, Introduction
Chronic kidney disease (CKD) is typically diagnosed based on a low estimated glomerular filtration rate (eGFR; <60 mL/min/1.73 m2) or other kidney-related symptoms, such as albuminuria [1]. CKD is one of the world’s most serious public health issues, with an incidence of 9.1% [2].
In 2017, 697.5 million cases of CKD at any stage were recorded [2]. Since 1990, the global prevalence of CKD has increased by 29%, whereas the age-standardized prevalence has not changed significantly (1.2%) [2]. Furthermore, this condition affects 35.8 million disability-adjusted life years [3]. Between 1990 and 2017, CKD-related mortality increased by an estimated 41.5%, independent of age [2]. Diabetes, the primary cause of CKD worldwide, is likely to intensify the burden of CKD [4]. The largest increase in the global burden of CKD will be attributed to the population of Asia [5,6]. The prevalence of diabetes is expected to more than double in South Asia alone between 2000 and 2035 [5,6]. The statistics show that as Asian populations age, their risk of developing CKD increases [7]. The available statistics on the current prevalence of CKD in this region are poor and do not allow for more reliable predictions of future events [7]. The high poverty levels in South Asia contribute significantly to the burden of CKD [2]. Several studies in East, South, and Southeast Asia have found significant disparities in CKD prevalence (4.7%–17.4%) [8–10]. Although age plays a major role in the development of CKD, understanding the magnitude of the burden imposed by an aging population is critical for successful screening and management [11]. According to research, women are more likely than men to acquire CKD and end-stage kidney disease [2]. Therefore, sex differences in CKD need to be considered to improve sex-specific CKD treatment.
It should be noted that comparisons of crude CKD prevalence rates are misleading because crude rates convey little information about a population’s health [12]. Furthermore, the use of crude rates ignores influencing factors, such as age and sex [12]. Therefore, factors causing variations across groups must be standardized [13]. Age and sex are two of the most widely used standardization factors [13]. The use of non-standardized rates to understand disease patterns across countries can be highly biased because differences in crude rates can be attributed to differences in age or sex [14]. For example, in the general population without major chronic diseases or risk factors for CKD, women may experience a slower rate of mean GFR decline than men. The sex disparities and GFR loss reported even in healthy people highlight the importance of an age- and sex-adjusted definition of CKD [14]. Because of differences in age or sex structure, it is critical to adjust for age and sex when comparing disease rates among countries [13,14].
This meta-analysis examined the differences in CKD incidence by age, sex, geographic location, and analytical methods used to calculate GFR. A broader understanding of these factors can help healthcare professionals, politicians, and the public determine interventions and research priorities.
Methods
This meta-analysis was conducted in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. Search engines, such as MEDLINE, PubMed, Embase, Cochrane Controlled Register of Trials, CINAHL, and Web of Science, were used to collect articles published until February 30, 2025. The search was conducted using Medical Subject Headings terms and free text words, including exp kidney failure, chronic (MeSH), exp renal insufficiency, chronic (MeSH), exp glomerular filtration rate (MeSH), chronic renal impairment, CKD, chronic kidney failure, chronic renal disease, reduced glomerular filtration rate, reduced eGFR, low glomerular filtration rate, low eGFR, low GFR, Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI), Modification of Diet in Renal Disease (MDRD), Cockcroft-Gault (CG), CKD, end-stage renal disease (ESRD), exp prevalence (MeSH), prevalence, exp epidemiologic studies (MeSH), exp case-control studies (MeSH), exp cohort studies, and exp cross-sectional studies (MeSH). We also searched for relevant articles by reviewing systematic reviews and their references. The titles and abstracts were evaluated by three reviewers using Rayyan (Rayyan Systems Inc.) [15]. A fourth reviewer resolved the disagreements. All studies that met the full-text criteria were examined and appraised using the inclusion and exclusion criteria.
