Changing etiological spectrum of cirrhosis in China: a systematic review and meta-analysis

Abstract

Background

Cirrhosis remains a major public health challenge in China and has historically been driven mainly by chronic viral hepatitis. However, alcohol-related liver disease and metabolic liver disease have received increasing attention in recent decades. Changes in hepatitis B vaccination, antiviral therapy, blood safety, alcohol consumption, and metabolic risk factors may have altered the reported etiological profile of cirrhosis. Existing evidence remains largely region-specific, hospital-based, or limited to particular time periods. This systematic review and meta-analysis aimed to characterize the reported etiological spectrum of cirrhosis in mainland China and to describe temporal and regional patterns.

Methods

We searched PubMed, Scopus, Embase, Web of Science, CNKI, Wanfang Data, and Sinomed for studies published from January 1990 to December 2025. Observational studies reporting the etiological distribution of cirrhosis in adults or children in mainland China were included. For adults, the principal etiological ranking was based on mutually exclusive categories, with single reported etiologies and explicitly reported mixed etiologies analyzed separately. Temporal and regional analyses used inclusive factor-related groupings, in which mixed etiologies contributed to each relevant etiological factor. For children, broad reported etiological categories were summarized because many pediatric studies did not fully subclassify post-hepatitic cirrhosis by viral subtype. Pooled proportions were estimated using random-effects models, and heterogeneity was assessed using the I² statistic.

Results

A total of 71 studies were included, comprising 59 studies contributing adult data and 12 studies contributing pediatric data. In adults, hepatitis B virus (HBV)-only cirrhosis was the most frequently reported single etiology, with a pooled proportion of 54.5% (95% CI: 51.2–57.8), followed by alcohol-only cirrhosis at 11.8% (95% CI: 10.1–13.6) and hepatitis C virus (HCV)-only cirrhosis at 5.9% (95% CI: 4.3–7.7). Cryptogenic cirrhosis accounted for 5.3% (95% CI: 4.1–6.6), and schistosomiasis-related cirrhosis accounted for 3.9% (95% CI: 0.7–9.4). In children, viral/post-hepatitic cirrhosis was the most frequently reported broad category, accounting for 47.9% (95% CI: 33.3–62.6). In inclusive temporal and regional analyses among adults, HBV-related cirrhosis remained the most frequently reported factor-related grouping across study periods and regions, while alcohol-related cirrhosis showed non-monotonic temporal variation and remained an important contributor in the recent period.

Conclusion

HBV remains the leading reported etiology of cirrhosis in mainland China, although the reported etiological profile has become more diverse over time. Alcohol-related cirrhosis, HCV-related regional burden, and under-recognized metabolic liver disease warrant further attention. Given substantial heterogeneity, these estimates should be interpreted as descriptive summaries rather than nationally uniform proportions. These findings support continued HBV prevention and treatment, expanded HCV screening, linkage to care, and direct-acting antivirals (DAAs) treatment coverage, improved ascertainment of metabolic liver disease, and more standardized reporting of mixed-etiology cirrhosis.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12876-026-05009-8.

Introduction

Cirrhosis is the consequence of chronic liver disease characterized by fibrosis and liver dysfunction [1], which can eventually lead to liver failure and substantially increase the risk of hepatocellular carcinoma (HCC). Globally, cirrhosis is a major cause of morbidity and mortality [2, 3], ranking as the 11th leading cause of death and accounting for approximately 1.3 million deaths annually. The global prevalence of compensated and decompensated cirrhosis combined has been estimated at approximately 1.5%–2.0% of the population, corresponding to more than 120 million people worldwide [3]. In China, which bears one of the world’s heaviest burdens of liver disease [4], recent estimates indicate that more than 7 million people are living with cirrhosis [5], largely driven by chronic viral hepatitis, alcohol use, and metabolic disorders. The age-standardized mortality rate of cirrhosis in China remains above the global average despite recent progress in prevention and treatment of underlying etiologies [6].

Cirrhosis has diverse causes, and understanding their distribution is critical for prevention and management. Chronic viral hepatitis remains the leading cause. In China, successive updates to national CHB guidelines [7]and the inclusion of first-line antivirals in national medical insurance have greatly improved the affordability and accessibility of HBV treatment. For HCV, direct-acting antivirals achieve cure rates above 95%, and their inclusion in national medical insurance since 2019 has reduced costs by more than 85% [8]. For both infections, further progress hinges on broader treatment coverage. Additionally, lifestyle and metabolic changes associated with rapid urbanization have contributed to higher rates of excessive alcohol consumption, nonalcoholic steatohepatitis (NASH), and metabolic syndrome, further driving the incidence of cirrhosis.^[6] These conditions are now emerging as important contributors to cirrhosis in China and worldwide. Furthermore, idiosyncratic drug-induced liver injury (DILI) [9], often associated with the use of certain herbal remedies and dietary supplements, represents a notable but under-recognized contributor to the burden of cirrhosis in some populations.

Despite the growing health impact of cirrhosis, comprehensive data on cause-specific mortality remain limited [10]. Most countries lack a dedicated cirrhosis registry, and mortality statistics rarely differentiate among underlying causes. To address this gap, the World Health Organization (WHO) recommends integrating cause-specific cirrhosis mortality—particularly those attributable to viral hepatitis, alcohol, and metabolic disorders—into national vital registration systems [11]. Until such surveillance systems become more complete, systematic synthesis of available epidemiological studies can provide useful evidence for understanding the etiological composition of cirrhosis and informing targeted prevention strategies. Therefore, this systematic review and meta-analysis aimed to characterize the reported etiological spectrum of cirrhosis in mainland China, separately for adults and children, and to describe temporal and regional patterns using prespecified etiological classification frameworks.

