Hypertension in chronic renal failure and end-stage renal disease patients treated with haemodialysis or peritoneal dialysis

Guidelines should be used for clinical decisions about individual patients. However, most of the available guidelines are based on randomized clinical trials which only provide a statistical prediction of the likely effect of an intervention, usually based on the ‘average’ outcome, aggregated across all the heterogeneous patients participating in the trial. Heterogeneity due to the stage of the disease, its severity, organ complication, comorbid conditions etc., together with exclusion of the patients with major complications from the trial, limit the information generated by such trials for a precise prediction of outcome for an individual patient. Although these limitations may be partially circumvented by the clinical experience and pathophysiological knowledge of clinicians or experts (in the case of a guideline), personal bias is almost inevitable. A comprehensive approach to the issue of hypertension has recently been published [1]. The present review summarizes the most important aspects discussed in that review and, in particular, addresses some specific topics whcih, in the opinion of the Medical Expert Group, are rather relevant. The prevalence of hypertension varies according to the degree and the cause of renal failure in chronic renal insufficiency (CRI) or according to the type of dialytic treatment, haemodialysis (HD) or peritoneal dialysis (PD). However, the prevalence of hypertension is much greater in the CRI population (60–100%) than in the general population. With the remarkable exception of Tassin, France, where >s90% of patients achieve normotension without anti-hypertensive medication [2], the control of hypertension in HD patients is rather poor despite administration of anti-hypertensive drugs. This poor control has also been documented with 24 h ambulatory blood pressure (BP) monitoring where a reversed diurnal BP pattern was demonstrated [3]. Hypertension resistance to both volume control and drugs has been shown in a subgroup of patients, although it remains debatable whether ‘dry weight’ was really achieved in all of these patients [4,5]. The well established concept that hypertension is a risk factor for death both in the general population and in dialysis patients has recently been questioned by independent studies [4–9] showing that low pre-dialysis systolic BP is associated with increased adjusted mortality risk. No association with elevated mortality risk has been observed for pre-dialysis systolic hypertension. A number of explanations may be offered for this rather unexpected finding: (i) it may be an iatrogenic effect due to overmedication; (ii) all the studies were carried out in the USA on patients with an overall poor prognosis. The influence of expected survival is indeed crucial. The beneficial effect of BP reduction may need several years to become visible, and this may not be detectable if life expectancy is only 2–3 years; (iii) older malnourished patients have lower BP and shorter survival; and (iv) the presence of overt or not fully developed heart failure or disease may lower BP and survival. The Cox proportional hazard used in some of these studies may be inadequate to correct for all these confounding factors. In these clinical conditions, it is still not known if hypertension is caused by inappropriate secretion of renin for the status of body sodium and water, or if other renal or extra-renal factors are involved. A correct answer to this question is crucial not only for the understanding of the pathogenesis of hypertension but also for a correct therapeutical approach. There are close interrelationships among the various feedback loops regulating BP (renal, hormonal, haemodynamic, neural and humoral) [10], so a change in any one of these will generate changes in the others. Such secondary changes are likely to be controlled simply by normalizing the primary change that is responsible for the shift in the set point of the overall system. An example of this mechanism may be the increase in plasma catecholamines after renal artery constriction or the increase in sympathetic activity which may occur in CRI [11]. Both events may contribute to the rise in BP. Experimentally, the increase in plasma catecholamines is corrected by removing the primary cause (renal artery constriction), which also normalizes body sodium and renin. Similarly, the sympathetic hyperactivity of CRI may be corrected by administration of an angiotensin-converting enzyme (ACE) inhibitor [11]. Conversely, the blockade of the effect of catecholamines at their receptor may correct some haemodynamic abnormalities but will not affect the primary renal abnormalities responsible for the overall clinical conditions. The well documented beneficial effects of inhibitors of the renin–angiotensin system (RAS) (mainly ACE inhibitors) on both the cardiovascular system and the progression of renal failure may be due to their ability to block one of the primary triggering mechanisms responsible for the secondary cascade of events leading to the different clinical manifestations of the disease. This interpretation is weakened by the lack of careful studies comparing the long-term effects of ACE inhibitors with that of blockers of the sympathetic system. The latter may be useful at least for controlling