Viewer Blinded to Treatment Allocation

CLINICAL RESEARCH www.jasn.org

A Trial of Extending Hemodialysis Hours and Quality of Life

† ‡ | Meg J. Jardine,* Li Zuo, Nicholas A. Gray ,§ Janak R. de Zoysa ,¶ †† Christopher T. Chan ,** Martin P. Gallagher ,* Helen Monaghan,* ‡‡ | †††† Stuart M. Grieve , §§ Rajesh Puranik, ¶¶ Hongli Lin,*** Josette M. Eris , ‡‡‡ |||||| Ling Zhang, Jinsheng Xu,§§§ Kirsten Howard, ¶¶¶ Serigne Lo,* Alan Cass,**** and †††† Vlado Perkovic,* on behalf of the ACTIVE Dialysis Steering Committee

*The George Institute for Global Health, ††Concord Clinical School, ‡‡Sydney Translational Imaging Laboratory, Charles Perkins Centre, Sydney Medical School, and ||||||School of Public Health, University of Sydney, Sydney, Australia; †Department of Renal Medicine, Concord Repatriation General Hospital, Sydney, Australia; ‡Department of Nephrology, Peking University People’s Hospital, Beijing, China; §Department of Renal Medicine, Nambour General Hospital, Nambour, Australia; |Sunshine Coast Clinical School, The University of Queensland, Brisbane, Australia; ¶Department of Nephrology, North Shore Hospital, University of Auckland, Auckland, New Zealand, **Division of Nephrology, University Health Network, Toronto, Canada; Departments of §§Radiology, |Cardiology, and †††Renal Medicine, Royal Prince Alfred Hospital, Sydney, Australia; ¶¶Specialist Magnetic Resonance Imaging, Newtown, Australia; ***First Afﬁliated Hospital of Dalian Medical University, Dalian, China; ‡‡‡Department of Nephrology, China-Japan Friendship Hospital, Beijing, China; §§§Fourth Hospital Afﬁliated to Hebei Medical University, Shijiazhuang, China; ¶¶¶Institute for Choice, University of South Australia, Sydney, Australia; ****Menzies School of Health Research, Charles Darwin University, Darwin, Australia; and ††††Department of Renal Medicine, Royal North Shore Hospital, Sydney, Australia

ABSTRACT The relationship between increased hemodialysis hours and patient outcomes remains unclear. We randomized (1:1) 200 adult recipients of standard maintenance hemodialysis from in-center and home-based hemodialysis programs to extended weekly ($24 hours) or standard (target 12–15 hours, maximum 18 hours) hemodialysis hours for 12 months. The primary outcome was change in quality of life from baseline assessed by the EuroQol 5 dimension instrument (3 level) (EQ-5D). Secondary outcomes included medication usage, clinical laboratory values, vascular access events, and change in left ventricular mass index. At 12 months, median weekly hemodialysis hours were 24.0 (interquartile range, 23.6–24.0) and 12.0 (interquartile range, 12.0–16.0) in the extended and standard groups, respectively. Change in EQ-5D score at study end did not differ between groups (mean difference, 0.04 [95% confidence interval, 20.03 to 0.11]; P=0.29). Extended hours were associated with lower phosphate and potassium levels and higher hemoglobin levels. Blood pressure (BP) did not differ between groups at study end. Extended hours were associated with fewer BP-lowering agents and phosphate-binding medications, but were not associated with erythropoietin dosing. In a substudy with 95 patients, we detected no difference between groups in left ventricular mass index (mean difference, 26.0 [95% confidence interval, 214.8 to 2.7] g/m2;P=0.18). Five deaths occurred in the extended group and two in the standard group (P=0.44); two participants in each group withdrew consent. Similar numbers of patients experienced vascular access events in the two groups. Thus, extending weekly hemodialysis hours did not alter overall EQ-5D quality of life score, but was associated with improvement in some laboratory parameters and reductions in medication burden. (Clinicaltrials.gov identifier: NCT00649298).

J Am Soc Nephrol 28: 1898–1911, 2017. doi: https://doi.org/10.1681/ASN.2015111225

Outcomes for patients receiving maintenance dial- Received November 12, 2015. Accepted November 28, 2016. ysis are poor, with mortality rates many times Published online ahead of print. Publication date available at greater than the general population1 and a substan- www.jasn.org. 20 tially reduced quality of life. The rise in mortality Correspondence: Prof. Meg Jardine, The George Institute for and cardiovascular events as kidney function falls,3 Global Health, PO Box M201, Missenden Road, NSW 2050, with amelioration after renal transplantation,4 has Australia. Email: [email protected] driven the hypothesis that increasing dialysis Copyright © 2017 by the American Society of Nephrology

1898 ISSN : 1046-6673/2806-1898 JAmSocNephrol28: 1898–1911, 2017 www.jasn.org CLINICAL RESEARCH exposure will improve outcomes for people with kidney group (Figure 2). First treatment commenced a mean of 37.4 failure. Observational studies from several countries5–9 have (median, 27) days after randomization in the extended group supported this hypothesis and suggest a strong relationship and 24.2 (median, 15) days in the standard group. between weekly dialysis hours and survival. The few random- An estimate of small molecule clearance, predominantly ized studies assessing increasing dialysis hours have suggested urea reduction ratio, was available in 152 (75.1%) participants improvements in a number of surrogate outcomes, but have also according to usual site practices (Supplemental Table 3) and raised concerns about possible adverse effects on dialysis vascu- included measures conducted within the 3-month time win- lar access,10 residual renal function,11 and long-term mortality.12 dow for baseline assessment in 117 participants (Supplemen- Quality of life is substantially worse in patients receiving tal Table 4). The mean urea reduction ratio, the measured dialysis compared with the general population,13 whereas in- single pool Kt/V, and the calculated standardized Kt/V were creased weekly dialysis hours are associated with better quality greater in the extended group than the standard group at each of life in observational studies; however, recent trials have not postrandomization assessment (average difference of 7.1% clearly defined the effects of dialysis hours on this outcome.14,15 [95% confidence interval (95% CI), 4.9 to 9.3; P,0.001], The presence and magnitude of any benefit is central to the 0.48 [95% CI, 0.27 to 0.68; P,0.001], and 1.29 [95% CI, assessment of cost-effectiveness for quality of life outcomes, 0.61 to 1.96; P=0.001], respectively). such as quality-adjusted life years, as extending dialysis hours increases treatment costs in most settings.16 Outcomes Quality of Life Objective In the extended dialysis hours group, quality of life assessed by Weaimed to conduct a prospectiverandomizedtrial to evaluate EQ-5D at 12 months decreased by 0.018, as compared with the effect of increasing weekly hemodialysis hours on quality 0.055 in the standard group (between-group difference at study of life over 12 months compared with standard hemodialysis. endadjustedforbaseline:0.037;95%CI,20.03 to 0.11; P=0.33; Figure 3, Supplemental Figure 1). The results were similar for the prespecified subgroup analyses defined by re- RESULTS gion, dialysis location, and dialysis vintage (Figure 4, P heterogeneity all .0.2). In addition, no effect from increased Patients dialysis hours was seen when EQ-5D scores from all follow-up From 2009 to 2013, 200 participants were randomized from 40 visits throughout the intervention period were considered sites. Baseline characteristics were similar across randomized using a linear mixed model (mean difference, 0.03 [95% CI, groups (Table 1). Treatment was discontinued before 12 20.03 to 0.09]; P=0.30), in the sensitivity analyses excluding months in 16 participants because of death (seven partici- participants who did not complete an end of study quality pants), transplantation (five participants), and withdrawal of of life assessment (mean difference, 0.03; 95% CI, 20.04 to consent (four participants) (Figure 1). Analyzable results from 0.10; P=0.33), or in the post hoc sensitivity analysis adjusted the EuroQol 5 dimension instrument (3 level) (EQ-5D) qual- for baseline presence of chronic heart failure, diabetes, di- ity of life assessments were performed in 195 participants, alysis vintage, or catheter use (mean difference, 0.03; 95% excluding the four who withdrew consent, and one who de- CI, 20.04 to 0.01; P=0.36). There was no suggestion of clined to answer one part of the questionnaire at baseline. trend in the evolution of quality of life scores (Supplemental Figure 2). Delivered Treatment There was a small but statistically significant improvement Median weekly dialysis hours received were doubled in the in the Physical Component Summary (PCS) score with ex- extended group (median, 24 hours; interquartile range tended hours, but no effect on Mental Component Summary [IQR], 24–24 hours) compared with standard group (median, (MCS) score of the 36-Item Short Form Health Survey (SF-36) 12 hours; IQR, 12–16 hours) at 12 months, which was pre- quality of life measure in the prespecified analysis from baseline dominantly achieved by longer treatment sessions rather than to 12 months (between-group difference at 12 months adjusted increased session frequency (Table 2). Both blood and dialysis for baseline: 2.68 [95% CI, 0.39 to 4.97; P=0.02] and 1.53 flow rates were lower in the extended group (Supplemental [95% CI, 21.17 to 4.23; P=0.27], respectively). Both PCS Table 1). Most patients were treated with high flux, synthetic and MCS significantly improved with extended hours when membranes (Supplemental Table 2). No participant was treated all follow-up visits were assessed (Figure 5). with the NxStage system. Patients predominantly dialyzed through the day, although an average of ten participants in the Other Outcomes extended group and three in the standard group were dialyzing Systolic and diastolic BPs were similar between the two groups, overnight at each follow-up visit. Complete and partial ad- whereas the requirement for BP-lowering agents was reduced herence to the randomized intervention was 74%–78% and in the extended group (mean reduction in number of agents, 78%–81%, respectively, in the extended hours groups and 20.29 [95% CI, 20.53 to 0.06]; P=0.01; Table 3). The pro- 95%–98% and 98%–99%, respectively, in the standard hours portion of participants not requiring BP-lowering agents was

