Home versus in‐centre haemodialysis for people with kidney failure

Abstract

Background

Home haemodialysis (HHD) may be associated with important clinical, social or economic benefits. However, few randomised controlled trials (RCTs) have evaluated HHD versus in‐centre HD (ICHD). The relative benefits and harms of these two HD modalities are uncertain. This is an update of a review first published in 2014. This update includes non‐randomised studies of interventions (NRSIs).

Objectives

To evaluate the benefits and harms of HHD versus ICHD in adults with kidney failure.

Search methods

We contacted the Information Specialist and searched the Cochrane Kidney and Transplant Register of Studies up to 9 October 2022 using search terms relevant to this review. Studies in the Register are identified through searches of CENTRAL, MEDLINE, and EMBASE, conference proceedings, the International Clinical Trials Registry Platform (ICTRP) Search Portal, and ClinicalTrials.gov. We searched MEDLINE (OVID) and EMBASE (OVID) for NRSIs.

Selection criteria

RCTs and NRSIs evaluating HHD (including community houses and self‐care) compared to ICHD in adults with kidney failure were eligible. The outcomes of interest were cardiovascular death, all‐cause death, non‐fatal myocardial infarction, non‐fatal stroke, all‐cause hospitalisation, vascular access interventions, central venous catheter insertion/exchange, vascular access infection, parathyroidectomy, wait‐listing for a kidney transplant, receipt of a kidney transplant, quality of life (QoL), symptoms related to dialysis therapy, fatigue, recovery time, cost‐effectiveness, blood pressure, and left ventricular mass.

Data collection and analysis

Two authors independently assessed if the studies were eligible and then extracted data. The risk of bias was assessed, and relevant outcomes were extracted. Summary estimates of effect were obtained using a random‐effects model, and results were expressed as risk ratios (RR) and their 95% confidence intervals (CI) for dichotomous outcomes and mean difference (MD) or standardised mean difference (SMD) and 95% CI for continuous outcomes. Confidence in the evidence was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach.

Meta‐analysis was performed on outcomes where there was sufficient data.

Main results

From the 1305 records identified, a single cross‐over RCT and 39 NRSIs proved eligible for inclusion. These studies were of varying design (prospective cohort, retrospective cohort, cross‐sectional) and involved a widely variable number of participants (small single‐centre studies to international registry analyses). Studies also varied in the treatment prescription and delivery (e.g. treatment duration, frequency, dialysis machine parameters) and participant characteristics (e.g. time on dialysis). Studies often did not describe these parameters in detail. Although the risk of bias, as assessed by the Newcastle‐Ottawa Scale, was generally low for most studies, within the constraints of observational study design, studies were at risk of selection bias and residual confounding.

Many study outcomes were reported in ways that did not allow direct comparison or meta‐analysis. It is uncertain whether HHD, compared to ICHD, may be associated with a decrease in cardiovascular death (RR 0.92, 95% CI 0.80 to 1.07; 2 NRSIs, 30,900 participants; very low certainty evidence) or all‐cause death (RR 0.80, 95% CI 0.67 to 0.95; 9 NRSIs, 58,984 patients; very low certainty evidence). It is also uncertain whether HHD may be associated with a decrease in hospitalisation rate (MD ‐0.50 admissions per patient‐year, 95% CI ‐0.98 to ‐0.02; 2 NRSIs, 834 participants; very low certainty evidence), compared with ICHD.

Compared with ICHD, it is uncertain whether HHD may be associated with receipt of kidney transplantation (RR 1.28, 95% CI 1.01 to 1.63; 6 NRSIs, 10,910 participants; very low certainty evidence) and a shorter recovery time post‐dialysis (MD ‐2.0 hours, 95% CI ‐2.73 to ‐1.28; 2 NRSIs, 348 participants; very low certainty evidence). It remains uncertain if HHD may be associated with decreased systolic blood pressure (SBP) (MD ‐11.71 mm Hg, 95% CI ‐21.11 to ‐2.46; 4 NRSIs, 491 participants; very low certainty evidence) and decreased left ventricular mass index (LVMI) (MD ‐17.74 g/m², 95% CI ‐29.60 to ‐5.89; 2 NRSIs, 130 participants; low certainty evidence). There was insufficient data to evaluate the relative association of HHD and ICHD with fatigue or vascular access outcomes.

Patient‐reported outcome measures were reported using 18 different measures across 11 studies (QoL: 6 measures; mental health: 3 measures; symptoms: 1 measure; impact and view of health: 6 measures; functional ability: 2 measures). Few studies reported the same measures, which limited the ability to perform meta‐analysis or compare outcomes.

It is uncertain whether HHD is more cost‐effective than ICHD, both in the first (SMD ‐1.25, 95% CI ‐2.13 to ‐0.37; 4 NRSIs, 13,809 participants; very low certainty evidence) and second year of dialysis (SMD ‐1.47, 95% CI ‐2.72 to ‐0.21; 4 NRSIs, 13,809 participants; very low certainty evidence).

Authors' conclusions

Based on low to very low certainty evidence, HHD, compared with ICHD, has uncertain associations or may be associated with decreased cardiovascular and all‐cause death, hospitalisation rate, slower post‐dialysis recovery time, and decreased SBP and LVMI. HHD has uncertain cost‐effectiveness compared with ICHD in the first and second years of treatment.

The majority of studies included in this review were observational and subject to potential selection bias and confounding, especially as patients treated with HHD tended to be younger with fewer comorbidities. Variation from study to study in the choice of outcomes and the way in which they were reported limited the ability to perform meta‐analyses. Future research should align outcome measures and metrics with other research in the field in order to allow comparison between studies, establish outcome effects with greater certainty, and avoid research waste.

Plain language summary

Is home haemodialysis better than in‐centre haemodialysis for people with kidney failure?

Key messages

‐ Home haemodialysis may be preferred by some patients. However, the research gives us very uncertain answers.

‐ We are unsure whether the better patient outcomes with home haemodialysis are because of the dialysis treatment itself or because patients receiving home haemodialysis are younger and less sick.

Why perform haemodialysis at home rather than in a dialysis centre?

Kidney failure is a common and increasingly prevalent public health problem, which results in increases in illness, death and healthcare costs. People with kidney failure require kidney replacement therapy (dialysis and kidney transplantation) to remove the accumulation of waste products in the blood, which in turn may assist with reducing symptoms such as fatigue, nausea and itching and may improve a person's overall quality of life. Unfortunately, some patients lack access to hospital dialysis care.

What did we want to find out?

People who are treated with haemodialysis at home may experience increased well‐being and might live longer. However, home haemodialysis may also increase the burden of healthcare for patients and families and increase technical problems for patients.

What did we do?

We searched for randomised and non‐randomised studies comparing home haemodialysis with hemodialysis treatment performed in a hospital or clinical setting. We compared and summarised the trials' results and rated our confidence in the information based on factors such as trial methods and size.

What did we find?

We found only one randomised study (where patients are randomly allocated to one treatment or the other) that compared home haemodialysis with in‐centre haemodialysis in nine patients. All other studies (39) were observational (where the treatment was not randomly assigned).

Home haemodialysis may be associated with outcomes including increased length of life, fewer hospital stays, higher chance of receiving a kidney transplant, shorter recovery time from dialysis itself and increased control of blood pressure. Patients receiving home haemodialysis tended to have more dialysis (more hours or more often). Some of the differences in outcomes for patients may have been due to factors that were not related to dialysis treatment since patients receiving home haemodialysis were younger and had fewer other illnesses.

What are the limitations of the evidence?

The small number and size of the studies were limitations in this review. Not all the studies provided data about the outcomes we were interested in, and we are unsure about the results.

How up‐to‐date is the evidence?

The evidence is up to date as of October 2022.

Summary of findings

Summary of findings 1. Summary of findings.

Outcomes	Relative effect (95% CI)	No. of participants (studies)	Certainty of evidence (GRADE)	Comments
Cardiovascular death	RR 0.92 (0.80 to 1.07)	30,900 (2)	Very low1 ⊕⊝⊝⊝	‐
All‐cause death	RR 0.80 (0.67 to 0.95)	58,984 (9)	Very low2 ⊕⊝⊝⊝	Studies varied from small single centre studies and large registry analyses. Follow‐up duration varied from less than two years to more than 10 years
Hospitalisation
All‐cause annual hospitalisation rate (admissions/patient‐year)	MD ‐0.50 admissions (‐0.98 to ‐0.02)	834 (2)	Very low3 ⊕⊝⊝⊝	‐
All‐cause annual hospitalisation days (days/patient‐year)	MD ‐1.90 days (‐2.28 to ‐1.53)	834 (2)	Low4 ⊕⊕⊝⊝	‐
Received kidney transplantation	RR 1.28 (1.01 to 1.63)	10,910 (6)	Very low5 ⊕⊝⊝⊝	Studies varied from small single centre studies and large registry analyses
Health‐related quality of life
Physical Component Summary	SMD 0.42 (0.10 to 0.73)	922 (5)	Very low6 ⊕⊝⊝⊝	All studies were cross‐sectional design. The time on dialysis varied substantially between studies (Table 2)
Mental Component Summary	SMD 0.10 (‐0.05 to 0.25)	922 (5)	Very low7 ⊕⊝⊝⊝	All studies were cross‐sectional design. The time on dialysis varied substantially between studies (Table 2)
Recovery time (hours)	MD ‐2.00 (‐2.73 to ‐1.28)	348 (2)	Low8 ⊕⊕⊝⊝	Included one longitudinal cohort study of patients transitioning dialysis modality, and reported mean of weekly responses during 8‐week period on each modality, and one cross‐sectional survey of prevalent dialysis patients (time current on modality not reported)
Blood pressure7
Systolic BP (mm Hg)	MD ‐11.78 (‐21.11 to ‐2.46)	491 (4)	Very low9 ⊕⊝⊝⊝	Included both longitudinal cohort and cross‐sectional studies. Exclusion of cross‐sectional studies led to similar relative effect with reduced heterogeneity (MD ‐19.47, 95% CI ‐24.40 to ‐14.54; I2 = 42%) Data from one RCT also indicated lower BP with HHD versus ICHD (155 ± 18 vs 169 ± 24 mm Hg)
Diastolic BP (mm Hg)	MD 1.81 (‐1.31 to 4.94)	383 (3)	Very low10 ⊕⊝⊝⊝	Included both longitudinal cohort and cross‐sectional studies Data from one RCT indicated lower BP with HHD versus ICHD (89 ± 6 vs 93 ± 9 mm Hg)

HHD: home haemodialysis; ICHD: in‐centre haemodialysis; 95% CI: 95% confidence interval; RR: risk ratio; MD: mean difference; SMD: standardised mean difference; BP: blood pressure; NRSI: non‐randomised study of intervention; RCT: randomised controlled trial

1. Downgraded due to imprecision of relative effect (CI includes the possibility of no effect or harm)
2. Downgraded due to considerable heterogeneity (I² = 84%)
3. Downgraded due to considerable heterogeneity (I² = 90%)
4. Downgraded based on 2 non‐randomised observational studies
5. Downgraded due to substantial heterogeneity (I² = 78%)
6. Downgraded due to substantial heterogeneity (I² = 74%) and risk of bias
7. Downgraded due to imprecision of relative effect (CI includes the possibility of no effect or harm) and risk of bias
8. Relative effect based on data from NRSIs, as only one RCT was identified during the systematic review
9. Downgraded due to considerable heterogeneity (I² = 81%)
10. Downgraded due to imprecision of relative effect, CI includes the possibility of no effect or harm

1. Vintage of patients included in Physical Component Summary and Mental Component Summary analysis.

	Metric	HHD	ICHD
Jayanti 2016 ICHD (197) HHD (91)	Median dialysis vintage (years)	3.47 (IQR 1.39, 6.82)	2.68 (IQR 1.05, 5.12)
Toronto Group 2002 ICHD (163) HHD (56)		Not reported	Not reported
Wong 2019a ICHD (253) HHD (41)	Mean duration of ESKD diagnosis (years)	7.5 ± 4.9	Not reported
Wright 2015 ICHD (29) HHD (22)	Time on dialysis (number of patients)
	6 to 12 months	7	3
	1 to 5 years	8	17
	5 to 10 years	6	6
	10 to 20 years	0	2
	> 20 years	1	1
Watanabe 2014 ICHD (34) HHD (46)	Time on dialysis (years)	7.4 ± 8.3	6.4 ± 5.7

ESKD: end‐stage kidney disease; HHD: home haemodialysis; ICHD: in‐centre haemodialysis; IQR: interquartile range

Background

Description of the condition

Kidney failure is a common and increasingly prevalent public health problem, which results in excess morbidity, death and healthcare costs (Bello 2017a; Bello 2019). People with kidney failure require kidney replacement therapy (KRT; dialysis and kidney transplantation) to address the physiological accumulation of fluid and metabolites, which in turn may assist with reducing uraemic symptoms (such as fatigue, anorexia, nausea and pruritus), improving health‐related quality of life (HRQoL) and prolonging survival (Cabrera 2017). Unfortunately, some patients lack access to dialysis care, and although kidney transplantation is associated with increased survival and quality of life (QoL) compared with dialysis (Laupacis 1996; Tonelli 2011; Wolfe 1999), only 22% of all patients with treated kidney failure around the world receive kidney transplantation (Bello 2019). The remaining patients are treated with either haemodialysis (HD) or peritoneal dialysis (PD) (Bello 2017a; Bello 2019). The median survival of patients with kidney failure on dialysis is considerably shorter than for patients with common types of cancer (e.g. breast, colorectal, prostate) (Naylor 2019). Moreover, kidney failure results in substantial financial costs to the health system, accounting for 2% to 3% of healthcare spending in higher‐resource countries, despite patients with kidney failure comprising 0.1% to 0.2% of the population (Bello 2017b).

Ascertaining the optimal means of delivering dialysis in terms of patient‐reported, clinical and health‐economic outcomes is important information for patients, people who support their care, clinicians and healthcare policymakers.

Description of the intervention

HD is the most commonly used dialysis modality, comprising 89% of all dialysis and 69% of KRT globally (Pecoits‐Filho 2020). HD can be performed either in a centre (e.g. hospital or satellite dialysis units) or the patient’s own home. The prevalence of home HD (HHD) use varies widely worldwide, with 14% of prevalent dialysis patients doing HHD in Aotearoa, New Zealand, in 2019 (ANZDATA 2021). On the other hand, in the USA, only 1.9% of prevalent dialysis patients were doing HHD in 2019 (USRDS 2021). This is despite the fact that home‐based dialysis therapies have significant cost benefits compared to in‐centre HD (ICHD), with a Canadian study estimating the total annual cost of ICHD to be $73,920 after the first year, compared to $45,203 for HHD (Klarenbach 2014).

ICHD is usually performed by a trained nurse, who sets up the equipment, inserts the needles, monitors the patient, and adjusts treatment parameters as needed throughout the treatment. ICHD prescriptions typically involve four to five hours of dialysis three times/week (ANZDATA 2021).

Patients wishing to do HHD usually undertake one to four months of training, which covers machine set‐up and basic maintenance, needling of their fistula or graft, or accessing their dialysis vascular catheter, troubleshooting and managing alarms on dialysis (Kidney Health Australia 2013; USRDS 2021). In addition, home assessment and modifications may be required to confirm patients have a suitable space to perform dialysis and store supplies, as well as an adequate power and water supply. As most patients performing HD in their own homes do so independently or with the assistance of a family member or support person, this allows considerably greater flexibility in treatment duration and frequency compared with those undergoing ICHD, who must fit into more rigid schedules (Kidney Health Australia 2013). Patients receiving HHD may also utilise extended hours regimens, such as nocturnal (6 to 10 hours overnight), extended hours (6 to 10 hours/session), or short daily dialysis (< 4 hours/session, performed on a daily basis).

How the intervention might work

Epidemiological studies have indicated that HHD treatment may be associated with improved patient outcomes compared to ICHD. However, there is likely selection bias since patients receiving HHD tend to be younger and have fewer comorbidities (Mailloux 1996; Woods 1996). HHD has also been associated with increased patient autonomy and QoL (Cases 2011; Walsh 2005). Since HHD enables increased dialysis hours and frequency compared to ICHD, there are a number of potential treatment‐related benefits compared to conventional HD, including increased small solute clearance (typically measured as Kt/V_urea), improved control of serum phosphate and blood pressure (BP) (FHN Trial Group 2010), reduced myocardial stunning (Jefferies 2011), and shorter interdialytic intervals which may mitigate the heightened risk of death associated with long (three‐day) interdialytic intervals (Foley 2011; Krishnasamy 2013). Augmented dialysis duration and/or frequency can also assist with reducing ultrafiltration rate requirements, which in turn have been associated with reduced death (Assimon 2016). Thus, it is possible that HHD improves survival compared to ICHD, perhaps through a reduction in cardiovascular events.

On the other hand, HHD can result in an increased burden on patients, families and support people (Gilbertson 2019; Iyasere 2016; Morton 2010; Suri 2011). In addition, due to reduced clinical oversight of dialysis technique, patients performing HHD may be at increased risk of complications, such as infection and thrombosis (FHN Trial Group 2010; Suri 2013). These potential risks need to be weighed against the potential benefits of HHD.

Why it is important to do this review

HHD may increase survival and QoL and is less costly to healthcare systems than ICHD (Klarenbach 2014; Walker 2014). However, these potential benefits need to be considered against the potential disadvantages of HHD, which include increased burden and risk of complications. Evaluation of research comparing HHD and ICHD is limited, and a previous systematic review identified a single randomised controlled trial (RCT) (Palmer 2014). This updated version of that systematic review includes randomised and non‐randomised studies of interventions (NRSIs), using the available evidence to inform shared decision‐making by clinicians and patients regarding HD modality choice and its effects on patient‐reported, clinical and surrogate outcomes and adverse events.

Objectives

To evaluate the benefits and harms of HHD versus ICHD in adults with kidney failure.

Methods

Criteria for considering studies for this review

Types of studies

All RCTs and quasi‐RCTs (RCTs in which allocation to treatment was obtained by alternation, use of alternate medical records, date of birth or other predictable methods) and NRSIs (observational studies with prospective and retrospective identification of participants, including registry studies) comparing HHD with ICHD in patients with kidney failure were eligible. In order for data to be from studies providing dialysis in a manner comparable to current practice, studies published from the year 2000 onwards were eligible (Marshall 2015).

Studies were required to include:

• Comparison of HHD and ICHD between two or more groups of participants or within the same group of participants over time.

• Groups of individuals formed by randomisation, quasi‐randomisation, time differences, location differences, or health professionals' or participants' preferences.

• Identification of participants, assessment before intervention, actions/choices leading to an individual becoming a member of a group and assessment of outcomes carried out before or after the study was designed (prospective or retrospective design).

• An assessment of the comparability between groups of potential confounders or not (e.g. non‐adjusted analyses).

Types of participants

Inclusion criteria

• Adults (≥ 18 years) with kidney failure receiving ICHD or HHD

• Incident and prevalent HD patients

• HD patients previously treated with PD

• Patients previously treated with kidney transplantation.

Exclusion criteria

• The review did not include data obtained from children or patients with acute kidney injury, as these patients are rarely treated with HHD due to anticipated recovery of kidney function.

Types of interventions

HD provided using any dialysis machine, dialysate, blood or dialysate flow rate, membrane type, dialysis dose (urea clearance), or vascular access type (central venous catheter (CVC), arteriovenous fistula (AVF) or arteriovenous graft (AVG)) was included. We included studies with any duration of dialysis and any frequency in either treatment arm.

• HHD was defined as any type of HD, haemodiafiltration, or haemofiltration carried out by the patient or caregiver at home.

• HHD included HD performed independently by patients (without the assistance of nursing or technical staff) in a community home or self‐care unit.

• ICHD included HD provided in a hospital unit, a private dialysis unit, or a satellite dialysis unit in which nursing or technical staff provided dialysis care. Patients provided with HD by nursing or technical staff in their own homes were considered ICHD.

• Studies evaluating PD as a home dialysis modality were excluded, except where data regarding participants receiving ICHD and HHD were disaggregated.

Types of outcome measures

Primary outcomes

Cardiovascular death: fatal myocardial infarction (MI), fatal stroke, sudden death, heart failure.

Secondary outcomes

Clinical outcomes

• All‐cause death

• Non‐fatal MI

• Non‐fatal stroke

• All‐cause hospitalisation: number of patients with one or more hospitalisation events

• Kidney transplantation

• Vascular access events

  • AVF/AVG intervention: surgical revision, thrombolysis/thrombectomy, fistulogram/fistuloplasty

  • CVC insertion/exchange

  • Infection

    • Local infection: exit‐site infection, cellulitis, abscess/collection

    • Systemic infection: bacteraemia

• Parathyroidectomy

Patient‐reported outcomes

• QoL: we considered and tabulated, where necessary, all reports of QoL outcomes using any instrument. Meta‐analyses were conducted when sufficient studies reported QoL outcomes using a single instrument, including measures of depression and household financial stress.

• End‐of‐treatment employment status: employed, unemployed, not eligible for employment

• Symptoms related to dialysis therapy: intradialytic cramping, hypotension, nausea, vomiting, headache

• Fatigue

• Recovery time

Health economics outcomes

• Cost‐effectiveness

Surrogate outcomes

• BP (systolic BP (SBP), diastolic BP (DBP), mean arterial pressure (MAP), pulse pressure measured in pre‐dialysis or other setting) (mm Hg)

• Left ventricular mass (LVM): described using any diagnostic tool, including magnetic resonance imaging or echocardiography (g; g/m²)

Search methods for identification of studies

Electronic searches

For this update, we searched the Cochrane Kidney and Transplant Register of Studies up to 7 October 2022 through contact with the Information Specialist using search terms relevant to this review. The Register contains studies identified from the following sources.

1. Monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL)

2. Weekly searches of MEDLINE OVID SP

3. Searches of kidney and transplant journals and the proceedings and abstracts from major kidney and transplant conferences

4. Searching of the current year of EMBASE OVID SP

5. Weekly current awareness alerts for selected kidney and transplant journals

6. Searches of the International Clinical Trials Registry Platform (ICTRP) Search Portal and ClinicalTrials.gov.

Studies contained in the Register are identified through searches of CENTRAL, MEDLINE, and EMBASE based on the scope of Cochrane Kidney and Transplant. Details of search strategies, as well as a list of hand‐searched journals, conference proceedings and current awareness alerts, are available on the Cochrane Kidney and Transplant website under CKT Register of Studies.

For non‐randomised studies, we searched MEDLINE (OVID) 1 January 2000 to 20 July 2020 and EMBASE (OVID) 1 January 2000 to 20 July 2020.

See Appendix 1 for search terms used in strategies for this and the previous review.

Searching other resources

1. Reference lists of review articles, relevant studies and clinical practice guidelines.

2. Contacting relevant individuals or organisations seeking information about unpublished or incomplete studies.

Data collection and analysis

Selection of studies

The search strategy described was used to obtain titles and abstracts of studies that may be relevant to the review. The search was performed unrestricted by language; non‐English articles were translated before assessment. Titles and abstracts were screened independently by two authors, who discarded studies that were not applicable; however, studies and reviews that might have included relevant data or information on studies were retained initially. Two authors independently assessed retrieved abstracts and, if necessary, the full text of these studies to determine which studies satisfied the inclusion criteria.

Data extraction and management

Data extraction was carried out independently by two authors using standard data extraction forms. Any further information required from the original author(s) was requested by written correspondence, and any relevant information obtained in this manner was included in the review. Studies reported in non‐English language journals were translated before assessment. Where more than one publication from one study existed, reports were grouped together, and the publication with the most complete data was used in the analyses (Higgins 2022). Where multiple publications reported the same outcome in the same or overlapping populations, or where we could not be certain that this did not occur, the publication with the most complete data was used in the analyses. Where relevant outcomes were only published in earlier versions, these data were used. Any discrepancies between published versions were highlighted. Disagreements were resolved by consultation with the authors.

Assessment of risk of bias in included studies

Randomised controlled trials

The following items were independently assessed by two authors using the Risk of Bias assessment tool version 2 (RoB 2) (Higgins 2022). The domains included in the assessment were as follows.

1. Bias arising from the randomisation process

2. Bias due to deviations from intended interventions

3. Bias due to missing outcome data

4. Bias in the measurement of the outcome

5. Bias in the selection of the reported result.

Non‐randomised studies of interventions

The Newcastle‐Ottawa Scale (NOS) (www.ohri.ca/programs/clinical_epidemiology/nosgen.pdf) for assessing the quality of NRSIs was used. The NOS was used to adjudicate the risk of bias in non‐randomised studies. For cohort studies, the NOS used a star scoring system based on the selection of study groups (four items), comparability between the study group and the control group (two items) and the ascertainment of the exposure or outcome of interest (three items), for a total maximum score of nine stars (Appendix 2). An adapted NOS was used for studies using a cross‐sectional design (Herzog 2013). Selection (maximum five stars), comparability of cohorts based on the design or analysis (maximum two stars) and outcome (maximum two stars) were evaluated for a total maximum score of 10 stars (Appendix 3).

• For case‐control studies, the following items were evaluated.

  • Selection: adequacy of definition, representativeness of the cases, selection of controls, definition of controls

  • Comparability: comparability of cases and controls based on the design or analysis

  • Exposure: ascertainment of exposure, same method of ascertainment for cases and controls, non‐response rate

• For cohort studies, the following items were evaluated.

  • Selection: representativeness of the exposed cohort, selection of the non‐exposed cohort, ascertainment of exposure, demonstration that outcome of interest was not present at the start of the study

  • Comparability: comparability of cohorts based on the design or analysis

  • Outcome: assessment of outcome, adequacy of follow‐up and duration of follow‐up

• For cross‐sectional studies, an adapted NOS was used (Herzog 2013). The following items were evaluated.

  • Selection: representativeness of the sample, sample size, comparability of non‐respondents, ascertainment of the exposure

  • Comparability: comparability of cohorts based on the design or analysis

  • Outcome: assessment of outcome, suitability of statistical testing.

Measures of treatment effect

For dichotomous outcomes (e.g. death, cardiovascular events, hospitalisation, vascular access adverse events), results were expressed as a risk ratio (RR) with 95% confidence intervals (CI). Where continuous scales of measurement were used to assess the effects of treatment (e.g. QoL scale, BP, doses of medication, haemoglobin, biochemical variables), the mean difference (MD) was used, or the standardised mean difference (SMD) if different scales had been used. For effect measures reported as means and standard errors (SE), we obtained the standard deviation (SD) using the calculation provided in the Cochrane Handbook (Higgins 2022). For effect measures reported as medians, ranges and/or interquartile ranges (IQR), we estimated means and SD using the approach described by Wan and colleagues (Wan 2014). Outcomes from RCTs and NRSIs were reported separately.

Meta‐analysis of change scores

Where data on both change‐from‐baseline and final value scores existed, we planned to combine data (e.g. LVM, BP) in a meta‐analysis using the (unstandardised) MD method (Higgins 2022). End‐of‐treatment values and change‐from‐baseline scores were placed in subgroups for clarity and summarised using random effects meta‐analysis.

Imputing standard deviation

When none of the above methods allowed calculation of the SD, we imputed change‐from‐baseline SD using an imputed correlation coefficient when sufficient data were available (Abrams 2005; Follmann 1992). If possible, we conducted sensitivity analyses to evaluate the effect of imputing missing SD data in our meta‐analysis.

Unit of analysis issues

We included only data from the first period of treatment in cross‐over studies (Higgins 2022). Data in different metrics were analysed by converting reported values to International System (SI) units. The final results were presented in SI units with conventional units in parentheses.

Dealing with missing data

If possible, data for each prespecified outcome were evaluated regardless of whether the analysis was based on intention‐to‐treat (ITT) or completeness of follow‐up. In particular, dropout rates were investigated and reported in detail (e.g. dropout due to discontinuation of dialysis modality, treatment failure, death, transplantation, withdrawal of consent or loss to follow‐up). Any further information required from the original author was requested by written correspondence (e.g. emailing the corresponding author). Any relevant information obtained in this manner was included in the review. We assessed all studies for risks of bias due to incomplete reporting of results. Evaluation of important numerical data, such as screened, randomised patients, as well as ITT, as‐treated and per‐protocol population was carefully performed. Attrition rates were investigated. Issues of missing data and imputation methods (e.g. last‐observation‐carried‐forward) were critically appraised.

Assessment of heterogeneity

We first assessed the heterogeneity by visual inspection of the forest plot. We then quantified statistical heterogeneity using the I² statistic, which describes the percentage of total variation across studies due to heterogeneity rather than sampling error (Higgins 2022). The following is a guide to the interpretation of I² values.

• 0% to 40%: might not be important

• 30% to 60%: may represent moderate heterogeneity

• 50% to 90%: may represent substantial heterogeneity

• 75% to 100%: considerable heterogeneity.

The importance of the observed value of I² depended on the magnitude and direction of treatment effects and the strength of evidence for heterogeneity (e.g. P value from the Chi² test or a CI for I²) (Higgins 2022).

Assessment of reporting biases

When there were at least 10 studies included in the meta‐analyses (Higgins 2022), we planned to test for asymmetries in the inverted funnel plots (i.e. for systematic differences in the effect sizes between more precise and less precise studies) using the original and modified Egger tests (Egger 1997) and the Begg and Mazumdar correlation test (Begg 1994). There are many potential explanations for why an inverted funnel plot may be asymmetric, including chance, heterogeneity, publication and reporting bias (Terrin 2005). We planned to refrain from judging funnel plot asymmetries based on visual inspection, as this has been shown to be misleading in empirical research (Lau 2006). Since our meta‐analyses did not include at least 10 studies for any of the outcomes evaluated, funnel plot asymmetries were not assessed.

Publication bias was also evaluated by testing the robustness of the results according to publications, namely publication as a full manuscript in a peer‐reviewed journal versus studies published as abstracts, letters or editorials.

Data synthesis

Data were summarised using the random‐effects model, but the fixed‐effect model was also used to ensure the robustness of the chosen model and susceptibility to outliers. We qualitatively summarised data where insufficient data were available for meta‐analysis. A qualitative review was conducted for adverse events and QoL outcomes in studies where a validated tool or metric was not used.

Subgroup analysis and investigation of heterogeneity

Subgroup analyses were planned to explore possible sources of heterogeneity (e.g. participants, interventions, and study quality). Heterogeneity among participants could be related to age and HD methods. Heterogeneity in treatments could be related to prior agent(s) used and the agent, dose, and duration of therapy.

Heterogeneity was planned to be investigated by analysing the data using subgroups according to the following parameters.

• Age (< 60 years versus ≥ 60 years)

• Presence of diabetes

• Presence of cardiovascular disease

• Study design (RCT versus NRSI)

• Methodological quality.

However, subgroup analyses were not done due to the small number of studies and insufficient data available.

Adverse effects were tabulated and assessed with descriptive techniques. Where possible, the risk difference (RD) with 95% CI was calculated for each adverse effect, either compared to no treatment or another agent.

Sensitivity analysis

Sensitivity analyses were done to explore the robustness of findings to key decisions in the review process. These were determined as the review process took place (Higgins 2022). Sensitivity analyses were undertaken to explore the influence of a study's risk of bias on the results.

• Repeating the analysis, excluding unpublished studies

• Repeating the analysis, taking account of the risk of bias, as specified above

• Repeating the analysis, excluding any very long or large studies to establish how much they dominate the results

• Repeat the analysis, excluding studies, using the following filters: diagnostic criteria, language of publication, source of funding (industry versus other), and country.

Summary of findings and assessment of the certainty of the evidence

We presented the main results of the review in Summary of findings tables. These tables present key information concerning the certainty of the evidence, the magnitude of the effects of the interventions examined, and the sum of the available data for the main outcomes (Schünemann 2022a). The Summary of findings tables also include an overall grading of the evidence related to each main outcome using the Grades of Recommendation, Assessment, Development and Evaluation (GRADE) approach (GRADE 2011). The GRADE approach defines the quality of a body of evidence as the extent to which one can be confident that an estimate of effect or association is close to the true quantity of specific interest. The quality of a body of evidence involves consideration of within‐trial risk of bias (methodological quality), directness of evidence, heterogeneity, precision of effect estimates and risk of publication bias (Schünemann 2022b). We initially intended to present vascular access complications and fatigue in the summary of findings. However, there was insufficient data to meta‐analyse these outcomes. We therefore presented the following outcomes in the 'Summary of findings' tables.

• Cardiovascular death

• All‐cause death

• All‐cause hospitalisation

• Kidney transplantation

• HRQoL

• Recovery time

• BP.

Results

Description of studies

The following section contains broad descriptions of the studies considered in this review. For further details on each individual study, please see Characteristics of included studies; Characteristics of excluded studies; Characteristics of studies awaiting classification.

Results of the search

The search of databases and registers was conducted on 7 October 2022 and identified 1280 records; an additional 34 records were identified through other sources. After duplicate records were removed, 1305 studies were screened, and 144 records were selected for full‐text review. Of these, 39 new studies were included (81 records), 34 new studies were excluded (48 records), and 14 studies (14 records) are awaiting classification (see Characteristics of studies awaiting classification). We also identified one new report of an existing excluded study.

We reassessed and deleted 12 previously excluded studies (wrong study design, not HHD versus ICHD, no outcomes of interest).

A total of 40 studies were included (82 reports; 1 RCT, 39 NRSIs), 35 studies were excluded (60 reports), and 14 studies are awaiting classification. The search results are summarised in (Figure 1).

1.

2024 flow diagram for study selection

Included studies

One RCT, already included in the previous version of this systematic review, was included (McGregor 2001). The search did not find any additional RCTs. The search yielded 81 reports of 39 eligible NRSIs. Three of these studies were published as conference abstracts, with no related full article published (Bragg‐Gresham 2018; Dumaine 2018; Ha 2018). Included studies were published between 2001 and 2022, with half of the studies being published from 2015 onwards.

Various study designs were seen across the included studies. Many were small single‐centre prospective studies, cross‐sectional studies and large retrospective registry studies. Details on the design, setting and size of studies are shown in Table 3.

2. Design, setting and size of included studies.

Study ID	Study setting	Study size
Ageborg 2005	Prospective cohort study Cross‐sectional Stockholm (Sweden)	HHD (5) Self‐care HD (6) Conventional ICHD (8)
Bragg‐Gresham 2018	Retrospective cohort study USRDS – data available in the Medical Evidence Report 2006‐2015	Incident dialysis patients (156,377)
Dumaine 2018	Prospective, mixed‐methods, pilot study Comparing transitions (from CKD to initiating dialysis or transition from one dialysis modality to another)	CKD to ICHD (5) CKD to PD (9) CKD to HHD (2) PD to ICHD (2) PD to HHD (4) ICHD to PD (4) ICHD to HHD (7)
Griva 2010	Cross‐sectional 2 dialysis units (affiliated with Royal Free and University College Hospitals), London (UK)	HHD (25) ICHD (52) CAPD (45) APD (23)
Ha 2018	Single‐centre, prospective cohort study Cross‐sectional Sydney (Australia) From 2015‐2017	HHD (27) PD (48) ICHD (113) Transplant recipients (85)
Hayhurst 2015	Cross‐sectional Single‐centre, Royal Preston Hospital (UK) May‐June 2014	CKD stage 3‐5 (17) ICHD (28) HHD (17) PD (17) Transplant recipients (21) Controls: matched by age and sex (50)
Jayanti 2016	Prospective study (BASIC‐HHD study); combined cross‐sectional and prospective study design Multi‐centre across 5 tertiary centres (UK)	HHD (91) Prevalent ICHD (197)
Kasza 2016	Observational cohort study ANZDATA Registry (Australia & New Zealand) 1 Oct 2003‐31 Dec 2011	ICHD CVC (7414) ICHD AVF (5729) HHD AVF (357) PD (6665) HHD CVC: not included in analysis (26)
Kojima 2012	Observational cohort study, retrospectively collected data Single‐centre (Kidney Disease Center, Saitama Medical University, Japan)	Patients transitioned from ICHD to HHD (54)
Krahn 2019	Population‐based retrospective cohort study à linked registry (CORR) and administrative data Ontario (Canada) 1 April 2006 to 31 March 2014	HHD (112) ICHD (9687) short‐daily/slow nocturnal ICHD (65) PD (2827)
Kraus 2007	Prospective, multicenter, open‐label, feasibility study 6 USA centres	Patients transitioning from ICHD to HHD with NxStage System One (32)
Krishnasamy 2013	Observational cohort study ANZDATA Registry (Australia & New Zealand) 1999‐2008	ICHD (9765) PD (4298) HHD (573)
Lee 2002	Prospective cohort study South Alberta Renal Program (Calgary, Canada) 1999‐2000	HHD (9) ICHD (88) Satellite ICHD (31) PD (38)
Lorenzen 2012	Retrospective, longitudinal, single‐centre study Kuratorium für Dialyse und Nierentransplantation (Hannover, Germany)	Patients on maintenance dialysis transferring from ICHD to short daily HHD (11)
Malmstrom 2008	Cross‐sectional (15 Oct 2004) and retrospective (year 2004) Single‐centre (Helsinki University Hospital), Helsinki (Finland)	HHD (33) Satellite ICHD (32)
McGregor 2001	Randomised crossover trial Single‐centre, New Zealand	HHD patients assigned to receive short in‐centre HD and long HHD in a randomised sequence (9)
Murashima 2010	Retrospective cohort study Single‐centre	Patients converted from CHD to HHD (12)
Nebel 2002	Retrospective cohort study Single‐centre, Cologne‐Merheim Hospital (Germany) 1990‐1999	HHD (37) Satellite ICHD (66) PD (69) Transplant recipients (72)
Nesrallah 2012	Multinational renal databases: International Quotidian Dialysis Registry (IQDR) [intensive] + DOPPS [conventional] Secondary IQDR data from REIN registry (France), Fresenius Medical Care North America, PROMIS database (British Columbia, Canada) France, USA, Canada Between Jan 2000 and Aug 2010	Intensive HD (420) Conventional HD (5646) Matched patients by country, duration of ESKD before study enrolment and propensity score Intensive HD (338) Conventional HD (1388)
Nitsch 2011	UK Renal registry England and Wales 1 Jan 1997‐31 Dec 2005	Incident HHD (225) Incident PD: matched by age and sex (900) Incident hospital ICHD: matched by age and sex (900) Incident satellite ICHD: matched by age and sex (450)
Piccoli 2004	Prospective cohort study Single‐centre SMOM Unit (satellite of a large university centre) Turin (Italy) Nov 1998‐Nov 2002	HHD (at home or in training) (42) Limited care ICHD (35)
Rydell 2016	Retrospective, observational case‐control study HHD: Lund University Hospital from 1 Jan 1983 to 31 Dec 2002 ICHD: Malmo General Hospital from 1 Jan 1978 to 31 Dec 2007 (Sweden) Data on dialysis from Swedish Renal Registry; survival data from Swedish Census	Matched according to sex, age, comorbidity and date of start HHD (41 from 118) IHD (41 from 377) Followed until death or 1 Jan 2013; median follow up duration 14 years (HHD) and 11 years (IHD)
Sands 2009	Retrospective study Fresenius Medical Services facilities (USA) 1 Nov 2006‐12 March 2007	Patients who transitioned from ICHD to HHD (29)
Saner 2005	Nested case‐cohort study; retrospective chart analysis – For each patient trained for HHD at the dialysis centre between 1970 and 1995 corresponding match searched from ICHD by retrospective chart analysis Single‐centre (University Hospital of Berne), Bern (Switzerland) From 1970 to 1995	HHD (58) ICHD: matched for sex, age, time of dialysis onset and renal disease category (58)
Suri 2015	Observational retrospective cohort study HHD patients from a large US dialysis provider’s administrative database – propensity score matched to contemporaneous USRDS patients Jan 2004‐Dec 2009	All adults who began DHD between 2004‐2009 HHD (1187) ICHD patients from USRDS (3173)
Tennankore 2022	Canadian Organ Replacement Registry (CORR) analysis 2005‐2014
Van Oosten 2018	National data (Netherlands) Health insurance claims data from 2012‐2014 Data validated with external database (Dutch Renal Registry – Renine)	HHD (197) ICHD (6463) CAPD (463) APD (477) Mix (281) Transplant recipients: living donor (1554) Transplant recipients: deceased donor (1275)
Watanabe 2014	Prospective cohort study Cross‐sectional Single‐centre (Saitama Medical University Hospital), Saitama (Japan) 2011	HHD (46) ICHD: matched for age, sex, cause of ESKD (34)
Wong 2019a	Cross‐sectional data of ESKD patients pooled from the following studies: Multi‐centre study in 2014‐2015 à KDQOL‐36 questionnaire in 356 Chinese adults receiving HD or PD* HHD: 41 patients at 3 public hospitals between May 2016 and Oct 2016 Hong Kong *Chen JY, Choi EPH, Wan EYF et al. Validation of the disease‐specific components of the kidney disease quality of Life‐36 (KDQOL‐36) in Chinese patients undergoing maintenance dialysis. PLoS One 2016;11: e0155188.	NHHD (41) PD (103) HICHD (135) CICHD (118)
Wong 2019b	Multi‐centre 3 public hospitals in Hong Kong Information from medical records + face‐to‐face interview with patients between May 2016 and Oct 2016	HHD (43) ICHD (170) PD (189)
Wright 2015	Cross‐sectional Pilot study 4 outpatient dialysis facilities located in Pennsylvania (USA)	HHD (22) ICHD (29) PD (26)
Xue 2015	Retrospective cohort study HHD: Virginia Lynchburg Dialysis Facility (USA) from 1997 to 2010 ICHD: Fresenius Medical Care North America facilities in Virginia (USA) from 1 Jan 2007 to 31 Dec 2010	HHD (63) ICHD: matched by age, gender, race, dialysis vintage, diabetes (121)
Yeung 2021	Prospective cohort study HHD: Single‐centre, Monash Medical Centre, Melbourne (Australia) ICHD: ANZDATA Registry Jan 2000‐June 2017	HHD (181) ICHD: matched by age, gender and cause of ESKD (413)
Zimbudzi 2014	Retrospective cohort study Single‐centre, Monash Medical Centre, Melbourne (Australia) Aug 2012‐Aug 2013	HHD (25) ICHD: satellite HD (25)

Note: refer to Table 4 for characteristics studies containing grouped reports.