the symptoms and cardiac consequences of hypertension, mainly in a subgroup of patients where ACE inhibitors, alone or in combination with calcium antagonists or a diuretic, cannot normalize BP. The following points are relevant both for correct understanding of the pathogenetic mechanisms of hypertension in CRI or dialysis patients and for appropriate therapeutic approaches: Experimental data on renovascular hypertension clearly show that the pathogenetic mechanisms of hypertension and the BP response to drugs change with time after the onset of renal artery stenosis [12]. In the initial stages, the pressor effect of renin and the efficiency of renin inhibitors on BP control are much greater than in later stages of hypertension. This time-dependency of pressor mechanisms has also been demonstrated in human hypertension. The pressor mechanisms at work in a patient with long-lasting renal disease thus differ from those in patients with a renal disease of recent onset. Such a difference must be taken into account in the discussion of pathogenetic or therapeutic issues. The demonstration that anephric patients on HD or PD only have a sodium-/volume-dependent hypertension may rule out the importance of many extra-renal mechanisms and suggests that one should concentrate on renal mechanisms. Some years ago, when bilateral nephrectomy in HD or PD patients was more frequent, this issue was reviewed [13] and it was concluded that, in all cases where volume dependency could be assessed, hypertension was caused by an excess of sodium and water. This conclusion is so important that it would be appropriate to have more data to support it. Transplant hypertension may be a very relevant clinical problem particularly when patients are given cyclosporin, because of the well documented pressor effect of this drug. Besides the mechanism discussed above and below, a genetic factor must also be considered here. It has been shown that some hypertension can be transplanted with the kidney, both in rats and in humans. In rats, this conclusion was reached with kidney cross-transplantation between genetically hypertensive strains of rats and their normotensive counterpart [14]. In humans, it was shown that recipients of kidneys removed from donors with hypertensive parents require more anti-hypertensive therapy to control their BP than recipients of kidneys removed from donors with normotensive parents. This ‘donor effect’ was seen only in recipients with normotensive parents and not in recipients with hypertensive parents, implying genetic co-adaptation and other genetic factors [15]. Whatever the interpretation of these findings may be, these data clearly support the relevance of genetic factors in determining the level of BP in chronic renal failure or end-stage renal disease. The prevalence of hypertension in PD patients is lower than that in HD patients, at least in the first years of treatment, although not all authors agree [16,17]. Achievement of ‘dry weight’ is more successful in PD patients, maybe because of the larger residual diuresis in PD patients or more favourable clearance of various humoral pressor factors. The more efficient rate of removal of sodium and water in PD than in HD may also play a role. The different prevalence of hypertension tends to disappear with longer duration of PD, which is also associated with the disappearance of residual diuresis. The latter event contributes to a more difficult control of body fluids in those patients on long-term PD than in those on short-term PD. Compared with short (3–4 h) HD, long (8 h) HD is associated with less cardiovascular morbidity (including hypertension) and longer patient survival [18,19]. It has been suggested that the beneficial effect on BP may not be directly dependent on volume control since normotensive patients on long HD may have the same volume indexes as hypertensive patients on short HD [20]. Unfortunately, many confounding factors were not properly controlled: diet, type of HD or lifestyle (French, Swedish, British and Dutch). These observations need to be confirmed by another better controlled study, but suggest that long HD may achieve better removal of some vasoactive substance(s) that potentially contribute to hypertension. Before accepting this explanation, however, one must also consider the ‘lag phenomenon’ [18], i.e. the finding that the fall in BP is delayed with respect to the normalization of extracellular fluid volume. This lag phenomenon not only becomes apparent when body fluid volume (or body weight) is plotted against BP during the first months of dialysis, but is also present in different clinical conditions involving reduction of body sodium, such as when a very low sodium diet is started in patients with essential hypertension or when patients with primary hyperaldosteronism are given aldosterone receptor antagonists. In both conditions the normalization of body fluids precedes the decrease in BP. It is likely that this lag phenomenon is increased in patients with long-lasting hypertension where the structural modifications of the cardiovascular system are more severe. This phenomenon is also consistent with the time-dependency of the pressor mechanisms, which was mentioned above in relation to experimental renal hypertension. A precise cause–effect relationship among the different pressor mechanisms triggered by decreased renal function has not