Inclusion and exclusion criteria
In this meta-analysis, the prevalence of CKD was investigated using both observational and quantitative studies. The studies confirmed CKD in patients with albuminuria/proteinuria or eGFR <60 mL/min/1.73 m2. Qualitative research, case reports/series, animal studies, and opinion surveys were excluded. Studies that only covered stages 1 and 2 of CKD were conducted in languages other than English, and those that reported on symptoms that did not meet the KDIGO (Kidney Disease: Improving Global Outcomes) criteria were excluded. CKD stages included stage 1 (eGFR, ≥90 mL/min/1.73 m2; albumin-to-creatinine ratio [ACR], >3 mg/mmol), stage 2 (eGFR, 60–89 mL/min/1.73 m2; ACR, >3 mg/mmol), stage 3A (eGFR, 45–59 mL/min/1.73 m2), stage 3B (eGFR, 30–44 mL/min/1.73 m2), stage 4 (eGFR, 15–29 mL/min/1.73 m2), and stage 5 (eGFR, <15 mL/min/1.73 m2). eGFR (mL/min per 1.73 m2) was calculated using the CKD-EPI or MDRD equations (Supplementary Table 1, available online).
Data analysis
Three reviewers retrieved data on patients, populations, interventions, comparators, and outcomes in accordance with the PICO (population, intervention, comparator, outcome) criteria. The country, sample size, age, sex, and comorbidities were recorded. The method used to calculate eGFR, chronicity, and geographical region was designated as intervention/exposure. The outcomes were CKD incidence in stages 1–5 and 3–5 (crude or adjusted). The Quality in Prognostic Studies (QUIPS) tool was used to assess the risk of bias in the included studies [16]. According to the Cochrane Methods Prognosis Group, the QUIPS tool is an effective method to evaluate prognostic research because it eliminates any potential bias in studies [16]. Three team members screened each study and determined its level of bias, which was classified as low, moderate, or high. To settle any potential disputes, an additional author was included in the review process. The QUIPS tool assesses prognostic factors, outcomes, statistical analysis and reporting, study confounding, participation, and attrition [16]. All studies, regardless of potential bias, were included.
For the meta-analysis, crude prevalence values rather than adjusted values were used. OpenMeta version 0.24.1 (Brown University) was used for the meta-analysis. The random-effects model was used in all meta-analyses because of significant heterogeneity (I2 > 50%, p < 0.05). Moreover, the random-effects model was used for the meta-analysis due to significant variations in methodology among the included studies, including different medical settings, patient age, and intervention approaches. More information regarding the random-effects model used in this study can be found in the Methods section of Supplementary Information (available online).
Funnel plots were used to assess publication bias. Asymmetry was assessed using Egger’s linear regression method. In cases where publication bias was clear, the trim-and-fill method [17] was used to produce adjusted estimates. The meta-analysis was carried out to evaluate the prevalence of CKD, comprising stages 3–5, in men and women from various countries and regions. The incidence of CKD stages 1–5 and 3–5 was also investigated in relation to MDRD and CKD-EPI. The CKD-EPI estimate was applied to studies reporting varying incidence rates using different GFR estimation methodologies.
The effects of various studies on the results were explored using sensitivity analysis, which classified leave-one-out approaches based on their effect size. Moreover, meta-regression analysis was used to establish the link between CKD prevalence and demographic variables, such as age, sex, and area. More information regarding the fitted model, regression estimates, t-statistics, and p-values associated with the regression estimates is provided in the Methods section of Supplementary Information (available online).
Direct comparisons of prevalence values were performed in this study. We explored a novel application of the two one-sided t test (TOST) to assess the equivalence of prevalence estimates in two populations [18]. For multiple comparisons using Dunn’s post hoc, the false discovery rate (FDR) was used as a method of conceptualizing the rate of type I errors in null hypothesis testing. In general, a smaller FDR indicates a lower probability of false positives, making the enrichment result statistically more robust.
