Global trends in chronic kidney disease-related mortality: a systematic review protocol

Study design

We will conduct a systematic review of studies reporting mortality for patients with CKD following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis for Protocols 2015 (PRISMA-P-2015)19 and the PRISMA 2020 methodological guidelines (PRISMA 2020).20 The review will include all original peer-reviewed observational studies published from 1990 to date involving patients with CKD, regardless of age, geographic area or stage of disease, that reported on their rate of mortality from any cause. All stages of CKD (stages G1–5) as defined by KDIGO (Kidney Disease-Improving Global Outcomes) criteria,21 patients undergoing dialysis (haemodialysis and peritoneal dialysis) and kidney transplant recipients. Operational definitions are specified in table 1. Studies that do not provide an operational definition of CKD will not be included. Clinical trials will also be excluded from this systematic review because they may not reflect the true mortality rates in the general population due to influences of the interventions or treatments being studied. Non-English language articles will be considered with translation.

Inclusion criteria

1. Studies conducted in patients with CKD (stages 1–5, dialysis and transplant) across all ages.

2. Studies reporting mortality from any cause in patients with CKD.

3. National- or regional-based cross-sectional or cohort studies.

4. Publication date from 1 January 1990.

5. No restriction on language.

Exclusion criteria

1. Studies conducted in patients with acute care settings, including critical illness and hospitalisation.

2. Studies that included patients with CKD as a subpopulation.

3. Studies conducted exclusively on patients with a specific cause of CKD.

4. Studies that do not provide an operational definition of CKD.

5. Clinical trials, review articles, case studies, case reports, case series, images, editorials, letters to editors and commentaries.

6. Studies where we cannot obtain relevant data.

7. Studies reporting only mortality rate before 1990.

8. Duplicate of articles with multiple publications.

Study process

This study will comprise four sections regarding CKD/dialysis status as follows:

1. Patients undergoing long-term haemodialysis, including in-centre and home haemodialysis.

2. Patients undergoing long-term peritoneal dialysis.

3. Kidney transplant recipients.

4. Patients with non-dialysis CKD defined by KDIGO criteria21

The data of each CKD/dialysis/transplant status population will be extracted, analysed and displayed separately.

Search strategy

The research librarian (JYK) collaborated in the development of the search strategy, which will involve an exhaustive exploration of electronic databases, as well as national and multiregional databases (online supplemental table S1), to identify eligible studies. The list of controlled vocabulary search terms (see online supplemental table S2) developed for the Ovid MEDLINE search strategy will be adapted for other databases. Any additional eligible publications will also be identified through manual searches. To ensure the inclusiveness of our research, we will also search for grey literature, including proceedings, organisational reports, policy papers and working papers, with the assistance of the research librarian (JYK).

Study outcomes

1. Primary outcome

   1. Temporal trend in mortality rates stratified by era of the study by decade (1990–2000, 2001–2010, 2011–current).

2. Secondary outcomes

   1. Differences in the mortality rate before and during the COVID-19 pandemic (after 2019).

   2. A distribution of attributable causes (pattern) of death.

   3. Temporal trends in mortality, including all-cause mortality versus CKD-related mortality, as well as across CKD stages, World Bank country classifications, income levels and demographic characteristics.

Data sources

Sources of data will include published studies in electronic databases (including Ovid MEDLINE, Ovid Embase, Cochrane Library (via Wiley), CINAHL, ProQuest Dissertations & Theses Global, Web of Science Core Collection and WHO ICTRP). The first 200 results from Google Scholar will also be evaluated for inclusion. Grey literature will include national and multiregional renal registries of CKD and KRT (online supplemental table S1), including the United States Renal Data System, European Renal Association-European Dialysis and Transplant Association Registry, United Kingdom Renal Registry, Australian and New Zealand Renal Registry, Latin American Dialysis and Transplant Registry, French Chronic Kidney Disease-Le Reseau Epidemiologie et Information en Nephrologie and Dialysis Outcomes and Practice Patterns Study. Kidney transplantation registries (Global Observatory on Donation and Transplantation, Scientific Registry of Transplant Recipients, Canadian Transplant Registry) and WHO reports will also be included.

Data collection and analysis

The study selection process is summarised in the PRISMA flow diagram (figure 1). We will use a two-stage collaborative review process. First, four reviewers (ST, AG, VKH and AA) will independently review titles and abstracts to screen the retrieved studies. In order for a study to be included or excluded in the full-text review, it must receive a favourable vote from at least two of the four reviewers. Then, full texts of selected studies will be examined by two reviewers (ST and AG) to determine their eligibility for inclusion. Studies will be included at each stage when there is consensus between the two reviewers (ST and AG). If there is a conflict between the two reviewers, an adjudicator (IO) will make the final decision. Reasons for exclusion will be reported.

Figure 1

Study selection process. *Other sources will include online publications, scientific abstracts, proceedings, organisational reports and policy papers.

Data extraction and management

Two reviewers will independently extract and enter the data into the standardised forms. Extracted data will include the type and scope of the study, publication year, first author, study site and country, study population, demographic information, comorbidities, stage of CKD, dialysis type, dialysis vintage, withdrawal from dialysis, kidney transplant status, kidney transplant vintage, COVID-19 infection, other infection, malignancy, death from any cause, disease-specific mortality and uncategorised mortality. In case of conflicts, an adjudicator will be responsible for resolving any discrepancies.

Assessment of risk of bias in included studies

The risk of bias in the included studies will be determined using the Cochrane Risk Of Bias In Non-randomized Studies-of Exposure (ROBINS-E) tool.22 23 The quality of evidence will be evaluated using the Grading of Recommendations Assessment, Development and Evaluation methodology.24 We will present the risk of bias for each study in a risk of bias summary table. Publication bias will be assessed using the regression-based test and inspecting funnel plots if more than 10 studies reported the data for each outcome.

Data synthesis and analysis

We will calculate relative risk values with 95% CIs to determine the rate of change in mortality over time and report secular trends by CKD status and regional groups. Missing data will be noted in the data extraction form and reported in the risk of bias section. Heterogeneity will be quantified using I² statistics in each analysis. If there is high statistical heterogeneity between studies, defined as I²>50%, we will describe the results and perform a narrative synthesis in accordance with the Synthesis Without Meta-analysis reporting guideline.25 If an adequate number of studies with reasonable methodological heterogeneity can be retrieved, we will perform meta-analyses using a random effects model in Stata/MP V.17.0 (StataCorp) and present the pooled proportions of included studies in a forest plot. Sensitivity analysis will be conducted to evaluate the influence of individual study.

Subgroup analyses will be conducted by age, gender, pre-COVID-19 and post-COVID-19 pandemic, cause of death, CKD/dialysis status (eg, non-dialysis, haemodialysis, peritoneal dialysis and kidney transplantation), country, region and income based on World Bank country classifications (eg, low, lower-middle, upper-middle and high income). We will perform meta-regression to identify potential explanations for the observed between-study heterogeneity.

Patient and public involvement

This is a study protocol for a systematic review, which does not require patient involvement. No additional patient data will be collected during the study.

Timeline

We plan to collect and organise our data between April and August 2023 and plan to analyse our data and compile the results from September to December 2023. From January to June 2024, we will engage in knowledge translation activities to disseminate our findings to the public (figure 2).

Figure 2

Project timeline.

Ethics and dissemination

A systematic review involves analysing existing data from previously published studies. Thus, ethical approval is not required. Our findings will be published in an open-access peer-reviewed journal and presented at scientific conferences.