Methods

Search strategy for identification of studies

This study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 (PRISMA 2020) reporting guideline [12], and the protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO ID: CRD420251125172). We conducted a comprehensive search of English databases—PubMed, Scopus, Embase, and Web of Science—and three Chinese databases—CNKI, Wanfang Data, and Sinomed—from January 1990 to December 2025. The PubMed strategy combined Medical Subject Headings (MeSH) and free-text terms related to “cirrhosis,” “etiology,” and “China.” The complete search strategies for all English and Chinese databases are provided in Supplementary Material 7. Titles and abstracts of retrieved records were independently screened by two reviewers. Full texts were then evaluated for eligibility. Reference lists of included studies were also searched for additional relevant articles, and disagreements were resolved by a third reviewer.

Eligibility criteria

Studies were included in this review if they met the following criteria:

1. The study participants were patients diagnosed with liver cirrhosis;

2. Studies reported the etiological distribution of cirrhosis and provided quantifiable case counts for each etiology; for adult studies, the total sample size was required to be at least 50 cases.

3. The articles were published in full text either in English or Chinese, from January 1990 to December 2025.

4. The study population was based in mainland China.

We excluded the following studies on the basis of titles, abstracts, and full-text review, as appropriate:

1. Publication types including letters, commentaries, editorials, systematic reviews/meta-analyses, narrative reviews, guidelines, animal or cellular experiments, conference abstracts, early-access items, corrections, retracted publications, and protocols;

2. Studies were excluded if cirrhosis was not the primary disease of interest. Studies restricted to selected clinical subgroups, such as hepatocellular carcinoma, liver transplantation candidates, surgical cohorts, or patients with specific cirrhosis-related complications, were also excluded because these populations may not reflect the general etiological spectrum of cirrhosis.

When multiple publications appeared to report overlapping populations from the same hospital or region during the same study period, the study with the largest sample size, the most complete etiologic information, or the longest study period was retained.

Data extraction and quality assessment

For each study, we extracted the first author, publication year, region, study period, sample size, study population, and reported cirrhosis etiologies. Etiologies were first recorded as reported in the original studies. Data extraction and quality assessment were independently performed by two reviewers using standardized forms and appropriate Joanna Briggs Institute (JBI) critical appraisal tools for observational prevalence/cross-sectional evidence [13], with disagreements resolved by a third reviewer.

For the adult main results table, sole etiologies and explicitly reported mixed etiologies were presented as mutually exclusive categories. For pediatric studies, broad reported categories, such as post-hepatitic cirrhosis and viral subtype-specific estimates were summarized separately when subtype information was available, because many studies did not fully subclassify post-hepatitic cases by virological subtype. The main characteristics and quality assessment results of the included studies are summarized in Supplementary Material 3 and 6.

Etiologic classification framework

To address differences in etiological reporting between adult and pediatric cirrhosis studies, etiologies were harmonized using age-specific classification frameworks. Adult and pediatric studies were analyzed separately because the etiological spectrum, reporting granularity, and clinical interpretation differed substantially between the two populations.

For adult studies, two prespecified etiological classification frameworks were used, each serving a different analytical purpose. For the principal adult etiological ranking, patient-level mutually exclusive categories were used. Under this framework, each patient was counted only once and assigned to either a single reported etiology, such as HBV-only cirrhosis, HCV-only cirrhosis, alcohol-only cirrhosis, NAFLD-related cirrhosis, primary biliary cholangitis, autoimmune cirrhosis, or cryptogenic cirrhosis, or to a distinct mixed-etiology category, such as HBV–alcohol, HBV–HCV, HBV–HDV, HCV–alcohol, or HBV–HCV–alcohol. Mixed-etiology cases were not redistributed into the corresponding single-etiology categories. This framework was used to avoid double counting and to provide the main adult etiological ranking.

For adult temporal and regional analyses, inclusive factor-related groupings were used. Under this framework, mixed-etiology cases contributed to each relevant etiological factor. For example, HBV–alcohol cirrhosis contributed to both the HBV-related and alcohol-related groupings, and HCV–alcohol cirrhosis contributed to both the HCV-related and alcohol-related groupings. Therefore, estimates from temporal and regional analyses were interpreted as factor-related proportions rather than mutually exclusive etiological distributions, and should not be summed across categories. Throughout the manuscript, “HBV-only,” “HCV-only,” and “alcohol-only” cirrhosis refer to mutually exclusive single-etiology categories, whereas “HBV-related,” “HCV-related,” and “alcohol-related” cirrhosis refer to inclusive factor-related groupings.

For pediatric studies, a separate cause-based framework was used because pediatric cirrhosis showed a more heterogeneous and fragmented etiological spectrum. Many pediatric etiologies were rare, and several specific diagnoses were reported in only one or a few cases. Therefore, analyzing each individual pediatric diagnosis separately would have produced very small subgroup counts and unstable estimates. To improve interpretability, pediatric etiologies were harmonized into six broad clinically meaningful cause-based categories according to the reported underlying cause of cirrhosis [14–16]: viral/post-hepatitic cirrhosis, cholestatic/biliary cirrhosis, genetic/metabolic/nutritional cirrhosis, vascular/portal hypertension-related cirrhosis, schistosomiasis-related cirrhosis or hepatic fibrosis, and cryptogenic/other/mixed etiologies.

