Impairment in Cardiac Autonomic Regulation Preceding Arterial Hypertension in Humans
Insights From Spectral Analysis of Beat-by-Beat Cardiovascular Variability
Background— Subjects in the upper-normal range of arterial pressure have an excess cardiovascular risk, which suggests that other factors, such as impaired autonomic regulation, might be implicated. This study was designed to assess whether noninvasive markers of cardiac and vascular autonomic regulation might already be altered in subjects with high-normal arterial pressure levels.
Methods and Results— We performed an observational study on a population comprising 300 subjects of both sexes with arterial pressure ranging from 90/60 to 210/120 mm Hg, who were divided into 3 groups (each n=100) with average systolic pressures of 103, 133, and 163 mm Hg. Autonomic regulation was inferred from spectral analysis of RR interval and systolic arterial pressure variability, considering rest and stand-induced changes, to account for sympathetic excitatory components. Significant alterations in markers of sinoatrial regulation (increased low-frequency normalized units, reduced high-frequency normalized units, and α-index) were already apparent in subjects in the second tertile, ie, those with arterial pressure within normal limits. Markers of vascular regulation instead showed significant alterations in the third (hypertensive) tertile. In response to standing, changes in markers of sinoatrial modulations were gradually reduced, whereas those of vascular regulation were increased. A tight link between progression of arterial pressure and the continuum of changes in autonomic markers as shown by simple correlation analysis appeared strongly affected by age and was spread across many spectral analysis–derived variables. Hypertensive autonomic dysregulation was particularly apparent in the youngest group.
Conclusions— RR-variability parameters might prove useful to assess, with longitudinal studies, the mechanistic role of autonomic impairment in the increased risk of prehypertensive conditions.
Received July 30, 2002; accepted September 9, 2002.
Expert national committees recommend that subjects with established hypertension, defined as arterial pressure values exceeding 140 mm Hg systolic and 90 mm Hg diastolic, should receive antihypertensive treatment to reduce their increased cardiovascular risk.1 Despite the fact that in population studies, even small reductions in arterial pressure load might appear useful, the clinical benefits of lowering arterial pressure are far from optimal, particularly with regard to coronary risk.2 Cardiovascular events such as sudden coronary death, myocardial infarction, and stroke might easily occur at pressure values below this threshold,3 and antihypertensive treatment might prove valuable, at least in selected cases, even in patients who do not reach the hypertensive limit.4 Accordingly, the usual dichotomous approach to hypertension, distinguishing between normotensive and hypertensive subjects, might be suboptimal. Limited compliance with individual variations, such as spontaneous or seasonal shifts in arterial and autonomic baseline values,5 the “white coat” effect,6 or measurement bias,7 would be particularly confounding in subjects with borderline values, contributing to uncertain diagnosis.8 Most recent guidelines suggest further dividing the pressure range,9 focusing also on high-normal values. The demonstration that subjects in the upper-normal range of pressure (notably, those with systolic arterial values between 130 and 135 mm Hg10) show an excess event rate argues in favor of a linear relationship11 between arterial pressure and cardiovascular risk, which supports the use of continuum models. Moreover, risk should be considered in the context of its multifactorial nature, which derives from interaction with other unfavorable genetic, behavioral, or metabolic elements, such as smoking, stress,12 sedentary lifestyle, dyslipidemia, or endothelial dysfunction.11 Among possible components of increased risk, studies showing the presence of several anomalies in autonomic markers such as reduced heart rate variability,13 decreased baroreflex sensitivity,14,15⇓ increased sympathetic activity,16,17⇓ and altered autonomic responsiveness18 in arterial hypertension support a mechanistic role of excitatory sympathetic circuits.19,20⇓ The present observational study was therefore designed to assess whether noninvasive markers of cardiac and vascular autonomic regulation, viewed as part of a continuum, might already be altered in subjects with high-normal arterial pressure levels.
This study considered 300 nonobese (body mass index <26) subjects (169 men, 131 women) whose sitting systolic arterial pressure (SAP) ranged from 90 to 210 mm Hg and whose diastolic arterial pressure ranged from 60 to 120 mm Hg. All subjects were nonsmokers and free of any disease or disturbance (except arterial hypertension), as determined by history, physical examination, and routine laboratory tests. Left ventricular hypertrophy was ruled out by ECG criteria. Half of the study population had been referred because of clinical diagnosis of arterial hypertension, whereas the second half was recruited from healthy hospital staff, medical students, and relatives. Subjects were instructed to avoid alcohol and caffeinated beverages for the 12 hours preceding the study, to abstain from heavy physical activity the day before, and after a light breakfast, to come to the laboratory between 8:30 am and 12:30 pm on the day of the study. All subjects were instructed about the study procedure and gave their informed consent. None were taking any medication (treated hypertensives had discontinued their usual therapy for at least 4 weeks). Our institutional review board approved the study protocol.
