Mo2003 Long-Term Prospective Observational Study of Cancer Risk Among Pantoprazole Users

Abstract

BACKGROUND: Streptococcus bovis bacteremia (SBB) has been associated with gastrointestinal (GI) malignancies, namely colorectal cancer. For this reason, screening colonoscopy for SBB patients is recommended in current practice. However, these recommendations are primarily based on data from case reports and case series. AIM: Determine the incidence of gastrointestinal malignancies in patients with SBB in a regional population-based setting. METHODS: We performed a retrospective cohort study on data from January 2000 December 2013 from a regional, integrated healthcare system in Southern California. All adult patients with positive S. bovis blood cultures were identified. New cases of GI malignancy were identified through an internal prospective cancer registry. Incidence rates (IR) of primary GI malignancies (e.g. colorectal, esophageal, gastric, small intestinal, pancreatic, hepatocellular carcinoma, cholangiocarcinoma, gallbladder, and appendiceal) among the SBB patients were compared to the ageand sex-matched reference population. A priori, we planned to evaluate the impact of S. bovis subspecies on the risk of malignancy. Patients were censored based on death, loss to follow-up, or study end. RESULTS: 356 cases (209 male, 58.7%) with a mean follow-up of 8.75 years (3,115 person-years) of SBB were identified, and among these, 54 new cases of primary GI malignancy were identified. The reference population consisted of 13,976,494 patients with a mean follow-up of 2.76 years (38,630,000 person-years), and among these, 33,208 primary GI malignancies developed. The age-sex adjusted incidence rates for the SBB and the reference populations to develop any GI malignancy were 1301.26 per 100,000 (95% CI 700.28-1902.23) and 85.94 per 100,000 (95% CI 84.99-86.89), respectively; the overall standardized incidence rate ratio (SIRR) for risk of any GI malignancy was 15.14 (95% CI 9.54-24.03; p<0.0001). More specifically, for colon cancer (n= 23 cases in the SBB population), the SIRR was 17.10 (95% CI 8.4134.76; p<0.0001). The highest SIRRs were associated with hepatocellular carcinoma (n=7, SIRR=20.49 (95% CI 8.01-52.40, p<0.0001)) and with gastric cancer (n=5, SIRR=25.75 (95% CI 5.21-127.28, p=0.0001)). We were unable to evaluate the impact of S. bovis subspecies as sub-speciation was not routinely performed. CONCLUSION: To our knowledge, this is the largest longitudinal cohort study to date to systematically evaluate the incidence of gastrointestinal malignancy in patients with SBB. We have confirmed that SBB is associated with an increased risk for colon cancer. However, the greatest increase in relative risk was for primary liver and stomach cancers.