When pediatric studies reported specific diagnoses, these were mapped to the corresponding broad cause-based category. When studies reported only broad categories, such as post-hepatitic cirrhosis, the original category was retained within the closest broad group. Because many pediatric studies did not fully subclassify post-hepatitic cirrhosis by viral subtype, broad viral/post-hepatitic cirrhosis and subtype-specific viral estimates were interpreted separately rather than as additive categories. This pediatric classification was intended to provide a descriptive summary of the major etiological patterns in children, rather than a detailed ranking of rare individual pediatric disorders.

Statistical analysis

Proportion meta-analyses were conducted separately for each etiology. For each etiology, the study-specific proportion was calculated as the number of cases attributed to that etiology divided by the total number of cirrhosis cases in studies reporting that etiology. Studies that did not report a specific etiology were excluded from the corresponding synthesis, whereas studies explicitly reporting zero cases were retained. Because each etiological estimate was based on studies reporting that specific etiology, denominators differed across etiologies, and pooled proportions were not summed across rows [17].

Because several etiologies were rare or included zero-event studies, proportions were Freeman–Tukey double-arcsine transformed before pooling [17, 18]. The transformed proportions were synthesized using inverse-variance random-effects models [19], with between-study variance estimated using the DerSimonian–Laird method in the primary analysis. Pooled estimates were subsequently back-transformed to the original proportion scale and reported as percentages with 95% confidence intervals. Sensitivity analyses using alternative random-effects estimators, including restricted maximum likelihood and Paule–Mandel methods, were performed for the major adult etiologies to assess the robustness of the findings. Prediction intervals were calculated for major adult etiologies to illustrate the expected range of etiological proportions across comparable study settings. Heterogeneity was assessed using I² and Cochran’s Q test, with P < 0.10 considered statistically significant. Because the included studies differed in geographic region, study period, clinical setting, and etiological ascertainment, pooled estimates were interpreted as descriptive summary estimates across heterogeneous study settings.

Subgroup analyses by geographic region and study period were conducted to describe regional and temporal patterns. For geographic subgroup analyses, provinces were grouped into North, Northeast, East, Central, South, Southwest, and Northwest China according to commonly used administrative regional classifications. For temporal subgroup analyses, studies were assigned to predefined calendar-period subgroups according to the patient enrollment or case collection period reported in the original articles. When etiology-specific data were reported for separate periods, these data were extracted separately and analyzed within the corresponding calendar-period subgroup. When only an overall enrollment or case collection interval was reported, the midpoint between the start and end years was used as the representative study year. To improve the stability and interpretability of subgroup estimates, calendar periods with very few contributing studies were combined with the adjacent period; therefore, the most recent period was presented as 2015–2025 in the main temporal analysis. Leave-one-out analyses are presented in Supplementary Material 4. Sensitivity analyses using alternative random-effects estimators, prediction intervals, funnel plots, Egger’s tests, and trim-and-fill analyses are presented in Supplementary Material 5.

Small-study effects and potential publication bias were assessed exploratorily by funnel plot inspection and, when at least 10 studies were available for a given etiology, by Egger’s regression test and trim-and-fill analysis [20, 21]. Data were recorded in Excel and analyzed using R version 4.2.0 with the meta package, including the metaprop function for proportion meta-analysis.

Results

The PRISMA flow diagram is shown in Fig. 1. Ultimately, 71 studies were included in the systematic review and meta-analysis, including 59 studies contributing adult data and 12 studies contributing pediatric data.

Fig. 1.

PRISMA flow diagram of study selection

The included studies covered different calendar periods, geographic regions, and clinical settings in mainland China. Most studies were hospital-based observational studies. The JBI quality assessment results are summarized in Supplementary Material 6. Overall, the included studies provided extractable etiological data, although many were retrospective and hospital-based, with variable reporting of diagnostic criteria and etiological ascertainment. These limitations were considered when interpreting the pooled estimates.

Cirrhosis in children

We synthesized pediatric data comprising 518 children overall. In pediatric studies, broad reported categories and viral subtype-specific estimates were interpreted separately.

Viral/post-hepatitic cirrhosis was the most frequently reported broad category, accounting for 47.9% (95% CI: 33.3–62.6; 12 studies). Other broad categories included cholestatic/biliary cirrhosis (18.3%, 95% CI: 10.3–27.7; 11 studies), genetic/metabolic/nutritional cirrhosis (13.7%, 95% CI: 9.6–18.4; 12 studies), cryptogenic/other/mixed etiologies (13.3%, 95% CI: 9.1–18.0; 10 studies), vascular/portal hypertension-related cirrhosis (7.8%, 95% CI: 4.8–11.5; 7 studies), and schistosomiasis-related cirrhosis/fibrosis (6.4%, 95% CI: 0.0–39.1; 2 studies). Details are shown in Table 1 and Supplementary Material 1.

Table 1.

