Abstract
Among patients on dialysis, hypertension is highly prevalent and contributes to the high burden of cardiovascular morbidity and mortality. Strict volume control via sodium restriction and probing of dry weight are first-line approaches for the treatment of hypertension in this population; however, antihypertensive drug therapy is often needed to control BP. Few trials compare head-to-head the superiority of one antihypertensive drug class over another with respect to improving BP control or altering cardiovascular outcomes; accordingly, selection of the appropriate antihypertensive regimen should be individualized. To individualize therapy, consideration should be given to intra- and interdialytic pharmacokinetics, effect on cardiovascular reflexes, ability to treat comorbid illnesses, and adverse effect profile. β-Blockers followed by dihydropyridine calcium-channel blockers are our first- and second-line choices for antihypertensive drug use. Angiotensin–converting enzyme inhibitors and angiotensin receptor blockers seem to be reasonable third–line choices, because the evidence base to support their use in patients on dialysis is sparse. Add-on therapy with mineralocorticoid receptor antagonists in specific subgroups of patients on dialysis (i.e., those with severe congestive heart failure) seems to be another promising option in anticipation of the ongoing trials evaluating their efficacy and safety. Adequately powered, multicenter, randomized trials evaluating hard cardiovascular end points are urgently warranted to elucidate the comparative effectiveness of antihypertensive drug classes in patients on dialysis. In this review, we provide an overview of the randomized evidence on pharmacotherapy of hypertension in patients on dialysis, and we conclude with suggestions for future research to address critical gaps in this important area.
- hypertension
- pharmacotherapy
- cardiovascular disease
- Adrenergic beta-Antagonists
- Angiotensin Receptor Antagonists
- Angiotensin-Converting Enzyme Inhibitors
- Antihypertensive Agents
- blood pressure
- Humans
- Mineralocorticoid Receptor Antagonists
- renal dialysis
- Sodium
Introduction
Hypertension in people with ESRD on chronic dialysis is often poorly controlled, highly prevalent, and difficult to diagnose (1,2). Elevated BP, particularly when recorded outside of dialysis with home or ambulatory BP monitoring, is clearly associated with increased risk of cardiovascular morbidity and mortality (3). Sodium and volume overload are the most important causes of hypertension in patients on dialysis; therefore, nonpharmacologic interventions, such as dietary sodium restriction, enhanced sodium removal with dialysis, and probing of dry weight, should be the initial therapies for achievement of adequate BP control (4,5). This therapeutic approach is challenging, because several factors, such as short delivered dialysis (6), missed dialysis sessions, patient nonadherence to the dietary recommendations for sodium restriction, and intradialytic sodium gain through high dialysate sodium prescription, may be barriers to the achievement of dry weight (7). Even after aggressive volume management, hypertension remains poorly controlled; in these patients, pharmacologic therapy is necessary to control BP. Most classes of antihypertensive drugs are useful for pharmacotherapy of hypertension in patients on dialysis, and a combination therapy is often required (4,5).
In this article, we discuss the choice of the appropriate antihypertensive regimen in the era of the currently available evidence from randomized studies evaluating the BP-lowering efficacy, safety, and cardioprotective and pharmacokinetic properties of antihypertensive agents in the dialysis population. In the absence of properly designed randomized trials, several aspects of pharmacologic and nonpharmacologic management of hypertension in patients on dialysis remain obscure. We conclude with critical gaps in the existing knowledge, provide directions for future research, and call for randomized trials among patients on dialysis.
Research Challenges in Designing Interventional Studies
Compared with the nondialysis population, BP among patients on dialysis is highly variable from pre- to postdialysis and from one day to the other (8). Peridialytic BP recordings cannot accurately reflect the actual BP burden over the entire interdialytic period and can easily miss the magnitude of treatment-induced reduction in BP (9). In addition, out-of-dialysis unit BP recordings were shown to be strongly associated with all-cause and cardiovascular mortality, contrary to the limited or no prognostic significance of BP recordings obtained within the dialysis unit (3,10,11). The use of out-of-dialysis BP recordings (home or ideally, interdialytic ambulatory BP) as outcome measures of the efficacy of antihypertensive drugs is critical. However, state of art out-of-dialysis BP recording to evaluate BP outcomes are available only in a few of the studies conducted so far (Table 1). Notably, use of ambulatory BP monitoring also has the advantage of recording BP during nighttime, providing additional information on the efficacy of drugs to restore the nondipping BP pattern and improve nocturnal hypertension. The blunted nocturnal BP fall is common in patients on dialysis and associated with acceleration of target organ damage (12) and increased mortality risk (13).