yet been established, although this sequence can be accepted as an established fact, at least in experimental renovascular hypertension. Soon after renal artery constriction, the increase in plasma renin and peripheral vascular resistance accounts for the concomitant BP increase only for a few days. Later, body fluid retention contributes to the hypertension by increasing plasma and extracellular fluid volume and cardiac output. A few weeks later, plasma renin, body fluid and cardiac output are back to normal and hypertension is associated with increased vascular resistance [21,22]. Sympathetic hyperactivity, or increased plasma catecholamines, ouabain-like factor (OLF), endothelial factors and vascular structural changes (and/or peripheral autoregulatory mechanisms of blood-flow adaptation) have all been suggested as mechanisms supporting this third phase of hypertension. But the fundamental finding is that, if renal artery stenosis is removed within a critical period of time, BP returns to normal. BP does not return to normal if the removal of renal artery stenosis is carried out after that critical period of time. Some of the secondary pressor mechanisms mentioned above may therefore become autonomous, independent of the primary mechanisms that triggered them. The lag phenomenon, which also reflects all these mechanisms with the provisos above mentioned, may be taken as evidence that, in the long run, the normalization of body sodium and fluids achieved and maintained with relatively low rates of fluid removal with long dialysis, as practised in Tassin, is sufficient to control BP without antihypertensive drugs. The rate of fluid removal is probably also crucial: below a certain critical rate, the body’s homeostatic feedback loop can adapt much better, thus facilitating the achievement of the ‘normal’ set point of body sodium, water and BP existing before renal disease, without triggering counter-regulatory mechanisms. Failure to understand these phenomena may generate a lot of confusion and unfruitful controversies. For instance, acute studies [23] demonstrated a poor correlation between interdialytic BP changes and body weight changes (as an expression of acute change in body fluids). This finding has been used as an argument against a relevant role of body fluids in BP regulation and in favour of a role of other factors such as sympathetic overactivity, nitric oxide, etc. [23]. However, considering both the lag phenomenon and the time dependency of the pressor mechanism including structural adaptation of vessels, this poor correlation with acute changes is not unexpected and cannot be used as an argument against the long-term pressor role of body fluids. In other words, the variety of pressor mechanisms should not be considered only as an expression of contrasting views on the pathogenesis of hypertension but also as an expression of the sequence of pathophysiological feedback loops, which come into play in the different phases of the disease. This sequence is also affected by the overall genetic and environmental background, modulating the heterogeneous pattern of organ complications or comorbid conditions affecting the individual patients. A correct definition of the role of the individual pressor mechanisms present in a given patient must include the concept of context dependency. That is, the phase, time, genetics (if possible), environment, and so on, must be considered when assessing the respective role of each individual pressor mechanism. Taking into consideration this concept of context dependency, sodium and water retention together with inappropriate production of renin still have a leading role in determining BP. Therefore, all manoeuvres aimed at reducing and/or normalizing body sodium and water and the pressor effect of renin should rank first in the therapy of hypertension associated with renal diseases or with dialysis. Patience must be allowed before evaluating the effect on BP of such a manoeuvre. During the fundamental manoeuvres aimed at achieving ‘dry weight’ in some patients, it may be helpful to add hypertensive agents (mainly inhibitors of the RAS or of the sympathetic system) temporarily, which may accelerate the process of BP normalization. In addition, it is not possible to exclude the possibility that the still functioning diseased kidney may, in some patients, directly activate the sympathetic nervous system. Drugs acting on the sympathetic nervous system may be required for this specific reason. Renal artery disease has recently come to be regarded as a common and increasingly frequent cause of CRI [24,25]. Improvement of diagnostic procedures and ageing of the population with vessel alteration secondary to atherosclerosis or diabetes may be responsible for such a phenomenon. Therefore, any evaluation of patients aged >50 years with hypertension and renal failure should at least include an echo Doppler study of the renal vascular tree. Patients with hypertension, renal failure and other signs of peripheral artery disease should undergo an angio magnetic resonance of the renal vessels, since they have a high risk of renal artery disease (around 30–40%) and are potentially treatable. The most recent guidelines and publications [26,27] recommend that the systolic and diastolic BP of patients with CRI should be maintained at ≤130 mm Hg and ≤80 mm Hg for BP, respectively. When proteinuria is