J Am Soc Nephrol 28: 1898–1911, 2017 xxxx 1899 CLINICAL RESEARCH www.jasn.org

Table 1. Baseline characteristics of participants of the ACTIVE Dialysis trial Characteristics Standard (n=100) Extended (n=100) Total (n=200) Age at randomization, yr Mean (SD) 51.6 (11.50) 52.1 (12.72) 51.8 (12.10) Median (Q1, Q3) 52.6 (45.2, 60.3) 53.2 (40.7, 62.6) 52.8 (42.0, 61.2) Sex, n (%) Men 70 (70.0) 69 (69.0) 139 (69.5) Women 30 (30.0) 31 (31.0) 61 (30.5) Primary cause of renal disease, n (%) Diabetic nephropathy 34 (34.0) 27 (27.0) 61 (30.5) Hypertension/vascular nephrosclerosis 11 (11.0) 11 (11.0) 22 (11.0) GN 34 (34.0) 41 (41.0) 75 (37.5) Reflux nephrology 5 (5.0) 3 (3.0) 8 (4.0) Polycystic kidney disease 7 (7.0) 5 (5.0) 12 (6.0) Other or unknown 9 (9.0) 13 (13.0) 22 (11.0) Comorbidity, n (%) Diabetes mellitus 39 (39.0) 34 (34.0) 73 (36.5) Hypertension 85 (85.0) 82 (82.0) 167 (83.5) Any cardiovascular disease 32 (32.0) 33 (33.0) 65 (32.5) Symptomatic ischemic heart disease 18 (18.0) 18 (18.0) 36 (18.0) Angina 11 (11.0) 9 (9.0) 20 (10.0) Acute myocardial infarction 2 (2.0) 2 (2.0) 4 (2.0) Previous coronary artery bypass graft/percutaneous 6 (6.0) 7 (7.0) 13 (6.5) transluminal coronary angioplasty Congestive heart failure 13 (13.0) 19 (19.0) 32 (16.0) Cerebrovascular disease 8 (8.0) 7 (7.0) 15 (7.5) Peripheral vascular disease 8 (8.0) 6 (6.0) 14 (7.0) Smoking status, n (%) Never smoked 56 (56.0) 60 (60.0) 116 (58.0) Past cigarette smoker 23 (23.0) 20 (20.0) 43 (21.5) Current cigarette smoker 21 (21.0) 20 (20.0) 41 (20.5) Country, n (%) Australia 28 (28.0) 30 (30.0) 58 (29.0) Canada 6 (6.0) 5 (5.0) 11 (5.5) China 62 (62.0) 62 (62.0) 124 (62.0) New Zealand 4 (4.0) 3 (3.0) 7 (3.5) Ethnicity, n (%) White 24 (24.0) 23 (23.0) 47 (23.5) Asian 66 (66.0) 65 (65.0) 131 (65.5) Other 10 (10.0) 10 (10.0) 20 (10.0) Not reported 0 (0.0) 2 (2.0) 2 (1.0) Number of dialysis sessions per wk, n (%) 2 1 (1.0) 1 (1.0) 2 (1.0) 3 82 (82.0) 86 (86.0) 168 (84.0) $4 17(17.0) 13(13.0) 30(15.0) Total number of hr on dialysis per wk Mean (SD) 14.1 (2.8) 13.6 (2.6) 13.9 (2.7) Median (Q1, Q3) 12.4 (12.0, 16.0) 12.0 (12.0, 15.0) 12.0 (12.0, 15.0) Duration on dialysis at enrolment, median (IQR) in yr 2.63 (0.97–6.74) 2.43 (0.67–5.04) 2.48 (0.72–6.00) Dialysis site at enrolmenta, n (%) Home 12(12.0) 11(11.0) 23(11.5) In-center/satellite 88 (88.0) 89 (89.0) 177 (88.5) Intended dialysis site for study treatmenta, n (%) Home 25(25) 26(26) 51(25.5) In-center/satellite 75 (75.0) 74 (74.0) 149 (74.5) Dialysis access, n (%) Native fistula 81 (81.0) 87 (87.0) 168 (84.0) Synthetic fistula 4(4.0) 3(3.0) 7(3.5)

1900 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 1898–1911, 2017 www.jasn.org CLINICAL RESEARCH

Table 1. Continued

Characteristics Standard (n=100) Extended (n=100) Total (n=200) Tunneled dialysis catheter 14 (14.0) 9 (9.0) 23 (11.5) Other 1(1.0) 1(1.0) 2(1.0) Dialysis cannulation method, n (%) Buttonhole technique 14 (14.0) 22 (22.0) 36 (18.0) Rope ladder technique 71 (71.0) 68 (68.0) 139 (69.5) Not applicableb 15 (15.0) 10 (10.0) 25 (12.5) Predialysis BP, mean (SD) in mmHg Systolic BP 139 (21.0) 141 (20.7) 140 (20.8) Diastolic BP 79 (13.6) 81 (13.9) 80 (13.8) Body mass index, median (IQR) in kg/m2 24.9 (22.5–29.3) 24.3 (21.2–27.9) 24.5 (21.9–29.0) Quality of life measured using EQ-5D, mean (SD) score 0.76 (0.25) 0.79 (0.23) 0.78 (0.24) Left ventricular mass index,c mean (SD) in g/m2 103.64 (30.68) 106.48 (36.39) 105.16 (33.72) aSome patients were recruited from home training programs and so although their dialysis site had been determined they were, strictly speaking, dialyzing within the institutional setting at enrolment. bThese patients were dialyzing via a venous dialysis catheter. cThe cardiac remodeling substudy included 44 participants in the standard arm and 51 in the extended arm. higher in the extended group at every time point in the study extended dialysis (mean potassium reduction, 0.35 [95% CI, after randomization (odds ratio for medication-free BP control 0.20 to 0.50] meq/L P,0.001; SI conversion factor to mmol/L: 1; at 12 months in the extended group, 2.25 [95% CI, 1.20 to and mean phosphate reduction, 0.77 [95% CI, 0.46 to 1.08] 4.19]; P=0.01). mg/dl; P,0.001; SI conversion factor to mmol/l: 0.323), The extended hours group had a higher mean hemoglobin whereas there was a reduction in the number of daily phos- level (average difference, 0.52 [95% CI, 0.18 to 0.86] g/dl; phate binder tablets in the extended dialysis group (mean P=0.003; SI conversion factor to g/L: 10) with no effect on reduction, 0.83 [95% CI, 0.04 to 1.61] tablets; P=0.04). The iron stores or dose of erythropoiesis-stimulating agents (Table proportion of participants requiring no phosphate binder use 3). Mean potassium and phosphate levels were lower with was higher in the extended group at every time point in the study after randomization (odds ratio for medication-free phosphate control at 12 months in the extended group, 5.75 [95% CI 1.85 to 17.81]; P=0.002). There was no overall difference in the change in left ventricular mass index among the 95 participants (described further in Supplemental Table 5) participating in the magnetic resonance imaging (MRI) substudy (mean change at 12 months, 26.018 [95% CI, 214.78 to 2.74] g/m2; P=0.18). Within the prespeciﬁed subgroups, extended dialysis hours was associated with a reduction in left ventricular mass index in participants who had been receiving dialysis for .6 months at baseline compared with those dialyzing for #6 months (mean change of 210.95 [95% CI, 220.45 to 21.44] g/m2; P=0.03, compared with 14.93 [95% CI, 26.87 to 36.72] g/m2; P=0.16, respectively; P heterogeneity =0.02, Supplemental Figure 3).

Safety During the study, the composite cardiovascular outcome occurred eight times in seven participants. Of these, three participants were in the extended group and four in the Figure 1. Participant ﬂow through the study. standard group. Overall, there were seven

J Am Soc Nephrol 28: 1898–1911, 2017 xxxx 1901 1902 LNCLRESEARCH CLINICAL ora fteAeia oit fNephrology of Society American the of Journal www.jasn.org

Table 2. Weekly received total hemodialysis hours and session frequency during the 12 months of study duration

Weekly hr Weekly Session Frequency Session Length

Measure Standard Extended Standard Extended Standard Extended Time Point Median (IQR) mo 0 12.0 (12.0–16.0) 12.0 (12.0–15.0) 3.0 (3.0–3.0) 3.0 (3.0–3.0) 4.0 (4.0–5.0) 4.0 (4.0–4.5) mo 3 12.0 (12.0–16.0) 24.0 (24.0–24.0) 3.0 (3.0–3.0) 3.0 (3.0–4.0) 4.0 (4.0–5.0) 6.0 (6.0–8.0) mo 6 12.0 (12.0–16.0) 24.0 (24.0–24.0) 3.0 (3.0–3.0) 3.0 (3.0–4.0) 4.0 (4.0–4.5) 6.0 (6.0–8.0) mo 9 12.0 (12.0–16.0) 24.0 (22.0–24.0) 3.0 (3.0–3.0) 3.0 (3.0–4.0) 4.0 (4.0–4.5) 6.0 (6.0–8.0) mo 12 12.0 (12.0–16.0) 24.0 (23.6–24.0) 3.0 (3.0–3.0) 3.0 (3.0–4.0) 4.0 (4.0–5.0) 6.0 (6.0–8.0) Mean (SD) mo 0 14.1 (2.7) 13.6 (2.6) 3.2 (0.4) 3.2 (0.5) 4.5 (0.7) 4.3 (0.6) mo 3 14.1 (2.8) 22.7 (4.4) 3.2 (0.4) 3.4 (0.5) 4.4 (0.7) 6.7 (1.4) mo 6 13.9 (2.6) 22.5 (4.8) 3.2 (0.4) 3.4 (0.5) 4.4 (0.7) 6.7 (1.4) mo 9 14.2 (3.9) 22.3 (4.6) 3.2 (0.5) 3.4 (0.6) 4.4 (0.8) 6.6 (1.5) mo 12 14.1 (4.2) 22.1 (4.3) 3.2 (0.5) 3.4 (0.5) 4.4 (1.0) 6.6 (1.5) Proportionsa ,12.0 12.0–18.0 18.1–23.9 24.0–28.0 .28.0 ,12.0 12.0–18.0 18.1–23.9 24.0–28.0 .28.0 2 3 $42 3 $4 ,4.0 4.0–6.0 6.1–8.0 .8.0 ,4.0 4.0–6.0 6.1–8.0 .8.0 mo 0, % 1.0 93.0 5.0 1.0 0.0 1.0 93.0 6.0 0.0 0.0 1.0 82.0 17.0 1.0 86.0 13.0 2.0 97.0 1.0 0.0 2.0 98.0 0.0 0.0 mo 3, % 1.0 94.9 2.0 2.0 0.0 0.0 17.0 3.0 76.0 2.0 0.0 81.6 18.4 0.0 59.0 39.0 2.0 96.9 1.0 0.0 0.0 50.0 46.0 2.0 mo 6, % 3.1 92.8 2.1 0.0 0.0 0.0 19.6 3.1 70.1 3.1 1.0 79.4 17.5 0.0 59.8 36.1 3.1 93.8 0.0 0.0 0.0 49.5 45.4 1.0 mo 9, % 2.1 92.6 2.1 1.1 1.1 0.0 20.4 5.4 68.8 3.2 1.1 77.7 20.2 0.0 58.1 39.8 4.3 92.6 2.1 0.0 0.0 52.7 43.0 2.2 mo 12, % 2.1 91.5 1.1 1.1 1.1 0.0 19.8 5.5 71.4 1.1 0.0 77.7 18.1 0.0 59.3 38.5 4.3 89.4 3.2 0.0 0.0 52.7 42.9 2.2 aProportion of participants in categories of weekly hemodialysis hours, weekly session frequency, and dialysis session duration at each study visit, according to randomized treatment (proportion of missing data not mScNephrol Soc Am J shown). 28: 1898 – 91 2017 1911, www.jasn.org CLINICAL RESEARCH

Figure 2. Total weekly dialysis hours at each visit in participants randomized to standard or extended weekly dialysis hours. The figure depicts the weekly hours at each visit according to randomization, with the size of the circles being proportional to the number of participants practicing the particular number of hours. * indicates the percentage of participants attending the study visit who were fully or partially adherent to the randomized allocation. Partially adherent was defined as receiving between 18 and 24 hours of weekly dialysis (i.e., more than the definition for standard hours but less than the definition for extended hours). FU, follow-up. deaths during the study follow-up (five in the extended group and achieving a large separation in hemodialysis hours, but without two in the standard group, P=0.44), including one death during an improvement in overall quality of life as measured by the dialysis in the extended group. There were 102 serious adverse EQ-5D instrument. Secondary effects upon clinical and labo- events in 65 participants reported with no difference between the ratory parameters with extended hemodialysis hours, and sug- groups (P=0.55, Table 4). gestions of benefit for other measures of quality of life, may be There were 49 dialysis vascular access–related adverse events considered worthwhile by some individuals requiring dialysis occurring in 30 (15%) participants (Table 4). Similar numbers of and could also portend clinical benefits that accrue over longer dialysis vascular access events occurred in each group, including time periods. the prespecified end points of fistula thrombosis, fistula throm- The strengths of the study include the use of a primary out- bosis or intervention, and fistula-related infection (Table 4). come of fundamental importance to patients, the multinational health care settings involved, the generalizable nature of thestudy design, and the central blinding of outcome assessments. Despite DISCUSSION successfully recruiting its target participant number, however, and being the largest trial of extended hemodialysis hours, A This randomized trial successfully allocated participants to Clinical Trial of IntensIVe (ACTIVE) Dialysis remains a relatively 12 months of extended or standard hemodialysis hours, small trial. Other limitations include the lack of feasibility for

