Background An exaggerated blood pressure response (ExBPR) to exercise has been shown to be predictive of future hypertension and left ventricular hypertrophy. The aim of this study was to test the reproducibility over time of ExBPR and to better characterize patients who consistently respond in this manner. Methods During periodical health examination, patients underwent routine treadmill exercise testing. ExBPR was said to be present if systolic blood pressure and/or diastolic blood pressure at peak exercise exceeded 200 mm Hg and 100 mm Hg, respectively. Over the past 25 years, 117 healthy patients with ExBPR performed 2 to 7 consecutive treadmill exercise tests. According to subsequent ExBPR, these patients were divided into a concordant group—those who had at least two-thirds repetitions of the ExBPR—and a discordant group—those who had less than two-thirds repetitions. For comparison, we identified patients who did not have ExBPR (control group). Results Of the 117 patients who had ExBPR, only 18 (15.4%) were in the concordant group in subsequent tests. No clinical features were found to characterize patients in the concordant group. In the two study groups, the variability of blood pressure measurements during stress was significantly greater than in the control group. Also, systolic blood pressure measurements at rest and after 3 minutes of recovery were significantly lower in the control group. Conclusions ExBPR to exercise is rarely reproducible, and there are no clinical findings characterizing those who consistently respond in this manner. Thus the prognostic importance of blood pressure response to exercise should be reconsidered. (Am Heart J 2001;141:1014-7.)