Unattended Office Heart Rate Measurement: A New Challenge in Clinical Practice?

In recent times, a large body of evidence has shown that about one-third of patients with hypertension have persistent tachycardia and that a high resting heart rate (RHR), independent of the elevated blood pressure (BP), is a potent additional risk factor for cardiovascular disease and mortality [1-4]. The pathophysiology of the detrimental effects of fast RHR has been well explored and documented by numerous experimental studies [5]. For this reason, RHR has been included among the cardiovascular risk factors in the latest Guidelines of the European Society of Hypertension, which recommend always measuring also RHR when evaluating BP levels [6]. However, many sources of variability, including psychic stimuli, environmental factors, and body position, may affect the assessment of heart rate in resting conditions [7-9]. To minimize the effect of these variables, a consensus document of the European Society of Hypertension recommends that the measurement of RHR in the doctor's office should be strictly standardized [9]. However, also in standardized conditions, both RHR and BP are influenced by the presence of healthcare personnel, and office measurements often overestimate the usual level of these hemodynamic variables as a consequence of the so-called white-coat effect [7, 8, 10]. To overcome this drawback, strategies for assessing RHR and BP out of the office in the absence of the doctor have been devised, which led to a wide use of self-measurement and of ambulatory monitoring techniques, which have been shown to provide prognostic information over and above office measurement [6]. Available data suggest that alike BP, RHR measured out of the office yields more meaningful clinical information than RHR measured by healthcare personnel [11].

The advent of oscillometric sphygmomanometry as a replacement for the auscultatory measurement led to an improvement of office BP measurement, eliminating some errors related to the observer and allowing the recording of multiple readings automatically [12]. However, in clinical practice, oscillometric measurements do not differ substantially from manual measurements as long as the medical staff remain in close proximity to the patient and thus cannot avoid the white-coat effect. A step forward with oscillometric BP measurement was taken with the introduction of devices capable of recording multiple BP and RHR readings automatically without the need to have a nurse or a doctor present during the measurements [13, 14]. The effects of removing the healthcare personnel with the associated reduction of anxiety have been well documented by several studies that showed that routine office BP is substantially higher than unattended BP and similar to awake ambulatory BP [13, 14]. In contrast, little is known on the comparability of unattended office RHR measurement with conventional RHR measurement and heart rate measured out of the office. The article by Sobieraj and Coll. published in this issue of JCH sheds light on this unexplored issue [15]. To assess the agreement between RHR during unattended measurement and other methods of measurement, the authors conducted a comparability study in a group of 110 participants referred for ambulatory BP monitoring. The variables investigated as comparators were office RHR, ambulatory heart rate, and RHR recorded with electrocardiography. Like the results obtained for BP measurements, RHR measured with the unattended modality (70.8 bpm) was significantly lower than oscillometric RHR measured by the medical staff (72.8 bpm) and similar to average daytime ambulatory heart rate (70.3 bpm). A slightly lower value was found for RHR measured with the electrocardiogram (69.1 bpm), but the difference was not statistically significant. The Bland-Altman plot showed that unattended RHR was 2.0 bpm lower than standard office RHR with a 95% confidence interval of −2.8 to −1.3 bpm. These data confirm that the measurement of hemodynamic variables obtained in the office in the absence of the observer may avoid at least partially the alerting reaction associated with the presence of the medical staff. This phenomenon was described long ago with the use of 24-h ambulatory monitoring, which documented that the doctor's visit is accompanied by an immediate rise in the patient's heart rate [7, 8]. This reaction, however, showed large differences between individuals and may differ according to BP status. In a study performed in our laboratory [10], heart rate increased by 4.6 ± 4.5 bpm during the doctor's visit in a group of hypertensive subjects, while it remained virtually unchanged in a normotensive group of subjects of control (0.3 ± 3.8 bpm, p < 0.001 versus hypertensives).