Magnitude of change in ambulatory BP in response to pharmacologic and nonpharmacologic interventions in patients on dialysis
Nonpharmacologic interventions (i.e., changes in dry weight or dialysate sodium) can easily confound the BP-lowering efficacy of antihypertensive drug therapy. The Dry-Weight Reduction in Hemodialysis Patients (DRIP) trial showed the net BP–lowering effect of probing of dry weight in patients on hemodialysis who are hypertensive (14). A reduction in postdialysis weight of 0.9 kg provoked a significant decrease of 6.9/3.1 mmHg in interdialytic ambulatory BP at 4 weeks. Because background antihypertensive therapy was not modified, BP lowering seen in the DRIP trial is solely attributable to dry weight reduction. In contrast, it is rather difficult to distinguish the BP-lowering effect of drug therapy from the BP reduction caused by dry weight adjustment in studies combining pharmacologic and nonpharmacologic BP–lowering interventions. In the Hypertension in Hemodialysis Patients Treated with Atenolol or Lisinopril (HDPAL) trial (15), changes in interdialytic ambulatory BP during follow-up were no different between the atenolol and the lisinopril groups. However, lisinopril-treated patients required greater dry weight reduction and administration of higher numbers of antihypertensive drugs as add-on therapy (15). This suggests that the BP-lowering efficacy of atenolol is superior to that of lisinopril.
Pharmacologic Management of Hypertension in Patients on Dialysis
Meta-analyses of randomized trials showed that BP lowering with the use of antihypertensive agents reduces cardiovascular morbidity and mortality in patients on dialysis (16,17), particularly when they are hypertensive (16). All antihypertensive drug classes can be used for pharmacotherapy of hypertension (4,5). The exception may be diuretics, which are generally ineffective in patients with low GFR. Echocardiographic studies showed that, among anuric patients, intravenous administration of high–dose loop diuretics causes minimal alterations in central hemodynamic parameters (18). Given the risk of ototoxicity, use of diuretics in anuric patients on dialysis is not recommended (4,5). These compounds may have a role in those with preserved residual renal function as a maneuver to limit interdialytic weight gain (19); whether such an approach is useful or detrimental has not been adequately examined.
Angiotensin–Converting Enzyme Inhibitors and Angiotensin Receptor Blockers
Blockade of the renin-angiotensin-aldosterone system (RAAS) is often recommended as first–line antihypertensive therapy for patients on dialysis by extrapolation of cardiovascular benefits of RAAS inhibitors in the general population (5). However, extrapolation of outcomes from the general population or those with CKD not on dialysis to those on dialysis is hazardous. For example, in the Fosinopril in Dialysis trial 397 patients on hemodialysis were randomly assigned to the angiotensin–converting enzyme inhibitor (ACEI) fosinopril (titrated up to 20 mg/d) or placebo for 48 months (20). This study enrolled patients with left ventricular hypertrophy (LVH), but hypertension was not an inclusion criterion. Although fosinopril significantly lowered predialysis BP relative to placebo in the subgroup of hypertensive participants, incidence of fatal and nonfatal cardiovascular events was no different between the active treatment and placebo groups (relative risk [RR], 0.93; 95% confidence interval [95% CI], 0.68 to 1.26) (20). In contrast, in a much smaller Japanese study that included only 80 patients on hemodialysis without overt cardiovascular disease, the angiotensin receptor blocker (ARB) candesartan improved cardiovascular event–free survival versus placebo (21). Similarly, Suzuki et al. (22) showed that, among 360 patients with hypertension on hemodialysis, ARB therapy (valsartan, candersartan, or losartan) reduced by 49% the risk of cardiovascular death, nonfatal myocardial infarction (MI), stroke, coronary revascularization, and hospitalized congestive heart failure (CHF) in comparison with control therapy not including ARBs or ACEIs (hazard ratio [HR], 0.51; 95% CI, 0.33 to 0.79). A meta-analysis of these three studies showed that ACEI and ARB use was not associated with significant reduction in the risk of future cardiovascular events (pooled RR, 0.66; 95% CI, 0.35 to 1.25), indicating the necessity for larger high–quality studies (23). The subsequent Olmesartan Clinical Trial in Okinawa Patients under Okinawa Dialysis Study Trial compared the effect of olmesartan (10–40 mg/d) versus control therapy not including RAAS inhibitors in 469 patients with hypertension on hemodialysis. Over a mean follow-up of 3.5 years, incidence of all-cause death, nonfatal stroke, MI, and coronary revascularization was identical in the ARB and control arms (HR, 1.00; 95% CI, 0.71 to 1.40) (24).
