Near Absence of Septic Complications Following Successful Therapeutic ERCP Justifies Selective Intravenous and Intracontrast Use of Antibiotics

Abstract

Background: Infection following therapeutic ERCP is the third most frequently reported complication behind pancreatitis and bleeding. The use of prophylactic antibiotics to prevent septic complications has in general been advocated, but has recently been questioned (Harris, Endoscopy 1999). Our institution has used a policy of employing intravenous and intracontrast antibiotics only in selected high-risk situations. We report the results of this strategy. Methods: From 1/91 to 10/04, 6695 ERCPs were entered prospectively into a database. Complications were sought at 1, 7 and 30 days and recorded prospectively. Noninfected cases undergoing diagnostic and therapeutic ERCP, who were considered to be at low risk for infection, regardless of jaundice or biliary obstruction, did not routinely receive antibiotics (Group A, n = 5525). Ceftazadime 1 gm or ciprofloxacin 400mg IV and gentamycin or tobramycin 10 mg/10cc of x-ray contrast were used for the following high-risk indications: hilar obstruction (benign or malignant), sclerosing cholangitis, pancreatic endotherapy for strictures, stones, pseudocysts. (Group B, n = 1170). Delayed infections due to stent occlusions were excluded. Results: Group A (low risk patients) 7/5525 (0.1%) experienced septic complications attributed to the ERCP procedure compared with 15/1170 (1.3%) in Group B (P < 0.0001). Infections in Group A were due to fever or cholangitis after stent placement (n = 4), incomplete stricture drainage (n = 2), and reactivation of an unsuspected undrained abscess (n = 1). Severe infectious complications were rare and only seen in the high risk group (6/1170 (0.5%) vs. 0% P < 0.0001): 5 infected pancreatic necrosis, 1 infected pseudocyst. All these cases had received appropriate pre-ERCP antibiotics by our protocol. The percent diagnostic procedures in Group A and B were 15.8% and 14.0%, respectively. Therapeutic ERCPs were successful in establishing drainage in 5550/5688 (98.1%). Conclusions: Selective use of intravenous and intracontrast antibiotics was associated with an exceedingly low incidence of post-ERCP infection. The high success rate of drainage may beneficially influence this risk and may be a more important factor than the use of antibiotics. The use of prophylactic antibiotics for all ERCPs, diagnostic and/or therapeutic, is difficult to justify based on this experience.