Prognostic factors in advanced pancreatic carcinoma. Multivariate analysis and predictive score of survival.

Study aim: To identify prognostic factors in advanced pancreatic cancer and to define a predictive score. Patients and method: One hundred and sixty six patients were included in this multicentre study. Seventeen covariables were prospectively collected for each patient. Covariables associated with survival (p < 0.10) were analysed by a stepwise Cox model. Results: Four prognostic factors were selected on multivariate analysis: pain (RR 1.5; CI: 1.1-2.0), ascites (RR 1.7; CI: 1.0-2.9), weight loss > 10 kg (RR 1.4; CI: 1.0-2.0), and metastases (RR 2.3: CI: 1.6-3.2). A score was defined by attributing a value of one for pain, ascites and weight loss, and two for metastases. Patients with a score > 2 had a median survival of 2 months (SE: 0.5), and patients with a score less than or equal to 2 had a median survival of 6 months (SE: 0.6) (Logrank p < 0.0001). Conclusion: The proposed score may be helpful in therapeutic decisions concerning surgery, palliative chemotherapy and/or radiotherapy. However, this score must be validated on an independent series of patients. (C) 2000 Editions scientifiques et medicales Elsevier SAS.