Prevalence of sapovirus infection among hospitalized pediatric patients in Asia: A systematic review and meta-analysis

Abstract

BACKGROUND

Acute gastroenteritis significantly affects children under five, with viral agents causing over 75% of cases. Sapovirus, from the Caliciviridae family, is increasingly recognized as a major cause of childhood diarrhea, especially in low and middle-income countries with high morbidity, hospitalization, and mortality rates.

AIM

To determine the prevalence of sapovirus infection among hospitalized pediatric patients in Asia.

METHODS

This study conducted a systematic review to investigate the presence of sapovirus in hospitalised pediatric patients suffering from acute gastroenteritis across Asian countries. The articles were retrieved from Scopus, PubMed, and Web of Science. The review followed PRISMA guidelines and was registered with PROSPERO (No. CRD420251029792). The quality of individual studies was assessed using Joanna Briggs Institute guidelines.

RESULTS

A total of 39 studies with 30800 observations and 567 sapovirus-positive cases were analyzed. The pooled prevalence of sapovirus in hospitalized pediatric patients was 1.73% (95% confidence interval: 1.33%-2.25%). Substantial heterogeneity was observed (I² = 84.9%), supporting the use of a random-effects model. Sensitivity analysis using 32 high-quality studies yielded a consistent prevalence of 1.74% (95% confidence interval: 1.31%-2.31%), reinforcing the robustness of the findings.

CONCLUSION

In Asian countries, sapovirus has been detected in hospitalized pediatric patients with a pooled prevalence rate of 1.73%, indicating its ongoing circulation and potential public health relevance in the region. Despite variability, consistent findings from sensitivity analysis underscore the need for enhanced surveillance and water quality monitoring to reduce public health risks.

Core Tip: Sapovirus is an underrecognized viral cause of acute gastroenteritis in hospitalized children across Asia. This systematic review and meta-analysis synthesizes evidence from 39 studies and demonstrates a pooled prevalence of approximately 1.7%, with consistent estimates across high-quality studies despite marked heterogeneity. The findings confirm ongoing sapovirus circulation in pediatric hospital settings and support the need for strengthened surveillance systems and preventive measures, particularly in regions with high enteric disease burden.

INTRODUCTION

Acute gastroenteritis (AGE) remains a leading health concern among children under five years of age. In over 75% of cases, viral pathogens are identified as the primary cause of acute infectious diarrhea, which typically presents with a sudden onset of diarrhea, often accompanied by symptoms such as vomiting, fever, abdominal pain, and dehydration[1,2]. It results from inflammation of the stomach and intestinal lining and is commonly caused by a variety of viral, bacterial, and parasitic pathogens[3]. Among these, viral agents-particularly Norovirus, Rotavirus, Astrovirus, Adenovirus, and sapovirus are the most frequently identified, especially in children under five years of age[4].

Sapoviruses (SaV) belonging to the Caliciviridae family, are increasingly recognized as significant etiological agents of AGE across all age groups, particularly in young children, the elderly, and immunocompromised individuals[5]. AGE remains a leading cause of morbidity and mortality worldwide, especially in low- and middle-income countries, where it contributes to considerable healthcare and economic burdens[6]. In Asia, AGE remains a major public health concern, with South Asia alone accounting for a substantial proportion of the nearly 450000 under-five diarrheal deaths reported globally each year[6,7].

SaV are non-enveloped, single-stranded positive-sense RNA viruses with a characteristic particle shape with cup-like depressions, commonly referred as the “Star of David”, as revealed by negative-stain electron microscopy[8,9]. It’s RNA genome is around 7 kb long and has a poly A sequence segment at the 3’ end. SaV, in contrast to noroviruses, encode the capsid protein near the vast non-structural polyprotein [open reading frame 1 (ORF1)][10]. The stop codon of ORF1 and the first AUG codon of ORF2 overlap by one or four nucleotides to form the junction that represents ORF1 and ORF2 of the norovirus. This results in a frameshift of ORF1. One 62 kDa polypeptide is encoded by the 3’ terminus of ORF1 classified into at least 19 genogroups (GI-GXIX), among which GI, GII, GIV, and GV are known to infect humans[11]. The viral genome encodes non-structural and structural proteins essential for replication and pathogenesis. Transmission primarily occurs via the fecal-oral route through contaminated food, water, surfaces, and person-to-person contact[12]. Sapovirus outbreaks have been reported in community settings such as childcare centres, schools, nursing homes, military camps, and healthcare facilities, often affecting vulnerable populations[13].

