Aims and background: We investigated the relation between diabetes and the risk of non-Hodgkin lymphoma, as epidemiological results are controversial and diabetes has been related to the risk of several neoplasms.
Patients and methods: We analyzed the combined dataset of two Italian case-control studies conducted in 1985-1997 and 1999-2002. Cases were 671 patients, aged <85 years, with incident, histologically confirmed non-Hodgkin lymphoma, and controls were 1799 patients admitted to hospitals for acute non-neoplastic conditions. Odds ratios were estimated using unconditional multiple logistic regression models including terms for age, center, sex, residence and educational level.
Results: No material association between diabetes and non-Hodgkin lymphoma risk was observed, with an odds ratio of 1.12 (95% confidence intervals, 0.70-1.77). No association was found in relation to age at first diagnosis of diabetes, years since diagnosis, or in younger and older subjects at diagnosis of non-Hodgkin lymphoma.
Conclusions: The results of the study allow to exclude a strong association between diabetes and non-Hodgkin lymphoma, although the small number of cases with diabetes leaves open the possibility of a moderate direct relation.