Attended, or unattended blood pressure, that is the question

It has been known for a long time that conditions under which blood pressure (BP) is measured are very important. Alam and Smirk were the first to study the effect of rest on blood pressure; they found that during 30 min of rest, SBP dropped by 15--20 mmHg. They wrote that ‘basal blood pressures cannot be obtained by repeated measurements made in a hospital ward, with or without screens around the bed; nor in individuals who suspect that the blood pressure measurements are only a prelude to further procedures. The environment of the individual must be entirely unstimulating. Restlessness on the part of the observer, noises outside the room, and displays of medical equipment may invalidate the result’ [1]. The presence of staff is one of the most important factors, which contributes to the BP increase. In the 1980s, Mancia et al. studied its effect, using invasive measurement in the radial artery [2]. When a physician unknown to the patient entered the room, BP increased by 27/15 mmHg whereas the increase of BP was about half when a nurse entered [3]. More recently, finger BP was measured noninvasively by the beat-to-beat method during the presence of a physician, and a smaller increase was found: by 12/6 mmHg in untreated grade 1 hypertensive individuals [4]. A new approach to measuring BP in the office was made possible when the devices for repeated automated measurements appeared in the market after the year 2000 [5]. Myers and Valdivieso [6] were among the first to show that blood pressure readings taken by the automated device (BpTRU) were lower than those taken by a physician. Since that time, a large number of studies with automated office blood pressure (AOBP), including those without presence of staff, have been done and they brought new insights on blood pressure measurement. Most of them show that unattended AOBP gives lower values than attended, the mean difference of SBP being 5.8 mmHg according the article published in this issue [7]. When unattended AOBP is compared with auscultatory office BP measurements, the difference is still larger: up to 15/8 mmHg [8]. The question raised now by researchers and by authors of guidelines is whether AOBP, particularly unattended, should be preferred over conventional office BP. Actually, results about AOBP are conflicting. Unattended AOBP values were found to be similar and more strongly related to awake ambulatory BP than other office BP measurements [9] but other studies did not confirm these findings [10,11]: in a study undertaken in stable treated hypertensive individuals [10], the associations of both 24-h and daytime systolic ambulatory BP with unattended automated and auscultatory office BP were very similar; all the correlation coefficients were between 0.38 and 0.53 with no significant difference between them. Regarding organ damage, carotid intima--media thickness was associated with unattended AOBP, but not with office BP measured by auscultatory method in 176 firemen [12]. Left ventricular mass was associated with unattended AOBP and 24-h ambulatory BP but not with attended AOBP. (There was a striking difference of 23 mmHg between systolic attended and unattended AOBP; [13]). However, Salvetti et al.[14] recently published a study on 564 outpatients undergoing a diagnostic work-up for known or suspect hypertension. The individuals underwent unattended and attended AOBP measurements using the same device (Omron HEM 9000Ai) in a random order; mean difference between systolic unattended and attended AOBP was 6.5 mmHg. Left ventricular mass index and carotid intima--media thickness were significantly associated with both AOBPs; the differences between these correlations were not statistically significant. This means that we do not have enough evidence that unattended AOBP would predict better hypertension-mediated organ damage. The first large prospective randomized study where AOBP was chosen as the main method of BP measurement was SPRINT [15]. It compared standard and intensive BP control with the goal to have SBP below 140 or below 120 mmHg, respectively, in nondiabetic patients without history of stroke. Although the results were unequivocally in favour of intensive control, there is still a lot of discussion, mainly about BP measurement, which is usually presented as unattended AOBP. Indeed, the AOBP was far from being strictly unattended. In the original protocol, it is stated that ‘seated blood pressure and pulse are measured at each clinic visit after a rest period using an automated device or manual devises if necessary. The preferred method is the automated device as it offers reduced potential for observer biases and decreased demand on staff in terms of training and effort in data collection.’ Later on, the SPRINT Manual of Procedures recommended that ‘the staff leave the room during the rest period but return to take the BPs at the end of the 5-min rest,’ but it did not require staff attendance or absence during the BP measurement [16]. This is a confusing recommendation as the centres were equipped with the Omron 907 devices, which start the BP measurement automatically and the attendance of staff is not necessary. According to the questionnaire data, which were collected from the centres after the publication of SPRINT and whose reliability is questionable, fully unattended AOBP was measured in 4082, that is, in 44% of individuals, whereas in the rest, the member of staff was present during either the resting period, or proper BP measurement, or both phases [16]. However, office BP in SPRINT was lower than in other studies; this was shown when the mean values of 24-h BP monitoring with corresponding office BP values in SPRINT were compared with the PAMELA and ELSA studies [17]. The probable reason for lower office BP values in SPRINT is that auscultatory measurements were not (or very rarely) done. Taken together, the SPRINT study does not give enough strong arguments for unattended AOBP in preference to other office BP