According to global disease burden estimates, sapovirus is documented as an emerging yet underreported agent of gastroenteritis in children under five, particularly in low- and middle-income countries. According to World Health Organization and affiliated surveillance data, sapovirus is a globally distributed enteric pathogen causing sporadic and institutional outbreaks of AGE in both community and healthcare settings. Following the decline in rotavirus infections due to vaccination, sapovirus has emerged as a notable viral cause of childhood diarrhea. World Health Organization emphasizes the need for enhanced molecular surveillance [real-time polymerase chain reaction (RT-PCR)], improved sanitation, and sustained global monitoring to reduce its public health burden[14]. Surveillance in all regions of the globe has linked sapovirus in both sporadic and community outbreaks[15]. Some prominent studies, such as those by GEMS or the Mal-ED cohort, have found that sapovirus remains among the top five viral agents causing moderate-to-severe diarrhea in children under five years[16]. Contrary to other enteric viruses, sapovirus tends to be present throughout the year, with seasonal peaks that vary from region to region[17]. This poses a challenge for the timely detection and implementation of preventive measures.

Although awareness of the viral agent continues to increase, so do the gaps in knowledge about sapovirus[18]. Differing diagnostic methods in each study, lack of standardized protocols, and poor routine screening for sapovirus-associated gastroenteritis contribute to this[19]. The level and type of sapovirus circulating in community/population may differ from one region to another, providing an added hurdle to detection; the varying levels of viral loads and host immune responses further complicate this[20]. Their co-infections with other common pathogens, such as rotavirus, norovirus, or certain bacteria, also complicate the diagnosis[20].

Data from community and hospital settings, especially in low- and middle-income countries, remain limited. A few systematic reviews and meta-analyses have previously explored the global distribution of sapovirus infections; the prevalence specifically among paediatric populations hospitalized for gastroenteritis in Asia is still uncertain[18,21]. Therefore, this study has been designed to perform an elaborated systematic review and meta-analysis to estimate pooled prevalence and/or incidence of sapovirus infections in hospitalized pediatric populations across Asia.

MATERIALS AND METHODS

This systematic review and meta-analysis was conducted using the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines, and this review is registered with the PROSPERO database (No. CRD420251029792).

Search strategy

Two reviewers independently searched for relevant articles published from 2000 to 2023 across three major databases: Scopus, PubMed, and Web of Science. The following search terms were used: “Sapovirus”, “HuSaV”, “prevalence”, “incidence”, “frequency”, “epidemiology”, “child”, “pediatric”, “infant”, “neonate”, “adolescent”, “hospitalization”, “inpatient”, “admitted”, “gastroenteritis”, “diarrhea”, “diarrhoea”, “stomach flu”, “intestinal infection”, and “Asia” (including all individual Asian country names); the complete search strings are provided in the Supplementary material, and the search was limited to English-language publications. Also, reference lists of all relevant original studies and reflective narrative articles were screened for relevant article.

Eligibility criteria

In this systematic review and meta-analysis, we included studies reporting the data of SaV prevalence among hospitalized pediatric patients in Asia.

Inclusion criteria

Population: Studies in pediatric patients (0-18 years) admitted with gastroenteritis or presenting with gastroenteritis symptoms where sapovirus is being investigated. Exposure: Exposure of interest in the presence of sapovirus in this study population as diagnosed by molecular methods (e.g., RT-PCR) or antigen/antibody detection assays. Geographical scope: Only studies from Asian countries to get regional data on sapovirus in pediatric hospital settings. Time period: Studies published from January 1, 2000 to December 31, 2023, covering the latest data on sapovirus epidemiology, diagnostics, and management in Asia.

Exclusion criteria

Research articles meeting the following criteria were excluded: (1) Studies done before 2000 or after 2023 were excluded; (2) Only hospitalized pediatric populations were included, and non-hospitalized pediatric populations were excluded; (3) Only research done in Asia was included to keep the study local and relevant to Asian settings; and (4) Case reports, editorials, commentary, conference abstracts, and review papers were excluded. We have conducted a comprehensive search for original research studies encompassing empirical data on the prevalence of sapovirus.

Study selection

All searches were uploaded to Google Spreadsheet, an online Google Workspace. Both the reviewers identified and removed duplicates. Both further reviewed the screen title and selected studies in search results against eligibility criteria. Then, both downloaded the full text of selected studies and did a full review to exclude more articles that met the exclusion criteria. Reasons for exclusion were documented for the excluded articles as full text.

Data extraction

Titles and abstracts of all retrieved citations were reviewed by two reviewers to exclude irrelevant studies. The full text of selected articles was then collected and reviewed by other reviewers and excluded if they did not meet the inclusion criteria. Finally, any disputes at any stage of the review were resolved by a third reviewer. Data was extracted on the year of publication, year of sampling, age of participants, diagnostic method, sample size number of SaV-positive cases, population (pediatric patients, hospitalized), study design, and study location. The data was then entered into a pre-designed Excel sheet.

