Protocol-based medical management of post-ERCP pancreatitis

Abstract

Background and aims: Although numerous studies have evaluated outcomes pertaining to endoscopic retrograde cholangio-pancreatography (ERCP) complications, studies evaluating outcomes of management of post-ERCP pancreatitis are scant. This study evaluated the effectiveness of a standard treatment protocol in management of post-ERCP pancreatitis.

Methods: This is a retrospective study of consecutive patients managed for post-ERCP pancreatitis, using a standard treatment protocol over a 3-year period. By protocol, patients received only intravenous fluids, narcotics, and analgesics for the first 24-72 h after admission. Oral intake was attempted when white cell count was normal or followed a downward trend, abdominal pain was absent or minimal without need for narcotics over a 12-h period, and serum lipase was less than three times normal range. For patients hospitalized beyond 72 h, an abdomen CT was obtained at days 4 and 10 to guide management. Intravenous antibiotics were administered only for patients with pancreatic necrosis. Jejunal feeding and a meperidine pump for pain control were initiated in symptomatic patients at day 4. Data on ERCP complications were collected prospectively and graded per consensus criteria. Effectiveness of the treatment protocol was evaluated by comparing clinical outcomes of patients managed by protocol versus those managed outside protocol.

Results: 45 of 1976 patients (2.3%) who underwent ERCP developed post-ERCP pancreatitis. Of the 45 (female 31; mean age 43 years) patients, 32 were managed by protocol and 13 outside protocol. Protocol based management was associated with less severe disease as compared with those managed outside protocol (crude odds ratio (OR) = 11.2; 95% confidence interval (CI) = 1.9-68.7; P = 0.002). One patient managed outside protocol died of severe pancreatitis. When compared with those managed outside protocol, the median duration of hospital stay (7 vs 3 days; P = 0.01), the use of CT (100% vs 15.6%; P < 0.001), and the use of antibiotics (50% vs 3.1%; P = 0.01) were significantly lower in those managed by protocol. By multiple logistic regression, protocol-based management was associated with less severe disease (adjusted OR = 18.7; 95% CI = 2.6-132.1; P = 0.003) when adjusted for age, comorbidity, endotherapy and pancreatic stenting.

Conclusions: A protocol-based management strategy was associated with less severe pancreatitis, shorter length of hospital stay, need for fewer imaging studies, and use of antibiotics. Prospective validation of these findings is justified.