Comparative Efficacy and Safety of BP-Lowering Pharmacotherapy In

CJASN ePress. Published on July 16, 2020 as doi: 10.2215/CJN.12201019 Article

Comparative Efﬁcacy and Safety of BP-Lowering Pharmacotherapy in Patients Undergoing Maintenance Dialysis A Network Meta-Analysis of Randomized, Controlled Trials

Ahmed M. Shaman,1,2 Brendan Smyth ,1,3 Clare Arnott,1,4 Suetonia C. Palmer,5 Anastasia S. Mihailidou,6,7 1,8 1,9 1 1 Due to the number of Meg J. Jardine, Martin P. Gallagher, Vlado Perkovic, and Min Jun contributing authors, the affiliations are Abstract listed at the end of Background and objectives Elevated BP is an important risk factor for cardiovascular disease, with a prevalence of this article. over 80% in patients undergoing maintenance dialysis. We assessed the comparative BP-lowering efficacy and the Correspondence: safety of BP-lowering drugs in patients undergoing maintenance dialysis. Dr. Min Jun, The George Institute for Design, settings, participants, & measurements We performed a frequentist random effects network meta-analysis Global Health, The of randomized, controlled trials evaluating BP-lowering agents in adult patients undergoing maintenance University of New dialysis. Electronic databases (CENTRAL, MEDLINE, and Embase) were systematically searched (up to August South Wales, Level 5, 1 King Street, 2018) for relevant trials. The main outcome was systolic BP reduction. Newtown, NSW 2042, Australia. Email: Results Forty trials (4283 participants) met our inclusion criteria. Angiotensin-converting enzyme inhibitors, mjun@georgeinstitute. b-blockers, calcium-channel blockers, and aldosterone antagonists lowered systolic BP to a greater extent than org.au placebo, with effect sizes ranging from 210.8 mm Hg (95% confidence interval, 214.8 to 26.7 mm Hg) for the aldosterone antagonists to 24.3 mm Hg (95% confidence interval, 27.2 to 21.5 mm Hg) for angiotensin- converting enzyme inhibitors. Aldosterone antagonists and b-blockers were superior to angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, calcium-channel blockers, and renin inhibitors at lowering systolic BP. Compared with angiotensin-converting enzyme inhibitors, aldosterone antagonists and b-blockers lowered systolic BP by 6.4 mm Hg (95% confidence interval, 211.4 to 21.4 mm Hg) and 4.4 mm Hg (95% confidence interval, 27.4 to 21.3 mm Hg), respectively. Systolic BP reduction was not different with angiotensin receptor blockers, a-blockers, and calcium-channel blockers compared with angiotensin-converting enzyme inhibitors. Renin inhibitors were less effective. Angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, and aldosterone antagonists incurred risks of drug discontinuation due to adverse events and hypotension.

Conclusions BP-lowering agents signiﬁcantly reduced systolic BP in patients undergoing maintenance dialysis. b-Blockers and aldosterone antagonists may confer larger reductions, although treatment with aldosterone antagonists may be limited by adverse events. CJASN 15: ccc–ccc, 2020. doi: https://doi.org/10.2215/CJN.12201019

Introduction dialysis is challenging as the benefits and tolera- Cardiovascular diseases account for approximately bility of BP-lowering medications may be different half of all deaths in people with kidney failure as has been observed for other cardiovascular receiving maintenance dialysis (1). Elevated BP is medications (5). an important risk factor for cardiovascular disease, A lack of head-to-head trials may limit the capacity and its prevalence rises to 86% among patients of conventional pairwise meta-analyses to evaluate undergoing maintenance dialysis (2). Although two the comparative efficacy and safety of different classes earlier meta-analyses of randomized, controlled of BP-lowering agents. Network meta-analysis uses trials showed cardiovascular and survival benefit direct and indirect evidence to simultaneously com- with BP-lowering pharmacotherapy (3,4), their pare available treatments (6). In this review, we comparative efficacy and safety are not well known examine the comparative efficacy and safety of BP- because these patients have typically been excluded lowering medications to reduce BP in patients with from large trials. Extrapolation of findings ob- kidney failure requiring dialysis using network served within patients not requiring maintenance meta-analysis. www.cjasn.org Vol 15 August, 2020 Copyright © 2020 by the American Society of Nephrology 1 2 CJASN

Materials and Methods Data Synthesis and Analyses We conducted a systematic review according to the We used random effects pairwise (10) and network meta- PRISMA guidelines (7). The study was preregistered with analyses within a frequentist environment to obtain summary PROSPERO (CRD42016035487). estimates of study outcomes. Network meta-analysis combines evidence about treatments from direct head-to-head trials and indirectly from studies that used a common comparator for Data Sources and Searches both treatments. For example, on the basis of direct compar- We searched MEDLINE, Embase, and the Cochrane isons of interventions A versus B and B versus C, one can Central Register of Controlled Trials from database in- investigate the effects of intervention A versus C using indirect ception up to August 2018 using medical subject headings comparisons. The direct and indirect comparisons are then and text words relevant to BP-lowering agents, kidney pooled to form a network effect (9). We assumed that eligible failure, and clinical trials designs (Supplemental Table 1) participants can be randomized to any of the network without date or language restriction. We hand searched interventions. We checked the transitivity assumption by reference lists from retrieved records and clinical trials investigating the distribution of potential effect modifiers registries to identify any additional studies. (age, baseline systolic BP, follow-up duration, sample size, The titles and abstracts of retrieved records were population, and study methodological quality) (11). We as- screened independently by two investigators (C.A. and sessed agreement between direct and indirect estimates in every A.M.S.) according to the review eligibility criteria on the closed loop of evidence using loop-specific and node-splitting basis of a standardized approach. Any disagreement on the approaches and for the entire network using design-by-treat- selection of studies was resolved in discussion with ment interaction model (global inconsistency test) (12,13). additional reviewers (M.J. and V.P.). For continuous outcomes, the mean differences and We included randomized, controlled trials that recruited corresponding 95% confidence intervals (95% CIs) were adults ($18 years old) requiring dialysis and assessed the calculated using end of trial mean values, their correspond- efficacy of any class of BP-lowering therapy: angiotensin- ing SD, and treatment arm size. For crossover trials, we converting enzyme (ACE) inhibitors, angiotensin receptor calculated the SEMs from paired t statistics using a method blockers (ARBs), a-blockers, b-blockers, calcium-channel described elsewhere (14). For dichotomous outcomes, blockers, diuretics, centrally acting vasodilators, aldoste- relative risks (RRs) and corresponding 95% CIs were calcu- rone antagonists, or direct renin inhibitors alone or in lated using total number of patients randomized in each combination compared with control (placebo or standard of group as the denominator. Evidence of statistical heteroge- care) or another class of BP-lowering agents. Trials aiming neity in estimates between studies beyond the level of chance to achieve specific BP targets using a range of therapies was estimated using the I2 statistic in the pairwise meta- were excluded. analysis (15). We investigated possible sources of heterogeneity by comparing summary results obtained from subsets Data Extraction of studies, which compared the efficacy of BP-lowering drugs Two authors (C.A. and A.M.S.) extracted and verified compared with control, grouped by class of BP-lowering data after entry into a spreadsheet. Collected data included drugs, age, proportion of men, baseline systolic BP, dialysis baseline characteristics (age, sex, comorbidities [diabetes, modality, type of BP measurements, median study size, and hypertension, baseline BP, and history of cardiovascular follow-upduration.Weassumedthesameheterogeneity disease], type of dialysis, and drug use); trial data (trial variance (t2) for all comparisons in the network meta-analysis design, sample size, follow-up duration, country, and year (16) and compared its value with outcome- and treatment- of publication); and outcomes data. specific empirical distribution of variances to assess the magnitude of heterogeneity in the entire network (17,18). Publication bias was assessed using comparison-adjusted Quality Assessment funnel plots for all active drugs against control (19). Statistical We assessed risk of bias in included trials using the fi analyses were performed with Stata, version 15.1 (Stata, Cochrane tool (8). Two authors (A.M.S. and B.S.) classi ed College Station, TX) using published methods (20–22). risk of bias for each study domain as low, unclear, or high. Any disagreement in risk of bias assessment was adjudi- cated by additional reviewers (M.J. and V.P.). Grading of Recommendations, Assessment, Development and Evalu- Results ation (GRADE) domains for network meta-analysis pro- Search Results and Characteristics of Included Studies The electronic literature search retrieved 6406 records, of posed by Salanti et al. (9) were used to assess confidence in which 40 randomized trials met our inclusion criteria each treatment comparison estimates. (Supplemental Figure 1). Supplemental Table 2 summarizes the characteristics of included studies, which were reported Outcomes between 1986 and 2016. Among the 40 included trials, 32 We collected data on BP and the method of measure- included patients receiving hemodialysis, 4 enrolled patients ment, heart rate, and adverse events. The main outcome receiving peritoneal dialysis, and 4 included both. Twenty- was mean difference in systolic BP measurements between three trials were placebo controlled, 6 trials compared an assigned treatment groups. Other outcomes included di- intervention with standard of care, and 11 trials were head- astolic BP, heart rate, discontinuation due to adverse to-head trials, of which 2 were three-arm trials. Thirty-seven events, hypotension, hyperkalemia, and serum potassium trials had a parallel design, two were crossover trials, and one concentration. was a two-by-two factorial study. CJASN 15: ccc–ccc, August, 2020 BP Lowering in Kidney Failure, Shaman et al. 3

The trials included a total of 4283 participants (sample size Supplemental Table 3). Of the included studies, 22 were range, 8–469 participants) with a median proportion of men double blinded, 15 were open label, and blinding was not of 61%. Study follow-up duration ranged from 2 weeks to specified in 3 trials. 3.5 years (median 6 months). The median age of trial For the systolic BP outcome, trials with some concerns participants was 57 years (range, 41–71 years). The mean regarding quality contributed 55% of direct evidence to the baseline systolic BP was 148 mm Hg (range, 123–189 mm network of BP-lowering drugs. Trials with no concern Hg). In 26 trials, mean baseline systolic BP was $140 mm Hg. regarding quality contributed 14% of direct evidence, Included trials reported BP measurements in several whereas trials with major concerns contributed 31% of ways: 19 (59%) studies reported predialysis BP measure- the direct evidence in the network (Supplemental Figure 3). ment, 5 (16%) reported clinic BP, 4 (13%) reported home BP, The GRADE confidence rating for effects of treatments 3 (9%) reported ambulatory BP measurement, and 1 (3%) on systolic BP is presented in Table 1. Confidence in trial measured postdialysis BP. network estimates varied; 1 comparison (b-blockers versus placebo) provided high confidence, 6 provided moderate confidence, 12 provided low confidence, and 9 provided Risks of Bias and Confidence Rating very low confidence. Overall, 14 trials (35%) were considered at low risk of There was no evidence of disagreement between direct bias; 10 (25%) were at unclear risk of bias, and 16 trials and indirect evidence for all outcomes, except for heart (40%) were at high risk of bias (Supplemental Figure 2, rate outcome (ACE inhibitors versus placebo; P50.05).

Table 1. Summary of conﬁdence in effect estimates of BP-lowering agents in lowering systolic BP

Type Conﬁdence Comparison Reasons for Downgrading of Evidence Rating

ACE inhibitors versus placebo Mixed Moderate Study limitationsa ARBs versus placebo Mixed Very low Study limitationsa; imprecisionb; inconsistencyc b-Blockers versus placebo Mixed High — Calcium-channel blockers versus placebo Mixed Moderate Study limitationsa Aldosterone antagonists versus placebo Mixed Low Study limitationsa; heterogeneityd a-Blockers versus placebo Indirect Very low Study limitationsa; imprecisionb; inconsistencye Renin inhibitors versus placebo Indirect Very low Study limitationsa; imprecisionb; inconsistencye ACE inhibitors versus ARBs Mixed Low Study limitationsa; imprecisionb ACE inhibitors versus b-blockers Mixed Moderate Study limitationsa ACE inhibitors versus calcium-channel blockers Mixed Moderate Study limitationsa ACE inhibitors versus a-blockers Indirect Very low Study limitationsa; imprecisionb; inconsistencye ACE inhibitors versus aldosterone antagonists Indirect Low Study limitationsa; inconsistencye ACE inhibitors versus renin inhibitors Indirect Low Study limitationsa; inconsistencye ARBs versus calcium-channel blockers Mixed Low Study limitationsa; imprecisionb ARBs versus renin inhibitors Mixed Low Study limitationsa; inconsistencyc ARBs versus a-blockers Indirect Very low Study limitationsa; imprecisionb; inconsistencye ARBs versus b-blockers Indirect Very low Study limitationsa; imprecisionb; inconsistencye ARBs versus aldosterone antagonists Indirect Low Study limitationsa; inconsistencye a-Blockers versus b-blockers Mixed Low Study limitationsa; imprecisionb a-Blockers versus calcium-channel blockers Indirect Very low Study limitationsa; imprecisionb; inconsistencye a-Blockers versus aldosterone antagonists Indirect Very low Study limitationsa; imprecisionb; inconsistencye a-Blockers versus renin inhibitors Indirect Low Study limitationsa; inconsistencye b-Blockers versus calcium-channel blockers Mixed Moderate Study limitationsa b-Blockers versus aldosterone antagonists Indirect Very low Study limitationsa; imprecisionb; inconsistencye b-Blockers versus renin inhibitors Indirect Low Study limitationsa; inconsistencye Calcium-channel blockers versus renin inhibitors Mixed Moderate Study limitationsa Calcium-channel blockers versus aldosterone Indirect Low Study limitationsa; inconsistencye antagonists Aldosterone antagonists versus renin inhibitors Indirect Low Study limitationsa; inconsistencye

ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker. aThe majority of evidence was at high or unclear risk of bias. bConﬁdence intervals for network estimates include values that would favor either treatment. cConﬁdence intervals for direct and indirect evidence provided different interpretation. dSubstantial heterogeneity present between the included trials. eOnly indirect evidence existed, and inconsistency could not be assessed. 4 CJASN

The “design-by-treatment” interaction models found no Hg (95% CI, 27.4 to 21.3 mm Hg), respectively. Systolic BP evidence for global inconsistency for all outcomes except reduction was not detectably different with ARBs, for serum potassium concentration outcome (P,0.001). a-blockers, and calcium-channel blockers compared with Low heterogeneity existed in the network for all outcomes ACE inhibitors. Renin inhibitors were less effective at except for diastolic BP, heart rate (high), and hypotension lowering BP compared with all other classes. (low to moderate) (Supplemental Tables 4 and 5). The comparison-adjusted funnel plot of placebo- controlled trials of BP-lowering drugs’ effect on systolic BP in patients receiving dialysis suggested no evidence of a Effects of BP-Lowering Pharmacotherapy small study effect in the network (Supplemental Figure 4), Systolic BP. Twenty-five randomized trials involving although few trials were available for each drug class. 1681 participants formed the network for the systolic BP We did not observe significant differences in pooled outcome (Figure 1). Four treatments (ACE inhibitors, estimates for systolic BP across any of the subgroups b-blockers, calcium-channel blockers, and aldosterone assessed (P value for heterogeneity for all subgroups antagonists) lowered systolic BP to a greater extent than .0.05) (Figure 3). placebo, with effect sizes ranging from 210.8 mm Hg Diastolic BP and Heart Rate. A total of 22 trials (1553 (95% CI, 214.8 to 26.7 mm Hg) for aldosterone antago- patients) contributed to diastolic BP outcome (Supplemental nists to 24.3 mm Hg (95% CI, 27.2 to 21.5 mm Hg) for Figure 11). b-Blockers showed an effect on diastolic BP com- ACE inhibitors (Figure 2). No separate effect was clearly pared with placebo (24.4 mm Hg; 27.1 to 21.7 mm Hg). identified for ARBs (23.0 mm Hg; 95% CI, 28.7 to 2.6 mm There was no detectable difference between other classes Hg), a-blockers (26.7 mm Hg; 95% CI, 214.1 to 0.7 mm against placebo or each other (Supplemental Figure 5). Hg), or renin inhibitors (6.0 mm Hg; 95% CI, 20.5 The network for the heart rate outcome is shown in to 12.4 mm Hg). Supplemental Figure 11. b-Blockers showed a reduction of Aldosterone antagonists and b-blockers lowered systolic heart rate by 21 beats per minute (95% CI, 226 to 216 beats BP to a greater extent than ACE inhibitors, ARBs, calcium- per minute), a more marked effect than any other class channel blockers, and renin inhibitors (Figure 2). For (Supplemental Figure 6). example, compared with ACE inhibitors, aldosterone Discontinuation Due to Adverse Events. Sixteen trials antagonists and b-blockers lowered systolic BP by (1996 patients) contributed to the network for the risk of 6.4 mm Hg (95% CI, 211.4 to 21.4 mm Hg) and 4.4 mm discontinuation due to adverse events (Figure 1). This risk