present, the target BP should be reduced to ≤125/75 mm Hg. Adequate control of BP certainly slows the progression of CRI and this effect is particularly evident in patients with increased urine protein excretion. Such a beneficial effect was first demonstrated in diabetic nephropathy but has also been proven in other forms of nephropathy. There is a general agreement on the importance of reducing dietary sodium intake in these patients to about 2–3 g/day, and to use loop diuretics alone or in combination with a thiazide diuretic if oedema refractory to loop diuretics is present. For some time, it has been debated whether certain classes of antihypertensive agents, independent of their antihypertensive effect, exert a beneficial effect on the preservation of renal function in CRI. Although this question is not yet settled completely, this additional effect can be tentatively assigned to ACE inhibitors [28,29], while more studies are needed for the nondihydropyridine calcium antagonists or angiotensin II receptor antagonists. The superiority of one ACE inhibitor over the others has not yet been demonstrated [30]. An important practical problem is the initial impairment of renal function that occurs when therapy with ACE inhibitors is started in patients with CRI. The magnitude of such an effect may vary from one patient to another. While the cause of such a decrease in glomerular filtration rate (GFR) is clear (relative decrease in glomerular capillary pressure due to vasodilatation of the efferent arteriole), the cause of individual variation in the fall in GFR is less clear. Certainly, the presence of renal artery disease, particularly when bilateral, is one important cause that precludes the use of these drugs. However, a remarkable decrease of GFR may also occur in patients without renal artery disease. Some type of intrarenal small artery lesions are probably present in such patients, but has not yet been proven. From the practical point of view, any patient with mild to moderate CRI who starts treatment with ACE inhibitor should have his/her plasma creatinine monitored for at least 14 days after initiation of therapy. An initial increase in plasma creatinine of ≤20–30% of the initial level is acceptable, but if plasma creatinine continues to rise above this critical level, ACE inhibitors should be withdrawn. Two weeks of monitoring creatinine closely should be long enough to detect initial impairment of renal function [31]. Creatinine should then be measured twice monthly for 2 months, then every 2 months thereafter. The critical serum creatinine above which ACE inhibitors should not be given or should be stopped is also controversial. Serum creatinine concentrations of 3 and 5 mg/dl, respectively, may be appropriate cutoff levels. In patients who are already being treated with ACE inhibitors and in whom the withdrawal of these drugs does not a decrease in plasma creatinine within the ACE inhibitor may be because these inhibitors have a effect on the cardiovascular system. ACE inhibitors are the drugs for hypertension in these clinical conditions. However, they may not be to normalize BP, may cause impairment of renal function and may have effects such as and this blockers or angiotensin II receptor antagonists may be used or in combination ACE inhibitors to achieve target BP. There are data supporting the that angiotensin II receptor their for patients in whom ACE inhibitors cause the effect of ACE inhibitors conditions such as artery disease, and may require other drugs which may on both hypertension and the comorbid for or artery disease, or for presence of or may be needed in some patients when the above drugs are to control the hypertension or The sympathetic hyperactivity present in some of these patients may provide the for these drugs when ACE inhibitors are present the control of BP, particularly in patients on short (3–4 h) HD, is very in of the of very anti-hypertensive drugs In hypertension is almost in patients on long (8 h) HD treatment is more and is not by many patients, so it would be to achieve the same of long dialysis by the present short dialysis in hypertensive patients From has been discussed it is probably not the of the HD but the rate and of fluid and sodium individual of the patients, that are most These individual the duration and of hypertension by secondary structural cardiovascular the genetic background, which the set point of the different homeostatic feedback loops and the structural response of the cardiovascular and the role of the The demonstration that only hypertension is present in anephric patients may the discussion about the of removing the at least in some patients. removal should be considered not only in of the of fluid removed during a given HD but also in of long it to achieve weight instance, weeks The is that, in patients achieving weight within a longer period of time, BP control should be much in the long A possible study should be carried out in of patient well for the mentioned The most patients to include are those with long-lasting and hypertension before HD and those with important cardiovascular structural This could also be used to which of the available anti-hypertensive drugs is more in the homeostatic feedback mechanisms, particularly during the first months of of the RAS or of the sympathetic nervous system may be given in for a certain period of time before achieving