Background: There are a lack of data on the associations of circulating levels of TNF alpha and IL-18 with myocardial infarction (MI), and on the extent of confounding by classical and inflammatory risk markers. Methods: We measured TNF alpha and IL-18 in plasma from 446 MI cases and 477 age- and sex-matched controls from North Glasgow. Results: TNF alpha and IL-18 were elevated in cases compared to controls (TNF alpha medians 0.99 pg/ml [interquartile range 0.65-1.64 pg/ml] versus 0.77 pg/ml [0.52-1.22 pg/ml], p < 0.0001; IL-18 medians 287 pg/ml [212-404 pg/ml] versus 271 pg/ml [200-373 pg/ml], p = 0.01). IL-18 was moderately associated with HDL cholesterol r = -0.22, triglycerides r = 0.16, and BMI r = 0.14 (p for all <= 0.003) in the control population, but not among cases. TNF alpha had few associations with classical risk factors among cases or controls. TNF alpha had a significant association with MI: odds ratio (OR) 1.66 (95% confidence interval; 1.10-2.50), comparing extreme thirds after adjusting for classical risk factors, which was reduced on further adjustment for other inflammatory markers (OR 1.47; 0.91-2.37). IL-18 showed no association by thirds after adjustment for classical risk factors (OR 1.07; 0.70-1.62). Conclusions: Circulating levels of IL-18 and TNF alpha were elevated in those with previous MI, but only TNF alpha retained an association after adjustment for classical risk factors. Independently elevated TNF alpha among those with previous MI may reflect cardiac expression of TNF alpha in ongoing myocardial remodeling. (C) 2009 Elsevier Ltd. All rights reserved.