The amplifier hypothesis: persisting dissent

This issue of the Journal of Hypertension contains an article by Wright et al. [1], which builds on a previous study by Wright and Angus [2], for which we also wrote a commentary [3]. As before, we were reviewers for this paper, and disagreed with Wright et al. [1] both in substance and in style. At this point, the Editor of the Journal could have rejected the paper, requested more extensive revisions or sent the paper to other reviewers for a ‘second opinion'. However, neither of these approaches would have brought the relevant issues out in the open. Instead, because the Journal strongly favours debate about important scientific issues between reputable scientists, Dr Zanchetti decided to accept the paper for publication but to have it accompanied by a commentary by the reviewers. We fully endorse this approach and are delighted to accept his invitation to state our point of view on the main issues of the paper by Wright et al. [1]. We wish to congratulate the authors on their dedication in obtaining the data which led to the current publication. It is only with their interpretation and conclusions that we take issue in the hope that, by open cordial debate with our friends, we can arrive at the correct consensus. Wright and co-workers used instrumented rabbits with hypertension induced by wrapping the kidneys to study the effects of vasodilator and constrictor drugs on haemodynamics in the presence and absence of neurohumoral block. They compared the results obtained with those obtained from sham-operated normotensive animals. The object was to provide further evidence to support their contention that the circulation of a hypertensive organism acts as an amplifier when stimulated due to pre-existing structural changes. The authors conclude that the enhanced mean arterial pressure and total peripheral resistance observed in wrap hypertension are in accord with an interaction between vascular geometry and the sum of altered neurohumoral plus local activity, plus a rarefaction component. Unfortunately, this is not the case because this conclusion cannot be substantiated from the measurements made. What has actually been measured is arterial blood pressure and cardiac output: other parameters have been calculated of necessity. In this context, we have values for total peripheral conductance (TPC); total peripheral resistance (TPR) is 1/TPC and an arbitrary averaged radius (r) is calculated as r = 1/4√TPR. Here enters the first confounding error: the authors state that this model of hypertension is associated with rarefaction. The implication is that some of the TPR increase in this model must be due to rarefaction; taking this at face value, there must have been an overestimation of the calculated reduction in r. A possible error in the estimation of r is particularly important since, as it stands, the calculation indicates that lumen diameter is reduced by a given amount from maximal dilation to maximal constriction in hypertension. This is contrary to the structurally induced exaggerated changes in lumen diameter in hypertension proposed by Folkow, as discussed previously [3]. Secondary forms of hypertension do result in structural changes in the resistance vasculature, namely wall thickening and luminal narrowing, a requirement for the vascular amplifier concept, and also rarefaction [4]. However, the contribution of these structural changes to the increased TPR is probably negligible given the fact that, in secondary models of hypertension, removal of the pressor stimulus leads to falls in pressure which are much too rapid to be mediated by a normalization of vascular structure [5]. In our commentary [3] on the previous paper by Wright and Angus [2] regarding the amplifier hypothesis, we reviewed the studies published in the 1980s and 1990s on blood pressure and TPR responsiveness in hypertensive versus normotensive animals. Differences in responsiveness were found to be agonist-specific (i.e. enhanced responses to some but not to others). We actually noted decreased blood pressure and TPR responses to phenylephrine in spontaneously hypertensive (SHR) versus Wistar-Kyoto rats over a wide age-range, without or with ganglionic blockade, the latter to eliminate differences in baroreflex buffering [6]. In contrast, SHR exhibited increased responses to methoxamine, but only at low (and not at higher) rates of infusion [6]. These findings are consistent with changes in function rather than in wall structure per se in determining overall TPR responsiveness in the whole animal. In the study by Wright and Angus [2] in hypertensive rabbits with ganglionic blockade, adenosine-induced decreases and methoxamine-induced increases in TPR became larger compared to sham-operated rabbits only at the higher rates of infusion (and therefore at large changes in TPR). We suggested that in this study changes in TPR around the (patho)physiological set point did not provide evidence for a TPR