Etiological spectrum of cirrhosis in children in China

Etiology	No. of studies	Cases (n)	Total sample size (N)	Pooled proportion % (95% CI)	I² (%)
Overall pediatric population	12	—	518	—	—
Viral/post-hepatitic cirrhosis	12	230	518	47.9 (33.3–62.6)	90.9
Cholestatic/biliary cirrhosis	11	100	484	18.3 (10.3–27.7)	82.5
Genetic/metabolic/nutritional cirrhosis	12	72	518	13.7 (9.6–18.4)	46.5
Cryptogenic/other/mixed etiologies	10	68	451	13.3 (9.1–18.0)	41.6
Vascular/portal hypertension-related cirrhosis	7	28	334	7.8 (4.8–11.5)	11.8
Schistosomiasis-related cirrhosis	2	20	177	6.4 (0.0–39.1)	96.8

This table presents pooled proportions for broad reported etiological categories among pediatric cirrhosis cases. Because many pediatric studies did not fully subclassify post-hepatitic cirrhosis by viral subtype, broad categories and viral subtype-specific estimates were interpreted separately. Denominators differ across etiologies because each pooled estimate was based on studies reporting that specific category; therefore, pooled proportions should not be summed across rows

Cirrhosis in adults

In adult studies, substantial between-study heterogeneity was observed for most etiologies; therefore, pooled estimates were calculated using a random-effects model. In the principal adult analysis, etiologies were summarized using mutually exclusive categories, with explicitly reported mixed etiologies retained as separate categories rather than redistributed into single-etiology groups.

Among mutually exclusive adult etiologies, HBV-only cirrhosis was the leading sole etiology in China, with a pooled proportion of 54.5% (53 studies; 126,308/187,721; 95% CI: 51.2–57.8). Alcohol-only cirrhosis ranked second, accounting for 11.8% (56 studies; 20,072/209,068; 95% CI: 10.1–13.6), followed by HCV-only cirrhosis at 5.9% (50 studies; 14,599/186,923; 95% CI: 4.3–7.7). Cryptogenic cirrhosis accounted for 5.3% (30 studies; 4,831/146,777), and schistosomiasis-related cirrhosis for 3.9% (16 studies; 2,039/100,502). Other sole etiologies were less common, autoimmune cirrhosis (3.5%), PBC (3.4%), cholestatic cirrhosis reported as a separate category (1.7%), NAFLD-related cirrhosis (1.2%), drug-induced cirrhosis (0.7%), Wilson’s disease (0.6%), cardiac cirrhosis (0.5%), and Budd–Chiari syndrome (0.4%). Mixed etiologies were also reported, particularly HBV–alcohol (8.4%), HBV–HCV (1.1%), and HBV–HDV (1.3%). Less frequently reported mixed categories included HBV–HEV (0.9%), HCV–alcohol (0.6%), and HBV–HCV–alcohol (0.8%). These mixed-etiology categories were retained as distinct mutually exclusive categories and were not redistributed into the corresponding single-etiology groups. Detailed forest plots for all etiological categories are presented in Supplementary Material 2.

Table 2.

Pooled proportions of mutually exclusive etiological categories among adults with cirrhosis in China

Etiology	No. of studies	Cases (n)	Total sample size (N)	Pooled proportion % (95% CI)	I² (%)
Overall population	59	-	212,150	-	-
Single reported etiology
HBV-only	53	126,308	187,721	54.5 (51.2–57.8)	99.4
Alcohol-only	56	20,072	209,068	11.8 (10.1–13.6)	99.2
HCV-only	50	14,599	186,923	5.9 (4.3–7.7)	99.5
Schistosomiasis	16	2,039	100,502	3.9 (0.7–9.4)	99.8
PBC	25	935	20,252	3.4 (2.1–5.0)	96.8
Autoimmune cirrhosis	31	10,420	188,021	3.5 (2.2–5.0)	99.5
Cholestatic cirrhosis	13	776	43,398	1.7 (1.0–2.6)	94.2
NAFLD	20	464	112,630	1.2 (0.6–2.1)	98.1
Drug-induced cirrhosis	21	836	122,310	0.7 (0.4–1.1)	95.6
Wilson’s disease	16	249	112,014	0.6 (0.4–1.0)	93.9
Cardiac cirrhosis	13	115	108,132	0.5 (0.2–0.9)	94.1
Budd–Chiari syndrome	15	64	15,691	0.4 (0.2–0.6)	56.4
Cryptogenic	30	4,831	146,777	5.3 (4.1–6.6)	98.6
Mixed reported etiologies
HBV–Alcohol	28	2,178	30,692	8.4 (5.6–11.7)	98.9
HBV–HCV	25	1,755	145,252	1.1 (0.8–1.4)	94.1
HBV–HDV	10	182	15,613	1.3 (0.4–2.7)	97.1
HBV–HEV	4	38	3,820	0.9 (0.2–1.9)	84.1
HCV–Alcohol	14	91	15,822	0.6 (0.3–1.1)	88.6
HBV–HCV–Alcohol	3	42	4,207	0.8 (0.0–2.5)	95.0

This table presents pooled proportions for reported mutually exclusive adult etiological categories in the principal adult analysis. The “Overall population” row represents the total number of adult cirrhosis cases across all adult studies. For each etiological category, the total sample size represents the sum of cirrhosis cases from studies that reported that specific etiology. Therefore, denominators differ across etiologies, and pooled proportions should not be summed across rows. Single reported etiologies and explicitly reported mixed etiologies were analyzed separately, and mixed-etiology cases were not redistributed into single-etiology categories

Temporal trends of cirrhosis in adults

Temporal analyses were based on inclusive burden-oriented groupings rather than mutually exclusive etiologic categories. Mixed etiologies could contribute to more than one factor-related grouping; therefore, these estimates should be interpreted as factor-specific burdens rather than mutually exclusive proportions.