After a preliminary 10-minute period, to allow for stabilization, a 10-minute rest period was obtained in all subjects, followed by a 7-minute period of active standing. During the last 2 periods, recordings were performed. The ECG (CM5) and respiratory signal were recorded with a 2-way radiotelemetry system (Marazza), whereas the arterial pressure waveform was continuously assessed noninvasively with a plethysmographic device (Finapres, Ohmeda), the average error of which was estimated to be ≈2.6±0.4 mm Hg, according to Bland and Altman,7 with a standard sphygmomanometer. Data were acquired with a personal computer equipped with an analogue-to-digital board (Data Translation) with an acquisition rate of 300 samples per channel per second.
From the simultaneous autoregressive spectral analysis of RR interval and SAP variability, a series of autonomic indexes was subsequently computed offline, as described previously,21–23⇓⇓ which provided markers of autonomic regulation of the sinoatrial (SA) node, the arterial vasculature, and overall baroreflex gain (α-index).15 In all subjects included in the study, respiratory rate coincided with the high-frequency (HF) component of RR variability and is therefore presented in Tables 1 and 2⇓⇓⇓.
Data are presented as mean±SEM (except age, which is shown as mean and range, Table 1⇑). The study population was divided into 3 groups of equal size (n=100) according to ranks of SAP, as measured by Finapres, to focus on the high-normal range of pressure. One-way ANOVA was used to assess the significance of differences between the 3 groups (Tables 1 and 2⇑⇑⇑), whereas paired t test was used to assess the effects of stand-induced changes (stand-rest). Simple and partial correlation analyses were used to estimate the relationship between SAP and time- and frequency-domain indices of cardiovascular variability, with further control for age and sex (Table 3). To demonstrate graphically the simultaneous influence of age, as in Figure 3, subjects were further divided into 3 age groups, and cells presented in every panel of the illustration provide data that refer to the youngest and eldest tertiles. (The middle age tertile is not included for clarity.) A P level <0.05 was considered significant.
An example of the spectral profile of RR variability and respiration in a normotensive and a hypertensive subject is shown in Figure 1. Notice in the latter case the clear shift in spectral distribution toward the LF component at rest and its diminished increase with standing up. Respiration, in both cases, clearly coincides with HF range.
Descriptive statistics at rest are presented in Table 1⇑. Average SAP (by Finapres) was respectively 103±1, 133±1, and 163±1 mm Hg in the 3 groups. With increasing pressure level, from tertile I to tertiles II and III, resting values of RR interval, RR variance, and both LFRR and HFRR in absolute units decreased significantly. Likewise, the center frequency of the LFRR component and the baricenter of the whole spectrum24 decreased significantly. With regard to components, LFRR in normalized units and LF/HF ratio in tertiles II and III were significantly higher than in tertile I, whereas HFRR in tertiles II and III was significantly lower than in tertile I. These changes overall indicate that with rising pressure groups, RR spectral power distribution shifts toward the LF end of the spectrum. Differences in markers of SA node modulation, slightly affected by age and sex, were greater between tertiles I and II than between tertiles II and III. Conversely, differences in autonomic indicators derived from arterial pressure, such as SAP variance and the LFSAP component, which increased with rising pressure levels, became more pronounced in tertile III. The α-index, a measure of baroreflex gain,15 was progressively reduced with rising pressure level across all tertiles. In brief, alterations in autonomic markers, particularly those that reflect SA node modulation, were already well apparent in tertile II, at a systolic arterial pressure well below the usual upper limit of normal (ie, 140 mm Hg), whereas changes in markers of sympathetic modulation of arterial pressure reached significance only in tertile III, well above the hypertensive limit.
The extent of changes induced by standing in heart rate variability parameters (Table 2⇑) was progressively reduced with rising pressure tertile, with these variations particularly evident for LFRR in normalized units and for HFRR in both normalized and absolute units. The standing-induced reduction in α-index was progressively smaller with rising pressure groups. Standing-induced changes in SAP-derived parameters (such as SAP variance and LFSAP component), conversely, increased with growing pressure tertile. Overall, responses to standing in indices of SA node and SAP regulation appeared divergent: whereas with growing pressure tertile, markers of RR modulation changed less with active orthostatism, the reverse occurred with markers of SAP modulation, which increased more.
