Heart Rate Variability and Chronic Heart Failure, From Bench to Bedside


For more than two decades short-term power spectral analysis of heart rate variability (HRV) has been used to assess autonomic control of heart rate. Recently it has also been used to stratify the risk of sudden death in subjects with chronic heart failure (CHF). Spectral analysis of RR variability in normotensive healthy subjects shows two distinct components of HRV: high-frequency power (HF),1 which synchronises with breathing activity (from 0.15 to 0.45Hz) and therefore reflects vagal modulation of the sinus node, and lowfrequency power (LF) oscillating around 0.1Hz1 influenced also, though not solely, by sympathetic modulation of the sinus node.1 Recording spectral variables during controlled breathing is useful for three reasons. First, this condition makes it easier to distinguish the two spectral components, HF and LF. Second, when allowed to breathe freely, patients with CHF tend to use a respiratory rhythm that alters the spectral profile. And last, LF values obtained during controlled breathing are predictive of sudden death.2 Among factors that strongly influence autonomic control of the sinus node and HRV are ageing, hypertension and CHF.3-8

Ageing reduces both components of HRV; in fact, sinus node of elderly subjects is less able to modulate heart rate. This decrease is caused by a reduction of vagal activity and by an increase of cathecolamine blood level. Even though all spectral components are reduced, LF component is more affected.3,5 Furthermore, this age-related decline in LF was confirmed in a cohort of elderly subjects observed over 10 years in the Framingham retrospective study.9 The Kaplan- Meier survival curves disclosed a direct relation between tertiles of LF and survival ├óÔé¼ÔÇ£ the tertile with LF <218 ms2 had a significantly higher overall risk of death than subjects in the first (LF >219.9 <415.5ms2) and second tertiles (LFP >415.8ms2).

Hypertension and salt intake induce an increase of LF linked to an increase of sympathetic activity.4,6 In these subjects HF and vagal modulation are reduced. These changes are strictly correlated to the first phases of hypertension ├óÔé¼ÔÇ£ in fact, autonomic control and spectral analysis components are similar in elderly subjects with and without hypertension. Those data suggest that the main influence on the autonomic control is caused by ageing and not by hypertension.

Eventually, CHF causes a marked increase in catecholamine concentrations, accompanied by a reduction in beta adrenergic receptor function, and a marked decrease in baroreflex sensitivity and vagal activity ├óÔé¼ÔÇ£ the more prominent these changes, the higher the risk of sudden death due to the onset of malignant ventricular arrhythmias. In short-term power spectral recordings these conditions reduce both spectral components of HRV markedly.3├óÔé¼ÔÇ£6 Rather than increasing LF of HRV, as one might expect, excessive sympathetic stimulation reduces it. Precisely why LF paradoxically decreased in those patients with CHF is unclear. One explanation might be over-saturation of spectral signal reflecting autonomic control of heart rate. Although spectral analysis provides a qualitative and quantitative assessment of RR oscillations, in the presence of sympathetic overactivity, for example in CHF, the sympathetic spectral component (LF) loses its characteristic oscillations at frequencies around 0.10Hz (Mayer waves) and therefore seems to diminish.

The current need for more information on LF of HRV arises from a prospective study conducted by La Rovere et al2 showing that LF <13ms2 is a risk factor for sudden cardiac death in patients with CHF. LF could become a useful tool to more precisely identify subjects with high risk of sudden death. In fact the data from the MADIT II showed that implanting a cardiac defibrillator in subjects with an ejection fraction of <30% but without recent symptoms of NYHA class IV is advantageous in terms of mortality, hence the severity of symptoms alone is apparently not a reliable criteria for evaluating the effective need to implant a defibrillator. Many investigators nonetheless underline that only some subjects with an ejection fraction <30% really need a defibrillator implanted. Hence, neither is the ejection fraction alone reliable in detecting a real risk of sudden cardiac death.

With regards to vagal modulation of the sinus node in patients with CHF, it seems to be depressed exactly as it is in elderly normotensive and hypertensive subjects. One explanation for this phenomenon is that ageing or the reduced physical activity that often accompanies ageing could diminish vagal modulation of the sinus node as it does in CHF. Because healthy elderly subjects and elderly patients with CHF have similar HF values, HF probably has no prognostic value in elderly persons with CHF. Low HF values might only have prognostic value in younger patients with CHF.

Power spectral analysis of heart rate, born as pure instrument of research, is changing into a clinic tool.