OBES SURG DOI 10.1007/s11695-014-1337-4
ORIGINAL CONTRIBUTIONS
Acute Pancreatitis in Patients After Bariatric Surgery: Incidence, Outcomes, and Risk Factors Arthi Kumaravel & Andrea Zelisko & Philip Schauer & Rocio Lopez & Matthew Kroh & Tyler Stevens
# Springer Science+Business Media New York 2014
Abstract Background The incidence of acute pancreatitis (AP) in bariatric surgery patients is not known. Ouraim was to determine the incidence, outcomes, and risk factors of AP in postbariatric surgery patients. Methods An historical cohort study was conducted of all patients who underwent Roux-en-Y gastric bypass, sleeve gastrectomy, adjustable gastric banding, and revisional procedures at our institution from January 2004 to September 2011. Patients who developed AP were identified by review of the
A. Kumaravel Department of Gastroenterology and Hepatology, A31, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA e-mail: [email protected] A. Zelisko Bariatric and Metabolic Institute, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA e-mail: [email protected] P. Schauer Bariatric and Metabolic Institute, M61, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA e-mail: [email protected] R. Lopez Quantitative Health Sciences, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA e-mail: [email protected] M. Kroh (*) Bariatric and Metabolic Institute, A100, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA e-mail: [email protected] T. Stevens Department of Gastroenterology and Hepatology, Q3, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA e-mail: [email protected]
electronic medical record. A nested case-control study using Cox regression analysis was done to identify risk factors. Results A total of 2695 patients underwent bariatric surgery. Twenty-eight patients (1.04 %) developed AP during a median follow-up of 3.5 years (interquartile range [IQR] 1.9–5.8). One patient had severe AP, and there was one AP-related death. In the case-control study, the only baseline variable that predicted postoperative AP was a prior history of AP. Three other variables identified after surgery were associated with AP: (1) rapid weight loss as measured by percent of excess weight loss (EWL) at the first post-operative visit, (2) abnormal findings on post-operative ultrasound (stones, sludge or ductal dilation), and (3) postoperative complications of bowel leak or anastomotic stricture. Conclusions The incidence of AP in this cohort is 1.04 %, which is higher than that reported for the general population (~17/100,000, 0.017 %). Most cases were clinically mild and managed conservatively with good outcomes. Rapid postoperative weight loss and the presence of gallstones or sludge on post-operative ultrasound were significant risk factors for AP. Keywords Acute pancreatitis . Bariatric surgery . Incidence . Outcomes . Risk factors Abbreviations AP Acute pancreatitis BISAP Bedside index of severity of AP BMI Body mass index CBD Common bile duct CT Computed tomography ERCP Endoscopic retrograde cholangiopancreatography EWL Excess weight loss LFT Liver function test NHANES National Health and Nutrition Examination Survey SOD Sphincter of Oddi dysfunction US Ultrasound
OBES SURG
Introduction The prevalence of obesity is increasing in the United States, as are related conditions like diabetes, hypertension, hypercholesterolemia, and sleep apnea. The most recent National Health and Nutrition Examination Survey (NHANES) reported the prevalence of obesity to be 35.7 % in 2010 [1]. Coinciding with the obesity epidemic, bariatric surgery has become increasingly common. An estimated 54.2 procedures per 100,000 adults were performed in 2008 [2]. Bariatric surgery patients may be at plausible risk for developing pancreatic and biliary disorders due to a number of potential mechanisms, including stones/sludge from rapid weight loss, ampulla stenosis, and sphincter of Oddi dysfunction. Although there are published case-series describing the clinical course of acute pancreatitis (AP) and related therapeutic interventions in bariatric patients [3–10]; there is no information on incidence or risk factors in this group. Our aim was to estimate the incidence of AP after bariatric surgery using our prospective database of patients who have undergone bariatric surgery at the Cleveland Clinic. We also sought to identify risk factors for AP that potentially could guide targeted preventive strategies.
