Comprehensive evaluation of medical conditions associated with risk of non-Hodgkin lymphoma using Medicare claims ("MedWAS")
A partir des données des registres américains des cancers, de la base Medicare et portant sur 52 691 cas et 200 000 témoins, cette étude présente une méthodologie pour évaluer l'association entre près de 6 000 pathologies et le risque de lymphome non hodgkinien
Background: Certain medical conditions affect risk of non-Hodgkin lymphoma (NHL), but the full range of associations is unknown. We implemented a novel method ("medical condition-wide association study," MedWAS) to comprehensively evaluate medical risk factors for NHL documented in administrative health claims. Methods: Using SEER-Medicare data, we conducted a case-control study comparing NHL cases (N=52,691, age 66+ years, with five subtypes: chronic lymphocytic leukemia/small lymphocytic lymphoma [CLL/SLL], diffuse large B-cell lymphoma [DLBCL], follicular lymphoma, marginal zone lymphoma [MZL], T-cell lymphoma [TCL]) to controls (N=200,000).We systematically screened for associations with 5926 medical conditions documented in Medicare claims more than one year before selection. Results: Fifty-five conditions were variously associated with NHL. Examples include well-established associations of human immunodeficiency virus, solid organ transplantation, and hepatitis C virus with increased DLBCL risk (odds ratios [ORs] 3.83, 4.27, and 1.74, respectively), and autoimmune conditions with DLBCL and MZL (e.g., ORs of 2.10 and 4.74, respectively, for Sjögren syndrome). Risks for all NHL subtypes were increased after diagnoses of non-melanoma skin cancer (ORs 1.19-1.55), actinic keratosis (1.12-1.25), or hemolytic anemia (1.64-4.07). Nine additional skin conditions increased only TCL risk (ORs 2.20-4.12). Diabetes mellitus was associated with increased DLBCL risk (OR 1.09). Associations varied significantly across NHL subtypes for 49 conditions (89%). Conclusion: Using an exploratory method, we found numerous medical conditions associated with NHL risk, and many associations varied across NHL subtypes. Impact: These results point to etiologic heterogeneity among NHL subtypes. MedWAS is a new method for assessing the etiology of cancer and other diseases.