Prognostic factors in the palliation of pancreatic cancer

Aim: Few patients with pancreatic cancer are eligible for resection. In the remainder, estimation of prognosis is important to optimise various aspects of care, including palliation of biliary obstruction and trial of chemotherapy. The aim is to evaluate the Prognostic significance of clinical and laboratory variables in patients with unresectable pancreatic cancer. Methods: Information was gathered retrospectively for 325 patients with unresectable pancreatic cancer who underwent palliative interventions, including surgical bypass, endoscopic or percutaneous stenting or who received supportive care only. Results: Histological proof was obtained in 182 patients (56%). Median survival was 5.7 months. Absence of therapeutic intervention, leukocytosis (WCC greater than or equal to 11 x 10(9)/l), gamma glutamyl transferase (gammaGT) > 165 U/L, prothrombin time ratio > 1.1, and C-reactive protein (CRP) greater than or equal to5 mg/dL were associated with shorter survival on univariate analysis. Only absence of therapeutic intervention, leukocytosis, and gammaGT > 165 U/L reached significance on multivariate analysis. In the 51 patients in whom serum CRP was available, CRP was the only significant predictor of survival on multivariate analysis. Conclusions: Leukocytosis, elevated gammaGT and raised CRP predict shorter survival and may help to guide the choice of palliative intervention for patients with unresectable pancreatic cancer. (C) 2002 Elsevier Science Ltd. All rights reserved.