Portal de Periódicos da CAPES

Arterial baroreflexes can determine long-termblood pressure

2004; Wiley; Volume: 89; Issue: 4 Linguagem: Inglês

10.1111/j.1469-445x.2004.t01-1-00053.x

1469-445X

Peter Sleight,

Blood Pressure and Hypertension Studies

This issue of Experimental Physiology includes a review of some recent experiments by Thrasher & colleagues, which cast doubt on the current standard view that baroreceptors are merely responsible for short-term heart rate and blood pressure control, and have no impact on long-term blood pressure control (Bray et al. 1999). These widely held concepts stem largely from the fact that ‘resetting’ of the baroreceptors to the new level of pressure occurs within hours or days, when the pressure is made to rise by some intervention which defeats the normal control mechanisms (McCubbin et al. 1956, Kreiger & Marseillan, 1966, Sleight et al. 1977). Thrasher (2002, 2003) reviews the work from his own laboratory in which he uses a novel dog model. This model uses surgical denervation of all aortic baroreceptors and one carotid sinus baroreceptor area, leaving the animal with one functional set of carotid baroreceptors. These baroreceptors are then chronically ‘unloaded’ by ligating the carotid artery below the innervated sinus. The response is a prompt and sustained (over 7 days) rise in mean arterial pressure (of some 20 mmHg systolic), after a 5-day control period in which there was no rise in arterial pressure. The pressure returned promptly (over 1 day) to the preligation systemic pressure on removal of the ligature. The author also reviews important new work from Lohmeier's group (2002, 2003). This group have examined the effect of the baroreflex on sodium excretion in animals with a split bladder. This enabled the examination of the urinary excretion of sodium from innervated versus denervated kidneys, in response to the vasoconstrictor and sodium retaining effect of continuously infused angiotensin-II. This infusion raised the mean arterial pressure by 30–35 mmHg in the dog. Sodium excretion from bilaterally innervated kidneys was equal on both sides in the absence of angiotensin infusion. In response to a 5-day infusion of angiotensin, sodium excretion increased only in the urine from an innervated kidney, compared to a denervated kidney, although both kidneys were exposed to the same increase in mean arterial pressure and plasma angiotensin-II concentration. This differential response was not present when the cardiopulmonary and baroreceptor areas were denervated. Thrasher also reviews other (different) experiments which support the view that baroreceptor resetting is far from complete – so that, although some resetting may take place – baroreflex control is still present, but this control is impaired (Osborn & Hornfeldt, 1998, Barrett et al. 2003). Some 30 years ago I too (as does Thrasher) criticised the experiments of Cowley and Guyton (Cowley et al. 1973), in which they claimed that, after baroreceptor denervation in the dog, there was a great increase in the variability of arterial pressure monitored over 24 h, but little increase in the level of mean arterial pressure, when averaged in a group of dogs studied after sino-aortic denervation. Figure 3, taken from the Cowley et al. (1973) publication, shows histograms of the beat-to-beat mean arterial pressure in conscious (but isolated) dogs. Panel (A) shows the tightly controlled narrow peak in one normally innervated dog, prepared with the spread out (highly variable) curve for the same dog after sino-aortic denervation. Panel (B) shows the records for all normal dogs. Panel (C) shows the histograms for all the sino-aortic denervated dogs. At the time, my reservations (Sleight et al. 1977; Sleight, 1979) were based on three main points. Although the individual mean arterial pressure of all the denervated dogs (panel C in the figure) showed great variability during the 24 h, the average 24 h mean arterial pressure was very different from dog to dog. Inspection of panel C appears to show two populations of dogs: (a) about half the dogs (the right of panel C) show that the peak of the 24 h histogram of arterial pressure is clearly higher than normal by some 30–50 mmHg, and (b) another group in which the peak of the mean arterial pressure histogram is well below that of the 24 h record for normal dogs (to the left of panel C). At this time I had been doing experiments recording from many intrathoracic nerves in the dog, including the first recordings from unmyelinated C-fibres coming from the heart, and myelinated fibres from the thoracic aorta and other areas (Sleight & Widdicombe, 1965; Coleridge et al. 1973). In the course of these recordings from the inferior cardiac nerve, and also from other