Resting Heart Rate “Kinetics”: Paradox or Means to Understand the Link Between Hypertension and Heart Failure?

High heart rate (HR) is associated with poor outcome in the general population1,2 and patients with cardiovascular diseases including coronary artery disease3 and heart failure (HF).4 In patients with HF with reduced left ventricular ejection fraction (LVEF), pharmacological HR reduction is beneficial, and HR is now regarded as a therapeutic target.5 However, HR is under the control of the autonomic nervous system and has a substantial diurnal variability depending not only on physical activity but also mental stress. This may be important because in contrast to the measurement of blood pressure, HR assessment is not well standardized. In practice, HR is typically measured in 2 ways: first, by the electrocardiogram (ECG), i.e., in the supine position and second, during measurement of blood pressure in the doctor’s office, i.e., in the sitting position. It is likely that for a given patient, HR measured in the supine and sitting position will differ since HR depends on posture. There is a physiological baroreflex-mediated rise in HR after getting up from the supine to the upright position. This HR rise may be blunted in the presence of diseases affecting the autonomic nervous system such as diabetes. This observation represents the basis for the use of HR as a test for the diagnosis of diabetic neuropathy. In this setting, the following parameters are of interest: (i) the decrease in HR within the first seconds after the baroreceptor-mediated rapid rise following getting up and (ii) the net difference between HR in the supine position and the stable HR plateau approximately 30 seconds following getting up.6 A similar test is not broadly established in patients with cardiovascular disease although autonomic dysfunction is common in these patients, particularly in those with HF. Only recently, our research group has shown that in patients with chronic HF, a small rise in HR after getting up from the supine to the upright position (ΔHR = HR in upright position − HR in supine position; cut-off 3 bpm) was an independent predictor of death or HF hospitalization.7 In the present issue of the Journal, Oba et al.8 present their data on a modified variant of ΔHR in the Japan Morning Surge-Home Blood Pressure study (J-HOP) population, a large population of asymptomatic subjects with at least 1 cardiovascular risk factor (90% with hypertension), i.e., a population at risk for the development of HF. They studied 2,972 patients in sinus rhythm, without β-blocker therapy and with an HR reading derived from an ECG (ECG HR) and an HR reading measured during blood pressure recording in the sitting position (office HR), and ΔHR was calculated as office HR − ECG HR. As expected, office HR was overall higher than ECG HR, and a lower ΔHR was associated with higher ECG HR and lower office HR. In addition, higher ΔHR was related to lower age, lower systolic blood pressure, higher diastolic blood pressure, and lower pulse pressure, while other patient characteristics including body mass index and diabetes prevalence were similar across ΔHR quintiles. In this study, no information on the association between ΔHR and clinical outcomes was presented, but B-type natriuretic peptide (BNP; available in all 2,972 patients) and left ventricular mass index (LVMI; available in 1,061 patients) were used as surrogate markers. The main finding was an inverse association between lower ΔHR and higher LVMI and higher BNP, and these associations remained statistically significant after adjusting for important covariates. To interpret this result, we must first answer the question whether these surrogate markers are relevant outcome variables. In patients with hypertension, LVMI predicts the occurrence of cardiovascular events9 and is regarded as an important subclinical target organ damage marker.10 In the general population, LVMI is an independent predictor of the future occurrence of HF, even in subjects with normal baseline LVEF and without a history of coronary artery disease.11 Importantly, left ventricular hypertrophy is not always attributable to hypertension alone, i.e., there is a load-dependent and load-independent left ventricular hypertrophy.12 In a landmark study by de Simone,11 both LVMI and excess left ventricular mass (defined as the amount of left ventricular mass not explained by age, gender, and stroke work) were found to predict the incidence of HF. BNP and the N-terminal part of its precursor peptide NT-proBNP are established predictors of mortality not only in patients with HF but also in the general population.13 At least 2 studies have also shown that NT-proBNP was associated with all-cause mortality in patients with hypertension.14,15 Importantly, LVMI and natriuretic peptides provide independent prognostic information. In the Multi-Ethnic Study of Atherosclerosis (MESA) with approximately 