This study examined the relationship between breath pentane and plasma lipid peroxide levels sampled simultaneously in patients with stable angina (n = 17), unstable angina (n = 23), and controls (n = 10). Plasma lipid peroxides were measured in venous blood as the adduct formed between thiobarbituric acid and malondialdehyde (MDA) using high performance liquid chromatography. Pentane was measured in end-expiratory air using gas chromatography. MDA concentrations in stable (1.81 +/- 0.84 mumol/l) and unstable (1.5 +/- 1.23 mumol/l) angina were not different. However, both groups had significantly (p < 0.005) elevated MDA levels compared to controls (0.41 +/- 0.26 mumol/l). Breath pentane was 0.20 +/- 0.12 nmol/l in controls and not different from stable angina (0.26 +/- 0.20 nmol/l) or unstable angina (0.15 +/- 0.07 nmol/l). When the data from all three groups were combined, there was no correlation between pentane and MDA (rho = -0.09, p = 0.54). In five of the unstable angina patients treated with balloon angioplasty, MDA in pulmonary arterial blood rose by 69 +/- 15% (p < 0.01), and breath pentane rose by 73 +/- 20% (p < 0.01) immediately after balloon deflation. One minute after balloon deflation MDA and pentane had returned to preinflation levels. The results suggest that basal levels of pentane are less useful than MDA as an index of lipid peroxidation in patients with coronary artery disease. However, breath pentane appears to be a sensitive index of reperfusion-induced lipid peroxidation.