AVF: arteriovenous fistula; AVG: arteriovenous graft; CKD: chronic kidney disease; CVC: central venous catheter; ESKD: end‐stage kidney disease; HD: haemodialysis; HHD: home haemodialysis; ICHD: in‐centre haemodialysis; PD: peritoneal dialysis

In some cases, there were multiple publications generated from the same trial or study, which we have grouped together (Kjellstrand 2008; Rydell 2019; Tablo IDE 2020). In other cases, there were multiple publications which were generated from different studies but were performed within the same or overlapping populations. Marshall 2021 conducted multiple registry studies based on the Australia and New Zealand Dialysis and Transplant (ANZDATA) Registry and reporting on death. Populations included in these studies were overlapping over the period from 1996 to 2017. NxStage‐USRDS 2012 compared the HHD patients registry of NxStage System One users (NxStage Medical Inc) to matched patients from the United States Renal Data System (USRDS) over the 2006 to 2012 period. Finally, there were instances of multiple studies being performed within the same clinical dialysis program over the same or overlapping time periods. Many studies comparing HHD to ICHD were conducted at the University Health Network (UHN) in Toronto, Canada, from 1993 to 2009. As we could not be certain that the populations included in these studies were not overlapping, all studies involving the UHN HHD unit were grouped as Toronto Group 2002. In the aforementioned situations of study grouping, data were extracted only from publications which reported outcomes relevant to this review. Where more than one publication from a study group reported a relevant outcome, the publication with the most complete data was used for analyses. Details on the design, setting and size of reports within study groups are shown in Table 4.

3. Characteristics of grouped reports within studies.

Study/report	Study design and setting	Population	Study period	Outcomes
Kjellstrand 2008
Kjellstrand 2008*	Observational retrospective cohort study 5 centres from USA and Europe	HHD or self‐care HD (265) ICHD (150)	1982 to Jun 2005	Survival (%)
Kjellstrand 2010	Observational retrospective cohort study 5 centres from USA and Europe	HHD (189) ICHD (73)	1982 to Jun 2005	Survival (%)
Marshall 2021
Marshall 2011	Observational retrospective cohort study ANZDATA Registry	HHD (3190) ICHD (21,968)	31 Mar 1996 to 31 Dec 2007	Death (HR for death)
Marshall 2013	Observational retrospective cohort study ANZDATA Registry (NZ only)	HHD (1532) ICHD (5647)	31 Mar 2000 to 31 Dec 2010	Death (HR for death)
Marshall 2014	Observational retrospective cohort study ANZDATA Registry (NZ only)	HHD (1547) ICHD (8713)	1 Jan 1997 to 31 Dec 2011	Death (HR for death)
Marshall 2016	Observational retrospective cohort study ANZDATA Registry	HHD (5764) ICHD (34,952)	31 Mar 1996 to 31 Dec 2012	Death (HR for death)
Marshall 2021*	Observational retrospective cohort study ANZDATA Registry	Complete cohort: 52097 Modality comparison cohort: HHD (1236), ICHD (29,548)	Complete cohort: 1998 to 2017 Modality comparison cohort: 2013 to 2017	Death (HR for death) Proportion cardiovascular death as cause of death
NxStage‐USRDS 2012
Kansal 2019	Observational retrospective cohort study NxStage Medical user registry linked to USRDS (USA)	HHD: NxStage (521) ICHD: USRDS (32,931)	2006 to 2012	Survival (%) Death (HR for death)
Weinhandl 2012*	Observational retrospective cohort study NxStage Medical user registry linked to USRDS (USA)	HHD: NxStage (1873) ICHD: USRDS (9365)	1 Jan 2005 to 31 Dec 2007	All‐cause death (HR for death) Cardiovascular death (HR for death) Infection death (HR for death) Interval‐specific death (HR for death)
Weinhandl 2015a*	Observational retrospective cohort study NxStage Medical user registry linked to USRDS (USA)	HHD: NxStage (3480) ICHD: USRDS (17,400)	1 Jan 2006 to 31 Dec 2009	All‐cause hospital admissions (RR) Cardiovascular hospital admissions (RR) Vascular hospital admissions (RR) All‐cause hospital duration (RR) Cardiovascular hospital duration (RR) Vascular hospital duration (RR)
Weinhandl 2015b	Observational retrospective cohort study NxStage Medical user registry linked to USRDS (USA)	HHD: NxStage (834) ICHD: USRDS (4170)	1 Jan 2007 to 30 Jun 2010	Death (HR for death) Cardiovascular death (HR for death) Infection death (HR for death)
Weinhandl 2015c	Observational retrospective cohort study NxStage Medical user registry linked to USRDS (USA)	HHD: NxStage (3560) ICHD: USRDS (17,800)	1 Jan 2007 to 30 Jun 2010	30‐day readmission after discharge for heart failure 30‐day readmission after discharge for hypertension
Weinhandl 2015d*	Observational retrospective cohort study NxStage Medical user registry linked to USRDS (USA)	HHD: NxStage (1368) ICHD: USRDS (6840)	1 Jan 2007 to 30 Jun 2010	Relative incidence of transplant
Rydell 2019
Rydell 2019a*	Observational retrospective cohort study Swedish Renal Registry, Swedish Inpatient Registry and Swedish Mortality Database	HHD (152) ICHD (608)	1991 to 2012	All‐cause death Median survival (years)
Rydell 2019b*	Observational retrospective cohort study Swedish Renal Registry, Swedish Inpatient Registry and Swedish Mortality Database	HHD (152) ICHD (608)	1991 to 2012	All‐cause annual hospital admission rate All‐cause hospitalization days per patient‐year
Tablo IDE 2020
Chertow 2020*	Prospective cross‐over study Multicentre (USA)	Cross‐over design (30)	Not specified	Recovery time (hours) EQ‐5D‐5L Sleep duration (hours)
Aragon 2020	Awaiting classification: contact made with author
Chahal 2020a	Awaiting classification: contact made with author
Chahal 2020b	Awaiting classification: contact made with author
Plumb 2020	Awaiting classification: contact made with author
Plumb 2021	Awaiting classification: contact made with author
Toronto Group 2002
Bergman 2008*	Controlled cohort study Toronto (Canada)	HHD (32) ICHD (42)	1993 to 2003	Dialysis or cardiovascular‐related admission (per patient year) Duration of dialysis or cardiovascular‐related admission (days per year) All‐cause hospitalisation (per patient year) Duration of all‐cause hospitalisation (days per year) Emergency department visits (per patient year)
Bugeja 2004	Prospective observational cohort study University Health Network, Toronto (Canada)	Cross‐over design (11)	Not specified	Systolic BP Diastolic BP
Cafazzo 2009	Cross‐sectional survey of prevalent patients Toronto (Canada)	HHD (56) ICHD (153)	Not specified	Appraisal of Self‐Care Agency (ASA scale) SF‐12 (Mental Component Summary, Physical Component Summary) Multidimensional Scale of Perceived Social Support Anxiety State (Spielberger) Anxiety Trait (Spielberger)
Chan 2002	Prospective observational cohort study University Health Network, Toronto (Canada)	Cross‐over design (6)	Since Oct 1997; no end date specified	Mean systolic BP Mean diastolic BP LVMI (g/m2)
Chan 2003	Prospective observational cohort study University Health Network, Toronto (Canada)	Cross‐over design (18)	Not specified	Systolic BP Diastolic BP 24‐hour systolic BP 24‐hour diastolic BP
Chan 2005*	Cohort study Cross‐sectional University Health Network, Toronto (Canada)	HHD (10) ICHD (12)	Not specified	Systolic BP Diastolic BP Mean BP LVMI (g/m2)
Chan 2005a	Cohort study University Health Network, Toronto (Canada)	Cross‐over design (10)	Not specified	Systolic BP Diastolic BP

*outcome data utilised in analyses

BP: blood pressure; HHD: home haemodialysis; HR: hazard ratio; ICHD: in‐centre haemodialysis; LVMI: left ventricular mass index; RR: risk ratio

Intervention groups

Most compared groups were parallel, while a few studies used a non‐randomised, single‐cross‐over design in which each participant served as their own control in situations where patients were transitioning from one modality to another (Kojima 2012; Lorenzen 2012; McGregor 2001; Murashima 2010; Sands 2009; Tablo IDE 2020). While all studies compared some form of HHD to ICHD, some studies specified additional subgroups based on treatment duration/timing (e.g. intensive, conventional, nocturnal, daily), setting (e.g. satellite centre, community house), staff involvement (e.g. self‐care, limited care) or vascular access (CVC versus AVF/AVG). Many studies did not specify the treatment duration or did not define “intensive” or “conventional” duration. Additionally, some studies included other comparison groups (e.g. people treated with PD, recipients of a kidney transplant, people with earlier stages of chronic kidney disease (CKD), and healthy controls). Details on intervention groups across studies are provided in Table 5.

4. Detailed intervention comparisons by study.

Study ID/report	HHD	ICHD	Other comparisons
Ageborg 2005	HHD (NS)	Self‐care ICHD ICHD: conventional	‐
Bragg‐Gresham 2018	HHD (NS)	ICHD (NS) Self‐care ICHD (NS)	CAPD CCPD
Dumaine 2018	Transition from ICHD to HHD	‐	Transition from CKD to ICHD Transition from CKD to PD Transition from CKD to HHD Transition from PD to IHD Transition from PD to HHD Transition from ICHD to PD
Griva 2010	HHD: conventional; 3 times/week)	ICHD (NS)	CAPD APD
Ha 2018	HHD (NS)	ICHD (NS)	PD Transplant recipient
Hayhurst 2015	HHD (NS)	ICHD (NS)	PD CKD stages 3‐5 Transplant recipients Controls (healthy)
Jayanti 2016	Intensive HHD: 30.8% conventional HHD; variable dialysis prescriptions	ICHD: conventional; 12 hours/week; 54% HDF	‐
Kasza 2016	HHD: AVF/AVG	ICHD: CVC ICHD: AVF/AVG	PD
Kjellstrand 2008	Intensive HHD: short daily	Intensive ICHD: short daily	Survival probabilities derived from 2005 USRDS incident HD population
Kojima 2012	Intensive HHD	ICHD: conventional	‐
Krahn 2019	HHD: short daily (2 to 3 hours; 6 to 7 times/week (awake)) or slow nocturnal (6 to 9 hours; 5 to 7 times/week) Canadian Organ Replacement Register	ICHD: conventional Intensive ICHD: short daily (2 to 3 hours; 6 to 7 times/week (awake)) or slow nocturnal (6 to 9 hours; 5 to 7 time/week Canadian Organ Replacement Register	PD (CAPD or APD)
Kraus 2007	Intensive HHD: 2 to 3 hours; 6 times/week (NxStage System One)	Intensive ICHD: 2 to 3 hours; 6 times/week (NxStage System One)	‐
Krishnasamy 2013	HHD: mix of intensive and conventional	ICHD: 97% conventional	PD
Lee 2002	HHD/self‐care: ≥ 4 hours/session; ≥ 3 times/week	ICHD: ≥ 4 hours/session; ≥ 3 times/week Satellite HD: ≥ 4 hours/session; ≥ 3 times/week	PD (CAPD and APD)
Lorenzen 2012	Intensive HHD: short daily	ICHD: conventional	‐
Malmstrom 2008	HHD: flexible schedule and length of dialysis	Self‐care satellite HD: 3 times/week	‐
Marshall 2021 / Marshall 2011	HHD: conventional Intensive HHD: frequent/extended; > 3 sessions of ≥ 4 hours or 3 sessions of > 6 hours or 5 sessions of ≥ 3 hours or > 5 sessions of ≥ 2 hours per week	ICHD: conventional Intensive ICHD: frequent/extended; > 3 sessions of ≥ 4 hours or 3 sessions of > 6 hours or 5 sessions of ≥ 3 hours or > 5 sessions of ≥ 2 hours per week	PD
Marshall 2021 / Marshall 2013	HHD (NS)	ICHD (NS)	Community house HD PD
Marshall 2021 / Marshall 2014	HHD (NS)	ICHD (NS)	PD
Marshall 2021 / Marshall 2016	Intensive HHD: any hours/session; ≥ 5 times/week Quasi‐intensive HHD: between conventional and intensive HHD: conventional; ≤ 3 times /week; ≤ 6 hours/session	Intensive ICHD: any hours/session; ≥ 5 times/week Quasi‐intensive ICHD: between conventional and intensive ICHD: conventional; ≤ 3 times /week; ≤ 6 hours/session	PD Deceased donor transplant recipient Living donor transplant recipient
Marshall 2021 / Marshall 2021	HHD (NS)	ICHD (NS)	CAPD APD
McGregor 2001	HHD: 6 to 8 hours/session; 3 times/week	ICHD: 3.5 to 4.5 hours/session; 3 times/week	‐
Murashima 2010	Intensive HHD: short daily (NxStage System One)	ICHD: 3 times/week)	‐
Nebel 2002	HHD (NS)	Satellite HD (NS)	PD (CAPD and APD) Kidney transplant recipient Inpatient acute ICHD (reported but not analysed due to low numbers)
Nesrallah 2012	Intensive HHD: ≥ 5.5 hours/session; 3 to 7 times/week; day or nocturnal International Quotidian Dialysis Registry (none with NxStage)	ICHD: conventional; < 5.5 hours/session; 3 times/week	‐
Nitsch 2011	HHD (NS)	ICHD (NS) Satellite HD (NS)	PD
NxStage‐USRDS 2012 / Kansal 2019	Intensive HHD: previously receiving PD; 92% prescribed 5 to 6 times/week	ICHD: previously receiving PD (USRDS)	‐
NxStage‐USRDS 2012 / Weinhandl 2012	Intensive HHD: daily (NxStage System One)	ICHD: conventional; 3 times/week (USRDS)	‐
NxStage‐USRDS 2012 / Weinhandl 2015a	Intensive HHD: daily; 5 to 6 times/week (NxStage System One)	ICHD: conventional; 3 times/week	‐
NxStage‐USRDS 2012 / Weinhandl 2015b	Intensive HHD: daily (NxStage System One)	ICHD: conventional (USRDS)	‐
NxStage‐USRDS 2012 / Weinhandl 2015c	Intensive HHD: daily (NxStage System One)	ICHD: conventional (USRDS)	PD
NxStage‐USRDS 2012 / Weinhandl 2015d	Intensive HHD: daily (NxStage System One)	ICHD: conventional (USRDS)	PD
Piccoli 2004	HHD: not daily Intensive HHD: daily)	Limited care ICHD: not daily Intensive limited care ICHD: daily	‐
Rydell 2016	HHD (NS)	ICHD (NS)	‐
Rydell 2019 / Rydell 2019a	HHD (NS)	ICHD (NS)	PD
Rydell 2019 / Rydell 2019b	HHD (NS)	ICHD (NS)	PD
Sands 2009	HHD (NS)	ICHD (NS)	‐
Saner 2005	HHD (NS): same prescription at home then in‐centre	ICHD (NS) same prescription at home then in‐centre	‐
Suri 2015	Intensive HHD: daily; 1.5 to 4.5 hours/session; > 5 times/week	ICHD: conventional	PD
Tablo IDE 2020 / Chertow 2020	Intensive HHD: 4 sessions/week	Intensive ICHD: 4 sessions/week	‐
Tennankore 2022	HHD Intensive HHD: frequent	ICHD Intensive ICHD: frequent	PD
Toronto Group 2002 / Bergman 2008	Intensive HHD: nocturnal	ICHD: conventional	‐
Toronto Group 2002 / Bugeja 2004	Intensive HHD: nocturnal; 6 to 8 hours/session; 5 to 6 times/week)	ICHD: conventional; 4 hours/session; 3 t imes/week	‐
Toronto Group 2002 / Cafazzo 2009	Intensive HHD: nocturnal; 6 to 8 hours/session; 4 to 6 times/week	ICHD: conventional; 4 hours/session; 3 times/week	Predialysis included in qualitative interviews
Toronto Group 2002 / Chan 2002	Intensive HHD: nocturnal	ICHD: conventional	‐
Toronto Group 2002 / Chan 2003	Intensive HHD: nocturnal; 8 to 10 hours/session; 6 times/week)	ICHD: conventional; 4 hours/session; 3 times/week	‐
Toronto Group 2002 / Chan 2005	Intensive HHD: nocturnal	ICHD: conventional	‐
Toronto Group 2002 / Chan 2005a	Intensive HHD: nocturnal	ICHD: conventional	‐
Van Oosten 2018	HHD (NS)	ICHD (NS)	CAPD APD Deceased donor transplant recipient Living donor transplant recipient Mix: multiple dialysis modalities in a year
Watanabe 2014	Intensive HHD: 3 to 5 hours/session; 5 to 6 times/week; no nocturnal treatment; large segment had previously undergone PD or PD+HD combined therapy; AVF for all	ICHD: 3 to 5 hours/session; 3 times/week; AVF for all	‐
Wong 2019a	Intensive HHD: nocturnal	ICHD: (NS); hospital‐based Satellite HD (NS); community in‐centre	PD
Wong 2019b	Intensive HHD: nocturnal	ICHD: (NS); hospital‐based	PD
Wright 2015	HHD (NS)	ICHD (NS)	PD
Xue 2015	Intensive HHD: frequent nightly	ICHD: conventional	‐
Yeung 2021	HHD: 6 to 8 hours/session; alternate days	ICHD: 4 to 5 hours/session; 3 times/week	‐
Zimbudzi 2014	Intensive HHD: > 75% 8 hours alternate days	Satellite HD patients on Category 1 transplant waitlist: 5 hours/session; 3 times/week	‐

AVF/AVG: arteriovenous fistula/arteriovenous graft; CCPD: continuous cycling peritoneal dialysis; CVC: central venous catheter; HHD: home haemodialysis; ICHD: in‐centre haemodialysis; NS: (duration and frequency) not specified; PD: peritoneal dialysis; APD: automated PD; CAPD: continuous ambulatory PD

Tablo IDE 2020 specifically evaluated the Tablo HD device at home and in‐centre, while NxStage‐USRDS 2012 (and related reports) evaluated the use of the NxStage dialysis system in HHD compared to ICHD based on matched cohorts from USRDS. In contrast, another study from the USA, where the NxStage System is more broadly used, specified that this device was not used in their study (Nesrallah 2012). However, most studies did not provide details on the type of machine used or other dialysis parameters (e.g. vascular access, blood flow rate).

Specific definitions and prescription details (e.g. duration, frequency) of the HHD and ICHD groups were not always provided. In registry analyses, which included incident patients starting HD, some studies defined groups according to the modality 90 days after dialysis initiation (Marshall 2021) or the current modality at the time of the event of interest (Krishnasamy 2013), whereas others did not describe how groups were defined. Moreover, modality transfers were addressed differently across studies. Some studies used ITT (dialysis modality modelled as fixed) and as‐treated (dialysis modality modelled as time‐varying) frameworks, whereas others censored patients at modality switch.

Reported outcomes

Most studies reported outcomes of interest from only one category (clinical, patient‐reported, health economics or surrogate outcomes) (Table 6). No studies specifically reported on non‐fatal MI, non‐fatal stroke, parathyroidectomy or fatigue. Employment after the commencement of dialysis was reported in Helantera 2012; however, this study included patients from the age of 15 years, and data for adult patients could not be extracted separately. The metrics used to report outcomes were highly variable across studies. For example, vascular access‐related outcomes were reported as vascular access event‐free survival (Piccoli 2004), access complications (Sands 2009), vascular access surgery (Saner 2005) and catheter‐related sepsis (Xue 2015). Similarly, hospitalisation was reported as hospital admission rate (NxStage‐USRDS 2012; Rydell 2019; Suri 2015; Tennankore 2022; Toronto Group 2002), mean number of hospitalisations/patient (Saner 2005), hospitalisation days (NxStage‐USRDS 2012; Rydell 2019; Suri 2015; Toronto Group 2002; Zimbudzi 2014), number of patients experiencing a hospitalisation event (Sands 2009), and number of hospitalisation events in the study (Suri 2015; Tennankore 2022; Zimbudzi 2014). Despite attempts to gain further data from authors, measures of variability (SD, standard errors or IQR) were unobtainable for continuous outcomes in many studies. Likewise, many studies provided the total number of events overall for dichotomous outcomes but did not provide the number of events per treatment arm. In these situations, the study outcomes were unable to be included in the meta‐analyses.

5. Reported outcomes of interest by study.

Study ID	Clinical outcomes	Patient‐reported outcomes	Health economics outcomes	Surrogate outcomes
Ageborg 2005	‐	SF‐36 (physical functioning, role‐physical, bodily pain, general health, vitality, social functioning, role‐emotional, mental health) Appraisal of Self‐Care Agency (ASA scale) Sense of Coherence (SOC) questionnaire	‐	‐
Bragg‐Gresham 2018	‐	Employment prior to dialysis initiation (6 months, %)	‐	‐
Dumaine 2018	‐	KDQOL‐SF (Symptoms and Problems, Effects of Kidney Disease, Burden of Kidney Disease, Physical Component Summary, Mental Component Summary)	‐	‐
Griva 2010	‐	Illness Perceptions Questionnaire (Identity score) Illness Effects Questionnaire Treatment Effects Questionnaire Beck Depression Inventory‐II Cognitive Depression Index	‐	‐
Ha 2018	‐	iPOS‐Renal	‐	‐
Hayhurst 2015	‐	Maximum activity score Total activity score Activity loss score	‐	‐
Jayanti 2016	‐	Recovery time (hours)	‐	Proportion of patients with systolic BP >115 mm Hg Proportion of patients with diastolic BP > 85 mm Hg
Kasza 2016	Unadjusted survival (number at risk) Death (HR for death)	‐	‐	‐
Kjellstrand 2008	Death (HR for death) Survival (HR by modality)		‐	‐
Kojima 2012	‐	‐	‐	Pre‐dialysis systolic BP LVMI
Krahn 2019	Unadjusted survival (number at risk)	‐	30‐day costs (CAD) Cumulative costs (CAD)	‐
Kraus 2007	‐	KDQOL‐SF	‐	Systolic BP Diastolic BP Pulse pressure
Krishnasamy 2013	Day of week as predictor of cardiac death (adjusted OR)	‐	‐	‐
Lee 2002	‐	‐	Annual total health‐care related costs per patient (USD) Annual outpatient dialysis costs (USD) Annual inpatient costs (USD) Total outpatient costs (USD) Annual physician billing costs (USD)	‐
Lorenzen 2012	‐‐		‐	Pre‐dialysis MAP Post‐dialysis MAP
Malmstrom 2008	‐	15D	Annual hospital costs (EUR) Annual total health‐care‐related costs per patient (EUR)	Pre‐dialysis systolic BP Post‐dialysis systolic BP Pre‐dialysis diastolic BP Post‐dialysis diastolic BP
Marshall 2021	Death (HR for death) Proportion cardiovascular death as cause of death	‐	‐	‐
McGregor 2001	‐	Symptoms and quality of life (uraemia‐related symptoms, physical suffering, interference with social activity, burden on families)	‐	Pre‐dialysis BP Post‐dialysis BP Ambulatory BP Symptomatic hypotension LVMI
Murashima 2010	‐	‐	‐	Pulse pressure Incidence of intradialytic hypotension (OR) Clinically significant hypotension during dialysis (OR)
Nebel 2002	‐	‐	Annual healthcare costs (DM)	‐
Nesrallah 2012	Death(HR for death)	‐	‐	‐
Nitsch 2011	1‐year survival (%) Survival after KRT start (HR) Long‐term survival (HR) Waitlisting for kidney transplantation before KRT start (OR) Waitlisting for kidney transplantation after KRT start (HR)	‐	‐	‐
NxStage‐USRDS 2012	All‐cause death (HR for death) Cardiovascular death (HR for death) Infection death (HR for death) Interval‐specific death (HR for death) All‐cause hospital admissions (RR) Cardiovascular hospital admissions (RR) All‐cause hospital duration (RR) Cardiovascular hospital duration (RR) 30‐day readmission after discharge for heart failure 30‐day readmission after discharge for hypertension Relative incidence of transplant Survival (%)	‐	‐	‐
Piccoli 2004	Adverse event‐free survival Vascular access event‐free survival	‐	‐	‐
Rydell 2016	Mean survival (years) Survival (%) Number of patients who received a kidney transplant	‐	‐	Mean systolic BP Mean diastolic BP
Rydell 2019	Median survival (years) All‐cause annual hospital admission rate Hospitalisation (days per year) Time to hospitalisation (years)	‐	‐	‐
Sands 2009	Hospitalisations (per 100 treatments) Arterial site access complications (per 100 treatments) Venous site access complications (per 100 treatments)	‐	‐	‐
Saner 2005	Survival (%) Cardiovascular death during study Vascular access surgery (per patient) All‐cause hospitalisation (per patient) Number of patients who received a kidney transplant during study	‐	‐	‐
Suri 2015	Composite hospitalisation (HR) Composite hospitalisation rate (per patient‐year) Time to first hospitalisation (HR) Cardiovascular hospitalisations (HR) Access‐related hospitalisations (HR) Access infection‐related hospitalisations (HR)	‐	‐	‐
Tablo IDE 2020	Recovery time (hours) EQ‐5D‐5L Sleep duration (hours)	‐	‐	‐
Tennankore 2022	All‐cause hospital admissions (per 1000 patient‐days)	‐	‐	‐
Toronto Group 2002	Dialysis or cardiovascular related admission (per patient year) Duration of dialysis or cardiovascular related admission (days per year) All‐cause hospitalisation (per patient year) Duration of all‐cause hospitalisation (days per year)	Modified Appraisal of Self‐Care Agency (ASA) subscale SF‐12 (Mental Component Summary, Physical Component Summary) Multidimensional Scale of Perceived Social Support Anxiety State (Spielberger State‐Trait Anxiety Inventory for Adults)		Mean systolic BP Mean diastolic BP Mean BP 24‐hour systolic BP 24‐hour diastolic BP LVMI
Van Oosten 2018	‐	‐	Annual RRT costs (EUR) Annual healthcare costs (EUR)	‐
Watanabe 2014	‐	SF‐36 (physical functioning, role‐physical, bodily pain, general health, vitality, social functioning, role‐emotional, mental health; Mental Component Summary, Physical Component Summary) KDQOL (Symptoms and Problems, Effects of Kidney Disease, Burden of Kidney Disease, Work Status, Cognitive Function, Quality of Social Interaction, Sexual Function, Sleep)	‐	‐
Wong 2019b	‐	SF‐12 (physical functioning, role‐physical, bodily pain, general health, vitality, social functioning, role‐emotional, mental health; Mental Component Summary, Physical Component Summary)	‐	‐
Wong 2019a	Annual hospitalisation utilisation (IRR)	‐	First year direct cost (HKD) Second year direct cost (HKD) Yearly indirect cost (HKD) First year societal cost (HKD) Second year societal cost (HKD) First year healthcare provider cost (HKD) Second year healthcare provider cost (HKD)	‐
Wright 2015	‐	KDQOL‐SF (Symptoms and Problems, Effects of Kidney Disease, Burden of Kidney Disease, Work Status, Cognitive Function, Quality of Social Interaction, Sexual Function, Sleep, Social support, Dialysis staff encouragement, Patient satisfaction, Overall health) SF‐12 (Mental Component Summary, Physical Component Summary) SUPPH (positive attitude, stress reduction, decision making)	‐	‐
Xue 2015	Death rate (per 100 patient‐months) Catheter‐related sepsis (per 100 patient‐months) Catheter‐related sepsis (HR)	‐	‐	‐
Yeung 2021	Death (HR for death) Transplantation rate	‐	‐	‐
Zimbudzi 2014	Number of patients hospitalised Hospital admissions (days per patient‐year) Mean length of stay hospitalisation (days)	‐	‐	‐

BP: blood pressure; HR: hazard ratio; IRR: incidence rate ratio; KDQOL‐SF: Kidney Disease Quality of Life Instrument Short Form; KRT: kidney replacement therapy; LVMI: left ventricular mass index; MAP: mean arterial pressure; OR: odds ratio;SF‐12: 12‐Item Short Form Survey; SF‐36: 36‐Item Short Form Survey; Short form; SUPPH: Strategies Used by People to Promote Health

Patient‐reported outcome measures were reported in a wide variety of ways, encompassing measures of QoL (SF‐12, SF‐36, KDQOL‐SF, SF‐6D, EQ‐5D‐3L, 15D), mental health (Beck Depression Inventory‐II (BDI‐II), Spielberger State‐Trait Anxiety Inventory; Cognitive Depression Index), symptoms (IPOS‐Renal), impact and view of health (Sense of Coherence (SOC) questionnaire, Strategies Used by People to Promote Health (SUPPH), Illness Perceptions Questionnaire, Illness Effects Questionnaire, Treatments Effects Questionnaire, Multidimensional Scale of Perceived Support), functional ability (ASA scale, activity score), and recovery time.

Due to variations in how outcomes were defined and reported by studies, pooled analysis was only possible for a few reported outcomes.

• Cardiovascular death (2 studies)

• All‐cause death (9 studies)

• All‐cause annual hospitalisation rate (2 studies)

• All‐cause hospitalisation days per patient‐year (2 studies)

• Kidney transplantation during the study period (6 studies)

• Physical functioning, role‐physical, bodily pain, general health, vitality, social functioning, role‐emotional, and mental health domains of the SF‐12 or SF‐36 (2 studies)

• Physical Component Summary and Mental Component Summary of the SF‐12 or SF‐36 (5 studies)

• Symptoms and problems, effects of kidney disease, burden of kidney disease, work status, cognitive function, quality of social interaction, sexual function and sleep domains of the KDQOL‐SF (2 studies)

• BDI‐II (2 studies)

• Recovery time (2 studies)

• Healthcare costs (4 studies)

• SBP (4 studies)

• DBP (3 studies)

• MAP (2 studies)

• LVM index (2 studies).

Outcomes that were reported in a manner permitting extraction and analysis are demonstrated in Table 7, Table 8, Table 9 and Table 10.

6. Extractable clinical outcomes by study.

Marshall 2021

Nesrallah 2012

Nitsch 2011

NxStage‐USRDS 2012

Piccoli 2004

Rydell 2016

Rydell 2019

Sands 2009

Saner 2005

Suri 2015

Tennankore 2022

Toronto Group 2002

Xue 2015

Yeung 2021

Zimbudzi 2014

Cardiovascular death

*

*

*

All‐cause death

*

*

*

*

*

*

*

*

Median survival (years)

*

Death rate: death/1000 patient‐years

*

All‐cause annual hospital admission rate: number of admissions/patient‐year

‐

*

‐

‐

*

Number of patients with one or more hospitalisation events

*

All‐cause hospitalisation days/patient‐year

‐

*

‐

*

‐

Waitlisted for kidney transplant after KRT start

*

Kidney transplantation during study period

*

*

*

*

*

Vascular access surgery: number/patient

‐

Vascular access complications, number of patients experiencing

*

‐

Vascular access failure rate/100 patient‐months

‐

Catheter‐related sepsis event rate/100 patient‐months

‐

*outcome data extractable and able to be analysed

‐outcome data extractable but unable to be analysed (e.g. Incomplete data reporting, no measure of variability reported or able to be obtained)

7. Extractable patient‐reported outcomes by study.

	Ageborg 2005	Griva 2010	Ha 2018	Hayhurst 2015	Jayanti 2016	Malmstrom 2008	Tablo IDE 2020	Toronto Group 2002	Watanabe 2014	Wong 2019a	Wright 2015
Physical Functioning (SF‐12; SF‐36)									*	*
Role: Physical (SF‐12; SF‐36)									*	*
Bodily Pain (SF‐12; SF‐36)									*	*
General Health (SF‐12; SF‐36)									*	*
Vitality (SF‐12; SF‐36)									*	*
Social Functioning (SF‐12; SF‐36)									*	*
Role: Emotional (SF‐12; SF‐36)									*	*
Mental Health (SF‐12; SF‐36)									*	*
Role‐Social Component Scale (SF‐36)									*
Physical Component Summary (SF‐12; SF‐36)					*			*	*	*	*
Mental Component Summary (SF‐12; SF‐36)					*			*	*	*	*
Symptoms and problems (KDQOL‐SF)									*		*
Effects of kidney disease (KDQOL‐SF)									*		*
Burden of kidney disease (KDQOL‐SF)									*		*
Work status (KDQOL‐SF)									*		*
Cognitive function (KDQOL‐SF)									*		*
Quality of social interaction (KDQOL‐SF)									*		*
Sexual function (KDQOL‐SF)									*		*
Sleep (KDQOL‐SF)									*		*
Social support (KDQOL‐SF)											*
Dialysis staff encouragement (KDQOL‐SF)											*
Patient satisfaction (KDQOL‐SF)											*
Physical functioning (KDQOL‐SF)											*
Role limitations ‐ Physical (KDQOL‐SF)											*
Pain (KDQOL‐SF)											*
General health (KDQOL‐SF)											*
Emotional well‐being (KDQOL‐SF)											*
Role limitations ‐ Emotional (KDQOL‐SF)											*
Social function (KDQOL‐SF)											*
Energy/fatigue (KDQOL‐SF)											*
Overall health (KDQOL‐SF)											*
SF‐6D										*
EQ‐5D‐5L							*
15D						*
BDI‐II		*			*
Anxiety state (Spielberger State‐Trait Anxiety Inventory for Adults)					*			*
Anxiety trait (Spielberger State‐Trait Anxiety Inventory for Adults)					*			*
Cognitive Depression Index		*
IPOS‐Renal			*
SOC questionnaire	*
SUPPH: positive attitude											*
SUPPH: stress reduction											*
SUPPH: decision making											*
Identity score: Illness Perceptions Questionnaire		*
Illness Effects Questionnaire		*
Treatment Effects Questionnaire		*
Multidimensional Scale of Perceived Social Support	*
ASA scale	*
Maximum activity score				*
Total activity score				*
Activity loss score				*
Recovery time (minutes)					*		*

ASA: Appraisal of Self‐Care Agency; BDI: Beck Depression Inventory; EQ‐5D‐5L: Euro‐QOL‐5‐dimension 5‐level; KDQOL: Kidney Disease Quality of Life questionnaire; SF: short form; IPOS: Integrated Palliative Outcome Score; SOC: Sense of Coherence; SUPPH: Strategies Used by People to Promote Health

8. Extractable health economics outcomes by study.

	Krahn 2019	Lee 2002	Malmstrom 2008	Nebel 2002	Van Oosten 2018	Wong 2019b
Annual direct healthcare costs in the first year of dialysis, USD (2021 reference value)	*	*	‐	‐	*	*
Annual direct healthcare costs in first year of dialysis in the second year of dialysis, USD (2021 reference value)	*					*

*outcome data extractable and able to be analysed

‐outcome data extractable but unable to be analysed (eg. Incomplete data reporting, no measure of variability reported or able to be obtained)

9. Extractable surrogate outcomes by study.

	Hayhurst 2015	Kojima 2012	Kraus 2007	Lorenzen 2012	Malmstrom 2008	McGregor 2001	Murashima 2010	Rydell 2016	Toronto Group 2002	Wong 2019a
SBP (mm Hg)		*			*	*		‐	*	*
DBP (mm Hg)					*	*		‐	*	*
MAP (mm Hg)	‐			*					*
MPP (mm Hg)			‐				*
LVMI (g/m2)		*							*

*outcome data extractable and able to be analysed

‐outcome data extractable but unable to be analysed (e.g. Incomplete data reporting, no measure of variability reported or able to be obtained)

DBP: diastolic blood pressure; LVMI: left ventricular mass index; MAP: mean arterial pressure; MPP: mean pulse pressure; SBP: systolic blood pressure

Excluded studies

The majority of records excluded at screening were due to ineligible study design because the study did not compare HHD versus ICHD. Following full‐text report retrieval and assessment for eligibility, most studies were excluded because the study did not compare HHD versus ICHD, as per the review's definition of these interventions.

See Characteristics of excluded studies.

Risk of bias in included studies

Randomised controlled trials

The risk of bias of the one included RCT (McGregor 2001) was assessed using the Risk of Bias assessment tool version 2 (RoB 2) and is shown in Table 11.

