Single‐minded about heart failure

Since the mid 1960's, when Karl-Erik Hagbarth and Åke Vallbo first inserted tungsten microelectrodes into their own nerves and heard over the loudspeakers the sound of ‘waves approaching the distant shore’ the identity of these waves as bursts of activity generated by sympathetic postganglionic axons has spawned thousands of studies that have taken advantage of this capacity to record sympathetic nerve activity directly in awake human subjects. Yet because the axons are unmyelinated, and the action potentials small, most intraneural recordings of human sympathetic nerve activity are multi-unit, and most investigators simply count the number of visibly identified bursts per minute (burst frequency) or per 100 heart beats (burst incidence) from the integrated or RMS-processed nerve signal. Indeed, many investigators do not bother with the underlying neural signal, being concerned only with the mean voltage record. Accordingly, the signal content within such a recording is rather crude. And this is why single-unit recordings from sympathetic neurones were developed: like unitary recordings from myelinated sensory axons (which can also be made through these same microelectrodes), they provide more information than a multi-unit recording. There is no doubt that the approach is technically demanding – after all, we are trying to record action potentials generated by individual C-fibres, which requires that the spikes be of sufficient (and invariant) amplitude that they stick out above the noise. The first single-unit recordings from type-identified sympathetic neurones were made by Macefield, Vallbo & Wallin in 1994, who reported the firing properties of individual muscle vasoconstrictor neurones, and all subsequent studies have reported the firing properties of identified neurones. Importantly, we now have a schema for analysing the firing properties of these neurones, in terms of (i) the mean firing frequency, (ii) the firing probability (the percentage of cardiac intervals in which a unit fires) and (iii) the number of spikes a unit generates per cardiac interval (for review see Macefield et al. 2002). While the absolute strength of a multi-unit recording cannot be determined – the signal amplitude depends on the proximity of the electrode tip to the active nerve fibres – this is not a problem for single-unit recordings: a spike is either present or absent, i.e. the information content is quantal in nature. So, in single-unit terms, differences in the strength of sympathetic nerve activity may be brought about in two ways: by differences in (i) the number of active neurones (and recruitment of previously silent neurones) or (ii) the mean firing frequency of the neurones (or a combination of both). A difference in mean firing frequency may, in turn, be due to (a) a difference in firing probability of active neurones, i.e. the percentage of heart beats in which the neurones are active or (b) a difference in the number of spikes the neurones discharge within the bursts (or a combination of both). Increases in multi-unit muscle sympathetic nerve activity (MSNA) have been observed in several cardiovascular pathologies, such as congestive heart failure (CHF) and the obstructive sleep apnoea syndrome (OSAS). However, a high level of MSNA does not necessarily mean there is any underlying pathology. Indeed, many healthy individuals have high MSNA at rest yet normal blood pressures and no evidence of cardiovascular disease. And this is where single-unit recordings can contribute more information: by looking for increases in firing probability, mean firing rate or multiple firing we can see whether an elevated level of MSNA reflects any underlying pathology. We have already shown that individual muscle vasoconstrictor neurones in healthy subjects with high resting levels of MSNA have low firing probabilities, low firing rates and a low incidence of multiple firing (Macefield & Wallin, 1999). We have also shown that these parameters change with pathology, but that the nature of the changes differs according to the type of pathology: CHF patients show a high firing probability and a high mean firing rate, but no increase in multiple firing (Macefield et al. 1999), while OSAS patients show a high firing probability and a high firing rate, comparable to those seen in CHF, but also show an increase in multiple firing (Elam et al. 2002). So, in contrast to healthy subjects with many sympathetic bursts, an increase in unit firing frequency contributes to the increase of multi-unit activity in both CHF and OSAS, but the increase in frequency is brought about by different means. In this issue of The Journal of Physiology, Hisayoshi Murai and colleagues (Murai et al. 2009), from the Graduate School of Medical Science at Kanazawa University in Japan, report data from their single-unit recordings from muscle vasoconstrictor neurones in patients with heart failure. What is novel in their study is that they show that a manoeuvre that increases MSNA – sustained isometric handgrip exercise – increases the firing probability and the incidence of multiple firing of single units more in the CHF patients than in the healthy controls. This suggests that increases in vasoconstrictor drive in CHF can be brought about by changes in the firing properties of neurones already active, which is not to say that additional neurones are not also recruited. Why is this important? Well, it tells us that – despite the high level of resting MSNA in CHF – a ceiling has not been reached: there is still the capacity to increase vasoconstrictor drive even more. To what extent further vasoconstriction is accomplished by the now apparent capacity for multiple firing, versus recruitment of neurones still silent, in patients with CHF is now a topic most worthy of additional study.