J Am Soc Nephrol 28: 1898–1911, 2017 xxxx 1903 CLINICAL RESEARCH www.jasn.org

overall quality of life as measured by the EQ-5D instrument. It remains possible the intervention was too short to assess benefit, although there was no evidence of progres- sion in the repeated measures. Similarly, it needs to be noted that adherence was not complete, particularly in the extended dialysis group. Although comparison is ham- pered by the different ways in which the trials have measured adherence, adherence appears to be similar in the ACTIVE Dial- ysis and Frequent Hemodialysis Network (FHN) trials17 which both appear lower than that reported in the Alberta trial.18 The relatively high quality of life scores at study entry common to ACTIVE Dialysis participants and to participants of home dialysis programs in general, may have limited the study’s potential to measurably improve quality of life but may also explain the association of extended hours with bet- Figure 3. Change in quality of life from study beginning to end was comparable for the ter quality of life previously reported from extended and standard dialysis hours groups. The mean change in EQ-5D score is observational analyses.19 Alternatively, the presented for each group firstly as a scatter plot depicting the individual change for high baseline scores may have limited the each participant and as a box plot representing the median (line), interquartiles (box), capacity of the EQ-5D instrument to detect and 10th and 90th percentiles (whiskers) of change. Positive numbers indicate EQ-5D meaningful improvement in quality of life score was higher at the end of the study than the beginning. (“ceiling effect”), although the instrument has been able to detect difference in ran- blinding participants and investigators, and the measurement domized participants with higher baseline scores,20 and can of cardiac imaging in a subset of participants. distinguish improvements in renal function delivered to peo- These results broaden our understanding of the effects of ple with ESRD through transplantation.21 Moreover, the wide extended hours of hemodialysis on quality of life, a key out- variation in EQ-5D–measured change in quality of life sug- come for patientsreceiving dialysis. The result makesit unlikely gests our study participants did not universally enjoy a high that extending hemodialysis hours will substantially increase quality of life. It remains possible the intervention might be effective in patients with substantially lower quality of life scores, such as those in sicker dialysis populations. Our findings do not conform to those suggested by the Alberta trial, the only similar completed trial at the time ACTIVE Dialysis was initiated, which suggested that extended weekly dialysis hours might lead to an improvement in EQ-5D–defined quality of life, with a nonsignificant increased score of 0.12 (95% CI, 20.01 to 0.25; P=0.06) in 52 participants from randomization to completion of the 6-month intervention. Our population differs from that in the Al- berta trial in that the majority of our participants were not drawn from people Figure 4. The impact of extended dialysis hours on the change in quality of life from study deemed suitable for and selecting home di- beginning to end was comparable for prespecified subgroups defined by region, dialysis lo- alysis, and in differences in the ethnic and cation, and time on dialysis. The figure shows the effect of extended dialysis hours on the primary cultural background of our participants. end point of change in EQ-5D quality of life at 12 months according to prespecified subgroups The EQ-5D instrument is designed for defined by region, dialysis location, and months on dialysis at commencement of the study. and validated in multiple cultural and

1904 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 1898–1911, 2017 www.jasn.org CLINICAL RESEARCH

the FHN trial of increased dialysis frequency within a standard hours regimen23 but not the FHN trial of nocturnal dialysis with extended hours.24 The estimation of minimal important differences in the scores of tools used to assess quality of life is complex and may vary according to the research context and disease state.25,26 There is no clear consensus on the minimal important difference for PCS and MCS scores in the dialysis context, although a difference of 3–5 has been suggested as a minimally important difference in PCS in the literature for CKD27 and chronic disease28—a little larger than the difference we found. The decision about whether these benefits are worth the additional commitment of extended dialysis hours is ultimately an individual one, and will likely vary from person to person. Our findings are broadly consistent with previous data regarding effects on secondary biochemical, clinical, and medication outcomes whilst providing greater preci- sion, but differences were observed for cardiac structure and vascular access event analyses. The ACTIVE Dialysis trial has suggested extended dialysis hours may allow for a reduction in at least some medication requirements, with similar or better control of BP,phosphate, potassium, and hemoglobin. These findings have been in- consistently reported in previous smaller studies, potentially because of insufficient power.18,24 Figure 5. Both PCS and MCS significantly improved with extended hours when all follow-up visits were assessed. Effect of extended hours on (A) physical and (B) mental Surprisingly and in contrast with some 18,23 fi quality of life at each study visit. Figures represent the mean (box) and SD (whiskers) previous studies, we did not nd any score at each time point for participants randomized to extended and standard weekly overall effects on left ventricular mass in- hemodialysis hours. The inserted text provides the estimate of the mean intervention dex. For pragmatic reasons largely related effect over all follow-up visits, adjusted for baseline, with positive values favoring the to access to MRI, the left ventricular mass extended dialysis group. index assessment was a substudy. Baseline and follow-up MRIs on 100 participants would have provided 90% power to ethnic settings, is the most widely used multiattribute utility detect a difference of 8.7 g/m2 in the change in left ventricular instrument in the world, and is the most widely used direct mass index and 80% power to detect a difference of 7.5 g/m2. multiattribute utility instrument in CKD.22 Our prespecified The actual (nonsignificant) difference observed was 6.02 subgroup testing did not indicate the results were influenced by (95% CI, 214.78 to 2.74) g/m2 and comparable with results either dialysis setting or country. Our results do not exclude an in previous similar trials. Differences in left ventricular mass effect on EQ-5D–defined quality of life from differences in index reported in previous extended weekly hemodialysis hours weekly dialysis hours below the median 12 hours practiced trials are 4.4 (95% CI, 210.8 to 2.1) g/m2 in the FHN nocturnal by our control group. trial and 8.1 (95% CI, 216.2 to 20.1) g/m2 in the Alberta trial, We did find statistically significant but clinically modest whereas the FHN trial of frequent weekly hemodialysis ses- evidence of benefit for SF-36 determined physical and mental sions reported a significant difference of 27.1 (95% CI, 212.0 components of quality of life, consistent with the findings of to 22.2) g/m.2,18,29 We did observe significant heterogeneity

J Am Soc Nephrol 28: 1898–1911, 2017 xxxx 1905 CLINICAL RESEARCH www.jasn.org

Table 3. Effect of extended dialysis hours on clinical measures and medication usage assessed at 3-month follow-up visits adjusted for baseline values Standard Extended Difference between Extended and Domain Parameter Baseline Follow-Up Baseline Follow-Up Standard hr P Valuea Mean Mean Mean Mean Dialysis, Mean (SD) (SD) (SD) (SD) Difference (95% CI)a BP control Systolic BP (mmHg) 138.8 (21.02) 137.0 (16.21) 140.7 (20.71) 134.6 (17.48) 22.88 (26.83 to 1.06) 0.15 Diastolic BP (mmHg) 78.73 (13.61) 77.42 (9.37) 80.65 (13.92) 76.37 (10.49) 21.64 (23.89 to 0.61) 0.15 BP-lowering agents 1.56 (1.36) 1.37 (1.05) 1.36 (1.27) 0.97 (1.01) 20.29 (20.53 to 20.06) 0.01 (number) Cardiac remodeling Left ventricular mass 103.6 (30.7) 100.4 (35.2) 106.5 (36.4) 96.8 (36.8) 26.02 (214.78 to 2.74) 0.16 index (g/m2) Electrolytes and Potassium (meq/L) 4.97 (0.73) 5.03 (0.59) 4.96 (0.76) 4.66 (0.55) 20.35 (20.50 to 20.20) ,0.001 mineral metabolism Phosphate (mg/dl) 5.49 (1.83) 5.49 (1.29) 5.75 (1.65) 4.76 (1.14) 20.78 (21.07 to 20.48) ,0.001 PTH (pg/ml) 319.3 (378.8) 373.6 (476.3) 409.5 (519.0) 404.7 (579.6) 211.87 (2135.96 to 112.21) 0.85 Phosphate binders 4.32 (3.68) 4.73 (3.80) 3.67 (3.31) 3.44 (3.67) 20.83 (21.61 to 20.04) 0.04 (number of tablets) Corrected calcium 9.13 (0.94) 9.06 (0.83) 9.08 (0.79) 9.25 (0.65) 0.20 (0.04 to 0.36) 0.01 (mg/dl) Hematology Hemoglobin (g/dl) 11.18 (1.57) 11.16 (1.30) 11.07 (1.48) 11.66 (1.38) 0.52 (0.18 to 0.86) 0.003 ESA dose (EPO units) 7256 (5235) 7108 (5561) 7686 (5684) 6369 (4860) 2955 (22145 to 234) 0.11 Ferritin (ng/ml) 349.5 (308.9) 395.1 (308.2) 420.8 (435.3) 420.4 (359.0) 225.28 (2104.9 to 54.3) 0.53 Transferrin saturation (%) 28.11 (13.87) 26.26 (9.44) 26.51 (14.80) 26.90 (9.98) 21.03 (24.18 to 2.11) 0.52 Body habitus and Weight (kg) 77.63 (19.30) 76.81 (18.34) 73.87 (20.76) 74.41 (19.63) 1.16 (0.21 to 2.10) 0.02 nutrition Waist:hip ratio (cm:cm) 0.94 (0.09) 0.94 (0.09) 0.92 (0.10) 0.93 (0.08) 20.00 (20.01 to 0.01) 0.60 Albumin (g/dl) 3.90 (0.45) 4.04 (0.36) 3.90 (0.47) 3.96 (0.40) 20.07 (20.16 to 0.02) 0.14 The following conversion factors can be used to obtain the stated SI units: potassium, 1 mmol/L; phosphate, 0.323 mmol/L; PTH, 0.105, pmol/L; calcium, 0.25 mmol/L; hemoglobin, 10 g/L; albumin, 10 g/dl. Details of numbers of patients with missing data are listed in Supplemental Table 6. PTH, parathyroid hormone; ESA, erythropoiesis stimulating agent; EPO, erythropoietin. aCardiac remodeling was only assessed at baseline and 12 months and was analyzed using normal linear regression adjusted for baseline value. All other estimates and P values are from a mixed linear regression adjusted by time (categorical) and baseline value. according to dialysis duration. The FHN nocturnal and daily baseline,10,18 greater proportion dialyzing in in-center/satellite trials had appeared to show greater reductions in left ventricular settings, and lower buttonhole cannulation usage (a technique mass in participants with negligible compared with some re- associated with increased infection31). Additionally, reports of sidual urine production, although the differences were not increased access event rates published during the conduct of statistically significant. Our finding that extended hemodial- ACTIVE Dialysis may have led to greater vigilance. Our results ysis hours was associated with reduced left ventricular mass suggest extended hours do not inevitably result in increased index for participants who had been on dialysis for .6 vascular access events. months but not for those with a shorter dialysis exposure, It is timely to reflect on the state of current knowledge of suggest people on dialysis for some time or with reduced re- the effects of extended hemodialysis hours. Feasible ran- sidual renal function may benefit more from extended dialysis domized studies remain difficult for this treatment. Any hours.29 beneficial effect upon overall quality of life as measured No clear effect on the risk of death or cardiovascular by the EQ-5D instrument, or physical quality of life as mea- eventswasobservedat12months.Thenumericalimbalance sured by the SF-36 PCS, may be modest and the interpreta- in deaths is potentially consistent with the unexplained tion of the clinical significance may vary from person to excess of deaths seen in the long-term follow-up of the person. With a total of 339 patients now recruited in the FHN nocturnal trial12 but further information will be pro- ACTIVE Dialysis, FHN nocturnal, and Alberta trials to test vided by planned long-term follow-up of ACTIVE Dialysis the effect of extended dialysis hours, it appears unlikely a participants. randomized trial will ever have the power to prove an effect Although a previous trial10 and some observational upon mortality. The situation may be different for the re- studies6,30 had raised concerns, ACTIVE Dialysis showed lated intervention of increasing session frequency where one no evidence of harm for vascular access events or interven- substantial trial of 245 participants designed to test in- tions. From this perspective, the distinguishing features of creased frequency rather than increased total weekly dialysis the ACTIVE Dialysis trial compared with previous studies hours, showed those randomized to six weekly dialysis ses- are lower dialysis session frequency, more native fistula usage at sions, who actually received total weekly hours similar to

1906 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 1898–1911, 2017 www.jasn.org CLINICAL RESEARCH