In addition, the age- or sex-standardized (aCKD or sCKD) prevalence of CKD was calculated using the indirect standardization method. Instead of using a single population structure as a standard and applying sets of rates to it to predict anticipated occurrences, this method uses a rate from a standard population (e.g., the prevalence of CKD in Asia) to determine the standardized CKD prevalence in each of the populations under consideration. The indirect standardization method was used to compute the number of CKD cases expected in Country B if it had the same age-specific CKD rates as Country A. The predicted number of CKD cases in Country B was derived by multiplying Country A’s age- or sex-specific rate by Country B’s population in the relevant age or sex group. The total number of patients with CKD in Country B was calculated by adding age and sex categories. Whereas indirect standardization estimates expected values by applying the age-specific rates of a reference population to the age structure of the subject population, then compares observed values to expected values, direct standardization applies the age-specific rates of the subject population to the age structure of the standard population to obtain an overall rate [13,19–21]. We used indirect standardization because we did not know the age-specific rates of the subject population and because some countries in this study had few events; if we had done direct standardization, the estimated rates would have been subject to significant sampling variation [13,19–21]. With indirect standardizations, we can use rates from a wide population as the standard, thus limiting the consequences of sampling errors [13,19–21].
Results
Search results and study characteristics
Of the 6,190 articles we searched, 405 were selected for full content. Fig. 1 shows the 42 studies [22–63] that met the inclusion criteria and involved 2,271,169 participants (978,394 men; 1,216,812 women). Table 1 presents the characteristics of the studies included in the meta-analysis. The meta-analysis included 13 studies from China [26–28,37,39,42,44,47,49,52,54,56,61], two from Taiwan [22,35], four from South Korea [24,29,31,63], eight from India [23,34,40,41,43,46,48,50], five from Iran [25,32,36,53,62], four from Japan [30,45,55,59], two from Sri Lanka [38,60], one from Nepal [57], one from Vietnam [33], one from Malaysia [51], and one from Bangladesh [58]. East Asia includes China, Taiwan, Korea, Japan, and Malaysia (participants, 2,171,927); South Asia includes India, Bangladesh, Nepal, Sri Lanka, and Vietnam (participants, 45,943); and the Middle East includes Iran (participants, 53,299). In 24 studies, the prognosis of CKD was evaluated by combining eGFR and proteinuria (or albuminuria), whereas in 18 studies, it was estimated only by eGFR or proteinuria. MDRD, CKD-EPI, and other equations were used to estimate eGFR in 15, 27, and two investigations, respectively. The average age of the participants (>18 years) in this meta-analysis was 52.3 ± 10.9 years. Women constituted 55.4% of the participants. Supplementary Table 2 (available online) summarizes the quality assessments of the included studies.
Meta-analysis of chronic kidney disease prevalence in Asia
The results of the meta-analyses regarding the crude prevalence of CKD in Asia, CKD stages 1–5 and 3–5, different countries, regions, age, sex, and methods used to calculate eGFR (CKD-EPI and MDRD equations) are presented in the Results section of Supplementary Information and Supplementary Figs. 1–27 (available online).
Prevalence of age- and sex-standardized chronic kidney disease in Asian countries relative to the Asian reference
The prevalence of aCKD and sCKD is presented in Tables 2 and 3. In comparison with the prevalence of aCKD in Asia (as the comparator), the prevalence of aCKD stages 1–5 was the highest in Nepal (4.6 times), followed by Bangladesh (2.8 times), Malaysia (2.3 times), Taiwan (2 times), China (83.8%), Sri Lanka (71% times), Japan (33.5%), Korea (17%), India (8.5%) and Vietnam (4.4%) (Table 2). South Asia (65.8%) and East Asia (34.6%) had a higher prevalence of aCKD stages 1–5 than the comparator (aCKD stages 1–5 prevalence throughout Asia). In comparison with the overall prevalence of aCKD stages 3–5 in Asia, Vietnam had the highest prevalence (24 times), followed by Malaysia (21.7 times), Sri Lanka (2.1 times), Iran (89%), Taiwan (70.9%), Japan (52.3%), South Asia (25.8%), China (25.2%), and Bangladesh (25.6%). Korea (66.7%), India (36.8%), and East Asia (3.9%) had lower prevalences of aCKD stages 3–5 compared to the comparator (the prevalence of aCKD stages 3–5 throughout Asia) (Table 2).
