1 Introduction

Chronic kidney disease (CKD) is a significant global health issue, leading to enormous morbidity and mortality. According to the recent Global Burden of Disease Study, there has been a significant increase in the rate of disability-adjusted life-years (DALYs) for CKD between 1990 and 2019, particularly in the age groups 50–74 years (from 1.6 to 2.3%; rank #14 to #8) and 75 years and older (from 1.6 to 2.5%; rank #14 to #9) [1]. In countries with lower and middle levels of socio-demographic index (SDI), the impact of age-standardised CKD is more pronounced, highlighting a significant disparity between the burden of CKD and the availability of adequate healthcare services [2]. A systematic review and meta-analysis conducted in 2016 investigated the worldwide prevalence of CKD based on stage, geographical location, gender and age group. The findings demonstrated a consistently high global CKD prevalence ranging from 11 to 13%, with the majority of cases being stage 3 (10.6%) [3]. Additionally, the study revealed that the CKD prevalence tend to increase with age, and is more common among females [3]. Despite developed regions displaying higher rates of CKD prevalence, the overall prevalence in develo** areas was also notably high, ranging from 8.66 to 17.96% [3]. In 2017, approximately 844 million people worldwide were affected by CKD, with a higher burden observed among socially disadvantaged and vulnerable populations [4, 5]. While the impact of kidney disease is well-documented in developed countries, the magnitude of CKD in develo** countries remains unclear [3,4,5,6]. A recent study revealed a wide variation in advanced-CKD (estimated glomerular filtration rate [eGFR] < 30 mL/min/1.73 m2) prevalence across Asian countries, with China and India having the highest proportion of adults with CKD [7].

In addition to the variation in the etiology of CKD among developed and develo** countries,, the underlying pathogenesis of CKD differs as well [8]. Traditional risk factors such as diabetes mellitus and hypertension share an equal proportion as the leading causes of CKD, both in developed and many develo** countries. However, glomerulonephritis is observed more frequent as cause of CKD in Asia and sub-Saharan Africa countries [8]. There are numerous factors considered to play a role in these differences in develo** countries, particularly the constant predominant of infectious diseases in low-income countries, as a consequence of poor sanitation, inadequate access to safe water, and high prevalence of disease-transmitting vectors [8, 9]. Furthermore, the influence of environmental pollution, pesticides, analgesic abuse, herbal medications, and use of unregulated food additives also plays a role in the development of CKD in develo** countries [8,9,10,11]. However, changes in lifestyles, rapid urbanization and globalization in develo** countries have augmented the demographic transition, which has directed to an intersection of disease burdens, where prevalence of infectious diseases remains high, yet prevalence and severity of metabolic disorders, such as diabetes and hypertension, increases [8]. Geographically, Southeast Asia is situated in a region that is highly susceptible to natural disasters such as earthquakes, floods, and environmental pollutions which could potentially impact the population’s health, including kidney health through the development of acute kidney injury (AKI)-associated CKD [12]. Moreover, Southeast Asia exhibits not only geographical diversity but also significant variation in socio-cultural and economic status. The included countries vary greatly in terms of wealth, ranging from very rich to extremely poor [13]. These factors have resulted in distinct health statuses and healthcare systems both within and between countries.

These settings emphasize the urgent need for early detection to target better prevention and treatment of CKD in Asia. However, the underlying etiologies of CKD have not been clearly identified in most regions in Asia, particularly in Southeast Asia. Therefore, this meta-analysis aims at providing summarized evidence on the underlying kidney disease in advanced CKD including those receiving kidney replacement therapy (KRT) in Southeast Asia.

2 Methods

2.1 Eligibility Criteria and Search Strategy

We conducted this meta-analysis following the Meta-analysis of Observational Studies in Epidemiology (MOOSE) reporting guidelines [14] (Supplementary Form 1), and the review protocol was registered on the International Prospective Register of Systematic Reviews (PROSPERO; Registration ID: CRD42022300786).

The search strategy was systematically conducted across several electronic databases: PubMed, Embase, Cochrane Library, and Scopus, from September 1 to November 10, 2022. In addition, we complemented the initial search strategy by adding potentially relevant studies from the national kidney registries and repository libraries until July 20, 2023. The search conducted on the Indonesia’s kidney registries and repository libraries included the use of Bahasa Indonesia language.

We included observational and interventional studies from the 11 countries in Southeast Asia, that is Brunei Darussalam, Cambodia, East Timor, Indonesia, Laos, Malaysia, Myanmar, the Philippines, Singapore, Thailand, and Vietnam. We included published studies that describe the underlying kidney disease in patients with advanced CKD, articles published in English or Bahasa Indonesia, as well as studies involving dialysis (hemodialysis/HD or peritoneal dialysis/PD) and kidney transplantation patients. We excluded articles from the meta-analysis if they were not original investigations, such as reviews, reports from conference abstracts, case reports/series, and study protocols. Additionally, any studies involving patients under 15 years old, individuals without chronic kidney disease, or those that included subjects that could potentially influence the distribution of the underlying kidney disease (e.g., studies that only recruited selected CKD population such as limited to lupus nephritis or diabetes population) were also excluded. In cases where multiple studies used the same cohort dataset, we selected the one that provided the most information to use in our analysis.

The search was not restricted to a specific publication year. NH independently performed the database search, following the guidance of expert librarians, using search terms related to the patient population of interest, specifically chronic kidney disease and its stages, countries in the Southeast Asia region, and kidney replacement therapies such as dialysis (hemodialysis or peritoneal dialysis) and kidney transplantation. The complete search strategy can be found in Supplementary Table S1.

2.2 Definition of Variables and Outcomes

Chronic kidney disease was defined as any abnormalities in kidney structure or function (eGFR < 60 ml/min/1.73 m2) lasting for at least three months, with implications for health [15]. Advanced CKD was defined as eGFR < 30 ml/min/1.73 m2 (KDIGO stage 4 and 5). The primary kidney disease was categorized into six different etiologies based on KDIGO criteria, that is diabetic kidney disease (DKD), hypertensive nephrosclerosis, glomerulonephritis, polycystic kidney disease (PKD), other, and unknown.

2.3 Data Extraction and Risk of Bias Assessment

Two researchers (NH, AI) independently reviewed titles and abstracts according to our in- and exclusion criteria. Any disagreements between the two researchers were resolved through consensus or by involving a third researcher (FW). The interrater reliability (IRR) analyses between the three researchers was conducted in two stages: initially and midway through the screening process (Supplementary Fig. 1). The intraclass correlation (ICC) statistics were used to determine the degree of correlation and agreement between the measurements of each researcher. We observed excellent reliability in the IRR assessment during both stages, with confidence intervals of 0.989–0.991 and 0.979–0.985 for the ICC (Supplementary Table S2).

Two authors (NH and FW) independently gathered data from each included study. We collected the following information: the first author and year of publication, country where the study was conducted, study design, primary outcome, the stage of chronic kidney disease (CKD), form of kidney replacement therapy (hemodialysis, peritoneal dialysis, kidney transplantation), the study population, sample size, study duration, demographic information (mean age, percentage of male/female sex), and the distribution of the primary cause of kidney disease.

Risk of bias assessment was conducted by two authors (NH and FW) independently. We used the Newcastle–Ottawa Quality Assessment Scale for observational studies (cross-sectional, cohort, and case–control) to evaluate the studies. A combination of seven and eight multiple-choice questions were used to score the studies in three areas: (1) selection of study participants (maximum of 5 stars in cross-sectional studies and 4 stars in cohort and case–control studies); (2) comparability between groups (maximum of 2 stars); (3) outcomes (maximum of 3 stars in cross-sectional studies and cohort) and exposure in case–control studies (maximum 3 stars). A total score of less than 3 indicates a high risk of bias, 4–6 indicates a moderate risk, and 7–9 indicates a low risk. For randomized trials, the Version-2 of the Cochrane risk of bias tool is used to assess potential bias. For non-randomized intervention studies, the Risk of Bias in Non-Randomized studies tool is used to evaluate bias.

2.4 Synthesis Methods and Data Analysis

The primary outcome was the overall distribution of primary renal disease in Southeast Asian countries. Our secondary outcomes include stratifications by country, time period (i.e., before and after the year 2000) and economic status according to the World Bank (i.e., high-income, upper-middle-income, and lower-middle-income country) [16]. Studies that report more than six different underlying kidney diseases will be categorized according to the mentioned etiology classification. The proportions were transformed using the Freeman-Tukey double arcsine transformation to calculate weighted pooled estimates. This provides confidence intervals for the pooled estimates and tests for significance based on normal approximation and variance stabilization. The heterogeneity of the studies was measured using the I2 statistics, where an I2 value greater than 50% indicates significant heterogeneity.

A sensitivity analysis, we repeated our main meta-analysis including only low risk of bias studies to assess robustness of the pooled distribution of CKD etiology.

We conducted our meta-analysis using metaprop in STATA software version 14 (StataCorp, Texas, USA).

3 Results

3.1 Characteristics of Study Population

We identified 1131 studies, of which 1096 were retrieved from database searches and 35 from national kidney registry and library repositories. After the initial screening process, 475 studies were eligible for full-text review. Of these, 81 studies met the inclusion criteria, involving 32,834 participants (Fig. 1. Selection of the included studies). Two countries, East Timor and Laos, did not have any available articles among the eleven countries in the region. Our analysis included 75 observational studies (38 cohort, 35 cross-sectional, 1 cross-sectional & case–control, and 1 cross-sectional & cohort study), 3 randomized controlled trials, and 3 registry data. The majority of the reports came from five countries: Thailand (21 studies, 25.9%; n = 12,215), Malaysia (19 studies, 23.5%; n = 3378), Singapore (18 studies, 22.2%; n = 13,013), Indonesia (11 studies, 13.6%; n = 1805), and Brunei Darussalam (5 studies, 6.2%; n = 1105). These studies were published between 1986 and 2022 in high-income to lower-middle-income countries (Table 1). The study population mainly consisted of dialysis patients (n = 17,533, 53%) and kidney transplant recipient (n = 9722, 30%), while the smallest proportion originated from the non-dialysis population (n = 1379, 4.2%) (Supplementary Table S3). Detailed characteristics of all included studies are presented in Table 2. The assessment of bias in the included studies (excluding three registry studies) showed that less than half of the studies (n = 34) had a low risk of bias.

Fig. 1
figure 1

Selection process for the included studies

Full size image
Table 1 Summary of included studies
Full size table
Table 2 Characteristics of the included studies
Full size table

3.2 Etiology of Advanced CKD Among South East Asia Countries

The pooled proportions for different underlying renal conditions were: diabetic kidney disease 29.2% (95%CI 23.88–34.78), glomerulonephritis 20.0% (95%CI 16.84–23.38), hypertensive nephrosclerosis 16.8% (95%CI 14.05–19.70), other 8.6% (95%CI 6.97–10.47), unknown 7.5% (95%CI 4.32–11.50), and polycystic kidney disease 0.7% (95%CI 0.40–1.16) (Table 3 and Supplementary Table S4). Heterogeneity between studies was observed in all etiological diagnosis (Supplementary Figure S2-7).

Table 3 Etiology of advanced-CKD in Southeast Asia
Full size table

Sensitivity analysis was conducted in 34 studies with low risk of bias, including a total of 12 218 patients, across five countries, namely Indonesia (1 study, n = 389), Malaysia (7 studies, n = 1664), Myanmar (1 study, n = 54), Singapore (9 studies, n = 4450), and Thailand (16 studies, n = 5661) (Table 4, Supplementary Figure S8-13).Overall, the results are similar to the main analysis. It is evident that the most common cause of CKD is still diabetic kidney disease (29.9%, 95%CI 21.97–38.41), followed by glomerulonephritis (19.4%, 95%CI 14.73–24.49), and hypertensive nephrosclerosis (13.4%, 95%CI 9.34–17.94). However, often the underlying disease was unknown (13.6%, 95%CI 6.82–22.11), possibly due to the majority of studies being from Thailand, where a high rate of unknown causes has been reported. Additionally, the prevalence of polycystic kidney disease remains low compared to the prevalence before sensitivity analysis (0.9% vs. 0.7%). Nevertheless, there is a significant heterogeneity observed between the studies after this analysis.

Table 4 Sensitivity analysis on the etiology of advanced-CKD in Southeast Asia
Full size table

3.3 Primary Renal Disease Across Countries

Results were stratified to determine the proportion of underlying kidney disease in 9 countries. The analysis revealed that diabetic kidney disease is most commonly found in Malaysia (40.3%, 95%CI 28.78–52.28), Singapore (39.2%, 95%CI 28.12–50.82), and Brunei Darussalam (37.9%, 95%CI 21.90–55.38) (Table 5). On the other hand, hypertensive nephrosclerosis is the main cause of kidney disease in Myanmar (53.7%, 95%CI 40.61–66.31), Indonesia (48.7%, 95%CI 37.73–59.77), and Cambodia (46.7%, 95%CI 41.89–51.54). Glomerulonephritis was found to be the primary cause in Vietnam (57.3%, 95%CI 51.74–62.80) and the Philippines (42.2%, 95%CI 18.70–67.59). Although the overall proportion of unknown causes is less than 10%, it remains prevalent in countries like Malaysia and Thailand (10.4%, 95%CI 5.24–17.04 and 23.6%, 95%CI 15.24 –33.04; respectively). Surprisingly, the prevalence of polycystic kidney disease is high in Vietnam (3.5%, 95%CI 1.66–5.94), with smaller percentages reported in Brunei Darussalam and Singapore (1.09%, 95% CI 0.00–4.46 and 1.95%, 95%CI 0.84–3.44; respectively), and remains low (prevalence less than one percent) in the other six countries (Table 5, Supplementary Figure S14-19).

Table 5 Stratification of the distribution of primary kidney diseases across countries
Full size table

3.4 Primary Kidney Diseases Before and After 2000

Further analysis was conducted to examine the occurrence of underlying kidney disease during two different time periods, specifically before and after 2000. Prior to 2000, three reports were completed from Singapore, Thailand, and Malaysia, while the majority of studies (96.3%, n = 78/81 [n = 30 427 subjects]) were carried out after 2000. The most common primary renal disease found before 2000 was unknown causes (35%, 95%CI 19.67–52.04), followed by glomerulonephritis (34.7%, 95%CI 14.97–57.70). After 2000, the majority of diagnoses were related to diabetic kidney disease (30%, 95%CI 24.59–35.97), although glomerulonephritis still accounted for a significant proportion of the study population (19.5%; 95%CI 16.23–22.93). During the two study periods, there was a notable increase in both hypertensive nephrosclerosis and diabetic kidney disease (from 4.9%, 95%CI 0.00–22.78 to 17.4%, 95%CI 14.56–20.41 and from 9.2%, 95%CI 0.00–33.01 to 30%, 95%CI 24.59–35.97, respectively). The prevalence of polycystic kidney disease and other etiology remained stable. In contrast, the occurrence of glomerulonephritis and unknown causes declined (from 34.7%, 95%CI 14.97–57.70 to 19.5%, 95%CI 16.23–22.93, and from 35%, 95%CI 19.67–52.04 to 6.8%, 95%CI 3.64–10.78, respectively) (Table 6, Supplementary Figure S20-25).

Table 6 Stratification of the distribution of primary kidney diseases before and after 2000
Full size table

3.5 Primary Renal Disease Across Countries’ Economic Status

The prevalence of primary renal diseases was further investigated based on the economic standing of the country. Most of the studies come from the upper-middle-income countries (51 studies; n = 17,398), while only 7 studies (n = 1,318) come from low-middle-income countries. Among upper-middle-income and high-income countries, diabetic kidney disease is the most common condition (26.8%, 95%CI 21.42–32.60 and 38.8%, 95%CI 29.33–48.79, respectively). On the other hand, glomerulonephritis is more prevalent in lower middle-income countries (33.8%, 95%CI 15.62–54.81), when compared to upper middle-income countries (15.5%, 95%CI 12.53–18.74), but it is also common in high-income nations (27%, 95%CI 19.56–35.22). Lower-middle-income countries report a relatively higher occurrence of hypertensive nephrosclerosis and other etiologies compared to upper-middle- and high-income countries. Additionally, polycystic kidney disease is reported more frequently in high-income countries compared to countries with lower income status (Table 7).

Table 7 Stratification of the distribution of primary kidney diseases across economic status
Full size table

4 Discussion

In this meta-analysis, we summarized the evidence using the best available data to estimate the prevalence of the different causes of CKD in Southeast Asian countries. Our results suggest that in the advanced-CKD population and those receiving kidney replacement therapy, diabetic kidney disease is the most prevalent etiology, followed by glomerulonephritis and hypertensive nephrosclerosis. The prevalence of polycystic kidney disease is relatively low. Diabetic kidney disease appears to have the highest prevalence in upper-middle and high-income countries and to have sharply increased after the year 2000. Nevertheless, the proportion of glomerulonephritis should not be disregarded.

Our study aimed to examine the variation in CKD etiology before and after the year 2000 in Southeast Asian countries, as well as among different socioeconomic groups [16]. Global trends indicate that non-communicable diseases have become more prevalent since the turn of the century, resulting in increased mortality rates [17]. Diabetes is a major contributor to this trend, with its increasing prevalence mirroring the rise in CKD prevalence. Furthermore, the longer life expectancies observed after the year 2000 increase the likelihood of develo** chronic diseases [17]. Likewise, advancements in health have been linked to economic growth through improvement in nutrition, enhancements to public health infrastructure like water purity and sanitation, and more effective medical treatments like antibiotics and vaccinations [18]. Subsequently, this situation is believed to impact the underlying kidney disease across various time periods and countries with differing levels of wealth.

In the entire cohort, it is evident that diabetes mellitus is the leading cause of primary kidney disease among studies. This trend is particularly observed in upper-middle and high-income countries. And a significant increase in the prevalence of diabetic kidney disease is observed after the year 2000. This observation resembles the trend in western countries like Europe and the US. The incidence of diabetic kidney disease in kidney failure patients receiving KRT is reported to be 23% (32.4 per million population) in Europe in 2021, up to 46.8% in the US in 2019 [19,20,21]. Diabetes mellitus is a non-communicable disease that has a major impact on global health, particularly in relation to kidney disease. Based on data from the Global Burden of Disease Study 2019 (GDB 2019), approximately two thirds of DALYs in 2019 from known causes were attributable to CKD caused by type 2 diabetes mellitus (T2DM) and hypertension, which also showed the largest increase in the age-standardized DALY rate [1, 22]. Between 2021 and 2045, it is expected that the global rate of adults living with T2DM would escalate by 46%; where Southeast Asia will provide the largest increase (68%), following Africa and the Middle East & North Africa [23, 24]. A study involving 10 Asian countries (China, Hong Kong, Indonesia, Malaysia, Pakistan, Philippines, Singapore, South Korea, Taiwan and Thailand) revealed that a significant proportion of hypertensive type 2 diabetic patients (58.6%) exhibit either micro or macroalbuminuria, which could lead to kidney complications [25]. Moreover, the epidemic of metabolic syndrome [26], is the primary cause of the rising burden of T2DM. This increase in T2DM has been linked to a rise in processed food consumption, a decline in physical activity, and an increase in sedentary behavior [26]. The so-called western lifestyle is associated with increased urbanization and industrial advancement on a worldwide scale. The prevalence of metabolic syndrome among Asian populations was reported to be about 10–30%, resembling that among Western countries [26,27,28,29,30]. Moreover, there is evidence linking the hazard of metabolic syndrome to an increased risk of develo** CKD. Additionally, it is estimated that 20–50% of T2DM patients will eventually develop DKD [31,32,33]. The rapid growth of the ageing population in Asia has also altered the disease epidemiology. From this standpoint, we strongly support the international recommendation to early identify the development of CKD in the high-risk population (such as those with T2DM) in Southeast Asia [34].

