Peritoneal dialysis versus haemodialysis for people commencing dialysis
Abstract
Objectives
This is a protocol for a Cochrane Review (intervention). The objectives are as follows:
This review aims to look at the benefits and harms of PD versus HD for people with kidney failure commencing dialysis.
This review will focus on RKF as it has been strongly associated with death, morbidity, and quality of life in dialysis patients. Although RKF is the main aspect evaluated in this review, other benefits and harms will also be explored. The outcomes included in this review are in line with the core outcomes identified by SONG‐HD and SONG‐PD (SONG 2017).
Background
Description of the condition
The kidneys play a vital role in preserving health through volume maintenance, uraemic toxin and solute clearance, and synthesis of hormones such as erythropoietin. Chronic kidney disease (CKD) is defined as sustained impairment in kidney function and/or the presence of kidney damage (e.g. proteinuria) for longer than three months, and can be categorised into five stages according to the level of estimated glomerular filtration rate (eGFR) (Levin 2013). Globally, 11% to 13% of the population is affected by CKD (Hill 2016). It also estimated that approximately 5 million people have kidney failure (Liyanage 2015), also known as CKD stage 5 (defined as eGFR < 15) who require kidney replacement therapy (KRT) in the form of dialysis or transplantation to sustain life. In 2015, 2.8 million patients with kidney failure were receiving dialysis worldwide, mostly in the form of haemodialysis (HD) (88% of patients receiving dialysis) and 700,000 had functioning kidney transplant (Fresenius Medical Care 2015). The number of patients receiving KRT is projected to exponentially increase to 5.4 million (3.9 to 7.6 million) by 2030 (Liyanage 2015), and these estimations are considered to be conservative as it does not account for those patients with kidney failure who are unable to access KRT resulting in premature death (2.3 to 7.1 million in 2010) (Liyanage 2015).
Patients receiving dialysis are at heightened risk of cardiovascular death (4 to 52 fold higher standardised risk) and premature death compared to the general population (Choi 2014; de Jager 2009; Jager 2017; Roberts 2011; Storey 2018; Wang 2016). The adverse outcomes observed among patients with kidney failure are thought to be driven by an increased prevalence of both traditional and non‐traditional risk factors including chronic inflammation, malnutrition, vascular calcification, left ventricular hypertrophy, atherosclerosis, and alterations of mineral metabolism (Benz 2018; Palit 2014). Moreover, patients with CKD often concurrently suffer from comorbidities that are associated with increased risk of cardiovascular events such as hypertension, diabetes, and dyslipidaemia. One patient‐related factor that has consistently and strongly been shown to be associated with improved survival in patients receiving dialysis is residual kidney function (RKF). For instance, patients receiving HD with preserved RKF at one year after commencing dialysis have been shown to be 30% less likely to die compared to those without RKF and to have a 31% lower risk of cardiovascular death (Shafi 2010). Similar results were also reported among patients receiving peritoneal dialysis (PD) (Bargman 2001; Paniagua 2002). The re‐analysis of the Canada‐USA (CANUSA) study has shown that an increase of 5 L/week/1.73 m2 in GFR was associated with a 12% lower risk of death in PD patients with RKF (Bargman 2001). One of the major benefits of RKF is improved volume management, which in turn may be beneficial for better blood pressure control, regression of left ventricular hypertrophy, and reduction of cardiovascular disease (Wang 2002). RKF is also associated with uraemic toxin clearance, especially protein‐bound and middle molecules (the latter not being easily removed across the HD or peritoneal membrane). Patients with RKF demonstrate more stable parameters of calcium, phosphorus, and haemoglobin reflecting better preservation of hormonal function of kidneys (Morduchowicz 1994; Penne 2011). Moreover, RKF is also associated with better outcomes in terms of quality of life in dialysis (Shafi 2010; Termorshuizen 2003).
Dialysis is also intrusive and can restrict life participation and quality of life, causing additional symptom burden and complications such as fatigue, depression (Murtagh 2007), and access‐related complications (e.g. infection).
Description of the intervention
Dialysis aims to mimic native kidney function by removing small solutes, uraemic toxins, and fluid from patients with kidney failure.
