Home versus in-centre haemodialysis for people with kidney failure

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).

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

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.

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.

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.

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).

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.