As shown in Fig. 2, HBV-related cirrhosis remained the dominant factor-related grouping across all calendar-period subgroups, with pooled proportions of 83.0% in ≤ 1999, 71.7% in 2000–2004, 61.2% in 2005–2009, 62.1% in 2010–2014, and 54.1% in 2015–2025. Alcohol-related cirrhosis, based on the inclusive alcohol-related grouping, showed non-monotonic temporal variation, with pooled proportions of 24.2%, 29.4%, 14.8%, 15.5%, and 19.1% across the same periods. HCV-related cirrhosis accounted for 5.6%, 10.2%, 6.8%, 7.2%, and 9.0%, respectively, with the highest estimate observed in 2000–2004. NAFLD-related cirrhosis remained relatively uncommon across calendar-period subgroups with available data, with period-specific estimates ranging from 0.8% to 4.4% and showing no clear monotonic temporal pattern. Immune-related etiologies increased from 0.4% in ≤ 1999 to 6.2% in 2015–2025, while PBC increased from 0.3% in ≤ 1999 to 6.6% in 2015–2025. Cryptogenic cirrhosis also showed temporal variation, with the highest pooled proportion observed in 2015–2025 (9.6%). Overall, the reported etiological composition of adult cirrhosis appeared to become more diverse over time, although HBV-related cirrhosis remained the most frequently reported factor-related grouping throughout the study period.

Fig. 2.

Temporal patterns of inclusive etiological groupings among adults with cirrhosis in mainland China, ≤ 1999–2025. Note: Temporal analyses were based on inclusive factor-related groupings. Mixed-etiology cases contributed to each relevant etiological factor; for example, HBV–alcohol cirrhosis contributed to both the HBV-related and alcohol-related groupings. Therefore, the proportions shown in this figure represent factor-related proportions and should not be summed across etiological categories. Calendar periods were assigned according to the reported study years or, when only aggregated data were available, to the period that best matched the original study period

Regional patterns of cirrhosis in adults

Regional analyses used the same inclusive factor-related framework. Therefore, regional estimates represent factor-related proportions rather than mutually exclusive etiological distributions, and alcohol-related cirrhosis refers to the inclusive alcohol-related grouping. As shown in Fig. 3, HBV-related cirrhosis was the most frequently reported factor-related grouping across all regions, with the highest pooled proportion in Northwest China (73.5%), followed by South China (70.3%), Southwest China (66.2%), and East China (63.7%). HCV-related cirrhosis had the highest reported pooled proportion in Northeast China (12.5%) and the lowest reported pooled proportion in East China (2.0%). Alcohol-related cirrhosis was more common in Southwest China (21.3%), East China (19.7%), Northeast China (17.2%), and North China (16.5%) than in Northwest China (10.9%), South China (10.6%), and Central China (9.1%). NAFLD-related cirrhosis remained uncommon across regions and peaked in Southwest China (1.7%). Immune-related etiologies were more frequent in North China (6.1%), Northwest China (4.4%), and Northeast China (4.2%), whereas South China showed the lowest pooled proportion (0.5%). Schistosomiasis-related cirrhosis was highly concentrated in Central China (15.5%) and East China (6.6%), representing a distinct regional burden. These findings indicate substantial geographic heterogeneity in the etiological composition of adult cirrhosis in mainland China.

Fig. 3.

Regional patterns of inclusive etiologic groupings among adults with cirrhosis in mainland China. Note: Regional analyses used inclusive factor-related groupings. Mixed-etiology cases were assigned to each relevant etiological factor; therefore, regional estimates represent factor-related proportions rather than mutually exclusive etiological distributions. The proportions shown in this figure should not be summed across categories. Provinces were grouped into North, Northeast, East, Central, South, Southwest, and Northwest China according to commonly used administrative regional classifications. Regional differences should be interpreted descriptively because included studies varied in clinical setting, study period, patient source, and etiological ascertainment

Publication bias and sensitivity analyses

Across the DerSimonian–Laird, restricted maximum likelihood, and Paule–Mandel random-effects estimators, the pooled estimates for the major mutually exclusive adult etiologies were broadly consistent, and the overall etiological ranking remained largely unchanged. For example, the pooled estimates for HBV-only cirrhosis were 54.49%, 54.25%, and 54.25% using the DerSimonian–Laird, restricted maximum likelihood, and Paule–Mandel methods, respectively. Corresponding estimates were 11.78%, 11.84%, and 11.84% for alcohol-only cirrhosis, and 5.90%, 5.90%, and 5.90% for HCV-only cirrhosis. Similar consistency was observed for the other major etiological categories. Detailed estimator-specific sensitivity analyses and prediction intervals are provided in Supplementary Material 5.

Prediction intervals were relatively wide for several major etiologies, reflecting between-study variability and supporting the interpretation of pooled proportions as descriptive summary estimates across heterogeneous study settings rather than as fixed proportions applicable to all populations. This heterogeneity was likely related to differences in clinical setting, geographical region, study period, patient source, etiological ascertainment, reporting practices, and the availability of serological, imaging, histological, or metabolic data. Therefore, the high I² values should be interpreted as reflecting clinical and methodological diversity rather than instability driven by any single study.

