. Tumor necrosis factor (TNF)-! has been implicated in pathophysiological processes in coronary artery disease (CAD). TNF receptor 2 is of particular interest in mediating such effects. The gene for this receptor (TNFRSF1B) has, moreover, been implicated in hypertension, elevated cholesterol and insulin resistance. TNFRSF1B is thus a worthy candidate in studies of the genetic basis of CAD. We therefore conducted a case-control study of a microsatellite marker with five alleles (CA13-CA17) in intron 4 of TNFRSF1B in 1006 well-characterized white patients with angiographically confirmed CAD and a control group of 183 healthy subjects. We found a strong association of the TNFRSF1B marker with CAD (L²=40, P=0.00000069). The frequency of the CA16 allele was 33% in CAD vs. 21% in control (odds ratio, OR, to have CAD for presence vs. absence of CA16 allele in CA16 homozygotes was 4.5, 95% CI 2.1–9.4, P<0.0001; in CA16 heterozygotes or was 1.3, 95% CI 0.94–1.89, P=0.10). The frequency of the major allele (CA15) was 43% in CAD vs. 56% in controls (in CA15 homozygotes OR 0.33, 95% CI 0.20–0.52, P<0.0001; in heterozygotes OR 0.41, 95% CI 0.26–0.63, P<0.0001). In a stepwise logistic regression model the CA16 allele was significantly associated with overweight (OR 1.44, 95% CI 1.0–1.9, P=0.027). Apolipoprotein A-I was elevated (P<0.0001), as was high-density lipoprotein (P=0.098), and severity of angina was decreased (P=0.024) as a function of genotype. Plasma soluble (s) TNF-R2 was 5.1-0.1 ng/ml in CAD vs. 3.2-0.1 in control (P<0.0001), 5.2-0.1 in the presence vs. 4.6-0.2 in the absence of vessel disease (P=0.009), and rose with increasing severity of angina: 4.2-0.2 (no angina), 5.0-0.1 (stable angina), 5.4-0.2 (unstable angina; P=0.003). sTNF-R2 was correlated with age, cholesterol, creatinine, fibrinogen, transforming growth factor # and homocysteine and was influenced by TNFRSF1B genotype. Thus genetic variation in or near the TNFRSF1B locus may predispose to CAD.