Is the blunted fall in nighttime heart rate a marker of subclinical cardiac damage?

Scanty evidence is available in support of the prognostic value of heart rate (HR) recorded during 24-hour ambulatory blood pressure monitoring (ABPM; ie, average 24-hour daytime and nighttime HR, as well as circadian HR variations) as predictors of target organ damage and cardiovascular events. In this issue of the Journal, Oba and colleagues1 report that a flattened circadian rhythm of HR is a marker of subclinical cardiac damage in patients at high cardiovascular risk, a setting not addressed in previous investigations. Before commenting in detail the strengths and limitations of the contribution by Oba and coworkers, available evidence in this research area may warrant some considerations. Since the pioneering report published by Levy and coworkers2 more than 70 years ago, growing evidence has supported the view that elevated values of HR may be regarded as important determinants of subclinical vascular damage and powerful predictors of cardiovascular morbidity, mortality, and all-cause death. The issue about the value of HR as an unhealthy risk factor useful to refine cardiovascular risk stratification or as a simple indicator of autonomic imbalance remains unsettled due to the complex interplay with a variety of traditional risk factors.3 The largest amount of information on the association of HR with cardiovascular and renal outcomes derives from studies where HR was assessed in resting conditions.4 Emerging although not univocal findings indicate that HR measured outside the medical environment during ABPM, likewise blood pressure (BP), may provide more reliable information than office measurement. Facila and colleagues, examining the relationship between resting and ambulatory HR with target organ damage in 566 patients with arterial hypertension, showed that elevated nocturnal HR values (>65 beats per min) were associated with the presence of organ damage; the same was not true for office and daytime HR.5 Whether HR during sleep is more representative of the hemodynamic load on the vascular system than HR during waking hours or in the doctor's office remains an issue of debate. In a cohort of 580 untreated essential hypertensive patients, we documented that prevalence rates of left ventricular (LV) hypertrophy (LVH), carotid atherosclerosis, and microalbuminuria were not increased across nighttime HR tertiles.6 Of note, average 48-hour, daytime, and nighttime HRs were similar in patients with no organ involvement and in those with one, two, or three target organ signs. Age, 48-hour systolic BP, and metabolic syndrome were independently associated with target organ; this was not the case for HR. The Ambulatory Blood Pressure in Referred Hypertensive Subjects: An International Database (ABP-International) study, a registry established to investigate the relationship of 24-hour ABPM and HR with morbidity and mortality in 7600 participants recruited in different centers from Italy, the United States, Japan, and Australia documented the superiority of ambulatory HR over office HR in predicting cardiovascular events.7 While office HR was no longer a significant predictor of outcome after adjusting for covariates (including systolic and diastolic BP), ambulatory HR turned out to be independently correlated with cardiovascular events; in particular, nighttime HR predicted combined fatal and nonfatal cardiovascular events more closely than 24-hour HR (P=.007 and P=.03, respectively). More limited information is available on the relationship between altered circadian HR rhythm with risk factors, markers of subclinical organ damage, and cardiovascular events. Framme and coworkers8 assessed the hypothesis that the differences from daytime to nighttime systolic BP and HR were smaller in obese (n=55) vs lean (n=55) children and adolescents. Nocturnal dipping in SBP was less pronounced in obese (16±6 mm Hg) compared with lean controls (21±6 mm Hg, P<.01), whereas HR changes from day to night were similar in both groups. A cross-sectional study carried out in 179 high-risk type 2 diabetic patients with proliferative retinopathy showed that nondipping HR (ie, <10% decrement of nighttime HR from daytime values) was independently associated with higher albumin-creatinine ratio, also after correcting for β-blocker use.9 The hypothesis that nondipping HR may represent a novel mechanism contributing to the association of hypertension with subclinical organ damage was previously addressed by our group in the setting of nondiabetic recently diagnosed hypertensive patients. We studied a cohort of 657 untreated essential hypertensive patients characterized by three patterns of day-night HR on the basis of two ABPM sessions carried out within 4 weeks: reproducible dipper (n=513), reproducible nondipper (n=76), and variable dipper (n=69).10 The three groups were not different with regard to age, sex, body size, metabolic variables, office and ambulatory BP, LV mass, common carotid intima-media thickness, carotid plaque, and microalbuminuria. In addition, day-night HR variations treated as continuous variables failed to correlate with all parameters of subclinical organ damage. To the best of our knowledge, the prognostic value of a blunted HR nocturnal dipping has been addressed in only