Abstract
BACKGROUND: Sepsis is an important cause of mortality among older adults in the United States. The association between sepsis and subsequent risk of cancer is poorly understood. METHODS: Using the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked database, we conducted a case-control study in US adults. We included 1801156 cases with a first cancer diagnosis in SEER during 1992-2013 (ages 66-115 years) and 200000 cancer-free controls from a 5% random sample of Medicare beneficiaries. Sepsis was identified using inpatient Medicare claims. Associations with sepsis were estimated using logistic regression. RESULTS: After correction for multiple comparisons, sepsis was significantly associated with increased risk for cancers of the colon (adjusted odds ratio [aOR] = 1.12), rectum (1.13), liver (1.47), lung (1.17), and cervix (1.52), as well as acute myeloid leukemia (AML, 1.19), chronic myeloid leukemia (1.54), and myelodysplastic syndrome (1.30). Inverse associations were observed for cancers of the breast (aOR = 0.86), prostate (0.75), kidney (0.90), and thyroid (0.68) and for melanoma (0.83), diffuse large B-cell lymphoma (0.89), and follicular lymphoma (0.65). Sepsis was significantly associated with the following 9 types of cancer in the period >5 years following sepsis diagnosis: thyroid, prostate, colon, rectum, lung, and liver and follicular lymphoma, melanoma, and AML. CONCLUSIONS: Sepsis is associated with increased or decreased risks for a small group of cancers. Factors that may explain these associations include etiologic effects. Other associations may reflect the presence of precursor conditions or patterns in ascertainment of cancer and screening.