RAAS blockade is suggested to exert beneficial actions on LVH and large artery stiffness, both of which are strong cardiovascular risk predictors in patients on dialysis (25,26). This notion is supported by small randomized studies showing that, compared with placebo or other therapies, ACEIs and ARBs caused greater reductions in left ventricular mass index (LVMI) and pulse wave velocity (PWV) (Table 2) (27,28). A previous meta-analysis of five randomized studies showed that ACEI or ARB use resulted in a significant reduction in LVMI, with a weighted mean difference of 15.4 g/m2 (95% CI, 7.4 to 23.3) relative to control therapy (23). Improvement in LVMI with ARB therapy was evident in another meta-analysis of six randomized studies (29). However, these benefits were not consistent in all relevant studies. In the recent Saving Residual Renal Function among Hemodialysis Patients Receiving Irbesartan Study, 12-month treatment with irbesartan was not superior to placebo in reducing LVMI and aortic PWV; notably, ARB-induced reductions in LVMI and PWV went in parallel with the changes in predialysis BP (30). Therefore, it remains unclear whether these beneficial actions on cardiac remodeling and arterial stiffness are BP mediated or BP independent (31).
Randomized studies evaluating the effects of angiotensin–converting enzyme inhibitors and angiotensin receptor blockers in patients on dialysis
Dual RAAS blockade may offer additive cardiovascular protection in specific subgroups of patients on dialysis (i.e., in patients with severe CHF). In a double-blind trial, 332 patients on hemodialysis with left ventricular ejection fraction (LVEF) <40% already treated with ACEIs were randomized to add-on telmisartan (40–80 mg/d) or placebo (32). After a mean follow-up of 35.5 months, telmisartan reduced cardiovascular mortality (HR, 0.42; 95% CI, 0.38 to 0.61) and hospitalizations for worsening CHF (HR, 0.38; 95% CI, 0.18 to 0.51). Incidence of hyperkalemia was similar in the telmisartan and placebo groups (32). Whether dual RAAS blockade is generalizable to the overall dialysis population requires additional research; this trial included patients with particular clinical characteristics (i.e., patients with CHF and reduced LVEF). Additionally, it could be argued that the low incidence of hyperkalemia with dual RAAS blockade in this trial could be partially attributed to the fact that study participants were receiving intensive RRT with four dialysis session per week and that most of them had a well functioning arteriovenous fistula as vascular access for dialysis (32).
The intradialytic kinetics of various ACEIs differ substantially from each other. With the exception of the nondialyzable fosinopril, most of the other ACEIs are removed during dialysis. In contrast, ARBs are nondialyzable, and no supplement doses are required (5,33).
In summary, the evidence base to support the use of ACEIs or ARBs among patients on dialysis for either cardioprotection or regression of LVH is thin. Thus, in our practice, we use them as third-line agents after blockers and dihydropyridine calcium-channel blockers (CCBs).
Mineralocorticoid Receptor Antagonists
Among patients not on dialysis with CHF and reduced LVEF, add-on therapy with mineralocorticoid receptor antagonists (MRAs) on top of background treatment with ACEIs, ARBs, and/or β-blockers effectively reduced cardiovascular morbidity and mortality (34). Efficacy of spironolactone and eplerenone among patients on dialysis is less well studied, mainly because of the concerns that any benefit from their use may be counteracted by excessive risk of life-threatening hyperkalemia. MRAs could theoretically reduce potassium removal through the intestine, which represents an important route of excretion, particularly in anuric patients on hemodialysis. A recent randomized study compared the safety profile of eplerenone (25–50 mg/d) versus that of placebo in 146 patients on hemodialysis. Over a 13-week follow-up, incidence of hyperkalemia (defined as predialysis serum potassium >6.5 mmol/L) was higher in eplerenone-treated patients (RR, 4.50; 95% CI, 1.0 to 20.2) (35). However, eplerenone was not associated with excess need for permanent drug discontinuation because of hyperkalemia or hypotension, which was the primary study end point (35). Whether potassium-binding resins could be helpful in maintaining the potassium balance in patients with dialysis-requiring CKD treated with spironolactone or eplerenone requires randomized trials to evaluate efficacy and safety of this strategy (36).