10. Risk of bias assessment using Risk of Bias assessment tool 2 (randomised controlled trials).

McGregor 2001
Domain	Signalling question	Response	Comments
Bias arising from the randomization process	1.1 Was the allocation sequence random?	Y	Sequence generation using random number table designed by statistician (data obtained from authors on request).
	1.2 Was the allocation sequence concealed until participants were enrolled and assigned to interventions?	Y	Treatment allocation assigned by statistician unaware of patient details (data obtained from authors on request)
	1.3 Did baseline differences between intervention groups suggest a problem with the randomization process?	N	Cross‐over study, patients acted as their own control
	Risk of bias judgement	Low	‐
Bias due to deviations from intended interventions	2.1.Were participants aware of their assigned intervention during the trial?	Y	Participants and personnel not blinded. Quality of life, BP, and echocardiography outcomes assessed by investigators unaware of treatment sequence (data obtained from authors on request).
	2.2.Were carers and people delivering the interventions aware of participants' assigned intervention during the trial?	Y
	2.3. If Y/PY/NI to 2.1 or 2.2: Were there deviations from the intended intervention that arose because of the experimental context?	N	‐
	2.4 If Y/PY to 2.3: Were these deviations likely to have affected the outcome?	NA	‐
	2.5. If Y/PY/NI to 2.4: Were these deviations from intended intervention balanced between groups?	NA	‐
	2.6 Was an appropriate analysis used to estimate the effect of assignment to intervention?	Y	‐
	2.7 If N/PN/NI to 2.6: Was there potential for a substantial impact (on the result) of the failure to analyse participants in the group to which they were randomized?	NA	‐
	Risk of bias judgement	Low	‐
Bias due to missing outcome data	3.1 Were data for this outcome available for all, or nearly all, participants randomized?	Y	Data available for all participants.
	3.2 If N/PN/NI to 3.1: Is there evidence that result was not biased by missing outcome data?	NA	‐
	3.3 If N/PN to 3.2: Could missingness in the outcome depend on its true value?	NA	‐
	3.4 If Y/PY/NI to 3.3: Is it likely that missingness in the outcome depended on its true value?	NA	‐
	Risk of bias judgement	Low
Bias in measurement of the outcome	4.1 Was the method of measuring the outcome inappropriate?	N	BP measured with electronic BP machine reading. LVMI measured by echocardiography.
	4.2 Could measurement or ascertainment of the outcome have differed between intervention groups?	N	Standardised method and equipment used to measure BP. Standardised method and timing for echocardiogram to be performed.
	4.3 Were outcome assessors aware of the intervention received by study participants?	N	Quality of life, BP, and echocardiography outcomes assessed by investigators unaware of treatment sequence (data obtained from authors on request).
	4.4 If Y/PY/NI to 4.3: Could assessment of the outcome have been influenced by knowledge of intervention received?	NA	‐
	4.5 If Y/PY/NI to 4.4: Is it likely that assessment of the outcome was influenced by knowledge of intervention received?	NA
	Risk of bias judgement	Low	‐
Bias in selection of the reported result	5.1 Were the data that produced this result analysed in accordance with a pre‐specified analysis plan that was finalized before unblinded outcome data were available for analysis?	NI	Sample size calculation and statistical tests reported in publication. However unable to confirm with author if the statistical analysis plan was formalised prior to data being available for analyses.
	5.2 ... multiple eligible outcome measurements (e.g. scales, definitions, time points) within the outcome domain?	N	Multiple time point assessment for BP but mean result reported. Single echocardiogram performed and data used.
	5.3 ... multiple eligible analyses of the data?	N	Single method of analysis.
	Risk of bias judgement	Some concerns	‐
Overall bias	Risk of bias judgement	Some concerns	Unable to blind investigators or participants to intervention, however outcome assessors were blinded. There was also potential bias arising from unmatched interventions (acetate vs bicarbonate buffer).

Allocation

The risk of bias arising from the randomisation process was assessed as low risk, as the treatment allocation sequence was random and concealed until participants were enrolled and assigned to interventions.

Blinding

Due to the nature of the interventions, participants and investigators could not be blinded to treatment assignment. However, outcomes were measured using a standardised method, and assessors were blinded to treatment allocation and sequence; thus, bias in the measurement of the outcome was low.

Incomplete outcome data

There was a low risk of bias due to deviations from intended interventions. There was no loss to follow‐up, so complete outcome data were obtained.

Selective reporting

This study aimed to evaluate outcomes in patients receiving long, slow HD versus standard dialysis, potentially introducing differences in outcomes between treatment arms based on treatment duration rather than location. However, since longer‐hours dialysis might be considered a feature of HHD, this was not considered to be a potential source of bias.

Other potential sources of bias

Although the study reported a sample size calculation and described the statistical tests utilised, there was not enough information to confirm if the statistical analysis plan had been finalised prior to the data being available for analysis. The study reported a mismatch in treatment interventions (acetate versus bicarbonate buffer), which may have reduced comparability between interventions.

Funnel plot analysis was not used to assess for evidence of small study effects, as no analysis included 10 studies or more.

Risk of bias of non‐randomised studies of interventions

Using the NOS for cohort studies, only one study was adjudicated at high risk of bias, with a total of 3/9 stars, as this study was reported as an abstract only and complete results have not been published (Dumaine 2018). The remainder of the reports of studies were adjudicated as low risk of bias in the selection and outcome domains, and only a few did not receive the maximum score in these domains due to issues related to the comparability of cohorts on design or analysis (Table 12).

11. Risk of bias assessment using the Newcastle‐Ottawa Scale (cohort studies).

Study ID/report	Selection				Comparability		Outcome			Total
	Representativeness of the exposed cohort (/1)	Selection of the non‐exposed cohort (/1)	Ascertainment of exposure (/1)	Demonstration that outcome of interest not present at start (/1)	Study controls for the main factor (/1)	Study controls for an additional factor (/1)	Assessment of outcome (/1)	Appropriate length of follow‐up (/1)	Adequacy of follow‐up (/1)
Bragg‐Gresham 2018	*	*	*	*	*	*	*	*	*	9
Dumaine 2018	*	*		*						3
Kasza 2016	*	*	*	*	*	*	*	*	*	9
Kjellstrand 2008 / Kjellstrand 2008	*	*	*	*	*	*	*	*	*	9
Kjellstrand 2008 / Kjellstrand 2010	*	*	*	*	*	*	*	*	*	9
Kojima 2012	*	*	*	*	*	*	*	*	*	9
Krahn 2019	*	*	*	*	*	*	*	*	*	9
Kraus 2007	*	*	*	*	*	*	*		*	8
Krishnasamy 2013	*	*	*	*	*	*	*	*	*	9
Lee 2002	*	*	*	*	*	*	*	*	*	9
Lorenzen 2012	*	*	*	*	*	*	*	*	*	9
Marshall 2021 / Marshall 2011	*	*	*	*	*	*	*	*	*	9
Marshall 2021 / Marshall 2013	*	*	*	*	*	*	*	*	*	9
Marshall 2021 / Marshall 2014	*	*	*	*	*	*	*	*	*	9
Marshall 2021 / Marshall 2016	*	*	*	*	*	*	*	*	*	9
Marshall 2021 / Marshall 2021	*	*	*	*	*	*	*	*	*	9
Murashima 2010	*	*	*	*	*	*	*	*	*	9
Nebel 2002	*	*	*	*			*	*	*	7
Nesrallah 2012	*	*	*	*	*	*	*	*	*	9
Nitsch 2011	*	*	*	*	*	*	*	*	*	9
NxStage‐USRDS 2012 / Kansal 2019	*	*	*	*	*	*	*	*	*	9
NxStage‐USRDS 2012 / Weinhandl 2012	*	*	*	*	*	*	*	*	*	9
NxStage‐USRDS 2012 / Weinhandl 2015a	*	*	*	*	*	*	*	*	*	9
NxStage‐USRDS 2012 / Weinhandl 2015b	*	*	*	*	*	*	*	*	*	9
NxStage‐USRDS 2012 / Weinhandl 2015c	*	*	*	*	*	*	*	*	*	9
NxStage‐USRDS 2012 / Weinhandl 2015d	*	*	*	*	*	*	*	*	*	9
Piccoli 2004	*	*	*	*	*	*	*	*	*	9
Rydell 2016	*	*	*	*	*	*	*	*	*	9
Rydell 2019
Rydell 2019a	*	*	*	*	*	*	*	*	*	9
Rydell 2019b	*	*	*	*	*	*	*	*	*	9
Sands 2009	*	*	*	*	*	*	*	*	*	9
Saner 2005	*	*	*	*	*	*	*	*	*	9
Suri 2015	*	*	*	*	*	*	*	*	*	9
Tablo IDE 2020 / Chertow 2020	*	*	*	*			*	*	*	7
Tennankore 2022	*	*	*	*	*	*	*	*	*	9
Toronto Group 2002 / Bergman 2008	*	*	*	*	*	*	*	*	*	9
Toronto Group 2002 / Bugeja 2004	*	*	*	*	*	*	*	*	*	9
Toronto Group 2002 / Chan 2002	*	*	*	*			*	*	*	7
Toronto Group 2002 / Chan 2003	*	*	*	*			*	*	*	7
Toronto Group 2002 / Chan 2005a	*	*	*	*			*	*	*	7
Van Oosten 2018	*	*	*	*	*		*	*	*	8
Wong 2019b	*	*	*	*	*	*	*	*	*	9
Xue 2015	*	*	*	*	*	*	*	*	*	9
Yeung 2021	*	*	*	*	*	*	*	*	*	9
Zimbudzi 2014	*	*	*	*			*	*	*	7

Using the adapted NOS for cross‐sectional studies, three reports of studies scored 5/10 stars, one report scored 6/10 stars, and seven reports scored 8/10 stars. Most risks of bias pertained to the sample size and comparability of non‐respondents (Table 13).

12. Risk of bias assessment using the adapted Newcastle‐Ottawa Scale (cross‐sectional studies).

Study / report	Selection				Comparability		Outcome		Total
	Representativeness of the sample (/1)	Sample size (/1)	Comparability of non‐respondents (/1)	Ascertainment of exposure (/2)	Study controls for the most important factor (/1)	Study controls for any additional factor	Assessment of outcome (/2)	Statistical test (/1)
Ageborg 2005	*			**			**		5
Griva 2010	*			**	*	*	**	*	8
Ha 2018	*			**			**		5
Hayhurst 2015	*			**			*	*	5
Jayanti 2016	*			**	*	*	**	*	8
Malmstrom 2008	*			**	*	*	**	*	8
Toronto Group 2002 / Cafazzo 2009	*			**	*	*	**	*	8
Toronto Group 2002 / Chan 2005	*			**	*	*	**	*	8
Watanabe 2014	*			**	*	*	**	*	8
Wong 2019a	*			**	*	*	**	*	8
Wright 2015	*	*		*			**	*	6

Effects of interventions

See: Table 1

See Table 1.

Cardiovascular death

Compared with ICHD, HHD had uncertain effects on cardiovascular death (Analysis 1.1 (2 studies, 30,900 participants): RR 0.92, 95% CI 0.80 to 1.07; I² = 0%; very low certainty evidence). The data for this outcome were mainly contributed by Marshall 2021, a large registry study. Data from Krishnasamy 2013 were not included in the meta‐analysis, firstly because this study focused on day‐of‐the‐week variability in cardiovascular death rather than overall cardiovascular death, and secondly, because it analysed ANZDATA Registry data over an overlapping time period with Marshall 2021. NxStage‐USRDS 2012 reported hazard ratios (HR) for cardiovascular death, which indicated no evidence of different cardiovascular death comparing HHD with ICHD (HR 0.92, 95% CI 0.78 to 1.09); however, we were unable to obtain sufficient data to allow inclusion in the meta‐analysis. When reviewing cardiovascular death as reported in all included studies (not just those included in the meta‐analysis), the risk of cardiovascular death appeared to be, in general, lower in patients receiving HHD compared to ICHD; however, cardiovascular death comprised a larger proportion of all deaths in those receiving HHD compared to ICHD (Table 14). Subgroup and sensitivity analyses could not be performed due to the small number of studies included in the analysis. The risk of bias, within the constraints of a NRSI, was considered to be low, with all included studies scoring 9/9 on the NOS (Table 12).

1.1. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 1: Cardiovascular death

13. Cardiovascular death as reported by studies.

Study ID	Reported outcome	Effect measure	Notes on analysis
Krishnasamy 2013	Odds of cardiac death by day of the week (Reference: odds of cardiac death for all days of the week)	Saturday: OR 1.23, 95% CI 0.74 to 2.03 (HHD) vs OR 0.85, 95% CI 0.76 to 0.95 (ICHD: ≤ 3 sessions/week) vs OR 1.27, 95% CI 0.40 to 3.47 (ICHD: > 3 sessions/week) Sunday: OR 0.54, 95% CI 0.34 to 0.86 (HHD) vs OR 0.88, 95% CI 0.78 to 0.96 (ICHD: ≤ 3 sessions/week) vs OR 0.79, 95% CI 0.25 to 2.51 (ICHD: > 3 sessions/week) Monday: OR 1.19, 95% CI 0.72 to 1.98 (HHD) vs OR 1.26, 95% CI 1.14 to 1.40 (ICHD: ≤ 3 sessions/week) vs OR 2.69, 95% CI 0.84 to 8.61 (ICHD: > 3 sessions/week) Tuesday: OR 1.04, 95% CI 0.68 to 1.60 (HHD) vs OR 1.08, 95% CI 0.97 to 1.20 (ICHD: ≤ 3 sessions/week) vs OR 1.45, 95% CI 0.49 to 4.23 (ICHD: > 3 sessions/week) Wednesday: OR 1.13, 95% CI 0.69 to 1.85 (HHD) vs OR 0.99, 95% CI 0.89 to 1.10 (ICHD: ≤ 3 sessions/week) vs OR 1.07, 95% CI 0.33 to 3.50 (ICHD: > 3 sessions/week) Thursday: OR 1.18, 95% CI 0.72 to 1.94 (HHD) vs OR 1.03, 95% CI 0.92 to 1.15 (ICHD: ≤ 3 sessions/week) vs OR 1.16, 95% CI 0.40 to 3.41 (ICHD: > 3 sessions/week) Friday: OR 0.90, 95% CI 0.54 to 1.51 (HHD) vs OR 0.95, 95% CI 0.85 to 1.06 (ICHD: ≤ 3 sessions/week) vs OR 1.12, 95% CI 0.36 to 3.47 (ICHD: > 3 sessions/week)	Analysis adjusted for age, sex, racial origin, body mass index, late referral, smoking status, chronic lung disease, coronary artery disease, cerebrovascular disease, peripheral vascular disease, diabetes mellitus, country of treatment (Australia or New Zealand), and centre size
Marshall 2021	Proportion of deaths due to cardiovascular disease	1998 to 2002: 59.2% (HHD) vs 48.9% (ICHD) 2003 to 2007: 64.9% (HHD) vs 55.6% (ICHD) 2008 to 2012: 74.5% vs 52.0% (ICHD) 2012 to 2017: 51.3% vs 49.5% (ICHD)	Analysed as 5‐year eras. Modelled as‐treated (i.e. time‐varying) modality, with a 90‐day lag in the attribution of death to modality
NxStage‐USRDS 2012	Cardiovascular death (Reference: ICHD)	HR 0.92, 95% CI 0.78 to 1.09 (intention to treat) HR 0.83, 95% CI 0.67 to 1.01 (as treated)	ICHD patients matched 5:1 for HHD patients. Matching variables included first date of follow‐up, demographic characteristics, and measures of disease severity
Saner 2005	Proportion of deaths due to cardiovascular disease	25.9% (HHD) vs 22.4% (ICHD)	Proportion of all deaths

CI: confidence interval; HHD: home haemodialysis; HR: hazard ratio; ICHD: in‐centre haemodialysis; OR: Odds ratio

All‐cause death

Compared with ICHD, HHD had uncertain effects on all‐cause death (Analysis 1.2 (9 studies, 58,984 participants): RR 0.80, 95% CI 0.67 to 0.95; I² = 84%; very low certainty evidence). The death rate was reported to be between 35 and 110 deaths/1000 patient‐years for HHD and between 57 and 127 for ICHD (Nesrallah 2012; NxStage‐USRDS 2012; Yeung 2021). The studies varied widely in size and included large registry analyses (Marshall 2021; Nesrallah 2012; NxStage‐USRDS 2012; Rydell 2019) as well as smaller single‐centre studies (Rydell 2016; Saner 2005; Xue 2015; Yeung 2021). The duration of follow‐up in these studies varied from less than two years to more than 10 years (Table 15).

1.2. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 2: All‐cause death

14. Follow‐up duration in studies assessing all‐cause death.

Study ID	HHD	ICHD
Marshall 2021	5‐, 7‐ and 10‐year windows	5‐, 7‐ and 10‐year windows
Nesrallah 2012	Median 1.8 years
NxStage‐USRDS 2012	Mean 1.8 years	Mean 1.7 years
Rydell 2016	Median 14.2 years	Median 10.8 years
Rydell 2019	Median 10.4 years	Median 7.0 years
Saner 2005	Mean 10.5 years	Mean 7.4 years
Xue 2015	20 months
Yeung 2021	Mean 4.44 years

Kjellstrand 2008, which used pooled data from five centres in the USA, Italy, France and the UK, reported that deaths at five years in patients receiving daily HHD were one‐third of those receiving daily ICHD and about two‐thirds of those treated with conventional ICHD; however, we were unable to extract data for meta‐analysis. When reviewing all‐cause death as reported in all included studies (not just those included in the meta‐analysis), the risk of death was consistently lower in patients receiving HHD compared to ICHD (Table 16). As was the case with the cardiovascular death outcome, subgroup and sensitivity analyses could not be performed due to the small number of studies included in the analysis. The risk of bias, within the constraints of a NRSI, was considered to be low, with all included studies scoring 9/9 on the NOS (Table 12).

15. Death as reported by studies.

Study ID	Reported outcome	Effect measure	Notes on analysis
Kasza 2016	Unadjusted pseudo‐survival curves	Demonstrated in figures, values unable to be extracted. Study reported that HHD with an AVF/AVG had better survival than any other exposure	ICHD with AVF/AVF vs ICHD with CVC vs HHD with AVF/AVG
	Time dependent HR for death (reference: ICHD patients with AVF/AVG)
Kjellstrand 2008	Cumulative survival	Demonstrated in figures, values unable to be extracted. Study reported that the 5‐year death of daily HD patients treating at home is one‐third and of those treated in centre	‐
Krahn 2019	Survival (%)	5‐year survival of 65% (HHD) vs 46% (ICHD)	Unadjusted
	Death (reference: ICHD)	HR 0.44, 95% CI 0.27 to 0.68 for HHD	Adjusted for age, sex, and comorbidity (Aggregated Diagnostic Groups count)
Marshall 2021	Death (reference: ICHD)	aHR 0.50, 95%CI 0.40 to 0.64 for HHD	Modelled as‐treated (i.e. time‐varying) modality, with a 90‐day lag in the attribution of death to modality
Nesrallah 2012	Death rate	6.1 deaths per 100 person‐years, 95% CI 2.6 to 8.2 (HHD) vs 10.5 deaths per100 person‐years, 95% CI 8.1 to 13.5 (ICHD)	Attributed all deaths to dialysis modality at index date, regardless of switches to other dialysis modalities. Stratified by matched set and country
	Death (eeference: ICHD)	HR 0.55, 95%CI 0.34 to 0.87 for intensive HHD
Nitsch 2011	1‐year survival	90%, 95% CI 87% to 91% (ICHD) vs 97%, 95% CI 94% to 99% (HHD patients of similar age and sex)	Unadjusted
	Long‐term survival (reference: HHD)	HR 1.06, 95% CI 0.55 to 2.04 for satellite HD	Time‐dependent variable for date of start of HHD, and baseline demographic variables. Follow‐up was not censored for kidney transplantation as it was entered as a time‐dependent variable. A time‐dependent variable for wait‐listing for kidney transplantation was included as a surrogate for general health status
NxStage‐USRDS 2012	All‐cause mortality for daily HHD (Reference: ICHD)	HR 0.87, 95%CI 0.78 to 0.97 (ITT) HR 0.82, 95%CI 0.72 to 0.94 (as treated)	ICHD patients matched 5:1 for HHD patients. Matching variables included first date of follow‐up, demographic characteristics, and measures of disease severity
Rydell 2016	Mean survival (years)	17.3 years (HHD) vs 13.0 years (ICHD)	Survival analysis was performed as ITT analysis, where patients were considered at risk also after changes to other modalities of KRT, including transplantation
	Survival (%)	5 years: 98% (HHD) vs 71% (ICHD) 10 years: 73% (HHD) vs 56% (ICHD)
Rydell 2019	Median survival (years)	18.5 years, IQR 10.4 – not available (HHD) vs 11.9 years, IQR 3.8 – not available (ICHD)	HHD patients matched with ICHD 1:4, using gender, Charlson Comorbidity Index, age (± 3 years) and date for start of KRT (± 3 years) were used as matching criteria. Matching was performed at day 0 of KRT. ITT, where patients were considered at risk also after switching to other KRT
	Survival (%)	5 years: 91% (HHD) vs 70% (ICHD) 10 years: 76% (HHD) vs 57% (ICHD) 20 years: 49% (HHD) vs 34% (ICHD)
Saner 2005	Survival (%)	5 years: 93% (HHD) vs 64% (ICHD) 10 years: 72% (HHD) vs 48% (ICHD) 20 years: 34% (HHD) vs 23% (ICHD)	Survival time was defined as the time from the initiation of the first dialysis treatment until death from any cause or the last date of follow‐up alive
Xue 2015	Death rate	First catheter: 0.00 events per 100 patient‐months (HHD) vs 0.4 (ICHD) All catheters: 0.26 events per 100 patient‐months (HHD) vs 0.33 (ICHD)	HHD patients matched with ICHD patients based on five variables: age (± 5 years), gender, race, dialysis vintage, and diabetes. Dialysis vintage was divided into seven categories: 1 day, > 1 to 30 days, > 1 to 3 months, > 3 to 12 months, > 1 to 2 years, > 2 to 5 years, and > 5 years. Adjustment for residual differences in age, gender, race, vintage, diabetes, as well as for primary cause of kidney failure(including diabetes, hypertension, glomerulonephritis, and polycystic kidney disease. All events after 20 months since the start of catheter were censored
Yeung 2021	Death rate	3.5 per 100 person‐years (HHD) vs 5.7 per 100 person‐years (ICHD)	ICHD patients matched 3:1 to HHD patients by age (within 5 years), gender and cause of kidney failure separated into glomerulonephritis, diabetes, hypertension and renovascular disease, reflux nephropathy and polycystic kidney disease. HR adjusted for BMI, smoking status, racial group and lung and vascular disease at dialysis commencement
	Death (reference: ICHD)	HR 0.49, 95% CI 0.30 to 0.80 for HHD

aHR: adjusted hazard ratio; AVF: arteriovenous fistula; AVG: arteriovenous graft; BMI: body mass index; CI: confidence interval; HD: haemodialysis; HHD: home haemodialysis; HR: hazard ratio; ICHD: in‐centre haemodialysis; IQR: interquartile range; ITT: intention to treat; KRT: kidney replacement therapy

All‐cause hospitalisation

Compared with ICHD, HHD had uncertain effects on the annual all‐cause admission rate (Analysis 1.3 ( 2 studies, 834 participants): MD ‐0.50 admissions/patient‐year, 95% CI ‐0.98 to ‐0.02; I² = 90%; very low certainty evidence). Three studies were excluded from the meta‐analysis as we were unable to obtain measures of variability; however, their reported mean annual hospital admission rates fell in the range between the two analysed studies, from 0.78 admissions/patient‐year in the HHD group to 1.54 admissions/patient‐year in the ICHD group (NxStage‐USRDS 2012; Suri 2015; Tennankore 2022).

1.3. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 3: All‐cause annual hospital admission rate (number of admissions/patient‐year)

Five studies reported all‐cause hospitalisation days/patient‐year; however, three studies were excluded from the meta‐analysis as we were unable to obtain measures of variability. Mean hospitalisation days/patient‐year may have been slightly lower in patients receiving HHD compared to ICHD (Analysis 1.4 (2 studies, 834 participants): MD ‐1.90, 95% CI ‐2.28 to ‐1.53; I² = 0%; low certainty evidence). Subgroup and sensitivity analyses could not be performed. The risk of bias, within the constraints of a NRSI, was considered to be low, with all included studies scoring 9/9 on the NOS (Table 12).

1.4. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 4: All‐cause hospitalisation days/patient‐year

Only Sands 2009 reported the number of patients experiencing one or more hospitalisation events, precluding meta‐analysis.

Kidney transplantation

Wait‐listing for kidney transplantation was reported in Nitsch 2011, which found that patients receiving ICHD were less likely to be wait‐listed than patients receiving HHD after adjusting for centre effect, year of start, age, gender, primary kidney disease, ethnicity and social deprivation (HR 0.56, 95% CI 0.44 to 0.79). Receipt of a kidney transplant during the study period was reported in six studies and was reported in a variety of ways in the individual studies, including relative incidence in HHD versus matched ICHD participants (1.22, 95% CI 1.06 to 1.66) (NxStage‐USRDS 2012), transplant rates/100 person‐years (9.5, 95% CI 7.6 to 12.1 for HHD versus 8.8. 95% CI 6.7 to 11.6 for ICHD) (Nesrallah 2012), and only the number of patients transplanted (Rydell 2016; Saner 2005; Xue 2015; Yeung 2021).

The meta‐analysis demonstrated HHD had an uncertain effect on the likelihood of receiving a transplant compared to ICHD (Analysis 1.5 (6 studies, 10,910 participants): RR 1.28, 95% CI 1.01 to 1.63; I² = 78%; very low certainty evidence). These studies varied substantially in size and design, from large registry analyses contributing most of the weight in the analysis (NxStage‐USRDS 2012) to much smaller single‐centre cohort studies (Saner 2005; Xue 2015). The risk of bias, within the constraints of a NRSI, was considered to be low, with all included studies scoring 9/9 on the NOS (Table 12).

1.5. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 5: Kidney transplantation during study period

Vascular access events

Vascular access interventions were reported in a variety of ways, and meta‐analysis could not be performed on any of the measures. One study reported the mean number of vascular access surgeries/patient (Saner 2005).

Two studies reported the number of patients experiencing a vascular access complication; however, these studies could not be meta‐analysed. Sands 2009 reported venous and arterial site complications separately, while Piccoli 2004 did not specify the site. Thus, we could not be certain that patients would not be included in the analysis more than once if they experienced both a venous and arterial site complication.

Piccoli 2004 reported vascular access failure rate, and Xue 2015 reported catheter‐related sepsis rate. It was unclear if vascular access complications occurred more frequently in patients receiving HHD or ICHD across these studies. Saner 2005 reported increased complications in patients receiving ICHD (3.9 versus 2.5 mean events/patient), Sands 2009 reported more frequent arterial site complications (0.25 versus 0.08 events/100 treatments) but fewer venous site complications in patients receiving HHD (0.08 versus 0.17 events/100 treatments), and two studies suggesting no difference in vascular access events (Piccoli 2004) or catheter‐related sepsis (Xue 2015). The risk of bias was low for all studies (Table 12).

Quality of life

QoL (including broader aspects such as mental health, symptoms, impact and view of health and functional ability) were reported using multiple different outcome measures across studies. While many of these tools measure different aspects of QoL and are thus not simply interchangeable, the low usage of numerous tools resulted in meta‐analysis being unable to be performed for many outcome measures. Two studies reported SF‐12 and SF‐36 domains of physical functioning, role‐physical, bodily pain, general health, vitality, social functioning, role‐emotional and mental health, and analyses generally indicated that HRQoL may have been slightly better in patients receiving HHD (Analysis 1.6; Analysis 1.7; Analysis 1.8; Analysis 1.9; Analysis 1.10; Analysis 1.11; Analysis 1.12; Analysis 1.13: 2 studies, 374 participants). It was uncertain if the Physical Component Score was higher in patients receiving HHD (Analysis 1.14 (5 studies, 922 participants): SMD 0.42, 95% CI 0.10 to 0.73; I² = 74%; very low certainty evidence); however, there was little or no difference in the Mental Component Score (Analysis 1.15 (5 studies, 922 participants): SMD 0.10, 95% CI ‐0.05 to 0.25; I² = 0%; very low certainty evidence).

1.6. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 6: Physical functioning (SF‐12; SF‐36)

1.7. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 7: Role: Physical (SF‐12; SF‐36)

1.8. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 8: Bodily pain (SF‐12; SF‐36)

1.9. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 9: General health (SF‐12; SF‐36)

1.10. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 10: Vitality (SF‐12; SF‐36)

1.11. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 11: Social functioning (SF‐12; SF‐36)

1.12. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 12: Role: Emotional (SF‐12; SF‐36)

1.13. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 13: Mental health (SF‐12; SF‐36)

1.14. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 14: Physical Component Summary (SF‐12; SF‐36)

1.15. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 15: Mental Component Summary (SF‐12; SF‐36)

Two studies reported KDQOL‐SF, and analyses indicated HRQoL may have been higher in most domains for patients receiving HHD compared to ICHD (Analysis 1.16; Analysis 1.17; Analysis 1.18; Analysis 1.19; Analysis 1.20; Analysis 1.21; Analysis 1.22; Analysis 1.23: 2 studies, 131 participants). The domains of social support, dialysis staff encouragement, patient satisfaction, physical functioning, role limitations ‐ physical, pain, general health, emotional well‐being, role limitations ‐ emotional, social function, energy/fatigue and overall health were reported only in Wright 2015, and so were not meta‐analysed. SF‐6D (Wong 2019a), EQ‐5D‐5L (Tablo IDE 2020) and 15D (Malmstrom 2008) were also only reported in single publications and were not meta‐analysed.

1.16. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 16: Symptoms and problems (KDQOL‐SF)

1.17. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 17: Effects of kidney disease (KDQOL‐SF)

1.18. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 18: Burden of kidney disease (KDQOL‐SF)

1.19. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 19: Work status (KDQOL‐SF)

1.20. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 20: Cognitive function (KDQOL‐SF)

1.21. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 21: Quality of social interaction (KDQOL‐SF)

1.22. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 22: Sexual function (KDQOL‐SF)

1.23. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 23: Sleep (KDQOL‐SF)

BDI‐II was reported in two studies, but meta‐analysis could not be performed as one study reported the mean and SD (Griva 2010), whereas the other reported only the frequency of scoring ranges as a categorical variable (Jayanti 2016). This was also the case for Spielberger State‐Trait Anxiety Inventory for Adults, where mean and SD were available for one study (Toronto Group 2002) but reported as a categorical variable in the other (Jayanti 2016).

Cognitive Depression Questionnaire, Illness Perceptions Questionnaire, Illness Effects Questionnaire, Treatment Effects Questionnaire (Griva 2010); IPOS Renal (Ha 2018), SOC questionnaire, Multidimensional Scale of Perceived Social Support (Ageborg 2005); SUPPH (Wright 2015); and Activity score (Hayhurst 2015) were each reported in only a single study and were not meta‐analysed. Appraisal of Self‐care Agency was reported in two studies; however, one study used a modified scale with a different scoring system, and the results could not be meta‐analysed (Ageborg 2005; Toronto Group 2002).

Most studies reporting QoL used a cross‐sectional design, with only one reporting longitudinal results for the EQ‐5D‐5L (Tablo IDE 2020). For several studies, questionnaires were mailed to potential participants, particularly those dialysing at home where convenience sampling could not be performed. Survey bias may, therefore, have been an issue. Dialysis vintage and duration on current modality varied between studies and were not reported in some; for example, for the Physical Component Score and the Mental Component Score outcomes, some studies reported duration on dialysis as a mean or median value, one study reported duration of ESKD diagnosis, and another reported duration on dialysis as a categorical variable (number of patients/time category) (Table 2). This may have impacted results since it is not known how long a patient must remain on a treatment modality before expecting to see either a beneficial or detrimental effect on QoL, but it is likely patients would require some time to become accustomed to a treatment. The risk of bias was low for the cohort study (Table 12) and low to moderate for the cross‐sectional studies, with four studies scoring 6 or less out of 10 on the adapted NOS (Table 13). Subgroup and sensitivity analyses could not be performed as there were few studies reporting the same metric of QoL.

Employment status

Employment could not be meta‐analysed in this review due to insufficient data. One study reported that self‐dialysis modality use was significantly associated with maintained employment at dialysis initiation compared to ICHD; however, the study was available in abstract format only, and the values for rates of employment were not reported (Bragg‐Gresham 2018). Employment status was the primary outcome of a cross‐sectional analysis of the Finnish Registry for Kidney Diseases; however, the study included participants aged from 15 to 64 years and, therefore, did not meet the review's inclusion criteria (Helantera 2012). For interest, since this review identified no other studies reporting this outcome, Helantera 2012 reported that patients receiving HHD (n = 57) had a significantly increased probability of being employed compared to patients receiving ICHD (n = 550) when adjusted for age, sex, cause of kidney failure, number of comorbid conditions and time since the start of KRT (prevalence rate ratio (PRR) 1.87, 95% CI 1.26 to 2.64). Furthermore, the study also found that patients treated with HHD as their last dialysis modality prior to kidney transplantation had an increased likelihood of being employed post‐transplant (PRR 2.14, 95% CI 1.68 to 2.74, reference group ICHD). To date, we have not been able to obtain data for the 18 years and older subgroup.

Recovery time

Recovery time was reported in two studies. The effect of HHD on recovery time was uncertain (Analysis 1.24 (2 studies, 348 participants): MD ‐2.0 hours, 95% CI ‐2.73 to ‐1.28; I² = 0%; low certainty evidence). These studies differed in design and analytical approach. One was a cross‐sectional survey, which included 288 patients from five UK centres and had a response rate of 94.2% (Jayanti 2016). Median dialysis vintage was 2.68 years (IQR 1.05 to 5.12) in patients receiving ICHD and 3.47 years (IQR 1.39 to 6.82) in patients receiving HHD; however, this study did not report how long patients had been on their current modality. The other study was a prospective cohort study of 30 prevalent HD patients who were transitioned from an eight‐week period of in‐centre to an eight‐week period of HHD, with a four‐week transition period (Tablo IDE 2020). In this study, recovery time was ascertained by a weekly questionnaire and within‐patient average scores for the in‐centre and in‐home periods were analysed. The risk of bias was considered to be low within the constraints of observational studies, with the cross‐sectional study scoring 8/10 on the adapted NOS and the cohort study scoring 7/9.

1.24. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 24: Recovery time

Cost‐effectiveness

Annual direct healthcare costs were analysed in four studies. Two studies reported costs year by year from dialysis initiation (Krahn 2019; Wong 2019b), one study reported costs for patients already established on dialysis for six months (Lee 2002), and another included all patients who received chronic dialysis during the time period of interest and did not specify vintage for inclusion (Van Oosten 2018). When cost in the first year of dialysis was analysed, it was uncertain if HHD was more cost‐effective than ICHD (Analysis 1.25 (2 studies, 10,077 participants): SMD ‐1.80, 95% CI ‐4.08 to ‐0.49; I² = 98%; very low certainty evidence). When data from the studies that did not specify dialysis vintage were included, this was also the case (4 studies, 13,809 participants: SMD ‐1.25, 95% CI ‐2.13 to ‐0.37, I² = 85%; very low certainty evidence). The effects of HHD on cost‐effectiveness in the second year of dialysis were uncertain (Analysis 1.26 (2 studies, 10,077 participants): SMD ‐2.30, 95% CI ‐6.94 to 2.34; I² = 100%; very low certainty evidence), which was also the case when data from studies that did not specify dialysis vintage were included in analysis (4 studies, 13,809 participants: SMD ‐1.47, 95% CI ‐2.72 to ‐0.21; I² = 88%; very low certainty evidence). The included studies were performed in Canada (Krahn 2019; Lee 2002), the Netherlands (Van Oosten 2018) and Hong Kong (Wong 2019b) between 2002 and 2019. Two further studies reported annual direct healthcare costs but were not included in the meta‐analysis due to unobtainable measures of variability (Malmstrom 2008; Nebel 2002). Nebel 2002 reported costs of treatment/year in Germany in 1999 and found that the cost of satellite HD was substantially greater than HHD (86,908 versus 59,591 DM, equivalent to 44,435 versus 30,468 EUR) (Bundesbank). Malmstrom 2008 also reported an increased cost of satellite HD compared to HHD in Finland, though the proportional difference between groups was less (39,781 versus 38,477 EUR).

1.25. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 25: Annual direct healthcare costs in first year of dialysis (currency as reported)

1.26. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 26: Annual direct healthcare costs in second year of dialysis (currency as reported)

Equivalent costs in US dollars (USD) with a December 2021 reference value are demonstrated in Table 17 for all studies reporting annual direct healthcare costs (including those not meta‐analysed). Reported costs varied widely across studies, even after calculating equivalent values, although HHD was consistently reported as more cost‐effective than ICHD in all studies. The costs of healthcare delivery would likely differ between countries and over time for various reasons unrelated to currency or inflation, and this should be taken into consideration when interpreting results. The risk of bias was low, with three studies scoring 9/9 and one scoring 8/9 on the NOS.

16. Annual healthcare costs reported in USD (December 2021 reference value).

Study ID	HHD	ICHD
Krahn 2019	$77,589.62	$130,579.58
Lee 2002	$48,006.82	$78,283.54
Malmstrom 2008	$70,113.6	$72,489.77
Nebel 2002	$36,057.27	$52,586.22
Van Oosten 2018	$137,263.28	$146,038.26
Wong 2019b	$14,132.96	$53,092.15

HHD: home haemodialysis; ICHD: in‐centre haemodialysis

Costs reported in studies in a different currency were converted to USD for the same year, using OECD data exchange rates (OECD 2022).

Cost from years prior to 2021 was subsequently converted to December 2021 value using the U.S. Bureau of Labour Statistics Consumer Price Index (CPI) Inflation Calculator (U.S. Bureau of Labor Statistics 2022)

Blood pressure

BP was reported in only one randomised cross‐over trial of nine patients, which showed it to be lower during the HHD phase versus ICHD (systolic BP 155 ± 18 versus 169 ± 24 mm Hg, diastolic BP 89 ± 6 versus 93 ± 9 mm Hg) (McGregor 2001).

SBP was analysed in three NRSIs, two of which reported pre‐dialysis BP measurements (Kojima 2012; Malmstrom 2008). The effect of HHD on SBP was uncertain (Analysis 1.27 (2 studies, 173 participants): MD ‐9.97 mm Hg, 95% CI ‐34.35 to 14.41; I² = 93%; very low certainty evidence). One study reported clinic BP measurements (Toronto Group 2002), and another did not specify how or when the measurements were taken (Wong 2019a). When these data were included in the analysis, HHD may lead to improved systolic BP control (4 studies, 491 participants: MD ‐11.78 mm Hg, 95% CI ‐21.11 to ‐2.46; I² = 81%; low certainty evidence).

1.27. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 27: Systolic blood pressure (NRSIs)

There was no significant difference in DBP (Analysis 1.28 (3 studies, 383 participants): MD 1.81 mm Hg, 95% CI ‐1.31 to 4.94; I² = 14.0%; very low certainty evidence). Rydell 2016 also reported no significant difference in DBP in patients receiving treatment with ICHD versus HHD; however, this study was not included in the analysis as DBP values were not obtainable. Two additional studies reported both SBP and DBP but were not included in the meta‐analysis due to unobtainable measures of variability (Kraus 2007; Lorenzen 2012).

1.28. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 28: Diastolic blood pressure (NSRIs)

The effect of HHD compared to ICHD on MAP was uncertain (Analysis 1.29 (2 studies, 44 participants): MD ‐7.01 mm Hg, 95% CI ‐12.57 to ‐1.46; I² = 23%; low certainty evidence). One additional study reported MAP but was not included in the analysis due to unobtainable measures of variability, although it also reported lower MAP in patients receiving HHD (90.02 versus 94.57 mm Hg) (Hayhurst 2015). Pulse pressure was reported in two studies (Kraus 2007; Murashima 2010) but was not analysed since measures of variability were only obtainable for one study (Murashima 2010).