Systolic blood pressure Discontinuation due to adverse events

ACE inhibitors α blockers (n = 10) β blockers (n=85) (n = 411) 1 2 Control 1 (n=988) β blockers 1 (n = 224) 3 6 4 Calcium-channel 5 3 3 blockers 1 Aldosterone antagonists (n = 246) ARB 3 (n = 86) Aldosterone antagonists 1 4 (n=417) 1 (n=426) 1 1 2 Renin inhibitors 1 1 (n = 49) Placebo ARB (n = 599) ACE inhibitors ACE inhibitors + ARB (n = 56) (n=105) (n=12)

Hypotension Hyperkalaemia ACE inhibitors ARB (n = 657) 1 (n = 170) Calcium-channel blockers (n = 123) β blockers (n = 100) 2 ARB 1 6 (n = 624) ACE inhibitors 4 Aldosterone antagonists 1 (n = 53) Aldosterone antagonists (n = 469) (n = 155) 2 7

Control Control (n = 994) (n = 1165)

Figure 1. | Diagram showing the networks of treatment comparison of BP-lowering agents for systolic BP and safety outcomes in the identiﬁed trials. The size of the node corresponds to the number of trials. The thickness of the line connecting two treatments corresponds to the number of patients. Numbers next to each line represent the number of trials that compared the connected treatments. Numbers in brackets correspond to the number of patients. ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker. CJASN 15: ccc–ccc, August, 2020 BP Lowering in Kidney Failure, Shaman et al. 5

Network Direct Patients Trials 2 Treatment vs comparator MD (95% CI) MD (95% CI) (N) (N) I BP lowering agents vs placebo ACE inhibitors -4.3 (-7.2, -1.5) -4.0 (-7.7, -0.3) 602 5 0.0% ARB -3.0 (-8.7, 2.6) 1.5 (-8.1, 11.1) 70 2 0.0% α blockers -6.7 (-14.1, 0.7) β blockers -8.7 (-10.9, -6.4) -8.9 (-11.9, -5.8) 192 3 0.0% Calcium-channel blockers -4.6 (-7.0, -2.2) -7.8 (-16.2, 0.5) 177 3 73.5% Aldosterone antagonists -10.8 (-14.8, -6.7) -8.5 (-16.4, -0.5) 162 4 66.3% Renin inhibitors 6.0 (-0.5,12.4) BP lowering agents vs ACE inhibitors ARB 1.3 (-4.6, 7.1) 1.9 (-7.1, 10.9) 24 1 α blockers -2.4 (-10.0, 5.3) β blockers -4.4 (-7.4, -1.3) -4.7 (-11.8, 2.4) 102 1 Calcium-channel blockers -0.3 (-3.5, 2.8) 1.3( -5.3, 7.9) 132 3 0.0% Aldosterone antagonists -6.4 (-11.4, -1.4) Renin inhibitors 10.3 (3.6,16.9) BP lowering agents vs ARB α blockers -3.6 (-12.7, 5.4) β blockers -5.6 (-11.4, 0.1) Calcium-channel blockers -1.6 (-7.2, 4.0) 0.3 (-9.2, 9.8) 24 1 Aldosterone antagonists -7.7 (-14.7, -0.8) Renin inhibitors 9.0 (3.4, 14.6) 10.6 (4, 17.2) 18 1 BP lowering agents vs α blockers β blockers -2.0 (-9.1, 5.1) -2.0 (-9.1, 5.1) 20 1 Calcium-channel blockers 2.0 (-5.3, 9.4) Aldosterone antagonists -4.1 (-12.6, 4.4) Renin inhibitors 12.6 (3.1, 22.2) BP lowering agents vs β blockers Calcium-channel blockers 4.1 (2.1, 6.0) 3.9 (1.7, 6.1) 120 1 Aldosterone antagonists -2.1 (-6.7, 2.6) Renin inhibitors 14.6 (8.2, 21.0) BP lowering agents vs Calcium-channel blockers Aldosterone antagonists -6.1 (-10.9, -1.4) Renin inhibitors 10.6 (4.4, 16.8) 8.0 (-0.2, 16.2) 74 1 BP lowering agent vs aldosterone antagonists Renin inhibitors 16.7 (9.1, 24.3) Overall BP lowering agents vs Placebo -6.4 (-9.4, -3.5) -6.4 (-9.2, -3.6) 1203 17 47.7%

-15 -10 -5 0 5 10 25 Favours Treatment Favours Comparator

Figure 2. | BP-lowering agents lowered systolic BP (millimeters of mercury) to a greater extent than placebo, and aldosterone antagonists and b-blockers may confer larger reductions. ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; 95% CI, 95% confidence interval; MD, mean difference. was significantly increased with aldosterone antagonists potassium concentrations by 0.32 meq/L (95% CI, 0.15 to (RR, 3.35; 95% CI, 1.32 to 8.49), ACE inhibitors (RR, 1.77; 0.49 meq/L) compared with control, by 0.40 meq/L (95% 95% CI, 1.09 to 2.87), and ARBs (RR, 1.57; 95% CI, 0.96 to CI, 0.08 to 0.71 meq/L) compared with ACE inhibitors, and 2.57) compared with control (Figure 4, Supplemental by 0.54 meq/L (95% CI, 0.20 to 0.88 meq/L) compared with Figure 7), whereas there was no detectably increased risk ARBs (Supplemental Figure 10). with b-blockers (RR, 0.81; 95% CI, 0.27 to 2.39). Limited data were available for other classes. Hypotension. Ten trials (1980 patients) contributed to Sensitivity Analyses the network for hypotension (Figure 1). Compared with Because of heterogeneity in some loops of evidence control, ACE inhibitors (RR, 6.62; 95% CI, 1.48 to 29.54) and involving calcium-channel blockers for systolic BP outcome ARBs (RR, 1.53; 95% CI, 0.94 to 2.48) increased the risk of (Supplemental Table 4), we removed the study by London hypotension (Figure 4). Calcium-channel blockers (RR, et al. (23) (baseline mean systolic BP was 189 mm Hg). 0.09; 95% CI, 0.02 to 0.50) and aldosterone antagonists (RR, Although loop-specific heterogeneity (t2)inallloops 0.18; 95% CI, 0.03 to 0.90) were superior to ACE inhibitors dropped to zero, the treatment estimates did not change. in this regard (Supplemental Figure 8). Limited data were In addition, for the serum potassium concentration out- available for other classes. come, we split the control node to placebo and standard of Hyperkalemia and Serum Potassium Concen- care separately. The global inconsistency test showed a tration. Fourteen trials (2514 patients) contributed to the nonsignificant P value of 0.91 with no changes in treatment network for hyperkalemia (Figure 1). No clear differ- estimates. In further sensitivity analysis, we excluded trials ences between medication classes for hyperkalemia risk with unclear or high risk of bias from the systolic BP were observed (Figure 4, Supplemental Figure 9). network, and overall results remained largely unchanged, In a network of 15 trials (959 patients) (Supplemental although some comparisons did not reach statistical sig- Figure 11), aldosterone antagonists increased serum nificance (Supplemental Figure 12). 6 CJASN

Subgroups MD (95%CI) p for heterogeneity Age over median (55 years) Yes -6.6 (-11.3, -2.0) 0.79 No -5.8 (-8.8, -2.8) Type of systolic BP measurement Pre-dialysis -5.3 (-8.5, -2.1) 0.20 Clinic -6.1 (-14.2, 2.1) Post dialysis -7.0 (-22.4, 8.4) Home -15.0 (-20.6, -9.4) ABPM -8.0 (-22.3, 6.3) Cardiovascular disease over median (33%) Yes -7.9 (-11.2, -4.6) 0.074 No -3.1(-6.3, 0.2) Diabetes mellitus over median (30%) Yes -7.0 (-10.0, -4.0) 0.45 No -4.2 (-10.6, 2.1) Baseline systolic BP (mmHg) 140 or more -5.9 (-10.2, -1.6) 0.83 Less than 140 -7.9 (-10.5, -5.3) Dialysis modality Haemodialysis -5.4 (-8.2, -2.6) 0.11 Mixed -12.9 (-17.7, -8.2) Study period over median (24 weeks) Yes -4.5 (-7.7, -1.2) 0.076 No -9.7 (-14.7, -4.7) Male over median (60%) Yes -5.1 (-7.9, -2.3) 0.10 No -9.8 (-15.8, -3.7) Risk of bias in included trials Low risk -6.7 (-10.2, -3.3) 0.97 Unclear risk -6.8 (-11.4, -2.2) High risk 0.0 (-19.4, 19.4) Sample size over median (48) Yes -6.0 (-9.5, -2.5) 0.91 No -7.3 (-12.5, -2.1) Year of publication Before 2006 -8.3 (-11.8, -4.8) 0.094 After 2006 -4.4 (-8.8, -0.1)

-20 020 favours Treatment favours Comparator

Figure 3. | Subgroup analyses for the effects of BP-lowering agents on systolic BP outcome compared with placebo showed no signiﬁcant differences in pooled estimates across the assessed groups. ABPM, ambulatory BP monitoring; 95% CI, 95% conﬁdence interval; MD, mean difference (millimeters of mercury).

Discussion irrespective of volume status, and increased sympa- This meta-analysis provides evidence that should guide thetic drive may be important mechanisms underpinning the use of BP-lowering agents in patients undergoing the increased BP observed in people with kidney failure, maintenance dialysis. The pooled analysis showed an providing a potential rationale for superior efficacy of overall significant effect of BP-lowering agents in lowering aldosterone antagonists and b-blockers in people receiving systolic BP and suggests that aldosterone antagonists and dialysis (26–28). It is noted, however, that our findings on b-blockers may produce greater reductions in systolic BP. the effects of aldosterone antagonists conflict with those The data also suggest that ACE inhibitors and calcium- reported in two recent trials (29,30), which showed no effect channel blockers have important BP-lowering effects. The on systolic BP with spironolactone compared with placebo. effects of a-blockers and ARBs were less precise. These data It is possible that smaller size and suboptimal quality of suggest that b-blockers and perhaps aldosterone antago- earlier studies may have contributed to an overestimation nists may be considered as BP-lowering agents of choice of the treatment effect. Caution is thus warranted when where they are tolerated for people with kidney failure interpreting these findings. The two ongoing clinical requiring maintenance dialysis. trials, ALCHEMIST (NCT01848639) and ACHIEVE There seem to be differences in the BP-lowering efficacy (NCT03020303), should help better define the effective- of different drug classes. Specifically, aldosterone antago- ness and safety of spironolactone in patients undergoing nists and b-blockers appear superior to other classes of BP- maintenance dialysis. lowering drugs at lowering systolic BP, whereas the effects Our study supports the use of b-blockers to lower BP in of ACE inhibitors and ARBs appear less potent. There is a patients undergoing maintenance dialysis. However, potential pathophysiologic rationale for reduced efficacy of b-blockers may be underused in clinical practice (31), agents targeting the renin-angiotensin system because renin and because included trials in our analysis are relatively is produced by the kidney, and levels may be lower in people small, future research is needed to evaluate the use of with kidney failure (24,25). Conversely, both increased b-blockers as first-line BP-lowering agents in this patient aldosterone levels (so called “relative hyperaldosteronism”), population. In addition, water-soluble b-blockers are CJASN 15: ccc–ccc, August, 2020 BP Lowering in Kidney Failure, Shaman et al. 7

Network Direct Patients Trials Safety outcomes/ Comparison 2 RR (95% CI) RR (95% CI) (N) (N) I

Discontinuation due to adverse events ACE inhibitor 1.77 (1.09, 2.87) 1.68 (1.03, 2.74) 183 3 0.0% Angiotensin receptor blocker 1.57 (0.96, 2.57) 1.66 (1.01, 2.73) 800 4 0.0% β blocker 0.81 (0.27, 2.39) 1.55 (0.11, 22.1) 163 2 67.8% Aldosterone antagonist 3.35 (1.32, 8.49) 4.05 (1.16, 14.20) 817 5 39.2% ACE inhibitor +ARB 0.90 (0.15, 5.31) All BP lowering agents vs control 1.70 (1.24, 2.33) 1.87 (1.20, 2.91) 1981 15 22.5%

Hyperkalaemia ACE inhibitor 1.03 (0.11, 9.55) 1.03 (0.11, 9.55) 137 2 0.0% Angiotensin receptor blocker 1.17 (0.63, 2.15) 1.16 (0.68, 2) 1249 4 0.0% β blocker 0.31 (0.02, 4.00) Aldosterone antagonist 1.63 (0.75, 3.57) 1.85 (0.89, 3.85) 928 7 9.1% All BP lowering agents vs control 1.37 (0.86, 2.21) 1.37 (0.94, 1.99) 2314 13 0.0%

Hypotension ACE inhibitor 6.62 (1.48, 29.54) 6.62 (1.57, 27.94) 107 1 Angiotensin receptor blocker 1.53 (0.94, 2.48) 1.41 (1.03, 1.94) 1310 6 0.0% Calcium-channel blocker 0.59 (0.24, 1.41) 0.59 (0.27, 1.28) 251 1 Aldosterone antagonist 1.17 (0.61, 2.22) 1.16 (0.66, 2.04) 312 2 0.0% All BP lowering agents vs control 1.42 (0.93, 2.17) 1.41 (0.97, 2.05) 1980 10 36.2%

