Evidence-based assessment of ERCP in the treatment of pancreatitis.

Abstract

This article reports the results of an evidencebased assessment of ERCP for the treatment of pancreatitis.1 Pancreatitis encompasses a number of distinct entities with differing etiologies, clinical expression, and treatment options. Each is addressed separately to the extent allowed by the available literature. Also, there are a number of different endoscopic techniques used for varying clinical situations. For the purposes of this paper, “ERCP” will refer to the spectrum of interventional endoscopic techniques that are used in the treatment of pancreatitis. METHODS The methods used in the systematic review are summarized in the Methods article included in this supplement2 and are described in detail in the fullevidence report.1 The protocol for this systematic review prospectively defined study objectives; search strategy; patient populations of interest; study selection criteria; outcomes of interest; data elements to be abstracted and methods for abstraction; and methods for study quality assessment. Briefly, the initial selection criteria for this systematic review were full-length reports of comparative studies published in English in peer-reviewed journals. A minimum of 25 patients per treatment arm were required. Assessment of study quality was adapted from that of the U.S. Preventive Health Services Task Force.3 Outcomes of interest included measures of technical success, clinical success, resource utilization, and procedure-related morbidity. Because of a paucity of literature, especially the lack of comparative trials, very few studies on the treatment of recurrent or chronic pancreatitis met the initial selection criteria. Therefore, the selection criteria were relaxed so that this question could be examined. Concurrently controlled studies comparing ERCP to a therapeutic alternative were included regardless of sample size. Single-arm observational studies (subject serves as own control) of ERCP treatment of chronic pancreatitis with a minimum of 25 patients were included if the study selected a well-defined population and used appropriate outcome measures. Baseline evaluation and 6-month follow-up data were required. Single-arm studies of an ERCP in pancreas divisum were also included subject to the above conditions, but regardless of sample size.