The cardioprotective properties of MRAs among patients on dialysis are supported by a number of randomized studies showing that these drugs culminate in sustained reduction of interdialytic BP and improvement in LVMI, LVEF, and carotid intima-media thickness (Table 3) (37–39). Two larger studies provided evidence that these beneficial effects on surrogate cardiovascular risk factors may be translated into survival benefits. In the Dialysis Outcomes Heart Failure Aldactone Study, 309 oligoanuric patients on hemodialysis were randomized to spironolactone (25 mg/d) or no add-on therapy for a mean follow-up of 3 years. Spironolactone lowered by 62% the risk of cardiovascular death or cardiovascular-related hospitalization (HR, 0.38; 95% CI, 0.17 to 0.83) (40). Incidence of serious hyperkalemia leading to drug discontinuation was as low as 1.9%. In the second study, 253 nonheart failure patients receiving hemodialysis or peritoneal dialysis were randomized to 2-year add-on therapy with spironolactone (25 mg/d) or placebo. Spironolactone reduced by 58% the occurrence of the composite primary outcome of cardiocerebrovascular death, observed cardiac arrest, and sudden death (HR, 0.42; 95% CI, 0.26 to 0.78) (41). These promising results need to be confirmed in larger, properly designed studies. A trial named the Aldosterone Antagonist Hemodialysis Survival Trial (42) that plans to recruit 825 patients on hemodialysis is underway and anticipated to shed more light on the potential harms and benefits of MRAs in patients on dialysis.
Randomized studies evaluating the effects of mineralocorticoid receptor antagonists in patients on dialysis
β-Blockers
Vulnerability of patients on dialysis to serious arrhythmias and sudden death together with excessive sympathetic overactivity make β-blockade an attractive therapeutic option for cardiovascular protection in this population (Tables 4–6) (43). In a double-blind fashion, 114 patients on hemodialysis with dilated cardiomyopathy (not necessarily hypertensive) were randomized to carvedilol (titrated up to 25 mg twice daily) or placebo. Over a mean follow-up of 2 years, carvedilol improved left ventricular systolic performance and reduced the risk of all-cause hospitalization (HR, 0.44; 95% CI, 0.25 to 0.77) and mortality (HR, 0.51; 95% CI, 0.32 to 0.82) (44). Cardioprotective properties of β-blockers are supported by the HDPAL trial which performed a head-to-head comparison between the β-blocker atenolol and the ACEI lisinopril (both administered thrice weekly postdialysis) in 200 patients with hypertension on hemodialysis with LVH (15). The HDPAL Trial was neutral with respect to the primary outcome; improvement in LVMI over the 12-month follow-up was similar in both arms. However, lisinopril-treated patients required more aggressive volume management and administration of higher numbers of antihypertensive drugs to achieve the prespecified home BP target of <140/90 mmHg. The HDPAL Trial was prematurely terminated because of excess risk of serious cardiovascular adverse events in the lisinopril group. Incidence of the combined end point of MI, stroke, hospitalized heart failure, and cardiovascular death was 2.29 times higher in lisinopril-treated patients than in atenolol-treated patients (incidence rate ratio, 2.29; 95% CI, 1.07 to 5.21) (15). In a secondary analysis of the HDPAL Trial, atenolol was shown to be superior to lisinopril in improving aortic PWV (45). The recently reported Beta-Blocker to Lower Cardiovascular Dialysis Events Trial failed to advance our knowledge on the cardioprotective role of β-blockade because of the low recruitment rate that resulted in a small sample size (46). Particularly, among 1443 patients screened, only 354 were eligible, 91 consented, and 72 entered the 6-week active treatment run–in period. Of these, only 49 participants (68%; 95% CI, 57% to 79%) tolerated carvedilol therapy (6.25 mg twice daily) during the run in and progressed to randomization (46). Narrow inclusion criteria and insufficient washout of background β-blockade therapy led to exclusion of high-risk patients who were more likely to benefit from the cardioprotective actions of carvedilol.