measurements. A common argument for the use of unattended AOBP is that it eliminates white-coat effect (WCE). It is probably only diminished, not eliminated, as the WCE may have two components: hospital effect and physician effect [18]. Moreover, WCE calculated classically, that is, as the difference between office and out-of-office BP measurement, is not the same as WCE in the office [4]. In any case, if only unattended AOBP were measured in the office, white-coat hypertension (WCH) would not be detected and the basic information on some 30% of individuals with office hypertension would be lost. It has been found in some studies that WCH had similar prognosis as normotension but there is major methodological limitation: information on antihypertensive treatment, which is often started during follow-up on the basis of office BP, is usually insufficient. Actually, the evidence is growing that WCH is associated with increased risk of cardiovascular events as summarized in the meta-analysis [19]. WCH was found to be associated with increased aortic stiffness [20]: it is possible that these individuals have frequent peaks in BP throughout the day, similar to those in the office, and they result in damage to the vascular wall. WCH is associated with an approximately three-fold risk of developing sustained hypertension within 10 years [21], and therefore, its diagnosis (and treatment in individual cases) may be regarded as early prevention of more severe forms of hypertension. The possibility of overtreatment in WCH is often discussed but the inverse problem may exist: in the HYVET study, performed in hypertensive individuals over 80 years, 50% of the relatively small subsample who had ambulatory BP monitoring undertaken had WCH [22]. As active treatment, compared with placebo, lowered several cardiovascular endpoints and total mortality in the whole cohort, we may speculate that individuals with WCH also benefited from the treatment. We would of course need a randomized placebo-controlled trial specially designed for WCH. The study HYVET2, the protocol of which was published, could give some results how to treat elderly individuals with this type of blood pressure [23]. We do not have robust evidence, but subanalyses of other studies yielded some interesting results: it has been found in two randomized trials, Syst-Eur and ELSA, that active treatment given to the individuals with WCH decreased only in-office, but not out-of-office BP [24,25], and therefore, the risk of too low BP does not seem big. It is clear that we need a prospective study, comparing prediction of cardiovascular outcomes by unattended AOBP and attended office BP; the available data are very preliminary [26]. In such a study, unattended AOBP must be measured as strictly unattended, in a separate room, undisturbed by phone, persons passing through, and so forth. Attended office BP must be clearly defined, mainly who measures blood pressure – a physician or another person – and whether the auscultatory or oscillometric method is used. In the case of oscillometric measurement, it must be clear whether the staff member is sitting close to the patient, just as during auscultatory measurement, or is only present in the room: the distance between them is the factor that makes auscultatory measurements higher than oscillometric ones. The staff member must not know the unattended AOBP values before attended office BP is measured and their order must be randomized. Such a study can be done only in specialized hypertension centres where the members of staff are interested in the methods of blood pressure measurement and are able to work under precisely defined conditions. When AOBP is measured in nonspecialized offices, the results may be confusing. A study performed in four general practices (’in a real life setting’) found that mean values of unattended AOBP and office BP measured with the auscultatory method were practically identical (n = 107), and the same applied for unattended and attended AOBPs (n = 51, [27]). This is in contradiction with most studies. Mancia and colleagues published results of a 12-year follow-up of 2051 individuals in the PAMELA study. They found a progressive increase in both cardiovascular and all-cause mortality from individuals in whom office, home, and ambulatory BP were all normal to those in whom one, two, or all three BPs were elevated, regardless of which BP was considered [28]. I believe this is the best way how to assess different methods of BP measurement: each has a prognostic value, and it adds to that of the other BPs. If a high-quality prospective study showed that strictly unattended AOBP carries prognostic information independent of conventional office BP measurement, it could become the fourth BP measurement, which should be taken into account. Indeed, the differences between attended and unattended office BP may be very large: up to +40 mmHg or, paradoxically, −10 mmHg for SBP [8]. Until we have the results of a prospective study, unattended BP measurement should be considered an ancillary method with low level of evidence, which gives us, in association with conventional office BP, clinically interesting information on how office BP changes with and without the member of staff present. This may be helpful especially in patients who refuse or are not able to have out-of-office BP measured, or in patients with high office BP where the differences are the largest. At present, it is appropriate to follow rather liberal recommendations of the 2018 ESC/ESH guidelines for the management of arterial hypertension: we may measure office BP with auscultatory or oscillometric methods, using automated or manual devices; when taking unattended AOBP measurements, we should be aware of limited evidence on its prognostic value [29]. Whichever method is used, the patient should be seated comfortably in a quiet environment for 5 min before beginning the BP measurements. In any case, we should not forget about WCH. ACKNOWLEDGEMENTS Conflicts of interest There are no conflicts of interest.