Quality assessment

A Joanna Briggs Institute critical appraisal tool was used to assess the quality of the papers[22]. A customized checklist of nine questions with yes/no/not applicable answers was used. Two reviewers independently appraised each paper using this tool to ensure rigor and reduce bias. Where there were discrepancies between the reviewers, these were resolved through discussion and consensus with a third reviewer. This ensured a thorough and balanced appraisal of each paper. Papers scoring five or more out of nine on this customized checklist were considered of superior quality, i.e., having sound methodology and reliability.

Statistical analysis

The meta-analysis examined the proportionate occurrence of human sapovirus-associated gastroenteritis among hospitalized pediatric patients in Asian countries. This analysis was based on the selected research articles, from which a pooled prevalence estimate was computed. Generalized linear mixed models with logit transformation were used to determine the effect sizes for each study, as these models are more suitable in meta-analysis of single prevalence studies. The heterogeneity between studies was assessed using the I² statistic and Cochran’s Q test. Due to significant methodological differences among the studies, a random effects model was applied for the analysis. The confidence intervals (CIs) for effect sizes of individual studies were calculated using the Clopper-Pearson method, with all effect sizes and pooled estimates presented as proportions with 95%CIs. Publication bias was evaluated by visually inspecting funnel plots and using the Begg and Mazumdar rank correlation test.

Data extraction was conducted using Microsoft Excel 2021, and the statistical analyses were performed using the Meta and Metafor packages of R 4.5.0 software. Subgroup analyses were performed to identify sources of heterogeneity, considering factors such as geographic variations, age groups, and diagnostic methods. A sensitivity analysis was performed, removing studies with quality scores of five or less.

To explore sources of heterogeneity, mixed-effects meta-regression was performed using a generalized linear mixed models with logit-transformed proportions. Study-level moderators included age group, World Bank income category, publication period, and sample size category. Maximum likelihood estimation was employed. The overall significance of moderators was assessed using the omnibus QM test.

RESULTS

Study selection

We initiated our meta-analysis and systematic review by thoroughly searching three well-known databases: Scopus, Web of Science, and PubMed. An initial search in the database yielded 352 publications-180 from PubMed, 78 from Web of Science, and 94 from Scopus (Figure 1). To maintain veracity and avoid duplicate entries, the records were carefully checked, resulting in the removal of 30 duplicates. Thus 323 unique articles were left for preliminary screening. The titles and abstracts of the remaining articles were scanned to exclude any studies not in keeping with the objectives of this review. From this categorization, 200 articles were found to be worthy of full-text assessment, as they all appeared to fulfil the inclusion criteria. All 200 full-text articles underwent a thorough review to assess their relevance, study design, methodology, and conformity with the defined eligibility criteria. After rigorous screening, 39 studies were subjected to quality control using Joanna Briggs Institute critical appraisal tool. We found all these 39 studies scored more than 50%, and 82% included study having quality score more than 60%, These 39 fully eligible studies are included in the final systematic review and meta-analysis. These studies formed a firm and comprehensive base for analysis, lending strength to the findings. Published from 2000 to 2023, these studies deal with sapovirus gastroenteritis affecting the hospitalized pediatric population. The included 39 studies varied in sample size from 62 to 6504 cases. Out of these, 26 studies concerned children under five years of age, while the remaining 13 involved children aged from 5 years to 18 years. Cases of gastroenteritis in hospital-admitted children were subjected to a meta-analysis, with the publication dates of the studies ranging from 2000 to 2023. The studies were conducted in various Asian countries, with the highest number of research contributions coming from China (n = 8), India (n = 5), China Taiwan (n = 3), Thailand (n = 13), Vietnam (n = 3), Japan (n = 2), Qatar (n = 1), Lebanon (n = 1), Philippines (n = 1), Iran (n = 1), and Hong Kong (n = 1).

Figure 1.

PRISMA flow diagram for study selection process.

Prevalence of sapovirus-associated gastroenteritis

A total of 39 studies, comprising 30800 individual observations and reporting 567 sapovirus-positive events, were included in the meta-analysis to estimate the pooled prevalence of sapovirus infection. Using a random-effects model based on a random intercept logistic regression approach with a logit transformation, the overall pooled prevalence was found to be 1.73% with a 95%CI ranging from 1.33% to 2.25% (Table 1)[23-61]. This indicates that, on average, approximately 1.73 out of every 100 individuals in the included study populations tested positive for sapovirus (Figure 2).

Table 1.