1.29. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 29: Mean arterial pressure

Left ventricular mass

HHD had an uncertain effect on LVM index, based on two NRSIs in which patients were transitioned from ICHD to HHD (Analysis 1.30 (2 studies, 130 participants): MD ‐18.13 g/m², 95% CI ‐32.08 to ‐4.17; I² = 26%; low certainty evidence). LVMI was also measured in one RCT, which reported no change during the course of the study; however, values of these measurements were not obtainable (McGregor 2001). The risk of bias was low within the constraints of a NRSI (Table 12).

1.30. Analysis.

Comparison 1: Home versus in‐centre haemodialysis, Outcome 30: Left ventricular mass index [g/m²]

Adverse events

One study reported that clinically significant hypotension (SBP ≤ 90 or DBP ≤ 55 mm Hg) was less likely when patients were treated with HHD compared to ICHD (odds ratio (OR) 0.36, 95% CI 0.16 to 0.81) (Murashima 2010). Various other adverse events that were not outcomes of interest were reported in a few studies (Table 18). Most studies did not specifically report adverse events, possibly due to observational cohort (including registry analysis) or cross‐sectional study design.

17. Other adverse events.

	HHD group	ICHD group
Kraus 2007	Patients reporting at least one AE: 13 Absolute number of AEs: 25 All AEs: 2.10 events/100 treatments Event rate/100 treatments Back pain: 0.00 Arthralgia: 0.17 Night cramps: 0.25 Neck pain: 0.00 Dysgeusia: 0.00 Dizziness: 0.00 Tremor: 0.00 Sinusitis: 0.08 Nausea: 0.17	Patients reporting at least one AE: 24 Absolute number AEs: 76 Overall AEs (defined as any unfavourable or unintended sign, symptom or disease temporally associated with the use of the device) significantly higher in the in‐centre group 5.3 events/100 treatments (P = 0.007) Event rate per 100 treatments Back pain: 0.21 Arthralgia: 0.42 Night cramps: 0.14 Neck pain: 0.70 Dysgeusia: 0.21 Dizziness: 0.141 Tremor: 0.21 Sinusitis: 0.07 Nausea: 0.17
Malmstrom 2008	Nausea: 3/33 patients Muscle cramps: 3/33 Headaches: 1/33 Hypotensive episodes: 3/33	Nausea: 1/32 patients Muscle cramps: 6/32 Headache: 1/32 Hypotensive episodes: 10/33
Sands 2009	18 patients had 40 AEs Overall 3.34 AEs/100 treatments	12 patients had 69 AEs Overall 5.84 AEs/100 treatments
Tablo IDE 2020	Overall 16 AEs (1.8%) 6 prespecified events (3 severe: 1 missed treatments resulting in hospital admission for overload, 2 falls), all considered unrelated to Tablowere	Overall 17 AEs (1.9%) 2 prespecified events (1 severe: fall), all considered unrelated to Tablo

AE: adverse events; HHD: home haemodialysis; ICHD: in‐centre haemodialysis; AE; adverse event

Discussion

Summary of main results

This systematic review identified one small RCT that compared the effects of shorter hours of ICHD with longer hours of HHD in nine prevalent HHD patients. The study found that BP was lower when patients were receiving HHD compared to ICHD.

Several NRSIs meeting our inclusion criteria were identified, varying in design from small single‐centre studies to large registry analyses. Meta‐analysis of NRSIs indicated uncertain differences in risk of all‐cause death and hospitalisation in patients receiving HHD compared to ICHD. The effects of HHD on kidney transplantation and recovery time were also uncertain. Effects on QoL were variable across different measures and, in some domains, did not reach statistical significance but generally favoured HHD. In a meta‐analysis, cost‐effectiveness was uncertain; however, the results of studies generally favoured HHD over ICHD. Although SBP, MAP, and LVM index were lower in patients receiving HHD compared to ICHD, there was no difference in DBP.

Overall completeness and applicability of evidence

In this review, treatment with HHD was associated with uncertain but possible benefits compared to ICHD, including death, QoL and cost benefit. Improvement in some surrogate outcomes was also seen. These benefits are certainly plausible; however, they may have been due at least in part to differences in how dialysis was delivered, including duration, frequency, blood flow rate, type of machine used and vascular access. Intensive dialysis regimens have been associated with reduced medication burden and improved biochemical parameters (Jardine 2017; Ok 2011) and reduced ultrafiltration rate, intradialytic hypotension and myocardial stunning (Jefferies 2011), which may, in turn, result in improved clinical outcomes. In many studies included in this review, dialysis treatments at home were longer, more frequent or both. There was also variation between studies in what was considered a conventional prescription, likely due to differences in standard practice from country to country. Likewise, while NxStage was used in some studies comparing HHD and ICHD, NxStage is not available or not widely used in some countries. Some studies reported treatment duration or parameters in little detail or did not describe it at all. While all of these factors have potentially impacted the generalisability of results, differences in dialysis prescription were not considered a source of bias in this review since the option to easily convert to intensive dialysis is one of the distinct benefits of HHD.

Studies varied in design and included prospective longitudinal cohort studies, retrospective cohort studies including registry analyses, cross‐sectional studies and one cross‐over RCT. Dialysis vintage was not always described in studies, nor was the duration of the current dialysis modality. Across included studies, some were conducted in patients already established on HHD, some in patients established on HD but new to HHD, and in others, patients were new to any form of dialysis. The experiences of different venues for HD treatment may be systematically different between these populations. Thus, varying vintage and familiarity with dialysis modality may have been an effect modifier for some outcomes. QoL data were derived nearly entirely from cross‐sectional studies in which patients had been treated with a dialysis modality for varying and, in some cases, unknown time periods.

Additionally, descriptions of selection and support for HHD (e.g. payment source for equipment, water and electricity; presence of support at home; financial incentives) were generally not described and could have led to residual confounding. The lack of standardised descriptions of selection and support for HHD limits the generalisability of the findings.

One of the main limitations of this review was the inability to meta‐analyse many outcomes due to the large number of ways outcomes were reported. Planned subgroup and sensitivity analyses could not be performed due to the small number of studies available for analysis for each outcome. Differences in outcome reporting were due to the use of different tools (e.g. QoL measures), as well as different reporting metrics (e.g. hospital admissions versus duration versus number of patients hospitalised). While multiple metrics for an outcome may be valuable since they often reflect different aspects of that outcome, the highly variable, limited or incomplete data reporting seen in this review inevitably limits comparability to other studies, ability to benchmark, and therefore contributes to research waste. This has been one of the motivating issues behind the Standardised Outcomes in Nephrology (SONG) Initiative, which establishes a set of core outcome measures to be reported in all trials in a given field.

Quality of the evidence

Overall, the risk of bias assessment for most studies was low following assessment by the NOS and adapted NOS. However, these observational studies were subject to unavoidable bias due to their study design. Patients receiving HHD were often younger and less comorbid than patients receiving ICHD, thus potentially leading to selection bias and confounding by indication. Some cohort studies attempted to account for potential differences by using patient matching techniques or reporting on patients transitioning from one modality to another, thus acting as their own control. However, many studies were cross‐sectional studies and did not adjust for other factors. It is possible that differences in outcomes reflected not just the effects of the treatment modality but the differing study populations.

The risk of bias was increased by the lack of comparability between treatment interventions, including differences in dialysate composition in the RCT by McGregor 2001, and differences in treatment prescription and delivery as previously outlined. The heterogeneity of studies also impacted the quality of evidence, and the overall quality for all outcomes was low or very low (Table 1).

Potential biases in the review process

While this review was conducted using standard Cochrane methodology, potential biases existed in the review process that may have limited the validity of the findings. First, only one RCT was included after searching and assessment according to our inclusion criteria, and the remainder of the included studies were NRSIs. Secondly, publication bias may have existed (i.e. bias caused by lack of publication with neutral or opposite effects). Due to a lack of sufficient data, we could not test for potential publication bias. Third, many outcomes could not be meta‐analysed due to high variability in reporting metrics utilised and reported data being insufficient for extraction. Beneficial effects or harms of HHD may not have been observed due to a lack of analysable data.

Agreements and disagreements with other studies or reviews

A 2003 review of HHD summarised data from 22 cohort studies and compared HHD versus hospital or satellite unit HD for people with kidney failure (Mowatt 2003). The authors reported that people treated with HHD generally experienced better QoL, lower hospitalisation, longer survival and increased likelihood of employment. HHD was also more cost‐effective than hospital or satellite HD. However, partners of these patients appeared to experience increased treatment burden and decreased satisfaction. Confounding by indication may have impacted the results. As was the case in the current review, details regarding dialysis prescription and delivery method were often unavailable when these may have impacted outcomes. Overall, the consistency and effect size of HHD on clinical outcomes (QoL, death, hospital admission) based on these analyses were uncertain. A systematic qualitative review of HHD versus ICHD identified 44 studies and found that clinical outcomes in patients receiving HHD were generally superior, including improved survival, cardiovascular parameters and QoL (Miller 2018). However, there were several limitations to the review, including data were mostly obtained from retrospective cohort studies, many being small observational studies, and prone to confounding since the patients receiving HHD were younger with fewer comorbidities, had increased pre‐dialysis education and were more likely to have been referred earlier to a nephrologist. Finally, a systematic review and meta‐analysis of HRQoL outcomes in HHD versus ICHD suggested nominally improved physical HRQoL in patients receiving HHD; however, included studies provided low‐quality evidence, and many had design issues (Bonenkamp 2020).

Intensive dialysis regimens are more easily implemented in the home setting since HHD has the potential to increase dialysis frequency and duration through greater treatment flexibility. While the location of dialysis (i.e. home versus in‐centre) was not the exposure of interest, a review comparing intensive dialysis with conventional HD suggested intensive dialysis may be associated with improved clinical outcomes. However, effect estimates were imprecise, and the effects of confounding could not be excluded (Mohr 2001). Similarly, reviews of 14 cohort studies of daily HD (Suri 2006) and 14 cohort studies of nocturnal HD (Walsh 2005) found that evidence, especially for hard clinical outcomes such as death, tended to be limited by small sample sizes for available studies, non‐comparable control groups, bias due to treatment selection and attrition, and insufficient data for potential risks. Finally, two RCTs compared increased frequency, duration of HD or both (FHN Trial Group 2010; Walsh 2006). Improvements in cardiac function and selected components of QoL and BP control indicated that longer hours of dialysis might improve patient‐relevant outcomes in adults treated with HD, although dialysis vascular access complications may be more frequent.

Authors' conclusions

Implications for practice.

Currently, available data indicate uncertain treatment benefits of HHD on patient‐relevant outcomes. However, this is based on low to very low certainty evidence and must be interpreted and applied cautiously.

Implications for research.

Given the existing poor survival and high symptom burden associated with HD treatment, an understanding of modifiable determinants of clinical outcomes in this population is needed. Non‐randomised evidence suggested that HHD was associated with better survival and QoL.

No further RCTs comparing HHD and ICHD have been performed since the prior version of this review, and we are not aware of any that are planned. Recently, there has been increased emphasis on shared decision‐making based on patients' own priorities, which is of great importance during the modality selection process. For some patients, freedom and flexibility may be a priority, leading them to favour HHD. On the other hand, some patients may prioritise reduced mental and treatment burden, leading them to prefer ICHD. As such, it may be unlikely that a large, well‐designed RCT of HHD versus ICHD will be performed in the future.

Therefore, future research in this area may continue to be mainly based on NRSIs. However, it is essential that future studies align outcome measures and metrics with other research in the field to allow comparison between studies, establish outcome effects with greater certainty, and avoid research waste.

What's new

Date	Event	Description
22 May 2024	Amended	Minor edit to plain language heading

History

Protocol first published: Issue 1, 2012
Review first published: Issue 11, 2014

Date	Event	Description
8 April 2024	New search has been performed	Inclusion criteria revised to also include non‐randomised studies. MEDLINE (OVID) and EMBASE (OVID) search strategies updated.
8 April 2024	New citation required and conclusions have changed	New studies included
4 December 2014	Review declared as stable	As of December 2014 this Cochrane Review was no longer being updated. There have been no new studies published on this topic in the past 12 years and there are currently no registered ongoing studies.

Appendices

Appendix 1. Electronic search strategies

Database	Search terms
CENTRAL	MeSH descriptor Hemodialysis, Home, this term only (home hemodialysis):ti,ab,kw in Clinical Trials (home haemodialysis):ti,ab,kw in Clinical Trials (home dialysis):ti,ab,kw in Clinical Trials MeSH descriptor Peritoneal Dialysis explode all trees (home):ti,ab,kw in Clinical Trials (#5 AND #6) (dialysis modalit*):ti,ab,kw in Clinical Trials (#6 AND #8) MeSH descriptor Renal Dialysis, this term only (#6 AND #10) MeSH descriptor Kidney Failure, Chronic, this term only (CKF or CKD or CRF or CRD):ti,ab,kw in Clinical Trials (predialysis or pre‐dialysis):ti,ab,kw in Clinical Trials (chronic kidney):ti,ab,kw or (chronic renal):ti,ab,kw in Clinical Trials (end‐stage renal):ti,ab,kw or (endstage renal):ti,ab,kw or (end‐stage kidney):ti,ab,kw or (endstage kidney):ab in Clinical Trials (ESRF or ESKF or ESRD or ESKD):ti,ab,kw in Clinical Trials (#12 OR #13 OR #14 OR #15 OR #16 OR #17) (#6 AND #18) (#1 OR #2 OR #3 OR #4) (#20 OR #7 OR #9 OR #11 OR #19)
MEDLINE (OVID)	Hemodialysis, Home/ home h?emodialysis.tw. home dialysis.tw. exp Peritoneal Dialysis/ and/3‐4 dialysis modalit$.tw. and/3,6 Renal Dialysis/ and/3,8 Kidney Failure, Chronic/ (CKF or CKD or CRF or CRD).tw. (predialysis or pre‐dialysis).tw. (chronic kidney or chronic renal).tw. (end‐stage renal or end‐stage kidney or endstage renal or endstage kidney).tw. (ESRF or ESKF or ESRD or ESKD).tw. or/10‐15 and/3,16 or/1‐3,5,7,9,17
EMBASE (OVID)	home dialysis/ home h?emodialys?s.tw. home dialys?s.tw. or/2‐3 hemodialysis patient/ and/4‐5 peritoneal dialysis/ and/4,7 chronic kidney failure/ (predialysis or pre‐dialysis).tw. or/9‐10 and/4,11 or/1‐3,6,8,12

Appendix 2. Newcastle‐Ottawa Scale (for cohort studies)

Selection (maximum 1 star for each numbered item)

1. Representativeness of the exposed cohort)

   1. truly representative of the average _______________ (describe) in the community*

   2. somewhat representative of the average ______________ in the community*

   3. selected group of users eg nurses, volunteers

   4. no description of the derivation of the cohort

2. Selection of the non exposed cohort

   1. drawn from the same community as the exposed cohort*

   2. drawn from a different source

   3. no description of the derivation of the non exposed cohort

3. Ascertainment of exposure

   1. secure record (e.g. surgical records)*

   2. structured interview*

   3. written self report

   4. no description

4. Demonstration that outcome of interest was not present at start of study

   1. yes*

   2. no

Comparability (maximum 2 stars)

1. Comparability of cohorts on the basis of the design or analysis

   1. study controls for _____________ (select the most important factor)*

   2. study controls for any additional factor* (this criteria could be modified to indicate specific control for a second important factor)

Outcome (maximum 1 star for each numbered item)

1. Assessment of outcome

   1. independent blind assessment*

   2. record linkage*

   3. self report

   4. no description

2. Was follow‐up long enough for outcomes to occur

   1. yes (select an adequate follow up period for outcome of interest)*

   2. no

3. Adequacy of follow up of cohorts

   1. complete follow up ‐ all subjects accounted for*

   2. subjects lost to follow up unlikely to introduce bias ‐ small number lost ⇢ ____ % (select an adequate %) follow up, or description provided of those lost)*

   3. follow up rate < ____% (select an adequate %) and no description of those lost) no statement

Appendix 3. Adapted Newcastle‐Ottawa Scale (for cross‐sectional studies)

Selection (maximum 5 stars)

1. Representativeness of the sample:

   1. Truly representative of the average in the target population* (all subjects or random sampling)

   2. Somewhat representative of the average in the target population* (non‐random sampling)

   3. Selected group of users

   4. No description of the sampling strategy

2. Sample size:

   1. Justified and satisfactory*

   2. Not justified

3. Non‐respondents:

   1. Comparability between respondents and non‐respondents characteristics is established, and the response rate is satisfactory*

   2. The response rate is unsatisfactory, or the comparability between respondents and non‐respondents is unsatisfactory

   3. No description of the response rate or the characteristics of the responders and the non‐responders

4. Ascertainment of the exposure (risk factor):

   1. Validated measurement tool**

   2. Non‐validated measurement tool, but the tool is available or described*

   3. No description of the measurement tool

Comparability (maximum 2 stars)

1. The subjects in different outcome groups are comparable, based on the study design or analysis. Confounding factors are controlled

   1. The study controls for the most important factor (select one)*

   2. The study control for any additional factor. *

Outcome (maximum 3 stars)

1. Assessment of the outcome:

   1. Independent blind assessment**

   2. Record linkage**

   3. Self report*

   4. No description

2. Statistical test:

   1. The statistical test used to analyse the data is clearly described and appropriate, and the measurement of the association is presented, including confidence intervals and the probability level (p value)*

   2. The statistical test is not appropriate, not described or incomplete

Data and analyses

Comparison 1. Home versus in‐centre haemodialysis.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Cardiovascular death	2	30900	Risk Ratio (IV, Random, 95% CI)	0.92 [0.80, 1.07]
1.2 All‐cause death	9	58984	Risk Ratio (M‐H, Random, 95% CI)	0.80 [0.67, 0.95]
1.3 All‐cause annual hospital admission rate (number of admissions/patient‐year)	2	834	Mean Difference (IV, Random, 95% CI)	‐0.50 [‐0.98, ‐0.02]
1.4 All‐cause hospitalisation days/patient‐year	2	834	Mean Difference (IV, Random, 95% CI)	‐1.90 [‐2.28, ‐1.53]
1.5 Kidney transplantation during study period	6	10910	Risk Ratio (M‐H, Random, 95% CI)	1.28 [1.01, 1.63]
1.6 Physical functioning (SF‐12; SF‐36)	2	374	Std. Mean Difference (IV, Random, 95% CI)	0.56 [‐0.03, 1.16]
1.6.1 SF‐12	1	294	Std. Mean Difference (IV, Random, 95% CI)	0.29 [‐0.05, 0.62]
1.6.2 SF‐36	1	80	Std. Mean Difference (IV, Random, 95% CI)	0.89 [0.43, 1.36]
1.7 Role: Physical (SF‐12; SF‐36)	2	374	Std. Mean Difference (IV, Random, 95% CI)	0.42 [‐0.12, 0.96]
1.7.1 SF‐12	1	294	Std. Mean Difference (IV, Random, 95% CI)	0.17 [‐0.17, 0.50]
1.7.2 SF‐36	1	80	Std. Mean Difference (IV, Random, 95% CI)	0.72 [0.26, 1.18]
1.8 Bodily pain (SF‐12; SF‐36)	2	374	Std. Mean Difference (IV, Random, 95% CI)	0.18 [‐0.39, 0.75]
1.8.1 SF‐12	1	294	Std. Mean Difference (IV, Random, 95% CI)	‐0.09 [‐0.42, 0.24]
1.8.2 SF‐36	1	80	Std. Mean Difference (IV, Random, 95% CI)	0.49 [0.04, 0.94]
1.9 General health (SF‐12; SF‐36)	2	374	Std. Mean Difference (IV, Random, 95% CI)	0.34 [0.07, 0.60]
1.9.1 SF‐12	1	294	Std. Mean Difference (IV, Random, 95% CI)	0.35 [0.02, 0.68]
1.9.2 SF‐36	1	80	Std. Mean Difference (IV, Random, 95% CI)	0.32 [‐0.12, 0.77]
1.10 Vitality (SF‐12; SF‐36)	2	374	Std. Mean Difference (IV, Random, 95% CI)	0.54 [0.27, 0.80]
1.10.1 SF‐12	1	294	Std. Mean Difference (IV, Random, 95% CI)	0.55 [0.21, 0.88]
1.10.2 SF‐36	1	80	Std. Mean Difference (IV, Random, 95% CI)	0.51 [0.06, 0.97]
1.11 Social functioning (SF‐12; SF‐36)	2	374	Std. Mean Difference (IV, Random, 95% CI)	0.56 [0.29, 0.83]
1.11.1 SF‐12	1	294	Std. Mean Difference (IV, Random, 95% CI)	0.51 [0.17, 0.84]
1.11.2 SF‐36	1	80	Std. Mean Difference (IV, Random, 95% CI)	0.67 [0.22, 1.13]
1.12 Role: Emotional (SF‐12; SF‐36)	2	374	Std. Mean Difference (IV, Random, 95% CI)	0.48 [‐0.84, 1.81]
1.12.1 SF‐12	1	294	Std. Mean Difference (IV, Random, 95% CI)	‐0.18 [‐0.51, 0.15]
1.12.2 SF‐36	1	80	Std. Mean Difference (IV, Random, 95% CI)	1.17 [0.69, 1.65]
1.13 Mental health (SF‐12; SF‐36)	2	374	Std. Mean Difference (IV, Random, 95% CI)	0.33 [‐0.23, 0.89]
1.13.1 SF‐12	1	294	Std. Mean Difference (IV, Random, 95% CI)	0.07 [‐0.26, 0.40]
1.13.2 SF‐36	1	80	Std. Mean Difference (IV, Random, 95% CI)	0.64 [0.19, 1.10]
1.14 Physical Component Summary (SF‐12; SF‐36)	5	922	Std. Mean Difference (IV, Random, 95% CI)	0.42 [0.10, 0.73]
1.14.1 SF‐12	4	842	Std. Mean Difference (IV, Random, 95% CI)	0.30 [0.04, 0.57]
1.14.2 SF‐36	1	80	Std. Mean Difference (IV, Random, 95% CI)	1.05 [0.58, 1.53]
1.15 Mental Component Summary (SF‐12; SF‐36)	5	922	Std. Mean Difference (IV, Random, 95% CI)	0.10 [‐0.05, 0.25]
1.15.1 SF‐12	4	842	Std. Mean Difference (IV, Random, 95% CI)	0.09 [‐0.07, 0.25]
1.15.2 SF‐36	1	80	Std. Mean Difference (IV, Random, 95% CI)	0.20 [‐0.24, 0.65]
1.16 Symptoms and problems (KDQOL‐SF)	2	131	Mean Difference (IV, Random, 95% CI)	3.96 [‐7.48, 15.40]
1.17 Effects of kidney disease (KDQOL‐SF)	2	131	Mean Difference (IV, Random, 95% CI)	11.61 [‐0.52, 23.75]
1.18 Burden of kidney disease (KDQOL‐SF)	2	131	Mean Difference (IV, Random, 95% CI)	‐0.65 [‐8.70, 7.40]
1.19 Work status (KDQOL‐SF)	2	131	Mean Difference (IV, Random, 95% CI)	33.32 [12.52, 54.11]
1.20 Cognitive function (KDQOL‐SF)	2	131	Mean Difference (IV, Random, 95% CI)	2.12 [‐3.10, 7.34]
1.21 Quality of social interaction (KDQOL‐SF)	2	131	Mean Difference (IV, Random, 95% CI)	4.82 [‐0.78, 10.43]
1.22 Sexual function (KDQOL‐SF)	2	131	Mean Difference (IV, Random, 95% CI)	8.62 [‐0.71, 17.95]
1.23 Sleep (KDQOL‐SF)	2	131	Mean Difference (IV, Random, 95% CI)	4.02 [‐2.19, 10.22]
1.24 Recovery time	2	348	Mean Difference (IV, Random, 95% CI)	‐2.00 [‐2.73, ‐1.28]
1.25 Annual direct healthcare costs in first year of dialysis (currency as reported)	4	13809	Std. Mean Difference (IV, Random, 95% CI)	‐1.25 [‐2.13, ‐0.37]
1.25.1 Year 1 of dialysis	2	10077	Std. Mean Difference (IV, Random, 95% CI)	‐1.80 [‐4.08, 0.49]
1.25.2 Dialysis vintage not specified	2	3732	Std. Mean Difference (IV, Random, 95% CI)	‐0.67 [‐1.59, 0.25]
1.26 Annual direct healthcare costs in second year of dialysis (currency as reported)	4	13809	Std. Mean Difference (IV, Random, 95% CI)	‐1.47 [‐2.72, ‐0.21]
1.26.1 Year 2 of dialysis	2	10077	Std. Mean Difference (IV, Random, 95% CI)	‐2.30 [‐6.94, 2.34]
1.26.2 Dialysis vintage not specified	2	3732	Std. Mean Difference (IV, Random, 95% CI)	‐0.63 [‐1.65, 0.38]
1.27 Systolic blood pressure (NRSIs)	4	491	Mean Difference (IV, Random, 95% CI)	‐11.78 [‐21.11, ‐2.46]
1.27.1 NRSIs, pre‐dialysis measurement	2	173	Mean Difference (IV, Random, 95% CI)	‐9.97 [‐34.35, 14.41]
1.27.2 NRSIs, clinic or other measurements	2	318	Mean Difference (IV, Random, 95% CI)	‐12.23 [‐17.92, ‐6.54]
1.28 Diastolic blood pressure (NSRIs)	3	383	Mean Difference (IV, Random, 95% CI)	1.81 [‐1.31, 4.94]
1.28.1 NRSIs, pre‐dialysis measurement	1	65	Mean Difference (IV, Random, 95% CI)	4.00 [‐1.06, 9.06]
1.28.2 NRSIs, clinic or other measurement	2	318	Mean Difference (IV, Random, 95% CI)	0.46 [‐3.51, 4.44]
1.29 Mean arterial pressure	2	44	Mean Difference (IV, Random, 95% CI)	‐7.01 [‐12.57, ‐1.46]
1.30 Left ventricular mass index [g/m2]	2	130	Mean Difference (IV, Random, 95% CI)	‐18.13 [‐32.08, ‐4.17]

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Ageborg 2005.

Study characteristics
Methods	Study design Cross‐sectional Study grouping: parallel Study characteristics Country: Sweden Setting: Home and in‐centre HD patients in a single Swedish centre Inclusion criteria: ability to read and understand the Swedish language; all the patients on HHD and all the self‐care HD patients were asked to participate; convenience sample of ICHD patients was recruited from patients who dialysed Monday, Wednesday and Friday afternoons in the dialysis unit
Participants	Baseline characteristics HHD Mean age: 47 years Male: 80% Mean education: 16 years Mean duration on dialysis: 34 months Primary disease Diabetes (20%); GN (60%) Self‐care HD Mean age: 47 years Male: 83% Mean education: 16 years Mean duration on dialysis: 42 months Primary disease Diabetes (0%); GN (17%) ICHD Mean age: 60 years Male: 88% Mean education: 12 years Mean duration on dialysis: 43 months Primary disease Diabetes (25%); GN (0%) Group differences ICHD patients were older (mean 60 vs. 47 years in both the HHD and self‐care groups). ICHD patients had lower education level (mean 12 years vs. 16 years in both the HHD and self‐care groups). Definitions HHD patients: trained by a nurse in centre for 8 weeks to manage their own treatment & have chosen to dialyse at home (n=5) Self‐care HD patients: trained by a nurse in centre for 8 weeks to manage their own treatment & carry out their treatment in the dialysis unit (n=6) ICHD patients: dialysed by nurses in a hospital satellite unit. All patients who dialysed between 1:30 PM and 8:00 PM on the same day were included in the study (n=8)
Interventions	HHD Location: own home Operator: patient Training: by nurse; 8 weeks in centre Self‐care HD Location: dialysis unit Operator: patient Training: by nurse; 8 weeks in centre ICHD Location: hospital satellite unit Operator: Nurse Training: None
Outcomes	Outcomes relevant to this review SF‐36 Physical Functioning SF‐36 Role‐Physical SF‐36 Bodily Pain SF‐36 General Health SF‐36 Vitality SF‐36 Social Functioning SF‐36 Role‐Emotional SF‐36 Mental Health Appraisal of Self‐Care Agency (ASA scale) Sense of Coherence (SOC) questionnaire Physical Component Summary (SF‐12) Mental Component Summary (SF‐12)
Identification	Additional information Sponsorship source: None reported Author name: Maria Ageborg Institution: Department of Nephrology, Karolinska University, Solna Email: maria.ageborg@karolinska.se Address: Home hemodialysis Unit, Karolinska University Hospital, SE‐17176 Stockholm, Sweden Other authors: Britt‐Louise Allenius, Claes Cederfjall
Notes	Contacted author to obtain values for SF‐36 (presented in graph only), awaiting response

Bragg‐Gresham 2018.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: parallel Statistical analysis: multivariable logistic regression Accounting for patient demographics, comorbidities, pre‐ESKD care (yes/no), and insurance status Study characteristics Country: USA Setting: United States Renal Data System (USRDS); Medical Evidence Report (CMS 2728) Inclusion criteria: incident dialysis patients ages 21 to 65 who were working (full or part‐time) 6 months prior to dialysis initiation in the United States Renal Data System (USRDS) using data available in the Medical Evidence Report (CMS 2728) Exclusion criteria: subjects who were institutionalized, non‐ambulatory or needing assistance
Participants	Baseline characteristics HHD In‐centre self‐care HD ICHD CAPD CCPD Group differences Patients electing self‐care modalities: higher prevalence of being employed 6 months prior to dialysis initiation (51%, 43%, 48%, and 53%, respectively for in‐centre self‐care, HHD, CAPD, and CCPD vs. 28% for HD patients) Self‐care patients: higher prevalence of pre‐ESKD nephrology care (about 80% vs 56% for HD)
Interventions	HHD Self‐care HD ICHD
Outcomes	Outcomes relevant to this review Employment prior to dialysis initiation (6 months), % Maintained employment (AOR) Outcomes reported Primary outcome: Odds of sustained employment during the 6‐month transition period to dialysis initiation (defined as employed at 6 months prior and at dialysis initiation) Electing self‐dialysis associated with maintained employment status (AORs ranged from 1.37 to 1.96, all P < 0.0001) compared to HD. No specific result was reported for HHD Pre‐ESKD care was associated with 42% higher odds of maintained employment (AOR‐1.42, 95% CI 1.39 to 1.45). Other characteristics associated with maintained employment included younger age, White race, non‐Hispanic ethnicity, few comorbidities, and private health insurance (P < 0.0001)
Identification	Additional information Sponsorship source: none reported Comments: abstract from the 38th Annual Dialysis Conference March 3–6, 2018 (Orlando, Florida); page 21 Authors name: Jennifer Bragg‐Gresham Institution: University of Michigan, Ann Arbor, MI, USA Email: jennb@med.umich.edu Address: not reported Other authors: Dori Schatell, Beth Witten, Yuxin Nie, Rajiv Saran
Notes	Abstract only. Contacted author to obtain data on maintained employment for each modality separately, awaiting response.

Dumaine 2018.

Study characteristics
Methods	Study design Prospective cohort study Study grouping: parallel group Design details Prospective, mixed‐methods, pilot study Over a 3‐month recruitment period, patients initiating dialysis were asked to complete a baseline assessment of HRQoL using the Kidney Disease Quality of Life‐36 Survey (KDQOL‐36), with a repeat assessment at 90 days Subgroup analysis: CKD to ICHD, CKD to HHD, CKD to PD, ICHD to HHD, ICHD to PD, PD to ICHD, and PD to HHD Study characteristics Country: Canada Setting: not specified Inclusion criteria: not specified; 33 patients were recruited (7 patients initiating ICHD (5 from CKD / 2 from PD), 13 patients initiating PD (9 from CKD / 4 from ICHD), and 13 patients initiating HHD (2 from CKD / 4 from PD / 7 from ICHD)) Exclusion criteria: not specified
Participants	Baseline characteristics HHD Mean age ± SD: 39.6 ± 19.6 years (7 patients transferring from ICHD to HHD) Overall Mean age ± SD: 54 ± 18 years Male: 75.8% Primary disease Diabetes (18.1%); GN (36.4%)
Interventions	HHD ICHD
Outcomes	Outcomes relevant to this review Symptoms and Problems (KDQOL) Effects of Kidney Disease (KDQOL) Burden of Kidney Disease (KDQOL) Physical Composite Score (KDQOL) Mental Composite Score (KDQOL) Notes: study intended to report the mean change at 90 days from baseline in patients who transitioned from IHD to HHD, however, this abstract included only baseline data Reported outcomes Primary outcome: Mean change in the score of each of the 5 domains of the KDQOL‐36 (“Symptoms/Problems”, “Effects of Kidney Disease”, “Burden of Kidney Disease”, “Physical Composite Score”, and “Mental Composite Score”)
Identification	Additional information Sponsorship source: not reported Authors name: C.S. Dumaine Institution: University of Calgary (now moved to University of Sasketchwan) Other authors: P. Ravani, M. Santana, J. MacRae
Notes	Abstract only. Contacted author to obtain follow‐up data after modality transition, awaiting response

Griva 2010.

Study characteristics
Methods	Study design Cross‐sectional Study grouping: parallel group Study characteristics Country: UK Setting: Two dialysis units affiliated with Royal Free and University College Hospitals (tertiary hospitals) Inclusion criteria: Receiving dialysis in two dialysis units affiliated with Royal Free and University College Hospitals, UK; ≥ 18 years; Maintained on the same dialysis modality for a minimum of 3 months; Fluent in English; Medically stable without acute medical or psychiatric problems Details of eligibility and recruitment procedures previously reported
Participants	Baseline characteristics HHD Mean age ± SD: 51.08 ± 12.11 years Male: 56% Mean education ± SD: 11.76 ± 4.13 years Primary disease Diabetes (0%); GN (20%) Time on KRT ± SD: 163.6 ± 84.26 months Time on dialysis: 88.44 ± 71.2 months Diabetes: 0% ICHD Mean age ± SD: 46.85 ± 16.02 years Male: 57.5% Mean education ±: 12.5 ± 6.34 years Primary disease Diabetes (9.6%); GN (25%) Time on KRT ± SD: 64 ± 60.45 months Time on dialysis ± SD: 38.94 ± 39.64 months Diabetes: 11.5% Group differences Final sample 145 patients (response rate 88.4%): hospital HD (n=52), home conventional HD (n=25) undergoing dialysis thrice weekly, CAPD (n=45), and APD (n=23) patients Significant differences in work status, income, time on KRT, time on dialysis, prevalence of diabetes, albumin Patients on conventional home HD had been on dialysis and renal replacement therapy for longer than other three groups Patients on CAPD had lower albumin levels and higher prevalence of diabetes Employment and income were lowest in CAPD and hospital HD patients
Interventions	HHD ICHD
Outcomes	Outcomes relevant to this review BDI‐II Cognitive Depression Index Identity score (IPQ) IEQ TEQ Reported outcomes ESRD‐SI Kt/V BDI‐II: higher score = depressive symptoms; 16+ = cutoff for significant depression symptoms in ESKD patients Cognitive Depression Index (10+ = cutoff for depression) IPQ: cognitive representation of illness; 5‐point scale; higher scores indicating stronger beliefs that ESKD is chronic, has serious consequences, and is amenable to control or cure IEQ: individuals’ perceptions about how illness interferes with or affects personal, social behaviours, and life in general; 8‐point Likert scale; higher scores signifying a greater level of disruptiveness; < 56 = mild disruptiveness, 56‐89 = moderate disruptiveness, > 89 = moderate to extreme disruptiveness TEQ: patients' perceptions of the physical and psychosocial disruption associated with their treatment rather than their illness; 20 items combined into a single score
Identification	Additional information Sponsorship source: This research was supported by grants from the Alexandros Onassis Foundation to Konstadina Griva and the Reita Lila Weston Institute for Neurological Studies, University College London Authors name: Konstadina Griva Institution: University College London Email: s.newman@ucl.ac.uk Address: Stanton Newman, Unit of Behavioural Medicine, University College London, Charles Bell House, 67‐73 Riding House Street, London W1W 7EJ, United Kingdom Other authors: Andrew Davenport, Michael Harrison, Stanton Newman
Notes

Ha 2018.

Study characteristics
Methods	Study design Cross‐sectional (for the outcomes of interest in this review) as part of a prospective cohort study Study grouping: parallel group Methods: symptom surveys of transplant, PD, ICHD and HHD patients were conducted using Renal Integrated Palliative Care Outcomes Scale (iPOS‐Renal) every 6 months from 2015‐2017 Study characteristics Inclusion criteria: KRT patients at the treating centre Exclusion criteria: not specified
Participants	Baseline characteristics HHD ICHD
Interventions	HHD ICHD
Outcomes	Outcomes relevant to this review iPOS‐Renal
Identification	Additional information Sponsorship source: none stated Country: Australia Setting: tertiary hospital Comments: abstract only, from conference Author name: J Ha Institution: St George Hospital, Sydney Other authors: A Hoffman, MA Brown
Notes	Results Responses: 1174 responses from 273 patients; overall response rate was 51% Symptom burden scores (maximum 68) for PD (16 ± 9) and ICHD patients (14 ± 9) were higher than transplants (9 ± 8) (P < 0.001) HHD (11 ± 7) and transplant patients had comparable symptom scores (P = 0.011) Lack of energy (75%), difficulty sleeping (59%) and poor mobility (55%) were among the commonest symptoms reported Transplant patients had a similar prevalence of depression (39% vs 46%) but less anxiety (42% vs 56%), compared to dialysis patients For transplant patients there were weak but significant (P < 0.01) relationships between symptom scores and a) time on dialysis prior to transplant, and b) eGFR

Hayhurst 2015.

Study characteristics
Methods	Study design Cross‐sectional Study grouping: parallel group Study methods 100 patients were asked to fill out the self‐created questionnaire before their clinic appointment at Royal Preston Hospital in May–June 2014 Two published questionnaires were combined to produce one physical activity questionnaire: ‘General Practice Physical Activity Questionnaire’ (GPPAQ) and ‘Human Activity Profile’ (HAP). Majority of the items and the scoring system taken from the validated HAP questionnaire; elements of the GPPAQ included to highlight the role of physical activity in the workplace (e.g. if the patient’s job involves vigorous work) Scoring produced two scores for each patient: 1. The total number of activities that the patient is able to perform ‐ total activity score (TAS); 2. The total number of activities that the patient has now stopped doing since their diagnosis with CKD (for the CKD not on KRT) or since starting their KRT ‐ activity loss score Maximum activity score (MAS), which is the maximum oxygen‐demanding activity the patient is still able to perform (as used in the HAP) Patient’s notes and the renal IT database (DiProton) Groups: approximately 20 patients from the following groups: CKD stages 3–5 not on any form of KRT, HHD, ICHD, PD and transplant (TX) patients vs. 50 age, sex‐matched controls Study characteristics Country: UK Setting: Royal Preston Hospital, NHS Inclusion criteria: CKD group required patients to be > 18 years, not to be on any form of KRT, between stages 3–5 CKD (eGFR < 60 mL/min) and known to the renal service for > 6 months prior to filling out the questionnaire. ESKD group on KRT included; >18 years, on RRT > 6 months Exclusion criteria: mobilising with frame; participating in a formal exercise program
Participants	Baseline characteristics Mean age ± SD: 60.82 ± 14.10 years Male: 61% Primary disease Diabetes (11%); GN (15%) Group differences Between‐group differences not reported Overall, ages ranged from 18–85 years old (mean age 60.82 ± 14.10); 39 patients were female and 61 were male Patients with CKD had been known to the renal service for an average of 51 ± 37.48 months (maximum 154 months, minimum 10 months) and patients on KRT for an average of 49.54 ± 48.73 months (maximum 192 months, minimum 6 months)
Interventions	HHD ICHD
Outcomes	Outcomes relevant to this review Maximum activity score Total activity score Activity loss score MAP
Identification	Additional information Sponsorship source: Not specified; no competing interests declared Authors name: William S. G. Hayhurst Institution: Royal Preston Hospital, Lancashire Teaching Hospitals NHS Foundation Trust, Preston, UK Email: william.hayhurst@student.manchester.ac.uk Address: Renal Department, Royal Preston Hospital, Lancashire Teaching Hospitals NHS Foundation Trust, Preston, UK Other authors: Aimun Ahmed
Notes	Unable to contact author to obtain measures of variability for outcomes.