Correlation Between SAP and Spectral Variables
To further characterize the continuum of changes, individual simple correlations between SAP and spectral analysis–derived indices were computed (see Table 3). The majority of autonomic indices, considering both their resting values and changes induced by stand, were highly significantly linked with SAP, albeit, not surprisingly,18 with moderate values of correlation (Figure 2 illustrates highest-ranking variables and the wide scatter of individual data points). Correlations were also controlled for age and for sex25 because of the strong link of age with SAP (r=0.528, P<0.001) and because of the reported possible influence of sex on autonomic dysregulation in hypertension.25 Many of the markers of SA modulation, such as RR variance, absolute LF RR, and HF RR, were thus no longer significantly linked to the continuum of SAP. Conversely, for others, such as the α-index and stand-induced changes in RR interval, LFRR (expressed in normalized units), the baricenter of the RR-interval spectrum, and LFSAP, the partial correlation, although indicating a clear influence of age but a minimal influence of sex, signaled an independent, highly significant link with SAP. After controlling for age and sex, resting RR interval, frequency of LFRR, and LFSAP power had a weakly significant link with SAP.
The combined effects of age and arterial pressure tertiles on most significant cardiac and vascular autonomic variables are depicted in Figure 3, which shows that a pressure-linked, age-dependent progressive derangement of SA node regulation was more apparent in the youngest tertile. Changes in autonomic markers of vascular regulation (eg, variance SAP) were more apparent in tertile III and were less sensitive to age.
This study provides new evidence in human hypertension of a progressive impairment in autonomic regulation that is already present in subjects with arterial pressure values below the upper limit of normalcy and is more apparent on markers of SA node regulation. The link between progression of arterial pressure and the continuum of changes in autonomic markers appears strongly affected by age and spreads across many variables, none of which can be reliably singled out as the most significant indicator. Impaired autonomic regulation of the SA node is particularly apparent in the youngest group.
Impaired autonomic regulation was assessed with a noninvasive approach, based on multiparametric evaluation, that considered indices derived from time- and frequency-domain analysis of RR and SAP beat-by-beat variability21–23⇓⇓ to simultaneously address multiple aspects of information26 about cardiac and vascular control. Notably, it must be stressed that because none of the techniques used in the present clinical study can provide a direct measure of autonomic nerve activity27 to the heart or blood vessels, only indirect inferences can be drawn from the present data, based on probability and error levels28 and on interpretative models,26 as shown by a large body of experimental and clinical evidence.23 The use of dynamic protocols26 that include, in addition to resting values, changes produced by sympathetic activation, eg, through standing-induced baroreceptor unloading29 or moderate exercise,15 might reveal autonomic disturbances that are not otherwise apparent.
The significant alterations in RR-variability parameters and in α-index values (a measure of baroreflex gain15) present at rest in the still normotensive tertile II subjects suggest a change in the interaction between the 2 autonomic arms that modulate the SA node toward sympathetic predominance and vagal withdrawal, which occurs at arterial pressure values below the upper limit of normalcy. This finding should be contrasted with the observation that a significant elevation in markers of vascular modulation (such as increased SAP variance and LFSAP) is only present in the third pressure tertile. Thus, although it is likely that LF oscillations in RR and SAP might both reflect similar oscillations in direct microneurographic recordings of efferent muscle sympathetic nerve activity, as suggested by studies using coherence analysis,27 it is possible that altered autonomic mechanisms might differently affect the SA node and the vasculature. In this context, a role for different end-organ responses, in addition to a significant modulation of age and, to a lesser extent, of sex,25 is suggested by prior studies12 on healthy students exposed to real-life long-term stress that induced diminished responses of the SA node and increased the response of the vasculature to stand-mediated sympathetic excitation.
The importance of a continuum of changes11 is emphasized by the significant correlation between arterial pressure and the majority of markers of RR and SAP variability, which also shows the large spread of information across RR-variability parameters, as previously reported for normotensive subjects.30 The use of partial correlations to control for the well-known effects of aging on autonomic regulation in hypertension25 leads to a loss of significant links between pressure levels and absolute, but not normalized, power of markers of autonomic regulation of the SA node. This finding further supports the different information imparted by normalized or absolute markers of autonomic oscillations.21–24⇓⇓⇓ Considering present data in the context of the different information26 provided by absolute and normalized powers of RR-variability components, we might hypothesize a clear and early disturbance of oscillatory properties of RR variability in human hypertension, already evident in the prehypertensive31 pressure levels of tertile II. This disturbance appears to be characterized by a shift in spectral power distribution toward the lower end of the frequency axis, as shown by normalized units (but not by absolute units, owing to the simultaneous reduction in variance), and further documented by a lower center frequency of LFRR and of the baricenter of the RR autospectrum.24 Although we did not specifically address the issue of sex influences,25 it appears from the present data that controlling in addition for sex might slightly modify our findings, if at all.
In conclusion, the present investigation revealed that in subjects with arterial pressure values in the upper normotensive range, noninvasive markers of cardiac and vascular regulation were already altered. These alterations, which were particularly apparent in the youngest group, suggest that disturbances in autonomic regulation that precede overt arterial hypertension might explain in part the increased risk of cardiovascular events even in hypertensive patients with optimal blood pressure control.