Characteristics of the studies included in meta-analysis

Ref.	Country	World bank category	Time-frame	Number of positive samples	Total number of samples	Diagnostic method used	Age
Al-Ali et al[37], 2011	Lebanon	Lower-middle income	2010	0	79	RT-PCR	Both
Borkakoty et al[60], 2023	India	Lower-middle income	2014-2016	24	407	RT-PCR	< 5 years
Cao et al[57], 2021	China	Upper-middle income	2015-2019	30	1181	RT-PCR	< 5 years
Chaimongkol et al[38], 2012	Thailand	Upper-middle income	2007	5	160	RT-PCR	< 5 years
Chaimongkol et al[42], 2014	Thailand	Upper-middle income	2007	5	160	RT-PCR	< 5 years
	Thailand	Upper-middle income	2010	2	109	RT-PCR	< 5 years
	Thailand	Upper-middle income	2011	0	298	RT-PCR	< 5 years
Chen et al[41], 2013	Taiwan	High income	2004-2011	6	755	RT-PCR	Both
Cheng et al[36], 2010	China	Upper-middle income	2001-2007	10	1195	RT-PCR	< 5 years
Guntapong et al[25], 2004	Thailand	Upper-middle income	2002-2003	12	321	RT-PCR	< 5 years
Gupta et al[43], 2015	India	Lower-middle income	2010-2012	5	169	RT-PCR	< 5 years
Gupta et al[50], 2018	India	Lower-middle income	2009-2010	1	144	EIA or RT-PCR	< 5 years
Hansman et al[24], 2004	Thailand	Upper-middle income	2000-2001	13	105	RT-PCR	< 5 years
Jin et al[30], 2008	China	Upper-middle income	2006-2007	10	1110	RT-PCR	< 5 years
Jin et al[34], 2009	China	Upper-middle income	2005-2007	6	544	ELISA, RT-PCR, or PCR	< 5 years
Kawano et al[26], 2007	Japan	High income	2000-2006	2	62	RT-PCR	< 5 years
Khamrin et al[28], 2007	Thailand	Upper-middle income	2002-2004	3	248	RT-PCR	Both
Khamrin et al[35], 2010	Thailand	Upper-middle income	2005	5	147	RT-PCR	< 5 years
Khamrin et al[49], 2017	Thailand	Upper-middle income	2012-2014	6	889	RT-PCR	Both
Kumthip et al[54], 2020	Thailand	Upper-middle income	2010-2018	50	3057	RT-PCR	Both
Lasure and Gopalkrishna[47], 2017	India	Lower-middle income	2007	6	778	RT-PCR	< 5 years
	India	Lower-middle income	2008	6		RT-PCR	< 5 years
	India	Lower-middle income	2009	3		RT-PCR	< 5 years
	India	Lower-middle income	2010	2		RT-PCR	< 5 years
	India	Lower-middle income	2011	4		RT-PCR	< 5 years
Li et al[33], 2009	Hong Kong	High income	2008	2	209	RT-PCR	Both
Li et al[45], 2015	China	Upper-middle income	2011-2013	5	508	RT-PCR	< 5 years
Li et al[46], 2016	China	Upper-middle income	2011-2012	30	461	RT-PCR	< 5 years
Liu et al[44], 2015	Philippines	Lower-middle income	2012-2013	29	417	RT-PCR	< 5 years
Mai et al[61], 2023	Vietnam	Lower-middle income	2016-2021	50	2317	RT-PCR	< 5 years
Malasao et al[31], 2008	Thailand	Upper-middle income	2000-2002	10	296	RT-PCR	< 5 years
Mohammadi et al[55], 2020	Iran	Upper-middle income	2015-2017	0	103	RT-PCR	< 5 years
Monica et al[29], 2007	India	Lower-middle income	2001-2004	18	350	RT-PCR	< 5 years
Nakanishi et al[32], 2009	Japan	High income	2003-2005	15	877	RT-PCR	Both
Nguyen et al[27], 2007	Vietnam	Lower-middle income	2002-2003	4	1010	RT-PCR	< 5 years
Phengma et al[59], 2022	Thailand	Upper-middle income	2019-2020	6	675	RT-PCR	< 5 years
Pongsuwanna et al[48], 2017	Thailand	Upper-middle income	2006-2008	12	1141	RT-PCR	Both
Abdel-Rahman et al[56], 2021	Qatar	High income	2016-2018	21	901	Film array gastrointestinal panel kit	Both
Supadej et al[52], 2019	Thailand	Upper-middle income	2015-2017	16	843	RT-PCR	Both
Thongprachum et al[40], 2013	Thailand	Upper-middle income	2006	0	156	RT-PCR	< 5 years
Trang et al[39], 2012	Vietnam	Lower-middle income	2007-2008	7	501	RT-PCR	< 5 years
Wu et al[58], 2021	Taiwan	High income	2008-2011	48	4055	RT-PCR	Both
	Taiwan	High income	2016-2017	56	2449	RT-PCR	Both
Xue et al[53], 2019	China	Upper-middle income	2013-2017	11	569	RT-PCR	Both
Yan et al[23], 2004	China	Upper-middle income	2001-2003	2	207	RT-PCR	Both
Yu et al[51], 2018	Taiwan	High income	2007-2011	5	396	RT-PCR	< 5 years
	Taiwan	High income	2012-2026	4	441	RT-PCR	< 5 years

RT-PCR: Real-time polymerase chain reaction; EIA: Enzyme immunoassay; ELISA: Enzyme-linked immunosorbent assay; PCR: Polymerase chain reaction.