Jayanti 2016.

Study characteristics
Methods	Study design Cross‐sectional Study grouping: parallel group Methods Combined cross‐sectional and prospective study design Data part of an ongoing programme of research (BASIC‐HHD) designed to investigate prospectively the facilitators and barriers to HHD (Barriers to successful implementation of care in HHD) within multicentre renal networks in UK Methodological details and scope of data collected in the BASIC‐HHD study presented in the protocol paper (Jayanti 2013) Recovery time (RT) data: cross‐sectional; 313 patients enrolled, 288 responded to the RT question (HHD n=91) Question on RT was posed at the same time as neuropsychometric evaluation of participants, by a member of the research team All other questionnaires were completed by hospital dialysis patients whilst on HD and by HHD patients in their own homes Study characteristics Country: UK Setting: BASIC‐HHD study: multicentre study of home and hospital HD across 5 tertiary centres in the UK Inclusion criteria: knowledge of the English language and visual intactness required to undertake neuropsychometric tests Recovery time data: Prevalent ICHD patients approached if they fulfilled eligibility criteria and were willing to undertake neuropsychometric assessments and complete study‐specific questionnaires; all HHD patients from each participating centre approached Exclusion criteria HHD group: self‐care patients, who were dialysing in‐centre, were excluded
Participants	Baseline characteristics HHD Mean age ± SD: 52.05 ± 11.82 years Male: 75.8% Median duration on dialysis (range): 3.47 years (1.39, 6.82) Mean Charlson comorbidity index: 3.97 (1.76) Diabetes: 12 (13.2%) Heart failure: 3 (3.3%) IHD: 20 (22.0%) Employed: 36 (39.6%) ICHD Mean age ± SD: 56.36 ± 14.44 years Male: 65.0% Median duration on dialysis (range): 2.68 years (1.05, 5.12) Mean Charlson comorbidity index: 4.63 (2.03) Diabetes: 60/195 (30.8%) Heart failure: 11 (5.6%) IHD: 53 (26.9%) Employed: 41/196 (20.9%) Overall Mean age ± SD: 55.00 ± 13.80 years Male: 68.4% Group differences The HHD cohort was significantly younger with a mean age of 52.05 years. These patients were in employment in significantly greater numbers and had a significantly lower Charlson Comorbidity Index (CCI) score and diabetes The HHD patients had a significantly longer dialysis vintage and received significantly longer median dialysis duration per week Compared with ‘in‐centre’ cohort, HHD patients had significantly higher serum albumin and serum bicarbonate levels, and a significantly higher proportion of patients achieved dialysis adequacy of standard Kt/Vurea>2.5 (55.4% vs 18.1%).
Interventions	HHD Duration: variable; based on clinical needs and preferences Frequency: variable; based on clinical needs and preferences; 30.8% received conventional HD 3 times/week Hours/week, median (IQR): 15.00 (12.00, 19.25) ICHD Location: hospital Duration: 4 hours/session Frequency: 3 times/week Hours/week, median (IQR): 12.00 (12.00, 12.00)
Outcomes	Outcomes relevant to this review Recovery time BDI‐II Spielberger Anxiety State‐Trait Inventory for Adults Physical Component Summary Mental Component Summary Outcome measures BDI‐II: 0–10, 11–15, 16–20, 21–25, 26–30, 31+; > 15 = highly predictive of a diagnosis of clinical depression State and Trait Anxiety Inventory: 20–29, 30–39, 40–49, 50+ Modified mini‐mental state (3MS): 94–100:1, 86–93:2, 81–85:3, 76–80:4, ≤ 75:5 Trail making test B scores RT question: ‘How long does it take you typically, to recover from a haemodialysis session?’ (in minutes or hours)
Identification	Additional information Sponsorship source: None stated Authors name: Anuradha Jayanti Institution: Central Manchester Hospitals NHS Trust, Manchester, UK, Email: anu.jayanti@cmft.nhs.uk Address: Dr Anuradha Jayanti, Department of Nephrology, Manchester Royal Infirmary, Oxford Road, Manchester, M13 9WL Other authors: Philip Foden, Julie Morris, Paul Brenchley, Sandip Mitra ‐ BASIC‐HHD study group
Notes	Contacted author to obtain more detailed outcome data, awaiting response. Questionnaires return rate Overall completion rate for the RT question: 94.2% Compiled validated questionnaires return rate ranged from 70–100% for the inventories, across all participating units Collective valid and complete responses averaged: 82%

Kasza 2016.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: parallel group Primary exposure: time‐varying dialysis modality (classified as PD, ICHD or HHD), with facility and home HD sub‐classified by VA type (CVC, AVF or AVG), starting from day 90 of dialysis. Methods ANZDATA analysis Periods of exposure to a particular dialysis modality and VA of < 90 days’ duration were assumed to be because of temporary interruptions to usual or intended modality/VA and were ignored Patients were censored at the time of a switch to home HD CVC Patients were followed until death, loss to follow‐up or 31 December 2011, with censoring at the time of kidney transplant or regain of kidney function Unadjusted pseudo‐survival curves (estimated using an extended Kaplan–Meier estimator) Study characteristics Country: Australia Setting: Australia and New Zealand Dialysis and Transplantation (ANZDATA) Registry analysis Inclusion criteria: all adult incident patients who commenced dialysis between 1 October 2003 and 31 December 2011 and underwent at least 90 days of dialysis Exclusion criteria: undergoing < 90 days of dialysis; patients with missing/extreme BMI at dialysis start; missing SCr at dialysis start; missing VA at day 90
Participants	Baseline characteristics HHD (Only HHD with AVF/AVG reported (n=357)) Mean age ± SD: 50.1 ± 11.2 years Male: 77.3% Primary disease Diabetes (17.9%); GN (39.8%); hypertension (6.4%); other (35.9%) Race NZ Maori/Pacific (7.8%); ATSI (0.6%); Asian (5.6%); White (86.0%) Smoker Current (11.2%); former (35.6%); never (53.2) Late referral: 4.8% BMI 15 to 19.9 (3.1%); 20 to 24.9 (22.1%); 25 to 29.9 (38.4%); 30 to 49.9 (36.4%) Comorbidities CAD (17.9%); type 1 diabetes (2.8%); type 2 diabetes (21.0%); PVD (10.1%); cerebrovascular disease (3.6%) SCr: 644 µmol/L (535, 800) ICHD (AVF/AVG) Mean age ± SD: 62.3 ± 14.1 years Male: 64.6% Primary disease Diabetes (35.6%); GN (21.9%); hypertension (13.6%); other (28.9%) Race NZ Maori/Pacific (7.4%); ATSI (8.6%); Asian (5.8%); White (78.2%) Smoker Current (12.1%); former (42.9%); never (45.0%) Late referral: 7.9% BMI 15 to 19.9 (5.4%); 20 to 24.9 (26.3%); 25 to 29.9 (32.2%); 30 to 49.9 (36.0%) Comorbidities CAD (41.8%); type 1 diabetes (2.6%); type 2 diabetes (43.9%); PVD (25.2%); cerebrovascular disease (14.6%) SCr: 624 µmol/L (485, 806) ICHD (CVC) Mean age ± SD: 61.7 ± 15.7 years Male: 59.9% Primary disease Diabetes (38.5%); GN (20.4%); hypertension (13.0%); other (28.0%) Race NZ Maori/Pacific (12.5%); ATSI (9.4%); Asian (6.5%); White (71.6%) Smoker Current (14.4%); former (41.6%); never (44.1%) Late referral: 39.1% BMI 15 to 19.9 (8.4%); 20 to 24.9 (29.3%); 25 to 29.9 (30.3%); 30 to 49.9 (32.0%) Comorbidities CAD (46.7%); type 1 diabetes (3.0%); type 2 diabetes (47.3%); PVD (29.5%); cerebrovascular disease (17.4%) SCr: 666 µmol/L (500, 896) Group differences HHD (n=383; 357 with AVF/AVG + 26 with CVC) vs. ICHD (n=13143; 5729 with AVF/AVG + 7414 with CVC) vs. PD (n=6665) At day 90, patients on HHD had less comorbidity, were less likely to have had late referral to a nephrologist and were younger on average than patients in other exposure categories ICHD patients with CVCs in use at day 90 had a higher proportion of late referrals to a nephrologist and were more likely to have comorbid conditions than other patients
Interventions	HHD Location: home ICHD (AVF/AVG) Location: facility ICHD (CVC) Location: facility
Outcomes	Outcomes relevant to this review HR for death (compared to ICD with AVF/AVG) Notes: Data shown in figure 3, but very limited data included reporting the actual HR values Unadjusted survival (number at risk)
Identification	Additional information Sponsorship source: None stated Authors name: Jessica Kasza Institution: Monash University, Melbourne, Australia Email: jessica. kasza@monash.edu Address: Dr Jessica Kasza, Department of Epidemiology and Preventive Medicine, Monash University, Melbourne, VIC 3004, Australia Other authors: Rory Wolfe, Stephan P McDonald, Mark R Marshall, Kevan R Polkinghorne
Notes	Contacted author to obtain more detailed data on outcomes by modality, awaiting response.

Kjellstrand 2008.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: parallel Methods: patients on SDHD followed for up to 25 years from 5 centres in USA and Europe Statistical analysis Cox proportional hazard analysis and backward stepwise Cox proportional hazard analysis were used to study the association of several variables with survival Patients were entered in the survival analysis at the date of their first daily HD and censored when they left short‐daily dialysis for any reason: transplantation, return to regular thrice‐weekly HD, or PD Primary exposure: unclear. Baseline characteristics provided for 2 weekly dialysis hours groups. Dialysis site (home vs. in‐center) was also evaluated Study characteristics Country: USA & Italy Setting: Pooled data from 5 USA and European centres Inclusion criteria: Adults with ESKD receiving short daily HD‐ either in‐centre or at home Group differences: In general, the patients were selected for daily ICHD because of more serious disease and poor dialysis tolerance, particularly after the long weekend without dialysis. Daily HHD was offered to patients to improve their quality of life and survival
Participants	Baseline characteristics Intensive HHD Mean age ± SD: 50 ± 14 years Secondary kidney disease: 35% Intensive ICHD Mean age ± SD: 54 ± 16 years Secondary kidney disease: 51% Overall Mean age ± SD: 51 ± 15 years Male: 68% Mean duration on dialysis ± SD: 4.9 ± 5.1 years Mean duration on SDHD ± SD: 2.1 ± 2.1 years Diabetes: 19% Hypertensive nephrosclerosis: 13% Dialysing in‐centre: 28% Primary kidney disease: 54% Secondary kidney disease: 40% Unknown kidney disease: 6%
Interventions	Intensive HHD Duration: 65 to 250 mins Frequency: 5 to 7 times/week Intensive ICHD Duration: 65 to 250 mins Frequency: 5 to 7 times/week
Outcomes	Outcomes relevant to this review HR for mortality (relative to facility HD) Notes: Including all 6 significant variables and era in backward stepwise Cox proportional hazard analysis
Identification	Additional information Sponsorship source: none stated Authors name: Carl Kjellstrand Institution: Loyola University, Chicago, Illinois Email: carl.kjellstrand@gmail.com Address: C. Kjellstrand, MD, PhD, FACP, FRCP(C), 965 SE Creekside Dr College Place, WA 99324, USA Other authors: Umberto Buoncristiani, George Ting, Jules Traeger, Giorgina B. Piccoli, Roula Sibai‐Galland, Bessie A. Young, Christopher R. Blagg
Notes	Contacted author to obtain more detailed outcome data, awaiting response

Kojima 2012.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: single cross‐over Methods: All participants were switched from their conventional HD (3 times for 4 to 5 h/week) to the intensified short HHD regimen (6 times for 3 to 5 h/week). During the following 6 months, quarterly follow‐up visits were conducted Primary exposure: dialysis modality (conventional in‐centre HD vs intensified short‐home HD) Statistical analysis: parameters during the follow‐up (baseline and 6 months) were compared by a nonparametric Friedman’s test for paired variables Study characteristics Country: Japan Setting: Kidney Disease Center, Saitama Medical University Inclusion criteria: patients starting HHD at the Kidney Disease Center, Saitama Medical University
Participants	Baseline characteristics Mean age ± SD: 49.2 ± 10 years Mean duration on dialysis ± SD: 4.0 ± 5.3 years Primary disease Diabetes (20%); GN (43%); nephrosclerosis (11%); others (4%)
Interventions	Intensive HHD Location: home (using Nikkiso DBB‐ 27 (Nikkiso Co., Tokyo, Japan) with a water treatment system MH‐ 500CX (Japan Water System Co., Tokyo, Japan) Training: at least 3 months Duration: 3 to 5 hours Frequency: 6 times/week ICHD Duration: 4 to 5 hours Frequency: 3 times/week
Outcomes	Outcomes relevant to this review Predialysis systolic BP (mm Hg) Left ventricular mass index (g/m2) Study outcomes Reported outcomes Primary outcome: Multiple surrogate parameters of known dialysis‐associated morbidities were assessed: Dialysis efficacy (creatinine, dry weight) BP (pre‐ and postdialytic systolic and diastolic blood pressures, antihypertensive drugs) Nutrition and inflammation (albumin, BMI) Bone metabolism (calcium, phosphate, iPTH, phosphate‐binding pharmacotherapy) Erythropoiesis (Hb, need of ESA, need for iron substitution therapy). ECG of the patients were recorded in the week before the start of treatment and at 6 months after the start of HHD (end diastole diameter of the left ventricular chamber, interventricular septum thickness, thickness of the left ventricular posterior wall, left ventricular mass index)
Identification	Additional information Sponsorship source: None mentioned Authors name: Eriko Kojima Institution: Kidney Disease Center, Saitama Medical University, Saitama, Japan Email: iromichi@saitama‐med.ac.jp Address: Hiromichi Suzuki, MD, PhD, Department of Nephrology, Saitama Medical University 38 Moroyamamachi, Irumagun, Saitama 350‐ 0495 (Japan) Other authors: Hitoshi Hoshi, Yusuke Watanabe, Tsuneo Takenaka, Hiromichi Suzuki
Notes	Unable to contact author to obtain measure of variability and more detailed outcome data.

Krahn 2019.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: parallel Study characteristics Country: Canada Setting: Ontario, Canada from Canadian Organ Replacement Register (CORR) ‐‐ Registry analysis Inclusion criteria: patients who initiated chronic dialysis at ages 18 to 105 years between 1 April 2006 and 31 March 2014 from the Canadian Organ Replacement Register (CORR) Exclusion criteria: evidence of previous dialysis (N = 2101); not Ontario residents or had missing age, invalid sex, or invalid provincial health insurance number (N = 102); died within 30 days after dialysis initiation (N = 346)
Participants	Baseline characteristics HHD Mean age ± SD: 51.1 ± 15.0 years Male: 73.2% Comorbidity (Aggregated Diagnostic Groups count): 9.2 (3.9) Rural residence: < 1% Diabetes within 2 years prior to dialysis initiation: 41.1% Kidney transplantation: 37.5% Death: 16.1% Intensive ICHD Mean age ± SD: 63.9 ± 16.8 years Male: 69.2% Comorbidity (Aggregated Diagnostic Groups count): 11.9 (4.1) Rural residence: < 1% Diabetes within 2 years prior to dialysis initiation: 64.6% Kidney transplantation: < 8% Death: < 48% ICHD Mean age ± SD: 66.8 ± 14.7 years Male: 60.6% Comorbidity (Aggregated Diagnostic Groups count): 11.1 (3.9) Rural residence: 11% Diabetes within 2 years prior to dialysis initiation: 56.0% Kidney transplantation: 8.7% Death: 51.1% Overall Mean age ± SD: 65.8 ± 14.9 years Male: 58.0% Comorbidity (Aggregated Diagnostic Groups count): 10.8 (3.8) Rural residence: 11.8% Diabetes within 2 years prior to dialysis initiation: 55.0% Kidney transplantation: 11.4% Death: 48.1%
Interventions	HHD Conventional or SD/SN ICHD Conventional HD SD/SN
Outcomes	Outcomes relevant to this review Unadjusted mean 30‐day costs by initial modality 30 day costs by as‐treated modality Unadjusted cumulative costs Adjusted cumulative costs Notes: Adjusted for age, sex, comorbidity Unadjusted survival (number at risk) Notes: censoring at kidney transplantation or end of follow‐up, but including modality switches. Definitions Initial modality Facility conventional HD: referred to as ‘facility HD’ (n=9687) Facility short daily HD (SD, performed 6 to 7 times per week, for 2 to 3 hours during awake time), or slow nocturnal HD (SN, performed 5 to 7 times per week, for 6 to 9 hours while sleeping), referred to as ‘SD/SN HD’ (n=65) Home HD, including conventional and SD/SN HD (n=112) PD (both CAPD (57%) and APD (43%) (n=2827) Costs Analysis from the perspective of the public payer, the Ontario Ministry of Health and Long‐Term Care, which insures all permanent residents of Ontario for medically‐necessary care under the publicly‐funded Ontario Health Insurance Plan (OHIP) Included costs paid by chronic kidney disease programs in Ontario, which receive funding from and/or have management agreements with the Ontario Renal Network through the publicly‐funded agency Cancer Care Ontario Unable to include costs for patients’ travel to dialysis clinics for dialysis or other medical care, because not readily available in the administrative data.
Identification	Additional information Sponsorship source: Funded by the Ontario Renal Network, Ontario, Canada. Dr. Krahn was supported by the F. Norman Hughes Chair in Pharmacoeconomics at the University of Toronto, and a Tier 1 Canada Research Chair in Health Technology Assessment. Dr. Garg was supported by the Dr. Adam Linton Chair in Kidney Health Analytics. Infrastructure support was provided by a Canada Foundation for Innovation Grant (THETA), and the Toronto General Hospital Research Institute. Some analyses were conducted at the Institute for Clinical Evaluative Sciences (ICES) Western site. The ICES is funded by an annual grant from the Ontario Ministry of Health and Long‐Term Care (MOHLTC). ICES Western is also supported by the Academic Medical Organization of Southwestern Ontario (AMOSO), the Schulich School of Medicine and Dentistry (SSMD), Western University, and the Lawson Health Research Institute (LHRI). Comments: ***parts of this paper were presented as a poster at the ISPOR Asia Pacific conference in Tokyo, Japan, September 8–11, 2018. The abstract is published in Value in Health 2018; 21(Suppl 2):S115.*** ‐‐‐ Krahn 2018 Authors name: Murray D. Krahn Institution: University Health Network, Toronto, ON, Canada Email: kbremner@uhnresearch.ca Address: Karen Bremner, Toronto General Hospital, 200 Elizabeth Street, Room EN10‐231, Toronto, ON M5G 2C4, Canada Other authors: Karen E. Bremner, Claire de Oliveira, Stephanie N. Dixon, Phil McFarlane, Amit X. Garg, Nicholas Mitsakakis, Peter G. Blake, Rebecca Harvey, and Petros Pechlivanoglou
Notes

Kraus 2007.

Study characteristics
Methods	Study design Prospective cohort study Study grouping: single cross‐over Study characteristics Country: USA Setting: multicentre (6 sites) Inclusion criteria: received a minimum of 2 weeks of HD with the System One in a centre environment before study initiation; ESKD, a life expectancy of at least 1 year, currently receiving HD at least 3 times/week for a minimum of 3 months; multispecialty team determined participants' appropriateness for HHD (defined by each centre); Each patient had to have an identifiable partner; ≥ 18 years; VA capable of a minimum blood flow rate of 350 mL/min and could be treated with a delivered spKt/V of 0.45 in ≤ 3.5 hours Exclusion criteria: eGFR > 6 mL/min/1.7 3m2 as estimated by 24‐hour urine in the presence of > 400 cm3 urine in 24 hours; liver disease; uncontrolled hypertension; symptomatic intradialytic hypotension; Hb < 10 g/dL; active infectious or inflammatory disease; documented noncompliance; malignancy other than superficial skin carcinomas
Participants	Baseline characteristics Overall Mean age: 51 years Male: 63.0% Race Black (19%%); White (75%); American Indian or Alaskan Native (3%); Asian (3%) Primary disease Diabetes (16%); GN (25%); hypertension (13%); PKD (13%); other (34%) Smoking Current (6%); previous (34%) Alcohol Current (22%); previous (16%) Anuric: 47% Group differences Same group (n=32) 31 were on thrice weekly dialysis with conventional dialysis, 5 at home, and 26 in‐centre One patient was on short‐daily dialysis at home with conventional dialysis equipment
Interventions	Intensive HHD Location: home Operator: patient and/or partner Training: during the in‐centre phase, qualified medical personnel, home dialysis nurses, and/or technicians trained each subject and the subject’s partner how to perform dialysis with the System One. Training consisted of individual sessions involving all phases of dialysis including setup, treatment, posttreatment care, machine alarms, and disaster preparedness. Subjects and their partners were required to successfully complete the System One Skills Checklist and achieve a score of 100% on a written exam before the first treatment of the Home Phase Duration: 2.3 to 3.2 hours Frequency: 6 times/week Intensive ICHD Location: in‐centre Operator: staff Training: patients were training for the home phase during the in‐centre phase Duration: 2.3 to 3.2 hours Frequency: 6 times/week
Outcomes	Outcomes relevant to this review Systolic BP Diastolic BP Mean pulse pressure
Identification	Additional information Sponsorship source: Supported by NxStage Medical Inc. Dr Kraus (principal investigator) was paid consulting fees by NxStage Medical for his role as medical monitor of this trial Authors name: Michael Kraus Institution: Indiana University Hospital, Indianapolis, Indiana, USA Email: mkraus@iupui.edu Address: M. Kraus, MD, Indiana University Hospital, 550 N University Boulevard, UH1115, Indianapolis, IN 46202, U.S.A. Other authors: John Burkart, Rebecca Hegeman, Richard Solomon, Norman Coplon, John Moran
Notes	Contacted author to obtain more detailed outcome data, awaiting response

Krishnasamy 2013.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: parallel Statistical analysis: main analysis was HD vs PD, but sensitivity analysis separating HHD was also performed Study characteristics Country: Australia and New Zealand Setting: Australia and New Zealand Dialysis and Transplantation (ANZDATA) ‐‐ Registry analysis Inclusion criteria: adult patients with ESKD receiving maintenance dialysis in Australia and New Zealand who died between January 1, 1999, and December 31, 2008 (modality at time of death: HD, n=10,338 (IHD=9765; HHD=573); PD, n=4298)
Participants	Baseline characteristics Overall Mean age ± SD: 67.4 ± 13.1 years Male: 59% Race White (78%); ATSI (9%); Maori and Pacific Islander (9%); Asian (3%); Other (2%) BMI Underweight (< 20): 12%; normal (20 to 24.9): 34%; overweight (25 to 29.9): 31%; obese (≥ 30): 23% Late referral: 26% Smoking Current (14%); former (43%); never (43%) Comorbidities Chronic lung disease (30%); coronary artery disease (74%); peripheral vascular disease (51%); cerebrovascular disease (35%); diabetes (43%) Primary disease Chronic GN (22%); diabetes (30%); renovascular (16%); PKD (4%); reflux (2%); miscellaneous (18%); unknown (7%) Country Australia (86%); New Zealand (14%) Centre size Small (< 360): 2%; small‐medium (360 to 699): 10%; medium‐large (700 to 839): 26%; large (≥ 840): 62% Group differences Patient characteristics not specified for in‐center HD vs. HHD (reported separately for HD and PD) Hours/session and frequency of dialysis sessions for in‐centre and home‐based HD differed
Interventions	HHD Location: home Duration: < 4 hours (6%); 4ours (22%); > 4 hours (72%) Frequency: > 3 times/week (19%); ≤ 3 times/week (81%) ICHD Location: facility Duration: < 4 hours (13%); 4 hours (55%); > 4 hours (33%) Frequency: ≤ 3 times/week (97%) Intensive ICHD Duration: < 4 hours (13%); 4 hours (55%); > 4 hours (33%) Frequency: > 3 times/week (3%)
Outcomes	Outcomes relevant to this review Day of week as predictor of cardiac death (Adj OR) Notes: odds of cardiac death for a given day of the week were compared with the odds of cardiac death for all days of the week Reported outcomes Primary outcome: Cardiac death after starting dialysis therapy, examined according to its timing (day of the week) Other outcomes: Timing of non‐cardiac deaths, including death due to vascular causes (cerebrovascular accident, aortic aneurysm rupture, bowel infarction, pulmonary embolus, gastrointestinal haemorrhage, dialysis access‐site haemorrhage, and other haemorrhage), infection, dialysis therapy withdrawal, accident, suicide, or miscellaneous causes Definitions Cause of cardiac death was reported to the registry by the patient’s attending nephrologist according to the following categories: myocardial ischemia (presumed), myocardial ischemia/infarction, pulmonary oedema, hyperkalemia, haemorrhagic pericarditis, hypertensive cardiac failure, other cause of cardiac failure, and cardiac arrest (cause uncertain; whether in or out of hospital). Deaths were reported as myocardial ischemia or infarction if there was direct evidence, such as dynamic electrocardiogram and cardiac enzyme level changes or autopsy report Deaths were classified as presumed myocardial ischemia if there was a strong clinical suspicion in the absence of such direct evidence. In this cohort, cardiac arrest (cause uncertain) and myocardial ischemia (presumed) were the codes for sudden death
Identification	Additional information Sponsorship source: No financial support for the study Comments: Dr Johnson is a consultant for Baxter Healthcare Pty Ltd and previously has received research funds from this company. He has also received speakers’ honoraria and research grants from Fresenius Medical Care and is a current recipient of a Queensland Government Health Research Fellowship. Dr Bannister is a consultant for Baxter Healthcare Pty Ltd. Dr Brown is a consultant for Baxter and Fresenius and has received travel grants from Amgen and Roche. Dr Polkinghorne has received speaking honoraria and travel grants from Amgen Australia. Assoc Prof Hawley has received research funding from Baxter Healthcare Pty Ltd and Fresenius Medical Care. She also has received travel grants from Amgen Australia. Dr McDonald is a consultant for Amgen Australia and Shire Australia. Assoc Prof Boudville previously has received research funds from Roche; travel grants from Roche, Amgen, and Jansen Cilag; and speaking honoraria from Roche. The remaining authors declare that they have no relevant financial interests. Authors name: Rathika Krishnasamy Institution: University of Queensland at Princess Alexandra Hospital, Brisbane Email: david_johnson2@health.qld.gov.au Address: David W. Johnson, Department of Nephrology, Level 2, ARTS Bldg, Princess Alexandra Hospital, Ipswich Rd, Woolloongabba, Brisbane Qld 4102, Australia Other authors: Sunil V. Badve, Carmel M. Hawley, Stephen P. McDonald, Neil Boudville, Fiona G. Brown, Kevan R. Polkinghorne, Kym M. Bannister, Kathryn J. Wiggins, Philip Clayton, and David W. Johnson
Notes	Data not extracted for meta‐analysis due to potential overlapping population with Marshall 2021.

Lee 2002.

Study characteristics
Methods	Study design Prospective cohort study Study grouping: parallel Data collection Patients followed up prospectively for 1 year Detailed information on baseline patient characteristics, including comorbidity, was collected Demographic information was assessed by means of a self‐administered questionnaire Presence of comorbid illness assessed by a trained research nurse as of the enrollment date by direct patient interview and complete review of patients’ inpatient and outpatient records Study characteristics Country: Canada Setting: Foothills Medical Centre, Calgary, Alberta, Canada Inclusion criteria: all patients treated locally with dialysis in July 1999 and who had been on dialysis longer than 6 months; approached for enrolment in a consecutive fashion from a randomly generated list of patients treated with each modality Exclusion criteria: inability to speak English or inability to be contacted after repeated attempts (n=110); did not wish to participate (n=56) Definitions: 332 patients eligible for the study, 166 enrolled (88 in‐centre HD; 31 satellite HD; 9 home/self‐care HD; 38 PD)
Participants	Baseline characteristics HHD Mean age: 55.6 years Male: 44.4% White: 88.9% Working: 33.3% Income: < $10,000 (25.0%); $10,000 to $30,000 (37.5%); > $30,000 (37.5%) Living arrangements Independently (66.7%); requiring part‐time home care (33.3%); requiring 24‐hour/long‐term care (0.0%) Education Less than grade 12 (33.3%); to grade 12/equivalent (44.4%); to post‐secondary (22.2%) Median months on dialysis (range): 54.7 months (11.4 to 120.8) Charlson comorbidity index Low risk (≤ 3): 44.4%; medium risk (4 to 5): 55.6%; high risk (≥ 6): 0.0% Mean Charlson comorbidity index: 3.4 (2.5 to 4.4) Comorbidities Coronary heart disease (11.1%); congestive heart failure (11.1%); PVD (22.2%); previous stroke (11.1%); lung disease (11.1%); diabetes (33.3%) Mean baseline Hb (range): 12.1 g/dL (11.3 to 12.9) Mean baseline albumin (range): 3.47 g/dL (3.26 to 3.67) Mean baseline Kt/V: 1.7 (1.5 to 1.9) Satellite HD Mean age: 63.7 years Male: 61.3% White: 71.0% Working: 12.9% Income: < $10,000 (9.5%); $10,000 to $30,000 (47.6%); > $30,000 (42.9%) Living arrangements Independently (65.5%); requiring part‐time home care (27.6%); requiring 24‐hour/long‐term care (6.8%) Education Less than grade 12 (24.1%); to grade 12/equivalent (37.9%); to post‐secondary (37.9%) Median months on dialysis (IQR): 28.8 months (23.6 to 34.1) Charlson comorbidity index Low risk (≤ 3): 51.6%; medium risk (4 to 5): 25.8%; high risk (≥ 6): 22.6% Mean Charlson comorbidity index: 3.8 (3.2 to 4.5) Comorbidities Coronary heart disease (32.3%); congestive heart failure (6.5%); PVD (19.4%); previous stroke (32.3%); lung disease (25.8%); diabetes (19.4%) Mean baseline Hb (range): 11.7 g/dL (11.3 to 12.0) Mean baseline albumin (range): 3.47 g/dL (3.36 to 3.57) Mean baseline Kt/V: 1.6 (1.5 to 1.6) ICHD Mean age: 61.8 years Male: 55.7% White: 76.1% Working: 11.4% Income: < $10,000 (20.9%); $10,000 to $30,000 (38.8%); > $30,000 (40.3%) Living arrangements Independently (61.5%); requiring part‐time home care (28.9%); requiring 24‐hour/long‐term care (9.6%) Education Less than grade 12 (17.6%); to grade 12/equivalent (40.0%); to post‐secondary (42.4%) Median months on dialysis (IQR): 41.2 months (31.9 to 50.5) Charlson comorbidity index Low risk (≤ 3): 38.6%; medium risk (4 to 5): 36.4%; high risk (≥ 6): 25.0% Mean Charlson comorbidity index: 4.3 (3.9 to 4.7) Comorbidities Coronary heart disease (34.1%); congestive heart failure (25.0%); PVD (25.0%); previous stroke (12.5%); lung disease (27.3%); diabetes (27.3%) Mean baseline Hb (range): 11.8 g/dL (11.6 to 12.1) Mean baseline albumin (range): 3.44 g/dL (3.36 to 3.51) Mean baseline Kt/V: 1.5 (1.5 to 1.6) PD Mean age: 57.7 years Male: 50.0% White: 71.1% Working: 29.0% Income: < $10,000 (23.5%); $10,000 to $30,000 (23.5%); > $30,000 (52.9%) Living arrangements Independently (58.3%); requiring part‐time home care (36.1%); requiring 24‐hour/long‐term care (5.6%) Education Less than grade 12 (30.6%); to grade 12/equivalent (41.6%); to post‐secondary (27.8%) Median months on dialysis (IQR): 36.2 months (22.2 to 50.1) Charlson comorbidity index Low risk (≤ 3): 36.8%; medium risk (4 to 5): 39.5%; high risk (≥ 6): 23.7% Mean Charlson comorbidity index: 4.3 (3.7 to 5.0) Comorbidities Coronary heart disease (39.5%); congestive heart failure (21.1%); PVD (31.6%); previous stroke (18.4%); lung disease (26.3%); diabetes (44.7%) Mean baseline Hb (range): 11.7 g/dL (11.1 to 12.4) Mean baseline albumin (range): 3.04 g/dL (2.88 to 3.19) Mean baseline Kt/V: 2.5 (3.7 to 5.0) Overall Mean age: 600.9 years Male: 54.8% White: 74.7% Working: 16.9% Income: < $10,000 (20.0%); $10,000 to $30,000 (36.2%); > $30,000 (43.9%) Living arrangements Independently (61.8%); requiring part‐time home care (30.6%); requiring 24‐hour/long‐term care (7.6%) Education Less than grade 12 (22.6%); to grade 12/equivalent (40.3%); to post‐secondary (37.1%) Median months on dialysis (IQR): 38.5 months (31.9 to 45.0) Charlson comorbidity index Low risk (≤ 3): 41.0%; medium risk (4 to 5): 36.1%; high risk (≥ 6): 22.9% Mean Charlson comorbidity index: 4.2 (3.9 to 4.5) Comorbidities Coronary heart disease (33.7%); congestive heart failure (19.9%); PVD (25.3%); previous stroke (17.5%); lung disease (25.9%); diabetes (30.1%) Mean baseline Hb (range): 11.8 g/dL (11.6 to 12.0) Mean baseline albumin (range): 3.35 g/dL (3.29 to 3.42) Mean baseline Kt/V: 1.8 (1.7 to 1.9) Group differences Non‐enrolled patients were similar in sex (55.6% men) but tended to be older (mean age, 63.8 years) and less likely to be on PD therapy (16.4% of non‐enrolled patients were on PD therapy) Patients treated with in‐centre and satellite HD were older than those treated with self‐care dialysis (home/self‐care HD and PD) Patients treated with self‐care dialysis were more likely to be working than those treated with in‐centre or satellite HD (P=0.08, chi‐square) Patients on home/self‐care HD had less comorbid disease than patients treated with the other modalities
Interventions	HHD Operator: patient, independent of nurses Training: training period of 6 weeks while on dialysis Duration: ≥ 4 hours Frequency: ≥ 3 times/week Satellite HD Operator: independently of nurses Training: training period of 6 weeks while on dialysis Duration: ≥ 4 hours Frequency: ≥ 3 times/week ICHD Duration: at least 4 hours Frequency: 3 times/week PD Training: 1 week Duration: based on total Kt/V target ≥ 2.1
Outcomes	Outcomes relevant to this review Annual total health‐care related costs per patient: $ Annual outpatient dialysis costs: $ Annual inpatient costs: $ Total outpatient costs: $ Annual physician billing costs: $ Outcome measures Costs considered included those related to outpatient dialysis care, inpatient care, outpatient non‐dialysis care, and physician claims From the perspective of the health care purchaser; included only direct health care–related costs Societal costs (i.e. time or patient transport costs) were excluded Costs were measured in Canadian dollars but reported in 2000 US dollars (US $1=CAN $1.45=British 0.68) Costs organised into 4 main categories: outpatient dialysis expenses, inpatient expenses, outpatient non‐dialysis expenses, and physician fees
Identification	Additional information Sponsorship source: Supported in part by The Alberta Heritage Foundation for Medical Research and The Center for Advancement of Health, Calgary Regional Health Authority, Calgary, Alberta, Canada Authors name: Helen Lee Institution: University of Calgary, Alberta, Canada Email: braden.manns@calgaryhealthregion.ca Address: Braden Manns, MD, Foothills Medical Centre, 1403 29th St NW, Calgary, Alberta, Canada, T2N 2T9 Other authors: Braden Manns, Ken Taub, William A. Ghali, Stafford Dean, David Johnson and Cam Donaldson
Notes	Incomplete data 16 patients died; 9 patients received a renal transplant; 4 patients switched dialysis modality (1 patient changed from ICHD to PD, and 3 patients switched from PD to ICHD); 13 patients were enrolled after the start of the fixed costing period

Lorenzen 2012.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: Single cross‐over Retrospective, longitudinal, single‐centre study at Kuratorium fur Dialyse und Nierentransplantation, a non‐profit dialysis provider associated with the Hannover Medical School Eleven patients on maintenance dialysis transferring from standard HD to short daily HHD. Before starting HHD, patients were instructed during a 3‐month training phase. All participants were switched from their conventional HD to the intensified short daily HHD regimen During the following 12 months, quarterly follow‐up visits with evaluation of parameters Study characteristics Country: Germany Setting: single‐centre study at Kuratorium fur Dialyse und Nierentransplantation Inclusion criteria: dialysed for at least 1 year prior to study (3 x 4 hours), using a native AVF, with a standard dialysis regimen of 4 hours, 3 times/week; only patients who satisfactorily completed the home dialysis training phase were included in the study Exclusion criteria: secondary hyperparathyroidism with a need of a calcium‐sensitising therapy; CVC; need of erythrocyte transfusion; severe arterial hypertension; severe comorbidities; compliance problems
Participants	Baseline characteristics Mean age: 47.9 ± 15.5 years Male: 73% Mean duration on dialysis: 23.3 ± 10.7 months
Interventions	Intensive HHD Training: 3‐month training phase; focused on the use of the dialysis machine, troubleshooting alarms, adjusting treatment parameters (including dialysate composition), dealing with emergencies (air embolism, accidental disconnection from the blood circuit), and proper vascular access cannulation/connection technique Duration: 3 hours Frequency: 6 times/week ICHD Duration: 4 hours Frequency: 3 times/week
Outcomes	Outcomes relevant to this review Pre‐dialysis MAP (mm Hg) Post‐dialysis MAP (mm Hg) Reported outcomes Multiple surrogate parameters of known dialysis‐associated morbidities were assessed Dialysis efficacy: creatinine, cumulative blood volume BP: pre‐and post‐dialytic MAP anti‐hypertensive drug score, serum and dialysate sodium levels Erythropoiesis: Hb, transferrin saturation, ferritin, need of ESA/kg body weight, need for iron substitution therapy Nutrition and inflammation: total protein, albumin, CRP, ‘dry’ body weight, body mass index Bone metabolism: Ca x Phos product, 25‐(OH)vitamin D, alkaline phosphatase, phosphate‐binding pharmacotherapy iPTH was measured using the Elecsys assay (Roche Diagnostics, Mannheim, Germany) Kt/V Antihypertensive drug index (10 x [daily dose/max. recommended dose]) Antihypertensive drug score: sum of the drug index/day
Identification	Additional information Sponsorship source: Not specified Authors name: Johan M. Lorenzen Institution: Hanover Medical School, Hanover, Germany Email: J.M.Lorenzen@gmail.com Address: Department of Medicine/Division of Nephrology and Hypertension, Hanover Medical School, Carl‐Neuberg‐Strasse 1, 30625 Hanover, Germany Other authors: Thomas Thum, Georg M. Eisenbach, Hermann Haller, Jan T. Kielstein
Notes

Malmstrom 2008.