Randomized studies evaluating the effects of β-blockers in patients on dialysis
Randomized studies evaluating the effects of calcium-channel blockers in patients on dialysis
β-Blockers represent a highly variable medication class with respect to their pharmacokinetic properties. Atenolol is easily removable during dialysis compared with lipophilic agents, such as propranolol (5,33). Intradialytic removal of β-blockers, apart from interfering with their efficacy, may also exacerbate the risk of serious arrhythmias and sudden death. In a propensity–matched retrospective cohort study of 6588 patients on hemodialysis, use of a high–dialyzable versus a low–dialyzable β-blocker was associated with 40% higher risk of all-cause death (47). Another concern around the use of β-blockers is the risk of exercise-induced or fasting hyperkalemia, an adverse effect that has been described for the nonselective β-blockers nadolol and labetalol and attributed to the role of β2-receptors in regulating potassium entry intracellularly (48). Finally, unfavorable metabolic effects of older versus newer β-blockers on serum lipid profile, insulin sensitivity, and risk of new-onset diabetes may have unintended cardiovascular consequences, which should be also taken into account (49).
CCBs
CCBs can effectively lower BP, even in the volume-expanded state (50), and are often used as combination therapy for management of hypertension in patients on dialysis. Tepel et al. (51) randomized 251 patients with hypertension on hemodialysis to receive amlodipine (5–10 mg/d) or placebo for 30 months. Amlodipine insignificantly improved survival relative to placebo but reduced by 47% the composite secondary end point of all-cause death, nonfatal stroke, MI, coronary revascularization, and angioplasty for peripheral vascular disease (HR, 0.53; 95% CI, 0.31 to 0.93) (51). Other studies suggested that dihydropyridine CCBs are as effective as ACEIs or ARBs in reducing oxidative stress and producing regression of both LVMI and carotid intima-media thickness (52,53). The pharmacokinetics of CCBs are unaltered among patients on dialysis, and these agents are generally nondialyzable (5,33).
Other Agents
The centrally acting α-agonist clonidine can effectively lower BP by reducing autonomic activation. Because of troublesome side effects, such as dry mouth, dizziness, and somnolence, this agent is reserved for patients with difficult to control hypertension (33). Guanfacine is also centrally acting but has less potential for causing sedation or dry mouth. Similarly, the direct vasodilators minoxidil and hydralazine are used as drugs of last resort in patients with multidrug-resistant hypertension. These agents cause reflex tachycardia, the control of which often requires the concomitant administration of β-blockers (33). A serious and not so rare side effect of minoxidil is pericardial effusion. Hydralazine has been associated with a lupus-like syndrome. Lastly, use of α-adrenergic blockers in patients on dialysis is frequently associated with episodes of orthostatic and intradialytic hypotension. These side effects make their use problematic; when chosen, bedtime dosing of these compounds is preferred (33).
Summary of Evidence and Clinical Practice Recommendations
Evidence to support the preferential administration of one antihypertensive drug class over another for pharmacologic management of hypertension in patients on dialysis is currently missing. Choice of the appropriate antihypertensive regimen should be individualized and on the basis of the BP-lowering efficacy of drugs, the BP–independent cardioprotective properties, the intra- and interdialytic pharmacokinetics, and the tolerability and side effect profile as well as the comorbidities of each patient. The use of β-blockers should be strongly considered on the basis of evidence suggesting that these agents likely offer the greatest cardioprotection (15,21,22,44), including their beneficial actions on regression of LVH and arterial stiffness (23,29,45). In the absence of randomized studies to compare the efficacy of different antihypertensive drug classes in patients with specific comorbidities or end organ damage conditions, any established indication of a certain drug class in the general population (i.e., β-blockers for post-MI) should be extrapolated to patients on dialysis as long as firm evidence will become available in this specific population (5). We use dihydropyridine CCBs as second-line agents when BP remains uncontrolled despite nonpharmacologic measures and β-blockade. ACEIs and ARBs are our third choices. Recent studies suggest that, among patients on dialysis, add-on administration of MRAs may offer additive cardiovascular protection without excessive risk of serious hyperkalemia (54). Safety and efficacy of MRAs in patients on dialysis are under investigation in ongoing trials; in the meantime, wide use of spironolactone and eplerenone cannot be recommended. Because management of hypertension often requires combination therapy to achieve adequate BP control, long–acting dihydropyridine CCBs are useful as combination therapy. Other drug classes, such as centrally acting agonists, direct vasodilators, and α-adrenergic blockers, are frequently associated with serious side effects, and their use should be limited to patients with difficult to control hypertension.
