Subclinical brain infarcts (SBI) are common in the elderly and are associated with covert neurologic and cognitive impairment. Although renal impairment is associated with accelerated cerebrovascular disease and an increased risk for clinically apparent brain infarct, few studies have examined the relationship between renal function and SBI, and these may have been limited by the inaccuracy of creatinine as a renal function marker. A cross-sectional study was performed among older adults in the Cardiovascular Health Study to examine associations between SBI and two serum markers of renal function: Serum creatinine (SCr) and cystatin C (CysC). Patients had cranial magnetic resonance imaging and renal markers measured in 1992 to 1993. Logistic regression was used to estimate the associations between renal function (estimated by 1/SCr and 1/CysC) and SBI, controlling for potential confounding factors. SBI were present in 789 (28.7%) of 2784 participants. A linear association with SBI was observed for 1/CysC (per 1-SD decrement; odds ratio [OR] 1.20; 95% confidence interval [CI] 1.09 to 1.32; P < 0.001) but not for 1/SCr (OR 1.08; 95% CI 0.98 to 1.19; P = 0.14), for which a quadratic U-shaped association was suggested (P = 0.004). In a model with both markers, 1/CysC was linearly associated with SBI (OR 1.26; P < 0.001), whereas 1/SCr was not (OR 1.06; P = 0.3). The prevalence of SBI was directly associated with quintile of CysC, whereas the association between SCr and SBI was U-shaped, with greater prevalence at high and low levels. Compared with creatinine, CysC, a novel marker of renal function, has a stronger and more direct association with SBI in the elderly.