HD is the most commonly utilised dialysis modality worldwide. Most patients receive HD in hospital or satellite dialysis facilities, with a smaller proportion of patients receiving it at home, although there is considerable variation between different regions and countries (ANZDATA Registry 2019; Saran 2019). HD requires the use of a dialyser as an artificial kidney. The dialyzer contains a semi‐permeable membrane which allows water and waste solutes to diffuse from the bloodstream into the dialysis fluid (Sam 2014). Typically, patients receiving HD undergo three dialysis treatments per week, each lasting four to five hours. To receive HD, patients require vascular access the form of an arteriovenous fistula/graft or central venous catheter.
In contrast to HD, PD is predominantly delivered in patients’ own homes and is performed by patients themselves or their caregivers. PD uses the peritoneum as a semi‐permeable membrane, allowing diffusion of solutes and fluid from the bloodstream into the peritoneal cavity. Treatment is delivered by exchanging PD solution several times per day through a PD catheter placed in the abdomen. Continuous ambulatory PD (CAPD) is the most common form of PD, using gravity to perform the PD exchanges manually. Usually four to five exchanges of dialysis fluid are required per day. Continuous cyclical PD (CCPD) is an automated form of PD (APD) which requires a machine to do the exchanges while the patients sleep.
How the intervention might work
The distinct differences in how HD and PD are delivered may translate into disparate benefits and harms for patients receiving dialysis. Several contemporary studies from different registries show that the survival between PD and HD was not different and improved over time. About 80% to 90% patients starting dialysis survive more than one year and 30% to 60% and in recent times more than five years (Haapio 2013; Lee 2009; Mehrotra 2011; Weinhandl 2010). The main cause of death remains cardiovascular disease, and infection is the most common cause of non‐cardiovascular death in dialysis patients (de Jager 2009; Storey 2018). In a recent meta‐analysis of 15 non‐randomised studies involving 4318 participants, the majority of included studies favoured PD over HD with respect to three domains of health‐related quality of life measures. However, since the pooled effect sizes were not statistically significant, it could not be concluded that PD was associated with better quality of life compared with HD (Queeley 2018). Also, PD promotes patient self‐management, and is reported to result in more flexibility, less requirement for technical support and lower electricity and other costs (Li 2017).
RKF is a significant contributor to the patient survival and overall health. It is reported that PD may lead to better preservation of RKF than HD (Jansen 2002; Lysaght 1991), possibly because PD is a continuous therapy that results in less marked fluctuations in fluid volume and serum electrolytes compared with the intermittently administered HD. For HD patients, those receiving extended hours HD at home (6 times/week nocturnal) have been reported to experience a more rapid decline in RKF (Daugirdas 2013), while another study showed that RKF declined comparably in nocturnal home HD and PD patients (Li 2018). Other investigators have suggested that incremental HD, online haemodiafiltration, high‐flux HD, the use of biocompatible membranes, and ultra‐pure dialysate may preserve RKF for a longer time period (McKane 2002; Schiffl 2013; Zhang 2014).
On the other side, maintenance of RKF in PD patients has been attributed to the possibility of subclinical fluid‐overload (Gunal 2001). In addition, PD has a higher technique failure rate than HD due to infection, ultrafiltration failure, and mechanical complications. PD‐related peritonitis remains a major complication and contributes to withdrawal and death from this therapy (Li 2016). A further disadvantage of PD is that due to the increased patient autonomy, PD may add a burden to patients and caregivers (Shimoyama 2003).
Why it is important to do this review
An existing meta‐analysis was published in 2004 but was limited to a comparison of outcomes comparing CAPD and HD (Vale 2004). As the review only identified one randomised controlled trial (RCT) which was only reported as an abstract, no conclusions were able to be drawn. The RCT was subsequently published following premature termination due to poor recruitment. It was only able to recruit 38 patients from 38 centres in the Netherlands over 3 years, and lacked sufficient statistical power for the primary outcome of quality‐adjusted life year score (Korevaar 2003). The proposed review will be an update of the previous review and will be inclusive of broader populations by incorporating data from patients receiving APD as well as CAPD and a number of forms of HD, and a more diverse range of outcomes including a comparison of residual kidney function. In the last decade, further evidence has been published and it is now timely to review the available evidence, including data from non‐randomised studies of intervention (NRSI).