Figure 2.

Forest plot of the prevalence of sapovirus in hospitalized pediatric patients in the Asian region. CI: Confidence interval.

These studies were heterogeneous primarily, as suggested by the value for I² of 84.9% (95%CI: 80.3%-88.5%), indicating that a significant proportion of the variability in effect sizes was due to between-study differences rather than sampling error. The heterogeneity variance (tau-squared, τ²) was estimated at 0.5540, with the standard deviation of underlying effects (tau, τ) as 0.7443. The H statistic was 2.58 (95%CI: 2.25-2.95), again strengthening the assertion of there being significant heterogeneity.

The statistical tests for heterogeneity yielded that there is heterogeneity in the estimates of sapovirus prevalence from the varying study settings: Wald Q-test, Q = 252.22, df = 38, P < 0.0001 and likelihood ratio test = 240.99, df = 38, P < 0.0001. The result necessitates the use of a random-effects model, as it aims to incorporate between-study variability.

Subgroup analysis results

Country: Subgroup analysis by country revealed varying sapovirus event rates. Thailand had the highest number of studies (k = 13) with a pooled rate of 1.81 (95%CI: 1.13-2.89; I² = 83.6%), China (k = 8) reported a rate of 1.54 (95%CI: 0.91-2.61; I² = 88.6%) followed by India (k = 5) with a notably higher pooled rate of 3.47 (95%CI: 2.21-5.42; I² = 66.4%), while Japan (k = 2) had a rate of 1.81 (95%CI: 1.13-2.89) with no observed heterogeneity (I² = 0.0%) (Table 2).

Table 2.

Subgroup analysis of the prevalence of sapoviruses infection among pediatric patients with gastroenteritis

	Categories	K events	Pooled prevalence (%) (95%CI)	Q value	I2%
Overall		567	1.7304 (1.3274-2.2530)		84.9
Country	Thailand	13	1.8077 (1.1255-2.8913)	73.36	83.6
	India	5	3.4731 (2.2098-5.4184)	11.89	66.4
	China	8	1.5423 (0.9089-2.6053)	61.52	88.6
	Japan	2	1.8104 (1.1284-2.8928)	0.72	0.0
	Qatar	1	2.3307 (1.5245-3.5481)	0.00	-
	Vietnam	3	1.1303 (0.5007-2.5317)	11.41	82.5
	Lebanon	1	0.0000 (0.0000-100.0000)	0.00	-
	China Taiwan	3	1.4699 (1.2295-1.7564)	3.90	48.7
	Philippines	1	6.9544 (4.8752-9.8288)	0.00	-
	Iran	1	0.0000 (0.0000-100.0000)	0.00	-
	Hong Kong	1	0.9569 (0.2395-3.7435)	0.00	-
Age	Below 5 years	26	1.9958 (1.3893-2.8594)	181.70	86.2
	Both	13	1.5234 (1.3463-1.7235)	14.92	19.5

CI: Confidence interval.

Single-study estimates were available for Qatar, Lebanon, Philippines, Iran, and Hong Kong, showing wide CIs due to limited data. The Philippines had the highest rate among these with 6.95 (95%CI: 4.88-9.83). Notably, Lebanon and Iran reported zero event rates, resulting in uninformative CIs of 0.00-100.00. Vietnam (k = 3) showed a lower rate of 1.13 (95%CI: 0.50-2.53; I² = 82.5%), while China Taiwan (k = 3) reported 1.47 (95%CI: 1.23-1.76) with low heterogeneity (I² = 48.7%). These findings point toward a high degree of geographic variation in sapovirus detection rates that the differences could have influenced in surveillance, population demographics, diagnostic methods, or environmental or socio-economic factors.

World bank income category: Subgroup analysis was performed to evaluate the impact of country income level on the pooled effect size using a random-effects model. The analysis included 7 studies from High-income countries (effect size: 1.57; 95%CI: 1.34-1.83; I² = 33.6%), 10 studies from lower-middle income countries (effect size: 2.17; 95%CI: 1.19-3.94; I² = 86.2%), and 22 studies from upper-middle income countries (effect size: 1.63; 95%CI: 1.14-2.33; I² = 84.5%). Although the point estimate appeared highest in the lower-middle income group. However, the test for subgroup differences revealed no statistically significant variation between the three categories (Q = 1.08, df = 2, P = 0.583).