0.02 0.5 1 2 4 30 Favours Treatment Favours Control

Figure 4. | A number of BP lowering agents were associated with adverse events compared with control, including treatment discontinuation and hypotension. ACE, angiotensin-converting enzyme; ARB, angiotensin receptor blocker; 95% CI, 95% confidence interval; RR, relative risk. dialyzable, and they need to be supplemented after di- potassium binders is a potentially practice-changing alysis. This is important because observational evidence strategy, which may facilitate the use of BP-lowering suggested possible harm with dialyzable compared with agents that might increase the risk of hyperkalemia as nondialyzable b-blockers (32). observed with patiromer in the phase 2 AMBER trial (42). Volume control is important for BP management. The superiority of b-blockers and aldosterone antago- Achieving dry weight or increased dialysis frequency nists in systolic BP lowering may translate to better and/or time have been shown to lower BP in patients cardiovascular morbidity and mortality risk reduction. undergoing hemodialysis (33–37). However, because few HDPAL study (43) showed that lisinopril was associated data were available, it is not clear how volume control in with 2.29 times higher cardiovascular events compared the included trials could potentially affect or modify BP- with atenolol in 200 predominately black patients under- lowering effects of BP-lowering drugs. going maintenance hemodialysis. However, both lisinopril It is important to consider the potential risks of BP- and atenolol improved left ventricular mass index, and the lowering agents in this population. We found the risk of BP-lowering effect was not significantly different. In our drug discontinuation due to adverse events was increased, analysis, b-blockers were superior at lowering systolic BP especially for ACE inhibitors, ARBs, and aldosterone (by 4.4 mm Hg) compared with ACE inhibitors (Figure 2), antagonists. Hypotension risk was increased with ACE consistent with these findings. Other meta-analyses report- inhibitors and ARBs, whereas aldosterone antagonists ed cardiovascular and mortality benefits with b-blockers increased serum potassium concentrations but not hyper- (39) and aldosterone antagonists (44), but no clear effects on kalemia. b-Blockers may also increase hyperkalemia and majorcardiovascularevents and mortality with ACE hypotension risk, but data were limited to evaluate these inhibitors and ARBs were observed in patients receiving outcomes in our study (38,39). These insights into the maintenance dialysis (45). nature of adverse outcomes may be particularly important This study has several strengths. It aggregates a com- in patients undergoing maintenance dialysis, in whom prehensive, methodologically rigorous overview of the intradialytic hypotension, hyperkalemia, and the use of available data, incorporating direct and indirect treatment low dialysate potassium have all been associated with effects of BP-lowering drugs against placebo or against worse outcomes (40,41). On a related note, the use of each other. It includes a substantial number of studies and a 8 CJASN

much larger number of participants than any previous support paid to her institution. Dr. Jun reports receiving grant study of BP lowering in patients undergoing maintenance support from National Health and Medical Research Council of dialysis. It also has some important limitations. As with Australia project grant 1148060 and unrestricted grant support from any systematic review, it is dependent on the quality of the VentureWise (a wholly owned commercial subsidiary of NPSMedi- included studies, which varied substantially as did the size cineWise) to conduct a commissioned project funded by AstraZeneca. and design of the included trials. Most of the data in the Dr. Perkovic has participated in advisory boards for both Relypsa and network for the outcome of systolic BP were derived from AstraZeneca and has received honoraria and consultation fees from trials with unclear or high risk of bias, although overall Tricida, Novartis, Amgen, Janssen, GlaxoSmithKline, Astellas, Boer- findings remained largely unchanged after excluding stud- ingerIngelheim,Baxter,MitsubishiTanabe,Retrophin,Merck,Abbvie, ies with unclear or high risk of bias. There were relatively Novo Nordisk, AstraZeneca,Gilead,Durect, Servier, Eli Lilly, Relypsa, few head-to-head studies, leading to limited network Pharmalink, Bayer, Bristol-Myers Squibb, and Tufts, with payments connectivity and power to detect statistical inconsistencies paid to his institution. All remaining authors have nothing to disclose. and publication bias. Limited data may have also affected the stability of the point estimates, including those reported Funding for aldosterone antagonists (46). The majority of trials were None. reporting predialysis BP measurements, and thus, subgroup analysis may be underpowered to detect important differ- Supplemental Material This article contains the following supplemental material online at ences between different BP measurements methods. This also limited our ability to draw conclusions regarding BP http://cjasn.asnjournals.org/lookup/suppl/doi:10.2215/CJN. variability, a predictor of adverse outcomes in patients with 12201019/-/DCSupplemental. kidney failure (47,48). Adverse events were also inconsis- Supplemental Figure1. Electronic searches in MEDLINE, Embase, and the Cochrane Central Register of Randomised Trials for trials on tently captured and reported, and thus, they should be BP-lowering drugs in adults with kidney failure requiring dialysis. interpreted with caution. Because included trials used fi different agents from different classes, we could not assess Supplemental Figure 2. Domain-speci c risk of bias assessment summary of included studies. the effect of the doses on transitivity assumption. However, most of the included agents had doses that are within the Supplemental Figure 3. Study limitations weighted by contribu- therapeutic ranges for BP control (49). Lastly, this analysis tion of direct estimates to the network of BP-lowering drugs for systolic BP outcome. focused on BP effects and did not evaluate the effects of BP- Supplemental Figure 4. Comparison-adjusted funnel plot of lowering agents on patient-reported outcomes. n5 In conclusion, our study provides important evidence placebo-controlled trials ( 17) of BP-lowering drugs effect on regarding the comparative efficacy and safety of different systolic BP in patients on dialysis. classes of BP-lowering agents in managing BP in patients Supplemental Figure 5. Network estimates of BP-lowering drugs undergoing maintenance dialysis. It suggests, within the effects on diastolic BP. limitations of the available evidence, that b-blockers and Supplemental Figure 6. Network estimates of BP-lowering drugs effects on heart rate. aldosterone antagonists may have greatest systolic BP- lowering effect, although aldosterone antagonists have a Supplemental Figure 7. Network estimates of BP-lowering drugs higher risk of adverse effects. Therefore, b-blockers should effects on discontinuation due to adverse events. Supplemental Figure 8. Network estimates of BP-lowering drugs be evaluated in further research as the first class of agents considered for the treatment of hypertension in this pa- effects on risk of hypotension. tient population. Supplemental Figure 9. Network estimates of BP-lowering drugs effects on the risk of hyperkalemia. Supplemental Figure 10. Network estimates of BP-lowering drugs Acknowledgments effects on serum potassium. Dr. Perkovic and Dr. Shaman designed the study; Dr. Gallagher, Supplemental Figure 11. Network of treatment comparison of BP- Dr. Jardine, Dr. Jun, and Dr. Mihailidou revised the protocol; lowering agents effects on diastolic BP, heart rate, and potassium Dr. Shaman carried out the literature search; Dr. Arnott and Dr. concentration outcomes. Shaman independently screened trials for inclusion and extracted Supplemental Figure 12. Network estimates of BP-lowering drugs fi and veri ed data; Dr. Shaman and Dr. Smyth independently as- effects on systolic BP in trials with low risk of bias (n59). sessed the quality of included trials; Dr. Shaman analyzed the data Supplemental Material. References. fi and prepared gures and tables with support from Dr. Palmer; Supplemental Table 1. Electronic search terms. Dr. Shaman drafted the paper; Dr. Jun, Dr. Palmer, and Dr. Perkovic Supplemental Table 2. Characteristics of included studies. revised the paper; and Dr. Arnott, Dr. Gallagher, Dr. Jardine, Dr. Jun, Supplemental Table 3. Risk of bias assessments of all in- Dr. Mihailidou, Dr. Palmer, Dr. Perkovic, Dr. Shaman, and Dr. Smyth cluded trials. fi approved the nal version of the manuscript. Supplemental Table 4. Assessment of loop-specific (in closed loops of evidence) and overall inconsistency and heterogeneity. Disclosures Supplemental Table 5. Assessment of agreement between direct Dr. Gallagher reports receiving honoraria from Shire and Amgen and indirect evidence using side-splitting approach for all outcomes. for speaking at scientific meetings. Dr. Jardine is responsible for research projects that have received unrestricted funding from Gambro, Baxter, CSL, Amgen, Eli Lilly, and Merck; has served on References advisory boards sponsored by Akebia, Baxter, and Boehringer In- 1. Saran R, Robinson B, Abbott KC, Agodoa LYC, Bhave N, Bragg- fi gelheim; and has spoken at scienti c meetings sponsored by Janssen, Gresham J, Balkrishnan R, Dietrich X, Eckard A, Eggers PW, Amgen, and Roche, with any consultancy, honoraria, or travel Gaipov A, Gillen D, Gipson D, Hailpern SM, Hall YN, Han Y, He CJASN 15: ccc–ccc, August, 2020 BP Lowering in Kidney Failure, Shaman et al. 9

K, Herman W, Heung M, Hirth RA, Hutton D, Jacobsen SJ, Jin Y, 19. Chaimani A, Salanti G: Using network meta-analysis to evaluate Kalantar-Zadeh K, Kapke A, Kovesdy CP, Lavallee D, Leslie J, the existence of small-study effects in a network of interventions. McCullough K, Modi Z, Molnar MZ, Montez-Rath M, Moradi H, Res Synth Methods 3: 161–176, 2012 Morgenstern H, Mukhopadhyay P, Nallamothu B, Nguyen DV, 20. Harris RJ, Deeks JJ, Altman DG, Bradburn MJ, Harbord RM, Sterne Norris KC, O’Hare AM, Obi Y, Park C, Pearson J, Pisoni R, JA: Metan: Fixed-and random-effects meta-analysis. Stata J 8: Potukuchi PK, Rao P, Repeck K, Rhee CM, Schrager J, Schaubel 3–28, 2008 DE, Selewski DT, Shaw SF, Shi JM, Shieu M, Sim JJ, Soohoo M, 21. Liu Y, Wang W, Zhang AB, Bai X, Zhang S: Epley and Semont Steffick D, Streja E, Sumida K, Tamura MK, Tilea A, Tong L, Wang maneuvers for posterior canal benign paroxysmal positional D, Wang M, Woodside KJ, Xin X, Yin M, You AS, Zhou H, vertigo: A network meta-analysis. Laryngoscope 126: 951–955, Shahinian V: US Renal Data System 2017 annual data report: 2016 Epidemiology of kidney disease in the United States [published 22. Chaimani A, Salanti G: Visualizing assumptions and results in correction appears inAmJKidneyDis 71:501,2018].AmJKidney network meta-analysis: The network graphs package. Stata J 15: Dis 71[Suppl 1]: A7, 2018 905–950, 2015 2. Agarwal R, Nissenson AR, Batlle D, Coyne DW,Trout JR, Warnock 23. London GM, Marchais SJ, Guerin AP,Metivier F,Safar ME, Fabiani DG: Prevalence, treatment, and control of hypertension in F, Froment L: Salt and water retention and calcium blockade in chronic hemodialysis patients in the United States. Am J Med 115: uremia. Circulation 82: 105–113, 1990 291–297, 2003 24. Man in ’t Veld AJ, Schicht IM, Derkx FH, de Bruyn JH, Schalekamp 3. Heerspink HJL, Ninomiya T, Zoungas S, de Zeeuw D, Grobbee MA: Effects of an angiotensin-converting enzyme inhibitor DE, Jardine MJ, Gallagher M, Roberts MA, Cass A, Neal B, (captopril) on blood pressure in anephric subjects. BMJ 280: Perkovic V: Effect of lowering blood pressure on cardiovascular 288–290, 1980 events and mortality in patients on dialysis: A systematic review 25. Leslie BR, Case DB, Sullivan JF,Vaughan ED Jr.: Absence of blood- and meta-analysis of randomised controlled trials. Lancet 373: pressure lowering effect of captopril in anephric patients. BMJ 1009–1015, 2009 280: 1067–1068, 1980 4. Agarwal R, Sinha AD: Cardiovascular protection with antihy- 26. Hung SC, Lin YP, Huang HL, Pu HF, Tarng DC: Aldosterone and pertensive drugs in dialysis patients: Systematic review and meta- mortality in hemodialysis patients: Role of volume overload. PLoS analysis. Hypertension 53: 860–866, 2009 One 8: e57511, 2013 5. Palmer SC, Navaneethan SD, Craig JC, Johnson DW, Perkovic V, 27. Converse RL Jr., Jacobsen TN, Toto RD, Jost CM, Cosentino F, Nigwekar SU, Hegbrant J, Strippoli GF: HMG CoA reductase Fouad-Tarazi F, Victor RG: Sympathetic overactivity in patients inhibitors (statins) for dialysis patients. Cochrane Database Syst with chronic renal failure. N Engl J Med 327: 1912–1918, 1992 Rev 9: CD004289, 2013 28. Bomback AS: Mineralocorticoid receptor antagonists in end- 6. Mills EJ, Thorlund K, Ioannidis JP: Demystifying trial networks and stage renal disease: Efficacy and safety. Blood Purif 41: 166–170, network meta-analysis. BMJ 346: f2914, 2013 2016 7. Hutton B, Salanti G, Caldwell DM, Chaimani A, Schmid CH, 29. Charytan DM, Himmelfarb J, Ikizler TA, Raj DS, Hsu JY,Landis JR, Cameron C, Ioannidis JP, Straus S, Thorlund K, Jansen JP, Mulrow Anderson AH, Hung AM, Mehrotra R, Sharma S, Weiner DE, C, Catala´-Lo´pez F, Gøtzsche PC, Dickersin K, Boutron I, Altman Williams M, DiCarli M, Skali H, Kimmel PL, Kliger AS, Dember DG, Moher D: The PRISMA extension statement for reporting of LM; Hemodialysis Novel Therapies Consortium: Safety and car- systematic reviews incorporating network meta-analyses of diovascular efficacy of spironolactone in dialysis-dependent health care interventions: Checklist and explanations. Ann Intern ESRD (SPin-D): A randomized, placebo-controlled, multiple Med 162: 777–784, 2015 dosage trial. Kidney Int 95: 973–982, 2019 8. Higgins JP, Altman DG, Gøtzsche PC, Ju¨ni P, Moher D, Oxman 30. Hammer F, Malzahn U, Donhauser J, Betz C, Schneider MP, AD, Savovic J, Schulz KF, Weeks L, Sterne JA; Cochrane Bias Grupp C, Pollak N, Sto¨rk S, Wanner C, Krane V; MiREnDa Study Methods Group; Cochrane Statistical Methods Group: The Co- Group: A randomized controlled trial of the effect of spi- chrane Collaboration’s tool for assessing risk of bias in rando- ronolactone on left ventricular mass in hemodialysis patients. mised trials. BMJ 343: d5928, 2011 Kidney Int 95: 983–991, 2019 9. Salanti G, Del Giovane C, Chaimani A, Caldwell DM, Higgins JP: 31. Weir MA, Herzog CA: Beta blockers in patients with end-stage Evaluating the quality of evidence from a network meta-analysis. renal disease: Evidence-based recommendations. Semin Dial 31: PLoS One 9: e99682, 2014 219–225, 2018 10. DerSimonian R, Laird N: Meta-analysis in clinical trials. Control 32. Weir MA, Dixon SN, Fleet JL, Roberts MA, Hackam DG, Oliver Clin Trials 7: 177–188, 1986 MJ, Suri RS, Quinn RR, Ozair S, Beyea MM, Kitchlu A, Garg AX: 11. Salanti G: Indirect and mixed-treatment comparison, network, or b multiple-treatments meta-analysis: Many names, many benefits, -Blocker dialyzability and mortality in older patients receiving many concerns for the next generation evidence synthesis tool. hemodialysis. J Am Soc Nephrol 26: 987–996, 2015 Res Synth Methods 3: 80–97, 2012 33. Agarwal R, Alborzi P, Satyan S, Light RP: Dry-weight reduction in 12. Veroniki AA, Vasiliadis HS, Higgins JP, Salanti G: Evaluation of hypertensive hemodialysis patients (DRIP): A randomized, con- inconsistency in networks of interventions. Int J Epidemiol 42: trolled trial. Hypertension 53: 500–507, 2009 332–345, 2013 34. Rocco MV, Lockridge RS Jr., Beck GJ, Eggers PW, Gassman JJ, 13. Dias S, Welton NJ, Caldwell DM, Ades AE: Checking consistency Greene T, Larive B, Chan CT, Chertow GM, Copland M, Hoy CD, in mixed treatment comparison meta-analysis. Stat Med 29: Lindsay RM, Levin NW, Ornt DB, Pierratos A, Pipkin MF, 932–944, 2010 Rajagopalan S, Stokes JB, Unruh ML, Star RA, Kliger AS, Kliger A, 14. Elbourne DR, Altman DG, Higgins JP, Curtin F, Worthington HV, Eggers P,Briggs J, Hostetter T,Narva A, Star R, Augustine B, Mohr P, Vail A: Meta-analyses involving cross-over trials: Methodological Beck G, Fu Z, Gassman J, Greene T, Daugirdas J, Hunsicker L, issues. Int J Epidemiol 31: 140–149, 2002 Larive B, Li M, Mackrell J, Wiggins K, Sherer S, Weiss B, 15. Higgins JP, Thompson SG, Deeks JJ, Altman DG: Measuring in- Rajagopalan S, Sanz J, Dellagrottaglie S, Kariisa M, Tran T, West J, consistency in meta-analyses. BMJ 327: 557–560, 2003 Unruh M, Keene R, Schlarb J, Chan C, McGrath-Chong M, Frome 16. White IR, Barrett JK, Jackson D, Higgins JP: Consistency and in- R, Higgins H, Ke S, Mandaci O, Owens C, Snell C, Eknoyan G, consistency in network meta-analysis: Model estimation using Appel L, Cheung A, Derse A, Kramer C, Geller N, Grimm R, multivariate meta-regression. Res Synth Methods 3: 111–125, Henderson L, Prichard S, Roecker E, Rocco M, Miller B, Riley J, 2012 Schuessler R, Lockridge R, Pipkin M, Peterson C, Hoy C, Fensterer 17. Rhodes KM, Turner RM, Higgins JP: Predictive distributions were A, Steigerwald D, Stokes J, Somers D, Hilkin A, Lilli K, Wallace W, developed for the extent of heterogeneity in meta-analyses of Franzwa B, Waterman E, Chan C, McGrath-Chong M, Copland M, continuous outcome data. J Clin Epidemiol 68: 52–60, 2015 Levin A, Sioson L, Cabezon E, Kwan S, Roger D, Lindsay R, Suri R, 18. Turner RM, Davey J, Clarke MJ, Thompson SG, Higgins JP: Pre- Champagne J, Bullas R, Garg A, Mazzorato A, Spanner E, Rocco dicting the extent of heterogeneity in meta-analysis, using em- M, Burkart J, Moossavi S, Mauck V, Kaufman T, Pierratos A, Chan pirical data from the Cochrane Database of Systematic Reviews. W, Regozo K, Kwok S; Frequent Hemodialysis Network (FHN) Int J Epidemiol 41: 818–827, 2012 Trial Group: The effects of frequent nocturnal home hemodialysis: 10 CJASN