Study characteristics
Methods	Study design Cross‐sectional (QoL outcomes); retrospective cohort study (other outcomes) Study grouping: parallel group Study characteristics Country: Finland Setting: single centre (Helsinki University Hospital) Definitions: The main difference between HHD and Satellite HD is the location where dialysis treatment takes place (i.e. home versus satellite) and the fact that in satellite HD, the patients are assisted by nurses Inclusion criteria: On 15 October 2004, a total of 65 patients attended self‐care HD in the area and were included in the study HHD (n=33) vs self‐care HD at satellite unit (n=32) The cost data were collected from those study patients who were on dialysis the whole calendar year of 2004 (23/33 HHD and 28/32 satellite HD patients) The HRQoL questionnaire was handed out to the patients on 15 October 2004, and 23/33 of HHD and 24/32 of satellite HD patients returned the questionnaire
Participants	Baseline characteristics HHD Mean age ± SD: 49.0 ± 9.9 years Male: 76% Primary disease PKD (9%); kidney disease (18%); diabetes (15%); other (15%); unspecified (9%) Comorbidities Coronary heart disease (12%); systemic amyloidosis (3%); diabetes (36%) Duration on HHD: 23.3 ± 21.9 months Satellite HD Mean age ± SD: 63.2 ± 14.0 years Male: 66% Primary disease PKD (16%); kidney disease (19%); diabetes (16%); other (16%); unspecified (19%) Comorbidities Coronary heart disease (22%); systemic amyloidosis (3%); diabetes (22%) Duration on HHD, months: not applicable Group differences Trained patients are then referred either to HHD or to one of two self‐care satellite HD units according to patient preference. HHD (n=33) vs. Self‐care HD (n=32) Travel and medication data available only for whole calendar years ‐ cost data collected from those study patients who were on dialysis the whole calendar year of 2004 (23/33 HHD and 28/32 satellite HD patients) HRQoL questionnaire handed out to the patients on 15 October 2004 23/33 of HHD and 24/32 of SHD patients returned the questionnaire In HHD the mean age of the responders and non‐responders did not differ [(50.2 and 46.5 years, P=not significant (ns)], whereas in satellite HD the responders were significantly older (67.0 years) than non‐responders (51.6 years), P < 0.05.
Interventions	HHD Location: home Operator: patient (or occasionally, an assistant) Training: trained to perform the dialysis alone; mean training time 4 to 6 weeks; only exceptionally, an assistant (spouse, another family member or a friend) is trained for the patient Duration: flexible; 16.2 ± 3.8 hours/week Frequency: Flexible (3.26 ± 0.44) Satellite HD Location: satellite unit Operator: assisted by nurses Training: Not specified Duration: 13.9 ± 1.1 hours/week Frequency: 3 times/week
Outcomes	Outcomes relevant to this review Predialysis systolic BP (mm Hg) Predialysis diastolic BP (mm Hg Post‐dialysis systolic BP (mm Hg) Post‐dialysis diastolic BP (mm Hg) Annual hospital costs (EUR) Annual total health‐care related costs/patient ($) 15D Reported outcomes Dialysis‐related adverse events Costs Direct medical costs of dialysis and hospital treatment Travel costs Costs of pharmaceuticals Total costs HRQoL Total 15D score 15 dimensions of the 15D instrument
Identification	Additional information Sponsorship source: This work was financially supported by the Finnish Office for Health Technology Assessment and a special governmental subsidy for health sciences research. Authors name: Raija K Malmström Institution: Porvoo Hospital, Helsinki and Uusimaa Hospital Group, Porvoo Email: raija.malmstrom@hus.fi Address: Raija K Malmstrom Sairaalatie 1, 06150, Porvoo, Finland Other authors: Risto P Roine, Anne Heikkila, Pirjo Rasanen, Harri Sintonen, RiittaMuroma‐Karttunen and Eero Honkanen
Notes	Unable to contact author to obtain measures of variability

Marshall 2021.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: Parallel Primary exposure: dialysis modality: facility HD, CAPD, APD, or HHD Primary outcome: death Statistical analysis Primary exposure of dialysis modality was defined 2 ways: Modelled as time‐varying, referring to a patient’s time‐updated treatment modality over the entire study period ‐‐> consistent with an as‐treated framework (“Did the exposure that the patient actually received affect mortality?”). Modelled as fixed from 90 days, referring to a patient’s initial treatment modality at baseline ‐‐> consistent with an intention‐to‐treat approach (“did exposure that the patient initially received affect mortality, irrespective of subsequent changes that occurred along the way?”) ‐ Modeling approaches: Cause‐specific proportional hazards models, censoring for kidney transplantation, return of kidney function, and loss to follow‐up Subdistribution proportional hazards (Fine and Gray) models, treating transplantation as a competing risk. For all comparative analyses, facility HD was the reference category Both modelling approaches have the effect of excluding patients who were not alive on dialysis at 90 days Stratified analyses by era, defined by the year of dialysis inception. The era was arbitrarily defined in 5‐year windows (i.e., 1998‐2002, 2003‐2007, 2008‐2012, 2013‐2017) Covariables: adjusted estimates of the effect of modality on mortality using available patient‐related risk factors: age, sex, ethnicity, primary kidney disease, eGFR at dialysis inception, late referral for nephrology predialysis care (< 3 months before dialysis inception), diabetes (none, type 1, type 2), BMI, medical comorbidity (coronary artery disease, PVD, cerebrovascular disease, chronic lung disease), and smoking Definitions Facility HD: dialysis at a staffed dialysis HD facility CAPD, APD, and HHD: dialysis in an unstaffed setting of a domiciliary or communal nature Study characteristics Country: Australia and New Zealand Setting: Australia and New Zealand Dialysis and Transplantation (ANZDATA) ‐‐ Registry analysis Inclusion criteria: all adult patients in ANZDATA who commenced dialysis in the 20 years to December 31, 2017
Participants	Baseline characteristics ICHD Median age (range): 63 years (51 to 72) Male: 61.4% Ethnicity ANZ European (69.9%); ATSI (10.3%); Asian (6.6%); NZ Maori (5.5%); Pacific People (5.3%); other (2.5%) Late referral: 24.2% Smoking: 13.8% Diabetes: type 1 (3.6%); type 2 (45.3%) Primary disease Diabetes (37.6%); GN (21.2%); hypertension/ischaemic (14.0%): PKD (5.5%); reflux (2.2%); other (19.7%) Comorbidities Coronary artery disease (41.5%); PVD (25.8%); cerebrovascular disease (14.9%); lung disease (17.6%) Median BMI (range): 27.5 (23.6 to 32.5) BMI ≥ 30: 35.6% Median eGFR: 6.5 (4.7 to 9.2) HHD Median age (range): 50 years (42 to 59) Male: 75.3% Ethnicity ANZ European (75.7%); ATSI (3.4%); Asian (7.2%); NZ Maori (5.9%); Pacific People (5.8%); other (3.1%) Late referral: 4.9% Smoking: 11.7% Diabetes: type 1 (3.7%); type 2 (25.7%) Primary disease Diabetes (21.9%); GN (35.2%); hypertension/ischaemic (6.8%): PKD (18.2%); reflux (4.8%); other (13.1%) Comorbidities Coronary artery disease (18.5%); PVD (10.1%); cerebrovascular disease (5.0%); lung disease (8.8%) Median BMI (range): 28.4(24.6 to 34.4) BMI ≥ 30: 41.6% Median eGFR: 6.1 (4.6 to 8.1)
Interventions	ICHD Location: staffed dialysis HD facility HHD Location: unstaffed setting of a domiciliary or communal nature
Outcomes	Outcomes relevant to this review Proportion cardiovascular death as cause of death HR for mortality (relative to facility HD), as‐treated HR for mortality (relative to facility HD), ITT
Identification	Additional information Sponsorship source: This work was supported in part by the Jacquot Family and unrestricted educational grants made through the Research Foundation of Royal Australasian College of Physicians. ANZDATA receives funding from the Australian Government Department of Health and Ageing, the New Zealand Ministry of Health, and Kidney Health Australia. General support for registry activities has been received from AMGEN Australia Pty Ltd, Novartis Pharmaceuticals Australia Pty Ltd, Janssen‐Cilag Pty Ltd, Fresenius Medical Care‐Australia Pty Ltd, Roche Products (Australia) Ltd, and Wyeth Australia Pty Ltd. The funders had no role in the study design, data collection, analysis, reporting, or decision to submit the manuscript for publication Authors name: Mark R. Marshall Institution: University of Auckland, Auckland, New Zealand Email: markrogermarshall@icloud.com Address: Mark R. Marshall, PO Box 29, Helensville 0840, New Zealand Other authors: Kevan R. Polkinghorne, Neil Boudville and Stephen P. McDonald
Notes	Data from Marshall 2021 publication analysed. Data from other publications not analysed due to potentially overlapping populations.

McGregor 2001.

Study characteristics
Methods	Study design Open‐label, prospective, cross‐over RCT Study duration: not reported Follow‐up period: 2 treatment phases, each of 8 weeks in random order Study characteristics Country: New Zealand Setting: single centre Inclusion criteria: patients on HHD for more than 6 months, to be dialysing > 6 hours, 3 times/week, to be on no antihypertensive drugs and to have a mean predialysis BP over the previous month of < 160/90 mm Hg Exclusion criteria: patients with diabetes mellitus, overt cardiac disease, prior nephrectomy, or any recent illness
Participants	Baseline characteristics Duration of HHD (range): 25.3 months (13 to 72) Residual kidney function: none Dialysis access: all native AVF Ethnicity: Caucasian (8); Polynesian (1) Number: 9 Mean age (range): 48 years (23 to 63) Sex (M/F): 4/5
Interventions	ICHD Volume‐controlled HD (Fresenius 2008A) machine, 3.5 to 4.5 h/session, 3 times/wk 1.5 m² cuprophan dialyser was used with bicarbonate buffer and blood flows of 300 to 350 mL/min HHD 6 to 8 hours dialysis 3 times/wk at home using a 0.8 m² cuprophan dialyser, acetate buffer, and a 200 mL/min blood flow
Outcomes	Outcomes relevant to this review BP: ambulatory and pre‐ and post‐dialysis, symptomatic hypotension ECG: left ventricular mass index, left ventricular function, left ventricular diastolic volume, diastolic function Symptoms and QoL: uraemia‐related symptoms, physical suffering, interference with social activity, burden on families
Identification	Additional information Funded by the Health Research Council of New Zealand and the Canterbury Medical Research Foundation, with additional support from Fresenius Health Care and Roche New Zealand Ltd
Notes	Contacted author to attempt to obtain measures of variability and further details from study, however data no longer available due to earthquake damage
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Sequence generation using random number table designed by statistician (data obtained from authors on request)
Allocation concealment (selection bias)	Low risk	Treatment allocation assigned by statistician unaware of patient details (data obtained from authors on request)
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not blinded
Blinding of outcome assessment (detection bias) All outcomes	Low risk	QoL, BP, and echocardiography outcomes assessed by investigators unaware of treatment sequence (data obtained from authors on request)
Incomplete outcome data (attrition bias) All outcomes	Low risk	0/9 (0%)
Selective reporting (reporting bias)	Low risk	All relevant outcomes systematically assessed
Other bias	High risk	Unmatched interventions (acetate versus bicarbonate buffer)

Murashima 2010.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: single cross‐over Statistical analysis: association between modality and intradialytic hypotension was estimated using conditional logistic regression models in which hypotension was the dependent variable, modality was the categorical predictor, and patient was the conditioning factor Study characteristics Country: USA Setting: Single home HD centre Inclusion criteria: > 18 years; ESKD treated at a single home HD facility; converted from conventional HD to SDHD during the calendar year 2007; on conventional HD for at least 1 month before conversion; on SDHD for at least 6 months following conversion Exclusion criteria: active infection or bleeding; underwent kidney transplantation during the study period; without records of BP measurement
Participants	Baseline characteristics Median age (range): 48 years (34 to 77) Male: 42% Race White (33%); African American (58%); Asian (8%) Primary disease Diabetes (25%); hypertension (33%); other (42%) Comorbidities Diabetes (33%); Hypertension (100%); coronary artery disease (42); PVD (17%) Group differences Each patient served as his or her own control
Interventions	ICHD Location: in‐centre Operator: staff Duration, median: between 3.5 and 4 hours Frequency: 3 times/week Intensive HHD Location: home (NxStage System One) Operator: patient Duration, median: 2.5 hours Frequency: daily
Outcomes	Outcomes relevant to this review Pulse pressure: mm Hg Incidence of intradialytic hypotension (OR) Clinically significant hypotension during HD (OR) Definitions Intradialytic hypotension was defined in 2 ways: Intradialytic hypotension by KDOQI BP criteria: frequency of the decrease in systolic BP of at least 20 mm Hg or decrease in MAP of at least 10 mm Hg (as there was no standard for documenting the symptoms of intradialytic symptoms) Clinically significant hypotension during dialysis: systolic BP of 90 or less or a diastolic BP of 55 or less at any time during a dialysis treatment
Identification	Additional information Sponsorship source: Davita Clinical Re‐search and the Davita Franklin Dialysis Unit for their support (unclear if financial or other) Authors name: Miho Murashima Institution: Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania, USA Email: joel.glickman@uphs.upenn.edu Address: J. D. Glickman, Renal Electrolyte and Hypertension Division, Hospital of the University of Pennsylvania, 1 Founders, 3400 Spruce Street, Philadelphia, PA 19104, USA Other authors: Dinesh Kumar, Alden M. Doyle, Joel D. Glickman
Notes

Nebel 2002.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: parallel Study characteristics Country: Germany Setting: KRT patients at a single centre (HHD, ICHD, PD, kidney transplant) Inclusion criteria: patients receiving dialysis treatment in Colonge‐Merheim between 1990 and 1999
Participants	Baseline characteristics HHD Mean age: 44.45 years Male: 70.3% Average duration of treatment: 47.43 months Diabetes: 0% ICHD Mean age: 47.43 years Male: 48.5% Average duration of treatment: 40.43 months Diabetes: 10.6% Group differences Larger proportion of male participants receiving HHD No patients with diabetes receiving HHD
Interventions	HHD ICHD
Outcomes	Outcomes relevant to this review Cost of dialysis treatment per year Reported outcomes Reported primary outcomes Cost (including inpatient and outpatient treatment (whether condition was related to KRT or not), outpatient care (including laboratory tests, investigations and imaging, medical fees), medications (immunosuppression, EPO and other medications), transport to outpatients and dialysis)
Identification	Additional information Sponsorship source: None stated Authors name: M Nebel Institution: Kuratorium fur Dialyse und Nierentransplantation e.V., Dialysezentrum Koln‐Merheim Email: Michael.Nebel@kfh.de Address: not available
Notes	Contacted author to obtain measures of variability, however advised data not available

Nesrallah 2012.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: parallel Primary exposure: dialysis modality (intensive HHD vs. conventional in‐centre HD) Statistical analysis Attributed death to dialysis modality at index date, regardless of switches to other dialysis modalities Propensity scores with logistic regression, regressing type of haemodialysis (intensive versus conventional) Matched patients by country, duration of ESKD before study enrollment, and propensity score, with up to 10 conventional haemodialysis patients for each intensive haemodialysis patient, using a “greedy‐matching” (nearest‐neighbour) algorithm Total follow‐up time was 3008 patient‐years Covariables: estimated propensity scores with logistic regression, regressing type of HD (intensive versus conventional) using the following covariates: age, sex, diabetes, myocardial infarction, congestive heart failure, cerebrovascular disease, cancer, race, and dry weight Study characteristics Country: France, USA, Canada Setting: two multinational renal databases on patients receiving intensive and conventional HD: the International Quotidian Dialysis Registry (IQDR), and the Dialysis Outcomes and Practice Patterns Study (DOPPS), respectively Inclusion criteria: ≥ 18 years at enrolment; Data obtained from IQDR and DOPPS For the study group, included patients receiving intensive HD, defined as≥ 5.5 hours/session (day or overnight), 3 to 7 sessions/week; performed at home For the comparator group, selected patients receiving conventional HD for < 5.5 hours/session, 3 sessions/week in a clinic or hospital setting
Participants	Baseline characteristics (after matching) Intensive HHD Mean age ± SD: 50.8 ± 12.4 years Male: 70.4% Race White (75.4%); Black (7.4%); other (17.2%) Country Canada (71.0%); France (14.5%); USA (14.5%) Mean duration ESKD ± SD: 5.7 ± 6.2 months Comorbidities Diabetes (28.1%); MI (13.0%); CHF (13.9%); PVD (12.4%); cerebrovascular disease (5.9%); COPD (6.8%); cancer (11.2%) ICHD Mean age ± SD: 52.3 ± 12.4 years Male: 64.2% Race White (72.5%); Black (11.5%); other (16.0%) Country Canada (71.0%); France (14.5%); USA (14.5%) Mean duration ESKD ± SD: 5.7 ± 3.1 months Comorbidities Diabetes (27.2%); MI (13.0%); CHF (15.1%); PVD (14.8%); cerebrovascular disease (5.9%); COPD (9.5%); cancer (12.7%)
Interventions	Intensive HHD Location: home Duration: at least 5.5 hours/session (day or overnight) Frequency: 3 to 7 times/week ICHD Location: clinic or hospital setting Duration: < 5.5 hours/session Frequency: 3 times /week
Outcomes	Outcomes relevant to this review HR for mortality (relative to conventional facility HD), matched sample, no censoring for modality switches Notes: matched sample HR for mortality (relative to conventional facility HD), matched sample, no censoring for modality switches (adjusted) Notes: matched sample ‐‐ Adjusted for variables not achieving <10% standardised difference after matching. Final model included the following covariates: age at index, sex, race, and diabetes Reported outcomes Primary outcome: all‐cause mortality
Identification	Additional information Sponsorship source: A.X.G. was funded by a clinician scientist award from the Canadian Institutes of Health Research. P.C.A. was supported by a career investigator award from the Heart and Stroke Foundation of Ontario. R.S.S. was supported by a randomized trials mentorship award from the Canadian Institutes of Health Research. This research was funded by unrestricted research grants from Baxter Healthcare Corporation, Gambro R&D, Fresenius Medical Care, and the Canadian Institutes for Health Research. R.M.L., A.X.G., and R.S.S. have an unrestricted research grant from Fresenius Medical Care Canada. Authors name: Gihad E. Nesrallah Institution: University of Western Ontario, London, Ontario, Canada Email: Gihad.Nesrallah@lawsonresearch.com Address: Dr. Gihad E. Nesrallah, London Kidney Clinical Research Unit, London Health Sciences Center, 800 Commissioners Road East, Room ELL‐101, London, Ontario N6A 4G5 Canada Other authors: Robert M. Lindsay, Meaghan S. Cuerden, Amit X. Garg, Friedrich Port, Peter C. Austin, Louise M. Moist, Andreas Pierratos, Christopher T. Chan, Deborah Zimmerman, Robert S. Lockridge, Cécile Couchoud, Charles Chazot, Norma Ofsthun, Adeera Levin, Michael Copland, Mark Courtney, Andrew Steele, Philip A. McFarlane, Denis F. Geary, Robert P. Pauly, Paul Komenda, and Rita S. Suri
Notes

Nitsch 2011.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: parallel Primary exposure: Dialysis modality (HHD vs PD/ICD/satellite HD) Statistical analysis Age and sex‐matched controls: Using frequency matching for age (using bands: 18–34, 35–44, 45–54, 55–64, 65–74, 75–84 and 85+ years old) and gender, we identified incident control patients (4 hospital HD patients, 4 PDs patients and 2 satellite HD patients per HHD patient) If more than the required number of controls were available, a computerized random number generator was used to randomly select the required number of controls Follow‐up data were collected until 31 December 2006, and these included date of death, date of recovery of kidney function, date of loss to follow‐up, dates of wait‐listing for transplantation, and modality changes (including transplantation) Descriptive analysis of time to wait‐listing for kidney transplantation, modality survival and survival was carried out using Kaplan–Meier graphs, calculating crude rates and log‐rank testing Censoring for death was performed at the end of the observation period or at the end of follow‐up (31 December 2006) except for patients who recovered within a year of starting their dialysis Adjusted for date of home HD start using a time‐dependent variable to avoid immortal time bias Final multivariable Cox models were adjusted for clustering in‐centre using robust standard errors Covariables: All adjusted models included terms for age, sex, primary renal disease, and year of start of dialysis Study characteristics Country: England and Wales (UK) Setting: UK Renal Registry (UKRR) ‐‐ registry analysis Inclusion criteria: all incident patients who commenced on HHD from 1 January 1997 until 31 December 2005 were identified from the UK Renal Registry (UKRR) database Exclusion criteria: < 18 years
Participants	Baseline characteristics HHD Mean age ± SD: 47.4 ± 13.9 years Male: 70.7% Ethnicity White (95.0%); Asian (3.5%); Black (1.0%); Chinese (0%); other (0.5%) Mean eGFR at start: 7.9 ± 6.3 Primary disease Diabetes (10.2%); GN (20.9%); hypertension (1.8%); PKD (18.2%); renal vascular (4.0%); pyelonephritis (10.2%); other (16.0%); uncertain (15.6%); missing (3.1%) Waitlisted over time after KRT start: 70.2% Waitlisted before KRT start: 12.9% ICHD Mean age ± SD: 48.3 ± 14.2 years Male: 70.7% Ethnicity White (76.1%); Asian (13.4%); Black (6.6%); Chinese (0.3%); other (3.6%) Mean eGFR at start: 8.6 ± 5.4 Primary disease Diabetes (22.3%); GN (11.6%); hypertension (6.1%); PKD (8.2%); renal vascular (3.1%); pyelonephritis (9.0%); other (15.7%); uncertain (20.6%); missing (3.4%) Waitlisted over time after KRT start: 42.4% Waitlisted before KRT start: 6.3% Satellite HD Mean age ± SD: 48.0 ± 14.5 years Male: 70.7% Ethnicity White (66.4%); Asian (11.6%); Black (14.1%); Chinese (0.3%); other (7.7%) Mean eGFR at start: 9.7 ± 7.3 Primary disease Diabetes (20.2%); GN (14.0%); hypertension (4.7%); PKD (6.9%); renal vascular (3.3%); pyelonephritis (5.8%); other (13.3%); uncertain (17.3%); missing (14.4%) Waitlisted over time after KRT start: 32.9% Waitlisted before KRT start: 7.3% Definitions Hospital HD: HD provided by a dialysis unit with renal inpatient facilities on the same site Satellite HD: HD provided by a dialysis unit with no inpatient renal facilities on‐site HHD patients: patients who started HHD after start of KRT
Interventions	HHD Location: home Operator: patient ICHD Location: dialysis unit with renal inpatient facilities on the same site Operator: staff Satellite HD Location: dialysis unit with no inpatient renal facilities on‐site Operator: unclear, possibly staff
Outcomes	Outcomes relevant to this review OR for waitlisting for kidney transplantation before KRT start, adjusted 1‐year survival, crude HR for long‐term survival HR for waitlisting for kidney transplantation after KRT start, adjusted Notes: Excluding Chinese patients and satellite haemodialysis due to violation of proportional hazards. HR for survival after RRT start, adjusted Notes: Adjusted for date of wait‐listing for transplantation, date of start HHD and date of transplantation as time‐changing variables. Results are shown for comparisons among HD patients, and separately for comparisons with PD patients. This was because hospital HD and PD were not directly comparable within the same analyses due to violation of the proportional hazards assumption Reported outcomes Primary outcome: time to wait‐listing for kidney transplantation, modality survival and patient survival
Identification	Additional information Sponsorship source: None mentioned. No conflict of interest declared. Authors name: Dorothea Nitsch Institution: London School of Hygiene and Tropical Medicine, London, UK Email: Dorothea.Nitsch@lshtm.ac.uk Other authors: Retha Steenkamp, Charles R.V. Tomson, Paul Roderick, David Ansell and Mark S. MacGregor
Notes

NxStage‐USRDS 2012.

Study characteristics
Methods	Weinhandl 2012 Study design: retrospective cohort study Study grouping: parallel Primary exposure: dialysis modality (daily HHD vs 3 times/week ICHD) Covariables: matched for age, cumulative hospital days, cumulative EPO dose, BMI, transplant waitlist registration, CHF, ESKD duration, race, cancer, primary ESKD cause, cerebrovascular disease, PVD, other cardiovascular diseases, diabetes, atherosclerotic heart disease, sex, dual Medicare/Medicaid eligibility Statistical analysis: for patients with possible interruptions of no longer than 1 month, we ignored gaps in treatment and merged the record pairs. For patients with interruptions of longer than 1 month, only the first record was retained. For each daily HHD patient, 5 thrice‐weekly in‐centre patients with matching characteristics were selected. Match quality was assessed with standardised differences; differences <10% indicated similarity Inclusion criteria Daily HHD patients: identified from a registry of NxStage SystemOne users maintained by NxStage Medical Inc. Beginning dates ranged from January 1, 2005, to December 31, 2007, and ending dates (as applicable) occurred as late as December 31, 2008. Able to be linked in USRDS database. Patients with five or six prescribed dialysis sessions/week, Medicare primary payer status during the 3 months preceding NxStage System One use initiation, or beginning of KRT during the 6 months preceding use initiation Thrice‐weekly in‐centre patients: all patients who were treated with in‐centre HD between January 1, 2005, and December 31, 2007 Weinhandl 2015a Study design: retrospective cohort study Study grouping: parallel Primary exposure: dialysis modality (daily HHD vs ICHD) Statistical analysis: stratified patients by year of daily HHD initiation: 2006‐2007 (era 1) or 2008‐2009 (era 2). Propensity score matching (5 IHD; 1 DHHD). Used both intention‐to‐treat and on‐treatment approaches during follow‐up. In the former, followed‐up patients from the index date (daily HHD patients) or date of in‐centre haemodialysis (matched patients) to the earliest of death, kidney transplantation, loss of Medicare primary payer coverage, and December 31, 2009 (era 1) or December 31, 2010 (era 2). In the latter, added the date exactly 2 months after cessation of daily HHD therapy (daily HHD patients) or in‐centre haemodialysis (matched patients) to the list of dates on which follow‐up may end; the 2‐month extension after a change in dialytic modality assigned late admissions to the first modality Inclusion criteria Daily HHD patients were identified from a registry of NxStage System One users maintained by NxStage Medical Inc. Beginning dates ranged from January 1, 2006, through December 31, 2009, and ending dates (when observed) occurred as late as December 31, 2010. For each era, the source cohort of thrice‐weekly in‐centre HD patients comprised those who were treated and had Medicare primary payer coverage for one or more days during the era Weinhandl 2015d Study design: retrospective cohort study Study grouping: parallel Primary exposure: dialysis modality (daily HHD vs ICHD) Statistical analysis: DHHD patients who initialed NxStage System One between 1 January 2007 and 30 June 2010 who were within 6 months of kidney failure diagnosis were matched 5‐to‐1 with ICHD patients from the USRDS database according the propensity score at dialysis initiation. Propensity scoring included demographic factors, comorbidity factors, and biochemistry. Patients were followed from daily HHD initiation date or matched index date, until earliest of kidney transplant, death or 31 December 2010 Study characteristics Country: USA Setting: USRDS and NxStage linked analysis
Participants	Baseline characteristics: Weinhandl 2012 HHD Mean age: 52.2 years Male: 64.2% Race Black (26.5%); other (73.5%) Primary disease Diabetes (27.3%); hypertension (19.3%); GN/PKD (30.3%); other/unknown (23.1%) Mean ESKD duration: 5.5 years Mean BMI: 28.1 Comorbidities Atherosclerotic heart disease (24.0%); cerebrovascular disease (8.3%); CHF (26.9%); PVD (20.9%); other cardiovascular disease (20.0%); cancer (9.1%); diabetes (40.6%) Dual Medicare/Medicaid eligibility: 23.2% Transplant waitlist registration: 35.0% ICHD Mean age: 53.2 years Male: 62.3% Race Black (28.3%); other (71.7%) Primary disease Diabetes (30.3%); hypertension (20.6%); GN/PKD (28.4%); other/unknown (20.7%) Mean ESKD duration: 5.1 years Mean BMI: 27.9 Comorbidities Atherosclerotic heart disease (22.7%); cerebrovascular disease (8.1%); CHF (27.1%); PVD (20.5%); other cardiovascular disease (17.9%); cancer (7.3%); diabetes (42.1%) Dual Medicare/Medicaid eligibility: 35.4% Transplant waitlist registration: 34.8% Baseline characteristics: Weinhandl 2015a HHD Mean age: 53.4 years Male: 65.3% Race Balck (26.8%); White (73.2%) Primary disease Diabetes (29.5%); hypertension (21.4%); GN/cystic kidney disease (27.0%); other/unknown (22.1%) Mean ESKD duration: 5.8 years Dual Medicare/Medicaid enrolment: 29% Comorbidities Cardiac disease NOS (31.7%); cerebrovascular disease (8.9%); CHF (34.8%); hypertension (40.4%); IHD (29.3%); peripheral arterial disease (24.2%); pulmonary heart disease (2.0%); cancer (10.3%); chronic pulmonary disease (11.6%); diabetes (46.1%) Mean BMI: 28.5 Kidney transplant waitlist registration: 34.6% ICHD Mean age: 53.6 years Male: 65.6% Race Balck (27.0%); White (73.0) Primary disease Diabetes (30.5%); hypertension (21.421.6); GN/cystic kidney disease (26.2%); other/unknown (21.6%) Mean ESKD duration: 5.4 years Dual Medicare/Medicaid enrolment: 29% Comorbidities Cardiac disease NOS (32.2%); cerebrovascular disease (9.2); CHF (35.7%); hypertension (41.0%); IHD (29.8%); peripheral arterial disease (24.2%); pulmonary heart disease (1.9%); cancer (10.4%); chronic pulmonary disease (11.6%); diabetes (47.4%) Mean BMI: 28.6 Kidney transplant waitlist registration: 32.9% Baseline characteristics: Weinhandl 2015d Abstract only; detailed information not provided
Interventions	HHD Location: home Frequency: 5 or 6 prescribed dialysis sessions/week ICHD Location: in‐centre Frequency: 3 times/week
Outcomes	Weinhandl 2012 Outcomes relevant to this review HR for all‐cause mortality (HHD vs matched ICD): ITT HR for cardiovascular mortality (HHD vs matched ICD): ITT HR for infection mortality (HHD vs matched ICD): ITT HR for all‐cause mortality (HHD vs matched ICD): as treated HR for cardiovascular mortality (HHD vs matched ICD): as treated HR for infection mortality (HHD vs matched ICD): as treated HR for interval‐specific death (HHD vs matched ICD): ITT HR for interval‐specific death (HHD vs matched ICD): as treated Reported outcomes Primary outcome: mortality Weinhandl 2015a Outcomes relevant to this review All‐cause hospital admissions, RR, ITT Cardiovascular hospital admission, RR, ITT Vascular hospital admission, RR, ITT All‐cause hospitalisation duration, RR, ITT Cardiovascular hospitalisation duration, RR, ITT Vascular access hospitalisation duration, RR, ITT All‐cause hospital admissions, RR, as treated Cardiovascular hospital admissions, RR, as treated Vascular access hospital admissions, RR, as treated All‐cause hospitalisation duration, RR, as treated Cardiovascular hospitalisation duration, RR, as treated Vascular access hospitalisation duration, RR, as treated Reported outcomes Primary outcome: risk of hospitalisation‐ all cause‐ cause‐specific‐ type‐specific. Hospital days Weinhandl 2015d Outcomes relevant to this review Incidence of kidney transplantation Reported outcomes Primary outcome: incidence of kidney transplantation
Identification	Additional information Sponsorship source: This work was supported by a grant from NxStage Medical Inc, Lawrence, Massachusetts. D.T.G. has provided consulting services to DaVita Clinical Research and A.J.C. to NxStage. Authors name: Eric D. Weinhandl Institution: Chronic Disease Research Group, Minneapolis Medical Research Foundation Email: eweinhandl@cdrg.org Address: Mr. Eric D. Weinhandl, Chronic Disease Research Group, Minneapolis Medical Research Foundation, 914 South 8th Street, Suite S‐206, Minneapolis, MN 55404. Other authors (Weinhandl 2012): Jiannong Liu, David T. Gilbertson, Thomas J. Arneson and Allan J. Collins Other authors (Weinhandl 2015): Kimberly M. Nieman, David T. Gilbertson and Allan J. Collins
Notes	Data from Weinhandl 2012 analysed for mortality outcome. Data from Weinhandl 2015 analysed for hospitalisation outcome. Data from Weinhandl 2015d analysed for kidney transplantation outcome. Data from other publications not analysed due to potential overlapping populations for the same outcomes. Contacted author to obtain more detailed outcome data, awaiting response.

Piccoli 2004.

Study characteristics
Methods	Study design Prospective cohort study Study grouping: parallel Statistical analysis Dialysis setting assessed by intention to treat analysis Dialysis schedule assessed by ITT and per protocol analyses If changed to catheter, considered as lost to follow up at that date If transplanted or changed centres, considered lost to follow up at that date Logistic regression model, two dependent variables were tested: freedom from vascular access failure or freedom from any adverse The time‐dependent Cox proportional hazard model was used to investigate the risk of the first adverse event and the risk of access failure event Only patients with vascular accesses other than central catheters were considered for the outcomes (If a patient was switched to a central venous catheter after the failure of an arteriovenous) Primary exposure Schedule: daily HHD vs non‐HHD Location: home or limited care unit Covariables: Age, sex, comorbidity, previous vascular events, schedule, setting of treatment (home, limited care), dialysis follow‐up, vascular access (native vs prosthetic, first vs subsequent) and setting of vascular access creation Study characteristics Country: Italy Setting: single centre (SMOM Unit, a satellite of a large university centre) Inclusion criteria: patients followed at the SMOM Unit from November 1998 to November 2002 (n=77; 42 at home and 35 in a limited care centre; 28 patients experienced at least one trial of daily HHD. Received/eligible for dialysis at home or limited care unit Exclusion criteria: contraindications for out‐of‐hospital dialysis
Participants	Baseline characteristics HHD Median age: 44.7 years Male: 69% Median age at KRT start: 40.6 years Overall follow‐up: 981.9 months Diabetes: 16.6% Cardiovascular comorbidities: 28.6% Other comorbidities: 59.5% Access Native AVF (76.2%); prosthetic AVF (16.6%); CVC (7.1%) Previous adverse events (vascular access): 19% Satellite HD Median age: 49.9 years Male: 66% Median age at KRT start: 47.7 years Overall follow‐up: 868.9 months Diabetes: 20% Cardiovascular comorbidities: 45.7% Other comorbidities: 58.6% Access Native AVF (77.1%); prosthetic AVF (20%); CVC (2.8%) Previous adverse events (vascular access): 40% Group differences No clinically relevant baseline differences between the subsets of daily HHD patients and patients treated with different schedules, with the notable exception of a longer dialysis follow‐up in patients that performed at least one trial of daily HHD Patients on HHD were younger and had a shorter history of KRT with less comorbidity than patients treated in the limited care unit
Interventions	HHD Location: home; the dialysis setting was assessed according to an ‘intention to treat’ analysis, that is, patients in training for HHD were placed in the home dialysis cohort Duration: duration of conventional dialysis ranged from 3.30 to 4.45 h Frequency: daily dialysis was designated as five to six sessions/week (2 to 3 h) Satellite HD Location: in‐centre Duration: duration of conventional dialysis ranged from 3.30 to 4.45 h Frequency: Daily dialysis was designated as five to six sessions per week (2 to 3 h).
Outcomes	Outcomes relevant to this review Adverse event‐free survival (ITT) Vascular access event‐free survival (ITT) Adverse event‐free survival (per protocol) Vascular event‐free survival (per protocol) Reported outcomes Primary outcomes Vascular access failure (defined as need for a new vascular access in the absence of sufficient blood flow to allow performing HD; the lack of response to mechanical declotting or urokinase in case of acute failure; and lack of response to pre‐emptive surgical treatment or angioplasty in case of subacute problems) All events in combination (need for surgical treatment, angioplasty, declotting with urokinase)
Identification	Additional information Sponsorship source: none mentioned. No conflict of interest declared Authors name: Giorgina Barbara Piccoli Institution: University of Turin, Italy Email: gbpiccoli@hotmail.com, giorgina.piccoli@unito.it Address: Giorgina B. Piccoli, Cattedra di Nephrologia, Departmente di Medicina Interna Corso Doguom 16‐10126 Torino Other authors: Francesca Bermond, Elisabetta Mezza, Manuel Burdese, Fabrizio Fop, Giovanni Mangiarotti, Alfonso Pacitti, Stefano Maffei, Guido Martina, Alberto Jeantet, Giuseppe Paolo Segoloni and Giuseppe Piccoli
Notes	Contacted author to obtain more detailed outcome data, awaiting response.

Rydell 2016.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: parallel Statistical analysis ITT analysis and on‐treatment analysis Sex, age, comorbidity and dialysis start date matched patients Primary exposure: dialysis modality (HHD vs ICHD) Other outcomes: factors related to survival: subsequent kidney transplantation, hyperphosphataemia, hypertension, anaemia, hypoalbuminaemia Covariables: sex, age, comorbidity and dialysis start date matched patients Study characteristics Country: Sweden Setting: single centre Inclusion criteria Incident patients (< 6 months on KRT prior to HHD training) starting HHD at Lund University Hospital from 1/1/1983‐31/12/2002, if an appropriate matched ICHD patient could be found Control ICHD patients were required to have the same sex, same level of comorbidity according to Davies Comorbidity Index, similar age (< 5 years difference) and similar start date of ICHD (< 5 years difference) Exclusion criteria: receiving ICHD for < 72 days (median training period) were not accepted as controls
Participants	Baseline characteristics HHD Median age: 51.5 years Male: 76% Davies comorbidity Grade 0 (71%); grade 1 (29%); grade 2 (0%) Primary disease GN (56%); PKD (27%); diabetes (2.4%); nephrosclerosis (2.4%); other (12%); unknown (0%) ICHD Median age: 53.9 years Male: 76% Davies comorbidity Grade 0 (71%); grade 1 (29%); grade 2 (0%) Primary disease GN (44%); PKD (22%); diabetes (12%); nephrosclerosis (5%); other (12%); unknown (5%)
Interventions	HHD ICHD
Outcomes	Outcomes relevant to this review Mean survival, years Survival, ITT (%) Received kidney transplant Survival, on‐treatment (%) Notes: excluding time after kidney transplantation Mean systolic BP Notes: BP measurement was not standardised and was performed under different circumstances Mean diastolic BP Notes: BP measurement was not standardised and was performed under different circumstances Reported outcomes Primary outcome: patient survival
Identification	Additional information Sponsorship source: None mentioned. Helena Rydell has received grants from Skåne Regional Council, The Southern Health Care Region in Sweden, Paul Frankenius Foundation and Swedish Society of Nephrology. Authors name: Helena Rydell Institution: Department of Nephrology Skåne University Hospital and Institution of Clinical Sciences Lund, Lund University Email: Helena.Rydell_Johnsen@med.lu.se Address: Department of Nephrology, Skåne University Hospital Lund, Barngatan 2 A, 22185 Lund (Sweden) Other authors: Naomi Clyne, Mårten Segelmark
Notes	Contacted author to obtain measures of variability, however advised data not available.

