In a population-based case-control study, we investigated the association of acute ischaemic stroke with lipoprotein(a) (Lp(a)) levels and apolipoprotein (Apo) (a) isoform size in subjects aged older than 70 years. A total of 163 patients with a first-ever-in-a-lifetime acute ischaemic/nonembolic stroke and 166 controls were included. Compared to controls, stroke patients exhibited higher Lp(a) concentrations (median value, 12.2 mg/dl versus 6.4 mg/dl, p < 0.001) and a higher frequency of small Apo(a) isoforms (44.2% versus 29.5%, p < 0.01). Multivariate logistic regression analysis showed a significant association of acute ischaemic stroke with Lp(a) levels [adjusted odds ratio (OR), 1.37, 95% CI (1.12-1.67); p = 0.002], and small Apo(a) isoform size [OR, 1.74 (1.10-3.03); p = 0.04]. Compared to subjects with Lp(a) levels in the lowest quintile, those within the highest quintile had a 3.2-times adjusted risk to suffer an acute ischaemic/nonembolic stroke (1.60-6.62, 95% CI; p < 0.001). Furthermore, analysis of interaction between lipid variables revealed that in the presence of elevated Lp(a) levels the inverse relationship between HDL-cholesterol levels and ischaemic stroke was negated [OR, 1.01 (1.00-1.03); p = 0.015]. Our study suggests that determination of Lp(a) levels and Apo(a) isoform size may be important in identifying elderly individuals at risk of ischaemic stroke independently of other risk factors and concurrent metabolic derangements.