The optimal schedule for self-home blood pressure monitoring

In 2008, the European Society of Hypertension (ESH) [1] and the American Heart Association [2] published guidelines for self-monitoring of blood pressure by patients at home (HBPM), providing almost identical recommendations with only minor differences. Both guidelines recommended HBPM for long-term monitoring of treated hypertension and also for the initial diagnostic phase of subjects with elevated blood pressure (BP) [1,2]. Moreover, a common recommendation was given for the optimal HBPM schedule, which should be based on a 7-day monitoring with duplicate morning and evening measurements, and the average should be calculated after discarding readings of the first day [1,2]. The amount of research work, discussions and debate on the optimal HBPM schedule is a typical example of ‘knowing more and more for less and less’, as Tom Pickering stated in the foreword of his book on ‘Ambulatory Monitoring and Blood Pressure Variability’ [3]. Several studies investigated the performance of HBPM in the diagnosis of hypertension phenotypes (sustained, white-coat, masked hypertension) in untreated and treated subjects, by taking ambulatory BP monitoring (ABPM) as reference method and implementing different HBPM schedules (Table 1) [4–18]. Given that hypertension remains the leading risk factor for death and disability globally [19] and HBPM is recommended for almost all subjects with elevated BP [1,2], the optimal HBPM schedule indeed is highly relevant for cardiovascular disease prevention.TABLE 1: Diagnostic performance of home blood pressure monitoring versus ambulatory blood pressure taken as reference (studies in untreated subjects)In this issue of the journal, Nunan et al.[18] present the results of a carefully conducted study in 203 untreated subjects investigating the optimal HBPM schedule by taking ABPM as reference. The main conclusion of the authors is that 5-day HBPM provides the optimal schedule, and there is no additional benefit by performing more measurements or by excluding the first day. Strengths of this study are: (i) it was conducted in primary care where the vast majority of hypertensive patients are managed, (ii) validated HBPM devices and a tele-monitoring system were used, preventing thereby the ‘Achilles heel’ of the method, which is the misreporting of HBPM values by patients [20], and (iii) the HBPM schedule recommended by current guidelines was the basis of investigation [1,2]. However, (i) these results are limited to untreated subjects; (ii) diagnostic disagreement of HBPM with ABPM does not necessarily imply that the former is problematic; and (iii) systolic daytime ABPM differed by 7.5 mmHg from systolic HBPM [18], which contrasts with previous reports [1,2] and might have affected the study results. A critical overview of the evidence on the optimal HBPM schedule should address three main research questions: (i) Which is the minimum number of HBPM readings required for a reliable evaluation? (ii) Should readings of the first HBPM day be discarded? (iii) What are the pitfalls of taking ABPM as reference? MINIMUM NUMBER OF HOME BLOOD PRESSURE READINGS It is well established knowledge that with any BP measurement method (office, home, or ambulatory) by increasing the number of readings: (i) the average value is reduced, (ii) the variability [standard variation (SD)] is reduced, (iii) the reproducibility is improved, and (iv) the prognostic ability improved [21–24]. Indeed, it has been shown that the first HBPM reading gives the highest value, the most unstable one (highest SD) and with the weakest prognostic ability [21,23,25]. Furthermore, the first HBPM day gives higher average value than the next days, with higher SD and weaker correlation with ABPM [21]. It is therefore clear that a single HBPM reading or day does not exhaust the potential of the method, and more readings should be obtained. On the other hand, it is reasonable to expect that after obtaining a certain amount of HBPM readings, additional readings will not provide additional benefit. Brook [26] reviewed 12 studies using different HBPM schedules and concluded that variation in schedule has little impact on improving the relationship with ABPM and the majority of the benefits derived from the method are achieved by obtaining only a few measurements. Ohkubo et al.[23] analyzed data from the Ohasama outcome study and showed progressive increase in the prognostic value of HBPM by increasing the number of measurements within the range of 1–14 and concluded that preferably more than 14 measurements should be recommended for better prediction of stroke risk. Stergiou et al.[25] reported an analysis of the Didima outcome study and showed progressive increase in the prognostic value of HBPM by averaging more measurements within the investigated range of 1–12. Johansson et al.[27] showed that a minimum of 4-day HBPM is reliable in terms of association with ABPM and target organ damage. Other cross-sectional studies showed progressive improvement in the stability of HBPM (reduction of average values and SD) and its association with ABPM by averaging more measurements, with little benefit after obtaining measurements of more than 3 days [21,28,29]. The findings by Nunan et al.[18] support the 3-day schedule as minimum (not optimal) ESH requirement for HBPM [1] by showing marginal improvement in sensitivity and specificity with more than 3 days, whereas 2-day schedule appeared to have significantly lower specificity. IMPACT OF THE FIRST HOME BLOOD PRESSURE MONITORING DAY Current guidelines in Europe [1], United States [2], and UK [30] recommended to discard HBPM measurements of the first day. There are at least three reasons why this recommendation appears to be reasonable. First, as mentioned above, several studies have shown that the first day gives higher and more unstable values, which are less closely correlated with ABPM and cardiovascular events risk [21,23,25,29]. Second, when averaging measurements of different days, the impact of including or excluding readings of the first day is progressively reduced by increasing the number of readings/days taken into account [21]. Thus, when 5-day HBPM is performed, as proposed by Nunan et al.