The frequent hemodialysis network nocturnal trial. Kidney Int 80: lisinopril: A randomized controlled trial. Nephrol Dial Transplant 1080–1091, 2011 29: 672–681, 2014 35. Jardine MJ, Zuo L, Gray NA, de Zoysa JR, Chan CT, Gallagher 44. Quach K, Lvtvyn L, Baigent C, Bueti J, Garg AX, Hawley C, Haynes MP, Monaghan H, Grieve SM, Puranik R, Lin H, Eris JM, Zhang R, Manns B, Perkovic V, Rabbat CG, Wald R, Walsh M: The safety L, Xu J, Howard K, Lo S, Cass A, Perkovic V; ACTIVE Dialysis and efficacy of mineralocorticoid receptor antagonists in patients Steering Committee; Paul: A trial of extending hemodialysis who require dialysis: A systematic review and meta-analysis. Am hours and quality of life. J Am Soc Nephrol 28: 1898–1911, J Kidney Dis 68: 591–598, 2016 2017 45. Liu Y, Ma X, Zheng J, Jia J, Yan T: Effects of angiotensin-converting 36. Chertow GM, Levin NW, Beck GJ, Depner TA, Eggers PW, enzyme inhibitors and angiotensin receptor blockers on cardio- Gassman JJ, Gorodetskaya I, Greene T, James S, Larive B, Lindsay vascular events and residual renal function in dialysis patients: A RM, Mehta RL, Miller B, Ornt DB, Rajagopalan S, Rastogi A, meta-analysis of randomised controlled trials. BMC Nephrol 18: Rocco MV,Schiller B, Sergeyeva O, Schulman G, Ting GO, Unruh 206, 2017 ML, Star RA, Kliger AS, Kliger AS; FHN Trial Group: In-center 46. Brignardello-Petersen R, Murad MH, Walter SD, McLeod S, hemodialysis six times per week versus three times per week Carrasco-Labra A, Rochwerg B, Schu¨nemann HJ, Tomlinson G, [published correction appears in N Engl J Med 364: 93, 2011]. Guyatt GH; GRADE Working Group: GRADE approach to rate the N Engl J Med 363: 2287–2300, 2010 certainty from a network meta-analysis: Avoiding spurious 37. Culleton BF, Walsh M, Klarenbach SW, Mortis G, Scott-Douglas judgments of imprecision in sparse networks. J Clin Epidemiol N, Quinn RR, Tonelli M, Donnelly S, Friedrich MG, Kumar A, 105: 60–67, 2019 Mahallati H, Hemmelgarn BR, Manns BJ: Effect of frequent 47. Sarafidis PA, Persu A, Agarwal R, Burnier M, de Leeuw P, Ferro C, nocturnal hemodialysis vs conventional hemodialysis on left Halimi JM, Heine G, Jadoul M, Jarraya F, Kanbay M, Mallamaci F, ventricular mass and quality of life: A randomized controlled trial. Mark PB, Ortiz A, Parati G, Pontremoli R, Rossignol P, Ruilope L, JAMA 298: 1291–1299, 2007 Van der Niepen P,Vanholder R, Verharr MC, Wiecek A, Wuerzner b 38. Shroff GR, Herzog CA: -Blockers in dialysis patients: A G, London GM, Zoccali C: Hypertension in dialysis patients: A nephrocardiology perspective. J Am Soc Nephrol 26: 774–776, consensus document by the European renal and cardiovascular 2015 medicine (EURECA-m) working group of the European renal as- 39. Badve SV, Roberts MA, Hawley CM, Cass A, Garg AX, sociation - European dialysis and transplant association (ERA- Krum H, Tonkin A, Perkovic V: Effects of beta-adrenergic an- EDTA) and the hypertension and the kidney working group of the tagonists in patients with chronic kidney disease: A systematic Europeansocietyofhypertension(ESH). JHypertens35: 657–676, review and meta-analysis. J Am Coll Cardiol 58: 1152–1161, 2017 2011 48. Sarafidis PA, Loutradis C, Karpetas A, Tzanis G, Bikos A, Raptis V, 40. Stefańsson BV, Brunelli SM, Cabrera C, Rosenbaum D, Anum E, Syrgkanis C, Liakopoulos V, Papagianni A, Bakris G, Parati G: The Ramakrishnan K, Jensen DE, Sta˚lhammar NO: Intradialytic hy- association of interdialytic blood pressure variability with car- potension and risk of cardiovascular disease. Clin J Am Soc diovascular events and all-cause mortality in haemodialysis pa- Nephrol 9: 2124–2132, 2014 tients. Nephrol Dial Transplant 34: 515–523, 2019 41. Karaboyas A, Zee J, Brunelli SM, Usvyat LA, Weiner DE, Maddux 49. Georgianos PI, Agarwal R: Pharmacotherapy of hypertension in FW, Nissenson AR, Jadoul M, Locatelli F, Winkelmayer WC, Port chronic dialysis patients. Clin J Am Soc Nephrol 11: 2062–2075, FK, Robinson BM, Tentori F: Dialysate potassium, serum potas- 2016 sium, mortality, and arrhythmia events in hemodialysis: Results from the Dialysis Outcomes and Practice Patterns Study (DOPPS). Received: October 8, 2019 Accepted: May 7, 2020 Am J Kidney Dis 69: 266–277, 2017 42. Agarwal R, Rossignol P,Romero A, Garza D, Mayo MR, Warren S, Ma J, White WB, Williams B: Patiromer versus placebo to enable Published online ahead of print. Publication date available at spironolactone use in patients with resistant hypertension and www.cjasn.org. chronic kidney disease (AMBER): A phase 2, randomised, double- “ blind, placebo-controlled trial. Lancet 394: 1540–1550, 2019 See related editorial, Drug Selection for Treating Hypertension in 43. Agarwal R, Sinha AD, Pappas MK, Abraham TN, Tegegne GG: Dialysis Patients: More to Consider than BP-Lowering Potency,” on Hypertension in hemodialysis patients treated with atenolol or pages xxx–xxx.

AFFILIATIONS

1The George Institute for Global Health, Faculty of Medicine, The University of New South Wales, Sydney, New South Wales, Australia 2Medication Safety Research Chair, Department of Clinical Pharmacy, College of Pharmacy, King Saud University, Riyadh, Saudi Arabia 3Sydney School of Public Health, Sydney Medical School, University of Sydney, Sydney, New South Wales, Australia 4Department of Cardiology, The Royal Prince Alfred Hospital, Sydney, New South Wales, Australia 5Department of Medicine, University of Otago Christchurch, Christchurch, New Zealand 6Department of Cardiology and Kolling Institute, Northern Sydney Local Health District, St. Leonards, New South Wales, Australia 7Department of Biomedical Sciences, Faculty of Medicine and Health Sciences, Macquarie University, Sydney, New South Wales, Australia 8Department of Renal Medicine, Concord Repatriation General Hospital, Sydney, New South Wales, Australia 9Concord Clinical School, Sydney Medical School, Faculty of Health and Medicine, University of Sydney, Sydney, New South Wales, Australia SUPPLEMENTAL MATERIAL TABLE OF CONTENTS

Supplemental Table 1. Electronic search terms. Supplemental Table 2. Characteristics of included studies. Supplemental Table 3. Risk of bias assessments of all included trials. Supplemental Table 4. Assessment of loop-specific (in closed loops of evidence) and overall inconsistency and heterogeneity. Supplemental Table 5. Assessment of agreement between direct and indirect evidence using side- splitting approach for all outcomes. ' Supplemental Figure 1. Electronic searches in MEDLINE, Embase, and the Cochrane Central Register of Randomised Trials for trials on blood pressure lowering drugs in adults with kidney failure requiring dialysis. Supplemental Figure 2. Domain-specific risk of bias assessment summary of included studies. Supplemental Figure 3. Study limitations weighted by contribution of direct estimates to the network of blood pressure lowering drugs for systolic blood pressure outcome. Supplemental Figure 4. Comparison-adjusted funnel plot of placebo-controlled trials (n=17) of blood pressure lowering drugs effect on systolic blood pressure in dialysis patients. Supplemental Figure 5. Network estimates of blood pressure lowering drugs effects on diastolic blood pressure. Supplemental Figure 6. Network estimates of blood pressure lowering drugs effects on heart rate. Supplemental Figure 7. Network estimates of blood pressure lowering drugs effects on discontinuation due to adverse events. Supplemental Figure 8. Network estimates of blood pressure lowering drugs effects on risk of hypotension. Supplemental Figure 9. Network estimates of blood pressure lowering drugs effects on the risk of hyperkalaemia. Supplemental Figure 10. Network estimates of blood pressure lowering drugs effects on serum potassium. Supplemental Figure 11. Network of treatment comparison of blood pressure lowering agents effects on diastolic blood pressure, heart rate and potassium concentration outcomes. Supplemental Figure 12. Network estimates of blood pressure lowering drugs effects on systolic blood pressure in trials with low risk of bias (n=9). Supplemental References 1 Supplemental Table 1. Electronic search terms

CENTRAL Embase MEDLINE 1. exp Antihypertensive Agents/ 1. exp Antihypertensive 1. exp Antihypertensive Agents/ 2. (antihypertensive$ adj (agent$ Agents/ 2. (antihypertensive$ adj (agent$ or drug)).tw. 2. (antihypertensive$ adj or drug)).tw. 3. exp Adrenergic alpha- (agent$ or drug)).tw. 3. exp Adrenergic alpha- Antagonists/ 3. exp Adrenergic alpha- Antagonists/ 4. Doxazosin.tw. Antagonists/ 4. Doxazosin.tw. 5. Prazosin.tw. 4. Doxazosin.tw. 5. Prazosin.tw. 6. Terazosin.tw. 5. Prazosin.tw. 6. Terazosin.tw. 7. exp Adrenergic alpha-2 6. Terazosin.tw. 7. exp Adrenergic alpha-2 Receptor Agonists/ 7. exp Adrenergic alpha-2 Receptor Agonists/ 8. Clonidine.tw. Receptor Agonists/ 8. Clonidine.tw. 9. Guanabenz.tw. 8. Clonidine.tw. 9. Guanabenz.tw. 10. Guanfacine.tw. 9. Guanabenz.tw. 10. Guanfacine.tw. 11. Methyldopa.tw. 10. Guanfacine.tw. 11. Methyldopa.tw. 12. Lofexidine.tw. 11. Methyldopa.tw. 12. Lofexidine.tw. 13. exp Diuretics, Potassium 12. Lofexidine.tw. 13. exp Diuretics, Potassium Sparing/ 13. exp Diuretics, Potassium Sparing/ 14. Eplerenone.tw. Sparing/ 14. Eplerenone.tw. 15. Amiloride.tw. 14. Eplerenone.tw. 15. Amiloride.tw. 16. Spironolactone.tw. 15. Amiloride.tw. 16. Spironolactone.tw. 17. Triamterene.tw. 16. Spironolactone.tw. 17. Triamterene.tw. 18. exp Angiotensin Receptor 17. Triamterene.tw. 18. exp Angiotensin Receptor Antagonists/ 18. exp Angiotensin Receptor Antagonists/ 19. Azilsartan.tw. Antagonists/ 19. Azilsartan.tw. 20. Candesartan.tw. 19. Azilsartan.tw. 20. Candesartan.tw. 21. Eprosartan.tw. 20. Candesartan.tw. 21. Eprosartan.tw. 22. Irbesartan.tw. 21. Eprosartan.tw. 22. Irbesartan.tw. 23. Losartan.tw. 22. Irbesartan.tw. 23. Losartan.tw. 24. Olmesartan.tw. 23. Losartan.tw. 24. Olmesartan.tw. 25. Telmisartan.tw. 24. Olmesartan.tw. 25. Telmisartan.tw. 26. Valsartan.tw. 25. Telmisartan.tw. 26. Valsartan.tw. 27. exp Angiotensin-Converting 26. Valsartan.tw. 27. exp Angiotensin-Converting Enzyme Inhibitors/ 27. exp Angiotensin-Converting Enzyme Inhibitors/ 28. Benazepril.tw. Enzyme Inhibitors/ 28. Benazepril.tw. 29. Captopril.tw. 28. Benazepril.tw. 29. Captopril.tw. 30. Cilazapril.tw. 29. Captopril.tw. 30. Cilazapril.tw. 31. Enalapril.tw. 30. Cilazapril.tw. 31. Enalapril.tw. 32. Fosinopril.tw. 31. Enalapril.tw. 32. Fosinopril.tw. 33. Lisinopril.tw. 32. Fosinopril.tw. 33. Lisinopril.tw. 34. Moexipril.tw. 33. Lisinopril.tw. 34. Moexipril.tw. 35. Perindopril.tw. 34. Moexipril.tw. 35. Perindopril.tw. 36. Quinapril.tw. 35. Perindopril.tw. 36. Quinapril.tw. 37. Ramipril.tw. 36. Quinapril.tw. 37. Ramipril.tw. 38. Trandolapril.tw. 37. Ramipril.tw. 38. Trandolapril.tw. 39. exp Adrenergic beta- 38. Trandolapril.tw. 39. exp Adrenergic beta- Antagonists/ 39. exp Adrenergic beta- Antagonists/ 40. Carvedilol.tw. Antagonists/ 40. Carvedilol.tw. 41. Labetalol.tw. 40. Carvedilol.tw. 41. Labetalol.tw. 42. Acebutolol.tw. 41. Labetalol.tw. 42. Acebutolol.tw. 43. Pindolol.tw. 42. Acebutolol.tw. 43. Pindolol.tw. 44. Penbutolol.tw. 43. Pindolol.tw. 44. Penbutolol.tw. 45. Atenolol.tw. 44. Penbutolol.tw. 45. Atenolol.tw.