Publication year: A temporal subgroup analysis was conducted to determine if the effect size shifted over time across three distinct publication periods: 2000-2010 (k = 14), 2011-2020 (k = 19), and 2021-2023 (k = 6). The pooled effect sizes for these periods were 1.82 (95%CI: 1.10-3.00), 1.52 (95%CI: 1.03-2.22), and 2.22 (95%CI: 1.47-3.34), respectively. High levels of heterogeneity persisted within each subgroup, with I² values ranging from 84.5% to 87.4%. The test for subgroup differences revealed no significant variation between the time periods (Q = 1.75, df = 2, P = 0.416). These results suggest that the reported effects have remained statistically stable over the last two decades, with no significant evidence of a temporal trend or “year-of-publication” bias.

Sample size: Subgroup analysis was conducted to examine the influence of study scale on the pooled results across three categories: 1-500 (k = 18), 501-1000 (k = 13), and > 1000 (k = 8). Small studies (1-500) demonstrated the largest effect size (effect size: 2.55; 95%CI: 1.56-4.15), whereas mid-sized (effect size: 1.39; 95%CI: 1.09-1.78) and large-scale studies (effect size: 1.30; 95%CI: 0.92-1.85) showed notably lower estimates. The test for subgroup differences yielded a P value of 0.0646 (Q = 5.48, df = 2). While this does not meet the conventional threshold for statistical significance (P < 0.05), it indicates a strong marginal trend suggesting that sample size may be a moderating factor, with smaller studies potentially overestimating the observed effect.

Age: When stratified by age group, the pooled event rate of sapovirus infection was found to be notably higher in the < 5 years subgroup, which included 26 studies, reporting a rate of 1.99 cases per 100 observations (95%CI: 1.39-2.86). This group also showed a high degree of heterogeneity among the studies (I² = 86.2%), indicating considerable variability in sapovirus detection rates within this young population. Such variability may stem from the study settings, diagnostic methods, hygiene conditions in the region, socioeconomic conditions, and the periods when samples are collected; all these factors influence the prevalence of infection in children under five.

On the other hand, the ‘both’ group, comprising studies that included participants from any wide or unknown age interval (k = 13), had a random pooled event rate of 1.52 per 100 observations (95%CI: 1.35-1.72) and showed very little heterogeneity (I² = 19.5%). This lower level of heterogeneity suggests that the estimates across these studies were relatively consistent and less influenced by external variations or study design differences.

The findings suggest that there may be a higher prevalence of sapovirus infections among children under five years old. This, in turn, is already conspicuous: This group is more prone to enteric pathogens due to its developing immune system and higher external exposure risks, which are further strengthened by improper hand hygiene, close contact in daycare settings, and the consumption of contaminated food or water. Such an age-specific trend rests with the published literature on viral gastroenteritis, wherein young children often constitute the bulk of cases and hospitalizations, especially in low-and middle-income countries.

Meta-regression analysis

The included moderators were jointly significant QM (df = 7) = 26.18, P = 0.0005 QM (df = 7) = 26.18, P = 0.0005, explaining part of the between-study variability, with residual heterogeneity remaining moderate (I² = 69.8%). Studies published during 2021-2023 showed significantly higher prevalence compared to earlier years (P = 0.049), and studies with sample size ≤ 500 reported significantly higher prevalence (P = 0.0001). Other variables, including age group and income category, were not statistically significant predictors.

Publication bias analysis

To assess the presence of potential publication bias in the included studies, both Begg’s rank correlation test and Egger’s linear regression test was performed. The rank correlation test, as proposed by Begg and Mazumdar (1993), yielded a Z value of -1.13 with a P value of 0.2606, indicating no statistically significant evidence of funnel plot asymmetry. The bias estimate was -93.00, with a SE of 82.67, meaning that small-study effects were probably not able to have a significant influence on the overall results.

Similarly, the Egger regression test on funnel plot asymmetry, which examines the relationship between effect size and standard error through linear regression, also failed to indicate significant bias. The test resulted in a t value of -1.30 with 37 degrees of freedom and a corresponding P value of 0.2026. The estimated bias coefficient was -1.0554 (SE = 0.8135). This analysis was performed using the standard error as the predictor variable and applying inverse-variance weighting, with a multiplicative residual heterogeneity variance (tau²) of 6.6228.