As mentioned above, a wealth of studies have documented that a high office RHR is associated with an increased risk of cardiovascular disease and mortality [1-4]. However, in some individuals, tachycardia may be a transient phenomenon related to the anxiety elicited by the observer, which can be avoided with out-of-office measurement. Another advantage of out-of-office over office RHR is that reproducibility is better for the former, as shown by the results of the HARVEST study [16]. These characteristics may account for the better prognostic value shown by ambulatory heart rate compared to office RHR in 7600 hypertensive patients from the ABP-International study followed for 5 years [11]. In a multivariable Cox model, 24-h and night-time heart rates emerged as the strongest predictors of fatal combined with nonfatal events with a hazard ratio of 1.11 (p = 0.031) and 1.13 (p = 0.007), respectively, for a 10-bpm increment of the ambulatory heart rates. In this study, office RHR was a weaker predictor of outcome than ambulatory heart rate and was no longer associated with the outcome after inclusion of systolic and diastolic BPs in the survival model. Similar results were obtained in the Copenhagen Holter study and other investigations after accounting for cardiovascular risk factors and other confounders [17]. Given the advantages of unattended over attended RHR, one would expect that also unattended RHR has better prognostic accuracy than office RHR, but unfortunately no information is available on the relationship between unattended RHR and adverse cardiovascular outcomes in comparison with office RHR.

Another limitation to the use of unattended RHR in clinical practice is the lack of a precise cutoff to distinguish between normal and high RHR. The same problem has been encountered with standard office RHR and has been the subject of much debate in the literature [18]. Most studies found a significant increase in risk for an office heart rate ≥ 80–85 bpm, which roughly corresponded to the upper quintile of the RHR distribution. However, as pointed out by Sobieraj and Coll in their article, also lower threshold levels were identified as limits of normality [15]. For example, the authors of the INVEST trial found an increased cardiovascular risk at RHR > 75 bpm [19]. However, it should be pointed out that the relationship between RHR and cardiovascular risk is a continuous one, and thus the adoption of a precise cutoff of normality is arbitrary. On the basis of an analysis performed within the frame of the SPRINT study, Sobieraj et al. claimed that an increase in cardiovascular risk was present for an unattended RHR > 70 bpm, which could thus be considered as a threshold for tachycardia [20]. However, Figure 3 of their article shows that in multivariable Cox models for the clinical composite endpoint event, there was a progressive increase in risk with increasing unattended RHR in both people with and without previous cardiovascular events. Thus, also for unattended RHR, the choice of a cutoff appears to be arbitrary. When studying the agreement of unattended office tachycardia with the other measurement modalities, the authors used both the 70 and the 80 bpm thresholds, which is a widely accepted level for RHR in clinical practice [18]. Although according to Kappa statistics the agreement between the attended and unattended RHR measurements was good, Cohen's kappa coefficient was < 0.80 for both RHR cutoffs [15], indicating that in some participants attended office tachycardia could not be confirmed by the unattended measurement. The better agreement of unattended RHR with daytime heart rate in the Sobieraj et al. study suggests that in some individuals office tachycardia may be due to an exaggerated alerting reaction at the time of RHR measurement.

There is no doubt that the increase in RHR that occurs during attended oscillometric measurement is the consequence of an alerting response of the patient, as documented by Grassi and Coll in an elegant laboratory experiment [21]. These authors showed that there is a concomitant increase in RHR and sympathetic nerve traffic to the skin vascular district during the attended measurement and that both increases can be offset when the measurement is made in the absence of the medical staff. However, until now, a comprehensive evaluation on the comparability of unattended RHR measurement with conventional office RHR and other measurement modalities was not available. The study by Sobieraj et al. has filled this gap by providing comparative values between these methods of measurement [15]. If unattended RHR data are to be used in clinical practice, measurements must closely adhere to the procedures recommended by scientific societies for the measurement of unattended BP, including multiple readings obtained with a fully automated oscillometric device while the subject rests alone in a quiet environment [6]. Although unattended RHR measurement may be helpful for identifying people with true tachycardia, the method has the same limitations as those reported for BP measurement [6]. The unattended modality requires appropriate spaces, proper instrumentation, and dedicated healthcare personnel, factors that limit its application not only in routine clinical practice but even in hypertension clinics. Normal RHR values are not available for this technique and are difficult to identify. According to Sobieraj and Coll, unattended RHR data should be the same as for daytime heart rate, but also for ambulatory heart rates there is much controversy about the upper limits of normality [22]. Finally, no evidence exists on the ability of unattended RHR to predict adverse cardiovascular outcomes in the general population, as the only available data have been obtained in the hypertensive population of the SPRINT trial [20].

The author declares no conflicts of interest.