Leave-one-out analyses indicated that no single study materially influenced the main pooled estimates. After sequentially excluding one study at a time, the pooled estimate ranged from 52.37% to 54.03% for HBV-only cirrhosis, from 10.93% to 11.53% for alcohol-only cirrhosis, and from 5.69% to 6.33% for HCV-only cirrhosis. Similar stability was observed for the other major adult etiological categories. Detailed leave-one-out analyses are provided in Supplementary Material 4.

Small-study effect assessments were considered exploratory because of the substantial between-study heterogeneity and the observational nature of the included studies. Egger’s regression test suggested funnel plot asymmetry for HBV-only cirrhosis, alcohol-only cirrhosis, cryptogenic cirrhosis, and NAFLD-related cirrhosis, whereas no statistically significant asymmetry was detected for HCV-only cirrhosis, schistosomiasis-related cirrhosis, primary biliary cholangitis, or autoimmune cirrhosis. In trim-and-fill analyses, imputed studies were observed for HBV-only, alcohol-only, and autoimmune cirrhosis, with adjusted pooled estimates of 58.58%, 9.44%, and 3.99%, respectively. No studies were imputed for HCV-only, cryptogenic, schistosomiasis-related, primary biliary cholangitis, or NAFLD-related cirrhosis. Detailed funnel plots, Egger’s tests, and trim-and-fill analyses are provided in Supplementary Material 4. Overall, these findings support the robustness of the main etiological pattern, particularly the predominance of HBV-only cirrhosis followed by alcohol-only cirrhosis, although the pooled estimates should be interpreted with consideration of between-study heterogeneity and the exploratory nature of small-study effect assessments.

Discussion

This systematic review and meta-analysis summarized the reported etiological spectrum of cirrhosis in mainland China across adult and pediatric populations. In adults, HBV-only cirrhosis remained the most frequently reported mutually exclusive etiology, followed by alcohol-only cirrhosis and HCV-only cirrhosis. Temporal and regional analyses based on inclusive factor-related groupings showed that HBV-related cirrhosis remained the dominant factor-related category across study periods and geographic regions, whereas alcohol-related cirrhosis showed non-monotonic temporal variation, with the highest pooled proportion observed in the 2020–2025 subgroup of the included studies. HCV-related cirrhosis constituted a persistent regional burden, particularly in Northeast China. These findings are broadly consistent with the 2025 Chinese cirrhosis guideline, which identifies viral hepatitis, alcoholic liver disease, and metabolic dysfunction-associated fatty liver disease as major etiologies of cirrhosis in China [7]. Overall, the etiological spectrum of cirrhosis in China remains strongly shaped by viral hepatitis, particularly HBV, but has become increasingly diversified with the growing contribution of alcohol-related, immune-related, and metabolic liver diseases.

HBV remains the leading reported cause of cirrhosis in China across both mutually exclusive and inclusive analytical frameworks. National seroepidemiological surveys reported that the prevalence of HBsAg in the general population decreased from 9.72% in 1992 to 5.86% in 2020, with the prevalence among children under five declining from 9.67% to 0.30% [22]. In our temporal analysis, HBV-related cirrhosis accounted for 83.0% in ≤ 1999, 71.7% in 2000–2004, 61.2% in 2005–2009, 62.1% in 2010–2014, and 54.1% in 2015–2025. This overall declining but still fluctuating pattern may be better understood in the context of the long natural history of chronic HBV infection and the cohort effects of HBV control in China. Universal infant vaccination, prevention of mother-to-child transmission, safer blood supply, and expanded antiviral therapy are likely to have contributed to reductions in new HBV infections and changes in the epidemiological profile of HBV infection [7, 22]. These population-level changes provide an epidemiological context for interpreting the observed pattern, although our study was not designed to directly estimate the causal impact of vaccination, prevention of mother-to-child transmission, safer blood supply, or antiviral therapy on cirrhosis burden. However, their effect on cirrhosis burden may take decades to become fully apparent, because cirrhosis usually develops after prolonged chronic infection. Many individuals infected with HBV before widespread vaccination and prevention of mother-to-child transmission may now be reaching middle or older age, when the cumulative risk of cirrhosis becomes higher. Therefore, even as new HBV infections decline, the burden of HBV-related cirrhosis may continue to persist among previously infected aging adults. In addition, improved HBV testing and etiological ascertainment in more recent studies, incomplete diagnosis and antiviral treatment coverage among adults, and differences in geographic coverage, hospital level, referral patterns, and patient case mix across study periods may have contributed to the observed variation. These findings suggest that HBV control has likely reduced the future pool of new infections, but a substantial residual burden of HBV-related cirrhosis may remain among chronically infected adults. Sustained vaccination and prevention of mother-to-child transmission should therefore be accompanied by expanded adult HBV screening, linkage to care, and antiviral treatment to further reduce the existing burden of HBV-related cirrhosis.