35% of subjects taking antihypertensive medication, patients with both left ventricular hypertrophy and elevated NT-proBNP or cardiac troponin had a higher risk of the occurrence of asymptomatic left ventricular dysfunction (reduced LVEF), incident HF, and cardiovascular death than those with left ventricular hypertrophy but without elevated biomarker level.16 Finally, guidelines highlight the importance of LVMI and BNP for the diagnosis of HF with preserved ejection fraction,5 a disease often caused by long-standing hypertension. Therefore, the association between lower ΔHR and higher LVMI and higher BNP as observed by Oba et al.8 is relevant, and we have to ask about possible mechanisms. A small ΔHR may be a simplified measure of HR variability and may indicate autonomic dysfunction in the context of more advanced subclinical cardiac dysfunction as reflected by higher LVMI and higher BNP, which would be in line with our findings on the prognostic impact of ΔHR in HF.7 Higher LVMI may be a surrogate for more advanced left ventricular systolic and diastolic dysfunction. However, this remains speculative since these parameters were not reported. Some aspects of the present study are not easy to understand, however, and may point to a higher complexity. Mathematically, a low or negative ΔHR is the result of a relatively high ECG HR and a relatively low office HR. Somewhat counterintuitively, both a higher ECG HR and a higher office HR were associated with lower LVMI and BNP. Interestingly, this is in line with recent findings from the MESA population where ECG HR predicted a deterioration of left ventricular strain and LVEF over a period of 5 years and incident HF independent of hypertension but where a higher ECG HR was also associated with lower LVMI.2 Oba et al.8 have suggested that a low HR may lead to higher LVMI via increased stroke volume and left ventricular end-diastolic pressure but that a low HR may be still vasculoprotective (lower sympathetic tone, less atherosclerosis) and via this mechanism (rather than via LVMI) may protect from left ventricular dysfunction and HF. This argumentation could at least in part also explain the inverse association between lower ECG HR and higher BNP. Given that left ventricular wall stress is thought to be the most important determinant of BNP release,17 the higher BNP with lower HR (and higher LVMI) based on the Laplace law (wall stress ~ pressure × radius/wall thickness) can be explained by a nonproportional increase in left ventricular pressure (pressure↑↑, wall thickness↑) or an eccentric left ventricular remodeling process (pressure and wall thickness unchanged, radius↑) or a combination of these mechanisms. To understand the whole picture, we must try to interpret ΔHR in the context of ECG HR (and office HR). In Figure 1, the relationship between ECG HR, ΔHR, and office HR with LVMI and BNP as semiquantitatively estimated from the data reported by Oba et al.8 is shown. It becomes clear that the correlation of certain ΔHR with LVMI and BNP differs if ECG HR differs. The subjects A and C and the subjects B and D have the same ΔHR but different ECG HR (and office HR) and thereby different LVMI and BNP given the inverse relationship between ECG HR (and office HR) with BNP and LVMI. Schematic representation of the key data by Oba et al.8 The scale on the left indicates the units for heart rate (HR), and the scale on the right indicates the units for left ventricular mass index (LVMI) and B-type natriuretic peptide (BNP). ΔHR = office HR − ECG HR. For discussion see text. While interpreting the intriguing findings reported by Oba et al.,8 we must realize that the study has a number of limitations: HR ECG and office HR were not measured at the same time but with an interval of up to 3 months. A variety of factors may have contributed to changes in HR during this period including lifestyle changes, incurrent diseases, and drugs. However, echocardiography and venipuncture were performed on the same day as the ECG was performed, and thus, the associations between ECG HR and LVMI and BNP, respectively, are reliable. The amount of echo data is very limited. Interpretation of the role of LVMI would have been facilitated if additional parameters of cardiac function such as LVEF, LV strain, LV diastolic function, and left atrial size had been available. And finally, the reported associations were relatively weak, and a number of unmeasured confounders may have played a role. Measurement of ΔHR is appealing because in contrast to LVMI and BNP assessment, it does not require an ultrasound machine or a laboratory but can be performed in the doctor’s office. The important work by Oba et al.8 has generated an interesting hypothesis. However, our attempts to interpret their findings were not conclusive in all aspects. Much work will be needed to understand ΔHR and to make it ready for clinical use. The authors declared no conflict of interest.