Why the study?
Does out-of-the-office blood pressure monitoring improve prognostic prediction of cardiovascular events compared to office blood pressure in the general and hypertensive populations?
Does out-of-the-office blood pressure monitoring improve prognostic prediction of cardiovascular events compared to office blood pressure in the general and hypertensive populations?
Answers are grounded in this paper and cite their sources.
Out-of-the-office blood pressure monitoring, particularly 24-hour, daytime, and night-time systolic blood pressure, provides important prognostic information for cardiovascular events that is often independent of conventional office measurements.
After the initial investigations by Perloff et al. [1,2], our knowledge on the prognostic significance of out-of-the-office blood pressure measurements has tremendously increased in the last decade [3–27]. Blood pressure was most often assessed by use of ambulatory blood pressure monitoring [1–7,9–25,27], although in some reports blood pressure was measured at the patients’ home [8,26,27]. Studies have been performed in the general population [5,6,8,9,14,19], in elderly [16,22], in hypertensive patients in general, with or without treatment [1–4,7,11,17,18,21,23,24,26,27], in older patients with systolic–diastolic or isolated systolic hypertension [12,13,15,20,25] and in refractory hypertension [10]. In the current issue of the journal, Björklund et al. [28] extend our knowledge by assessing the prognostic significance of ambulatory blood pressure in 70-year-old men who participated in a prospective longitudinal population-based follow-up study in Uppsala, Sweden. These studies, particularly those on 24-h ambulatory blood pressure monitoring, allow us to assess the prognostic significance of various aspects of out-of-the-office blood pressure: (i) average 24-h blood pressure; (ii) daytime blood pressure; (iii) night-time blood pressure; (iv) the nocturnal blood pressure decline or the night-day blood pressure ratio; (v) out-of-the-office pulse pressure; (vi) blood pressure variability and (vii) the prognostic significance of special types of hypertension, such as white-coat hypertension and masked hypertension. Whereas most studies addressed a limited number of these blood pressure characteristics, Björklund et al. [28] report on all aspects except the categorical analysis, which they considered at least partly in a previous article [22]. When the data of Björklund et al. [28] are considered together with those from the other reports, the results appear to be quite consistent for some blood pressure variables, but not for others. However, the studies may differ in various aspects, apart from the study population. These include the quality of the office blood pressure measurement; the technique of ambulatory blood pressure monitoring; the definition of daytime and night-time blood pressure; the outcome variables, which mostly include cardiovascular mortality, myocardial infarction and stroke, sometimes also total mortality, heart failure, angina pectoris, transient ischaemic attack and/or peripheral vascular disease; the statistical analysis; the confounding factors that have been taken into account in the multivariate analyses, and the adjustment for other blood pressure measurements. Despite these differences among studies, it may now be appropriate to review what has been achieved in the last decade. In this short overview, outcome is related to cardiovascular mortality in the Japanese population study and to various aggregates of cardiovascular events in the other populations. Twenty-four hour blood pressure Björklund et al. [28] report that 24-h systolic blood pressure (SBP) but not diastolic blood pressure (DBP) significantly predicts an aggregate of cardiovascular events in a population of elderly men, after adjustment for antihypertensive treatment, hyperlipidemia, diabetes mellitus and smoking. This is compatible with the findings in older men and women with hypertension in general [17] or with isolated systolic hypertension [12]. In a Japanese population study of at least 40-year-old subjects [6], and in investigations in younger hypertensives, both 24-h SBP and DBP appeared to carry significant prognostic information [17,18,21,23]. However, when Björklund et al. [28] included office SBP in the multivariate model, the predictive power of 24-h SBP was no longer significant [28]. This contrasts with other reports in which the prognostic significance of 24-h SBP [6,12,18,21,23] and/or 24-h DBP [18,21,23] did persist after adjustment for office blood pressure. Daytime blood pressure Similar to the findings for 24-h blood pressure, Björklund et al. [28] observed that daytime SBP, but not DBP, independently predicted cardiovascular morbidity, except when office blood pressure was added to the model. The results were similar in previous studies in elderly hypertensives [12,17] except that the prognostic power of daytime SBP was independent of office blood pressure when this was tested [12]. Again, in studies in younger subjects, SBP and DBP appeared to be both predictive [6,17,18,23] and independent of office blood pressure when tested [18,23]. Night-time blood pressure In 70-year-old men from Uppsala, night-time SBP, but not DBP, was a significant independent predictor for cardiovascular events [28], as was also observed in elderly hypertensives [12,17]. In contrast to observations in elderly individuals with systolic hypertension [12], significance was lost after control for office blood pressure in the Swedish population study [28]. In the Japanese general population, only night-time DBP predicted cardiovascular mortality [6]. In younger hypertensives, night-time SBP and DBP were both predictive [11,18,23] but when office blood pressure was added to the model, night-time blood pressure was no longer significant [18], or only systolic blood pressure remained significant [23]. In two studies, one in elderly individuals with systolic hypertension [12] and one in older patients who were registered in one general practice [27], daytime and night-time blood pressure were entered in the same multivariate model. In both studies, night-time blood pressure was found to be a stronger predictor of risk than daytime blood pressure. Night–day