Heterogeneity and sensitivity analysis findings

Sensitivity analysis was conducted by selecting studies with quality scores of 6 and above to test the robustness and reliability of the overall sapovirus prevalence estimates. This analysis was based on 32 quality studies with a total of 27570 observations and 506 sapovirus-positive events. The pooled event rate from these studies, calculated using a random-effects model, was 1.74 cases per 100 observations (95%CI: 1.31-2.31) (Figure 3).

Figure 3.

Funnel plot with transformed prevalence on the X-axis and standard error of transformed proportions on the Y-axis.

Despite limiting the analysis to higher-quality studies, substantial heterogeneity persisted among the included studies, with τ² = 0.5283, τ = 0.7268, and an I² value of 85.1% (95%CI: 80.0%-88.9%), indicating that most of the variability in effect sizes was due to genuine differences across studies rather than random error. The test for heterogeneity remained highly significant (Wald Q = 208.05, df = 31, P < 0.0001; likelihood ratio test Q = 192.90, df = 31, P < 0.0001), further confirming inconsistency across the included datasets.

It may be inferred from the findings that, when the analyses were restricted to studies with higher methodological quality, the estimated sapovirus detection rate remained consistent with the overall analyses, thus reinforcing the stability and validity of the pooled prevalence estimate.

DISCUSSION

Growing evidence points to sapovirus as the cause of AGE in young children. The virus is a serious public health risk, especially in underdeveloped nations with crowded cities and poor access to medical treatment. This systematic review and meta-analysis offer an in-depth assessment of the prevalence of sapovirus infections in pediatric and general populations across Asia, based on 39 studies comprising 30800 observations and 567 sapovirus-positive events. The pooled prevalence was calculated by applying the random-effects model and estimated to be around 1.73% (95%CI: 1.33%-2.25%), indicating that sapovirus, which are less studied than their counterparts such as norovirus and rotavirus, still play an important role in the cause of AGE, especially for children below five years of age. The pooled prevalence of sapovirus among people with AGE symptoms was 3.4%, according to a previous meta-analysis conducted globally. The highest frequency was found in children under the age of five (4.4%) and among people living in communal settings (7.1%)[18].

The prevalence of sapovirus in pediatric gastroenteritis hospitalizations in Asia, as reported in this meta-analysis, is relatively low compared to other enteric viruses (Table 1), with a pooled prevalence of 1.73% (95%CI: 1.33%-2.25%) compared to norovirus, rotavirus, and adenovirus, and similar to astrovirus, though the estimate was stable across the sensitivity analyses, suggesting this is an epidemiological finding rather than a methodological bias. The low prevalence of sapovirus likely reflects the high burden of rotavirus and norovirus in Asian pediatric populations, especially among hospitalized children, as diagnostic practices are geared towards major enteric pathogens, and differences in study period, geographic location, and timing of sample collection also influence prevalence estimates, because sapovirus circulation varies by region and season and may be underestimated in small or short-term studies.

This pooled prevalence estimate highlights sapovirus as an emerging pathogen of concern in Asia. While the burden appears to be less significant than that posed by other prominent enteric viruses (Table 3)[62-65], the sustained observation of the virus in several nations and across multiple age groups tends to indicate that casual circulation (intermittent, inefficient spread without sustained transmission chains or epidemic expansion) may be taking place. The detection of sapovirus cases within hospital settings underscores its relevance to public health. Despite the limited overall prevalence, gastrointestinal infections are generally recognised as posing increased risk in vulnerable populations such as infants, immunocompromised individuals, and individuals living in settings with compromised sanitation or hygiene. The I² statistic was 84.9%, and between-study variation (τ²) was 0.5540, indicating substantial heterogeneity among the included studies. That is, the I² statistic indicates that over 80% of the variation in the prevalence estimates is due to the differences between studies, rather than merely due to random error. Such heterogeneity is common in meta-analyses of infectious disease prevalence, mainly when studies differ in diagnostic techniques, geographic locations, sampling designs, or population parameters. A random-effects model was employed to account for such variation.

Table 3.

Prevalence of other enteric virus in Asian population

Viruses	Geographic scope	Population	Pooled prevalence/genotype distribution	Ref.
Norovirus	East, South, and Southeast Asia	Children with acute gastroenteritis	GII.4 approximately 53.6% of GII infections; other genotypes frequently > 10% (GII.2, GII.3, GII.7, GII.17)	[62]
Rotavirus	Multiple Asian countries	Children with acute gastroenteritis	G3P[8] 24%; G9P[8] 11%; G1P[8] 11%; G2P[4] 8%	[63]
Human astrovirus	Southeast Asia	Children with acute gastroenteritis	Approximately 2.5% (95%CI: 1.8-3.5)	[64]
Human adenovirus	Asia	Pediatric gastroenteritis cases	Approximately 7%	[65]
Sapovirus	Southeast Asia	Children with acute gastroenteritis	Approximately 2.3% (95%CI: 1.7-3.3)	[21]
Sapovirus	Asia	Hospitalized pediatric patients	1.73% (95%CI: 1.33-2.25); sensitivity analysis: 1.74% (95%CI: 1.31-2.31)	(Present meta-analysis)

CI: Confidence interval.