Pharmacokinetic properties and dialyzability should be taken into account in the choice and dosing of antihypertensive drugs. The patterns and rhythms of BP during the intra- and interdialytic periods are additional factors that should be considered. In patients prone to intradialytic hypotension, short-acting agents, such as hydralazine, before dialysis should not be administered. There is little support to withhold antihypertensive therapy otherwise before dialysis. Patients experiencing paradoxical BP elevation during or immediately postdialysis also exhibit sustained background hypertension over the entire interdialytic period (55); accordingly, the target of therapy in these patients remains the achievement of adequate 44-hour BP control and not simply the normalization of the intradialytic hemodynamic response (56). Therapeutic approach of intradialytic hypertension is not substantially differentiated from the overall treatment of hypertension in patients on dialysis; dry weight should be probed (57), and if volume control alone is ineffective, antihypertensive drug therapy may be needed. Small studies seem to suggest that improvement in endothelial dysfunction through newer vasodilating β-blockers may be of benefit in these individuals (58). Patients with sustained interdialytic hypertension should receive long-acting agents (56). A maneuver to reduce the high pill burden and enhance the patients’ compliance is supervised thrice weekly administration of long-acting agents immediately postdialysis (59,60). This is strongly supported by the HDPAL Trial, in which administration of atenolol and lisinopril in a three times per week schedule culminated in sufficient home BP control over the course of the trial (15). Patients with a nondipping BP pattern during nighttime may benefit from bedtime dosing of at least one antihypertensive agent. Whether this strategy would reduce cardiovascular outcomes remains unknown.
Perspectives and Directions for Future Research
One critical gap in our knowledge is whether BP control is associated with better outcomes in patients on dialysis. Randomizing patients to two different BP levels using out-of-office BP monitoring to assess cardiovascular outcomes would be a top research priority. It remains unclear whether one class of antihypertensive medications is superior to another in improving cardiovascular outcomes in patients on dialysis. Adequately powered, multicenter, randomized trials evaluating hard cardiovascular end points, ideally with identical BP targets in all treatment arms, are urgently warranted to elucidate the comparative effectiveness of major antihypertensive drug categories. Such types of studies may also clarify whether cardiovascular protection with antihypertensive therapy is BP mediated or whether specific drug categories offer cardioprotection beyond their BP-lowering properties. Another critical issue that remains unclear is the optimal levels at which BP should be targeted in patients on dialysis. Studies randomizing patients on dialysis to different BP goals, preferably with administration of similar BP–lowering agents in the randomized groups, are anticipated to elucidate whether a more aggressive BP–lowering strategy is associated with better or worse outcomes.
Given the high variability of BP among patients on dialysis (8), we need to know whether conventional peridialytic BP recordings are sufficient to guide the management of hypertension in this population. A small study randomized 60 patients on hemodialysis to home BP–guided versus predialysis BP–guided management of hypertension. Patients randomized to home BP–guided therapy achieved a significantly higher reduction in ambulatory systolic BP at study completion (61). Larger studies evaluating additional clinically relevant outcomes can clarify the value of home BP monitoring as a simple approach to optimize the management of hypertension in dialysis.
Disclosures
P.I.G. has no conflicts of interest to disclose. R.A. has consulted for several pharmaceutic companies that make antihypertensive drugs, including Merck, Takeda, Novartis, Daiichi Sankyo, Abbvie, Bayer HealthCare, and Johnson and Johnson.
Acknowledgments
This work was supported by independent research grant 5I01CX000829-04 from the Veterans Affairs Research and Administration (to R.A.).
Footnotes
Published online ahead of print. Publication date available at www.cjasn.org.
See related commentary, “Commentary on Pharmacotherapy of Hypertension in Patients on Chronic Dialysis” on pages 2076–2077.
- Copyright © 2016 by the American Society of Nephrology