intrathoracic nerves, it was apparent that many of these nerves were a mixture of afferent fibres (some clearly with the characteristics of myelinated baroreceptor fibres from the aortic arch) and some sympathetic efferent fibres coming from the cervical sympathetic ganglia. I therefore postulated (Sleight, 1986) that one possible explanation for the very disparate levels of pressure recorded over 24 h from individual dogs, after aortic & carotid denervation in Cowley and Guyton's experiments, was that in some cases, the thoracic nerves sectioned during the denervation of the aortic baroreceptor area resulted more in a sympathectomy to the heart (leading to lower arterial pressure), whereas in others the denervation may have been more of cardiac and aortic arch afferents (leading to higher postoperative pressures). The animals in Cowley and Guyton's experiments were carefully isolated during the recordings of 24 h arterial pressure, and were therefore subjected to a minimum of the environmental stimulae which might well have led to larger rises in arterial pressure in free-ranging denervated animals (Cowley et al. 1973). Inspection of their Fig. 4 (reproduced here as Fig. A1) shows also that the averaged 24 h mean arterial pressure of all the animals is in fact raised some 10–11 mmHg compared with normally innervated dogs. Today we know that even a few millimetres of pressure difference between one person and another can lead to a substantial increase in future hazards (Lewington et al. 2002). Frequency distribution curves of 24-h continuous recordings of mean arterial blood pressure in normal and sino-aortic denervated dogs (kept in isolation) A, individual dog before and after denervation. B, composite overlay of 10 normal dogs. C, composite overlay of 12 denervated dogs. Note: (i) The tight control of beat-to-beat mean arterial pressure over 24 h in a normally innervated dog isolated from environmental stimuli, compared with the highly variable pressures in the same dog after sino-aortic baroreceptor denervation (A). (ii) The very similar pressure histograms from 12 normal dogs kept in isolation, tightly centred around the normal mean arterial pressure close to 100 mmHg (B), (iii) the sino-aortic denervated dogs have each a much greater spread of beat-to-beat mean arterial pressures (now centred around an average MAP of 112 mmHg) (C). (iv) The average arterial pressures of individual dogs appear to form two groups, with a cluster of 6 dogs (out of 12) which show clear neurogenic hypertension, with average mean arterial pressures in the range of 125–160 mmHg. (right part of C). However some 2–3 dogs have average pressures below the normal 100 mg MAP, perhaps due to efferent sympathetic denervation (see text). (From Cowley et al. 1973; with permission; http://lww.com) I therefore believed that Cowley and Guytons' elegant, and at first sight critical, experiments had been misinterpreted, largely because those unfamiliar with the anatomical paths of the thoracic and aortic nerves do not realize that a clean surgical de-afferentation of the thoracic aortic baroreceptors is not possible without considerable interruption of efferent sympathetic fibres. Persson et al. (1987, 1988) further addressed this topic by distinguishing the separate effects of cardiopulmonary versus cardiopulmonary plus baroreceptor denervation. The latter procedure clearly produced neurogenic hypertension in conscious dogs, whereas the former had no effect on the arterial pressure level or its variability. The extent to which the pressor effects of denervation are mitigated by plasticity in the central nervous system – with any remaining fibres or other receptor areas compensating the effects of denervation – is another important consideration, as discussed by Thrasher. Muscle sympathetic nerve activity is influenced powerfully by baroreceptor input, especially during falling pressure (Eckberg et al. 1985). It is also influenced by other inputs from intrathoracic (low pressure) baroreceptors and of course, central nervous influences such as awakening or arousal – beyond the scope of this review. In 1972 I spent a sabbatical year in Paul Korner's department in the University of Sydney and together with an American graduate student, J. L. Robinson, compared the behaviour and physiological characteristics (pressure threshold, saturation pressure, and the slope of the linear part of the pressure/firing rate curve) in single carotid baroreceptor fibres from dogs with long-term renal clip hypertension, compared with intact normal dogs (Sleight et al. 1977). Although we confirmed, in both single fibre and multifibre preparations, the original work of McCubbin et al. (1956)– who used multifibre recording – that the carotid baroreceptors were reset to operate at a higher mean arterial pressure, and a higher threshold, than receptors from normal dogs, the single fibre recordings were able to show that the ‘reset’ receptors were clearly and significantly less sensitive to both dynamic and static pressure steps than the receptors from the normotensive dogs. We concluded this paper with a statement, ‘This would indicate that, at blood pressures above resting levels, further increases in pressure will be less efficiently buffered in the hypertensive dogs.’ (Sleight et al. 1977). Furthermore we found that the nerve discharge during pulsatile pressure was also significantly less in dogs with hypertension than in normotension (0.11 ± 0.01 and 0.17 ± 0.01 impulses per s per mmHg in hypertensive and normal dogs, respectively, P < 0.025). Finally we found that the time course of baroreceptor resetting in the dog lagged considerably behind the rise in arterial pressure. We did not observe any resetting in a dog that had been hypertensive for only 4 days. It appeared to take 5 days or longer before any resetting of the carotid baroreceptors was apparent in hypertensive dogs (Sleight et al. 1977). Baroreceptor resetting and loss of sensitivity is probably of importance in human hypertension. There are reports of hypertension resulting from damage to the carotid baroreceptor area in man (Pickering & Sleight, 1977; Gribbin et al. 1971). We re-studied one individual, previously normotensive, who had undergone bilateral carotid nerve section in 1964, in a (mistaken) attempt to help his long-standing asthma (Pickering & Sleight, 1977). The rationale for this now abandoned intervention was that bronchodilation would follow denervation of the carotid chemoreceptors. The patient's intra-arterial pressure immediately before surgery was 120/80 mmHg. Immediately postoperatively his intra-arterial pressure rose to 200/120 mmHg. When we studied him some 20 years after this operation he resembled a subject with fixed essential hypertension. Over the years after carotid denervation his clinical blood pressures had at first been very variable – resembling Cowley and Guyton's dogs, but later was persistently high. When we used our Oxford Phenylephrene method (Smyth et al. 1969; Bristow et al. 1969) to test his baroreflex response to an induced rise in pressure, we found no reflex cardiac slowing (Pickering & Sleight, 1977). Such experiments of nature (or rather of surgery) are now fortunately rare. It is nevertheless of interest that the carotid sinus is the commonest area for atheroma in human atherosclerosis (Heath et al. 1973). Under these circumstances what should be an elastic area, well able to sense stretch, is clearly an impaired barosensor, and this may therefore play some part in the development of the rise in arterial pressure with age in man (Eckberg & Sleight, 1992). The clinical importance of this re-examination of the importance of the arterial baroreflex for longer-term control blood pressure is that it might lead to novel pharmacological methods for increasing baroreflex gain. We know that this occurs in animals and man with some drugs, e.g. ACE inhibitors (Lee & Lumbers, 1981), and clonidine (Sleight et al. 1975; Sleight & West, 1975). Clonidine was a promising treatment for hypertension but was discarded because the central effects which led to the increase in baroreflex gain could not be separated from other central effects which caused drowsiness. Today's chemists and molecular biologists may be able to devise ways of separating such troublesome side-effects from effects which increase baroreflex gain. We also now know that the impaired baroreflex gain seen in some pathological states in man, e.g. cardiac failure (Eckberg et al. 1971), or after myocardial infarction (La Rovere et al. 1998), may be improved by exercise training (Coats et al. 1992). Yoga techniques, or religious rituals (such as repeating the Ave Maria prayer) may, by slowing respiration, improve baroreflex gain and increase measures of vagal tone (Bernardi et al. 2001). The very influential systems modelling approach to circulatory control proposed by Guyton and colleagues (which underemphasized the role of the arterial baroreflex in long-term blood pressure control) (Guyton et al. 1974) now seems flawed in two key respects (i) it erroneously suggested that baroreceptor ‘re-setting’ (in response to a prolonged rise in arterial pressure) was complete, and (ii) it underestimated or ignored the important baroreflex control of sodium excretion. Thrasher's review of more recent data suggests that arterial baroreflexes do indeed help determine longer term levels of blood pressure, although I would like to see his baroreceptor unloading model followed for longer than 1 week.