Our research has demonstrated that glomerulonephritis is prevalent in a similar manner to earlier findings from Southeast Asia [35, 36]. This finding remains consistent even after conducting sensitivity analysis and when comparing different countries, study periods, and income levels. However, due to the significant heterogeneity in the studies, we should interpret this result with caution. For instance, the prevalence of glomerulonephritis is particularly high in Vietnam and the Philippines (over 40%), while it consistently stays around 10% in other countries, except for Myanmar and Indonesia (less than 10%). Similarly, there are notable variations in the occurrence rate of unknown causes across nations. This difference may be attributed to factors beyond sample size and demographics, such as variations in healthcare infrastructure, availability of diagnostic equipment for kidney biopsies, as well as the number of kidney specialists and pathologists. Nevertheless, it is important to acknowledge that Southeast Asia has distinct characteristics compared to other regions worldwide, including diverse economic levels that can influence healthcare systems. The ongoing prevalence of infectious diseases (such as Malaria, Leptospirosis, Dengue, Hepatitis, HIV etc.) and the potential risk of kidney diseases caused by traditional medicines or environmental toxins should also be considered. IgA nephropathy has been identified as a major cause of glomerulonephritis in Asia [37,38,39]. This study presents a relatively high rate of glomerulonephritis, which may reflect the actual estimate of the disease prevalence in the population due to the aforementioned reasons. Additionally, the under-reporting of glomerulonephritis incidence, particularly in cases where etiological diagnosis cannot be performed due to late-stage chronic kidney disease or inadequate healthcare facilities, should also be taken into account. In contrast to the international recommendations that suggest performing urine albumin to creatinine ratio (uACR) tests in high-risk individuals, these findings emphasize the importance of identifying this specific population group and conducting simple urinalysis tests to detect not only proteinuria but also hematuria or abnormal sediment/cast associated with glomerular diseases.

Another interesting observation is the consistently low prevalence of autosomal dominant polycystic kidney disease (ADPKD) in this cohort, although a higher prevalence is observed among high-income countries. Even comparing to patients receiving KRT, the prevalence of ADPKD in our cohort was much lower than in the Europe (9%) and in the US (9.5% in dialysis and 15.7% in kidney transplant population) [40]. Autosomal dominant polycystic kidney disease affects all racial groups worldwide and is the most common monogenetic disorder of the kidney [41]. The lower prevalence observed in our study may be attributed to under-reporting, which could be due to the inability to perform etiological diagnosis, especially in lower-middle-income countries with limited healthcare facilities. In addition, it might reflect the limited access to nephrology or dialysis facilities.

This meta-analysis included studies with unidentified sources of bias and it was challenging to measure these publication bias. Nonetheless, we made several attempts to reduce the possibility of bias. Studies with a wide range of publication dates were taken into account in this meta-analysis, which also included articles that were published in both English and Bahasa Indonesia languages. Two reviewers individually looked into each study selection procedure. The pooled prevalence from sensitivity analysis was estimated from studies with moderate to high quality.

To our best knowledge, this study is the first to identify the underlying kidney disease distribution in the advanced-CKD population of Southeast Asia by conducting an extensive search across multiple databases and including a large sample size. The manuscript makes a valuable contribution to understanding CKD etiologies in Southeast Asia. However, it is important to acknowledge certain limitations: we included observational, investigational, and registry studies without considering their methodology; not all Southeast Asian countries were included in the study; and the establishment of etiological diagnosis was not clearly defined in each study, which may have led to incorrect classification of some causes. Moreover, the small number of study participants in certain countries making comparisons between countries less reliable. Finally, we acknowledge that relevant studies could be excluded due to language restrictions and there were possible language bias exist.

In conclusion, our meta-analysis indicates that in the advanced-CKD population and patients receiving kidney replacement therapy in Southeast Asia, diabetic kidney disease is the most prevalent etiology, followed by glomerulonephritis and hypertensive nephrosclerosis. The prevalence of diabetic kidney disease is highest in upper-middle-income and high-income countries and has significantly increased after 2000. To reduce the burden of the disease, it is essential to implement an effective screening program for high-risk populations (those with diabetes mellitus). This program should consider building infrastructure, improving access to diagnostic tests for early detection of CKD, and increasing public awareness about kidney health and risk factors.