Associations of Coffee Drinking with Systemic Immune and Inflammatory Markers

Abstract

Background: Coffee drinking has been inversely associated with mortality as well as cancers of the endometrium, colon, skin, prostate, and liver. Improved insulin sensitivity and reduced inflammation are among the hypothesized mechanisms by which coffee drinking may affect cancer risk; however, associations between coffee drinking and systemic levels of immune and inflammatory markers have not been well characterized.

Methods: We used Luminex bead-based assays to measure serum levels of 77 immune and inflammatory markers in 1,728 older non-Hispanic Whites. Usual coffee intake was self-reported using a food frequency questionnaire. We used weighted multivariable logistic regression models to examine associations between coffee and dichotomized marker levels. We conducted statistical trend tests by modeling the median value of each coffee category and applied a 20% false discovery rate criterion to P values.

Results: Ten of the 77 markers were nominally associated (P trend < 0.05) with coffee drinking. Five markers withstood correction for multiple comparisons and included aspects of the host response namely chemotaxis of monocytes/macrophages (IFNγ, CX3CL1/fractalkine, CCL4/MIP-1β), proinflammatory cytokines (sTNFRII), and regulators of cell growth (FGF-2). Heavy coffee drinkers had lower circulating levels of IFNγ [odds ratios (OR), 0.35; 95% confidence intervals (CI), 0.16-0.75], CX3CL1/fractalkine (OR, 0.25; 95% CI, 0.10-0.64), CCL4/MIP-1β (OR, 0.48; 95% CI, 0.24-0.99), FGF-2 (OR, 0.62; 95% CI, 0.28-1.38), and sTNFRII (OR, 0.34; 95% CI, 0.15-0.79) than non-coffee drinkers.

Conclusions: Lower circulating levels of inflammatory markers among coffee drinkers may partially mediate previously observed associations of coffee with cancer and other chronic diseases.

Impact: Validation studies, ideally controlled feeding trials, are needed to confirm these associations.

Figure 1

Summary of P-values for trend from tests of multivariable-adjusted association with coffee drinking for all analytes included in the analysis (N=77 markers). P-values for trend for analytes detectable in less than 50% of samples computed using Wald statistic of regression model parameters for dichotomized analyte level (undetectable vs. detectable). P-values for trend for all other analytes were computed using Wald statistic of regression model parameters for dichotomized analyte level (