Rydell 2019.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: parallel Statistical analysis Two matched control sets by matching HHD patients separately with PD (1:3) and ICHD (1:4) patients Gender, Charlson Comorbidity Index, age (± 3 years) and date for start of KRT (± 3 years) were used as matching criteria Matching was performed at day 0 of KRT Two different survival analyses were performed: ITT: patients were considered at risk also after switching to other KRT. As renal diagnosis was not considered in the matching, it was included in a bivariable Cox regression ITT survival analysis Sensitivity analysis was performed with censoring at recovery of native renal function Two different per‐protocol analyses were performed “on initial KRT” analysis with censoring for all switches from the group modality “on dialysis treatment” analysis with censoring at kidney transplantation, recovered native kidney function and end of study For hospital admission outcomes, only admissions from day 90 after KRT were included in analyses and follow‐up was analyses according to two approaches Per protocol analysis where, only admissions that started while the patients were still on their initially KRT were included ITT analysis, where all admissions until the end of the study (31st December 2013) were included Primary exposure: Dialysis modality (HHD vs ICHD vs PD). HHD, ICHD or PD as initial KRT were defined as the modality registered in the Swedish Renal Registry (SRR) at day 90 after start of KRT Study characteristics Country: Sweden Setting: Swedish Renal Registry (SRR) Inclusion criteria: ≥ 18 years; registered in the SRR and starting KRT between January 1st 1991 and December 31st 2012 were eligible for inclusion if they fulfilled the criteria of HHD, ICHD or PD as initial KRT HHD: first KRT, a patient was not allowed to have received PD before day 90 PD: a patient was not allowed to have had HHD before day 90 ICHD: no other KRT was allowed during the first year after start of RRT, except for transplantation after day 90 Exclusion criteria: A failing kidney transplant or a period of recovered kidney function before day 90
Participants	Baseline characteristics HHD Median age: 50.2 years Male: 82% Charlson comorbidity index Grade 0 (63%); grade 1 (28%); grade 2 (8%); grade 3 (2%) Primary disease Diabetes (10%); GN (30%); hypertension (6%); PKD (15%); pyelonephritis (4%); other (28%); unspecified (6%) ICHD Median age: 50.1 years Male: 82% Charlson comorbidity index Grade 0 (63%); grade 1 (28%); grade 2 (8%); grade 3 (2%) Primary disease Diabetes (20%); GN (25%); hypertension (7%); PKD (10%); pyelonephritis (3%); other (23%); unspecified (12%) PD Median age: 50.1 years Male: 82% Charlson comorbidity index Grade 0 (63%); grade 1 (28%); grade 2 (8%); grade 3 (2%) Primary disease Diabetes (27%); GN (28%); hypertension (5%); PKD (9%); pyelonephritis (3%); other (20%); unspecified (69%)
Interventions	HHD ICHD PD
Outcomes	Outcomes relevant to this review Median survival, censored at end of study, years Median survival, censored at end of study and recovered kidney function, years Median survival, on initial KRT modality only, years Median survival, on treatment only, years All‐cause annual hospitalisation rate (per protocol and ITT) Hospitalisation days per year (per protocol and ITT) Time to hospitalisation (per protocol and ITT), years Reported outcomes Primary outcome: patient survival
Identification	Additional information Sponsorship source: Helena Rydell has received grants from Skåne Regional Council, The Southern Health Care Region, Paul Frankenius Foundation and Swedish Society of Nephrology Authors name: Helena Rydell Institution: Department of Clinical Sciences Lund, University, Skane University Hospital, Nephrology Lund, Lund, Sweden Email: Helena.rydell@sll.se Address: Department of Clinical Sciences Lund, University, Skane University Hospital, Nephrology Lund, Lund, Sweden Other authors: Kerstin Ivarsson, Martin Almquist, Mårten Segelmarkand Naomi Clyne
Notes	Contacted author to obtain measures of variability, however advised data not available.

Sands 2009.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: single cross‐over Primary exposure: dialysis modality Study characteristics Country: USA Setting: patients who transitioned from ICHD to HHD in an outpatient dialysis facility Inclusion criteria: transitioned from ICHD to HHD on the 2008K@home HD machine when used in the home setting; minimum 25 HD treatments during each of the ICHD and home periods; at least 3 in‐centre paired pre‐ and post‐dialysis BUN determinations within 6 months immediately before conversion to at‐home HD and at least 3 paired home pre‐ and post‐dialysis BUN determinations within 6 months immediately following the HHD start date; documented evidence of completed training for patient and lay partner; minimum of 3 months of available in‐centre medical records that immediately preceded the HHD start date for each identified patient (in‐centre training records were considered part of the available in‐centre medical record); minimum of 3 months of available medical records immediately after the HHD start date for each identified patient; patient used the 2008K@home for HHD during the at‐home period; patient was dialysed on2008H or 2008K HD machines during the in‐centre period Exclusion criteria: unable to satisfy the inclusion criteria; insufficient evidence of training for patient or partner; patients using or requiring nurse‐assisted (healthcare professional‐assisted) HD while at home; medical records that were shipped off‐site and were not available for analysis during the retrospective review; the in‐centre period was conducted at a non‐Fresenius Medical Services facility
Participants	Baseline characteristics Mean age: 50 years Male: 76% Race White (79%); Black (14%); other )7%) Diabetes: 31% Access AVF (65.5%); AVG (20.7%); CVC (13.8%) Mean vintage at HHD start: 4.3 years Mean post‐dialysis weight: 96 kg
Interventions	HHD Location: home (2008K@home HD machine) Operator: patient and lay partner Training: HHD training program on the 2008K@home of approximately 8 weeks for patients and lay partners ICHD Location: outpatient dialysis facility (2008 series HD machine) Operator: professional dialysis personnel
Outcomes	Outcomes relevant to this review Hospitalisations per 100 treatments Access complications (arterial site), number per 100 treatments Notes: type of complications included not specified Access complications (venous site), n per 100 treatments Notes: type of complications included not specified Reported outcomes Standard weekly Kt/V Safety was assessed by reviewing the medical records for evidence of 30 different AEs as documented in the medical records and treatment sheets, from patient complaints and from medical device reports reported over the 3‐month period before (in‐centre) and after (home) patients transitioned from ICHD to HHD. The AEs included all hospitalisations, access complications, allergic reactions, arrests, blood loss > 100 mL, blood/body fluid/chemical exposures, fevers, injuries, loss of consciousness, procedure variances, and all equipment or disposables malfunctions consistent with a list of up to 30 routinely monitored dialysis‐related clinical variances in all Fresenius Medical Care outpatient HD facilities
Identification	Additional information Sponsorship source: None stated, however all authors employed by Fresenius Authors name: Jeffrey J. Sands Institution: Renal Therapies Group, Fresenius Medical Care North America, Celebration, Florida Email: Jeffrey.Sands@fmc‐na.com Address: Jeffrey J. Sands, MD, Fresenius Medical Care, NA, 231 Celebration Blvd., Celebration, FL 34747 Other authors: Eduardo Lacson Jr., Norma J. Ofsthun, Janet C. Kay and Jose A. Diaz‐Buxo
Notes	Unable to contact author to obtain more detailed outcome data

Saner 2005.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: parallel Primary exposure: dialysis modality (HHD vs ICD) Statistical analysis: for each patient trained for HHD at the dialysis centre between 1970 and 1995 (n=103), a corresponding match was searched from the ICHD patients by retrospective chart analysis. After univariable analysis, possible predictive factors (treatment modality, gender, smoking status, marital status, Khan comorbidity index and renal disease) were entered into a multivariable Cox proportional hazards model Study characteristics Country: Switzerland Setting: single centre Inclusion criteria: all patients starting HHD at the University Hospital of Berne between 1970 and 1995. For each dialysis patient treated at home, a corresponding match was searched from the HD patients treated in the centre during the same period by retrospective chart analysis. Each matched pair of patients had to have the same sex and age (± 5 years), to start HD treatment at the same time (± 2 years) and to have the same underlying cause of kidney disease causing ESKD (i.e. GN, pyelonephritis, analgesic nephropathy, PKD, renal vascular disease, unknown); patients eligible for matching were treated by HD for ≥ 3 months prior to any change ofKRT; intermittent PD for ≤ 1 month was accepted Exclusion criteria: treated at any time by HHD or self‐care dialysis were excluded as possible matches for the HHD treatment group
Participants	Baseline characteristics HHD Male: 67% Married: 84% Mean age at dialysis start: 50.1 years Diabetes: 0% Smoker: 27.6% History of: MI (5.2%); cerebrovascular disease (1.7%); peripheral arterial disease (3.4%) TB: 5.2% COPD: 12.1% Khan comorbidity index Low (86%); medium (5%); high (9%) ICHD Male: 67% Married: 70% Mean age at dialysis start: 50.6 years Diabetes: 1.7% Smoker: 34.5% History of: MI (6.9%); cerebrovascular disease (5.2%); peripheral arterial disease (5.2%) TB: 6.9% COPD: 12.1% Khan comorbidity index Low (74%); medium (17%); high (9%)
Interventions	HHD Duration: Same dialysis treatment parameters and time prescriptions were used at home and in the centre ICHD Duration: Same dialysis treatment parameters and time prescriptions were used at home and in the centre
Outcomes	Outcomes relevant to this review Received kidney transplant All‐cause hospitalisation: number/patient Vascular access surgery: number/patient Survival (%) Cardiovascular death during study Reported outcomes Primary outcome: patient survival
Identification	Additional information Sponsorship source: supported by grant no. 3200‐049585 from the Swiss National Science Foundation Authors name: Esther Saner Institution: University of Bern Email: uehlinger@mph.unibe.ch Address: Dominik Uehlinger, MD, Division of Nephrology/Hypertension, University of Bern, Freiburgstrasse, 3010 Bern – Inselspital, Switzerland Other authors: Dorothea Nitsch, Claude Descoeudres, Felix J. Frey and Dominik E. Uehlinger
Notes	Contacted author to obtain more detailed outcome data, awaiting response

Suri 2015.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: parallel Primary exposure: dialysis modality (daily HHD, conventional ICHD, PD) Statistical analysis: propensity scoring using age, sex, race, BMI, diabetes, hypertension, CHF, cerebrovascular disease, PVD, COPD, cancer, prior transplant, medical insurance coverage, albumin, and Hb. Then matched between one and three PD patients to each DHD patient by the following: the propensity score (caliper width=0.02), first KRT start date (5‐year intervals from 1995 to 2009), duration of ESKD before index date (index date minus first KRT start date: 1– 3months, 3–6 months, 6–12 months, 12–24 months, 24–48 months, 48–72 months, 72–96 months, > 96 months), age (3‐year intervals), sex, race, body mass index (5 kg/m2 intervals), congestive heart failure, cancer, and cerebrovascular disease. Matching on first KRT start date and vintage. Matched 1–3 conventional ICHD patients to each DHD patient by propensity score (caliper width=0.02), first KRT start date, age (3‐year intervals), sex, race, BMI (5 kg/m2 bins), albumin (1.5 g/dL intervals), congestive heart failure, diabetes, and prior transplant Study characteristics Country: USA Setting: registry analysis (large US dialysis provider and USRDS) Inclusion criteria: ≥ 18 years who began daily HHD (> 5 days/week, 1.5 to 4.5 h/day) between January 2004 and December 2009 in large US dialysis provider database. Matched ICD and PD controls selected from USRDS, ≥ 18 years, insured by Medicare as primary payer Exclusion criteria: daily HHD patients who received PD, and PD controls who received daily HHD at any time (ICHD controls who received PD first were not excluded). Patients not in the time window, not Medicare before index, non‐independent living, missing race, missing comorbidity, BMI > 50 or <1 6 or missing, albumin < 1.0 g/dL, Hb < 5 g/dL, prior transplants > 2, follow‐up < 30 days
Participants	Baseline characteristics HHD Mean age: 50.3 years Male: 67.6% Smoker: 7.2% Mean BMI: 29.6% Race White (68.8%); Black (28.1%); other (3.0%) Mean ESKD duration: 3.5 years Access AVF (10.6%); AVG (2.0%); CVC (34.9%); unknown (52.5%) Transplant history Waitlisted (39.0%); 0 prior transplants (91.3%); 1 prior transplant (8.4%); 2 prior transplants (0.2%) Comorbidities Diabetes (22.9%); hypertension (82.1%); CHF (18.1%); IHD (5.5%); cerebrovascular disease (4.8%); PVD (8.3%); COPD (5.3%); cancer (5.8%) ICHD Mean age: 50.8 years Male: 67.6% Smoker: 7.3% Mean BMI: 29.8% Race White (68.0%); Black (29.1%); other (3.0%) Mean ESKD duration: 3.5 years Access AVF (9.4%); AVG (2.1%); CVC (33.4%); unknown (55.0%) Transplant history Waitlisted (38.1%); 0 prior transplants (93.5%); 1 prior transplant (6.4%); 2 prior transplants (0.1%) Comorbidities Diabetes (23.5%); hypertension (81.6%); CHF (18.3%); IHD (5.8%); cerebrovascular disease (5.5%); PVD (9.5%); COPD (5.2%); cancer (3.6%)
Interventions	Intensive HHD Duration: 1.5 to 4.5 h/day Frequency: > 5 days/week ICHD
Outcomes	Outcomes relevant to this review HR for composite hospitalisation Notes: composite of all hospitalisations from the index date to the end of follow‐up owing to the prespecified causes of cardiovascular, infectious, access‐related, and bleeding Composite hospitalisation rate/patient‐year HR for time to first hospitalisation Notes: accounting for the competing events of death and transplantation HR for cardiovascular hospitalisations HR for access‐related hospitalisations HR for access infection‐related hospitalisations Reported outcomes Primary outcome: time to all hospitalisations owing to the prespecified causes (cardiovascular, infectious, access‐related, and bleeding). Events of transplantation, death, recovery of kidney function, and loss to follow‐up were censored; switches from daily HHD or PD to an alternative dialysis modality were censored at 90 days after the switch to reduce bias from informative censoring
Identification	Additional information Sponsorship source: This study was funded by a peer‐reviewed grant from the Baxter Extramural Grant Program. Authors name: Rita S. Suri Institution: Centre Hospitalier de l’Université de Montréal (CHUM) Email: rsuri.kidney@gmail.com Address: Rita S. Suri, Department of Medicine, Section of Nephrology, Centre de Recherche, Centre Hospitalier de l’Université de Montréal (CHUM), Room 3148, Hôpital Saint‐Luc, 1058 St Denis St, Montreal, Quebec H2X 3J4, Canada Other authors: Lihua Li and Gihad E. Nesrallah
Notes	Contacted author to obtain more detailed outcome data, awaiting response

Tablo IDE 2020.

Study characteristics
Methods	Study design Prospective cohort study Study grouping: cross‐over The trial consisted of four treatment phases during which Tablo was utilised: Run‐in phase of one‐week in‐centre In‐centre phase of 32 treatments (approximately eight weeks), during which the dialysis staff managed the treatments Transition phase of up to 4 weeks to train the patient or care partner to manage the dialysis Final in‐home phase of 32 treatments (approximately 8 weeks) Study characteristics Country: USA Setting: Investigational Device Exemption (IDE) study Inclusion criteria: adult patients (age 18 to 75 years) with ESKD treated with maintenance HD (either ICD or HHD) who consistently achieved a single pool Kt/Vurea ≥ 1.2 and who were stable for at least 3 months with vascular access providing a blood flow of at least 300 mL/min; participants were expected to be able to adhere to the trial protocol including a willingness to do HHD and the ability to train on Tablo Exclusion criteria: inability to read English or Spanish, a persistent pre‐dialysis systolic BP < 100 mm Hg or > 180 mm Hg despite maximal therapy, New York Heart Association Class III or IV heart failure or an ejection fraction < 30%, and life expectancy < 12 months and/or presence of other ongoing serious illness, as determined by the site investigator
Participants	Baseline characteristics Established on HHD Mean age ± SD: 49.8 ± 13.0 years Male: 62% Ethnicity/race Not Hispanic/Latino (69%); Hispanic/Latino (23%); not reported (8%) White (62%); Black or African American (38%) Mean weight ± SD: 92.1 ± 16.5 kg Mean BMI ± SD: 32.6 ± 6.1 Access AVF (69%); CVC (15%); AVG (15%) Comorbidities Coronary artery disease (46%); diabetes (46%); carotid artery disease (23%); peripheral artery disease (8%); arrhythmia (23%); current smoker (0%); former smoker (23%); hypertension (100%) New on HHD Mean age ± SD: 54.2 ± 10.4 years Male: 65% Ethnicity/race Not Hispanic/Latino (71%); Hispanic/Latino (29%); not reported (0%) White (53%); Black or African American (47%) Mean weight ± SD: 92.5 ± 17.7 kg Mean BMI ± SD: 31.3 ± 4.2 Access AVF (82%); CVC (12%); AVG (6%) Comorbidities Coronary artery disease (35%); diabetes (71%); carotid artery disease (18%); peripheral artery disease (24%); arrhythmia18(23%); current smoker (24%); former smoker (24%); hypertension (94%) Definitions Baseline characteristics are based on patients' previous HD experience (ICHD and HHD). Study protocol meant that all then did ICHD for a number of weeks, then HHD for a number of weeks, and study aimed at comparing those
Interventions	HHD Location: home Frequency: 4 times/week ICHD Location: in‐centre Frequency: 4 times/week
Outcomes	Outcomes relevant to this review Recovery time (hours) Euro‐QOL‐5‐dimension 5‐level (EQ‐5D‐5L)
Identification	Additional information Sponsorship source: Dr. Aragon is an employee of Outset. Drs. Chertow, Alvarez, and Prichard are advisors to Outset. Authors name: Glenn M. Chertow Institution: Stanford University Email: gchertow@stanford.edu Address: Glenn M. Chertow, Division of Nephrology, Stanford University School of Medicine, 1070 Arastradero Road, Suite 311, Palo Alto, CA 94304, USA Other authors: Luis Alvarez, Troy J. Plumb, Sarah S. Prichard, Michael Aragon
Notes

Tennankore 2022.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: parallel Primary exposure: dialysis modality (ICHD, HHD, frequent ICHD, short daily/nocturnal HHD, PD) Covariables: adjusted for baseline variables at start of dialysis including age, sex, race, cause of CKD, comorbidities (diabetes, coronary artery disease, CHF, cerebrovascular disease, PVD, malignancy, chronic obstructive lung disease), facility, Hb, albumin Statistical analysis Hospitalisations were attributed to the previous treatment if they occurred within 30 days of a treatment change For all patients we excluded any admissions that occurred within 7 days of the first dialysis treatment (i.e. at initiation of maintenance HD) Hospitalisations did not include same‐day procedures, outpatient procedures or isolated emergency department presentations without admission The total proportion and rate of hospitalisations (proportion/total follow‐up time) were described for each modality overall and stratified by day of the week Hospitalisation rates were calculated as total count/cumulative time at risk, reported as count/1000 patient days and included those individuals who experienced no admissions Differences in the proportion of patients experiencing a Monday/Tuesday admission with all other days of the week were compared using a generalised linear model with binomial distribution Follow‐up time was from the start of initiation of maintenance dialysis and censored at the first of either death or kidney transplantation Study characteristics Country: Canada Setting: Canadian Organ Replacement Register (CORR) registry analysis Inclusion criteria: cohort study of The Canadian Organ Replacement Register (CORR) from 1 Jan 2005 to 31 Dec 2014 Exclusion criteria: patients residing in Manitoba and Quebec were excluded from the analysis; the former due to an inability to link hospitalisation data and the latter due to a lack of inclusion in CORR for the years covering the cohort
Participants	Baseline characteristics HHD Median age: 51 years Male: 68% Primary disease Diabetes (32%); GN (21%); other (15%); ischaemic/renovascular (13%); unknown (< 5%); hereditary (17%) Race White (67%); Black (14%); other (15%); unknown (4%) BMI < 18.5 (< 5%); 18.5 to < 25 (29%); 25 to < 30 (29%); > 30 (39%) Comorbidities Cardiac disease (11%); CHF (6%); cerebrovascular disease (4%); PVD (9%); chronic obstructive lung disease (< 5%); prior malignancy (7%); diabetes (10%) ICHD Median age: 68 years Male: 61% Primary disease Diabetes (37%); GN (12%); other (22%); ischaemic/renovascular (19%); unknown (< 6%); hereditary (5%) Race White (72%); Black (4%); other (20%); unknown (5%) BMI < 18.5 (4%); 18.5 to < 25 (34%); 25 to < 30 (31%); > 30 (32%) Comorbidities Cardiac disease (32%); CHF (26%); cerebrovascular disease (14%); PVD (17%); chronic obstructive lung disease (12%); prior malignancy (14%); diabetes (14%) Intensive HHD Median age: 50 years Male: 71% Primary disease Diabetes (25%); GN (29%); other (16%); ischaemic/renovascular (7%); unknown (5%); hereditary (17%) Race White (65%); Black (9%); other (24%); unknown (< 5%) BMI < 18.5 (< 5%); 18.5 to < 25 (35%); 25 to < 30 (31%); > 30 (31%) Comorbidities Cardiac disease (13%); CHF (6%); cerebrovascular disease (6%); PVD (9%); chronic obstructive lung disease (< 5%); prior malignancy (12%); diabetes (10%) Intensive ICHD Median age: 66 years Male: 62% Primary disease Diabetes (36%); GN (10%); other (23%); ischaemic/renovascular (14%); unknown (13%); hereditary (5%) Race White (65%); Black (9%); other (24); unknown (< 5%) BMI < 18.5 (< 5%); 18.5 to < 25 (31%); 25 to < 30 (30%); > 30 (37%) Comorbidities Cardiac disease (28%); CHF (31%); cerebrovascular disease (13%); PVD (18%); chronic obstructive lung disease (11%); prior malignancy (13%); diabetes (21%) Definitions The categories were: HD (3 to 6 hours, 2 to 4 times/week) Frequent HD (inclusive of nocturnal HD; > 6 hours, 5 to 6 nights/week and short daily HD; 2 to 3 hours, 5 to 7 days/week) PD (inclusive of CAPD and APD) A separate variable included in the CORR captured home versus in‐centre location for each HD modality
Interventions	HHD Location: home Duration: 3 to 6 hours Frequency: 2 to 4 times/week ICHD Location: facility Duration: 3 to 6 hours Frequency: 2 to 4 times/week Intensive HHD Duration: inclusive of nocturnal HD (> 6 hours, 5 to 6 nights/week) and short daily HD (2 to 3 hours, 5 to 7 days/week) Intensive ICHD Duration: Inclusive of nocturnal HD (> 6 hours, 5 to 6 nights/week) and short daily HD (2 to 3 hours, 5 to 7 days/week)
Outcomes	Outcomes relevant to this review Risk of a Monday/Tuesday admission, adjusted ‐ inclusive of all admissions from index date (OR) Risk of a Monday/Tuesday admission, adjusted ‐ admissions within first year (OR) All‐cause hospital admissions, admissions/1000 patient days, overall Risk of a Monday/Tuesday admission, adjusted ‐ inclusive of all confounding variables (OR) Reported outcomes Primary outcome: All‐cause hospitalisations (rate and day of the week) Three pre‐specified secondary analyses: admissions for cardiovascular causes, dialysis‐associated infections and all infections using ICD‐10 codes mapped to definitions previously established using ICD‐9 codes
Identification	Additional information Sponsorship source: None Authors name: Karthik Tennankore Institution: Dalhousie University, Nova Scotia Health Authority, Halifax, NS, Canada Email: ktennankore@gmail.com Address: Division of Nephrology, Department of Medicine, Dalhousie University, Nova Scotia Health Authority, 5820 University Avenue, Room 5070 Dickson Center B3H 1V8 Other authors: Annie‐Claire Nadeau‐Fredette, Kara Matheson, Christopher Chan, Emilie Trinh, and Jeffrey Perl
Notes	Contacted author to obtain more detailed outcome data and measures of variability, awaiting response from statistician

Toronto Group 2002.

Study characteristics
Methods	Chan 2005 Study design: retrospective cohort study Study grouping: parallel Statistical analysis: Mann‐Whitney U‐test was used for comparison of continuous variables between two groups Inclusion criteria: Age and gender‐matched patients were studied: ESKD patients on ICHD and ESKD patients on nocturnal HD Nocturnal HD group represented patients who had been on this mode of dialysis for a minimum duration of 1 year Exclusion criteria: none of the patients had any acute illness or symptomatic cardiovascular disease (including congestive failure and acute coronary syndrome) Bergman 2008 Study design: prospective cohort study Study grouping: parallel For each NHD patient (study group): 1‐2 control subject(s), matching for age (± 5 years), gender, ESKD vintage (± 2 years), status of diabetes and controlled for comorbidities as assessed by the modified Charlson index Inclusion criteria Study group: consecutive patients who were converted to nocturnal HHD at the University Health Network between 1999‐2003 Control group: drawn from the 302 patients undergoing conventional HD at the University Health Network or one of its self‐care satellite dialysis centres Cafazzo 2009 Study design cross‐sectional survey of prevalent HD patients Two attempts were made to survey eligible patient cohorts over a 3‐month period ICHD patients were provided the survey in‐centre, whereas nocturnal HD patients received the survey by mail Survey consisted of 122 questions SF‐12 to determine general self‐perceived health Modified Appraisal of Self‐Care Agency subscale to measure a patients’ ability for self‐care Spielberger State‐Trait Anxiety Inventory for Adults as a measure of trait anxiety Multidimensional Scale of Perceived Social Support to measure the perceived social support from patients’ family and friends Specific questions related to nocturnal HD included the patients’ perceptions of nocturnal HHD, their likelihood of nocturnal HD adoption, and their perceived barriers to adoption Inclusion criteria HD patients from the University Health Network in Toronto Required to have a permanent address and a working knowledge of English (intent was to elicit a response from a “typical” ICHD population being considered for noctural HHD) Exclusion criteria Medical contraindications to nocturnal HHD Life expectancy < 6 months Physical and/or visual impairments limiting the ability to conduct home HD Mental or psychiatric diagnoses that would prevent them from living independently Qualitative ethnographic interviews Purposive sampling method for recruitment with specific criteria for inclusion; sought “typical” patients who could be found in a non‐randomised qualitative study The principal investigator for the study prepared a roster of recommended patients A series of ethnographic interviews of 3 ESKD patient groups was undertaken: nocturnal HD patients, ICHD patients, and predialysis patients The interview guide was structured around the concepts of the Health Belief Model using the ethnographic interview approach. Lines of questioning were around the informant’s: Perceived susceptibility to and severity of their disease condition Overall perceived threat to their health Perceived benefits and barriers to adopting nocturnal HD Level of self‐efficacy in performing nocturnal HHD Process for taking action External factors that affect their behaviour or decision‐making Study characteristics Country: Canada Setting: Toronto General Hospital, University Health Network, Toronto
Participants	Baseline characteristics: Chan 2005 HHD Mean age: 42 years Male: 70% Primary disease Diabetes (20%); GN (30%); hypertension (30%); congenital (30%) ICHD Mean age: 41 years Male: 67% Primary disease Diabetes (25%); GN (17%); hypertension (33%); congenital (25%) Baseline characteristics: Bergman 2008 HHD Mean age: 43 years Male: 60% Dialysis vintage: 9 years Charlson Comorbidity Index: 3.2 Primary disease Diabetes (6%); hypertension (3%); GN (41%); PKD (22%); reflux/obstruction (9%); other (13%) ICHD Mean age: 44 years Male: 64% Dialysis vintage: 8 years Charlson Comorbidity Index: 3.21 Primary disease Diabetes (10%); hypertension (10%); GN (45%); PKD (0%); reflux/obstruction (4%); other (24%) Group differences Nocturnal HHD (n=32) vs conventional HD (n=42) No differences in age, gender distribution, dialysis vintage, follow‐up period, distribution of ESKD etiologies or comorbidity indices‐ Biochemical indices (Hb, ferritin, potassium, bicarbonate, pre‐dialysis creatinine, albumin, calcium, phosphate, parathyroid hormone) similar between the cohorts at baseline Baseline characteristics: Cafazzo 2009 HHD Mean age: 47.0 years Male, %: 60% Diabetes: 12.5% Hypertension: 50% Cardiac disease: 10.7% Cancer: 7.1% ICHD Mean age: 55.4 years Male: 56.3% Diabetes: 31.4% Hypertension: 57.5% Cardiac disease: 22.9% Cancer: 6.5% Group differences Nocturnal HHD patients were younger than ICHD patients No significant differences with respect to gender or educational attainment Compared with ICHD patients, the majority of nocturnal HHD patients were nonimmigrant (60.7% versus 35.8%, P = 0.01) Nocturnal HHD patients had lower prevalence of diabetes and tended to have lower burden of heart disease Prevalence of hypertension and cancer were similar
Interventions	Chan 2005 HHD Location: home Duration: 6 to 8 hours Frequency: 5 to 6 times/week ICHD Location: in‐centre Duration: 4 hours Frequency: 3 times/week Bergman 2008 Nocturnal HHD Location: home Operator: patient Training: 4 hours, 3 times/week during training Regimen 8 to 10 hours, 5 to 6 nights/week Blood flow rates of 200 to 300 mL/min Dialysate flow of 350 mL/min F80 polysulfone dialysers (Fresenius) or Polyflux (polyamide) dialysers (Gambro) Mean follow‐up ± SD: 26 ± 3 months ICHD Location: University Health Network or one of its self‐care satellite dialysis centres Operator: nurse Training: none Regimen 4 hours, 3 times per week Blood flow rate prescribed at 400 mL/min and maximized at the nurses' discretion Dialysate flow rate 500 to 750 mL/min F80 polysulfone dialysers (Fresenius) or Polyflux (polyamide) dialysers (Gambro) Mean follow‐up ± SD: 23 ± 3 months Cafazzo 2009 Nocturnal HHD Location: home Operator: patient Training: not specified Regimen: 6 to 8 hours, 4 to 6 nights/week ICHD Location: hospital Operator: nurse Training: none Regimen: 4 hours, 3 times/week
Outcomes	Chan 2005 Outcomes relevant to this review Systolic BP (mm Hg) Diastolic BP (mm Hg) Mean BP (mm Hg) LVMI (g/m2) Reported outcomes Primary outcome: differences in endothelial progenitor cell number and migratory function among normal controls, ICHD, and nocturnal HHD patients LVMI, BP, predialysis urea, and other dialysis‐related biochemical parameters were also compared between ICHD and nocturnal HHD cohorts Bergman 2008 Outcomes relevant to this review Dialysis or cardiovascular‐related admissions (days per year) Dialysis or cardiovascular‐related admissions (per patient‐year) All‐cause hospitalisation (per patient‐year) Duration of all‐cause hospitalisation (per patient‐year) Emergency department visits (per patient‐year) Reported outcomes Primary outcome: composite endpoint of dialysis or cardiovascular‐related admissions rate 1 year before and 2 years after conversion to nocturnal HHD vs ICHD controls during the same time period Secondary outcomes: Within‐subject and between‐subject differences in all‐cause hospitalization rate; visits to emergency room; reasons and duration of hospitalization; dialysis‐related biochemical parameters Cafazzo 2009 Outcomes relevant to this review Physical Component Summary (SF‐12) Mental Component Summary (SF‐12) Modified Appraisal of Self‐Care Agency (ASA) ‐ ability for self‐care Multidimensional Scale of Perceived Social Support Anxiety State (Spielberger State‐Trait Anxiety Inventory for Adults) Anxiety Trait (Spielberger State‐Trait Anxiety Inventory for Adults)
Identification	Additional information: Chan 2005 Sponsorship source: This study was supported by the Heart and Stroke Foundation of Canada (Operating Grant NA 5571). Authors name: Christopher T. Chan Institution: Toronto General Hospital, Toronto Email: christopher.chan@uhn.on.ca Address: C. T. Chan, Toronto General Hospital, 200 Elizabeth St., 8N‐842, Toronto, Ontario, Canada M5G 2C4 Other authors: Shu Hong Li and Subodh Verma Additional information: Bergman 2008 Sponsorship source: Heart and Stroke Foundation Operating Grant in Aid (NA 5571); Physician Services Incorporated Foundation (PSI ‐‐ 06‐30) Comments: Toronto NHD Authors name: A. Bergman Institution: Department of Medicine, Division of Nephrology, The Toronto General Hospital ‐ University Health Network, University of Toronto Email: christopher.chan@uhn.on.ca Address: Christopher T. Chan, 200 Elizabeth Street, 8N room 842, Toronto, ON, M5G 2C4, Canada Additional information: Cafazzo 2009 Sponsorship source: Bell University Labs at the University of Toronto ‐‐ Grant funding (87285) Comments: Toronto NHD Authors name: Joseph Cafazzo Institution: Toronto General Hospital, University Health Network Email: joe.cafazzo@uhn.on.ca Address: Joseph Cafazzo, Centre for Global eHealth innovation, Toronto General Hospital, 190 Elizabeth Street, 4th Floor, R. Fraser Elliott Building Other authors: Kevin Leonard, Anthony C. Easty, Peter G. Rossos and Christopher T. Chan
Notes	Patient recruitment 66 nocturnal HHD patients, 56 nocturnal HHD responded (85%) 290 ICHD patients, 199 eligible ICHD patients, 153 ICHD responded (77%) Studies conducted within nocturnal HHD program and University Health Network, where patients receiving ICHD were trained to perform nocturnal HHD Data from Chan 2005 analysed for BP and LVMI outcome Data from Bergman 2008 analysed for hospitalisation outcome Data from Cafazzo 2009 analysed for quality of life outcomes. Data from other publications in study group not analysed due to potentially overlapping populations for the same outcome

Van Oosten 2018.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: Parallel Primary exposure: KRT modality ICHD HHD CAPD APD Multiple dialysis modalities in a year (Mix group) Kidney transplant from living donor Kidney transplant from deceased donor Statistical analysis To accommodate for differences in total treatment time between incident, full year on dialysis and deceased patients, we calculated the cost of 4 weeks of treatment (4‐week costs) per treatment state, as the sum of yearly costs divided by total treatment time (TTT) in days * 28 days. Validated the number of patients per modality in an external database, the Dutch Renal Registry (Renine) Study characteristics Country: Netherlands Setting: Health insurance claims analysis Inclusion criteria: ≥ 19 years in the Vektis claims database who had at least one health insurance claim related to KRT; on chronic KRT; dialysis patients were selected using health claims in the year 2014 and kidney transplant patients were identified using claims in the period January 1, 2012 to December 31, 2014 Exclusion criteria: incidental KRT use (e.g. acute dialysis); patients with unjustified (erroneous) claims
Participants	Baseline characteristics ICHD Mean age: 69.6 years Male: 59% Mean number of comorbidities: 1.0 HHD Mean age: 58.3 years Male: 65% Mean number of comorbidities: 0.7 Definitions Costs directly related to KRT: based on diagnosis‐related group codes (DRG’s), were identified and included all costs of the dialysis procedure (including access surgery and hospitalisation for access surgery), the kidney transplant (including donor expenses) as well as the pre‐ and post‐transplant care. KRT‐related costs include all medications used during dialysis (e.g.EPO, phosphate binders), staff costs, including physician fees, laboratory assessments and other diagnostics as included in KRT clinical guidelines (e.g. chest X‐ray). Also, equipment and devices needed, e.g. dialysis machines for home dialysis, are included. Finally, overhead costs, e.g. for water and energy are included Non‐KRT costs: remaining in‐ and outpatient DRG costs not directly related to KRT, such as primary care, mental healthcare, medication, medical devices, transportation, healthcare costs incurred abroad and other healthcare costs. Note that these non‐KRT costs may incur dialysis‐related costs as well, e.g. transportation costs to and from the dialysis center, but these costs cannot be attributed with 100% certainty to KRT. Definition of comorbidities Respiratory diseases: Asthma and COPD Auto‐immune diseases: Crohn's disease/Colitis Ulcerosa, psoriasis and rheumatism Cystic fibrosis/pancreatitis Diabetes type I and II Neuropathic pain Mental disorders: Psychosis, Alzheimer's and addiction and depression Epilepsy Glaucoma Heart disorder HIV/AIDS Cancer: Hormone‐sensitive tumours and metastasis Brain and Spinal cord disorders and injuries and multiple sclerosis Parkinson's disease Pulmonary (arterial) hypertension Thyroid disorders
Interventions	ICHD HHD
Outcomes	Outcomes relevant to this review Annual KRT costs (EUR) Annual healthcare costs (EUR) Reported outcomes Primary outcome: KRT‐related and KRT‐unrelated average annual healthcare costs
Identification	Additional information Sponsorship source: This work was financed by a grant from the Dutch Kidney Foundation to National Institute of Public Health and the Environment and the Institute of Medical Technology Assessment of Erasmus University Rotterdam Comments: Authors name: Sigrid M. Mohnen Institution: National Institute of Public Health and the Environment Email: sigrid.mohnen@rivm.nl Other authors: Manon J. M. van Oosten, Jeanine Los, Martijn J. H. Leegte, Kitty J. Jager, Marc H. Hemmelder, Susan J. J. Logtenberg, Vianda S. Stel, Leona Hakkaart‐van Roijen, G. Ardine de Wit
Notes	Abstract published in 2018 (van Oosten). Full paper published in 2019 (Mohnen and van Oosten as equal first authors).

Watanabe 2014.

Study characteristics
Methods	Study design Cross‐sectional Study grouping: parallel Study characteristics Country: Japan Setting: Single centre (Saitama Medical University Hospital) Inclusion criteria: prevalent HD patients at Saitama Medical University Hospital from January to April 2011, < 75 years, could perform activities of daily living independently and fill out the questionnaire independently Exclusion criteria: Started HD therapy or changed HD modality within the previous 3 months
Participants	Baseline characteristics HHD Mean age: 54.0 years Male: 87% Time on dialysis: 6.4 years Time on current therapy: 2.7 years Primary disease Diabetes (17.4%); GN (43.5%): hypertension (7.4%); other (21.7%) Comorbidities IHD (2.2%); CHF (4.3%); atrial fibrillation (4.3%); hypertension (84.8%); PVD (0%); diabetes (17.4%) KRT New to dialysis (30.4%); prior conventional HD (34.8%); prior kidney transplant (0%); prior PD or hybrid therapy (34.8%) ICHD Mean age: 57.1 years Male: 76.4% Time on dialysis: 7.4 years Time on current therapy: 6.4 years Primary disease Diabetes (20.6%); GN (29.4%): hypertension (32.4%); other (17.6%) Comorbidities IHD (8.8%); CHF (5.9%); atrial fibrillation (2.9%); hypertension (73.5%); PVD (2.9%); diabetes (20.6%) KRT New to dialysis (85.3%); prior conventional HD (NA); prior kidney transplant (5.9%); prior PD or hybrid therapy (8.8%)
Interventions	HHD Location: home; using the Nikkiso DBB‐27 (Nikkiso Co., Tokyo, Japan) with a water treatment system MH‐500CX (Japan Water System Co., Tokyo, Japan) Operator: patients, with support from caregiver Training: at least 3 months Duration: 3 to 5 hours/session; during the day Frequency: 5 to 6 times/week ICHD Location: facility Duration: 3 to 5 hours/session Frequency: 3 times/week
Outcomes	Outcomes relevant to this review SF‐36 Physical Functioning SF‐36 Role‐Physical SF‐36 Bodily Pain SF‐36 General Health SF‐36 Vitality SF‐36 Social Functioning SF‐36 Role‐Emotional SF‐36 Mental Health Physical Component Scale (SF‐36) Mental Component Scale (SF‐36) Role‐social component scale (SF‐36) Symptoms and Problems (KDQOL) Effects of Kidney Disease (KDQOL) Burden of Kidney Disease (KDQOL) Work status (KDQOL) Cognitive function (KDQOL) Quality of social interaction (KDQOL) Sexual function (KDQOL) Sleep (KDQOL) Reported outcomes Measures: QoL was assessed via the Short Form Health Survey—Version 2 (SF‐36 v2) and the Kidney Disease Quality of Life Instrument—Short Form (KDQOL‐SF)
Identification	Additional information Sponsorship source: None Authors name: Yusuke Watanabe Institution: Saitama Medical University Email: iromichi@saitama‐med.ac.jp Address: H. Suzuki, MD, PhD, Department of Nephrology, Saitama Medical University, 38 Morohongo, Moroyama‐machi, Iruma‐gun, Saitama 350–0495, Japan Other authors: Yoichi Ohno, Tsutomu Inoue, Hiroshi Takane, Hirokazu Okada, Hiromichi Suzuki
Notes

Wong 2019a.