Alcohol-related cirrhosis represented the second most frequently reported mutually exclusive adult etiology and showed a non-monotonic rather than strictly linear temporal pattern in the inclusive analyses.Alcohol-related cirrhosis accounted for 24.2% in ≤ 1999, 29.4% in 2000–2004, 14.8% in 2005–2009, 15.5% in 2010–2014, and 19.1% in 2015–2025. Regional analysis showed that alcohol-related cirrhosis was more frequent in Southwest China (21.3%), East China (19.7%), Northeast China (17.2%), and North China (16.5%) than in Northwest, South, and Central China. These findings may reflect regional variation in drinking patterns, socioeconomic conditions, healthcare access, referral patterns, and clinical recognition of alcohol-related liver disease. Although the temporal estimates fluctuated across calendar-period subgroups, alcohol-related cirrhosis remained an important contributor in the recent period, supporting the need for systematic alcohol-use screening [23], brief behavioral intervention in clinical settings, treatment of alcohol use disorder, and strengthened public health policies aimed at reducing harmful alcohol consumption [24].

HCV-only cirrhosis accounted for 5.9% (95% CI: 4.3–7.7) across 50 studies in the mutually exclusive adult analysis, while HCV-related cirrhosis showed the highest pooled proportion in Northeast China (12.5%) and the lowest pooled proportion in East China (2.0%) in the regional inclusive analysis. This regional concentration indicates that HCV-related cirrhosis should not be deprioritized simply because its pooled proportion was lower than that of HBV-related or alcohol-related cirrhosis. The public health importance of HCV lies not only in its burden but also in its treatability. China’s National Action Plan for Eliminating Hepatitis C as a Public Health Threat (2021–2030) set explicit targets for HCV RNA confirmation and antiviral treatment coverage [25], and price negotiations have substantially improved DAA affordability. Nevertheless, national DAA treatment uptake remained low in the early years after DAA availability [26]. Expanded HCV screening, HCV RNA confirmation, linkage to care, and broader DAA treatment coverage should therefore remain core components of cirrhosis prevention, particularly in regions with higher HCV-related burden.

NAFLD-related cirrhosis accounted for a relatively small pooled proportion in the mutually exclusive adult analysis (1.2%, 95% CI: 0.6–2.1), and temporal estimates remained low across calendar-period subgroups with available data, ranging from 0.8% to 4.4%. This finding should be interpreted cautiously [27]. Unlike HBV- or HCV-related cirrhosis, which can usually be supported by serological testing, NAFLD-related cirrhosis requires more comprehensive clinical information [28]. According to previous guidelines, NAFLD is generally diagnosed based on hepatic steatosis and exclusion of significant alcohol consumption and other secondary causes of steatosis, whereas MAFLD emphasizes hepatic steatosis with overweight/obesity, type 2 diabetes, or other metabolic abnormalities [29–31]. Therefore, identifying NAFLD/MAFLD-related cirrhosis at the cirrhotic stage requires information on prior or imaging-detected steatosis, metabolic risk factors, alcohol intake, and exclusion or differentiation of competing etiologies. From the perspective of disease burden, Younossi et al. [32] reported a global NAFLD prevalence of approximately 25.24% in the general population, with frequent comorbid obesity, type 2 diabetes, hyperlipidemia, hypertension, and metabolic syndrome. Disease-burden modelling projected that the number of patients with NAFLD in China would increase from approximately 244 million in 2016 to 315 million in 2030, with about 751,000 incident NAFLD-related decompensated cirrhosis cases during 2016–2030 [33]. Thus, the low reported proportion of NAFLD-related cirrhosis in the present review may partly reflect under-recognition in routine clinical practice and incomplete etiological reporting, given the more complex ascertainment required for NAFLD-related cirrhosis. 

This issue is also relevant to cryptogenic cirrhosis, which accounted for 5.3% of adult cases in our mutually exclusive analysis (4,831/146,777; 30 studies; 95% CI: 4.1–6.6; I² = 98.6%). Previous patient-level studies have suggested a possible overlap between cryptogenic cirrhosis and unrecognized NASH. Caldwell et al. reported [34] that 73% of patients with cryptogenic cirrhosis had type 2 diabetes and/or obesity, and Poonawala et al. reported [35] obesity and type 2 diabetes in 55% and 47% of cryptogenic cirrhosis patients, respectively. In selected liver transplantation or end-stage liver disease populations, NASH has been estimated to account for approximately 50%–75% of cryptogenic cirrhosis, although this estimate remains uncertain and population-specific [36]. However, these estimates cannot be directly applied to our meta-analysis to reassign cryptogenic cases to NAFLD/MAFLD-related cirrhosis. Given that typical features of steatosis and steatohepatitis may attenuate or disappear in advanced cirrhosis [37], cryptogenic cirrhosis was retained as a separate etiological category. Accordingly, the pooled proportion of NAFLD-related cirrhosis should be interpreted as the proportion explicitly reported in the original literature, while acknowledging possible underestimation when metabolic, imaging, alcohol-use, and histological information are incomplete.

Mixed etiologies were also clinically important. Reported mixed categories included HBV–alcohol (8.4%), HBV–HCV (1.1%), HBV–HDV (1.3%), HBV–HEV (0.9%), HCV–alcohol (0.6%), and HBV–HCV–alcohol (0.8%). Such overlaps are clinically important because multiple hepatic insults may accelerate disease progression and complicate prevention strategies. However, mixed etiologies were inconsistently reported across original studies [7]. The separation of mutually exclusive mixed categories in the principal adult ranking and the use of inclusive factor-related groupings in temporal and regional analyses allowed us to avoid double counting while preserving clinically meaningful burden-oriented information. Future epidemiological studies should report single and mixed etiologies explicitly so that both patient-level etiological ranking and factor-specific prevention priorities can be more accurately assessed.

