The baroreflex bandwagon: time to get off?

A recent editorial in this journal asked ‘Should we still use nitrovasodilators to test baroreflex sensitivity?’[1]. An even better question is ‘Should we still continue to test baroreflex sensitivity?'. I shall suggest that the time has come to jump off what might be called the ‘baroreflex bandwagon'. Since Smyth et al. [2] first measured the bradycardia produced in man by an intravenous bolus of a pressor drug, many different modifications and methods have been devised to examine arterial baroreceptor responses [3]. An enormous literature has grown up around the measurement of what is now called ‘baroreflex sensitivity'(BRS) – the ratio between changes in heart rate and applied or spontaneous changes in transmural pressure at the peripheral arterial baroreceptor structures. The apparent precision of the measurement is illusory. Many variables can influence it, including: the viscous and elastic properties of the arterial walls in which baroreceptor sensory endings are embedded; the tonic influence that efferent sympathetic innervation exerts on the local smooth muscle of the baroreceptor organs; the direct effect of the test drug on that smooth muscle; the direct effect of the test drug on the neural sensory endings themselves; the integrity of the afferent nerves concerned; the receptive state of the medullary nuclei, especially the solitary tract nucleus, dorsal nucleus of the vagus and the nucleus ambiguus; ascending neural influences on these nuclei from somatic afferents; descending neural influences upon these nuclei from higher centres, especially the hypothalamus and parabrachial nuclei; when a drug is used, the effect of the drug on those parts of the central nervous system not protected by the blood–brain barrier (notably the area postrema); tonic influences from the area postrema and higher centres on the caudal and rostral ventrolateral nuclei; the parasympathetic innervation of the effector systems, principally the vagus in heart rate control; the sympathetic innervation of the heart and blood vessels, involving synapses in the spinal cord influenced by descending tonic neural activity from the medullary raphe nuclei and influencing sympathetic discharge from the intermediolateral cell column; and the metabolic state of the heart muscle itself (intrinsic heart rate can be substantially influenced by local metabolic factors). Depressed baroreflex sensitivity has been recorded in all varieties of clinical and experimental hypertension. For this reason alone, it is unlikely that it has any significant aetiological role. When experimental hypertension has been acutely reversed, the decreased BRS rapidly returns to normal, for example, in rat 2K1C hypertension, after removal of the clip [4]. The carotid sinus baroreceptors in dogs are well known to become reset to higher pressures in chronic renal hypertension. This makes it difficult to believe that baroreceptor dysfunction could play any fundamental part in long-term blood pressure control. Tuck [5] opined that ‘most evidence indicates that baroreflex abnormalities in essential hypertension are a secondary event'. Zanchetti and Mancia [6] agreed and concluded that readjustments of cardiovascular baroreceptor reflexes which occur in every variety of hypertension appear to be ‘a consequence rather than a cause of hypertension'. Experimental denervation of the carotid sinus baroreceptors only raises mean blood pressure transiently, though there is a much increased spread of individual pressure values throughout the day [7]. More extensive peripheral baroreceptor denervation can raise mean arterial pressure permanently, though with a considerable increase in heart rate. Extensive peripheral baroreceptor denervation in a group of dogs raised average blood pressure from 101 to 123 mmHg while average resting heart rate rose from 85 to 124 beats/min [8]. Complete elimination of inhibitory afferent information by destruction of the solitary tract nuclei has produced fulminating neurogenic hypertension in rats [9] and chronic labile hypertension in cats [10] and dogs [11]. I have reviewed many similar experiments [12]. All such manoeuvres which interfere with afferent arterial baroreceptor information and produce labile neurogenic hypertension are accompanied by tachycardia and substantially increased efferent sympathetic cardiovascular tone. Such animal models do not resemble classic essential hypertension. A reduction in BRS comparable with that observed in every variety of clinical, animal genetic and experimental hypertension has also been reported in a vast number of other situations, including: in black Africans; in coronary artery disease; acute myocardial infarction; exercise; congestive heart failure; cardiomyopathies; volume overload cardiac hypertrophy; pulmonary hypertension; sino-aortic denervation; cardiac transplantation; multiple system atrophy; endogenous depression; pregnancy; severe hypoxia; sleep apnoea; prolonged supine bed rest; sodium depletion; mental stress; surgical