Methods An historical cohort study was conducted utilizing our prospective bariatric surgery registry. All patients undergoing bariatric surgery at our institution are prospectively included in this IRB approved registry. Those who underwent bariatric surgery or revisional bariatric surgery from January 2004 to September 2011 were included in the current study. After additional approval from the IRB, the electronic medical records of these patients were reviewed to identify those who developed AP following surgery. AP was defined based on the presence of two of the following criteria: (1) abdominal pain characteristic of AP, (2) serum amylase and/or lipase ≥3 times the upper limit of normal, and (3) computed tomography (CT) or ultrasound (US) findings characteristic of AP [11]. Demographic data and information regarding presentation, symptoms, length of stay, complications, diagnostic workup, and therapeutic interventions were collected. Patients were classified as having severe AP based on the revised Atlanta criteria [12]. Next, using the cohort of post bariatric surgery patients, a nested case-control study was conducted to identify risk factors for development of AP. For each case of AP, two controls matched forage, gender, and pre-operative body mass index (BMI) were randomly selected from all patients who underwent bariatric surgery in this time period and did not have AP. Information regarding type of bariatric procedure, date of procedure, post-operative complications, rapidity of
weight loss, family history of AP, and comorbid conditions was collected for all patients and controls from the medical records. The following variables before and after surgery were also collected: medication use, liver function test (LFT) results, abdominal US findings, smoking history, alcohol history, and interventions such as cholecystectomy and endoscopic retrograde cholangiopancreatography (ERCP). All patients undergoing bariatric surgery at our institution are seen by the surgeon, a psychiatrist and nutritionist prior to surgery. In addition, the patients are seen by a medical specialist if they have significant co-morbidities or BMI >50. Patients are required to stop all alcohol intake and smoking prior to surgery. The following laboratory data are drawn preoperatively and then every 6 months post-operatively for 24 months: complete blood count, complete metabolic profile, serum folate, serum iron and total iron binding capacity, vitamin D, vitamin B12, vitamin B1, and glycosylated hemoglobin level (HBA1C). Patients routinely undergo a right upper quadrant US prior to surgery unless they have had a previous cholecystectomy and an EGD if they have a history of gastro-esophageal reflux disease. Routine cholecystectomy was not performed at the time of surgery. Only patients with symptomatic gallstones underwent cholecystectomy at the time of bariatric surgery. Patients with asymptomatic gallstones and normal imaging were treated with ursodiol post operatively. Patients are seen in clinic at 1 week, 1 month, 3 months, 6 months, 9 months, 12 months and 18 months after surgery. They are then followed on an annual basis. Patients are put on histamine receptor 2 blocker or proton pump inhibitor for 3– 6 months post-operatively if they are not already on these medications and ursodiol for 3–6 months if the patient has not undergone a cholecystectomy prior to or at the time of surgery. Patients are required to take supplements of multivitamins and minerals, vitamin B12, iron, vitamin C, calcium and vitamin D. Analysis Data are presented as mean±standard deviation, median [25th, 75th percentiles] or N (%). Time-to-event analysis was used to assess incidence of post bariatric surgery AP as well as factors associated with this. Follow-up time was defined as time from bariatric surgery to diagnosis of AP or last-follow-up visit if no AP. A Kaplan–Meier plot was constructed to study incidence of AP after bariatric surgery. In addition, univariable and multivariable Cox regression analysis was performed; hazards marginal model was used to account for the correlation between matched subjects. An automated stepwise variable selection method performed on 1000 bootstrap samples was used to choose the final model. EWL (%) at first post-surgery follow-up and abnormal right upper quadrant US post-surgery were included in the models
OBES SURG
