Accelerated atherosclerosis in diabetes has been suggested as being due to an enhanced oxidative modification of LDL. We hypothesized that the titers of autoantibodies against oxidized LDL (oxLDL) may be increased in patients with non-insulin-dependent diabetes mellitus (NIDDM) and that they may contribute to various manifestations of atherosclerosis among such patients. In a 10-year follow-up study of 91 newly diagnosed NIDDM patients and 82 nondiabetic control subjects, autoantibodies against oxLDL (expressed as the ratio of autoantibodies against oxLDL and native LDL) were measured at baseline and after 10 years. Quantitative ultrasonography to examine the intimal-medial thickness of the common carotid artery (a morphological index of arterial wall injury) and carotid bifurcation was performed at the 10-year examination. The relationship of autoantibodies against oxLDL to the occurrence of cardiovascular death, fatal and nonfatal myocardial infarction, stroke, and any cardiovascular event as well as to the intimal-medial thickness of the common carotid artery and carotid bifurcation was evaluated. Associations between these autoantibodies and metabolic variables (fasting glucose, glycosylated hemoglobin, insulin, and serum lipids) in NIDDM patients were also examined. Autoantibodies against oxLDL did not differ between NIDDM and control subjects (NIDDM: baseline, 1.63 and 0.61 to 23.6; 10-year examination, 1.64 and 0.06 to 59.0; control group: baseline, 1.84 and 0.13 to 36.0; 10-year examination, 1.50 and 0.25 to 8.29; median and range, P = .62, baseline; P = .45, 10 year). In both groups, the titers of these autoantibodies measured at baseline and after 10 years significantly correlated with each other (r = .63 for the diabetic and r = .51 for the control group, respectively, P < .001 for each). The frequency of all cardiovascular events was markedly higher in the NIDDM group than in the control group, but autoantibodies against oxLDL had no significant association with any of these events, including cardiovascular mortality. At the 10-year examination the intimal-medial thickness of the common carotid artery (1.24 +/- 0.36 versus 1.06 +/- 0.30 mm, P = .002) and carotid bifurcation (2.11 +/- 0.73 versus 1.77 +/- 0.82 mm, P = .01) were greater in NIDDM patients than in control subjects, but autoantibodies did not show any association with the intimal-medial thicknesses in either the diabetic or control groups. Autoantibodies against oxLDL indicate the presence of oxidatively modified LDL in vivo, but their titers in the serum do not seem to associate with the excess cardiovascular mortality, morbidity, or intimal-medial thickness of the carotid artery.