The sCKD stages 1–5 were lower in Nepal (57.8%), Taiwan (36.4%), Korea (35.5%), South Asia (19.3%), and Vietnam (10.7%) than in Asia (Table 3). The prevalence of sCKD stages 1–5 was greater in Bangladesh (54.8%), India (27.5%), Malaysia (24.4%), China (8.5%), East Asia (3.3%), and Japan (1.3%) than in the comparator (the prevalence of sCKD stages 1–5 throughout Asia). The prevalence of sCKD stages 3–5 was lower in Korea (63.5%), Taiwan (37.7%), China (17.4%), and East Asia (3.4%) than in the reference group (Asia). In comparison with the reference group (Asia), the prevalence of sCKD stages 3-5 was greater in Iran (67.6%), Sri Lanka (30.5%), South Asia (19.8%), Japan (18.7%), and India (9%) (Table 3).
Discussion
According to this meta-analysis, the prevalence of CKD in Asia was 17% in stages 1–5 and 7.7% in stages 3–5. We found that differences in age, sex, country, and eGFR calculation method may contribute to the discrepancy in CKD prevalence in Asia.
We found that older Asians (>60 years) had a higher prevalence of CKD (stages 1–5 and 3–5) than people aged <0 years. Kampmann et al. [64] reported that the prevalence of CKD stages 3–5 in the Danish population increased with age. The risk of developing CKD increases with age, with approximately half of stage 3a+ CKD cases occurring after the age of 70 years [65]. We found that the mean age of patients with CKD stages 3–5 was 73.2 years. The higher prevalence of stages 3–5 CKD in older patients may also imply a higher risk of poor outcomes during the early stages of the disease [65,66]. The decrease in eGFR during normal aging requires distinguishing between age-related decline and kidney damage (by modifying thresholds) [66,67]. Changing the cutoff value may result in different estimates of CKD prevalence in older adults, which is essential to identify individuals most likely to develop the disease and improve treatment outcomes [66,67]. Therefore, it is important to conduct research to improve screening in older adults to prevent the development and progression of CKD. Furthermore, the prevalence of CKD stages 3–5 was 2.5 times greater in CKD-EPI-based studies compared with MDRD-based studies (five out of the eight studies with older persons used the CKD-EPI method to evaluate GFR). These findings imply an overdiagnosis, and the validity of eGFR cutoffs in older patients should be evaluated before drawing assumptions [66].
We found that Asian men and women had similar prevalence rates of CKD stages 1–5 and 3–5. A recent analysis comprising 119 publications with 29,159,948 participants reported that the prevalence of CKD stages 1–5 was similar in men and women worldwide; however, women had a higher prevalence of CKD stages 3–5 (6.4% vs. 7.5%) [68]. This meta-analysis showed that in South Asia, men had a higher prevalence of CKD stages 1–5 (19.0% vs. 17.2%) and 3–5 (10.2% vs. 8.0%) than women. Notably, there was no difference in the prevalence of CKD stages 1–5 and 3–5 among men and women in East Asia. It appears that men in South Asia may progress faster to ESRD [2]. South Asian men are more likely to be diagnosed with CKD and less likely to undergo specialist screening and treatment, contributing to the disease’s higher prevalence [68,69]. Furthermore, there is evidence of considerable sex inequalities in CKD treatment access, and modifications are required to provide equitable access [2]. These findings indicate disparities in CKD prevalence among men and women in South and East Asia.
We found that South Asia had a higher prevalence of CKD stages 1–5 than East Asia. The lower prevalence of CKD stages 1–5 in East Asia could be attributed to quick changes in sociodemographic and lifestyle characteristics, as well as higher incomes, a greater number of physicians or nephrologists per patient, and more frequent health checks, all of which may influence disease diagnosis and progression [70,71]. In East Asia, there are 26.5 physicians (95% confidence interval [CI], 19.5–35.1) per 10,000 people, whereas Southeast Asia has 7.3 (95% CI, 5.0–10.2) per 10,000 individuals [71,72]. This study showed that the prevalence of CKD stages 3–5 was greater among South Asians than among East Asians. East Asians may have better treatment options, medical attention, and more effective drugs to prevent progression to the advanced stages of CKD. There is evidence that healthcare systems in South Asia are challenging for patients [71]. Furthermore, a higher incidence of CKD stages 1–5 and 3–5 may reflect a potential incompatibility between high-risk disease and treatment options in South Asia [70]. Furthermore, because the majority of the studies included in this analysis were conducted in rural and suburban areas [26,32,34,37,38,40,42,43,49–51,58,60], the increased prevalence of CKD and CKD (stages 3–5) in South Asia could be attributed to risk factors such as inadequate treatment, a lack of targeted health care, agricultural practices, and environmental pollution [70]. Moreover, the lack of general practitioners and the large distances between nurses and pharmacists for dosage monitoring can make it difficult for people in South Asia to receive therapy [71].