A total of 2,695 patients underwent bariatric surgery at the Cleveland Clinic during the study period. These patients were followed for a median of 3.5 years (interquartile range [IQR] 1.9–5.8 years). Twenty-eight (1.04 %) developed AP. The baseline characteristics of the patients with AP are shown in Table 1. The most common surgical procedure was gastric bypass (78.6 %), followed by sleeve gastrectomy (10.7 %) and adjustable gastric banding (7.1 %). The mean preoperative BMI was 47.4 kg/m2 and the mean BMI at the time of admission for AP was 38.3 kg/m2. The median time from surgery to diagnosis of AP was 1.5 years (0.11–2.8). Only one patient had a prior history of AP. The most common identified etiologies were gallstones in eight patients and post-ERCP in three patients. The indications for ERCP were cholangiocarcinoma, common bile duct (CBD) stone and ductal dilation. Alcohol related pancreatitis was not
seen in this cohort as patients had to stop all alcohol consumption to be eligible for bariatric surgery. However, alcohol use may be an under recognized cause of AP in these patients as one patient who denied alcohol use at the time of admission for AP was treated for alcohol dependence a year later. Eleven patients had idiopathic AP. Patients with idiopathic AP underwent a right upper quadrant US or other imaging to look for biliary pathology, had review of their medications and alcohol intake and a triglyceride level checked. Fifteen patients presented with elevated LFTs just prior to or at the time of admission for AP. A liver biopsy had been done at the time of bariatric surgery in 14 of these patients. The pathologic findings showed normal parenchyma in4 patients, steatosis in six patients, steatohepatitis with fibrosis in three patients and cirrhosis in one patient. Biliary pathology (two patients with CBD stones, two patients with CBD dilation, four patients with gallstones, one patient with gall bladder sludge, one patient with cholangiocarcinoma) was found on imaging and presumed to be the cause of abnormal LFTs in ten patients. Of the remaining five patients, only two had steatohepatitis with fibrosis on liver biopsy, whilst the others had normal pathology or only mild steatosis. The LFTs were normal in all these patients at the time of the liver biopsy, with the elevation detected at the time of the AP episode, indicating a possible biliary etiology such as sphincter of Oddi dysfunction (SOD) or ampullary stenosis. One patient with unexplained LFT elevations underwent an ERCP for suspected SOD and did not have any recurrent attacks. The outcomes of the patients with AP are shown in Table 2. Twenty-three patients had mild AP, with a bedside index of severity of AP (BISAP) score on admission of 0 or 1. Only
Table 1 Baseline characteristics of patients with AP
Table 2 Outcomes of patients with AP
and demographics, medication use and other clinical characteristics were considered for inclusion; the additional variable with the highest inclusion rate was included in the final model. A p value
ORIGINAL CONTRIBUTIONS
Acute Pancreatitis in Patients After Bariatric Surgery: Incidence, Outcomes, and Risk Factors Arthi Kumaravel & Andrea Zelisko & Philip Schauer & Rocio Lopez & Matthew Kroh & Tyler Stevens
# Springer Science+Business Media New York 2014
Abstract Background The incidence of acute pancreatitis (AP) in bariatric surgery patients is not known. Ouraim was to determine the incidence, outcomes, and risk factors of AP in postbariatric surgery patients. Methods An historical cohort study was conducted of all patients who underwent Roux-en-Y gastric bypass, sleeve gastrectomy, adjustable gastric banding, and revisional procedures at our institution from January 2004 to September 2011. Patients who developed AP were identified by review of the
A. Kumaravel Department of Gastroenterology and Hepatology, A31, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA e-mail: [email protected] A. Zelisko Bariatric and Metabolic Institute, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA e-mail: [email protected] P. Schauer Bariatric and Metabolic Institute, M61, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA e-mail: [email protected] R. Lopez Quantitative Health Sciences, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA e-mail: [email protected] M. Kroh (*) Bariatric and Metabolic Institute, A100, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA e-mail: [email protected] T. Stevens Department of Gastroenterology and Hepatology, Q3, Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA e-mail: [email protected]
