Large 24-h blood pressure (BP) variability and an excessive drop in BP during nighttime are associated with a higher risk of cardiovascular events. Data are lacking regarding the prognostic significance of variability in BP measured during office visits. We analyzed the relationship between office BP variability and the risk of brain infarction in elderly patients receiving antihypertensive therapy. Patients who experienced their first-ever stroke at the age of 60 years or over were registered in the study. At least 2 sex- and age-matched control patients were registered for each case patient. Office BP at each clinic visit and known cardiovascular risk factors were recorded. The BP variability was defined as the variation coefficient (VC) of office BP. In this report, we analyze the data of brain infarction patients. The VC of both systolic and diastolic BPs was significantly higher in the brain infarction patients than in the control patients. Higher office BP variability was associated with a higher risk of brain infarction after adjustment for BP level and other confounding factors. Regarding diastolic BP, the association of brain infarction with the maximal value for the difference of office BPs taken at any consecutive two visits (Max-deltaBP) or the difference between the highest and lowest values of office BP (BP-range) recorded during a 1-year period prior to the event was also significant. In conclusion, a retrospective case-control study suggested that office BP variability was an independent predictor of brain infarction. Either the Max-deltaBP or the BP-range may be surrogate indices of diastolic BP variability.