CENTRAL Embase MEDLINE 46. Betaxolol.tw. 45. Atenolol.tw. 46. Betaxolol.tw. 47. Bisoprolol.tw. 46. Betaxolol.tw. 47. Bisoprolol.tw. 48. Celiprolol.tw. 47. Bisoprolol.tw. 48. Celiprolol.tw. 49. Metoprolol.tw. 48. Celiprolol.tw. 49. Metoprolol.tw. 50. Nebivolol.tw. 49. Metoprolol.tw. 50. Nebivolol.tw. 51. Sotalol.tw. 50. Nebivolol.tw. 51. Sotalol.tw. 52. Nadolol.tw. 51. Sotalol.tw. 52. Nadolol.tw. 53. Propranolol.tw. 52. Nadolol.tw. 53. Propranolol.tw. 54. Timolol.tw. 53. Propranolol.tw. 54. Timolol.tw. 55. exp Calcium Channel 54. Timolol.tw. 55. exp Calcium Channel Blockers/ 55. exp Calcium Channel Blockers/ 56. Amlodipine.tw. Blockers/ 56. Amlodipine.tw. 57. Bepridil.tw. 56. Amlodipine.tw. 57. Bepridil.tw. 58. Clevidipine.tw. 57. Bepridil.tw. 58. Clevidipine.tw. 59. Diltiazem.tw. 58. Clevidipine.tw. 59. Diltiazem.tw. 60. Felodipine.tw. 59. Diltiazem.tw. 60. Felodipine.tw. 61. Lacidipine.tw. 60. Felodipine.tw. 61. Lacidipine.tw. 62. Lercanidipine.tw. 61. Lacidipine.tw. 62. Lercanidipine.tw. 63. Levamlodipine.tw. 62. Lercanidipine.tw. 63. Levamlodipine.tw. 64. Isradipine.tw. 63. Levamlodipine.tw. 64. Isradipine.tw. 65. Nicardipine.tw. 64. Isradipine.tw. 65. Nicardipine.tw. 66. Nifedipine.tw. 65. Nicardipine.tw. 66. Nifedipine.tw. 67. Nimodipine.tw. 66. Nifedipine.tw. 67. Nimodipine.tw. 68. Nisoldipine.tw. 67. Nimodipine.tw. 68. Nisoldipine.tw. 69. Verapamil.tw. 68. Nisoldipine.tw. 69. Verapamil.tw. 70. exp Diuretics/ 69. Verapamil.tw. 70. exp Diuretics/ 71. Bumetanide.tw. 70. exp Diuretics/ 71. Bumetanide.tw. 72. Ethacrynic acid.tw. 71. Bumetanide.tw. 72. Ethacrynic acid.tw. 73. Furosemide.tw. 72. Ethacrynic acid.tw. 73. Furosemide.tw. 74. Piretanide.tw. 73. Furosemide.tw. 74. Piretanide.tw. 75. Torsemide.tw. 74. Piretanide.tw. 75. Torsemide.tw. 76. Bendroflumethiazide.tw. 75. Torsemide.tw. 76. Bendroflumethiazide.tw. 77. Chlorothiazide.tw. 76. Bendroflumethiazide.tw. 77. Chlorothiazide.tw. 78. Chlorthalidone.tw. 77. Chlorothiazide.tw. 78. Chlorthalidone.tw. 79. Indapamide.tw. 78. Chlorthalidone.tw. 79. Indapamide.tw. 80. Hydrochlorothiazide.tw. 79. Indapamide.tw. 80. Hydrochlorothiazide.tw. 81. Methyclothiazide.tw. 80. Hydrochlorothiazide.tw. 81. Methyclothiazide.tw. 82. Metolazone.tw. 81. Methyclothiazide.tw. 82. Metolazone.tw. 83. Guanadrel.tw. 82. Metolazone.tw. 83. Guanadrel.tw. 84. Guanethidine.tw. 83. Guanadrel.tw. 84. Guanethidine.tw. 85. Reserpine.tw. 84. Guanethidine.tw. 85. Reserpine.tw. 86. Aliskiren.tw. 85. Reserpine.tw. 86. Aliskiren.tw. 87. exp Vasodilator Agents/ 86. Aliskiren.tw. 87. exp Vasodilator Agents/ 88. Diazoxide.tw. 87. exp Vasodilator Agents/ 88. Diazoxide.tw. 89. Hydralazine.tw. 88. Diazoxide.tw. 89. Hydralazine.tw. 90. Minoxidil.tw. 89. Hydralazine.tw. 90. Minoxidil.tw. 91. Nitroprusside.tw. 90. Minoxidil.tw. 91. Nitroprusside.tw. 92. (ace adj2 inhibitor$).tw. 91. Nitroprusside.tw. 92. (ace adj2 inhibitor$).tw. 93. (angiotensin adj2 receptor 92. (ace adj2 inhibitor$).tw. 93. (angiotensin adj2 receptor antagonist$).tw. 93. (angiotensin adj2 receptor antagonist$).tw. 94. beta block$.tw. antagonist$).tw. 94. beta block$.tw. 95. zofenopril.tw. 94. beta block$.tw. 95. zofenopril.tw. 96. imidapril.tw. 95. zofenopril.tw. 96. imidapril.tw. 97. alprenolol.tw. 96. imidapril.tw. 97. alprenolol.tw. 98. bucindolol.tw. 97. alprenolol.tw. 98. bucindolol.tw. 3

CENTRAL Embase MEDLINE 99. carteolol.tw. 98. bucindolol.tw. 99. carteolol.tw. 100. oxprenolol.tw. 99. carteolol.tw. 100. oxprenolol.tw. 101. aranidipine.tw. 100. oxprenolol.tw. 101. aranidipine.tw. 102. azelnidipine.tw. 101. aranidipine.tw. 102. azelnidipine.tw. 103. barnidipine.tw. 102. azelnidipine.tw. 103. barnidipine.tw. 104. benidipine.tw. 103. barnidipine.tw. 104. benidipine.tw. 105. cilnidipine.tw. 104. benidipine.tw. 105. cilnidipine.tw. 106. efonidipine.tw. 105. cilnidipine.tw. 106. efonidipine.tw. 107. nilvadipine.tw. 106. efonidipine.tw. 107. nilvadipine.tw. 108. nitrendipine.tw. 107. nilvadipine.tw. 108. nitrendipine.tw. 109. pranidipine.tw. 108. nitrendipine.tw. 109. pranidipine.tw. 110. calcium channel blocker*.tw. 109. pranidipine.tw. 110. calcium channel blocker*.tw. 111. angiotensin receptor 110. calcium channel blocker*.tw. 111. angiotensin receptor blocker*.tw. 111. angiotensin receptor blocker*.tw. 112. angiotensin converting blocker*.tw. 112. angiotensin converting enzyme inhibitor*.tw. 112. angiotensin converting enzyme inhibitor*.tw. 113. Renin Inhibitor*.tw. enzyme inhibitor*.tw. 113. Renin Inhibitor*.tw. 114. endothelin receptor 113. Renin Inhibitor*.tw. 114. endothelin receptor antagonist$.tw. 114. endothelin receptor antagonist$.tw. 115. endothelin receptor antagonist$.tw. 115. endothelin receptor blocker$.tw. 115. endothelin receptor blocker$.tw. 116. endothelin inhibitor$.tw. blocker$.tw. 116. endothelin inhibitor$.tw. 117. or/1-116 116. endothelin inhibitor$.tw. 117. or/1-116 118. exp Renal Dialysis/ 117. or/1-116 118. exp Renal Dialysis/ 119. exp Kidney Failure, Chronic/ 118. exp Renal Dialysis/ 119. exp Kidney Failure, Chronic/ 120. dialysis.tw. 119. exp Kidney Failure, Chronic/ 120. dialysis.tw. 121. $dialysis.tw. 120. dialysis.tw. 121. $dialysis.tw. 122. Renal Insufficiency, Chronic/ 121. $dialysis.tw. 122. Renal Insufficiency, Chronic/ 123. (end-stage kidney or end-stage 122. Renal Insufficiency, 123. (end-stage kidney or end- renal or endstage kidney or Chronic/ stage renal or endstage kidney endstage renal).tw. 123. (end-stage kidney or end- or endstage renal).tw. 124. (ESKD or ESKF or ESRD or stage renal or endstage 124. (ESKD or ESKF or ESRD or ESRF).tw. kidney or endstage renal).tw. ESRF).tw. 125. (chronic kidney adj3 (stage 5 124. (ESKD or ESKF or ESRD or 125. (chronic kidney adj3 (stage 5 or stage V)).tw. ESRF).tw. or stage V)).tw. 126. exp continuous ambulatory 125. (chronic kidney adj3 (stage 5 126. or/118-125 peritoneal dialysis/ or stage V)).tw. 127. randomized controlled trial.pt. 127. hemodialysis.tw. 126. exp continuous ambulatory 128. controlled clinical trial.pt. 128. exp hemodialysis/ peritoneal dialysis/ 129. randomized.ab. 129. (CAPD or CCPD or APD).tw. 127. hemodialysis.tw. 130. placebo.ab. 130. h?emodialysis.tw. 128. exp hemodialysis/ 131. clinical trials as topic/ 131. exp end stage renal disease/ 129. (CAPD or CCPD or 132. trial.ti. 132. or/118-131 APD).tw. 133. randomly.ab. 133. 117 and 132 130. h?emodialysis.tw. 134. or/127-133 131. exp end stage renal disease/ 135. animals/ not (humans/ and 132. or/118-131 animals/) 133. random:.tw. 136. 134 not 135 134. double-blind:.tw. 137. 117 and 126 and 136 135. placebo:.mp. 136. or/133-135 137. 117 and 132 and 136

Supplemental Table 2. Characteristics of included studies

Study Country Design Duratio Populati No of Male Mean Interven Interven Control Baseline BP n on patients (%) age tion 1 tion 2 SBP measure (weeks) (years) (mmHg) ment London France parallel 52 HD 24 58 54 perindop nitrendip 177 ABPM et al ril 2-4 ine 20-40 (1994) mg mg od (1) weekly Perfect New Factorial 24 mixed 107 68 49 enalapril Placebo 138 clinic (1997)(2) Zealand 2.5-5 mg OD Nakamot Japan parallel 12 PD 36 58 benzapril valsartan amlodipi 154 pre- o et al or or ne 2.5-10 dialysis (2004) enalapril candesart mg/day (3) 2.5-10 an 20-80 mg/day mg/day or 4-12 mg/day Suzuki et Japan parallel 52 HD 33 61 65 enalapril losartan losartan 170 clinic al (2004) 10mg/da 100mg/d + (4) y ay enalapril losartan 100mg/d ay + enalapril 10mg/da y

Study Country Design Duratio Populati No of Male Mean Interven Interven Control Baseline BP n on patients (%) age tion 1 tion 2 SBP measure (weeks) (years) (mmHg) ment Matsumo Japan parallel 24 HD 27 56 54 imidapril Placebo 150 pre- to et al 2.5mg/da dialysis (2006) y (5) Yu et al Taiwan parallel 52 HD 46 65 47 ramipril Placebo 123 clinic (2006) 1.25- (6) 2.5mg 3 times/we ek on nonD day FOSIDI France parallel 104 HD 397 52 67 fosinopri Placebo 148 pre- AL l 20 dialysis (2006) mg/day (7) (13.2(5.5 ) mg/day) Ordaz- Mexico parallel 12 HD 25 52 41 enalapril Placebo 152 clinic Medina 10 mg et al BD (2010) (8) Yilmaz Turkey parallel 52 HD 92 57 52 ramipril amlodipi 156 ABPM et al 5- ne 5- (2010) 10mg/da 10mg/da (9) y y

Study Country Design Duratio Populati No of Male Mean Interven Interven Control Baseline BP n on patients (%) age tion 1 tion 2 SBP measure (weeks) (years) (mmHg) ment Reyes- Mexico parallel 52 PD 60 60 46 enalapril losartan 133 Marin et 10 50mg/da al (2012) mg/day y (10) Ottosson Sweden parallel 6 HD 20 90 67 candesart Placebo 149 pre- et al an 4-16 dialysis (2003) mg (11) Suzuki et Japan parallel 156 HD 360 59 60 losartan Standard 155 pre- al (2008) 50-150 of care dialysis (12) mg/day candesart an 8-12 mg/day valsartan 80- 160mg/d ay Cice et al Italy parallel 156 HD 332 54 63 telmisart Placebo 125 pre- (2010) an dialysis (13) 80mg/da y OCTOP Japan parallel 182 HD 469 62 60 olmesart Standard 159 pre- US an 10- of care dialysis (2013) 40mg/da (14) y

Study Country Design Duratio Populati No of Male Mean Interven Interven Control Baseline BP n on patients (%) age tion 1 tion 2 SBP measure (weeks) (years) (mmHg) ment SAFIR Denmark parallel 52 HD 82 68 62 irbesarta Placebo 147 pre- (2014) n 150- dialysis (15) 300mg/d ay Satirapoj Thailand parallel 12 HD 33 61 55 valsartan Standard 147 pre- et al 80-320 of care dialysis (2014) mg/day, (16) average 136 mg/day Sun et al China parallel 52 HD 65 48 58 losartan bisoprolo 153 home (2016) 50mg/da l (17) y 5mg/day Meltzer USA parallel 12 HD 20 45 50 prazosin proprano 162 pre- et al 8.3±2.2 lol 123 ± dialysis (1984) mg/day 39 (18) in two mg/day divided in two doses divided doses Cice et al Italy Crossove 2 HD 60 70 52 bisoprolo nifedipin 146 pre- (1997) r l 10 e 20 mg dialysis (19) mg/day BD Cice et al Italy parallel 6 HD 46 54 carvedilo Placebo 158 pre- (1998) l 50 dialysis (20) mg/day

Study Country Design Duratio Populati No of Male Mean Interven Interven Control Baseline BP n on patients (%) age tion 1 tion 2 SBP measure (weeks) (years) (mmHg) ment Cice et al Italy parallel 52 HD 114 61 55 carvedilo Placebo 134 pre- (2001) l 25 mg dialysis (21) BD HDPAL USA parallel 52 HD 200 66 53 atenolol lisinopril 152 home (2014) 25mg- 10mg- (22) 100mg 40mg TIW TIW after after dialysis dialysis BLOCA ANZ parallel 52 mixed 49 63 59 carvedilo Placebo 136 post- DE l 6.25- dialysis (2016) 25mg (23) BD London France parallel 16 HD 39 49 57 nitrendip Placebo 189 pre- et al ine 20 dialysis (1990) mg od- (24) bid Cice et al Italy parallel 12 HD 48 60 46 diltiazem Placebo 127 pre- (1999) 60 mg dialysis (25) BD Tepel et Germany parallel 76 HD 251 63 61 amlodipi Placebo 140 pre- al (2008) ne 10 dialysis (26) mg/day Hausberg Germany parallel 26 HD 23 70 47 moxonidi Placebo et al ne 0.3 (2010) mg/day (27)