Table 4. Safety outcomes Deaths and Serious Adverse Events Standard (n=100), n (%) Extended (n=100), n (%) Total (n=200), n (%) P Value Total number of deaths 2 (2.0) 5 (5.0) 7 (3.5) 0.44 Patients with at least one serious adverse eventa 30 (30.0) 35 (35.0) 65 (32.5) 0.55 Occurrence of serious adverse eventsb 47 55 102 Occurrence of cardiovascular eventsb,c 448 Serious adverse event diagnosisa Acute myocardial infarctiond 1 (1.0) 0 (0.0) 1 (0.5) 1.00 Stroke 2 (2.0) 1 (1.0) 3 (1.5) 1.00 Cardiovascular death 1 (1.0) 2 (2.0) 3 (1.5) 1.00 Noncardiovascular deathe 0 (0.0) 3 (3.0) 3 (1.5) 0.25 Infection (not related to dialysis access) 4 (4.0) 5 (5.0) 9 (4.5) 1.00 Other events (not related to dialysis access) 20 (20.0) 23 (23.0) 43 (21.5) 0.73 ESRD-related 5 (5.0) 7 (7.0) 12 (6.0) Cardiac and peripheral vascular disease 7 (7.0) 3 (3.0) 10 (5.0) Gastrointestinal 4 (4.0) 5 (5.0) 9 (4.5) Musculoskeletal 4 (4.0) 1 (1.0) 5 (2.5) Neurologic 2 (2.0) 3 (3.0) 5 (2.5) Kidney/urologic 0 (0.0) 4 (4.0) 4 (2.0) Medication-related 1 (1.0) 1 (1.0) 2 (1.0) Ophthalmologic 0 (0.0) 2 (2.0) 2 (1.0) Carcinoma 1 (1.0) 0 (0.0) 1 (0.5) Dental 0 (0.0) 1 (1.0) 1 (0.5) Trauma 0 (0.0) 1 (1.0) 1 (0.5) Miscellaneous 0 (0.0) 2 (2.0) 2 (1.0) Fistula-related adverse events Occurrence of fistula-related adverse eventsb 25 24 49 Participants with at least one fistula-related 15 (15.0) 15 (15.0) 30 (15.0) 1.00 adverse eventa Thrombosis of fistula requiring revisiona 2 (2.0) 7 (7.0) 9 (4.5) 0.17 Thrombosis othera 2 (2.0) 0 (0.0) 2 (1.0) 0.50 Stenosis of fistula requiring revisiona 9 (9.0) 7 (7.0) 16 (8.0) 0.80 Stenosis othera 2 (2.0) 3 (3.0) 5 (2.5) 1.00 Infection related to accessa 2 (2.0) 2 (2.0) 4 (2.0) 1.00 Othera 3 (3.0) 2 (2.0) 5 (2.5) 1.00 Prespecified vascular access composites Vascular access failurea,f 4 (4.0) 8 (8.0) 12 (6.0) 0.26 Vascular access intervention or failurea 12 (12.0) 14 (14.0) 26 (13.0) 0.60 Vascular access infectiona 2 (2.0) 2 (2.0) 4 (2.0) 1.00 aCounts correspond to the number of subjects who experienced a specific serious adverse event at least once. bCounts correspond to the number of events. cCardiovascular events include acute myocardial infarction, stroke, and cardiovascular death. dIncluding one fatal acute myocardial infarction. eNoncardiovascular deaths included two deaths related to pneumonia and one related to liver failure. fVascular access failure was defined as access thrombosis or fistula revision where revision is defined as a procedure that requires a new anastomosis. those of ACTIVE control participants, had relatively better outcomes to be detected, such as the cluster randomized left ventricular mass at 12 months32 and a significant survival Time to Reduce Mortality in End Stage Renal Disease benefit at a median of 3.6 years in observational follow-up.33 (TiME) trial.34 Clearly, longer follow-up of cohorts is essential, to clarify long-term mortality and any legacy effects upon secondary outcomes. Future studies might focus upon clinical set- CONCISE METHODS tings where short dialysis hours are routine, producing a relatively large increment for additional hours, and could Full details of the study design have been described elsewhere.19 target hospitalization and medication outcomes to make In brief, ACTIVE Dialysis was a multicenter, parallel-arm, ran- an economic case for extending dialysis hours. Alterna- domized trial comparing the effects of extended with standard tively, very large trials with a small difference in dialysis hemodialysis hours on quality of life and other outcomes over hours may be feasible and allow smaller effects on important 12 months.

J Am Soc Nephrol 28: 1898–1911, 2017 xxxx 1907 CLINICAL RESEARCH www.jasn.org

Participants routine practices, before dialysis and not at the first weekly session Adult patientsrequiring maintenancehemodialysis in 40participating after a long break), and medication usage (BP-lowering, phosphate- centers in Australia, China, Canada, and New Zealand were eligible. binding, and erythropoiesis-stimulating agents). Safety outcomes in- Inclusion criteria19 included incident or prevalent dialysis patients cluded serious adverse events, dialysis vascular access events, and not currently receiving extended hours (.18 hours per week), with a survival. The end points and relevant conversion formulae are de- life expectancy of at least 6 months, and no planned renal transplan- scribed in the prespecified Statistical Analysis Plan.36 tation within 12 months. Participants were recruited from both Quality of life questionnaires were administered via telephone by home- and hospital-based in-center (and satellite) programs, with a central interviewer blinded to treatment allocation. Change in left dialysis setting determined before randomization. Participants were ventricular mass index was assessed at baseline and 12 months in the recruited only if they had decided on their dialysis program (home subset of participants who did not have specific contraindications at versus in-center/satellite), although they could be recruited while centers where cardiac magnetic resonance imaging was available. they were still in training programs for home dialysis. Sites were Cardiac imaging was collected with a 1.5T magnet and analyzed cen- selected for their ability to offer both interventions. Availability of trally according to a prespecified protocol. Adequacy of scanning MRI center was not a criterion for site selection but, where possible, capacity at each site was assessed on a test scan or the first study proximate MRI centers were recruited to perform cardiac MRI testing scan for each site before approval for ongoing scans. Laboratory val- on participants without contraindications. Participants were ex- ues were derived locally. Details of delivered dialysis were collected cluded from cardiac MRI assessment if they had individual contra- contemporaneously from dialysis treatment sheets. Standardized Kt/V indications or were recruited from an MRI-exempted site. was calculated using a fixed volume model.37 Study visits, including quality of life interviews and clinical and biochemical assessments Randomization and Masking were conducted at 3-monthly intervals, with continuous reporting of Consenting participants were randomized in a 1:1 ratio via acentralweb- serious adverse events and fistula-related adverse events. Sites were based interface using a minimization algorithm stratified according to asked to report all access-related adverse events, with prespecified access geographical region (Australia/New Zealand versus China/Canada), di- events including vascular access failure and the composite of vascular alysis setting (in-center/satellite versus home), and dialysis duration at access failure and vascular access intervention. Vascular access failure study entry (#6 months versus .6 months). Participants and physi- was defined as access thrombosis (including access abandonment) and cians were not blinded to study allocation (as this was not feasible), access revision where revision was defined as a procedure requiring a however, end points were assessed in a blinded fashion. new anastomosis. Vascular access intervention was defined as including any radiologic or surgical access intervention and not just those Intervention requiring a new anastomosis. Participants assigned to standard hemodialysis hours were scheduled to dialyze for at least 12 and no more than 18 hours per week. Those Ethics and Oversight assigned to extended hours were scheduled to undergo hemodialysis The study was approved by the EthicsCommitteeofNorthernSydney for at least 24 hours per week. By protocol, both groups received a Central Coast Health (reference: HREC/09/HARBR/26), with each minimum of three sessions per week, but beyond that there was no center obtaining additional approval in accordance with local practice. requirements on session frequency or duration, as long as the schedule Written consent was obtained from all participants, who received no fulfilled the requirements for the allocated total weekly hemodialysis reimbursement. The study was overseen by an independent Data hours. All other aspects of care were provided according to routine site Safety Monitoring Board, which undertook regular review of prede- practice, including the prevailing practices of time-determined ses- fined safety parameters and overall study conduct. sion prescription regardless of solute clearance measures, with a summary of expectations for routine care and monitoring of patients Statistical Analyses provided as guidance.19 The study was designed to enroll 200 patients to provide .90% power After randomization, participants were to commence allocated to detect an absolute difference of 0.10 in baseline-adjusted health- treatment within 1 month and continued for 12 months unless they related quality of life over 12 months between the intervention and died, withdrew consent, or were transferred to a setting that could control groups. The power calculation assumed a mean EQ-5D score not offer both therapies. Participants unable or unwilling to adhere of 0.70 in the standard dialysis group, a common SD of 0.22, and two- to randomized treatment were asked to continue planned study sided a=0.05. These assumptions were on the basis of a previous assessments. study of extended dialysis with a comparable patient population.18 The study also had 90% power with 100 participants completing Outcome Measures cardiac MRI imaging to detect a difference between the groups in The primary study outcome was the difference in quality of life at the change in left ventricular mass index from baseline to 12 months study end, adjusted for baseline, as measured by the EQ-5D.35 of 8.7 g/m2. The differences in left ventricular mass index in previous Secondary end points included cardiovascular effects (change in randomized controlled trials of extended weekly hemodialysis hours left ventricular mass index, change in BP,cardiovascular events), other have been means of 24.4 (95% CI, 210.8 to 2.1) and 28.1 (95% CI, 216.2 quality of life measures (including SF-36–based PCS and MCS), pa- to 20.1) g/m,2,18,29 and in a trial of frequent weekly hemodialysis tient adherence, laboratory outcomes (collected according to site sessions was 27.1 (95% CI, 212.0 to 22.2) g/m.2,29

1908 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 1898–1911, 2017 www.jasn.org CLINICAL RESEARCH