[18], or with 7-day HBPM as is the full schedule recommended by the ESH [1], the impact of the first day is expected to be negligible. On the other hand, when 3-day HBPM is available, which is the minimum requirement recommended by the ESH [1] and not uncommon in clinical practice, the impact of the first day is expected to be significant as the first day contributes to the average by one third. Thus, exclusion of the first day is advisable when 3-day or 4-day HBPM is available, whereas this is probably not needed when 5-day or more HBPM is available. However, this appears to be an impractical and confusing recommendation for clinical practice. Third, the ‘first-day phenomenon’ is not only present in the initial diagnostic phase, but appears to persist in repeated HBPM sessions in treated subjects [31]. In a randomized crossover study comparing two antihypertensive drugs, the second HBPM day had greater precision than the first day in demonstrating differences in drug efficacy [31]. Therefore, a universal recommendation should be provided to discard readings of the first day in the initial and follow-up HBPM sessions in untreated and treated subjects, which is more important when 3–4 HBPM days are available and less so with 5 or more days. This recommendation is supported by the data by Nunan et al.[18], because the average HBPM of days 2–7 gave better (at least arithmetically) sensitivity, specificity, positive and negative predictive value, and diagnostic agreement (percentage and kappa statistic) than HBPM of days 1–7. Moreover, the average HBPM of days 2–5 gave better sensitivity, positive and negative predictive value, and diagnostic agreement (percentage and kappa statistic) than HBPM of days 1–5. Had fewer HBPM days been taken into account in that study (e.g. 3 days which is the minimum ESH requirement [1]), the benefit of excluding day 1 might have been larger. The authors accepted that ‘the largest proportion of correctly classified participants was obtained from the average of HBPM days 2–5’, which discards the first day [18]. PITFALLS OF AMBULATORY BLOOD PRESSURE MONITORING AS REFERENCE METHOD ABPM is an attractive technique because it provides detailed information on the BP behavior during different daily activities, at work, at home, and during sleep, and has strong data on its prognostic ability [32]. However, although ABPM provides much more reproducible values than clinic BP [22], its reproducibility is imperfect. A study in 133 untreated subjects with elevated clinic BP who performed ABPM twice, 2 weeks apart, always on a routine workday, showed 59% of the subjects diagnosed as having hypertension on the basis of the initial ABPM and 62% on the second one [7]. However, there was disagreement in hypertension diagnosis between the two recordings in 25 individuals (19%; 14/133 subjects classified as hypertensives only on the first ABPM and 11 only on the second) [7]. These data suggest that when comparing hypertension diagnosis based on HBPM versus ABPM (taken as reference method), the diagnostic disagreement is at least partially attributed to imperfect reproducibility of the methods (applies to both HBPM and ABPM). In fact, the diagnostic disagreement between the two methods (Table 2) [6,7,10,11,15,16,22,33–41] is very close to the diagnostic reproducibility of the methods, which implies that they may be regarded as interchangeable in the evaluation of subjects with elevated BP.TABLE 2: Reproducibility of ambulatory and home blood pressure and diagnostic agreement in hypertension diagnosisThe fact that the available evidence and current guidelines suggest that the same threshold should be used for HBPM and daytime ABPM (135/85 mmHg for systolic/diastolic BP) [1,32] is rather surprising. ABPM and HBPM are similar methods in that they both provide multiple BP measurements taken in the usual environment and daily activities of each individual. However, there are important differences between the two methods, because ABPM is performed at different postures (standing, sitting, and lying) and environment (work, home, and during sleep), whereas HBPM is taken only in the sitting posture at home. These differences explain the fact that despite the similar average levels of HBPM and daytime ABPM, the latter has significantly wider BP range [33]. Another contributor to differences in HBPM and ABPM values is the use of different devices, which was the case in almost all the comparative trials with rare exceptions [33]. More importantly, the Pressioni Arteriose Monitorate E Loro Associazioni (PAMELA) general population study showed that among subjects with white-coat hypertension, those with low HBPM and ABPM had lower cardiovascular mortality than those with only one of them being low [42]. These findings suggest that the two methods reflect similar but also different aspects of the BP profile and therefore have a complementary rather than competitive role in the evaluation of subjects with elevated BP. Thus, full diagnostic agreement between ABPM and HBPM should not be expected. It is important to mention that when using only daytime ABPM as Nunan et al.[18] did in their study, which in fact is what is recommended in UK where the study was conducted [30], the full potential of the technique is not exhausted. This is because daytime ABPM gives less reproducible values than the 24-h average [22] and nocturnal ABPM is omitted, which appears to be the best aspect of the method in terms of prognostic ability [32]. For these reasons, the recent ESH guidelines on ABPM recommend to take into account the 24-h average rather than daytime alone for making the diagnosis of white-coat and masked hypertension [32]. REMAINING RESEARCH QUESTIONS There is international consensus for the optimal HBPM schedule to be applied in the management of hypertension in clinical practice [1,2,30]. However, despite the considerable amount of relevant published work, indeed there are some aspects that remain uncertain. Future analyses need to address the following questions: (i) what is the impact of the first day when the minimum recommended HBPM schedule (3 days) is obtained; (ii) whether the diagnostic accuracy of HBPM differs in untreated compared with treated hypertensives; (iii) whether the HBPM schedule performs the same in children and adolescents as in the adults; (iv) what is the difference in diagnostic agreement of HBPM with ABPM when 24-h instead of daytime ABPM is used. ACKNOWLEDGEMENTS Conflicts of interest There are no conflicts of interest.