Study Country Design Duratio Populati No of Male Mean Interven Interven Control Baseline BP n on patients (%) age tion 1 tion 2 SBP measure (weeks) (years) (mmHg) ment Gross et USA Crossove 2 HD 8 38 53 spironola Placebo 143 pre- al (2005) r ctone dialysis (28) 50mg BD Taheri et Iran parallel 26 HD 16 69 58 spironola Placebo al (2009) ctone (29) 25mg/da y Vukusich Chile parallel 104 HD 53 64 58 spironola Placebo 149 pre- et al ctone 50 dialysis (2010) mg thrice (30) weekly on non- dialysis days Taheri et Iran parallel 24 PD 18 56 54 spironola Placebo al (2012) ctone 25 (31) mg every other day Zaripova Russia parallel 24 HD 71 spironola Standard et al ctone 25 of care (2011) mg/day (32) DOHAS Japan parallel 156 HD 309 66 68 spironola Standard 150 (2014) ctone of care (33) 25mg/da y

Study Country Design Duratio Populati No of Male Mean Interven Interven Control Baseline BP n on patients (%) age tion 1 tion 2 SBP measure (weeks) (years) (mmHg) ment Ni et al China parallel 12 mixed 76 59 55 spironola Placebo 146 clinic (2014) ctone 25- (34) 50 mg/day Feniman- Brazil parallel 24 HD 17 53 54 spironola Placebo 137 ABPM De- ctone Stefano 12.5-25 et al mg/day (2015) (35) Ito et al Japan parallel 104 PD 158 72 57 spironola Standard 136 clinic (2015) ctone 25 of care (36) mg/day PHASE Canada parallel 13 HD 154 62 63 eplereno Placebo 146 (2015) ne 25-50 (37) mg/day Lin et al China parallel 104 mixed 253 60 71 spironola Placebo 143 (2016) ctone 25 (38) mg/day

Study Country Design Duratio Populati No of Male Mean Interven Interven Control Baseline BP n on patients (%) age tion 1 tion 2 SBP measure (weeks) (years) (mmHg) ment Ito et al Japan parallel 12 HD 18 67 59 aliskiren ARB 142 pre- (2014) 150 (valsarta dialysis (39) mg/day n 80mg n=12 or telmisart an 40 mg n=5 or olmesart an 20 mg n=1) Kuriyam Japan parallel 24 HD 74 88 61 aliskiren amlodipi 164 home a et al 150mg/d ne 5 (2014) ay mg/day (40) ABPM= ambulatory blood pressure monitoring. ANZ= Australia and New Zealand. BP= blood pressure. HD= haemodialysis. PD= peritoneal dialysis. SBP= systolic blood pressure.

Supplemental Table 3. Risk of bias assessments of all included trials

Random Blinding Allocation Blinding sequence Blinding (outcome Double- Incomplete Selective Baseline Summary Study concealm (investiga ITT generatio (patients) assessmen blind? outcome reporting imbalance quality ent tors) n t) London et Unclear Double- Low risk Low risk Low risk Low risk Low risk Unclear risk Yes Low risk Low risk al (1994) risk blinded of bias Perfect Double- Unclear Low risk Low risk Low risk Low risk Low risk Low risk Unclear risk Yes Low risk (1997) blinded risk of bias Nakamoto Unclear Unclear Unclear Unclear Unclear Unclear Not et al Low risk Yes Low risk Low risk risk of risk risk risk risk risk specified (2004) bias Suzuki et Unclear Open- Unclear High risk Low risk High risk High risk High risk High risk No Low risk al (2004) risk label risk of bias Matsumot Unclear Unclear Unclear Unclear Unclear Not Unclear High risk o et al High risk No Low risk risk risk risk risk risk specified risk of bias (2006) Unclear Yu et al Unclear Unclear Unclear Double- Low risk Low risk Low risk Yes Low risk Low risk risk of (2006) risk risk risk blinded bias FOSIDIA Double- High risk Low risk Low risk Low risk Low risk Low risk Unclear risk Yes Low risk High risk L (2006) blinded of bias Ordaz- Unclear Double- Low risk Medina et Low risk Low risk Low risk Low risk Low risk Yes Low risk Low risk risk blinded of bias al (2010) Yilmaz et Unclear Unclear Open- High risk High risk High risk High risk Unclear risk No Low risk Low risk al (2010) risk risk label of bias Reyes- Unclear Open- High risk Marin et Low risk High risk High risk High risk Low risk Yes Low risk Low risk risk label of bias al (2012) 13

Random Blinding Allocation Blinding sequence Blinding (outcome Double- Incomplete Selective Baseline Summary Study concealm (investiga ITT generatio (patients) assessmen blind? outcome reporting imbalance quality ent tors) n t) Ottosson Unclear Unclear Double- High risk et al Low risk Low risk Low risk High risk No Low risk High risk risk risk blinded of bias (2003) Suzuki et Unclear Open- High risk Low risk High risk High risk High risk Low risk No Low risk Low risk al (2008) risk label of bias Cice et al Double- Low risk Low risk Low risk Low risk Low risk Low risk Low risk Yes Low risk Low risk (2010) blinded of bias Unclear OCTOPU Unclear Unclear Open- High risk High risk Low risk Low risk Yes Low risk Low risk risk of S (2013) risk risk label bias SAFIR Double- Low risk Low risk Low risk Low risk Low risk Low risk Low risk Yes Low risk Low risk (2014) blinded of bias Satirapoj Unclear Open- Unclear High risk et al Low risk High risk High risk High risk Low risk Yes Low risk risk label risk of bias (2014) Sun et al Open- High risk Low risk Low risk High risk High risk Low risk Unclear risk No Low risk Low risk (2016) label of bias Meltzer et Unclear Unclear Open- Unclear Unclear High risk High risk High risk High risk Unclear risk Unclear al (1984) risk risk label risk risk of bias Cice et al Unclear Unclear Open- High risk High risk High risk High risk Low risk Yes Low risk Low risk (1997) risk risk label of bias Unclear Cice et al Unclear Unclear Unclear Double- Unclear Low risk Low risk Unclear risk No Low risk risk of (1998) risk risk risk blinded risk bias Cice et al Unclear Double- Low risk (2001,200 Low risk Low risk Low risk Low risk Low risk Yes Low risk Low risk risk blinded of bias 3) 14

Random Blinding Allocation Blinding sequence Blinding (outcome Double- Incomplete Selective Baseline Summary Study concealm (investiga ITT generatio (patients) assessmen blind? outcome reporting imbalance quality ent tors) n t) HDPAL Unclear Unclear Open- Low risk Low risk Low risk Low risk Low risk Yes Low risk Low risk (2014) risk risk label of bias BLOCAD Double- Low risk Low risk Low risk Low risk Low risk Low risk Low risk Yes Low risk Low risk E (2016) blinded of bias London et Unclear Unclear Double- Unclear Low risk Low risk Low risk Low risk Low risk Yes Low risk al (1990) risk risk blinded risk of bias Unclear Cice et al Unclear Unclear Unclear Double- Low risk Low risk Low risk Yes Low risk Low risk risk of (1999) risk risk risk blinded bias Tepel et al Double- Unclear Low risk Low risk Low risk Low risk Low risk Low risk Low risk Yes Low risk (2008) blinded risk of bias Hausberg Unclear Unclear Unclear Unclear Unclear Unclear Not Unclear Unclear et al Low risk Yes risk of risk risk risk risk risk specified risk risk (2010) bias Gross et al Unclear Double- Low risk Low risk Low risk Low risk Low risk Low risk Yes Low risk Low risk (2005) risk blinded of bias Taheri et Unclear Unclear Double- Unclear High risk Low risk Low risk Low risk High risk No Low risk al (2009) risk risk blinded risk of bias Vukusich Unclear Unclear Unclear Double- et al Low risk Low risk Low risk High risk No Low risk Low risk risk of risk risk blinded (2010) bias Unclear Taheri et Unclear Unclear Double- Unclear Low risk Low risk Low risk Unclear risk No Low risk risk of al (2012) risk risk blinded risk bias Zaripova Unclear Unclear Open- Unclear Unclear High risk et al High risk High risk High risk Unclear risk Unclear risk risk label risk risk of bias (2012) 15

Random Blinding Allocation Blinding sequence Blinding (outcome Double- Incomplete Selective Baseline Summary Study concealm (investiga ITT generatio (patients) assessmen blind? outcome reporting imbalance quality ent tors) n t) Unclear DOHAS Unclear Unclear Open- Unclear Low risk Low risk Low risk Low risk Yes Low risk risk of (2014) risk risk label risk bias Ni et al Unclear Unclear Double- Low risk Low risk Low risk Low risk Low risk Yes Low risk Low risk (2014) risk risk blinded of bias Feniman- Unclear De- Unclear Unclear Double- Unclear Low risk Low risk Low risk Unclear risk No Low risk risk of Stefano et risk risk blinded risk bias al (2015) Ito et al Unclear Unclear Open- High risk High risk High risk Low risk Low risk Yes Low risk Low risk (2015) risk risk label of bias PHASE Double- Low risk Low risk Low risk Low risk Low risk Low risk Low risk Yes Low risk Low risk (2015) blinded of bias Lin et al Double- Low risk Low risk Low risk Low risk Low risk Low risk Low risk Yes Low risk Low risk (2016) blinded of bias Ito et al Unclear Open- Unclear High risk Low risk High risk High risk High risk Low risk Yes Low risk (2014) risk label risk of bias Kuriyama Unclear Unclear Open- High risk et al High risk High risk Low risk High risk No Low risk Low risk risk risk label of bias (2014) ITT=intention-to-treat analysis.

Database searches (n=6406) MEDLINE (n=1791) EMBASE (n=3028) Cochrane CENTRAL (n=1587)

Duplicates (n=2163)

Abstract review (n=4243)

4108 excluded Not original investigation (e.g. review, editorial, commentary), Not a randomized controlled trials, animals and in vitro studies Not dialysis population, not relevant interventions

Full article review (n=135)

95 excluded Not a randomized controlled trial (n=17) Outcome of interest not reported (n=27) Not relevant intervention (n=33) Not dialysis population (n=10) Multiple reports of the same study (n=8)

40 trials were included: - 25 trials reported data on systolic BP outcome. - 15 trials did not report on systolic BP and were included for other outcomes.

Supplemental Figure 1. Electronic searches in MEDLINE, Embase, and the Cochrane Central Register of Randomised Trials for trials on blood pressure lowering drugs in adults with end-stage kidney disease requiring dialysis.

Random sequence generation (selection bias) Allocation concealment (selection bias) Blinding of participants (performance bias) Blinding of investigators (performance bias) Blinding of outcome assessment (detection bias) Incomplete outcome data (attrition bias) Selective reporting (reporting bias) Summary assessments of risk of bias

0% 25% 50% 75% 100% Low risk of bias Unclear risk of bias High risk of bias

Supplemental Figure 2. Domain-specific risk of bias assessment summary of included studies.

5% 5% 2% 11% 4%

7% 9%

ARB vs CCB β blocker vs placebo 4% 14%

11% 4% 16% ► No concern 14% ► Some concern 55% ► Major concern 31%

Supplemental Figure 3. Study limitations weighted by contribution of direct estimates to the network of blood pressure lowering drugs for systolic blood pressure outcome.

ACE=angiotensin-converting enzyme. ARB=angiotensin-receptor blockers. CCB= calcium- channel blockers.

Supplemental Table 4. Assessment of loop-specific (in closed loops of evidence) and overall inconsistency and heterogeneity.

Inconsistency factor Loop heterogeneity P for global Network Outcome Closed loop of evidence (95% CI) (tau2) inconsistency test heterogeneity (tau2) ARB-CCB-Placebo 9.23 (0.00,35.39) 28.58 ACEI-ARB-Placebo 3.75 (0.00,18.02) 0.00 0.96 ACEI-CCB-Placebo 3.09 (0.00,13.11) 4.77 BB-CCB-Placebo 2.62 (0.00,17.35) 15.73 Systolic blood 0.00 (low ARB-CCB- Renin pressure 2.3 (0.00,16.53) 0.00 heterogeneity) inhibitors ACEI-BB-CCB 2.15 (0.00,10.65) 0.00 ACEI-ARB-CCB 1.59 (0.00,17.61) 0.00 ACEI-BB-Placebo 0.14 (0.00,7.09) 0.00 ACEI-BB-Placebo 2.82(0.00,8.54) 2.16 ARB-CCB-Renin 2.4(0.00,14.72) 0.00 1.0 inhibitors Diastolic blood ACEI-ARB-CCB 2.05(0.00,16.72) 0.00 3.9 (high heterogeneity) pressure ARB-CCB-Placebo 1.84(0.00,32.80) 101.53 ACEI-CCB-Placebo 1.39(0.00,13.05) 18.54 ACEI-ARB-Placebo 0.32(0.00,14.21) 7.56 ACEI-ARB-Placebo 7.23(0.21,14.24) 0.00 0.72 ACEI-CCB-Placebo 6.26(0.00,14.96) 0.00 Heart rate ACEI-ARB-CCBs 3(0.00,12.32) 0.00 3.7 (high heterogeneity) BB-CCB-Placebo 0.49 (0.00,13.97) 0.00 ARB-CCB-Placebo 0.26(0.00,8.47) 0.00 Discontinuation due to ACEI-ARB-Control 1.94 (0.00,4.87) 0.00 0.20 0.0 (low heterogeneity)

adverse event ACEI-ARB-ACEI+ARB . 0.00 Hypotension no closed loops - - 0.64 0.0 (low heterogeneity) Hyperkalaemia no closed loops - - 0.22 0.0 (low heterogeneity) Potassium no closed loops - - <0.001 0.0 (low heterogeneity) concentration Aldosterone antagonists vs. placebo comparison did not form a closed loop of evidence and all the evidence for this comparison come from aldosterone antagonists vs. placebo trials, i.e. no indirect evidence was calculated for this comparison. ACEI = angiotensin-converting enzyme inhibitors. ARB = angiotensin receptor blockers. BB = β blockers. CCB = calcium channel blockers.

Supplemental Table 5. Assessment of agreement between direct and indirect evidence using side-splitting approach for all outcomes.

