To reach a consensus on ambulatory blood pressure (ABP) as a predictor of target-organ damage (TOD), morbidity and mortality.
The members of task force III wrote this article in preparation for the Seventh International Consensus Conference (23–25 September 1999). This article was amended after the meeting to reflect the consensus reached at the conference.
Points of consensus
In most studies, TOD in essential hypertension was more closely associated with ABP than it was with clinic blood pressure, the mean weighted correlation coefficients for the relationship of left ventricular mass with blood pressure being 0.50/0.44 (24 h systolic/diastolic blood pressure) and 0.35/0.32 (clinic systolic/diastolic blood pressure), respectively. The above correlation coefficients vary among studies, possibly because of different standardizations of clinic blood pressure measurements and ways of selecting subjects, among other reasons. The closeness of the association between clinic blood pressure and left ventricular mass increases with the numbers of clinic measurements of blood pressure and visits to a clinic. Thus, the variance of left ventricular mass explained by ABP in addition to that explained by clinic blood pressure diminishes with the number of clinic blood pressure readings. The proportion of variability of left ventricular mass that is directly accounted for by the day-night difference in blood pressure is 15% at the most. Thus, the advantage of ABP over clinic blood pressure appears to be, at least in part, a result of the greater number of measurements over the 24 h. It might also depend, however, on the information offered by ambulatory blood pressure monitoring (ABPM) on daily-life variations in blood pressure. TOD appears to be more closely associated with ABP than it is with clinic blood pressure for the subjects with reproducible ABP tracings, but not for those with poorly reproducible tracings. The probability of developing sustained clinic hypertension at follow-up seems to be better predicted by clinic blood pressure on several occasions over a 6-month period than it is by ABP at baseline, although, when also ABPM is repeatedly performed at follow-up, its ability to predict clinical outcomes of hypertensive patients remains superior to that of repeated clinic blood pressure measurements. ABPM of the elderly appears feasible and is tolerated well. A blunted day-night fall in blood pressure ('non-dipping') seems to be harmful, while evidence regarding the potentially harmful effect of extreme dipping is still limited. Authors of the Syst-Eur study recently demonstrated the prognostic value of ambulatory systolic blood pressure and in particular, of night-time blood pressure, in assessing old subjects with isolated systolic hypertension. The assessment of variability of blood pressure has been shown to provide a further prediction of cardiovascular risk and the potentially prognostic value of beat-to-beat variability assessed non-invasively (using a Finapres or Portapres device) needs further study. In the published event-based studies, the prognostic value of ABP recorded during a single session was superior to that of clinic blood pressure. Since the authors of published event-based prognostic studies compared ABP with only a few clinic measurements of blood pressure, it is not known how many visits or measurements of blood pressure (and at what cost) would equate to a single session of ABPM in terms of prediction of cardiovascular events. ABPM might allow one to identify a subset with 'normal' ABP (white-coat or isolated clinic hypertension). Daytime ABP levels < 135 mmHg systolic and 85 mmHg diastolic can be defined as normal and values < 130/80 mmHg could be defined as optimal. Cardiovascular risk for subjects with normal ABP seems to be lower than that for those with abnormally high ABP. Long-term observational and intervention studies concerning subjects with white-coat hypertension are needed. Among subjects with abnormally high ABP, cardiovascular risk seems to be inversely associated with the day-night difference in blood pressure and directly associated with ambulatory pulse pressure. The potential additional prognostic value of self-measured blood pressure needs further investigation.