Pediatric cirrhosis showed an etiological profile distinct from that observed in adults [14, 15], supporting the need for age-specific interpretation. Unlike adult studies, in which major etiologies were more consistently reported and could be summarized using mutually exclusive and inclusive frameworks, pediatric studies involved a broader and more fragmented spectrum of causes. Many pediatric etiologies, particularly congenital, cholestatic, genetic, metabolic, nutritional, vascular, and rare disease-related conditions, were reported in only one or a few cases. Therefore, pediatric etiologies were grouped into six broad cause-based categories to improve interpretability and avoid unstable estimates from excessive subdivision [15]. Among 518 children from 12 contributing studies, viral/post-hepatitic cirrhosis was the most frequently reported broad category, while cholestatic/biliary, genetic/metabolic/nutritional, vascular/portal hypertension-related, schistosomiasis-related, and cryptogenic/other/mixed etiologies also contributed to the pediatric spectrum. Because many pediatric studies reported post-hepatitic cirrhosis without complete viral subtype information, subtype-specific viral estimates should be interpreted separately rather than as additive components of the broad viral/post-hepatitic category. These findings suggest that pediatric cirrhosis should not be viewed simply as an extension of the adult etiological spectrum, and the pediatric estimates should be regarded as exploratory because of the limited number of studies, small sample size, and heterogeneous reporting practices.

Several limitations should be acknowledged. First, most included studies were hospital-based and retrospective, which may bias results toward more severe or advanced cases and limit generalizability to the broader community. Second, substantial heterogeneity was observed across most etiologies, likely reflecting differences in geographic region, study period, clinical setting, diagnostic criteria, and etiological ascertainment. Therefore, the pooled estimates should be interpreted as summary estimates across heterogeneous study settings rather than as uniform national proportions. Third, etiological classification varied across the original studies. Some studies used broad categories, such as post-hepatitic cirrhosis, and cryptogenic cirrhosis may have included undiagnosed NAFLD, autoimmune liver disease, alcohol-related liver disease, or other less well-recognized etiologies. Etiological misclassification may not have affected all etiologies equally. NAFLD-related cirrhosis was particularly vulnerable to under-recognition and misclassification as cryptogenic cirrhosis, whereas viral hepatitis-related etiologies were generally more likely to be identified through serological testing. Fourth, the numbers of contributing studies were small for several rare etiologies and for some pediatric categories, resulting in imprecise estimates and unstable heterogeneity statistics. Finally, although the use of mutually exclusive categories and inclusive burden-oriented groupings helped reduce double counting and preserve epidemiologically meaningful information, inconsistent reporting of mixed etiologies across studies posed challenges for standardized synthesis.These challenges also highlight the need for standardized etiological reporting in future cirrhosis epidemiology studies.

In conclusion, HBV remains the leading reported etiology of cirrhosis in mainland China, although the reported etiological profile has become more diverse over time. Alcohol-related cirrhosis showed non-monotonic temporal variation and remained an important contributor in the recent period, while HCV-related cirrhosis remained regionally important and potentially preventable. NAFLD-related cirrhosis was infrequently reported in the included studies, but its true contribution may be underestimated because of under-recognition and inconsistent etiological ascertainment. Given the substantial heterogeneity and predominance of hospital-based studies, these findings should be interpreted as descriptive summaries rather than nationally representative estimates. Continued HBV prevention and treatment, expanded HCV screening and DAA treatment coverage, systematic alcohol-use assessment, improved ascertainment of metabolic liver disease, and standardized reporting of mixed etiologies are needed to improve cirrhosis prevention and surveillance in China.

Abbreviations

CHB

Chronic hepatitis B

CI

Confidence interval

CNKI

China National Knowledge Infrastructure

DAA

Direct-acting antiviral

DILI

Drug-induced liver injury

HBV

Hepatitis B virus

HBsAg

Hepatitis B surface antigen

HCC

Hepatocellular carcinoma

HCV

Hepatitis C virus

HDV

Hepatitis D virus

HEV

Hepatitis E virus

JBI

Joanna Briggs Institute

MAFLD

Metabolic dysfunction-associated fatty liver disease

MASLD

Metabolic dysfunction-associated steatotic liver disease

MeSH

Medical Subject Headings

NAFLD

Nonalcoholic fatty liver disease

NASH

Nonalcoholic steatohepatitis

PBC

Primary biliary cholangitis

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-Analyses

PROSPERO

International Prospective Register of Systematic Reviews

RNA

Ribonucleic acid

WHO

World Health Organization

Authors’ contributions

C.H. and P.L. conceived the study. C.H. and P.L. independently performed the literature screening. X.F. resolved discrepancies during the screening process. C.H. and P.C. conducted the data analysis. J.L. supervised the project and interpreted the results. C.H. and X.L. drafted the manuscript. All authors critically reviewed and approved the final manuscript.

Funding

This work was supported by the Technical Guidance and Capacity Building for Prevention and Treatment of Chronic Viral Hepatitis (Grant No. 102393250020050000001).

Data availability

All data analyzed in this study were extracted from publicly available published articles. The extracted dataset and analytic code are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable. This study is a systematic review and meta-analysis of published literature and did not involve new data collection from human participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.