A method for using MR to evaluate the effects of cardiovascular disease on the brain: the cardiovascular health study.

PURPOSE To do a pilot study for the Cardiovascular Health Study (a population-based, longitudinal study of coronary heart disease and stroke in adults 65 years of age and older designed to identify risk factors related to cerebrovascular disease, particularly stroke): (a) to determine the feasibility of adding brain MR to the full-scale study; (b) to evaluate the reliability of standardized MR image interpretation in a multicenter study; and (c) to compare the prevalence of stroke determined by MR with that by clinical history.

METHODS Protocol-defined MR studies were performed in 100 subjects with clinical histories of stroke and 203 subjects without reported histories of stroke. MR scans were independently evaluated by two trained neuroradiologists for the presence of small (< or = 3 mm) and large (> 3 mm) "infarctlike" lesions. The sizes of the cerebral sulci and lateral ventricles and the extent of white matter disease were graded on a scale of 0 to 9.

RESULTS Eighty percent of the Cardiovascular Health Study participants who were invited to undergo MR studies agreed to do so; 95% of those agreeing to the procedure successfully completed the exams. Intrareader and interreader reliability of infarctlike lesion identification was high for large lesions (kappa, 0.71 and 0.78, respectively) but not for small lesions (kappa, 0.71 and 0.32, respectively). Relaxed intrareader and interreader kappa scores for sulcal and ventricular sizes and extent of white matter disease were greater than 0.8 MR evidence of infarctlike lesions was present in 77% of the participants with histories of stroke but was also present in 23% of the participants without clinical histories of stroke. Seventy-nine percent of the infarctlike lesions were larger than 3 mm.

CONCLUSIONS This preliminary study indicates that a large, prospective, epidemiologic study of elderly subjects using MR scans of the brain for identification of cerebrovascular disease is feasible and that the interpretative results are reproducible, and suggests that MR evidence of stroke is more prevalent than reported clinical history of stroke.