Predictors of improvement following endoscopic management of chronic pancreatitis

Background: Endoscopic therapy is an accepted treatment for pain in chronic pancreatitis, in addition to traditional management. However, guidelines for selecting treatment options have yet to be established. This study addresses 2 questions: (a) which patients are likely to benefit from endoscopic therapy and (b) does improvement after endoscopic therapy predict benefit from surgery? Methods: Between 8/91 and 9/95, 47 patients with chronic pancreatitis and pain were managed by endoscopic therapy (stent ± sphincterotomy). Data were collected by chart and data base review. Patients or their referring physicians were contacted by phone and questioned about symptoms. 1 patient lost to follow-up was excluded from this analysis. Of 46 patients remaining, 24 were female and 22 male, mean age 49 years (range 23 to 80 years). Data were analyzed using regression analysis and chi square. Results: For all patients and all treatments 63% of patients improved, 26% were unchanged and 10% had worsening pain. Similar response rates were found in the 13 patients who had endoscopic therapy followed by surgery (drainage procedure ± resection). Half the patients who responded to endoscopic therapy (eg trial of stenting) did not improve following surgery. The subgroup of patients with pancreas divisum had a good response to endoscopy with 6/7 reporting pain relief. Multiple variables including disease duration, frequency of hospitalization, pain severity, use of narcotics, alcohol abuse, ductal morphology and treatment modality did not predict benefit from endoscopic therapy. Only 2 variables showed a trend towards positive outcome: male gender (p=0.05) and intermittent pain (p=0.06). Conclusion: In the endoscopic therapy of chronic pancreatitis, gender and pain pattern are potential predictors of outcome. Intermittent pain responded better than chronic pain. Although the numbers are small there was no trend to suggest that endoscopic therapy will reliably predict response to surgery. Further studies prospectively evaluating larger patient populations are indicated.