Study characteristics
Methods	Study design Cross‐sectional Study grouping: parallel Data sources: a multicentre cohort study was conducted in 2014–2015, which evaluated the psychometric properties of the KDQOL‐36 in 356 Chinese adults receiving maintenance HD or PD; 43 ESKD patients on HHD selected by convenience sampling were recruited into the second cohort at the time when the patients attended the regular follow‐up consultation at renal specialist outpatient clinics at three public hospitals in Hong Kong, between May 2016 and October 2016 Primary exposure: Dialysis modality (PD vs ICHD vs HHD vs community in‐centre HD (Satellite HD)) Statistical analysis Unadjusted analyses on the HRQOL and health utility scores were tested using one‐way ANOVA with Tukey’s Post‐hoc test for multiple comparisons. Adjusted linear regression models were performed to estimate the differences in HRQOL and health utility scores between HHD and other modes of dialysis, accounting for the sociodemographic data and laboratory results collected at baseline assessment of two cohort studies Sociodemographic characteristics, including gender, age, education level, marital status and working status, were adjusted for confounders Laboratory results included blood haemoglobin, albumin, serum calcium, phosphorus, urea, low‐density lipoprotein–cholesterol, fasting glucose, SCr and eGFR Clinical characteristics, including duration of ESKD diagnosis and the number of dialysis sessions/week, were not adjusted in regression analyses because that information was not fully available in the pooled dataset Study characteristics Country: Hong Kong Setting: Pooled cohort analysis from 2 previous studies Multicentre cohort study in 2014‐2015 in Chinese adults receiving maintenance HD or PD Home HD patients selected by convenience sampling at 3 public hospitals Inclusion criteria Multicentre cohort: undergoing PD, hospital in‐centre HD or community in‐centre HD; some patients undergoing HD in community centres were recruited from a patient‐reported outcome survey when evaluating the quality of care of the HD public‐private partnership (HDPE) programme; details of patient recruitment and data collection for the first cohort have been reported elsewhere Home HD cohort: ESKD patients on Home HD selected by convenience sampling recruited at the time when the patients attended the regular follow‐up consultation at renal specialist outpatient clinics at three public hospitals in Hong Kong between May 2016 and October 2016 Exclusion criteria: not reported
Participants	ICHD Mean age: 56.4 years Male: 57% Education No formal schooling (8.1%); primary (31.1%); secondary (51.9%); tertiary (8.9%) Married: 59.8% Working: 16.3% Mean duration ESKD: NA Mean treatment sessions/week: 2.3 HHD Mean age: 47.9 years Male: 67.4% Education No formal schooling (0%); primary (4.7%); secondary (72.1%); tertiary (23.3%) Married: 69.8% Working: 60.5% Mean duration ESKD: 7.5 years Mean treatment sessions/week: 3.7 Satellite HD Mean age: 56.8 years Male: 66.1% Education No formal schooling (6.8%); primary (26.3%); secondary (54.2%); tertiary (12.7%) Married: 58.5% Working: 14.5% Mean duration ESKD: NA Mean treatment sessions/week: 2.5
Interventions	ICHD Location: hospital HHD Location: home Operator: patient Duration: nocturnal Frequency: 3 to 6 sessions/week Satellite HD Location: satellite centre
Outcomes	Outcomes relevant to this review SF‐12 Physical Functioning SF‐12 Role‐Physical SF‐12 Bodily Pain SF‐12 General Health SF‐12 Vitality SF‐12 Social Functioning SF‐12 Role‐Emotional SF‐12 Mental Health Physical Component Summary (SF‐12) Mental Component Summary (SF‐12) SF‐6D Reported outcomes Primary outcome: HRQOL using the Chinese (Hong Kong) version SF‐12 version 2 (SF‐12v2)
Identification	Additional information Sponsorship source: This work was supported by Small Project Funding (Ref. no:201409176109), the University of Hong Kong, the Health and Medical Research Fund (Ref. no: 13142451) and the Commissioned Research on Enhanced Primary Care Study(Ref. no EPC‐HKU‐2), Food and Health Bureau, HKSAR Authors name: Carlos K.H. Wong Institution: Department of Family Medicine and Primary Care, The University of Hong Kong. Email: carlosho@hku.hk Address: Dr Carlos KH Wong, Department of Family Medicine and Primary Care, The University of Hong Kong. Rm 1‐01, 1/F, Jockey Club Building for Interdisciplinary Research, 5 Sassoon Road, Pokfulam, Hong Kong SAR, China. Other authors: Julie Y. Chen, Samuel K.S. Fung, Wai Kei Lo, Sing Leung Lui, Tak Mao Chan, Yuk Lun Cheng, Irene Kong, Eric Y.F. Wan and Cindy L.K. Lam
Notes

Wong 2019b.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: Parallel Primary exposure: dialysis modality (PD, ICHD, nocturnal HHD) Primary outcome: Cost (direct and indirect) in Hong Kong Dollars (HKD) Statistical analysis Annual usage of health service resources, in terms of general outpatient clinic (GOPC), specialist outpatient clinic (SOPC), hospitalisations and accident and emergency (A&E) visits, under three dialysis modalities was compared. Count data models such as multiple Poisson regressions or negative binomial regressions were performed to determine the effect of dialysis modality on health service utilisation, controlling for sociodemographic and clinical characteristics, including the Charlson comorbidity index. Negative binomial regression was used to encounter the overdispersed count data of annual GOPC, SOPC, hospitalisation and A&E attendance rate. Mean direct medical costs, direct non‐medical costs, indirect costs and all healthcare and societal costs of ESKD patients undergoing three dialysis modalities were compared through oneway analysis of variance and multiple comparisons by the least significant difference method Linear regressions were performed to estimate the effect of gender, age, comorbidities and primary ESKD cause on the total healthcare and societal costs in the initial and second years by dialysis modalities, respectively. Total annual healthcare and societal costs were stratified with regards to subgroups of gender, age groups (< 55/55+ years), comorbidity (Charlson comorbidity index < 6/6+) and primary cause of ESKD (glomerulonephritis/others or unknown). Mean annual healthcare and societal costs (and 95% CIs) in the initial and second years of patients on PD, hospital HD and home HD per subgroup were calculated Study characteristics Country: Hong Kong Setting: 3 hospitals in the Hospital Authority (Princess Margaret Hospital, Tung Wah Hospital and Queen Mary Hospital) Inclusion criteria: All ESKD patients undergoing KRT in the renal specialist clinic or renal unit of three hospitals in the Hospital Authority (Princess Margaret Hospital, Tung Wah Hospital and Queen Mary Hospital) were invited. Subjects were included in the study if they satisfied all of the following criteria: ≥ 18 years; life expectancy of at least 6 months; confirmed diagnosis of ESKD and classified with stages 4–5 CKD according to the KDIQO guideline; starting dialysis on or before year 2015 and, under hospital‐based HD for at least 1 year OR under nocturnal home HD for at least 1 year OR under PD for at least 1 year AND given informed consent to take part in the study Exclusion criteria: patients who did not complete the questionnaire were excluded from the study
Participants	Baseline characteristics HHD Mean age: 47.9 years Male: 67.4% Mean education No formal school (0%): primary school (4.7%); secondary school (72.1%); tertiary school (23.3%) Married: 69.8% Monthly cost of medication related to ESKD, mean (HKD): 214.7 Weekly expenditure in transportation to hospitals, mean (HKD): 22.1 Duration of ESKD diagnosis, mean: 7.5 years Primary disease Diabetes (4.7%); hypertension (4.7%); GN (30.2%); PKD (16.3%); lupus (2.3%); Alport's (0%); unknown (41.9%) Kidney transplant: 18.6% Working status: 60.5% Charlson comorbidity index: 4.0 ICHD Mean age: 60.8 years Male: 50.6% Mean education No formal school (10.0%): primary school (32.9%); secondary school (44.1%); tertiary school (12.9%) Married: 68.2% Monthly cost of medication related to ESKD, mean (HKD): 269.3 Weekly expenditure in transportation to hospitals, mean (HKD): 32.4 Duration of ESKD diagnosis, mean: 8.3 years Primary disease Diabetes (18.6%); hypertension (4.8%); GN (18.6%); PKD (7.2%); lupus (3.6%); Alport's (1.2%); unknown (45.5%) Kidney transplant: 13.2% Working status: 14.1% Charlson comorbidity index: 6.2
Interventions	HHD ICHD
Outcomes	Outcomes relevant to this review Direct cost, first year (HKD) Direct cost, second year (HKD) Indirect cost, yearly (HKD) Societal cost, first year (HKD) Societal cost, second year (HKD) Healthcare provider cost, first year (HKD) Healthcare provider cost, second year (HKD) Annual hospitalisation utilisation (IRR)
Identification	Additional information Sponsorship source: This study was supported by the Health and Medical Research Fund (grant number 13142451), Food and Health Bureau, Government of the Hong Kong SAR Authors name: Carlos K.H. Wong Institution: Department of Family Medicine and Primary Care, The University of Hong Kong Email: carlosho@hku.hk Other authors: Julie Chen, Samuel K.S. Fung, Maggie M.Y. Mok, Yuk Lun Cheng, Irene Kong, Wai Kei Lo, Sing Leung Lui, Tak Mao Chan and Cindy L.K. Lam
Notes

Wright 2015.

Study characteristics
Methods	Study design Cross sectional Study grouping: parallel Methods Participants were recruited using flyers in each dialysis facility Patients on ICHD received the instruments at the time of their HD treatment. Patients on PD and HHD received the instruments at the time of their monthly clinic visit. Patients were given the option of completing the instruments at the dialysis unit or at another location of their choice Completed instruments were returned in a sealed envelope, either via a collection box in each dialysis unit or via U.S. mail using a stamped/addressed envelope provided with the study instruments Primary exposure: Dialysis modality (ICHD vs HHD vs PD) Statistical analysis Primary data analyses consisted of a series of one‐way ANOVAs Study characteristics Country: USA Setting: Four outpatient dialysis facilities located in Philadelphia, Pennsylvania, and the surrounding environs Inclusion criteria: At least 18 years, but not older than 89; community‐dwelling; no physical self‐care limitations; receiving some form of KRT at least six months; dialysis prescription consisting of one of the following: At least 3 ICHD treatments/week At least 3 HHD treatments/week Daily PD Exclusion criteria: did not complete study instruments in their entirety
Participants	Baseline characteristics HHD Male: 59% Race White (64%); African American (36%); Asian (0%); Hispanic (0%); other (0%) Age range 20 to 29 years (0%); 30 to 39 years (18%); 40 to 49 years (27%); 50 to 59 years (9%); 60 to 69 years (23%); 70 to 79 years (5%); 80 to 89 years (9%) Time on dialysis 6 to 12 months (32%); 1 to 5 years (36%); 5 to 10 years (27%); 10 to 20 years (0%); > 20 years (5%) Education < 8th grade (0%); < 4 years high school (5%); high school graduate (18%); < 4 years college (27%); college graduate (18%); graduate school (32%) Employment status Retired (23%); disability (32%); unemployed (0%); full‐time (14%); part‐time (27%); home maker (5%); medical leave (0%) ICHD Male: 55% Race White (14%); African American (36%); Asian (7%); Hispanic (3%); other (7%) Age range 20 to 29 years (3%); 30 to 39 years (34%); 40 to 49 years (14%); 50 to 59 years (17%); 60 to 69 years (17%); 70 to 79 years (3%); 80 to 89 years (3%) Time on dialysis 6 to 12 months (10%); 1 to 5 years (59%); 5 to 10 years (31%); 10 to 20 years (7%); > 20 years (3%) Education < 8th grade (3%); < 4 years high school (3%); high school graduate (48%); < 4 years college (31%); college graduate (10%); graduate school (3%) Employment status Retired (14%); disability (41%); unemployed (17%); full‐time (7%); part‐time (14%); home maker (0%); medical leave (7%) Group differences There was a statistically significant difference between the groups in terms of race/ethnicity. There were proportionately more African‐American patients in the HD group and more White patients in the HHD group.
Interventions	HHD ICHD
Outcomes	Outcomes relevant to this review SF‐36 Physical Functioning SF‐36 Role‐Physical SF‐36 Bodily Pain SF‐36 General Health SF‐36 Vitality SF‐36 Social Functioning SF‐36 Role‐Emotional SF‐36 Mental Health Physical Component Summary (SF‐12) Mental Component Summary (SF‐12) SUPPH positive attitude SUPPH stress reduction SUPPH decision making Symptoms and Problems (KDQOL) Effects of Kidney Disease (KDQOL) Burden of Kidney Disease (KDQOL) Work status (KDQOL) Cognitive function (KDQOL) Quality of social interaction (KDQOL) Sexual function (KDQOL) Sleep (KDQOL) Social support (KDQOL) Dialysis staff encouragement (KDQOL) Overall health (KDQOL) Patient satisfaction (KDQOL) Physical functioning (KDQOL) Role limitations ‐ Physical (KDQOL) Pain (KDQOL) General health (KDQOL) Emotional well‐being (KDQOL) Role limitations ‐ Emotional Social function (KDQOL) Energy/fatigue (KDQOL) Reported outcomes Primary outcome: Quality of life (KDQOL‐SF 1.3) General health status was also assessed via the first two questions contained in the Kidney Disease and Quality of Life instrument Self‐efficacy was measured using the Strategies Used by People to Promote Health (SUPPH) questionnaire
Identification	Additional information Sponsorship source: None reported Authors name: Linda S. Wright Institution: Thomas Jefferson University Hospital, Philadelphia, PA Email: linda.wright@jefferson.edu Other authors: Linda Wilson
Notes

Xue 2015.

Study characteristics
Methods	Study design Prospective cohort study Study grouping: parallel Matching: NHHD patients were matched to up to two eligible ICHD patients on these bases. Fifty‐eight NHHD patients were matched with 2 ICHD patients while 5 NHHD patients could be matched with only 1 ICHD patient, resulting in 63 NHHD patients and 121 ICHD patients (2×58+1×5 = 121) included in the analysis. Matching considered five variables: Age (± 5 years) Gender Race Dialysis vintage (seven categories: 1 day, > 1 to 30 days, > 1 to 3 months, > 3 to 12 months, > 1 to 2 years, > 2 to 5 years, and > 5 years) Diabetes Follow‐up: Follow‐up was censored at 20 months for the first catheter because there were only 3 catheters left in place in the control group at 20 months Study characteristics Country: USA Setting: NHHD program from 1997 to 2010 at University of Virginia Lynchburg Dialysis Facility in Virginia & IHD patients admitted to Fresenius Medical Care North America facilities in Virginia and surrounding states from January 1, 2007 to December 31, 2010 Inclusion criteria: 63 patients from the NHHD program from 1997 to 2010 at University of Virginia Lynchburg Dialysis Facility in Virginia; 6285 ICHD patients admitted to Fresenius Medical Care North America facilities in Virginia and surrounding states from 1 January 2007 to 31 December 2010
Participants	Baseline characteristics HHD Mean age ± SD: 52.8 ± 14.7 years Male: 57.1% Mean dialysis vintage ± SD: 38.5 ± 49.6 months Race White (57.1%); Black (42.9%) Comorbidities Diabetes (20.6%); hypertension (20.6%); GN (19.0%); PKD (6.3%) ICHD Mean age ± SD: 53.8 ± 14.1 years Male: 57.9% Mean dialysis vintage ± SD: 37.0 ± 46.1 months Race White (57.0%); Black (43%) Comorbidities Diabetes (21.5%); hypertension (41.3%); GN (9.1%); PKD (3.3%) Group differences 63 NHHD patients matched with 121 IHD patients Similar in age, dialysis vintage, gender, race and diabetes status (i.e., matching variables) Hypertension more common in the ICHD group (41.3%) vs NHHD group (20.6%; P = 0.005) 19.0% of NHHD patients that had GN vs only 9.1% of IHD patients (P = 0.05) No significant difference in the frequency of polycystic kidney disease between groups
Interventions	Intensive HHD ICHD
Outcomes	Outcomes relevant to this review Catheter‐related sepsis (first catheter), per 100 patient‐months Notes: For first catheter used Catheter‐related sepsis (all catheters), per 100 patient‐months Death rate (first catheter), per 100 patient‐months Death rate (all catheters), per 100 patient‐months HR for catheter‐related sepsis (first catheter), unadjusted Notes: CIs shown in Figure 3 but values not provided HR for catheter‐related sepsis (first catheter), adjusted Notes: adjusted for age, hypertension and GN CIs shown in Figure 3 but values not provided HR for catheter‐related sepsis (all catheters), unadjusted Notes: CIs shown in Figure 3 but values not provided HR for catheter‐related sepsis (all catheters), adjusted Notes: adjusted for age, hypertension and GN CIs shown in Figure 3 but values not provided Reported outcomes Primary outcome: combined incidence of catheter‐related bacteraemia (i.e., positive blood culture) and/or a clinical diagnosis of sepsis (primary ICD‐9 codes 790.7 and 995.91 and other related codes for sepsis) from a hospitalisation episode
Identification	Additional information Sponsorship source: No external funding. As of publication writing, authors N‐CL and EL are employees of Fresenius Medical Care, North America. SMB is an employee of Davita and has served as an advisor to Amgen, C.B. Fleet Company and Proctor & Gamble. SMB has received speaking honoria from Fresenius Medical Care North America, and his spouse is employed by Astra Zeneca. RSL sits on the Machine Medical Advisory Board for Fresenius Medical Care North America (Sorbent Machine). HX has nothing to disclose Authors name: Hui Xue Institution: Kaiser Medical Center, San Diego, California, USA Email: hui.shue@gmail.com Address: Division of Nephrology, Department of Medicine, Kaiser Medical Center, 4510 Viewridge Ave, San Diego, CA 92123, U.S.A. Other authors: Nien‐Chen Li, Eduardo Lacson Jr, Steven M. Brunelli, Robert S. Lockridge
Notes

Yeung 2021.

Study characteristics
Methods	Study design Prospective cohort study Study grouping: parallel Study characteristics Country: Australia Setting: home and satellite HD patients in a tertiary centre Inclusion criteria Incident adult (≥ 18 years) HHD patients treated at a tertiary centre from 1 January 2000 to 30 June 2017, who had biochemical data available within 12 months of commencing HHD Patients were included if they commenced or returned to dialysis after a failed kidney transplant on or after 1 January 2000 and completed HHD training within 90 days after the commencement of the training programme without modality change Controls included all facility HD patients from a tertiary centre who commenced dialysis on or after 1 Janurary 2000 to 30 June 2017 and were never treated with HHD, and were matched 3:1 to each HHD patient by age (within 5 years), gender and cause of ESKD
Participants	Baseline characteristics HHD Mean age: 48.9 years Male: 76.2% Race Caucasian (79.6%); ATSI (0%); Asian (11.0%); Pacific (4.4%); Maori (1.1%); other (3.3%) Ever smoker: 41.7% Late referral: 16.1% Primary cause of ESKD GN (42%); diabetes (20.4%); hypertension (6.1%); PKD (12.2%); reflux (6.6%); other (11.5%) Mean BMI: 30.6 kg/m2 Comorbidities Chronic lung disease (6.2%); vascular disease (24.0%); diabetes (30.9%) ICHD Mean age: 49.5 years Male: 71.9% Race Caucasian (71.4%); ATSI (1.4%); Asian (17.2%); Pacific (17.2%); Maori (0.7%); other (5.3%) Ever smoker: 47.0% Late referral: 21.2% Primary cause of ESKD GN (36.3%); diabetes (24.9%); hypertension (7.3%); PKD (9.0%); reflux (5.6%); other (16.9%) Mean BMI: 27.6 kg/m2 Comorbidities Chronic lung disease (6.29.0); vascular disease (37.5%); diabetes (32.0%) Group differences No significant between‐group differences in average age, gender, race, smoking status, late referral, or cause of ESKD BMI significantly higher in HHD patients (30.6kg/m2 vs 27.2kg/m2 (P < 0.001)) Significantly higher prevalence of vascular disease at baseline in ICHD vs HHD patients (P = 0.001) No differences in prevalence of chronic lung disease or diabetes
Interventions	HHD Location: home Duration: 6 to 8 hours/session Frequency: alternate days Blood flow rate: 200 to 250 mL/min Dialysate flow rate: 300 mL/min ICHD Location: facility Duration: 4 to 5 hours/session Frequency: 3 times/week Blood flow rate: 300 to 350 mL/min Dialysate flow rate: 500 mL/min
Outcomes	Outcomes relevant to this review HR for survival (HHD vs matched ICHD): unadjusted model, adjusted model, competing risk model Transplantation rate Reported outcomes Reported outcomes Primary outcomes: all‐cause mortality (censored for kidney transplantation); kidney transplantation Secondary outcomes: average biochemical levels (Hb, corrected calcium, phosphate and PTH); proportion of patients in each group with a Hb < 100 g/L and phosphate >1.8 mmol/L; graft survival 6 months post‐transplant
Identification	Additional information Sponsorship source: none stated Authors name: E Yeung Institution: Monash Medical Centre, Medicine, Clayton, Australia Email: Not available Address: Not available Other authors: P Kerr, K Polkinghorne
Notes

Zimbudzi 2014.

Study characteristics
Methods	Study design Retrospective cohort study Study grouping: parallel Sampling: "Randomised" sampling to obtain 25 patients per group. Patients were randomly selected from the HHD list and from the Category 1 satellite HD by a staff member who was not involved in the project Study characteristics Country: Australia Setting: Major referral metropolitan public hospital (Monash Medical Centre, Melbourne, Australia) Inclusion criteria: 25 patients per group were randomly selected from the HHD list and from the Cat1 satellite HD (satellite HD patients on a Category 1 transplant waiting list) by a staff member who was not involved in the project. From August 2012 to August 2013 at a major referral metropolitan hospital (Monash Medical Centre, Melbourne, Australia) Exclusion criteria: none stated Definitions More than 75% of HHD patients dialysed for 8 hours every second day, while the majority of the satellite patients dialysed for 5 hours, 3 times/week
Participants	Baseline characteristics HHD Mean age (range): 53.6 years (35 to 76) Male: 64% Primary disease Diabetes (16%); GN (4%); PKD (8%); IgA (20%); reflux (16%); Goodpastures (8%); vasculitis (12%); hypertension (8%); other (0%); FSGS (8%) Satellite HD Mean age (range): 47.4 years (25 to 69) Male: 60% Primary disease Diabetes (28%); GN (12%); PKD (12%); IgA (8%); reflux (16%); Goodpastures (0%); vasculitis (4%); hypertension (0%); other (16%); FSGS (4%) Group differences No difference between the mean ages (P = 0.08) and the time they had been on dialysis (P = 0.3)
Interventions	HHD Location: home Operator: patient Training: HHD training unit is strategically located 20 minutes away from the main hospital; clientele scattered around suburban and regional areas. HHD candidates normally dialysed in the in‐center unit or in satellite units before they commence HD training; on rare occasions, new patients Satellite HD Location: Regional satellite units affiliated to the hospital
Outcomes	Outcomes relevant to this review Number of patients hospitalised Hospital admission, days/patient‐year Mean length of stay hospitalisation, days Risk of admission for vascular event, absolute risk Admission due to vascular event, OR Reported outcomes Hospitalisation patterns Mean length of stay in hospital + Days/patient‐year Admissions (admissions/patient‐year) Absolute risk + OR of an admission due to a vascular event Lab values (K, PO4, Hb, albumin, creatinine)
Identification	Additional information Sponsorship source: None stated Authors name: Edward Zimbudzi Institution: Monash Health, Monash Medical Centre, Melbourne, Australia Email: edward.zimbudzi@southernhealth.org.au Address: Department of Nephrology, Monash Health, Monash Medical Centre, Clayton, VIC 3168, Australia Other authors: Reggie Samlero
Notes	Unable to contact author to obtain measures of variability and more detailed outcome data

AE: adverse event; APD: ambulatory peritoneal dialysis; ATSI: Aboriginal and Torres Strait Islander; AVF: arteriovenous fistula; AVG: arteriovenous graft; BDI: Beck Depression Inventory; BMI: body mass index; BP: blood pressure; BUN: blood urea nitrogen; CAS: coronary artery disease; CAPD: continuous ambulatory peritoneal dialysis; CCPD: continuous cycling peritoneal dialysis; CHF: congestive heart failure; CI: confidence interval; CKD: chronic kidney disease; CVC: central venous catheter; ECG: echocardiogram; eGFR: estimated glomerular filtration rate; EPO: erythropoietin; ESA: erythropoietin‐stimulating agent; ESKD: end‐stage kidney disease; ESRD‐SI: End‐Stage Renal Disease Severity Index; FSGS: focal segmental glomerulosclerosis; GN: glomerulonephritis; Hb: haemoglobin; HD: haemodialysis; HHD: home haemodialysis; HR: hazard ratio; HRQoL: health‐related quality of life; ICHD: in‐centre haemodialysis; IEQ: Illness Effects Questionnaire; IHD: ischaemic heart disease; IPQ: Illness Perception Questionnaire; iPTH: intact parathyroid hormone; IQR: interquartile range; ITT: intention‐to‐treat; KDQOL‐36: Kidney Disease Quality of Life‐36 questionnaire; KRT: kidney replacement therapy; Kt/V: dialysis adequacy; LVMI: left ventricular mass index; MAP: mean arterial pressure; NA: not applicable; PD: peritoneal dialysis; PKD: polycystic kidney disease; PVD: peripheral vascular disease; QoL: quality of life; RCT: randomised controlled trial; SCr: serum creatinine; SDHD: short daily haemodialysis; SF: short form; TB: tuberculosis; TEQ: Treatment Effects Questionnaire; VA: vascular access

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Benain 2007	Wrong study design
Benain 2015	Wrong study design
Bernieh 2020	Not HHD vs ICHD
Blagg 2006	Not HHD vs ICHD
Connor 2011	Wrong study design
Derrett 2017	Not HHD vs ICHD
Eneanya 2019	Not HHD vs ICHD
FREEDOM 2009	Not HHD vs ICHD
Gonzalez‐Perez 2005	Wrong study design
Gorham 2019	Wrong study design
Howard 2009	Wrong study design
Ipema 2010	Wrong outcome
Ipema 2012	Wrong outcome
Ipema 2014	Not HHD vs ICHD
Johansen 2009	Not HHD vs ICHD
Komenda 2012	Wrong study design
Kubisiak 2018	Wrong study design
Lee 2017	Not HHD vs ICHD
Lockridge 2011	Wrong study design
London Daily/Nocturnal 2003	Not HHD vs ICHD
Loos‐Ayav 2008	Not HHD vs ICHD
MacRae 2010	Not HHD vs ICHD
Marshall 2015	Not HHD vs ICHD
McDonald 2009	Not HHD vs ICHD
McFarlane 2002	Not HHD vs ICHD
Mehrotra 2016	Not HHD vs ICHD
Mohr 2001	Wrong study design
Quintaliani 2000	Not HHD vs ICHD
Rivara 2018	Not HHD vs ICHD
Seto 2007	Not HHD vs ICHD
Shen 2019	Not HHD vs ICHD
Ting 2003	Not HHD vs ICHD
Walsh 2006	Not HHD vs ICHD
Yang 2015	Not HHD vs ICHD
Yuen 2011	Wrong study design

HHD: home haemodialysis; ICHD: in‐centre haemodialysis

Characteristics of studies awaiting classification [ordered by study ID]

De Smet 2007.

Methods	Study design Open‐label, controlled before‐and‐after study Study duration: 4 weeks Study characteristics Country: Belgium Setting: not stated
Participants	Prevalent stable HD patients receiving standard HD Switched to nocturnal HD (n = 6) Remained on standard HD (n = 8)
Interventions	Standard HD Duration: not reported, 3 times/week Nocturnal HD 8 hours/session, frequency not reported
Outcomes	Biochemical parameters Calcium, phosphorous, urea, eKt/Vurea, urea reduction ratio, normalised protein catabolic rate)
Notes	Abstract‐only publication Contacted author to confirm location of conventional versus nocturnal HD treatments, awaiting response

Fadem 2011.

Methods	Study design Cross‐sectional survey Open invitation on American Association of Kidney Patients website and email invitations to member database Country: USA Setting: online
Participants	Patients (n = 821) and caregivers (n = 56) of patients with kidney failure
Interventions	ICHD HHD (nocturnal, short daily, conventional thrice‐weekly) Kidney transplant PD
Outcomes	Satisfaction with current KRT Satisfaction with education received
Notes	Unable to contact author to ascertain if additional outcome data available

Hanly 2001.

Methods	Study design Prospective cohort study Country: Canada Setting: single centre
Participants	Prevalent adult HD patients who switched from conventional ICHD to nocturnal HHD between November 1993 and November 1998 (n = 14)
Interventions	Conventional ICHD 4 hours/session, 3 times/week Nocturnal HHD 8 to 10 hours/session, nightly
Outcomes	Sleep apnoea
Notes	Unable to contact author to ascertain if additional outcome data available

Helantera 2012.

Methods	Study design Cross‐sectional retrospective cohort study Analysis of national kidney registry data combined with employment statistics from the Finnish government Country: Finland Setting: Finnish Registry for Kidney Diseases
Participants	Prevalent dialysis and kidney transplant patients in Finland aged 15 to 64 years at the end of 2007 (n = 2637)
Interventions	ICHD HHD PD: CAPD; APD Kidney transplant
Outcomes	Employment status
Notes	Contacted author to obtain data for adult (≥ 18 years) subgroup, awaiting response

Kannampuzha 2010.

Methods	Study design Cross‐sectional study Country: Canada Setting: single centre
Participants	Prevalent adult (18 to 75 years) HD patients on their respective modality for a minimum 6 months, who were medically stable Conventional ICHD (n = 15) Nocturnal HHD (n = 15) Healthy controls Hospital staff (n = 15)
Interventions	ICHD 4 hours/session, 3 times/week Nocturnal HHD 6 to 8 hours/session, 5 to 7 times/week
Outcomes	Glutathione, cysteine, homocysteine, C‐reactive protein, haematocrit concentrations Riboflavin status Food diary Anthropometric indices (biceps skinfold, triceps skinfold, subscapular skinfold, supra iliac skinfold, mid‐arm muscle circumference, dry body weight, height, BMI, subjective global assessment)
Notes	Unable to contact author to ascertain if additional outcome data available.

Mitchell 2020.

Methods	Study design Cross‐sectional survey Country: USA Setting: not reported
Participants	Patients with kidney failure (n = 69) and CKD stage 4‐5 (n =25)
Interventions	ICHD HHD Conservative kidney management
Outcomes	Patient Activation Measure (PAM‐13®)
Notes	Abstract‐only publication Contacted author to ascertain if additional outcome data available, awaiting response

Morton 2010.

Methods	Study design Qualitative analysis of semi‐structured interviews Country: Australia Setting: not reported
Participants	Prevalent adult dialysis and transplant patients Satellite HD (n = 52) ICHD (n = 8) Home HD (n = 4) CAPD (n = 8) APD(n = 5) Kidney transplant (n = 18)
Interventions	Satellite HD ICHD Home HD CAPD APD Kidney transplant
Outcomes	Patient views about treatment of stage 5 CKD
Notes	Contacted author to ascertain if additional outcome data available, awaiting response

Painter 2012.

Methods	Study design Open‐label cohort study Study duration: 6 months Country: USA Setting: multi‐centre (5 sites)
Participants	Prevalent conventional HD patients from 3 centres, and transplant recipients from 2 centres Sedentary healthy control subjects recruited from kidney donors (> 1 year post‐donation with normal kidney function), matched to patient groups by age decade, sex and general physical activity
Interventions	Group 1 Patients who were treated with conventional HD (3 to 4 hours, 3 times/week) and did not change modality (n = 13) Group 2 Changed from conventional HD to daily HD (3 hours 5 to 6 days/week, performed at home (n = 10) Group 3 Transplant received a living donor kidney transplant (n = 20) Control Healthy controls (n = 36)
Outcomes	SF‐36 KDQOL
Notes	Contacted author to confirm location of conventional HD treatments, awaiting response

Parker 2014.

Methods	Study design Retrospective cohort study All in‐centre HD,PD and HHD patients were identified from Greater Manchester East sector renal network Data were retrieved from a customized renal database, clinic and discharge letters with cross‐validation from the general practitioner when needed Country: UK Setting: Greater Manchester East sector renal network
Participants	Prevalent dialysis patients Exclusion criteria: on dialysis < 2 months, current inpatient or undertaking HHD training, > 80 years
Interventions	ICHD HHD PD
Outcomes	Medication pill burden
Notes	Contacted author to ascertain if additional outcome data available, awaiting response.

Pellicano 2010.

Methods	Study design Cross‐sectional study Country: Australia Setting: single centre
Participants	Prevalent HHD patients (n = 28) Alternate day HD, 5 hours/session (n = 7) Alternate night nocturnal HD, 8 hours/session (n = 21) Conventional ICHD patients (n = 28) Patient groups matched on age, gender, diabetic status and duration of KRT
Interventions	HHD ICHD
Outcomes	Total body protein Lean body mass Fat mass and distribution Biochemical nutritional and inflammatory parameters (albumin, transthyretin, urea, creatinine, haemoglobin, high sensitivity c‐reactive protein, erythrocyte sedimentation rate, transferrin, phosphate, calcium, total cholesterol and interleukin‐6)
Notes	Contacted author to ascertain if additional outcome data available, awaiting response.

Poon 2015.

Methods	Study design Retrospective cohort study Study duration: 2 years Country: Hong Kong Setting: two centres
Participants	Prevalent adult (18 to 59 years) HD patients who had received their current modality for at least 2 years Exclusion criteria: chronic blood loss, thalassemia, untreated vitamin B12 or folate deficiency, haematological diseases or marrow disorders causing anaemia, active malignancy, chronic infection, chronic inflammatory conditions, or severe hyperparathyroidism (serum parathyroid hormone > 200 pmol/L)
Interventions	Nocturnal HHD 4 to 5 hours/session, alternate days (n = 23) ICHD 4 to 5 hours/session, 2 to 3 sessions/week (n = 25)
Outcomes	Haemoglobin level Erythropoiesis‐stimulating agent requirement
Notes	Contacted author to ascertain if additional outcome data available, awaiting response

Sikkes 2009.

Methods	Study design Prospective non‐randomised cohort study Study duration: 1 year Stable patients on ICHD were followed for 1 year after conversion to nocturnal HHD Country: Netherlands Setting: single centre
Participants	Prevalent adult HD patients who were medically stable at least 3 months who were able to perform nocturnal HHD and agreed to do so, and who lived with a spouse or other partner who could assist them (n = 14)
Interventions	Conventional ICHD Parameters not described Nocturnal HHD 8 hours/session, 6 times/week)
Outcomes	Appetite Body weight Dietary intake Laboratory results (urea, calcium, phosphorus,albumin, potassium, C‐reactive protein parathyroid hormone) Single pool Kt/Vurea Normalised protein catabolic rate
Notes	Contacted author to ascertain if additional outcome data available, awaiting response.

Thomson 2013.

Methods	Study design Retrospective cohort study Country: Canada Setting: single centre
Participants	All patients in the HHD program of the Southwestern Ontario Regional Renal Program, from 1985 to December 31, 2011 All patients were on intermittent conventional HD prior to transition to HHD
Interventions	Conventional ICHD 1.5 to 4 hours/session, ≤ 4 times/week (n = 14) Short daily HD 1.5 to 4 hours/session, ≥ 5 times/week (n = 25) Frequent nocturnal HD ≥ 6 hours/session, ≥ 5 times/week (n = 25) Intermittent nocturnal HD ≥ 6 hours/session, ≤ 4x/week (n = 8)
Outcomes	QTc interval
Notes	Contacted author to ascertain if additional outcome data available, awaiting response

Yong 2014.

Methods	Study design Open‐label, uncontrolled longitudinal study Parameters measured prior to commencement and 3 to 6 months after conversion to quotidian home HD Country: Australia Setting: multi‐centre
Participants	Stable conventional HD patients commencing quotidian HHD between 2012 and 2013 (n = 14)
Interventions	Conventional ICHD Quotidian HHD
Outcomes	Inflammatory markers (IL‐12, IL‐18) Arterial stiffness (augmentation index, pulse wave velocity, pulse pressure)
Notes	Abstract‐only publication Author confirmed data available on systolic and diastolic blood pressure. Requested data from author to include in analysis, awaiting response

APD: automated peritoneal dialysis; BMI: body mass index; CAPD: continuous ambulatory peritoneal dialysis; CKD: chronic kidney disease; HD: haemodialysis; HHD: home haemodialysis; ICHD: in‐centre haemodialysis; KDQOL: Kidney Disease Quality of Life Questionnaire ; KRT: kidney replacemetn therapy; PD: peritoneal dialysis; SF‐36: 36‐Item Short Form Survey

Differences between protocol and review

2024 update

Non‐randomised studies have been included.

We clarified that studies evaluating peritoneal dialysis as a home dialysis modality were eligible for inclusion, where data regarding participants receiving in‐centre and home haemodialysis could be extracted separately. We initially sought to include the clinical outcome of wait‐listing for a kidney transplant; we also elected to include the outcome of receipt of a kidney transplant, as this was reported in a number of studies. Similarly, we initially sought to include pre‐dialysis systolic and diastolic blood pressure (BP) but elected to use a more inclusive BP outcome by also extracting systolic and diastolic BP taken at times other than pre‐dialysis (e.g. clinic measurements) mean arterial pressure (MAP) and pulse pressure.

Contributions of authors

1. Draft the protocol: MC, IE

2. Study selection: MC, IE

3. Extract data from studies: MC, IE

4. Enter data into RevMan: MC, IE

5. Carry out the analysis: MC

6. Interpret the analysis: all authors

7. Draft the final review: MC, IE

8. Review the final review for intellectual content: all authors

9. Disagreement resolution: YC, RK, DJ

10. Update the review: SP, GFMS

Sources of support

Internal sources

• No sources of support provided

External sources

• No sources of support provided

Declarations of interest

• Melissa S Cheetham: has received travel support from Amgen and is a current recipient of a Queensland Advancing Clinical Research Fellowship

• Isabelle Ethier: none known

• Rathika Krishnasamy: has received speaker’s honoraria, consultancy fees, research grants and travel support from Baxter Healthcare, and travel support from Amgen

• Yeoungjee Cho: has received research grants and speaker’s honoraria from Baxter Healthcare and Fresenius Medical Care

• Suetonia C Palmer: none known

• David W Johnson: has received consultancy fees, research grants, speaker’s honoraria and travel sponsorships from Baxter Healthcare and Fresenius Medical Care, consultancy fees from Astra Zeneca, Bayer, and AWAK, speaker’s honoraria from ONO and BI & Lilly, and travel sponsorships from Ono and Amgen. He is a current recipient of an Australian National Health and Medical Research Council Leadership Investigator Grant

• Jonathan C Craig: none known

• Paul Stroumza: none known

• Luc Frantzen: none known

• Jorgen Hegbrant: serves on the Board of Directors of NorrDia AB and provides consultancy services to Triomed AB

• Giovanni FM Strippoli: none known

Edited (no change to conclusions)