Risk and protective factors for small intestine neuroendocrine tumors: A prospective case-control study

Background: The incidence of small intestine neuroendocrine tumors (SI-NETs) is increasing, but few studies have investigated risk factors for their occurrence, suggesting that family history (FH) of any cancer, smoking and previous cholecystectomy are associated with an increased risk. Such studies investigated small series or examined cancer registries without direct interviews. Aim: We therefore aimed at clarifying risk and protective factors for the occurrence of sporadic SI-NETs. Subjects and Methods: We performed a multicenter case-control study. Patients with a histologic diagnosis of SI-NETs were prospectively evaluated, excluding familial syndromes. Controls with non-neoplastic/non-chronic disorders seen at gastrointestinal outpatients clinics were matched for sex and age (4:1). All subjects were directly interviewed by means of a specific questionnaire on potential risk and protective factors. Cases and controls were compared by Fisher's test or Student's t test for categorical or continuous variables. Explanatory variables were analyzed by simple logistic regression analysis. A multiple logistic regression analysis was performed with an Enter model; p 21 alcohol units per week (7.4 vs. 3.8%) were all significantly more frequent in SI-NET patients than in controls. Multivariate analysis showed that FH of CRC (OR 2.23, 95% CI 1.29-3.84, p = 0.003), FH of breast cancer (OR 2.05, 95% CI 1.13-3.69, p = 0.01) and smoking (OR 1.47, 95% CI 1.07-2.03, p = 0.01) and in particular heavy smoking (OR 1.94, 95% CI 1.29-3.84, p = 0.0008) were associated with an increased risk for carcinoid occurrence, while use of aspirin can be considered a protective factor (OR 0.20, 95% CI 0.06-0.65, p = 0.008). Conclusion: FH of colorectal and breast cancer as well as smoking seem to be risk factors for the development of SI-NETs, while use of aspirin might be a protective factor. These factors partially overlap with those associated with CRC, but are different from those previously associated with pancreatic neuroendocrine tumors. These findings may suggest that the mechanisms of carcinogenesis for endocrine cells in different sites can be specific and similar to those of their exocrine counterparts.