difference in blood pressure or night–day blood pressure ratio The night–day ratio of SBP and DBP did not carry any prognostic information in the study by Björklund et al. [28] and similar results were also obtained by Khattar et al. [11,17] and Clement et al. [23] in hypertensive patients. By contrast, Staessen et al. [12] found that the night–day ratio of SBP was significantly related to cardiovascular endpoints independent of the 24-h blood pressure. Ohkubo et al. [19] reported that a greater decline in nocturnal blood pressure was associated with a lower cardiovascular mortality, and that there was no significant interaction with 24-h blood pressure. In a case–control analysis, Verdecchia et al. [3] found that the nocturnal decline of blood pressure was significantly smaller in hypertensive women with subsequent cardiovascular events than in women without events, and this was confirmed in a subsequent analysis [4] in which non-dipping women appeared to have a worse prognosis than dipping women; this was not the case in men. Few studies have looked at the prognostic significance of extreme dippers and of reverse or inverted dippers. In older Japanese hypertensives, Kario et al. [15] observed that the relative risk for stroke was greater in extreme dippers than in dippers and non-dippers; by contrast, cardiovascular mortality was similar in extreme dippers and dippers in the Ohasama population study [5]. Furthermore, the results on inverted dippers were not consistent [5,15]. Pulse pressure The question arises of whether pulse pressure obtained from ambulatory blood pressure monitoring is a better predictor than office pulse pressure. Björklund et al. [28] observed that this was indeed the case for 24-h and daytime pulse pressure, but not for night-time pulse pressure in elderly men. In older patients with systolic hypertension, ambulatory pulse pressure was also a better predictor than conventional pulse pressure [20]; moreover, the prognostic significance of ambulatory pulse pressure was independent of 24-h mean or systolic blood pressure. In agreement with studies on conventional pulse pressure measurements, Khattar et al. [17] found that 24-h, daytime and night-time pulse pressure were significant predictors in elderly hypertensives, but not in younger patients; however, they did not include office pulse pressure, or mean or systolic blood pressure in the models. In younger treated hypertensives, neither office nor ambulatory pulse pressure independently predicted the risk of cardiovascular events according to a model including ambulatory SBP or DBP [23]. Blood pressure variability In studies on ambulatory blood pressure monitoring, the variability of blood pressure is usually assessed by the standard deviation, respectively, for daytime and night-time blood pressure. In the elderly Uppsala population, SBP and DBP variability carried significant prognostic information for daytime, but not night-time, blood pressure and significance only persisted for SBP, after adjustment for 24-h SBP [28]. The results were similar in the Ohasama population study, where only daytime SBP variability was independently predictive for cardiovascular mortality [14]. By contrast, measures of blood pressure variability did not predict cardiovascular events in studies on hypertensive patients [11,17]. White-coat hypertension and masked hypertension The prognostic significance of out-of-the-office blood pressure can also be assessed by categorizing patients according to office blood pressure and ambulatory or home blood pressure. There is little doubt that patients with hypertension in the office and out-of-the-office have a worse prognosis than patients with white-coat hypertension or isolated clinic hypertension [4,7,10, 13,16,18,27]. A few studies have compared the outcome of patients with white-coat hypertension with true normotensives (i.e. individuals with normal office and ambulatory or home blood pressure). Whereas prognosis was similar in the two groups in four studies [4,15,26,27], one study found that outcome was worse in white-coat hypertensives than in true normotensives [24]. Most convincing are two reports in which normotensives and hypertensives were drawn from the same study population and in which patients with masked hypertension were excluded from the normotensive group [26,27]. It appears that approximately 10% of subjects who are normotensive in the clinic have high blood pressure when measured out-of-the-office. Two recent studies report that the outcome of these so-called masked hypertensives is worse than that of true normotensives [22,26]. Conclusions Average 24-h, daytime and night-time blood pressure are predictive of cardiovascular events, but this was only the case for SBP in most studies performed in elderly individuals. Independence from office blood pressure was shown in most studies when tested, but several studies have failed to perform or report this important analysis. The results on the prognostic significance of the decline in nocturnal blood pressure or on the night–day ratio are not consistent, and when a lower night–day ratio was associated with a better prognosis, independence from average 24-h blood pressure was not always tested. Ambulatory pulse pressure appeared to predict cardiovascular events in elderly individuals, sometimes better than office blood pressure; however, independence from ambulatory mean or SBP was not often assessed. Data on blood pressure variability are scarce and inconsistent. Finally, there is little doubt that patients with white-coat hypertension have a better prognosis than patients with sustained hypertension and data are emerging indicating that their prognosis may be similar to that of true normotensives. More data are needed on the prognostic significance of masked hypertension. However, one limitation of this overview is that most studies reported on a selected number of blood pressure characteristics and we do not know if these characteristics were not analysed or just not reported, possibly introducing bias. Pooling of databases for further analyses and meta-analyses may yield more precise and powerful conclusions.
No takes yet. Share an insight, caveat, or question.
Fagard et al. (2004) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: