% CI 6%-31%). SGS detected before 3 months had higher EBD success rates, 78% (95% CI 52%-92%), p=0.03, than after 3 months, 66% (95% CI 47%-81%), p=0.10. Distal SGS had a EBD success rate of 63% (95% CI 35%-84%), p=0.36, while there were insufficient data for proximal SGS. In two studies, endoscopic stents were used after unsuccessful EBD with an overall endoscopic success rate of 100%. Conclusion: This is the first meta-analysis examining the overall success rate of EBD for treatment of SGS. This method is a safe and effective minimally-invasive alternative to surgical revision, and should be used as first-line therapy for SGS. Early stenosis appears to respond better to EBD, and long-term symptomatic stenosis, in which balloon dilation fails, may respond to endoscopic stents or salvage surgery. (Figure Presented)

Introduction: Bariatric surgery is associated with improvement of obesity related co-morbidities. Some studies suggest that these patients may be susceptible to alcohol abuse due to alteration in metabolism. Aim of this study was to assess the impact of bariatric surgery on high risk alcohol use in a meta-analysis of the published data. Methods: MEDLINE and Embase were searched from inception through September 2016 with MeSH terms "alcohol", "substance abuse", "outcomes", and "bariatric surgery". Dual extraction and quality assessment of studies was performed independently by two authors. Inclusion criteria were retrospective and prospective studies reporting long-term outcomes (>12 months) after bariatric procedures, with >100 patients in a cohort, and reporting rates of alcohol use as per the AUDIT, dSCID criteria and data from inpatient alcohol treatment. Primary outcomes included the pooled event rate and 95% confidence interval (95% CI) for significant alcohol use after bariatric surgery, new-onset significant alcohol use after bariatric surgery and odds ratio comparing rates of alcohol abuse before and after surgery. Results: 565 citations were identified;12 prospective and 16 retrospective or cross-sectional cohort studies including 15,714 subjects were analyzed. Average follow-up was 2.6 years(range 0.08-10 years). Average age of participants ranged from 26.5-50.1 years, with 76.8% female. Majority of patients had undergone Roux-en-Y gastric bypass and fewer had laparoscopic gastric banding/other restrictive procedures. New-onset significant alcohol use was identified in 8% of patients after bariatric surgery . 19% of patients had significant alcohol use prior to bariatric surgery, event rate 0.19(95% CI 0.12-0.28,18 studies,4697 patients). 23% of patients had significant alcohol abuse after bariatric surgery; event rate 0.23(95% CI 0.14-0.34, I2=97%, Tau2=1.5,19 studies,4552 patients). Patients had higher odds of significant alcohol use after surgery than before, OR 1.5(95% CI 1.07-2.11, I2=89%, Tau2=0.3,11 studies,3370 patients). Conclusion: Bariatric surgery is associated with significant increase in moderate to high-risk alcohol use/abuse with a new-onset incidence of 8%. Additionally, 19% of patients overall reported problem drinking prior to surgery. Thus, better selection of patients and further studies to identify modifiable factors are needed to decrease the substantial burden of high-risk alcohol use after bariatric surgery. (Table Presented)

Introduction: Inflammatory bowel disease (IBD) therapy is continuously evolving with novel drugs targeting various inflammatory pathways. Ustekinumab (USK), a monoclonal antibody inhibiting the IL-12/23 pathway, was approved in September 2016 to treat moderate-to-severe Crohn's disease (CD). While safety data in IBD is limited, USK has been used to treat other autoimmune diseases with favorable safety profiles. We aimed to establish rates of adverse events (AE) and demonstrate non-inferiority of AE of USK compared to placebo and other biologics. Methods: MEDLINE, PubMed and Embase databases were searched in May 2017 using terms "ustekinumab" and "clinical trials." Two authors independently performed quality assessment and dual extraction. Randomized control trials comparing USK to placebo or other biologics regardless of disease were included. The primary outcome was the odds ratio (OR) of AE of USK vs placebo, expressed as pooled OR and 95% confidence interval (CI). Secondary outcomes included OR of mild/moderate and serious AE (SAE) in USK vs placebo, USK vs biologics, and low vs high-dose USK, respectively (Table 2). A sub-analysis of outcomes in CD trials was performed. Random effects meta-analysis was performed for all outcomes. Results: 16 papers with 6756 subjects (44% female) were included (Fig 1). Infections were the most common AE (Table 1). The OR of serious AE in USK vs placebo was 0.76 (95% CI 0.56-1.03, Fig 2). The OR of mild-to-moderate AE in the USK vs placebo was 1.12 (95% CI 1.01-1.24), suggesting increased risk of mild/moderate AE with USK (Fig 3). However, this was no longer significant after sub-analysis of the three CD trials. Analysis of 5 trials comparing low vs high-dose USK revealed an OR of 0.96 (95% CI 0.46-2.04) for SAE and 1.17 (95% CI 0.98-1.39) for mild-to-moderate AE. Use of USK was not associated with increased AE compared to other biologics, with OR of 0.91 (95% CI 0.61-1.35) for SAE and 0.98 (95% CI 0.85-1.13) for mild/moderate AE. Heterogeneity was low for all calculations. Conclusion: USK has a comparable safety profile to placebo and other biologics in the treatment of various diseases, although we did find a mildly elevated risk of mild/moderate AE with USK; however, this (Figure Presented) was not seen in CD trials. The favorable safety profile of USK is of clinical importance with the advent of USK in CD and ongoing clinical trials for ulcerative colitis. More data on long-term safety data in the IBD population is needed