All analyses were performed on an intention-to-treat basis. All Baxter Clinical Evidence Council Investigator Initiated research grant continuous end points were analyzed using linear regression adjusted (2012). Additional support came from an unrestricted grant from for baseline values. Methods forhandling missingdataare described in Baxter International Inc., with funding for the vanguard phase from the Supplemental Material. Sensitivity analyses were conducted that an NHMRC program grant (358395). M.J. is supported by a Career excluded those who did not complete an end of study quality of life Development Fellowship from the National Health and Medical assessment. Prespecified subgroups were defined by region (China Research Council of Australia, and the National Heart Foundation. versus Australia, New Zealand, and Canada), dialysis location (in- V.P. is supported by an NHMRC Senior Research Fellowship. center/satellite versus home), and dialysis vintage (#6 months versus The primary results were presented at the American Society of .6 months dialysis duration). Laboratory outcomes were reported Nephrology Renal Week Conference, Philadelphia, PA, November 11– for all follow-up visits (adjusted for baseline values) on the rationale 16, 2014. that these outcomes might be expected to change promptly in re- The funding sources had no role in the design and conduct of the sponse to the intervention. Secondary outcomes repeatedly measured study, the collection, management, analysis, and interpretation of the over time were modeled using a linear mixed model. Binary repeated data, nor the preparation, review, or approval of the manuscript. measurement was analyzed with logistic generalized estimating equa- ACTIVE Dialysis Steering Committee members: Hongli Lin (First tions and count end points with a negative binomial model. Post hoc Affiliated Hospital of Dalian Medical University, Dalian, China), sensitivity analyses calculated the primary outcome adjusted for base- Josette Eris (Concord Repatriation General and Royal Prince Alfred line characteristics of chronic heart failure, diabetes mellitus, dialysis Hospitals, Sydney, Australia), Paul Snelling (Concord Repatriation vintage, and dialysis catheter use. All analyses were conducted using General and Royal Prince Alfred Hospitals, Sydney, Australia), Ling SAS Enterprise Guide 5.1 (SAS Institute Inc.). Zhang (China-Japan Friendship Hospital, Beijing, China), Jinsheng Xu (Fourth Hospital Affiliated to Hebei Medical University, Shijiazhuang, China), Junli Zhang (The Chinese PLA Shanghai 85th Role of the Funding Source The design, conduct, analysis, and interpretation of the ACTIVE Hospital, Shanghai, China), Carmel M. Hawley (Princess Alexandra Dialysis trial was the sole provenance of the ACTIVE Dialysis Steering Hospital, Brisbane, Australia), Carolyn Van Eps (Princess Alexandra Committee. The study funders had no role in the study design, col- Hospital, Brisbane, Australia), Bruce A. Cooper (Royal North Shore fi lection, analysis, and interpretation of the data, the writing of the Hospital, Sydney Australia), Yinhui Li (Fourth Hospital Af liated to – report, nor the decision to submit the finished manuscript for pub- Jilin University FAW General Hospital, Changchun, China), Michael lication. The corresponding author had full access to all of the data in G. Suranyi (Liverpool Hospital, Sydney, Australia), Jeffrey Wong the study and had final responsibility for the decision to submit the (Liverpool Hospital, Sydney, Australia), Li Zuo (Peking University paper for publication. First Hospital, Beijing, China), Yuqing Chen (Peking University First Hospital, Beijing, China),Ying Li (Third Hospital Affiliated to Hebei Medical University, Shijiazhuang, China), Nicholas Gray (Nambour General Hospital, Nambour, Australia), Christopher T. Chan CONCLUSIONS (Toronto General Hospital, Toronto, Canada), Janak de Zoysa (North Shore Hospital and Auckland City Hospital, Auckland, New In conclusion, the extended hours dialysis regimen used in Zealand), Mei Wang (Peking University People’s Hospital, Beijing, ACTIVE Dialysis does not improve overall quality of life over China), Jiang Lei (Peking University Shenzhen Hospital, Shenzhen, 12 months, as measured by the EQ-5D instrument, but may China), Zhangsuo Liu (The First Affiliated Hospital of Zhengzhou improve specific subdomains of quality of life, as measured by University, Zhengzhou, China), Paul Collett (Sydney Adventist Hos- SF-36, and may therefore be considered worthwhile by some pital, Sydney, Australia), Lanfen Xue (Shijiazhuang First People’s individuals. Improvements in BP control, phosphate control, Hospital, Shijiazhuang, China), Nan Chen (Shanghai Ruijin Hospital, and hemoglobin levels were observed, but the long-term ef- Shanghai, China), Yonghui Mao, (Beijing Hospital, Beijing, China), fects on cardiovascular events and death remain to be clarified. Rongshan Li (The Second Affiliated Hospital of Shanxi Medical University, Taiyuan, China), Caili Wang (First Affiliated Hospital of Inner Mongolia, Baotou Medical College, Baotou, China), Wen Huang ACKNOWLEDGMENTS (Beijing Tongren Hospital, Beijing, China), Ping Fu (West China Hos- pital, Sichuan University, Chengdu, China), Eugenia Pedagogos (Royal We gratefully acknowledge the contribution of the A Clinical Trial of Melbourne Hospital, Melbourne, Australia), Kai Wang (Civil Aviation IntensIVe (ACTIVE) Dialysis Investigators and Study Coordinators, General Hospital, Beijing, China), Shengrong Zhang (Beijing Hospital of the project coordinating central team and the magnetic resonance Traditional Chinese Medicine, Beijing, China), Dwarakanathan imaging centers (see Supplemental Material), and most particularly, Ranganathan (Royal Brisbane Hospital, Brisbane, Australia), Jijun Li the study participants. M.J. had full access to all of the data in the (First Affiliated Hospital of Chinese PLA General Hospital, Beijing, study and takes responsibility for the integrity of the data and the China), Bicheng Lui (Zhongda Hospital, Southeast University, accuracy of the data analysis. Nanjing, China), Michael Copland (Vancouver General Hospital, This work was supported by a National Health and Medical Re- University of British Columbia, Vancouver, Canada), Jeffrey Perl search Council (NHMRC) of Australia project grant (571045), and a (St. Michael’s Hospital, Toronto, Canada), Peter G. Kerr (Monash

J Am Soc Nephrol 28: 1898–1911, 2017 xxxx 1909 CLINICAL RESEARCH www.jasn.org

Medical Centre, Melbourne, Australia), Rebecca Pellicano (Monash and short daily compared to conventional hemodialysis: A USRDS Medical Centre, Melbourne, Australia), Girish Talaulikar (Canberra study. Kidney Int 76: 984–990, 2009 9. Weinhandl ED, Liu J, Gilbertson DT, Arneson TJ, Collins AJ: Survival in Hospital, Canberra, Australia), Richard Yu (Royal Hobart Hospital, daily home hemodialysis and matched thrice-weekly in-center hemo- Hobart, Australia), Neil Boudville (Royal Perth Hospital, Perth, dialysis patients. JAmSocNephrol23: 895–904, 2012 Australia), Peter Mount (Austin Health, Melbourne, Australia), 10. Suri RS, Larive B, Sherer S, Eggers P, Gassman J, James SH, Lindsay RM, Murthy Divi (Gold Coast Hospital), Alan Cass (The George Institute Lockridge RS, Ornt DB, Rocco MV, Ting GO, Kliger AS; Frequent He- for Global Health, Sydney, Australia), John Agar (Geelong Hospital, modialysis Network Trial Group: Risk of vascular access complications – Geelong, Australia), Kirsten Howard (University of Sydney, Australia), with frequent hemodialysis. JAmSocNephrol24: 498 505, 2013 11. Daugirdas JT, Greene T, Rocco MV, Kaysen GA, Depner TA, Levin NW, Ashley Irish (Royal Perth Hospital, Perth, Australia), Martin P Chertow GM, Ornt DB, Raimann JG, Larive B, Kliger AS; FHN Trial Gallagher (The George Institute for Global Health and Concord Group: Effect of frequent hemodialysis on residual kidney function. Clinical School, University of Sydney, Sydney, Australia), Meg J Kidney Int 83: 949–958, 2013 Jardine (The George Institute for Global Health, University of 12. Rocco MV, Daugirdas JT, Greene T, Lockridge RS, Chan C, Pierratos A, Sydney, Sydney, Australia, Sydney Adventist Hospital, Sydney, Lindsay R, Larive B, Chertow GM, Beck GJ, Eggers PW, Kliger AS; FHN Trial Group: Long-term effects of frequent nocturnal hemodialysis on Australia, and Concord Repatriation General Hospital, Sydney, mortality: The Frequent Hemodialysis Network (FHN) nocturnal trial. Australia), and Vlado Perkovic (The George Institute for Global Am J Kidney Dis 66: 459–468, 2015 Health, University of Sydney, Sydney, Australia and Royal North 13. Fukuhara S, Lopes AA, Bragg-Gresham JL, Kurokawa K, Mapes DL, Shore Hospital, Sydney, Australia). Akizawa T, Bommer J, Canaud BJ, Port FK, Held PJ; Worldwide Dialysis Outcomes and Practice Patterns Study: Health-related quality of life among dialysis patients on three continents: The dialysis outcomes and practice patterns study. Kidney Int 64: 1903–1910, 2003 DISCLOSURES 14. Pierratos A: Nocturnal home haemodialysis: An update on a 5-year M.J. hasreceived an unconditional researchgrant from GambroAB.N.G. has experience. Nephrol Dial Transplant 14: 2835–2840, 1999 received speaker’s fees from Baxter Healthcare. J.d.Z. has received an unre- 15. Van Eps CL, Jeffries JK, Johnson DW, Campbell SB, Isbel NM, Mudge stricted educational grant from Baxter Healthcare. C.T.C. has received funding DW, Hawley CM: Quality of life and alternate nightly nocturnal home from the National Institute of Diabetes and Digestive and Kidney Diseases, the hemodialysis. Hemodial Int 14: 29–38, 2010 Baxter Extramural Grant Program, and the Clinical Evidence Council pro- 16. Komenda P, Gavaghan MB, Garﬁeld SS, Poret AW, Sood MM: An gram. V.P. has previously served on a Baxter Extramural Grant Committee. economic assessment model for in-center, conventional home, and more frequent home hemodialysis. Kidney Int 81: 307–313, 2012 17. Rocco MV, Daugirdas JT, Greene T, Lockridge RS, Chan C, Pierratos A, Lindsay R, Larive B, Chertow GM, Beck GJ, Eggers PW, Kliger AS; FHN REFERENCES Trial Group: Effects of randomization to frequent nocturnal hemodialysis on long-term mortality: Frequent hemodialysis network nocturnal 1. U.S. Renal Data System: USRDS 2013 annual data report: Atlas of trial [Abstract]. JAmSocNephrol24: 443A, 2013 chronic kidney disease and End-Stage Renal Disease in the United 18. Culleton BF, Walsh M, Klarenbach SW, Mortis G, Scott-Douglas N, States. National Institutes of Health, Bethesda, MD, 2013. Quinn RR, Tonelli M, Donnelly S, Friedrich MG, Kumar A, Mahallati H, 2. Mazairac AH, de Wit GA, Penne EL, van der Weerd NC, de Jong B, Hemmelgarn BR, Manns BJ: Effect of frequent nocturnal hemodialysis Grooteman MP, van den Dorpel MA, Buskens E, Dekker FW, Nubé MJ, vs conventional hemodialysis on left ventricular mass and quality of life: – Ter Wee PM, Boeschoten EW, Bots ML, Blankestijn PJ; CONTRAST A randomized controlled trial. JAMA 298: 1291 1299, 2007 investigators: Changes in quality of life over time–Dutch haemodialysis 19. Jardine MJ, Zuo LI, Gray NA, de Zoysa J, Chan CT, Gallagher MP, patients and general population compared. Nephrol Dial Transplant Howard K, Hertier S, Cass A, Perkovic V; ACTIVE Dialysis Steering Committee: Design and participant baseline characteristics of ‘AClin- 26: 1984–1989, 2011 ical Trial of IntensiVE Dialysis’: The ACTIVE Dialysis Study. Nephrology 3. Gansevoort RT, Matsushita K, van der Velde M, Astor BC, Woodward M, (Carlton) 20: 257–265, 2015 Levey AS, de Jong PE, Coresh J; Chronic Kidney Disease Prognosis 20. Varnﬁeld M, Karunanithi M, Lee CK, Honeyman E, Arnold D, Ding H, Consortium: Lower estimated GFR and higher albuminuria are associ- Smith C, Walters DL: Smartphone-based home care model improved ated with adverse kidney outcomes. A collaborative meta-analysis of use of cardiac rehabilitation in postmyocardial infarction patients: Re- general and high-risk population cohorts. Kidney Int 80: 93–104, 2011 sults from a randomised controlled trial. Heart 100: 1770–1779, 2014 4. McDonald S: Deaths, ANZDATA Registry Report 2011, Australia and 21. Alvares J, Almeida AM, Szuster DA, Gomes IC, Andrade EI, Acurcio FA, New Zealand Dialysis and Transplant Registry. Adelaide, South Australia, Cherchiglia ML: [Factors associated with quality of life in patients in 2011 renal replacement therapy in Brazil]. Cien Saude Colet 18: 1903–1910, 5. Innes A, Charra B, Burden RP, Morgan AG, Laurent G: The effect of 2013 long, slow haemodialysis on patient survival. Nephrol Dial Transplant 22. Wyld M, Morton RL, Hayen A, Howard K, Webster AC: A systematic – 14: 919 922, 1999 review and meta-analysis of utility-based quality of life in chronic kidney 6. Jun M, Jardine MJ, Gray N, Masterson R, Kerr PG, Agar JW, Hawley CM, disease treatments. PLoS Med 9: e1001307, 2012 van Eps C, Cass A, Gallagher M, Perkovic V: Outcomes of extended- 23. Chertow GM, Levin NW, Beck GJ, Depner TA, Eggers PW, Gassman JJ, hours hemodialysis performed predominantly at home. Am J Kidney Gorodetskaya I, Greene T, James S, Larive B, Lindsay RM, Mehta RL, Dis 61: 247–253, 2013 Miller B, Ornt DB, Rajagopalan S, Rastogi A, Rocco MV, Schiller B, 7. Pierratos A: Daily nocturnal home hemodialysis. Kidney Int 65: 1975– Sergeyeva O, Schulman G, Ting GO, Unruh ML, Star RA, Kliger AS; FHN 1986, 2004 Trial Group: In-center hemodialysis six times per week versus three 8. Johansen KL, Zhang R, Huang Y, Chen SC, Blagg CR, Goldfarb- times per week. NEnglJMed363: 2287–2300, 2010 Rumyantzev AS, Hoy CD, Lockridge RS Jr., Miller BW, Eggers PW, 24. Rocco MV, Lockridge RS Jr., Beck GJ, Eggers PW, Gassman JJ, Greene Kutner NG: Survival and hospitalization among patients using nocturnal T, Larive B, Chan CT, Chertow GM, Copland M, Hoy CD, Lindsay RM,