electronic medical record. A nested case-control study using Cox regression analysis was done to identify risk factors. Results A total of 2695 patients underwent bariatric surgery. Twenty-eight patients (1.04 %) developed AP during a median follow-up of 3.5 years (interquartile range [IQR] 1.9–5.8). One patient had severe AP, and there was one AP-related death. In the case-control study, the only baseline variable that predicted postoperative AP was a prior history of AP. Three other variables identified after surgery were associated with AP: (1) rapid weight loss as measured by percent of excess weight loss (EWL) at the first post-operative visit, (2) abnormal findings on post-operative ultrasound (stones, sludge or ductal dilation), and (3) postoperative complications of bowel leak or anastomotic stricture. Conclusions The incidence of AP in this cohort is 1.04 %, which is higher than that reported for the general population (~17/100,000, 0.017 %). Most cases were clinically mild and managed conservatively with good outcomes. Rapid postoperative weight loss and the presence of gallstones or sludge on post-operative ultrasound were significant risk factors for AP. Keywords Acute pancreatitis . Bariatric surgery . Incidence . Outcomes . Risk factors Abbreviations AP Acute pancreatitis BISAP Bedside index of severity of AP BMI Body mass index CBD Common bile duct CT Computed tomography ERCP Endoscopic retrograde cholangiopancreatography EWL Excess weight loss LFT Liver function test NHANES National Health and Nutrition Examination Survey SOD Sphincter of Oddi dysfunction US Ultrasound
OBES SURG
Introduction The prevalence of obesity is increasing in the United States, as are related conditions like diabetes, hypertension, hypercholesterolemia, and sleep apnea. The most recent National Health and Nutrition Examination Survey (NHANES) reported the prevalence of obesity to be 35.7 % in 2010 [1]. Coinciding with the obesity epidemic, bariatric surgery has become increasingly common. An estimated 54.2 procedures per 100,000 adults were performed in 2008 [2]. Bariatric surgery patients may be at plausible risk for developing pancreatic and biliary disorders due to a number of potential mechanisms, including stones/sludge from rapid weight loss, ampulla stenosis, and sphincter of Oddi dysfunction. Although there are published case-series describing the clinical course of acute pancreatitis (AP) and related therapeutic interventions in bariatric patients [3–10]; there is no information on incidence or risk factors in this group. Our aim was to estimate the incidence of AP after bariatric surgery using our prospective database of patients who have undergone bariatric surgery at the Cleveland Clinic. We also sought to identify risk factors for AP that potentially could guide targeted preventive strategies.
Methods An historical cohort study was conducted utilizing our prospective bariatric surgery registry. All patients undergoing bariatric surgery at our institution are prospectively included in this IRB approved registry. Those who underwent bariatric surgery or revisional bariatric surgery from January 2004 to September 2011 were included in the current study. After additional approval from the IRB, the electronic medical records of these patients were reviewed to identify those who developed AP following surgery. AP was defined based on the presence of two of the following criteria: (1) abdominal pain characteristic of AP, (2) serum amylase and/or lipase ≥3 times the upper limit of normal, and (3) computed tomography (CT) or ultrasound (US) findings characteristic of AP [11]. Demographic data and information regarding presentation, symptoms, length of stay, complications, diagnostic workup, and therapeutic interventions were collected. Patients were classified as having severe AP based on the revised Atlanta criteria [12]. Next, using the cohort of post bariatric surgery patients, a nested case-control study was conducted to identify risk factors for development of AP. For each case of AP, two controls matched forage, gender, and pre-operative body mass index (BMI) were randomly selected from all patients who underwent bariatric surgery in this time period and did not have AP. Information regarding type of bariatric procedure, date of procedure, post-operative complications, rapidity of
weight loss, family history of AP, and comorbid conditions was collected for all patients and controls from the medical records. The following variables before and after surgery were also collected: medication use, liver function test (LFT) results, abdominal US findings, smoking history, alcohol history, and interventions such as cholecystectomy and endoscopic retrograde cholangiopancreatography (ERCP). All patients undergoing bariatric surgery at our institution are seen by the surgeon, a psychiatrist and nutritionist prior to surgery. In addition, the patients are seen by a medical specialist if they have significant co-morbidities or BMI >50. Patients are required to stop all alcohol intake and smoking prior to surgery. The following laboratory data are drawn preoperatively and then every 6 months post-operatively for 24 months: complete blood count, complete metabolic profile, serum folate, serum iron and total iron binding capacity, vitamin D, vitamin B12, vitamin B1, and glycosylated hemoglobin level (HBA1C). Patients routinely undergo a right upper quadrant US prior to surgery unless