OBJECTIVE This study was conducted to evaluate the role of high serum lipoprotein(a) levels in a group of patients with a relatively early onset of cerebral infarction as a whole and in a subgroup with the perforating artery occlusion subtype of cerebral infarction. METHODS Fifty-four patients with cerebral infarction, the onset of which was before age 65 years (37 men, 17 women; mean age, 61.9 +/- 7.7 years) were examined in this study. When patients with atrial fibrillation were excluded to omit cardiac embolic strokes from analysis, the group consisted of 45 patients. The patients were classified into two subtype groups, the perforating artery occlusion group and the cortical artery occlusion group, by using magnetic resonance imaging. Lipoprotein(a) levels were measured by an enzyme-linked immunosorbent assay. Four biochemical variables (serum levels of lipoprotein(a), high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and triglycerides) and other potential risk factors such as hypertension, diabetes mellitus, smoking habits, alcohol intake, and family history were analyzed by stepwise logistic regression to determine the independent and significant risk factors for cerebral infarction without atrial fibrillation. RESULTS The incidence of subjects with serum lipoprotein(a) levels > or = 42.6 mg/dL, which is the 95th percentile level of the control subjects, was significantly increased in the total cerebral infarction group (P < .025) and the perforating artery occlusion group (P < .025) compared with the control group. In addition, by using stepwise logistic regression analysis in the total and perforating artery occlusion patient groups we identified three independent and significant risk factors: hypertension, a low high-density lipoprotein cholesterol level, and a high serum lipoprotein(a) level. In the cortical artery occlusion group, the sample size was not large enough for the statistical analysis. Diabetes mellitus is the only known factor that correlates with serum lipoprotein(a) levels, but there were no significant correlations between serum lipoprotein(a) levels and history of diabetes mellitus or fasting blood sugar. CONCLUSIONS These findings indicate that high serum lipoprotein(a) levels are an independent risk factor in the development of cerebral infarction when subjects with atrial fibrillation were excluded from the total group and the perforating artery occlusion subtype group.