1910 Journal of the American Society of Nephrology J Am Soc Nephrol 28: 1898–1911, 2017 www.jasn.org CLINICAL RESEARCH

Levin NW, Ornt DB, Pierratos A, Pipkin MF, Rajagopalan S, Stokes JB, 30. Van Eps CL, Jones M, Ng T, Johnson DW, Campbell SB, Isbel NM, Unruh ML, Star RA, Kliger AS, Kliger A, Eggers P, Briggs J, Hostetter T, Mudge DW, Beller E, Hawley CM: The impact of extended-hours home Narva A, Star R, Augustine B, Mohr P, Beck G, Fu Z, Gassman J, Greene hemodialysis and buttonhole cannulation technique on hospitalization T, Daugirdas J, Hunsicker L, Larive B, Li M, Mackrell J, Wiggins K, Sherer rates for septic events related to dialysis access. Hemodial Int 14: 451– S, Weiss B, Rajagopalan S, Sanz J, Dellagrottaglie S, Kariisa M, Tran T, 463, 2010 West J, Unruh M, Keene R, Schlarb J, Chan C, McGrath-Chong M, 31. Muir CA, Kotwal SS, Hawley CM, Polkinghorne K, Gallagher MP, Frome R, Higgins H, Ke S, Mandaci O, Owens C, Snell C, Eknoyan G, Snelling P, Jardine MJ: Buttonhole cannulation and clinical outcomes Appel L, Cheung A, Derse A, Kramer C, Geller N, Grimm R, Henderson in a home hemodialysis cohort and systematic review. Clin J Am Soc L, Prichard S, Roecker E, Rocco M, Miller B, Riley J, Schuessler R, Nephrol 9: 110–119, 2014 Lockridge R, Pipkin M, Peterson C, Hoy C, Fensterer A, Steigerwald D, 32. Chertow GM, Levin NW, Beck GJ, Daugirdas JT, Eggers PW, Kliger AS, Stokes J, Somers D, Hilkin A, Lilli K, Wallace W, Franzwa B, Waterman E, Larive B, Rocco MV, Greene T; Frequent Hemodialysis Network (FHN) Chan C, McGrath-Chong M, Copland M, Levin A, Sioson L, Cabezon E, Trials Group: Effects of randomization to frequent in-center hemodialysis Kwan S, Roger D, Lindsay R, Suri R, Champagne J, Bullas R, Garg A, on long-term mortality: Frequent hemodialysis daily trial [Abstract]. JAm Mazzorato A, Spanner E, Rocco M, Burkart J, Moossavi S, Mauck V, Soc Nephrol 24: 442A, 2013 Kaufman T, Pierratos A, Chan W, Regozo K, Kwok S; Frequent Hemo- 33. Chertow GM, Levin NW, Beck GJ, Daugirdas JT, Eggers PW, Kliger AS, dialysis Network (FHN) Trial Group: The effects of frequent nocturnal Larive B, Rocco MV, Greene T; Frequent Hemodialysis Network (FHN) home hemodialysis: The frequent hemodialysis network nocturnal trial. Trials Group: Long-term effects of frequent in–center hemodialysis. Kidney Int 80: 1080–1091, 2011 JAmSocNephrol27: 1830–1836, 2016 25. Crosby RD, Kolotkin RL, Williams GR: Deﬁning clinically meaningful 34. US National Institutes of Health NCT02019225. Available at https:// change in health-related quality of life. JClinEpidemiol56: 395–407, clinicaltrials.gov/ct2/show/NCT02019225. Accessed November 25, 2003 2016 26. Revicki D, Hays RD, Cella D, Sloan J: Recommended methods for de- 35. EuroQol Group: EuroQol–a new facility for the measurement of health- termining responsiveness and minimally important differences for related quality of life. Health Policy 16: 199–208, 1990 patient-reported outcomes. J Clin Epidemiol 61: 102–109, 2008 36. ACTIVE Dialysis Steering Committee: 2014. ACTIVE Dialysis statistical 27. Leaf DE, Goldfarb DS: Interpretation and review of health-related analysis plan. Available at: www.georgeinstitute.org/projects/active- quality of life data in CKD patients receiving treatment for anemia. dialysis-a-clinical-trial-of-intensive-dialysis. Accessed November 25, Kidney Int 75: 15–24, 2009 2016 28. Samsa G, Edelman D, Rothman ML, Williams GR, Lipscomb J, Matchar 37. National Kidney Foundation: KDOQI clinical practice guideline for D: Determining clinically important differences in health status mea- hemodialysis ddequacy: 2015 update. Am J Kidney Dis 66: 884–930, sures: A general approach with illustration to the Health Utilities Index 2015 Mark II. Pharmacoeconomics 15: 141–155, 1999 29. Chan CT, Greene T, Chertow GM, Kliger AS, Stokes JB, Beck GJ, Daugirdas JT, Kotanko P, Larive B, Levin NW, Mehta RL, Rocco M, Sanz J, Yang PC, Rajagopalan S; Frequent Hemodialysis Network Trial Group: Effects of frequent hemodialysis on ventricular volumes and left This article contains supplemental material online at http://jasn.asnjournals. ventricular remodeling. Clin J Am Soc Nephrol 8: 2106–2116, 2013 org/lookup/suppl/doi:10.1681/ASN.2015111225/-/DCSupplemental.

J Am Soc Nephrol 28: 1898–1911, 2017 xxxx 1911 ACTIVE Primary Results paper Online Supplementary Material

Contents ACTIVE Primary Results paper Online Supplementary Material ...... 1 ACTIVE Dialysis Study Team ...... 2 Methods ...... 4 Imputation ...... 4 Equation for stdKt/v ...... 4

Appendix Tables Appendix Table 1. Blood flow rate (mls/min) and dialysis flow rate (mls/min) by treatment group over study period ...... 5 Appendix Table 2. Details of dialysis treatment characteristics used throughout the study ...... 6 Appendix Table 3. Number of patients with measures of small molecule clearance at baseline and throughout follow up by dialysis location1 ...... 7 Appendix Table 4. Impact of extended dialysis hours on measures of small molecule clearance among participants with baseline measures ...... 7 Appendix Table 5. Baseline characteristics of participants contributing to MRI analyses of the ACTIVE Dialysis Trial ...... 8 Appendix Table 6. Number (%) of participants with missing data for secondary outcomes ...... 10

Appendix Figures

Appendix Figure 1. Distribution of baseline EQ5D scores in ACTIVE participants according to randomisation and location in China or Australia/Canada/New Zealand ...... 11 Appendix Figure 2. Evolution of EQ5D scores through the study with the average effect from all follow-up visits ...... 12 Appendix Figure 3. Impact of extended dialysis hours on change in left ventricular mass index at 12 months according to prespecified subgroup analyses defined by region, dialysis setting and months on dialysis at baseline ...... 13

Page 1

ACTIVE Dialysis Study Team Steering Committee: Hongli Lin(First Affiliated Hospital of Dalian Medical University, Dalian, China), Josette Eris (Concord Repatriation General and Royal Prince Alfred Hospitals, Sydney, Australia), Paul Snelling (Concord Repatriation General and Royal Prince Alfred Hospitals, Sydney, Australia), Ling Zhang (China-Japan Friendship Hospital, Beijing, China), Jinsheng Xu (Fourth Hospital Affiliated to Hebei Medical University, Shijiazhuang, China), Junli Zhang (The Chinese PLA Shanghai 85th Hospital, Shanghai, China), Carmel Hawley (Princess Alexandra Hospital, Brisbane, Australia), Carolyn Van Eps (Princess Alexandra Hospital, Brisbane, Australia), Bruce Cooper (Royal North Shore Hospital, Sydney Australia), Yinhui Li (Fourth Hospital Affiliated to Jilin University -FAW General Hospital, Changchun, China), Michael Suranyi (Liverpool Hospital, Sydney, Australia), Jeff Wong (Liverpool Hospital, Sydney, Australia), Li Zuo (Peking University First Hospital, Beijing, China), Ying Li (Third Hospital Affiliated to Hebei Medical University, Shijiazhuang, China), Nicholas Gray (Nambour General Hospital, Nambour, Australia), Christopher T Chan (Toronto General Hospital, Toronto, Canada), Janak de Zoysa (North Shore Hospital and Auckland City Hospital, Auckland, New Zealand), Mei Wang (Peking University People’s Hospital, Beijing, China), Jiang Lei (Peking University Shenzhen Hospital, Shenzhen, China), Zhangsuo Liu(The First Affiliated Hospital of Zhengzhou University, Zhengzhou, China), Paul Collett (Sydney Adventist Hospital, Sydney, Australia), Lanfen Xue(Shijiazhuang First People’s Hospital, Shijiazhuang, China), Nan Chen (Ruijin Hospital Shanghai Jiaotong University, Shanghai, China), Yonghui Mao(Beijing Hospital, Beijing, China), Rongshan Li(The Second Affiliated Hospital of Shanxi Medical University, Taiyuan, China), Caili Wang (First Affiliated Hospital of Inner Mongolia, Baotou Medical College, Baotou, China), Wen Huang (Beijing Tongren Hospital, Beijing, China), Ping Fu (West China Hospital, Sichuan University, Chengdu, China), Eugenia Pedagogos (Royal Melbourne Hospitals, Melbourne, Australia), Kai Wang(Civil Aviation General Hospital, Beijing, China), Shengrong Zhang(Beijing Hospital of Traditional Chinese Medicine, Beijing, China), Dwarakanathan Ranganathan (Royal Brisbane Hospital, Brisbane, Australia), Jijun Li(First Affiliated Hospital of Chinese PLA General Hospital, Beijing, China), Bicheng Liu(Zhongda Hospital, Southeast University, Nanjing, China), Michael Copland (Vancouver Coastal Health, Vancouver, Canada), Jeffrey Perl (St Michael’s Hospital, Toronto, Canada), Peter Kerr (Monash Medical Centre, Melbourne, Australia), Rebecca Pellicano (Monash Medical Centre, Melbourne, Australia), Girish Talaulikar (Canberra Hospital, Canberra, Australia), Richard Yu (Royal Hobart Hospital, Hobart, Australia), Neil Boudville (Royal Perth Hospital, Perth, Australia), Peter Mount (Austin Health, Melbourne, Australia), Murthy Divi (Gold Coast Hospital), Imad Haloob (Auckland City Hospital), Alan Cass (The George Institute for Global Health, Sydney, Australia), John Agar (Geelong Hospital, Geelong, Australia), Kirsten Howard (University of Sydney, Australia), Ashley Irish (Royal Perth Hospital, Perth, Australia)

Steering Committee Executive: Vlado Perkovic (Steering Committee Chair); Meg Jardine (Study Director), Alan Cass; Chris Chan (Regional Lead Canada); Janak de Zoysa (Regional Lead New Zealand); Martin Gallagher; Nicholas Gray (Regional Lead Australia); Zuo Li (Regional Lead China)