Direct Standard Indirect Standard Standard Outcome Side Difference P value estimates error estimates error. error. ACEI-Placebo 4.01186 1.887974 4.748869 2.286428 -0.73701 2.965164 0.80 ACEI-ARB 1.894487 4.820632 0.89501 3.818895 0.999477 6.149051 0.87 ACEI-β blockers -4.70139 2.58363 -4.18304 1.923293 -0.51835 3.220966 0.87 ACEI-CCB 1.343631 3.290392 -0.82741 1.836203 2.17104 3.768251 0.57 ARB-Placebo -1.63793 5.115792 5.187827 3.477899 -6.82575 6.186041 0.27 ARB-CCB 0.300003 4.822033 -2.60319 3.524782 2.903196 5.972946 0.63 ARB-Renin Systolic 10.59749 3.362012 4.824083 5.417452 5.773405 6.375496 0.37 blood inhibitors α blockers-β blockers pressure -2 3.609678 -15.3588 200.0456 13.35879 200.1433 0.95 * β blockers-Placebo 8.852225 1.53188 8.457964 1.726304 0.394261 2.307983 0.86 β blockers-CCB 3.89945 1.130249 4.568475 2.05124 -0.66902 2.341872 0.78 CCB-Placebo 5.436714 2.128656 3.600421 2.495653 1.836293 3.558225 0.61 CCB-Renin 7.998312 4.262446 13.78184 4.73911 -5.78353 6.372505 0.36 inhibitors ALDO-Placebo ...... ACEI-Placebo 0.848348 1.921043 2.544202 2.229023 -1.69585 2.942216 0.56 ACEI-ARB -0.40164 5.355629 -0.16529 3.119034 -0.23635 6.197074 0.97 ACEI-β blockers -1.70027 2.527595 -3.8482 2.427576 2.147928 3.504563 0.54 Diastolic ACEI-CCB -0.80711 2.666879 -0.96117 3.055745 0.154064 4.049435 0.97 blood pressure ARB-Placebo 1.364134 3.55327 2.173941 3.828507 -0.80981 5.189358 0.88 ARB-CCB -1.8 5.34946 -0.24237 3.124403 -1.55762 6.195047 0.80 ARB-Renin 5.59959 2.847764 3.783104 5.155734 1.816486 5.889747 0.76 inhibitors

α blockers-β blockers -1 3.441318 -7.80411 200.0688 6.80411 200.1383 0.97 * β blockers-Placebo 4.873786 1.555929 2.722145 3.090398 2.151641 3.501714 0.54 CCB-Placebo 2.869727 3.102839 2.038153 2.770071 0.831574 4.024667 0.84 CCB-Renin 4.999044 3.977951 6.822081 4.343092 -1.82304 5.888967 0.76 inhibitors ALDO-Placebo ...... ACEI-Placebo -5.88078 2.352137 0.651982 2.261786 -6.53277 3.263162 0.045 ACEI-ARB -2.09945 2.558684 -8.26061 2.219672 6.161154 3.38747 0.069 ACEI-CCB -0.66272 2.143051 -7.34159 3.911949 6.678876 4.47188 0.14 ARB-Placebo 3.033585 1.952983 -0.53574 3.47376 3.569329 3.95103 0.37 Heart rate ARB-CCB 2.700005 3.589857 3.123915 3.677171 -0.42391 5.138935 0.93 β blockers-Placebo 20.62708 2.892543 19.13711 6.629586 1.489967 7.15432 0.84 β blockers-CCB 20.16617 6.206428 21.608 3.58629 -1.44183 7.168406 0.84 CCB-Placebo 1 3.998048 -1.82377 2.884053 2.823767 4.929721 0.57 CentralAct-Placebo ...... ACEI-Control -0.51677 0.250801 -2.45176 1.477647 1.934993 1.49878 0.20 Discontinua ACEI-ARB * -1.94575 1.455656 -0.01038 0.357449 -1.93537 1.498901 0.20 tion due to ACEI-ACEI+ARB * -0.84725 0.886403 3.022737 2.993614 -3.86999 2.997559 0.20 adverse events ARB-Control -0.50638 0.254693 1.429147 1.477142 -1.93553 1.498939 0.20 ARB-ACEI+ARB * 1.098612 1.581139 -2.77245 1.954636 3.871062 2.997877 0.20 *All evidence for these comparisons comes from trials that directly compare them. . = missing values indicate only direct evidence existed and inconsistency between direct and indirect evidence could not be assessed. ACEI = angiotensin-converting enzyme inhibitors. ALDO = aldosterone antagonists. ARB = angiotensin receptor blockers. CCB = calcium channel blockers. CentralAct = centrally acting vasodilators.

0 2.5 5 7.5 Standard error of effect size effect of error Standard 10 -20 -10 0 10 20 Effect size centered at comparison-specific pooled effect (yiXY-µXY) ACEI ARB βB CCB ALDO

Supplemental Figure 4. Comparison-adjusted funnel plot of placebo-controlled trials (n=17) of blood pressure lowering drugs effect on systolic blood pressure in dialysis patients.

Each point represents the difference between the trial-specific effect size and its respective comparison-specific summary effect. ACEI= angiotensin converting enzyme inhibitors. ARB= angiotensin receptor blockers. βB= β blockers. CCB= calcium channel blockers. ALDO= aldosterone antagonists.

Direct Trials Network Patients I2 Treatment vs comparator MD (95% CI) MD (95% CI) (N) (N) BP lowering agents vs Placebo ACE inhibitor -1.5 (-4.3, 1.4) -1.4 (-5.1, 2.3) 602 5 40.7% ARB -1.7 (-6.8, 3.4) -2.2 (-10.7, 6.2) 70 2 41.8% α blocker -3.4 (-10.7, 3.9) β blocker -4.4 (-7.1, -1.7) -5.2 (-6.9, -3.4) 192 3 0.0% CCB -2.3 (-6.4, 1.8) -6.9 (-25.5, 11.7) 87 2 89.4% Aldosterone antagonist -3.7 (-8.2, 0.8) -3.7 (-7.5, 0.1) 146 3 0.0% Renin inhibitor 3.5 (-2.5, 9.5) BP lowering agents vs ACE inhibitor ARB -0.2 (-5.5, 5.0) -0.4 (-10.1, 9.3) 24 1 α blocker -1.9 (-9.5, 5.6) β blocker -2.9 (-6.3, 0.4) -1.7 (-6.1, 2.7) 102 1 CCB -0.9 (-4.7, 3.0) -0.6 (-5.2, 4) 132 3 0.0% Aldosterone antagonist -2.2 (-7.6, 3.1) Renin inhibitor 5.0 (-1.1, 11.0) BP lowering agents vs ARB α blocker -1.7 (-10.5, 7.1) β blocker -2.7 (-8.3, 2.9) CCB -0.6 (-5.9, 4.6) -1.8 (-11.6, 8.0) 24 1 Aldosterone antagonist -2.0 (-8.8, 4.8) Renin inhibitor 5.2 (0.5, 9.9) 5.6 (1.9, 9.3) 18 1 BP lowering agents vs α blocker β blocker -1.0 (-7.7, 5.7) -1.0 (-6.5, 4.5) 20 1 CCB 1.1 (-7.1, 9.3) Aldosterone antagonist -0.3 (-8.8, 8.3) Renin inhibitor 6.9 (-2.4, 16.2) BP lowering agents vs β blocker CCB 2.1 (-2.6, 6.8) Aldosterone antagonist 0.7 (-4.5, 6.0) Renin inhibitor 7.9 (1.4, 14.4) BP lowering agents vs CCB Aldosterone antagonist -1.4 (-7.5, 4.7) Renin inhibitor 5.8 (0.2, 11.4) 5.0 (-1.7, 11.7) 74 1

Renin inhibitor vs aldosterone antagonist 7.2 (-0.3, 14.7) Overall BPLP vs placebo -3.1 (-5.2, -1.0) -3.2 (-5.6, -0.8) 1097 15 65.4%

-15 -10 -5 0 5 10 20 favours Treatment favours Comparator

Supplemental Figure 5. Network estimates of blood pressure lowering drugs effects on diastolic blood pressure.

BP= blood pressure. ACE= angiotensin converting enzyme. ARB= angiotensin receptor blockers. CCB= calcium-channel blockers. MD= mean difference (mmHg). BPLP= blood pressure lowering pharmacotherapy

Treatment vs Comparator Network Direct Patients Trials I2 MD (95% CI) MD (95% CI) (N) (N) BP lowering agents vs placebo ACE inhibitor 3.1 (-1.0, 7.2) 5.9 (1.2, 10.6) 73 2 0.0% ARB -2.2 (-5.8, 1.4) -3.4 (-5.1, -1.8) 388 2 0.0% β blocker -20.6 (-25.7, -15.5) -18 (-29.6, -6.5) 149 2 77.9% CCB 0.8 (-3.5, 5.0) -1 (-7.3, 5.3) 39 1 Centrally acting drug -0.1 (-8.2, 8.0) -0.1 (-7.2, 7) 23 1

BP lowering agents vs ACEI ARB -5.3 (-9.6, -1.0) -2.1 (-7.4, 3.2) 26 1 β blocker -23.7 (-29.5, -17.8) CCB -2.3 (-6.4, 1.8) 0.3 (-3.1, 3.8) 50 2 0.0% Centrally acting drug -3.2 (-12.2, 5.9)

BP lowering agents vs ARB β blocker -18.4 (-23.8, -12.9) CCB 3.0 (-1.5, 7.5) 2.7 (-2.8, 8.2) 26 1 Centrally acting drug 2.1 (-6.7, 11.0)

BP lowering agents vs β blocker CCB 21.4 (15.5, 27.3) 20.2 (8.8, 31.6) 120 1 Centrally acting drug 20.5 (10.9, 30.1)

BP lowering agents vs CCB Centrally acting drug -0.9 (-10.0, 8.3)

BPLP vs Placebo -3.5 (-10.8, 3.8) 672 8 94.4%

-20 -10 0 10 20 Favours treatment Favours comparator

Supplemental Figure 6. Network estimates of blood pressure lowering drugs effects on heart rate.

BP= blood pressure. ACE= angiotensin converting enzyme. ARB= angiotensin receptor blockers. CCB= calcium-channel blockers. MD= mean difference (beats per minute). BPLP= blood pressure lowering pharmacotherapy.

Network Direct Patients Trials Treatment vs comparator I2 RR (95% CI) RR (95% CI) (N) (N)

BP lowering agents vs control ACE inhibitor 1.77 (1.09, 2.87) 1.68 (1.03, 2.74) 183 3 0.0% ARB 1.57 (0.96, 2.57) 1.66 (1.01, 2.73) 800 4 0.0% β blocker 0.81 (0.27, 2.39) 1.55 (0.11, 22.10) 163 2 67.8% aldosterone antagonist 3.35 (1.32, 8.49) 4.05 (1.16, 14.20) 835 6 39.2% ACE inhibitor+ ARB 0.90 (0.15, 5.31)

BP lowering agents vs ACE inhibitor ARB 0.89 (0.45, 1.75) 0.14 (0.01, 2.48) 22 1 β blocker 0.46 (0.14, 1.50) aldosterone antagonist 1.89 (0.66, 5.40) ACE inhibitor+ ARB 0.51 (0.09, 2.84) 0.33 (0.04, 2.73) 22 1

BP lowering agents vs ARB β blocker 0.52 (0.16, 1.70) aldosterone antagonist 2.14 (0.75, 6.11) ACE inhibitor+ ARB 0.58 (0.09, 3.56) 0.33 (0.02, 7.39) 22 1

BP lowering agents vs β blocker aldosterone antagonist 4.14 (0.99, 17.30) ACE inhibitor+ ARB 1.12 (0.14, 8.92)

BP lowering agents vs aldosterone antagonist ACE inhibitor+ ARB 0.27 (0.04, 1.99)

Overall BP lowering agents vs control 1.70 (1.24, 2.33) 1.87 (1.20, 2.91) 1981 15 22.5%

0.03 0.25 0.5 1 2 4 18 favours Treatment favours Comparator

Supplemental Figure 7. Network estimates of blood pressure lowering drugs effects on discontinuation due to adverse events.

BP= blood pressure. ACE= angiotensin converting enzyme. ARB= angiotensin receptor blockers. CCB= calcium-channel blockers. RR= relative risk.

Network Direct Patients Trials Treatment vs comparator I2 RR (95% CI) RR (95% CI) (N) (N)

BP lowering agents vs control ACE inhibitor 6.62 (1.48, 29.54) 6.62 (1.57, 27.94) 107 1 ARB 1.53 (0.94, 2.48) 1.41 (1.03, 1.94) 1310 6 0.0% CCB 0.59 (0.24, 1.41) 0.59 (0.27, 1.28) 251 1 Aldosterone antagonist 1.17 (0.61, 2.22) 1.16 (0.66, 2.04) 312 2 0.0%

BP lowering agents vs ACE inhibitor ARB 0.23 (0.05, 1.11) CCB 0.09 (0.02, 0.50) Aldosterone antagonist 0.18 (0.03, 0.90)

BP lowering agents vs ARB CCB 0.38 (0.14, 1.05) Aldosterone antagonist 0.77 (0.34, 1.71)

BP lowering agents vs CCB Aldosterone antagonist 1.99 (0.67, 5.92)

Overall BPLP vs control 1.42 (0.93, 2.17) 1.41 (0.97, 2.05) 1980 10 36.2%

0.01 0.5 1 2 30 favours Treatment favours Comparator

Supplemental Figure 8. Network estimates of blood pressure lowering drugs effects on risk of hypotension.

ACE= angiotensin converting enzyme. ARB= angiotensin receptor blockers. BP= blood pressure. BPLP= blood pressure lowering pharmacotherapy. CCB= calcium-channel blockers. RR= relative risk.

Network Patients Trials I2 Treatment vs comparator RR (95% CI) Direct RR (95% CI) (N) (N)

BP lowering agents vs control ACE inhibitor 1.03 (0.11, 9.55) 1.03 (0.11, 9.55) 137 2 0.0% ARB 1.17 (0.63, 2.15) 1.16 (0.68, 2.00) 1249 4 0.0% β blocker 0.31 (0.02, 4.00) MCRA vs Control 1.63 (0.75, 3.57) 1.85 (0.89, 3.85) 928 7 9.1%

BP lowering agents vs ACE inhibitor ARB 1.13 (0.11, 11.44) β blocker 0.30 (0.08, 1.06) 0.30 (0.09, 1.06) 200 1 Aldosterone antagonist 1.59 (0.15, 16.87)

BP lowering agents vs ARB β blocker 0.26 (0.02, 3.69) Aldosterone antagonist 1.40 (0.62, 3.19)

BP lowering agents vs β blocker Aldosterone antagonist 5.30 (0.36, 77.20)

Overall BPLP vs control 1.37 (0.86, 2.21) 1.37 (0.94, 1.99) 2314 13 0.0%

0.01 0.5 1 2 17 favours Treatment favours Comparator

Supplemental Figure 9. Network estimates of blood pressure lowering drugs effects on the risk of hyperkalaemia.

ACE= angiotensin converting enzyme. ARB= angiotensin receptor blockers. BP= blood pressure. BPLP= blood pressure lowering pharmacotherapy. CCB= calcium-channel blockers. RR= relative risk.

Network Direct Patients Trials Treatment vs comparator 2 MD (95% CI) MD (95% CI) (N) (N) I BP lowering agents vs control ACE inhibitor -0.08 (-0.33, 0.18) -0.10 (-0.28, 0.08) 201 4 0.0% ARB -0.22 (-0.51, 0.07) -0.23 (-0.37, -0.09) 113 2 0.00% CCB -0.36 (-1.16, 0.45) Centrally-acting vasodilator -0.20 (-0.97, 0.57) -0.20 (-0.91, 0.51) 23 1 Aldosterone antagonist 0.32 (0.15, 0.49) 0.31 (0.09, 0.53) 524 6 76.8% Renin inhibitor -0.22 (-1.13, 0.69)

BP lowering agents vs ACE inhibitor ARB -0.15 (-0.54, 0.24) CCB -0.28 (-1.04, 0.48) -0.28 (-0.97, 0.41) 24 1 Centrally-acting vasodilator -0.12 (-0.94, 0.69) Aldosterone antagonist 0.40 (0.08, 0.71) Renin inhibitor -0.14 (-1.01, 0.73)

BP lowering agents vs ARB CCB vs ARB -0.13 (-0.99, 0.72) Centrally-acting vasodilator 0.02 (-0.81, 0.85) Aldosterone antagonist 0.54 (0.20, 0.88) Renin inhibitor 0.01 (-0.95, 0.96)

BP lowering agents vs CCB Centrally-acting vasodilator 0.16 (-0.96, 1.27) Aldosterone antagonist 0.68 (-0.15, 1.50) Renin inhibitor 0.14 (-0.28, 0.56) 0.14 (-0.14, 0.42) 74 1

BP lowering agents vs centrally-acting vasodilator Aldosterone antagonist 0.52 (-0.27, 1.31) Renin inhibitor -0.02 (-1.21, 1.18)

BP lowering agents vs aldosterone antagonist Renin inhibitor -0.54 (-1.46, 0.39)

Overall BPLP vs control 0.10 (-0.06, 0.27) 0.10 (-0.11, 0.31) 861 13 85.8%

-2 -1 0 1 2 favours Treatment favours Comparator

Supplemental Figure 10. Network estimates of blood pressure lowering drugs effects on serum potassium.