three studies. Verdecchia and colleagues, who investigated this issue in 1942 initially untreated uncomplicated patients with essential hypertension, reported that each 10% lower decrement of HR from day to night entailed a 1.30 relative risk of all-cause death (95% confidence interval, 1.02–1.65; P=.04), after adjustment for confounders including 24-hour systolic BP.11 Also, the extent of HR decrement from day to night had an independent inverse association with total cardiovascular morbidity after adjustment for age, diabetes, and LVH. This association, however, was largely attenuated when average 24-hour systolic BP was added to the model. In keeping with the above-mentioned findings, data collected in a population of 3957 patients referred for ABPM by Ben-Dov and colleagues12 revealed that nocturnal HR decrements from awake values, as well as HR levels during sleep, were independently associated with all-cause death. Of note, mortality analysis according to deciles of the different HR variables showed a robust linear relationship with HR decrements, with a hazard ratio of 2.67 (1.31–5.47) for the lowest vs the highest decile. In a subsample of patients not taking β-blocking treatment, nondipping HR was still associated with a higher mortality rate. Finally, in a sample of uncomplicated 457 asymptomatic hypertensive patients with (n=200) or without (n=257) type 2 diabetes, Eguchi and researchers13 documented that participants with nondipping HR had a 2.4-fold higher risk of cardiovascular events than their counterparts with preserved HR dipping. It should be noted that, at variance from previous cited papers, nondipping HR was not associated with all-cause mortality. The study by Oba and associates,1 conducted in 940 high-risk patients with prevalent hypertension (>90%) enrolled in the Japan Morning Surge Home Blood Pressure (J-HOP) trial, suggests that the association between nondipping HR and increased circulating levels of brain natriuretic peptide (BNP), a validated biomarker of cardiac stress and overload, may contribute to explain the link between alterations in circadian HR profile and adverse cardiovascular prognosis. Nondipping HR phenotype, defined according to current literature criteria, was found in 213 participants (22.6%). These patients, when compared with their counterparts with normal nocturnal HR dip, were older (mean age 67.3±10.4 vs 64.8±11.6 years, P<.001), more frequently treated with β-blockers (33.3% vs 12.1%, P<.001), had lower 24-hour, daytime, and nighttime diastolic BPs and higher daytime HR (73±9 beats per min vs 68±10 beats per min, P<.001). Subclinical target organ was assessed by measuring BNP, LVM indexed to body surface area (LVM index), brachial-ankle pulse wave velocity, and urinary albumin-creatinine ratio. Neither pulse wave velocity nor urinary albumin-creatinine ratio, established indexes, respectively, of large artery stiffness and endothelial dysfunction, were related with nondipping HR pattern. Patients classified as HR nondippers exhibited a nonsignificantly higher LVMI (+4 g/m2, P=.12) compared with dippers. Sex-based analysis revealed substantial sex-related differences in the relationship between heart structure and nondipping HR. Indeed, this pattern was associated with increased LVMI in women (111±32 g/m2 vs 104±27 g/m2, P=.03) but not in men (117±33 g/m2 vs 118±32 g/m2, P=.9). After exclusion of participants treated with β-blockers, however, LVMI difference was no more significant in women (109±27 g/m2 vs 104±27 g/m2, P=.13). Brain natriuretic peptide provides noninvasive information about LV filling pressures and is increasingly used for the diagnosis of LV dysfunction and heart failure.14 In this study, at difference from other markers of target organ damage, BNP was significantly higher in HR nondippers than in dippers (20.5 pg/mL vs 18.3 pg/mL, P<.001). The association between nondipping HR and elevated BNP levels was significant in both sexes. When the analysis was restricted to patients not taking β-blocking treatment, however, the difference in BNP concentrations was markedly attenuated (0.7 pg/mL, P=.046). Altogether, the findings provided by this study in favor of a close association between a flattened circadian HR rhythm and subclinical organ damage in high-risk patients are weak and not univocal. Crude pulse wave velocity, urinary albumin-creatinine ratio, and adjusted LVMI were similar in the two groups, and only log-transformed BNP was independently correlated with nondipping HR phenotype in multiple regression analysis. The strengths of the study by Oba and colleagues1 are the rigorous method of investigation and the appropriate statistical analysis of an issue largely unexplored. The hypothesis that a sympathetic overactive state, as reflected by blunted HR dipping during sleep, may have adverse effects on the cardiovascular system deserves to be further investigated. Larger studies in different clinical settings and more sophisticated investigations on cardiac damage using innovative imaging diagnostic tools such as myocardial tissue Doppler and two- or three-dimensional strain analysis detecting early changes in myocardial mechanics15 will provide more compressive information on this issue and provide indications for therapeutic interventions. The authors report no conflicts of interest.