Income, inflammation and cancer mortality: a study of U.S. National Health and Nutrition Examination Survey mortality follow-up cohorts

Abstract

BACKGROUND: To estimate the relationship between inflammatory biomarkers and cancer mortality in a nationally representative sample of the U.S. population while controlling for education, occupation, and income. METHODS: Data were obtained from the U.S. National Health and Nutrition Examination Survey from 1988 to 1994 (N = 7817) and 1999-2002 (N = 2344). We fit Cox proportional hazard models to examine the relationship between C-reactive protein (CRP) and fibrinogen with cancer mortality. RESULTS: In the full Cox multivariate model, clinically raised CRP was associated with cancer mortality in NHANES 1988-1994 (> 0.99 mg/dL: 95%CI: 1.04-2.13). However, across two inflammatory biomarkers (CRP and Fibrinogen), two NHANES time periods (1998-1994 and 1999-2002) and three income levels (12 strata in total), Hazard ratio confidence intervals did not include the null only for one association: CRP and cancer mortality among low income participants from 1988 to 1994 (HR = 1.83, 95% CI: 1.10-3.04). CONCLUSIONS: We find evidence that only in one unique stratum is earlier life CRP, and not fibrinogen, associated with prospective cancer mortality. After more complete control for socioeconomic confounding, CRP and fibrinogen do not predict cancer mortality in most subpopulations.