Borderline hypertension is characterized by pressor hyperreactivity to mental stress. However, it has not been shown whether blood pressure hyperresponsiveness is a temporary phenomenon due to situational anxiety or a stable feature of the borderline hypertensive state. We therefore evaluated the long-term stability of invasively assessed blood pressure and central hemodynamic responses to mental stress in 10 young male subjects with borderline hypertension recruited from a population screening. Two identical 10-min mental arithmetic stress tests were performed 50 to 62 months apart (median, 4 years 8 months). Intraarterial blood pressure was monitored continuously before, during, and after stress. Cardiac output was measured by the indocyanine green dye dilution technique and indexed for body surface area. Total peripheral resistance index was computed from cardiac index and mean arterial pressure. During the 4-year follow-up period, none of the central hemodynamic parameters had changed significantly, either with respect to rest or stress levels. Test-retest variability of blood pressure measures was low, and errors of measurement ranged between 4.8 and 8.2 mm Hg for blood pressure levels at rest and during stress. Mental arithmetic induced highly significant blood pressure increments on both occasions (ANOVA, P < .0001 throughout). Pressor responses were somewhat but not significantly lower during the second test. Errors of measurement for absolute blood pressure reactivity ranged between 3.9 and 7.1 mm Hg. Intersession correlation coefficients for blood pressure levels attained during stress were above r = 0.75 throughout (P < or = .01).(ABSTRACT TRUNCATED AT 250 WORDS)