ACE= angiotensin converting enzyme. ARB= angiotensin receptor blockers. BP= blood pressure. BPLP= blood pressure lowering pharmacotherapy. CCB= calcium-channel blockers. MD= mean difference (mEq/L).

Diastolicblood pressure Heart rate Potassium concentration

ACE inhibitor (n = 467) Renin inhibitor β blocker(n=134) CCB (n = 103) ARB 1 (n = 40) α blocker Calcium-channel blocker (n = 56) 1 1 (n = 10) (n = 44) 1 Centrally-acting vasodilator 1 1 2 2 (n = 11) 1 2 1 1 Renin inhibitor 5 2 (n = 49) 1 β blocker Placebo 1 ACE inhibitor ACE inhibitor (n = 206) (n = 340) 4 (n = 63) (n = 113) 1 3 Control 1 (n = 426) 6 Aldosterone antagonist 1 3 1 CCB 2 (n = 78) 2 2 (n= 141) Aldosterone antagonist Placebo Centrally acting vasodilator ARB ARB (n = 264) (n = 546) (n = 11) (n = 204) (n = 61)

Supplemental Figure 11. Network of treatment comparison of blood pressure lowering agents effects on diastolic blood pressure, heart rate and potassium concentration outcomes.

The size of the node corresponds to the number of trials. The thickness of the line connecting two treatments corresponds to the number of patients. Numbers next to each line represent the number of trials that compared the connected treatments. Numbers in brackets correspond to the number of patients. ACE= angiotensin converting enzyme. ARB= angiotensin receptor blockers. CCB= calcium-channel blockers.

Network Direct Patients Trials I2 Treatment vs comparator MD (95% CI) MD (95% CI) (N) (N)

BP lowering agents vs placebo ACE inhibitors -6.3 (-13.1, 0.5) -8.3 (-17.8, 1.3) 132 2 0% ARB 2.0 (-10.0, 14.0) 2.0 (-9.0, 13.0) 56 1 - β blockers -10.2 (-15.1, -5.3) -9.9 (-13.3, -6.4) 149 2 0% CCB -6.8 (-17.8, 4.2) -13.6 (-37.4, 10.2) 129 2 87% Aldosterone antagonists -13.5 (-19.6, -7.4) -14.6 (-20.0, -9.3) 92 2 0%

BP lowering agents vs ACE inhibitors ARB 8.3 (-5.5, 22.1) β blockers -3.9 (-10.1, 2.3) -4.7 (-11.8, 2.4) 102 1 - CCB -0.5 (-10.9, 9.9) -0.7 (-19.4, 18.0) 24 1 - Aldosterone antagonists -7.2 (-16.9, 2.5)

BP lowering agents vs ARB β blockers -12.2 (-25.1, 0.7) CCB -8.8 (-25.1, 7.5) Aldosterone antagonists -15.5 (-28.9, -2.1)

BP lowering agents vs β blockers CCB 3.4 (-8.2, 15.0) Aldosterone antagonists -3.3 (-11.2, 4.6)

BP lowering agent vs CCB Aldosterone antagonists -6.7 (-20.6, 7.2)

Overall BP lowering agents vs Placebo -9.1 (-13.3, -5.0) -9.0 (-13.5, -4.5) 558 9 54%

-25 -4 0 4 25 Favours Treatment Favours Comparator

Supplemental Figure 12. Network estimates of blood pressure lowering drugs effects on systolic blood pressure in trials with low risk of bias (n=9).

BP= blood pressure. ACE= angiotensin converting enzyme. ARB= angiotensin receptor blockers. CCB= calcium-channel blockers. MD= mean difference (mmHg).

Supplemental References

1. London GM, Pannier B, Guerin AP, Marchais SJ, Safar ME, Cuche J-L: Cardiac hypertrophy, aortic compliance, peripheral resistance, and wave reflection in end- stage renal disease. Comparative effects of ACE inhibition and calcium channel blockade. Circulation, 90: 2786-2796, 1994

2. Robson R, Collins J, Johnson R, Kitching R, Searle M, Walker R, Douglas J, Leary J, Whalley G, Sharpe N: Effects of simvastatin and enalapril on serum lipoprotein concentrations and left ventricular mass in patients on dialysis. The Perfect Study Collaborative Group. Journal of nephrology, 10: 33-40, 1997

3. Nakamoto H, Kanno Y, Okada H, Suzuki H: Erythropoietin resistance in patients on continuous ambulatory peritoneal dialysis. Adv Perit Dial, 20: 111-116, 2004

4. Suzuki H, Kanno Y, Kaneko K, Kaneko M, Kotaki S, Mimura T, Takane H: Comparison of the effects of angiotensin receptor antagonist, angiotensin converting enzyme inhibitor, and their combination on regression of left ventricular hypertrophy of diabetes type 2 patients on recent onset hemodialysis therapy. Therapeutic Apheresis and Dialysis, 8: 320-327, 2004

5. Matsumoto N, Ishimitsu T, Okamura A, Seta H, Takahashi M, Matsuoka H: Effects of imidapril on left ventricular mass in chronic hemodialysis patients. Hypertension research, 29: 253, 2006

6. Yu W-C, Lin Y-P, Lin I-F, Chuang S-Y, Chen C-H: Effect of ramipril on left ventricular mass in normotensive hemodialysis patients. American journal of kidney diseases, 47: 478-484, 2006

7. Zannad F, Kessler M, Lehert P, Grünfeld J, Thuilliez C, Leizorovicz A, Lechat P, Investigators F: Prevention of cardiovascular events in end-stage renal disease: results of a randomized trial of fosinopril and implications for future studies. Kidney international, 70: 1318-1324, 2006

8. Ordaz-Medina SM, González-Plascencia J, del Campo FM, Rojas-Campos E, Montañez-Fernández JL, Espinoza-Gómez F, Cueto-Manzano AM: Is systemic inflammation of hemodialysis patients improved with the use of enalapril? Results of a randomized, double-blinded, placebo-controlled clinical trial. ASAIO Journal, 56: 37-41, 2010

9. Yilmaz R, Altun B, Kahraman S, Ozer N, Akinci D, Turgan C: Impact of amlodipine or ramipril treatment on left ventricular mass and carotid intima-media thickness in nondiabetic hemodialysis patients. Renal failure, 32: 903-912, 2010 33

10. Reyes-Marín FA, Calzada C, Ballesteros A, Amato D: Comparative study of enalapril vs. losartan on residual renal function preservation in automated peritoneal dialysis. A randomized controlled study. Rev Invest Clin, 64: 315-321, 2012

11. Ottosson P, Attman P-O, Ågren A-C, Samuelsson O: Candesartan cilexetil in haemodialysis patients. Clinical drug investigation, 23: 545-550, 2003

12. Suzuki H, Kanno Y, Sugahara S, Ikeda N, Shoda J, Takenaka T, Inoue T, Araki R: Effect of angiotensin receptor blockers on cardiovascular events in patients undergoing hemodialysis: an open-label randomized controlled trial. American Journal of Kidney Diseases, 52: 501-506, 2008

13. Cice G, Di Benedetto A, D'Isa S, D'Andrea A, Marcelli D, Gatti E, Calabrò R: Effects of telmisartan added to angiotensin-converting enzyme inhibitors on mortality and morbidity in hemodialysis patients with chronic heart failure: a double-blind, placebo-controlled trial. Journal of the American College of Cardiology, 56: 1701-1708, 2010

14. Iseki K, Arima H, Kohagura K, Komiya I, Ueda S, Tokuyama K, Shiohira Y, Uehara H, Toma S, Group OCTiOPUO: Effects of angiotensin receptor blockade (ARB) on mortality and cardiovascular outcomes in patients with long-term haemodialysis: a randomized controlled trial. Nephrology Dialysis Transplantation, 28: 1579-1589, 2013

15. Peters CD, Kjaergaard KD, Jensen JD, Christensen KL, Strandhave C, Tietze IN, Novosel MK, Bibby BM, Jensen LT, Sloth E: No significant effect of angiotensin II receptor blockade on intermediate cardiovascular end points in hemodialysis patients. Kidney international, 86: 625-637, 2014

16. Satirapoj B, Leelasiri K, Supasyndh O, Choovichian P: Short-term administration of an angiotensin II receptor blocker in patients with long-term hemodialysis patients improves insulin resistance. J Med Assoc Thai, 97: 574-581, 2014

17. Sun F, Song Y, Liu J, Ma LJ, Shen Y, Huang J, Zhou YL: Efficacy of losartan for improving insulin resistance and vascular remodeling in hemodialysis patients. Hemodialysis International, 20: 22-30, 2016

18. Meltzer VN, Goldberg AP, Tindira CA, Naumovich AD, Harter HR: Effects of prazosin and propranolol on blood pressure and plasma lipids in patients undergoing chronic hemodialysis. Am J Cardiol, 53: 40A-45A, 1984

19. Cice G, Di AB, Tagliamonte E, Ferrara L, Sorice P, Iacono A: Effectiveness and tolerability of bisoprolol vs. nifedipine in uremic patients with ischemic cardiopathy in dialysis treatment. Cardiologia (Rome, Italy), 42: 397-403, 1997

20. Cice G, Tagliamonte E, Ferrara L, Di AB, Iacono A: Complex ventricular arrhythmias and carvedilol: efficacy in hemodialyzed uremic patients. Cardiologia (Rome, Italy), 43: 597-604, 1998

21. Cice G, Ferrara L, Di Benedetto A, Russo PE, Marinelli G, Pavese F, Iacono A: Dilated cardiomyopathy in dialysis patients--beneficial effects of carvedilol: a double-blind, placebo-controlled trial. J Am Coll Cardiol, 37: 407-411, 2001

22. Agarwal R, Sinha AD, Pappas MK, Abraham TN, Tegegne GG: Hypertension in hemodialysis patients treated with atenolol or lisinopril: a randomized controlled trial. Nephrology Dialysis Transplantation, 29: 672-681, 2014

23. Roberts MA, Pilmore HL, Ierino FL, Badve SV, Cass A, Garg AX, Isbel NM, Krum H, Pascoe EM, Perkovic V: The β-Blocker to Lower Cardiovascular Dialysis Events (BLOCADE) feasibility study: a randomized controlled trial. American Journal of Kidney Diseases, 67: 902-911, 2016

24. London GM, Marchais SJ, Guerin AP, Metivier F, Safar ME, Fabiani F, Froment L: Salt and water retention and calcium blockade in uremia. Circulation, 82: 105- 113, 1990

25. Cice G, Tagliamonte E, Ferrara L, Di AB, Sorice P, Iacono A: Efficacy of diltiazem in uremic hemodialyzed patients with isolated diastolic dysfunction and dialysis hypotensive crisis. Cardiologia (Rome, Italy), 44: 289-294, 1999

26. Tepel M, Hopfenmueller W, Scholze A, Maier A, Zidek W: Effect of amlodipine on cardiovascular events in hypertensive haemodialysis patients. Nephrology Dialysis Transplantation, 23: 3605-3612, 2008

27. Hausberg M, Tokmak F, Pavenstädt H, Krämer BK, Rump LC: Effects of moxonidine on sympathetic nerve activity in patients with end-stage renal disease. Journal of hypertension, 28: 1920-1927, 2010

28. Gross E, Rothstein M, Dombek S, Juknis HI: Effect of spironolactone on blood pressure and the renin-angiotensin-aldosterone system in oligo-anuric hemodialysis patients. American journal of kidney diseases, 46: 94-101, 2005

29. Taheri S, Mortazavi M, Shahidi S, Pourmoghadas A, Garakyaraghi M, Seirafian S, Eshaghian A, Ghassami M: Spironolactone in chronic hemodialysis patients 35

improves cardiac function. Saudi Journal of Kidney Diseases and Transplantation, 20: 392, 2009

30. Vukusich A, Kunstmann S, Varela C, Gainza D, Bravo S, Sepulveda D, Cavada G, Michea L, Marusic ET: A randomized, double-blind, placebo-controlled trial of spironolactone on carotid intima-media thickness in nondiabetic hemodialysis patients. Clinical Journal of the American Society of Nephrology: CJN. 09421209, 2010

31. Taheri S, Mortazavi M, Pourmoghadas A, Seyrafian S, Alipour Z, Karimi S: A prospective double-blind randomized placebo-controlled clinical trial to evaluate the safety and efficacy of spironolactone in patients with advanced congestive heart failure on continuous ambulatory peritoneal dialysis. Saudi Journal of Kidney Diseases and Transplantation, 23: 507, 2012

32. Зарипова И, Есаян А, Нимгирова А, Каюков И: Динамика концентрации калия в сыворотке крови на фоне «Трехкомпонентной» и «Двухкомпонентной» фармакологической блокады ренин_ангиотензин_ альдостероновой системы у больных на постоянном гемодиализе. Нефрология, 15, 2011

33. Matsumoto Y, Mori Y, Kageyama S, Arihara K, Sugiyama T, Ohmura H, Yakushigawa T, Sugiyama H, Shimada Y, Nojima Y: Spironolactone reduces cardiovascular and cerebrovascular morbidity and mortality in hemodialysis patients. Journal of the American College of Cardiology, 63: 528-536, 2014

34. Ni X, Zhang J, Zhang P, Wu F, Xia M, Ying G, Chen J: Effects of spironolactone on dialysis patients with refractory hypertension: a randomized controlled study. The Journal of Clinical Hypertension, 16: 658-663, 2014

35. Feniman-De-Stefano GMM, Zanati-Basan SG, De Stefano LM, Xavier PS, Castro AD, Caramori JCT, Barretti P, Jorge da Silva Franco R, Martin LC: Spironolactone is secure and reduces left ventricular hypertrophy in hemodialysis patients. Therapeutic advances in cardiovascular disease, 9: 158-167, 2015

36. Ito Y, Mizuno M, Suzuki Y, Tamai H, Hiramatsu T, Ohashi H, Ito I, Kasuga H, Horie M, Maruyama S: Long-term effects of spironolactone in peritoneal dialysis patients. Journal of the American Society of Nephrology: ASN. 2013030273, 2013

37. Walsh M, Manns B, Garg AX, Bueti J, Rabbat C, Smyth A, Tyrwhitt J, Bosch J, Gao P, Devereaux P: The safety of eplerenone in hemodialysis patients: a noninferiority randomized controlled trial. Clinical Journal of the American Society of Nephrology: CJN. 12371214, 2015 36

38. Lin C, Zhang Q, Zhang H, Lin A: Long‐term effects of low‐dose spironolactone on chronic dialysis patients: a randomized placebo‐controlled study. The Journal of Clinical Hypertension, 18: 121-128, 2016

39. Ito T, Ishikawa E, Fujimoto N, Okubo S, Ito G, Ichikawa T, Nomura S, Ito M: Effects of aliskiren on blood pressure and humoral factors in hypertensive hemodialysis patients previously on angiotensin II receptor antagonists. Clinical and Experimental Hypertension, 36: 497-502, 2014

40. Kuriyama S, Yokoyama K, Hara Y, Sugano N, Yokoo T, Hosoya T: Effect of aliskiren in chronic kidney disease patients with refractory hypertension undergoing hemodialysis: a randomized controlled multicenter study. Clinical and experimental nephrology, 18: 821-830, 2014