Antibiotics to Decrease Post ERCP Cholangitis

Description

OBJECTIVES AND PURPOSE

The aim is to determine whether a brief course of antibiotics following therapeutic ERCP can reduce post-ERCP cholangitis in patients for whom antibiotics are not already indicated.

STUDY DESIGN

The study will be a prospective, randomized trial consisting of 452 patients who are scheduled to undergo therapeutic ERCP at the LAC+USC Medical Center for standard indications. Patients undergoing ERCP for therapy of bile duct problems including choledocholithiasis, malignant obstruction, jaundice, and bile leak will be eligible. Those with mandatory antibiotic requirement will be excluded.

Patients will be randomly assigned in a 1:1 ratio using a computer generated randomization schedule and allocation will be concealed.

Patients will be randomized during the ERCP procedure, once decompression of the duct is achieved (either obstruction cleared or stent placed), to ensure that patients who are at increased risk (because of inability to decompress the duct with a stent or clearance) will be excluded from the study. Those randomized to the antibiotics arm will receive intravenous antibiotics (ceftriaxone 1gm) immediately following the ERCP procedure and will take oral antibiotics (levofloxacin 500mg) once daily for 3 subsequent days. This regimen is chosen to most closely reflect the actual clinical practice in our center. Those in the no antibiotics arm will not receive antibiotics. However, if they develop findings of cholangitis or other infection they will be treated with antibiotics. Comprehensive data including procedure indication will be recorded.

The primary outcome will be the development of post-ERCP cholangitis. Post-ERCP cholangitis will be defined by the 2013 Tokyo Guidelines

The secondary outcomes will include length of hospital stay and adverse events attributable to antibiotic use such as allergic reactions or diarrhea. All patients will be assessed 1, 3, and 7 days after the procedure by in-person visits for inpatients and telephone calls for outpatients. Patients who have fever, pain, jaundice, or signs of an adverse antibiotic outcome will be evaluated as would be done per standard clinical care.

Randomization scheme. Patients will be randomly assigned using a computer generated randomization schedule with concealed 1:1 allocation to receive a prophylactic course of antibiotics or no antibiotics.

ASSESSMENT OF EFFICACY AND SAFETY

Side effects/Toxicities to be monitored.

Major adverse outcomes which could be associated with antibiotic therapy will be recorded including unexpected allergies (such as rash or angioedema), adverse drug reactions (such as diarrhea, nausea). Patients who have diarrhea will be tested for Clostridium difficile colitis. The overall rate of these antibiotic associated events is estimated at 0.5-2%.2,13-15

CRITERIA FOR EVALUATION AND ENDPOINT DEFINITIONS

The outcome status (in terms of rates of cholangitis and pancreatitis as well as antibiotic toxicity and adverse events) of all eligible patients will be reported. All eligible patients who begin treatment will be included in the analysis of survival and time-to-failure.

Endpoint Definitions

Primary outcome: The primary outcome will be the development of DEFINITE post-ERCP cholangitis in <7 days.

Post-ERCP cholangitis will be defined by the 2013 Updated Tokyo Guidelines (TG13) which are the international standard.12 As the patients in this study are by definition undergoing therapeutic ERCP, criterion C will have been fulfilled.

1. Systemic Inflammation
2. Cholestasis
3. Imaging-biliary dilation and/or evidence of etiology

Secondary Outcome #1: We will compare the proportions of mild, moderate, and severe cholangitis in the two groups based on the Tokyo Guidelines (see below)12

Secondary Outcome #2: "Suspected cholangitis", as defined by the Tokyo guidelines (see table 1 above) within 7 days of ERCP.

Secondary Outcome #3: The requirement for a repeat procedure as defined by ERCP, percutaneous drainage, surgery for cholangitis within 7 days of ERCP.

Secondary Outcome #4: The length of hospital stay after ERCP.

Secondary Outcome #5: The rate of post-ERCP pancreatitis, defined as: new onset upper abdominal pain, an amylase or lipase ≥3 times the upper limit of normal, and hospitalization for ≥2 nights following ERCP.

Secondary Outcome #6: The rate of allergic reactions, serious adverse events, and C. difficile colitis attributable to antibiotic use.

The time period of all outcomes will be within 7 days of ERCP. Inpatient or telephone assessments will be done at 1, 3, and 7 days after ERCP for all patients. Any patients who are suspected to have pancreatitis or cholangitis will be advised to return to the endoscopy area immediately for a physical examination as well as vitals signs and laboratory assessment (see telephone form). In cases where there is potentially evidence of cholangitis or pancreatitis, the PI and study team will define whether the clinical constellation is most suggestive of cholangitis, pancreatitis or both processes.

STATISTICAL CONSIDERATIONS

The goal of this randomized study to determine whether a prophylactic antibiotic course can prevent cholangitis following therapeutic ERCP. Review of 200 patients at LAC USC in whom antibiotic course was given revealed a rate of cholangitis of 1%. A meta-analysis of randomized trials of antibiotics to prevent ERCP cholangitis suggested a rate of 5.8% in the control (no antibiotics group) with the largest individual trial reporting a rate of 6%.8 Pubished rates for cholangitis without antibiotics of approximately 6%. Thus at an α=0.05, β=0.8 and assuming attrition of 5% we estimate that a sample size 452 (226 per group) will be adequate to detect a statistically significant difference in the rate of post-ERCP cholangitis (assuming a difference of 1% versus 6%) within 7 days.

Intent-to-treat analyses of dichotomous outcomes will be compared using a Fischer's exact test and continuous outcomes will be compared using a Wilcoxan rank sum or T tests, depending on distribution. Multivariate logistic regression will be used for sensitivity analysis for the primary outcome if there is any imbalance in baseline characteristics, as well as if there are differential patterns of missing data or adherence between groups.

Details