Study Sites: (Auckland City Hospital, New Zealand) Janak de Zoysa, Imad Haloob, Grace Muyoma; (Austin Health, Australia) Pascal Bisscheroux, Peter Mount, Kathy Paizis, Matthew Robert; (Canberra Hospital, Australia) Patricia Johnson, Girish Talaulikar, Giles Walters; (Beijing Hospital of Ministry of Health, China) Yonghui Mao, Ying Sun, Songlan Wang; (Beijing Hospital of Traditional Chinese Medicine, China) Yufang Duan, Yuan Meng, Mingxia Sun, Shengrong Zhang; (Beijing Tongren Hospital Capital Medical University, China) Wen Huang, Wen Zhang; (China-Japan Friendship Hospital, China) Duo Li, Qian Liu, Ling Zhang, Yumei Zhang; (Civil Aviation General Hospital, China) Wang Bei, Kai Wang, Hongying Yang; (Concord Repatriation General Hospital and Royal Prince Alfred Hospitals, Australia) Jenny Burman, Josette Eris, Martin Gallagher, Samantha Hand, Meg Jardine, Shaundeep Sen, Paul Snelling; (Diamond Health Centre Vancouver, Canada) Katy Vela, Michael Copland, Jeffery Gardner, Matthew Paquette, Daniel Rogers, Lina Sioson; (FAW General Hospital (Jilin University Fourth Hospital), China) Gen Li, Yinhui Li, Ke Ma; (First Affiliated Hospital of Chinese PLA General Hospital, China) Fengkun Chen, Jijun Li; (First Affiliated Hospital of Dalian Medical University, China) Dan Liu, Hongli Lin, Yanling Sun, Jing Wang; (First Affiliated Hospital of Inner Mongolia, Baotou Medical College, China) Liping Liu, Lei Nan, Yingjin Shi, Caili Wang; (Gold Coast Hospital, Australia) Murthy Divi, Sandra Lawrence; (Hebei Medical University Fourth Hospital, China) Yaling Bai, Lei He, Jinsheng Xu, Shenglei Zhang; (Hebei Medical University Third Hospital, China) Qingqing Duan, Ying Li, Lihong Zhang; (Liverpool Hospital, Australia) Josephine Chow, Angela Makris, Ananthakrishnapuram Aravindan, Bruce Cleland, Margaret Gilbert, Bruce Hall, Kenneth Howlin, Andrew Jeffreys, Glenda Rayment, Stephen Timothy Spicer, Michael Suranyi, Jeffery Wong, Belinda Yip; (Monash Medical Centre, Australia) Fiona Brown, Peter Kerr, Jyotsna Nandkumar, Rebecca Pellicano, Kevan Polkinghorne, Maryanne Quinn; (Nambour General Hospital, Australia) Victoria Campbell, Carolyn Clark, Nicholas Gray, Peter Hollett, Page 2

Colleen Johnston, Kumar Mahadevan, Euan Noble, Andrea Pollock; (North Shore Hospital, New Zealand) Janak de Zoysa, Grace Muyoma; (Peking University First Hospital, China) Xuyang Cheng, Xiaomei Li, Li Zuo, Yuqing Chen; (Peking University Peoples Hospital, China) Liangying Gan, Mei Wang, Hongjie Yang; (Peking University Shenzhen Hospital, China) Lei Jiang, Jinnan Liao, Yue Zhang, Lei Zheng; (Princess Alexandra Hospital, Australia) Scott Campbell, Amanda Coburn, Carmel Hawley, Jean Helyar, Nicole Isbel, David Johnson, Joanna Sudak, Carolyn Van Eps; (Royal Brisbane and Womens Hospital, Australia) Helen Healy, Julie Owens, Dwarakanathan Ranganathan, Sharadchandra Kanneyalal Ratanjee, Rebecca Russo; (Royal Hobart Hospital, Australia) Melanie Clark, Roslyn Hamilton, Sharonne Hennessy, Matthew Jose, Geoffrey Kirkland, Gail Read, Richard Yu; (Royal Melbourne Hospital, Australia) Kathy Nicholls, Kym Boekel, Gloria Bruno, Peter Hughes, Michael Lian, Eugenia Padagogos, Matija Raspudic, Rowan Walker; (Royal North Shore Hospital, Australia) Bruce Cooper, Cheryl Macadam, Vlado Perkovic, Muh Geot Wong; (Ruijin Hospital Shanghai Jiaotong University School of Medicine, China) Chen Nan, Jiajie Wu, Xiaonong Chen; (Shijiazhuang First People's Hospital, China) Ling Cui, Na Liu, Lanfen Xue, Guoxin Zhang; (Sir Charles Gairdner Hospital, Australia) Neil Boudville, Helen Herson, Susan Pellicano; (St Michael's Hospital, Canada) Sheen Anthony, Stella Curvelo, Niki Dacouris, Strauss Diane, Jenny Hucule, Perl Jeff, Murji KarimaRosa Marticorena, Mauricio Medrano, Ron Wald; (Sydney Adventist Hospital, Australia) Paul Collett, Meg Jardine, Vlado Perkovic, Karen Sutherland, Debbie Talafua; (The Chinese Peoples Liberation Army Shanghai 85th Hospital, China) Hongdi Huan, Chunlai Lu, Junli Zhang, Jingjing Song; (The First Affiliated Hospital of Zhengzhou University, China) Zhangsuo Liu, Xiaoqing Lu, Pei Wang; (The Second Hospital of Shanxi Medical University, China) Qingjiang Hu, Rongshan Li, Xinyan Liu; (Toronto General Hospital, Canada) Christopher Chan, Margaret McGrath- Chong, Katherine Tsong; (West China Hospital of Sichuan University, China) Li Zhou, Ping Fu; (Zhong Da Hospital Southeast University, China) Min Gao, Bicheng Liu, Lichen Xu.

Management Committee: Vlado Perkovic (Steering Committee Chair); Meg Jardine (Study Director); Alan Cass; Martin Gallagher; Helen Monaghan (Senior Project Manager)

Central Co-ordinating Centre (Sydney and Beijing) (Data Management, The George Institute for Global Health, Sydney, Australia) Manuela Armenis, Bijini Bahuleyan, Dominic Byrne, Louise Ham, Leonie Pearson, Suzie Steley; (International Coordinating Centre, The George Institute for Global Health, Sydney, Australia) Rebecca Anderson, Ana Carreras, Alan Cass, Swamy Chintapatla, Eleanora Fjalling, Moya Flancman, Martin Gallagher, Jason Healey, Josephine Hsu, Meg Jardine, Min Jun, Tam Le, Nicola Martin, Mamta Merai, Helen Monaghan, Sue Murray, Sue Murray, Ayana Ono, Vlado Perkovic, Josephine Raley, Tanya Scotcher, Brendan Smyth, Amanda Wang; (Radiology, SMRI, Sydney, Australia; Sydney Medical School, University of Sydney, Australia; Peking University First Hospital, Beijing China) Stuart Grieve, Chi-Jen Hsu, Helen Lackey, Sarah Leighton, Raj Puranik, Yufeng Xu; (Regional Coordinating Centre, George Clinical, Beijing, China) Ying Guo, Ruolan Jia, Tingting Lei, Jicheng Lv, Zhang Qian, Li Shi, Lei Ting Ting, Xiaoting Wei, Xinfang Xie, Damin Xu, Fanli Zhou, Hui Zhou; (Statistics, The George Institute for Global Health, Sydney, Australia) Laurent Billot, Severine Bompoint, Stephane Heritier, Serigne Lo, Jayanthi Mysore, Kha Vo. DSMB: Laurent Billot (unblinded statistician) (The George Institute for Global Health, Sydney, Australia), Gavin Becker (The Royal Melbourne Hospital, Melbourne, Australia), Andrew Forbes (Monash University, Melbourne, Australia), Adeera Levin (University of British Columbia, Vancouver, Canada), Marcello Tonelli (Chair) (University of Alberta, Edmonton, Canada), Zhihong Liu (Nanjing University School of Medicine, Nanjing, China)

Page 3

Methods

Imputation Single imputation was performed for EQ-5D defined quality of life if the measure was missing in accordance with the EQ-5D User Guide (www.euroqol.org). If the patient died before the 12 months follow-up then quality of life was scored as 0. If the patient reached the end of the study duration before 12 months or if the quality of life score at 12 month was missing then the last available measure was used. There was no imputation for LVMI as there was only one post-randomisation measure.

Equation for stdKt/v

Standardised kt/v was calculated using the following equations described in the 2015 Update of the KDOQI Clinical Practice Guideline for Hemodialysis Adequacy1 to estimate true std Kt/v from spKt/v. The fixed volume model was used as measures of residual native kidney clearance were not collected in this study. Residual native clearance is likely to have been negligible for most participants given the median duration on dialysis of 2.5 years at recruitment.

1. Daugirdas JT, Depner TA, Inrig J, Mehrotra R, Rocco MV, Suri RS, Weiner DE, Greer N, Ishani A, MacDonald R, Olson C, Rutks I, Slinin Y, Wilt TJ, Rocco M, Kramer H, Choi MJ, Samaniego-Picota M, Scheel PJ, Willis K, Joseph J, Brereton L. KDOQI Clinical Practice Guideline for Hemodialysis Adequacy: 2015 Update. Am. J. Kidney Dis. 2015;66(5):884-930.

Page 4

Appendix Table 1. Blood flow rate (mls/min) and dialysis flow rate (mls/min) by treatment group over study period

Standard Extended Mean (SD) Median (IQR) Mean (SD) Median (IQR) p-value Blood flow rate (mls/min) Baseline 278.6 (36.8) 280.0 (250.0; 300.0) 279.2 (39.8) 288.0 (250.0; 300.0) Follow up 283.8 (48.4) 280.0 (250.0; 300.0) 261.6 (42.9) 250.0 (230.0; 300.0) 0.0010

Dialysate flow rate (mls/min) Baseline 503.9 (84.5) 500.0 (500.0; 500.0) 514.0 (69.7) 500.0 (500.0; 500.0) Follow up 502.3 (71.7) 500.0 (500.0; 500.0) 478.4 (64.4) 500.0 (500.0; 500.0) 0.0215

P-values are from a Linear mixed model with random intercept including randomisation and time categories with baseline as reference

Page 5

Appendix Table 2. Details of dialysis treatment characteristics used throughout the study

Standard Extended Total Characteristic N=100 N=100 N=200 Membrane Asahi 3.0% 1.0% 2.0% Baxter 6.0% 7.0% 6.5% Baxter then Bellco 1.0% 0.0% 0.5% Bellco 0.0% 1.0% 0.5% D-Tec 1.0% 1.0% 1.0% Fresenius 67.0% 70.0% 68.5% Fresenius then Nipro 1.0% 0.0% 0.5% Gambro 9.0% 9.0% 9.0% Nipro 6.0% 2.0% 4.0% Toray 1.0% 0.0% 0.5% Wego 2.0% 2.0% 2.0% Not reported 3.0% 7.0% 5.0%

Flux High 89.0% 82.0% 85.5% Low 7.0% 10.0% 8.5% High then low 1.0% 0.0% 0.5% Low then high 0.0% 1.0% 0.5% Not reported 3.0% 7.0% 5.0%

Biocompatibility Cellulose 4.0% 3.0% 3.5% Synthetic 93.0% 90.0% 91.5% Not reported 3.0% 7.0% 5.0%

Characteristics of dialysis treatment were mostly constant throughout the study. Proportions of those whose dialysis treatments changed during the study are indicated. Seven participants were treated with cellulose membranes throughout the study: 1 in Australia, 1 in Canada and 5 in China.

