Orthostatic hypotension (OH) is a potential risk factor for adverse cardiovascular events, but OH is highly variable and may not be detected on a single occasion. To assess the relation between intra-individual variability of systolic orthostatic blood pressure change (DeltaSBP) and cardiovascular outcomes, an algorithm was developed to identify DeltaSBP instability using repeated supine and standing BP measurements. A cohort of 673 nursing home residents underwent baseline postural BP measurements (supine to 1 minute of standing, four times in a single day) and were followed for up to 2 years. Two groups (stable vs. unstable) were identified based on an analysis of DeltaSBP variance components. Differences in outcomes were compared via Cox survival analysis. At baseline 12.6% were unstable, defined as a one standard deviation difference of at least 20.2 mmHg between DeltaSBP readings. Unstable subjects were more likely to have OH on at least one measurement (systolic BP drop of 20 mmHg or more; 85% vs. 36%, respectively) and to be on psychotropic medication at baseline (47% vs 35%) (P-values <0.001). Other characteristics (including previous stroke) did not differ. During a mean follow-up of 10.3 months, stroke incidence was higher in unstable subjects (13.1% vs. 4.9%; P = 0.012), but ischemic heart disease and mortality rates were not significantly different (respectively, 13.5% vs. 7.4%, P = 0.115; 14.8% vs. 10.7%, P = 0.178). Survival analyses (adjusted for age, sex, psychotropic medications, body mass index, ischemic heart disease, and supine systolic pressure) confirmed a higher risk of stroke in unstable subjects (relative risk = 3.7, 95% CI: 1.6-8.4). Highly variable orthostatic BP measures may reflect impaired BP regulatory mechanisms in elders with occult cerebrovascular disease, or may directly affect cerebral blood flow. Orthostatic BP variability may be a better indicator of future stroke than a single supine or orthostatic change measure.