stress; liver failure; obstructive sleep apnoea; global cerebral ischaemia; ovariectomy; high spinal cord section; nigrostriatal dopamine pathway ablation; Guillain–Barré syndrome; Chagas’ disease; Parkinson's disease; amyotrophic lateral sclerosis; type I diabetes; uraemia; and bilateral hind-limb ischaemia. Various drugs can also reduce apparent BRS, including: angiotensin II infusion; atropine; corticotrophin-releasing factor; adrenocorticotrophic hormone; adenosine; some angiotensin converting enzyme inhibitors; nitric oxide; nitrous oxide; smoking; endothelin infusion; dobutamine infusion; deoxycorticosterone salt administration; caffeine; intrathecal opioids; μ-opioid agonists; epidural anaesthesia; halothane; propofol; streptozotocin; systemic bupivacaine; cocaine; lidocaine; alpha-blocking drugs; benzodiazepines; diltiazem; metoclopramide; and alcohol. Do BRS measurements shed any light on the aetiology of hypertension? Consideration of all the possible confounding factors and the vast number of conditions in which BRS may be reduced makes naive the suggestion that ‘impairment of the baroreflex . . . may have an important influence on long-term arterial pressure’[13]. Unless some more precise measurement has been made of some component of the reflex path, such a statement is meaningless, even though there is much evidence to suggest that the changes which determine the rise of blood pressure lie within the central nervous system [12]. BRS measurement has two major attractions: it causes little disturbance to subjects and it invariably produces a (publishable) numerical result of some sort. There is no limit to the number of circumstances in which BRS can be examined. I suggest that further measurements of BRS will not shed any light on the aetiology of essential hypertension. The best way to bring the ‘baroreflex bandwagon’ to a halt is for grant administrators and journal referees to look critically at any new research proposal or submission which has the measurement of BRS as its main thrust. Referees need to ask exactly what new cardiovascular insight BRS measurement can provide. The potential clinical value of BRS measurement BRS is reduced by any form of cardiac damage and is being increasingly used to assess prognosis and therapy. Another related technique analyses long-term spontaneous reciprocal changes in blood pressure and heart rate. Studies have been made by power spectral or sequential analysis of variations in blood pressure and R–R interval, and by analysis of complex-demodulated oscillations in these variables. Analysis suggests that much of the spontaneous non-respiratory low frequency (0.1 Hz) blood pressure oscillation can be accounted for by baroreceptor reflexes, without the need to invoke an intrinsic brain oscillator. Myocardial infarction acutely reduces BRS. It later recovers and is virtually back to normal by 3 months [14]. These changes are not too surprising since a damaged myocardium is inevitably going to lose some of its former control range for heart rate. Efferent sympathetic nervous activity is increased after heart damage. BRS is inversely related to the probability of ventricular fibrillation during experimental myocardial ischaemia [15]. A low BRS is associated with increased mortality [16]. There are other measures of cardiac damage, such as decreased heart rate variability, low ejection fraction, ventricular premature beats, blood pressure reduction, altered breathing patterns, elevation of cardiac enzyme concentrations and of plasma noradrenaline, and reduction of urine volume and urine sodium with a rise in plasma creatinine. The prognostic value of these criteria (some much simpler than measuring BRS) needs rigorous comparison with measurements of BRS. Bigger et al. [16] suggested that the relative utility of measurements of heart rate variability and of BRS in predicting sudden cardiac death ‘needs to be evaluated in a large prospective study'. The larger ‘ATRAMI’ study has led the way to a more precise determination of the best prognostic indices to guide therapy [17]. BRS provides a rough clinically relevant measure of autonomic nervous system damage (e.g. in diabetes mellitus, in elderly patients complaining of dizzy turns and falls, or in Chagas’ disease). By indirectly reflecting the increased sympathetic nervous activity in heart failure, it can provide a semi-quantitative measure of functional cardiac impairment, even in mild failure [18]. It has been used to identify and quantify the reinnervation of the heart after transplantation. But these are obvious and banal situations in which, for example, the Valsalva manoeuvre usually provides enough information for most clinical purposes. Conclusion BRS measurement may have a part to play in assessing post-infarct prognosis and guiding treatment, though it needs to be critically compared with simpler methods. On the other hand, I suggest that further measurements of BRS, by themselves, can contribute nothing to understanding hypertension and they should be abandoned. C. John Dickinson