In dialysis, RKF is an important outcome in itself and is also associated with other outcomes relevant to all stakeholders. Therefore, reviewing the impact of HD versus PD on RKF is relevant to better informing shared decision‐making between clinicians, patients, and their caregivers regarding dialysis modality choice.
Objectives
This review aims to look at the benefits and harms of PD versus HD for people with kidney failure commencing dialysis.
This review will focus on RKF as it has been strongly associated with death, morbidity, and quality of life in dialysis patients. Although RKF is the main aspect evaluated in this review, other benefits and harms will also be explored. The outcomes included in this review are in line with the core outcomes identified by SONG‐HD and SONG‐PD (SONG 2017).
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 non‐RCTs (prospective and retrospective identification pf participants) comparing PD to HD treatments for people with kidney failure commencing dialysis will be included. All types of studies included will be used to answer all research questions.
Types of participants
Inclusion criteria
Participants to be considered in this review will be adults and children with kidney failure, who are commencing long‐term dialysis (HD or PD) treatment.
Exclusion criteria
We will exclude studies evaluating combined HD and PD strategies (i.e. “hybrid” dialysis). The previously published review (Vale 2004) already covered comparison of outcomes between PD (CAPD) and HD. Since 2000, there have been many developments in dialysis technology. To ensure that comparison is between contemporary PD and HD and to minimise the risk of confounders, studies published before the year 2000 will be excluded.
Types of interventions
Studies comparing PD and HD will be included in this review.
-
Intervention: Incident patients with kidney failure commenced on PD
-
Comparator: Incident patients with kidney failure commenced on HD.
There is possibility of confounding, particularly for NRSI. The main confounders for this review include, but are not restricted to, age, sex, presence of comorbidities, primary kidney disease, prescribing practices and health policies, socio‐economic status, and employment. Important co‐intervention also include the healthcare setting (in‐centre versus home) associated with the modality choice.
Types of outcome measures
This review will not exclude studies based on non‐reporting of outcomes of interest.
The outcomes selected include the relevant SONG core outcome sets as specified by the Standardised Outcomes in Nephrology initiative (SONG 2017).
Primary outcomes
-
To assess whether initiating dialysis with PD, compared to HD, differs in regards to residual kidney function (measured or eGFR)
Secondary outcomes
-
To assess whether initiating dialysis with PD, compared to HD, impacts:
-
Residual urine output
-
Death (any cause)
-
Cardiovascular death
-
Infection‐related death
-
-
Technique survival (number of patients remaining on the initial mode of dialysis at the end of study)
-
Cardiovascular disease (myocardial infarction/ acute coronary syndrome/heart failure)
-
Infection (blood stream infections, peritonitis)
-
Life participation (ability to participate in life‐related activities, e.g. work, hobbies)
-
Fatigue
-
For each of the specific outcomes in the review, all measurement methods and time points will be included in the review. For example, if kidney function is measured and reported using different metrics, the data will be analysed by calculating standardised mean difference (SMD). If an outcome was measured using a different tool (e.g. life participation, fatigue), this will be quantitatively reported if feasible or qualitatively described. Outcomes measured at different time points (for example at six months after dialysis initiation) will be captured in subgroup analyses.
Search methods for identification of studies
Electronic searches
Randomised controlled trials
We will search the Cochrane Kidney and Transplant Register of Studies through contact with the Information Specialist using search terms relevant to this review. The Register contains studies identified from the following sources:
-
Monthly searches of the Cochrane Central Register of Controlled Trials (CENTRAL)
-
Weekly searches of MEDLINE OVID SP
-
Searches of kidney and transplant journals, and the proceedings and abstracts from major kidney and transplant conferences
-
Searching of the current year of EMBASE OVID SP
-
Weekly current awareness alerts for selected kidney and transplant journals
-
Searches of the International Clinical Trials Register (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 handsearched journals, conference proceedings and current awareness alerts, are available on the Cochrane Kidney and Transplant website.