Page 6

Appendix Table 3. Number of patients with measures of small molecule clearance at baseline and throughout follow up by dialysis location1

In-center and satellite-based participants Home-based participants Baseline Follow up2 Baseline Follow up N=149 N=145 N=51 N=49 Urea Reduction Ratio (%) 78 (52.3%) 99 (68.3%) 32 (62.7%) 26 (53.1%) Kt/V 47 (31.5%) 66 (45.5%) 28 (54.9%) 25 (51.0%)

1. ACTIVE used values obtained from routine clinical practice. Not all participants had a measure of small molecular clearance measured every 3 months. The table depicts the numbers of patients who had measures of small molecular clearance during the study. 2. Numbers in followup exclude participants who had withdrawn or died before follow-up assessments

Appendix Table 4. Impact of extended dialysis hours on measures of small molecule clearance among participants with baseline measures

p- Standard Extended Between-Group difference value

n mean (sd) n mean (sd) Mean difference (95% CI)

Urea Reduction Ratio (%) Baseline 58 69.67 ( 10.29) 52 72.35 ( 8.52) 2.67 ( -0.92, 6.27) 3-month 55 69.62 ( 9.29) 50 78.06 ( 8.84) 8.44 ( 4.92, 11.96) 6-month 52 69.65 ( 8.63) 45 79.82 ( 6.89) 10.17 ( 6.99, 13.35) 9-month 48 68.40 ( 9.58) 47 77.68 ( 8.22) 9.29 ( 5.65, 12.92) 12-month 49 69.49 ( 8.97) 52 75.85 ( 9.51) 6.36 ( 2.70, 10.01) Average intervention effect1 7.12 ( 4.91, 9.33) <.0001

Single pool Kt/v Baseline 42 1.39 ( 0.35) 33 1.43 ( 0.31) 0.04 ( -0.11, 0.19) 3-month 41 1.47 ( 0.40) 34 1.96 ( 0.74) 0.50 ( 0.23, 0.76) 6-month 39 1.41 ( 0.33) 35 1.95 ( 0.69) 0.54 ( 0.29, 0.78) 9-month 36 1.40 ( 0.30) 34 1.93 ( 0.56) 0.54 ( 0.33, 0.75) 12-month 37 1.40 ( 0.26) 33 1.83 ( 0.61) 0.43 ( 0.21, 0.65) Average intervention effect1 0.48 ( 0.27, 0.68) <.0001

Standardised Kt/V2 Baseline 42 2.53 ( 0.42) 33 2.76 ( 0.85) 0.23 ( -0.07, 0.53) 3-month 41 2.72 ( 0.55) 34 3.81 ( 1.39) 1.10 ( 0.63, 1.57) 6-month 39 2.56 ( 0.46) 35 4.25 ( 3.86) 1.69 ( 0.45, 2.93) 9-month 36 2.83 ( 1.51) 34 3.65 ( 0.92) 0.83 ( 0.23, 1.43) 12-month 37 2.83 ( 1.40) 33 3.62 ( 1.20) 0.79 ( 0.17, 1.42) Average intervention effect1 1.29 ( 0.61, 1.96) 0.0003

1. Average effect of intervention for all followup visits adjusted for baseline. 2. Calculated from measuring spKt/V using a fixed volume model. Residual kidney function data was not collected in the study.

Page 7

Appendix Table 5. Baseline characteristics of participants contributing to MRI analyses of the ACTIVE Dialysis Trial

Standard Extended Total Characteristics (N = 44) (N = 51) (N = 95) Age at randomisation (years) Mean (SD) 50.2 (12.08) 50.0 (13.23) 50.1 (12.64) Median (Q1, Q3) 51.0 (42.0, 58.9) 50.5 (39.6, 62.1) 50.9 (40.2, 59.4)

Gender Male 26/44 (59.1%) 31/51 (60.8%) 57/95 (60.0%) Female 18/44 (40.9%) 20/51 (39.2%) 38/95 (40.0%)

Primary cause of renal disease Diabetic Nephropathy 17/44 (38.6%) 11/51 (21.6%) 28/95 (29.5%) Hypertension/Vascular Nephrosclerosis 4/44 (9.1%) 5/51 (9.8%) 9/95 (9.5%) Glomerulonephritis 16/44 (36.4%) 23/51 (45.1%) 39/95 (41.1%) Reflux Nephrology 1/44 (2.3%) 1/51 (2.0%) 2/95 (2.1%) Polycystic Kidney Disease 1/44 (2.3%) 4/51 (7.8%) 5/95 (5.3%) Other or unknown 5/44 (11.4%) 7/51 (13.8%) 12/95 (12.7%)

Co-morbidity Diabetes Mellitus 20/44 (45.5%) 15/51 (29.4%) 35/95 (36.8%) Hypertension 36/44 (81.8%) 40/51 (78.4%) 76/95 (80.0%) Any Cardiovascular disease 14/44 (31.8%) 12/51 (23.5%) 26/95 (27.4%) Symptomatic Ischaemic Heart Disease 6/44 (13.6%) 5/51 (9.8%) 11/95 (11.6%) Angina 3/44 (6.8%) 2/51 (3.9%) 5/95 (5.3%) Acute Myocardial Infarction 1/44 (2.3%) 1/51 (2.0%) 2/95 (2.1%) Previous Coronary Artery Bypass Graft/ 3/44 (6.8%) 2/51 (3.9%) 5/95 (5.3%) Percutaneous Transluminal Coronary Angioplasty Congestive Heart Failure 7/44 (15.9%) 5/51 (9.8%) 12/95 (12.6%) Cerebrovascular Disease 2/44 (4.5%) 3/51 (5.9%) 5/95 (5.3%) Peripheral Vascular Disease 5/44 (11.4%) 1/51 (2.0%) 6/95 (6.3%)

Smoking status Never smoked 27/44 (61.4%) 32/51 (62.7%) 59/95 (62.1%) Past cigarette smoker 9/44 (20.5%) 9/51 (17.6%) 18/95 (18.9%) Current cigarette smoker 8/44 (18.2%) 10/51 (19.6%) 18/95 (18.9%)

Country Australia 11/44 (25.0%) 18/51 (35.3%) 29/95 (30.5%) Canada 2/44 (4.5%) 3/51 (5.9%) 5/95 (5.3%) China 29/44 (65.9%) 30/51 (58.8%) 59/95 (62.1%) New Zealand 2/44 (4.5%) 0/51 (0.0%) 2/95 (2.1%)

Ethnicity Caucasian 7/44 (15.9%) 12/51 (23.5%) 19/95 (20.0%) Aboriginal or Torres Strait Islander 0/44 (0.0%) 1/51 (2.0%) 1/95 (1.1%) Maori 1/44 (2.3%) 2/51 (3.9%) 3/95 (3.2%) Pacific Islander 3/44 (6.8%) 2/51 (3.9%) 5/95 (5.3%) Asian 31/44 (70.5%) 32/51 (62.7%) 63/95 (66.3%)

Page 8

Standard Extended Total Characteristics (N = 44) (N = 51) (N = 95) Indian 0/44 (0.0%) 0/51 (0.0%) 0/95 (0.0%) Other 2/44 (4.5%) 1/51 (2.0%) 3/95 (3.2%) Not reported 0/44 (0.0%) 1/51 (2.0%) 1/95 (1.1%)

Number of dialysis sessions per week 2 1/44 (2.3%) 0/51 (0.0%) 1/95 (1.1%) 3 39/44 (88.6%) 44/51 (86.3%) 83/95 (87.4%) 4 or more 4/44 (9.1%) 7/51 (13.8%) 11/95 (11.7%)

Total number of hours on dialysis per week Mean (SD) 13.4 (2.58) 13.6 (2.50) 13.5 (2.53) Median (Q1, Q3) 12.0 (12.0, 15.0) 12.0 (12.0, 15.0) 12.0 (12.0, 15.0)

Duration on dialysis at enrolment, median (IQR) in years 2.20 (1.04, 5.19) 2.43 (0.54, 5.35) 2.24 (0.71, 5.35)

Dialysis site at enrolment At home 12/44 (27.3%) 13/51 (25.5%) 25/95 (26.3%) Institution 32/44 (72.7%) 38/51 (74.5%) 70/95 (73.7%)

Intended Dialysis site for study treatment Home 6/44 (13.6%) 3/51 (5.9%) 9/95 (9.5%) Institution (Satellite Centre/Hospital) 38/44 (86.4%) 48/51 (94.1%) 86/95 (90.5%)

Dialysis access Native arteriovenous fistula 35/44 (79.5%) 45/51 (88.2%) 80/95 (84.2%) Synthetic fistula 1/44 (2.3%) 1/51 (2.0%) 2/95 (2.1%) Tunnelled dialysis catheter 7/44 (15.9%) 4/51 (7.8%) 11/95 (11.6%) Non-tunnelled dialysis catheter 1/44 (2.3%) 1/51 (2.0%) 2/95 (2.1%)

Dialysis cannulation method The Buttonhole technique 8/44 (18.2%) 10/51 (19.6%) 18/95 (18.9%) The Rope Ladder technique 28/44 (63.6%) 36/51 (70.6%) 64/95 (67.4%) Dialysis Catheter 8/44 (18.2%) 5/51 (9.8%) 13/95 (13.7%)

Pre-dialysis blood pressure Systolic Blood Pressure, mean (SD) in mmHg 143.5 (20.04) 141.9 (15.94) 142.6 (17.87) Diastolic Blood Pressure, mean (SD) in mmHg 81.3 (13.29) 81.3 (11.85) 81.3 (12.47)

Body Mass Index, median (IQR) in kg/m2 24.3 (23.0, 28.9) 24.1 (21.6, 26.6) 24.2 (22.4, 27.7)

Left ventricular mass index2, mean (SD) in g/m2 103.64 (30.68) 106.48 (36.39) 105.16 (33.72)

Page 9

Appendix Table 6. Number (%) of participants with missing data for secondary outcomes

Parameter No baseline measure No follow-up measure

n=200 n=1941 Systolic Blood Pressure (mmHg) 0 (0.0%) 3 (1.5%) Diastolic Blood Pressure 0 (0.0%) 3 (1.5%) (mmHg) BP lowering agents (number) 0 (0.0%) 2 (1.0%) Potassium (meq/L) 3 (1.5%) 2 (1.0%) Phosphate (mg/dL) 3 (1.5%) 3 (1.5%) PTH (pg/mL) 25 (12.5%) 6 (3.1%) Phosphate binders (number of 0 (0.0%) 2 (1.0%) tablets) Calcium (mg/dl) 3 (1.5%) 3 (1.5%) Hemoglobin (g/dl) 2 (1.0%) 2 (1.0%) ESA dose (EPO units) 2 (1.0%) 2 (1.0%) Ferritin (ng/ml) 46 (23.0%) 27 (13.9%) Transferrin saturation (%) 65 (32.5%) 45 (23.2%) Weight (kg) 1 (0.5%) 3 (1.5%) Waist:hip ratio (cm:cm) 0 (0.0%) 4 (2.1%) Albumin (g/dL) 22 (11.0%) 15 (7.7%)

1. Numbers eligible for follow-up assessments excludes participants who had withdrawn or died before follow-up assessments

Page 10

Appendix Figure 1. Distribution of baseline EQ5D scores in ACTIVE participants according to randomisation and location in China or Australia/Canada/New Zealand

China China Standard Extended 40

10 t n e

c 0 r Non-China Non-China e

P Standard Extended 30

0 0.00 0.25 0.50 0.75 1.00 0.00 0.25 0.50 0.75 1.00 EQ5D at baseline

Page 11

Appendix Figure 2. Evolution of EQ5D scores through the study with the average effect from all follow-up visits Mean EQ5D score (primary) by visit (mean and std) 1 .1

1 .0

0 .9

0 .8

0 .7

0 .6

0 .5

0 .4 Average effect 0.03 (-0.03, 0.09), p=0.30 0 .3 B a s e lin e 3 - m o n th 6 - m o n th 9 - m o n th 1 2 - m o n th VISIT

Last run: 02OCT2014 10:14

Page 12

Appendix Figure 3. Impact of extended dialysis hours on change in left ventricular mass index at 12 months according to prespecified subgroup analyses defined by region, dialysis setting and months on dialysis at baseline

Page 13