Compared with the comparator (aCKD prevalence across Asia), we found that Nepal, Bangladesh, Malaysia, and South Asia had the highest prevalence of aCKD stages 1–5, whereas the highest prevalence of aCKD stages 3–5 was in Vietnam, Malaysia, Sri Lanka, and Iran. The prevalence of sCKD stages 1–5 was the highest in Bangladesh, India, and Malaysia. Furthermore, Iran and Sri Lanka had the highest prevalence of sCKD stages 3–5. The increased prevalence of aCKD stages 1–5 in the Middle East (Iran) and South Asia could be attributed to the fact that research in these regions was primarily conducted in rural and suburban settings. Furthermore, Southern Asian countries do not have universal healthcare coverage, and healthcare is primarily delivered by private organizations with limited government assistance [72]. In the Iranian public health system, there is no accurate routine CKD screening program, and most patients are not recognized in a timely manner [73]. In contrast, this study revealed that South Korea had the lowest prevalence of aCKD and sCKD stages 1–5 and 3–5. The Korean studies included in this study were primarily conducted in urban regions with better incomes, health insurance coverage, healthcare access, or the availability of more hospitals and nephrologists per patient [72]. Moreover, South Korea has made it mandatory for all citizens to participate in the National Health Insurance Service, with the exception of those who are eligible for the Medical Benefit System, a medical aid program for low-income citizens [72]. A recent study found that the prevalence of CKD was lower in Korea (8.2%) than in the United States (10.0% and 13.1%) and China (10.8%) [32]. Among the factors related to CKD, the Korean population’s body mass index (BMI) is comparatively low [74]. In South Korea, the average BMI for both sexes is 23.2 kg/m2, whereas Iranian men and women have higher BMIs (25.5 and 27.6 kg/m2, respectively) [74,75]. This may also contribute to the decreased prevalence of CKD in South Korea. These findings imply that there could be a disparity between the high risk of CKD progression and the level of treatment offered in different countries [72].
The present study revealed a comparatively high incidence of aCKD (stages 1–5, 33.5%; stages 3–5, 52.3%) in Japan compared with the comparator (i.e., Asia). The global prevalence of CKD ranges from 9.1% to 13.4% [2]. The high incidence of CKD in Japan may be attributable to factors such as aging, improved healthcare facilities, more centers for kidney biopsy diagnostic testing, and greater numbers of kidney experts and pathologists [72,73]. Moreover, the longer life expectancy of the Japanese (84.3 years) may explain why CKD rates are more prevalent in Japan [76]. Furthermore, this finding implied that there could be a disparity between the high risk of developing this condition and the level of treatment offered [72].