Although the cross-sectional design is an important intrinsic limitation, the present study suggests that RR-variability parameters, as in the case of ischemic heart disease32 and heart failure,33 might provide a useful functional outcome34 to assess the efficacy of integrated management35 of the increased risk of prehypertensive conditions.10 Accordingly, hypertension treatment should not be limited to achieving a reduction in blood pressure but should also aim to treat the mechanisms underlying high blood pressure states.
Supported in part by Agenzia Spaziale Italiana–COFIN 2000/2001 (to Dr Pagani). The typing assistance of Giovanna Macciò is gratefully acknowledged.
- ↵Oparil S. Arterial hypertension. In: Bennet CJ, Plum F, eds. Cecil - Textbook of Medicine. Philadelphia, Pa: WB Saunders; 1999: 256–271.
- ↵van de Borne P, Montano N, Zimmerman BG, et al. Relationship between repeated measures of hemodynamics, muscle sympathetic nerve activity and the spectral oscillations. Circulation. 1997; 96: 4326–4332.
- ↵Julius S, Mejla A, Jones K, et al. “White coat” versus “sustained” borderline hypertension in Tecumseh, Michigan. Hypertension. 1990; 16: 617–623.
- ↵Lucini D, Norbiato G, Clerici M, et al. Hemodynamic and autonomic adjustments to real life stress conditions in humans. Hypertension. 2002; 39: 184–188.
- ↵Singh JP, Larson MG, Tsuji H, et al. Reduced heart rate variability and new-onset hypertension: insights into pathogenesis of hypertension: the Framingham Heart Study. Hypertension. 1998; 32: 293–297.
- ↵Bristow JD, Honour AJ, Pickering GW, et al. Diminished baroreflex sensitivity in high blood pressure. Circulation. 1969; 39: 48–54.
- ↵Pagani M, Somers VK, Furlan R, et al. Changes in autonomic regulation induced by physical training in mild hypertension. Hypertension. 1988; 12: 600–610.
- ↵Greenwood JP, Stoker JB, Mary D. Single-unit sympathetic discharge: quantitative assessment in human hypertensive disease. Circulation. 1999; 100: 1305–1310.
- ↵Anderson EA, Sinkey CA, Lawton WJ, et al. Elevated sympathetic nerve activity in borderline hypertensive humans: evidence from direct intraneural recordings. Hypertension. 1989; 14: 177–183.
- ↵Malliani A, Montano N. Emerging excitatory role of cardiovascular sympathetic afferents in pathophysiological conditions. Hypertension. 2002; 39: 63–68.
- ↵Pagani M, Lombardi F, Guzzetti S, et al. Power spectral analysis of heart rate and arterial pressure variabilities as a marker of sympathovagal interaction in man and conscious dog. Circ Res. 1986; 58: 178–193.
- ↵Malliani A, Pagani M, Lombardi F, et al. Cardiovascular neural regulation explored in the frequency domain. Circulation. 1991; 84: 482–492.
- ↵Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Heart rate variability: standards of measurement, physiological interpretation, and clinical use. Circulation. 1996; 93: 1043–1065.
- ↵Jasson S, Medigue C, Maison-Blanche P, et al. Instant power spectrum analysis of heart rate variability during orthostatic tilt using a time-/frequency-domain method. Circulation. 1997; 96: 3521–3526.
- ↵Barnett S, Morin R, Kiely D, et al. Effects of age and gender on autonomic control of blood pressure dynamics. Hypertension. 1999; 33: 1195–1200.
- ↵Pagani M, Montano N, Porta A, et al. Relationship between spectral components of cardiovascular variabilities and direct measures of muscle sympathetic nerve activity in humans. Circulation. 1997; 95: 1441–1448.
- ↵Haken H. Synergetics: An Introduction. Nonequilibrium Phase Transitions and Self-Organization in Physics, Chemistry and Biology. 3rd ed. Berlin, Germany: Springer Verlag; 1983.
- ↵Malliani A, Pagani M, Furlan R, et al. Individual recognition by heart rate variability of two different autonomic profiles related to posture. Circulation. 1997; 96: 4143–4145.
- ↵Tank J, Jordan J, Diedrich AM, et al. Genetic influences on baroreflex function in normal twins. Hypertension. 2001; 37: 907–910.
- ↵Galinier M, Pathak A, Fourcade J, et al. Depressed low frequency power of heart rate variability as an independent predictor of sudden death in chronic heart failure. Eur Heart J. 2000; 21: 475–482.
- ↵Clancy CM, Eisenberg JM. Outcomes research: measuring the end results of health care. Science. 1998; 282: 245–246.