Introduction: Adrenal insufficiency (AI) is a clinical entity frequently unrecognized due to its nonspecific symptoms. It may lead to increased risk of systemic illness in conditions that precipitate shock, such as hypovolemia. The incidence of AI and its effects in patients with upper gastrointestinal non-variceal bleeding (UGINVB) is unknown. We aimed to determine the impact on the clinical outcomes of patients presenting with UGINVB in AI vs non-AI population. Methods: A retrospective cohort study was done using the United States 2013 National Inpatient Sample, the largest publicly available inpatient database in the United States. All patients with ICD-9 CM codes for a primary diagnosis of UGINVB were included. No exclusion criteria were used. The primary outcome was in-hospital mortality. Secondary outcomes were morbidity measured by intensive care unit (ICU) admission, shock, multi-organ dysfunction (MOD) and malnutrition; resource utilization measured by abdominal ultrasound (US), abdominal CT scan (ACT), length of hospital stay (LOS), total parenteral nutrition (TPN) use, total hospitalization charges (TOC), esophagogastroduodenoscopy (EGD) and colonoscopy. Patients were classified as AI or non-AI diagnosis based on ICD-9 CM codes. Odds ratios and means were calculated using multivariate regression analysis, after being adjusted for age, sex, race, Charlson Comorbidity Index, median income in the patient's zip code, hospital region, rural location, size and teaching status. Results: A total of 138,700 patients with UGINVB were included. Mean age was 64.3 years and 44.9% were female. 470 (0.34%) had AI diagnosis. In-hospital mortality rate was 1.74% overall, 4.2% in AI patients and 1.7% in non-AI patients. Adjusted odds ratios, means and P values are shown in Table 1. On multivariate analysis, patients with AI had increased mortality (OR 2.64, 95% CI 0.92 - 7.58; P=0.07) compared to the non-AI population. Regarding morbidity, ICU admission, shock, malnutrition and MOD were more common in AI patients. Resource utilization, total charges, LOS, and TPN use were higher in the AI group; the use of ACT/US and need for EGD/colonoscopy were the same in both groups. Conclusion: Adrenal insufficiency is associated with increased morbidity in patients admitted with UGINVB. These patients present an increase in resource hospitalization, particularly hospitalization charges, length of stay and TPN use. Special attention needs to be taken in these population due to worse overall outcomes

Introduction: The need to increase adenoma detection rate (ADR) for colorectal cancer screening has ushered in enhanced visualization devices, designed to mechanically (Endocuff, Endorings and G-Eye) or optically (G-Eye, FUSE and EWAVE) improve conventional colonoscopy (CC). Cap-assisted colonoscopy (CAC) was the first device in the market but studies have not shown benefit over CC. In recent years, newer enhanced technology devices (NTD) have become available. Our aims were: 1) to compare the ADR, adenoma miss rate (AMR) and polyp detection rate (PDR) between CC and NTD 2) to compare the ADR, AMR and PDR between mechanical and optical NTD. Methods: MEDLINE and Embase databases were searched from inception through May 2017 for manuscripts or conference abstracts reporting ADR, AMR and PDR with the available NTD. Data on CAC was collected and analyzed separately. Randomized controlled trials and high-quality casecontrol studies in adults with >10 subjects were included, with dual extraction of data. Primary outcomes included pooled ADR, AMR and PDR odds ratio (OR) with 95% confidence interval (95% CI) between CC and NTD. AMR was assessed in RCTs with tandem colonoscopies. Secondary outcomes included cecal intubation rates (CIR), complications, cecal intubation time (CIT), and total colonoscopy time (TCT). Results: Of 247 citations identified without CAC, 47 studies with 17,543 subject were eligible for >=1 analyses. Mean age was 61.3 years (range 37.1-76.0) and 50.3% were males. The overall OR for ADR/ PDR was higher with NTD than CE (Fig 1). Sub-analysis between NTD showed a higher ADR (OR1.44 vs 1.18, P=0.03) and PDR (OR1.63 vs 1.16, P=0.01) for mechanical compared to optical NTD. Comparison of each NTD to CC showed an improved ADR/PDR with Endocuff and higher ADR with G-Eye, but not with FUSE (Table 1). No significant ADR difference was found between CAC and CC. The overall AMR with NTD was lower than CC (Fig 2). Mechanical NTD had lower AMR compared to optical NTD (OR0.11 vs 0.33, P<0.01). There was no difference in CIR (OR1.03, 0.81-1.30). The CIT was 7.33 min (+/-2.5) in NTD vs 7.39 min (+/-2.88) in CC. The TCT was 18.54 (+/-5.92) in NTD and 19.01 (+/-6.31) in CC. Major complications were uncommon (9 in NTD and 10 CC). Conclusion: Newer endoscopic technologies are an effective option to increase ADR/PDR, and decrease AMR, particularly mechanical NTD. Overall, the risk of major complications is similar to CC. Further comparisons between devices are warranted