amplifier. Enhanced responses at the extremes of the dose–response curves on their own may be considered an ‘amplifier’ function for arterial hypertrophy, but can hardly be considered relevant for the development or maintenance of most hypertension. In the study by Wright et al. [1] in the present issue of the journal, hypertensive and normotensive rabbits were studied without and with ganglionic blockade, as well as with more complete ‘neurohumoral blockade'. The authors state that the study has four main new findings. However, two of these ‘new findings’ are clearly not findings: (i) Finding 1: ‘the systemic vasculature in wrap hypertensive rabbits behaves like a TPR amplifier which is responsible for the enhanced rise in MAP'. Whether a TPR amplifier (if indeed present, see below) is responsible for an enhanced rise in mean arterial pressure is an interpretation, and not a finding. (ii) Finding 3: ‘an hypothesis to explain this is discussed below'. Obviously, an hypothesis is not a finding. Now, let us look at the two actual findings. Is there indeed evidence for a TPR amplifier which could contribute to a rise in MAP, and are the baroreflex effects on constrictor tone similar in the hypertensive and normotensive rabbits? For the first finding, we would like to draw the reader's attention to Figure 1, which is redrawn from the paper by Wright et al. [1]. After neurohumoral blockade, the two lower doses of adenosine and methoxamine change TPC by approximately 60–70% in the sham-rabbits, but cause no change at all in the hypertensive rabbits. This finding is strongly indicative of decreased responsiveness of overall arterial smooth muscle and thereby TPR in the hypertensive rabbits. Despite the apparent non-linearity of the overall dose–response curve in the hypertensive rabbits, the authors calculate slopes anyway (see Fig. 2 in their paper) without providing a statistical validation, and smooth the curves somewhat by removing the two lower doses. Removal of doses as ‘subthreshold’ is rather arbitrary, particularly if only one group exhibits this threshold but not the other (see their Results section). Even so, it is readily apparently from their Figs 1 and 2 that, in normotensive rabbits, there is a good linear relationship, but not in the hypertensive rabbits which exhibit a different relationship (see redrawn Fig. 1 in this commentary). Thus, a more careful (skeptical?) analysis of the data suggests an opposite conclusion for responsiveness around the set-point of TPR: decreased rather than increased responsiveness of TPR to stimuli in the hypertensive rabbits. In a follow-up on our previous commentary, Folkow [7] states ‘To use a somewhat wild parallel: gravity is always there, and is not made ‘‘obsolete’’ because birds, and aeroplanes, can fly. To do so they must engage substantial ‘functional factors’ and trust the presence of air to offset the ever-present gravity – which ultimately takes them all down again'. In this context, one may postulate that in mild to moderate hypertension a structural amplifier may be present, but (in contrast to gravity) in vivo in the whole organism is only a minor player around the set-point of cardiovascular homeostasis, and is not necessarily able to even offset functional changes in arterial smooth muscle regulation. A more important role may be present in larger ranges of blood pressure (e.g. severe hypertension).Fig. 1: Effects of intra-atrial infusions of adenosine and methoxamine on total peripheral conductance (TPC) in sham and wrap-hypertensive rabbits with neurohumoral blockade. Adapted from Wright et al. [1] with permission. As another main new finding, Wright et al. [1] state that baroreflex effects on constrictor tone are similar in the normotensive and hypertensive rabbits. Closer review of the data in their Figure 1 suggests otherwise. In rabbits with intact baroreflex, the lower two doses of adenosine and methoxamine are similarly ineffective in changing TPR. In contrast, as stated already above, in the absence of a functional arterial baroreflex, these lower doses are still ineffective in the hypertensive rabbits, but now are effective in the normotensive rabbits. This pattern of changes does not prove, but is consistent with more effective baroreflex buffering of sympathetic tone in the normotensive versus hypertensive rabbits, as has been shown in other hypertension models [8]. In summary, the studies by Wright et al. [1] are important extensions of their previous studies in addressing the in-vivo relevance of a concept based on in-vitro findings. As we stated previously [3], in their analyses and interpretations, the authors retain a commitment to a hypothesis disputed by others, and we submit that the body of evidence from both in-vitro and in-vivo findings remains weighted against their position. Acknowledgement Frans H.H. Leenen is a career investigator of the Heart and Stroke Foundation, Ontario, Canada.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.