OBJECTIVES: Pancreatic cancer is the 4^thleading cause of adult cancer mortality in the United States with a 5-year survival rate of approximately 6%. ACE inhibitors and ARBs have been shown to inhibit tumor angiogenesis in pancreatic cells in both in-vitro and murine model studies. Previous research studies exploring a relationship between these agents and patients with pancreatic cancer used several different methods leading to unclear results. We conducted a large population-based study to examine overall survival differences in patients with pancreatic cancer taking ACE inhibitors or ARBs.  METHODS: We used the Italian Emilia-Romagna Region (RER) health care database to conduct a retrospective cohort study following approximately 4 million adults (≥18) from 2003 to 2011. The RER database captures fully-linkable demographic, hospital discharge, outpatient pharmacy, and procedure codes for all residents of the region. Patients were classified by resection and metastases based on surgical procedures and secondary malignancies. We used a Cox-proportional hazard model for each time-dependent medication exposure to minimize potential immortal-time-bias. RESULTS: We identified a total of 8,281 patients diagnosed with pancreatic cancer from 2003 to 2011 in the RER. After adjusting for covariates, the Cox-proportional hazards model using time-dependent exposure suggested that, among pancreatic cancer patients, those exposed to ACE inhibitors or to ARBs experienced a significantly decreased all-cause mortality risk as compared to those who were not (HR=0.90, 95% CI: 0.85 – 0.95 and HR=0.92, 95% CI: 0.86 – 0.98, respectively). CONCLUSIONS: Our population-based study showed that the use of ACE inhibitors and ARBs was associated with reduced overall mortality in patients with pancreatic cancer. Although modest, such benefits may be of great clinical importance because of the low overall survival rate seen in patients with pancreatic cancer and the relative safety of ACE inhibitors and ARBs. These relationships should be further examined in other populations using similar modeling techniques.