See Appendix 1 for search terms used in strategies for this review.
Non‐randomised studies of interventions
MEDLINE (OVID) and EMBASE (OVID) will be searched using the search strategies in Appendix 1.
Searching other resources
-
Reference lists of review articles, relevant studies, and clinical practice guidelines.
-
Contacting relevant individuals/organisations seeking information about unpublished or incomplete studies.
-
Grey literature sources (e.g. abstracts, dissertations and theses), additional to those already included in the Cochrane Kidney and Transplant Register of Studies, will not be searched.
Data collection and analysis
Selection of studies
The search strategy described will be used to obtain titles and abstracts of studies that may be relevant to the review. The titles and abstracts will be screened independently by two authors, who will discard studies that are not applicable, however studies and reviews that might include relevant data or information on trials and NRSI will be retained initially: any prospective and retrospective observational studies (including registry studies) comparing outcomes in incident patients with kidney failure receiving HD and PD will be included. Two authors will independently assess retrieved abstracts and, if necessary, the full text of these studies to determine which studies satisfies the inclusion criteria. Disagreements will be resolved in consultation with a third author.
Data extraction and management
Data extraction will be carried out independently by two authors using standard data extraction forms. Disagreements will be resolved in consultation with a third author. Studies reported in non‐English language journals will be translated before assessment. Where more than one publication of one study exists, reports will be grouped together and the publication with the most complete data will be used in the analyses. Where relevant outcomes are only published in earlier versions these data will be used. Any discrepancy between published versions will be highlighted.
Basic characteristics of each study (including details of participants and criteria for selection, interventions and comparators, outcomes, study methods) will be collected. Risk of bias, including confounders and co‐interventions, will also be recorded. Pre‐specified confounders and co‐interventions of interest include age, sex, presence of comorbidities, primary kidney disease, prescribing practices and health policies, socio‐economic status, employment, and healthcare setting (in‐centre versus home).
For every outcome, the outcome domain or title, the measurement tool or instrument, the metric used to characterise participant’s results, the method of aggregation and the timing of outcome measurements will be extracted.
The estimate of intervention effect, the measure of precision and information about derivation of the estimate (e.g. control for specified confounders) will be recorded. If both unadjusted and adjusted intervention effects are reported, the adjusted effects will be preferred. If multiple adjusted estimates of intervention effect are reported in NRSI, the estimate minimizing the risk of bias due to confounding will be chosen. Statistical advice will be sought about whether reported information may be transformed to consistently compare the effect measure across studies using standard software for analysis.
Assessment of risk of bias in included studies
Outcomes obtained from the NRSI could be driven by confounders and co‐interventions previously described. This information will be extracted and considered to inform conduct of subgroup analyses. Moreover, if feasible, analyses will be conducted separately by study design (RCT versus NRSI) to examine for potential impact of unaccounted confounders. Risk of bias assessment will be conducted in line with the Cochrane Handbook for RCT and by using Newcastle‐Ottawa Scale (NOS). This will be assessed independently by two investigators (JP, IE) and YC/DJ will review any discrepancies.
Randomised controlled trials
The following items will be independently assessed by two authors using the risk of bias assessment tool (Higgins 2011) (see Appendix 2).
-
Was there adequate sequence generation (selection bias)?
-
Was allocation adequately concealed (selection bias)?
-
Was knowledge of the allocated interventions adequately prevented during the study?
-
Participants and personnel (performance bias)
-
Outcome assessors (detection bias)
-
-
Were incomplete outcome data adequately addressed (attrition bias)?
-
Are reports of the study free of suggestion of selective outcome reporting (reporting bias)?
-
Was the study apparently free of other problems that could put it at a risk of bias?
Non‐randomised studies of interventions
The Newcastle‐Ottawa Scale (NOS) (www.ohri.ca/programs/clinical_epidemiology/nosgen.pdf) for assessing quality of non‐randomised studies will be used (see Appendix 3).
-
For case control studies the following items will be evaluated.
-
Selection (adequacy of definition, representativeness of the cases, selection of controls, definition of controls)
-
Comparability (comparability of cases and controls on the basis of 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 will be 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 start of study)
-
Comparability (comparability of cohorts on the basis of the design or analysis)
-
Outcome (assessment of outcome, adequacy of follow‐up and duration of follow‐up).
-
Measures of treatment effect
For dichotomous outcomes (e.g. death (any cause), cardiovascular death, infection‐related death) results will be expressed as risk ratio (RR) or odd ratio (OR) with 95% confidence intervals (CI). Where continuous scales of measurement are used to assess the effects of treatment (e.g. RKF, residual urine output, fatigue, hospitalisation), the mean difference (MD) will be used, or the SMD if different scales have been used. Outcomes from RCTs and NRSI will be reported separately. To ease interpretation of the results from SMD, it will be expressed in the units of one of the specific measurement instruments used by the included studies (for example, SMD for RKF will be expressed as GFR by re‐estimation).
Unit of analysis issues
If the review is to include cluster RCTs, the unit of analysis will be at the same level as the allocation, using a summary measurements from each cluster.
Dealing with missing data
Any further information required from the original author will be requested by written correspondence (e.g. emailing corresponding author/s) and any relevant information obtained in this manner will be included in the review. Evaluation of important numerical data such as screened, randomised patients as well as intention‐to‐treat, as‐treated and per‐protocol population will be carefully performed. Attrition rates, for example drop‐outs, losses to follow‐up and withdrawals will be investigated. Issues of missing data and imputation methods (for example, last‐observation‐carried‐forward) will be critically appraised (Higgins 2020).
Assessment of heterogeneity
We will first assess the heterogeneity by visual inspection of the forest plot. We will quantify statistical heterogeneity using the chi squared and I² statistic, which describes the percentage of total variation across studies that is due to heterogeneity rather than sampling error (Higgins 2003). A guide to the interpretation of I² values will be as follows.
-
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 depends 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 confidence interval for I²) (Higgins 2020).
Assessment of reporting biases
If possible, funnel plots will be used to assess for the potential existence of small study bias (Higgins 2020).
Data synthesis
Data will be pooled using the random‐effects model but the fixed‐effect model will also be used to ensure robustness of the model chosen and susceptibility to outliers.
Subgroup analysis and investigation of heterogeneity
Subgroup analysis will be used to explore possible sources of heterogeneity (e.g. participants, interventions and study quality including setting, such as single‐centre versus multi‐centre). Heterogeneity among participants could be related to age (children versus adults) and primary kidney disease (diabetic versus non‐diabetic). Heterogeneity in treatments could be related to prior treatment(s) used and the modality, dose, and duration of therapy. Therefore, subgroup analysis will be conducted to evaluate the source of heterogeneity according to the following.
-
Setting
-
Single‐centre versus multi‐centre
-
-
Study design
-
RCT versus NRSI
-
-
Population
-
Adults versus children
-
-
Primary kidney disease
-
Diabetic versus non‐diabetic
-
-
HD access
-
Fistula/graft versus catheter
-
Sensitivity analysis
We will perform sensitivity analyses in order to explore the influence of the following factors on effect size.
-
Repeating the analysis excluding unpublished studies
-
Repeating the analysis taking account of risk of bias, as specified
-
For each of the outcome, especially where significant heterogeneity is observed, sensitivity analysis according to risk of bias (high vs. low) will be performed
-
-
Repeating the analysis excluding any very long or large studies to establish how much they dominate the results
-
Repeating the analysis excluding studies using the following filters: diagnostic criteria, language of publication, source of funding (industry versus other), and country.
'Summary of findings' tables
We will present 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 (Schunemann 2020a). The 'Summary of findings' tables also include an overall grading of the evidence related to each of the main outcomes using the GRADE (Grades of Recommendation, Assessment, Development and Evaluation) approach (GRADE 2008; GRADE 2011). The GRADE approach defines the certainty 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. This will be assessed by two authors. The certainty 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 (Schunemann 2020b). We plan to present the following outcomes in the 'Summary of findings' tables.
-
RKF
-
Residual urine volume
-
Death (any cause)
-
Cardiovascular death
-
Infection‐related death
-
-
Technique survival
-
Fatigue
-
Duration and episodes of hospitalisation