The current study found that CKD-EPI-based studies had a higher prevalence of CKD (all stages) than MDRD-based studies (17.6% vs. 14.9%). Furthermore, we found that CKD stages 3–5 were more common in MDRD-based research (8.3%) than in CKD-EPI-based studies (6.6%). The prevalence estimates of CKD can differ based on the research design, blood creatinine and urine albumin tests, eGFR calculation method, and CKD definitions/classifications [31]. The CKD-EPI equation is thought to produce a more precise estimation of eGFR than the MDRD equation, especially when higher eGFR levels [77,78] and multi-ethnic groups are considered [79]. Matsushita et al. [78] reported that the CKD-EPI equation was associated with fewer patients diagnosed with CKD than the MDRD. Furthermore, previous studies have demonstrated that the CKD-EPI equations may accurately predict the probability of death and ESRD, outperforming MDRD [78,80]. A previous study found that MDRD may underestimate eGFR ≥60 mL/min/1.73 m2 [80]. The CKD-EPI was developed in 2009 to address the limitations of the MDRD equation [81]. CKD-EPI is more effective than MDRD for identifying eligible patients with CKD, particularly those with stages 3 and 4 [77]. However, one common shortcoming of the MDRD and CKD-EPI equations is that data are predominantly collected from Caucasians and African Americans [82–84]. There are limited data on the effectiveness of the revised CKD-EPI equation in Asian populations [85]. However, earlier studies have explored the performance of MDRD and CKD-EPI in different Asian populations (India, Korea, and Pakistan), with CKD-EPI consistently outperforming MDRD in measuring eGFR [77,85–87]. In recent years, an Asian racial coefficient has been introduced to validate this equation [82–84]. In studies of Chinese and Japanese populations [82–84], modifying the MDRD race coefficient resulted in considerable improvements in eGFR estimates. Therefore, using appropriate methodologies based on race and CKD stage may lead to improved diagnosis and interpretation.
This study has a few limitations. Most studies included in this review were cross-sectional, which may limit our knowledge of the chronic nature of CKD. Because of the paucity of longitudinal data in the studies included in this analysis, we could not determine temporal or political patterns in the prevalence and dynamic geographical distribution of CKD among Asian patients across time. Another limitation of this meta-analysis was the inclusion of biased studies (18 studies with a moderate risk of bias and 16 with a high risk of bias). The most problematic domain was study confounding, with 35 of the 42 studies showing moderate to high risk of bias in this domain. The study design and analysis did not account for important potential confounders, such as sex, age, educational level, income, marital status, and comorbidities. The reported prevalence of CKD was probably biased by confounding factors because the relationships between CKD prevalence and these potential confounders were not adjusted, resulting in values that do not reflect the true relationship. In this case, a false positive (Type I) error or misleading conclusion that the dependent variables are causally related to the independent variable can occur [88]. Thus, confounding poses significant risks to the validity of causal inferences (internal validity) [88]. To address this issue, we presented the aCKD and sCKD in this study. Furthermore, most of the included studies did not aim to reflect the general population of each country. Some studies only included residents of rural and suburban areas, whereas others only included urban residents, implying that sampling may influence the prevalence of CKD in a particular country. To correctly represent the CKD status in each country, additional research, including a wider range of social populations, is required.
In addition, population-level estimates of CKD prevalence are influenced by variations in screening methodologies, eGFR measurements, CKD classifications, and limited resources in some countries. Therefore, pooled estimates and differences attributable to demographic characteristics should be evaluated using these variables. The included studies applied different equations for GFR estimates, even in those that used CKD-EPI estimation (CKD-EPI 2009, CKD-EPI 2012cys, CKD-EPI 2012sCr-cys, and Chinese or Japanese modified CKD-EPI), making comparison difficult [89]. Furthermore, these eGFR equations can be influenced by non-GFR-related parameters, such as muscle mass, which impact men and women differently, especially during aging [14]. Standardizing the eGFR equations and modifications is critical for maintaining consistency. Furthermore, China and India, the two most populous countries, do not have adequate data on national studies and CKD registries [76]. Therefore, the limited availability of nationally representative data does not rule out the possibility of within-country variations [76]. Furthermore, the number of publications in some countries was limited, implying that the prevalence of CKD may have been underestimated. Furthermore, several studies have calculated eGFR using only a single creatinine measurement and have excluded follow-up measurements for establishing CKD. Because of the high number of false-positive results, the prevalence of CKD may have been overestimated.
This study revealed that the prevalence of CKD varies greatly across Asia, with significant differences based on sex and geographical location. Furthermore, discrepancies in screening methodologies, eGFR measures, and resource availability in certain regions may impact population-level estimations of CKD prevalence. Therefore, in this context, understanding aggregate estimates and deviations caused by demographic factors is important. Further research is needed to determine the origins of these discrepancies, as well as evidence-based screening measures to reduce the overall disease burden.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Supplement 1
Print