they have had a previous cholecystectomy and an EGD if they have a history of gastro-esophageal reflux disease. Routine cholecystectomy was not performed at the time of surgery. Only patients with symptomatic gallstones underwent cholecystectomy at the time of bariatric surgery. Patients with asymptomatic gallstones and normal imaging were treated with ursodiol post operatively. Patients are seen in clinic at 1 week, 1 month, 3 months, 6 months, 9 months, 12 months and 18 months after surgery. They are then followed on an annual basis. Patients are put on histamine receptor 2 blocker or proton pump inhibitor for 3– 6 months post-operatively if they are not already on these medications and ursodiol for 3–6 months if the patient has not undergone a cholecystectomy prior to or at the time of surgery. Patients are required to take supplements of multivitamins and minerals, vitamin B12, iron, vitamin C, calcium and vitamin D. Analysis Data are presented as mean±standard deviation, median [25th, 75th percentiles] or N (%). Time-to-event analysis was used to assess incidence of post bariatric surgery AP as well as factors associated with this. Follow-up time was defined as time from bariatric surgery to diagnosis of AP or last-follow-up visit if no AP. A Kaplan–Meier plot was constructed to study incidence of AP after bariatric surgery. In addition, univariable and multivariable Cox regression analysis was performed; hazards marginal model was used to account for the correlation between matched subjects. An automated stepwise variable selection method performed on 1000 bootstrap samples was used to choose the final model. EWL (%) at first post-surgery follow-up and abnormal right upper quadrant US post-surgery were included in the models
OBES SURG
A total of 2,695 patients underwent bariatric surgery at the Cleveland Clinic during the study period. These patients were followed for a median of 3.5 years (interquartile range [IQR] 1.9–5.8 years). Twenty-eight (1.04 %) developed AP. The baseline characteristics of the patients with AP are shown in Table 1. The most common surgical procedure was gastric bypass (78.6 %), followed by sleeve gastrectomy (10.7 %) and adjustable gastric banding (7.1 %). The mean preoperative BMI was 47.4 kg/m2 and the mean BMI at the time of admission for AP was 38.3 kg/m2. The median time from surgery to diagnosis of AP was 1.5 years (0.11–2.8). Only one patient had a prior history of AP. The most common identified etiologies were gallstones in eight patients and post-ERCP in three patients. The indications for ERCP were cholangiocarcinoma, common bile duct (CBD) stone and ductal dilation. Alcohol related pancreatitis was not
seen in this cohort as patients had to stop all alcohol consumption to be eligible for bariatric surgery. However, alcohol use may be an under recognized cause of AP in these patients as one patient who denied alcohol use at the time of admission for AP was treated for alcohol dependence a year later. Eleven patients had idiopathic AP. Patients with idiopathic AP underwent a right upper quadrant US or other imaging to look for biliary pathology, had review of their medications and alcohol intake and a triglyceride level checked. Fifteen patients presented with elevated LFTs just prior to or at the time of admission for AP. A liver biopsy had been done at the time of bariatric surgery in 14 of these patients. The pathologic findings showed normal parenchyma in4 patients, steatosis in six patients, steatohepatitis with fibrosis in three patients and cirrhosis in one patient. Biliary pathology (two patients with CBD stones, two patients with CBD dilation, four patients with gallstones, one patient with gall bladder sludge, one patient with cholangiocarcinoma) was found on imaging and presumed to be the cause of abnormal LFTs in ten patients. Of the remaining five patients, only two had steatohepatitis with fibrosis on liver biopsy, whilst the others had normal pathology or only mild steatosis. The LFTs were normal in all these patients at the time of the liver biopsy, with the elevation detected at the time of the AP episode, indicating a possible biliary etiology such as sphincter of Oddi dysfunction (SOD) or ampullary stenosis. One patient with unexplained LFT elevations underwent an ERCP for suspected SOD and did not have any recurrent attacks. The outcomes of the patients with AP are shown in Table 2. Twenty-three patients had mild AP, with a bedside index of severity of AP (BISAP) score on admission of 0 or 1. Only
Table 1 Baseline characteristics of patients with AP
Table 2 Outcomes of patients with AP
and demographics, medication use and other clinical characteristics were considered for inclusion; the additional variable with the highest inclusion rate was included in the final model. A p value