To evaluate the influence of study quality on pooled prevalence, a sensitivity analysis was performed including only studies with a quality score ≥ 6 (k = 32). The resulting prevalence remained similar (1.74%, 95%CI: 1.31%-2.31%) with comparable heterogeneity (I² = 85.1%). This consistency suggests that study quality did not substantially bias the overall estimate and supports the robustness of the main findings.

Both Begg’s rank correlation test and Egger’s linear regression test were performed to explore the presence of publication bias. Neither test indicated any significant asymmetry in the funnel plot (Begg’s: Z = -1.13, P = 0.2606; Egger’s: T = -1.30, P = 0.2026), suggesting no influence from publication bias on the findings. The possible limitation remains with unpublished data, especially from regions with limited surveillance; however, it would hardly alter the observed pattern.

Moreover, the presence of statistically significant heterogeneity among studies was supported by both the Wald test (Q = 252.22, P < 0.0001) and the likelihood ratio test (Q = 240.99, P < 0.0001). These results urge us to interpret the pooled estimate with caution and place greater importance on subgroup analyses to figure out the sources of heterogeneity. Certainly, regional sapovirus prevalence variations were further discussed in the subgroup analysis by country. The highest prevalence was recorded in India (3.47%, 95%CI: 2.21%-5.42%), followed by the one-study report from the Philippines (6.95%). Thereafter, steady albeit less common detections were reported in China (1.54%), Thailand (1.81%), and Japan (1.81%). In contrast, countries such as Iran and Lebanon were reported to have a prevalence of 0%. However, these findings were based on single studies, which may not necessarily mean that these countries entirely lack sapovirus, rather indicate under-detection, or that the sample sizes in these studies were too small to detect its presence.

The test for subgroup differences was highly significant (Q = 67.50, df = 10, P < 0.0001), indicating that geographical location has a significant influence on sapovirus prevalence. These differences may be driven by local climatic factors, variation in circulating genogroups, population density, diagnostic capacity, and socioeconomic conditions affecting hygiene and sanitation.

Age-stratified analysis revealed higher sapovirus prevalence in studies focused on children under five years (1.99%, 95%CI: 1.39%-2.86%) compared to studies involving mixed age groups (1.52%, 95%CI: 1.35%-1.72%). This trend supports existing literature indicating that young children are disproportionately affected by enteric viral infections due to their immature immune systems and higher susceptibility to fecal-oral transmission. The ability of sapovirus to cause asymptomatic infections or be involved in co-infections could further obscure age-specific associations.

Another crucial differentiating factor is the diagnostic method applied. Most studies have used RT-PCR/quantitative polymerase chain reaction for sapovirus detection, which varies in sensitivity and specificity based on the design of primers/probes and the targeted gene regions. Only a few studies employed droplet digital polymerase chain reaction, which offers enhanced sensitivity, especially for low viral load samples, and provides absolute quantification without the need for standard curves.

Meta-regression analysis demonstrated that the included moderators were jointly significant in explaining heterogeneity; however, substantial variability persisted (I² = 69.8%). Studies published between 2021 and 2023 reported significantly higher sapovirus prevalence compared to earlier studies (P = 0.049). This temporal trend may be attributed to advancements in diagnostic methods and increased surveillance capacity in recent years. Studies with sample sizes of 500 or fewer participants reported significantly higher prevalence (P = 0.0001), suggesting potential selection bias, as smaller studies often target outbreak settings or high-risk populations. Age group and income category were not significant predictors of prevalence, indicating that sapovirus circulates uniformly across pediatric age groups and socioeconomic strata in hospital settings. The moderate residual heterogeneity implies that additional unmeasured factors, such as diagnostic methods, regional differences, and seasonal patterns, continue to influence prevalence estimates.

The detection of sapovirus cases in various settings and age groups indicates that it may still be an important but unrecognized agent of AGE. The diverse genogroups and increasing reports of sapovirus outbreaks, especially in childcare centres and closed settings, call for greater attention in public health monitoring. Additionally, since sapovirus detection is not included in most routine panels, the actual burden is grossly underestimated. Enhanced molecular surveillance with high-end sensitive techniques of droplet digital polymerase chain reaction or metagenomic sequencing along with inclusion of sapovirus in syndromic testing panels are essential for impingement assessment. Region-wise diagnostic assays and genotyping tools also need to be developed, given the genetic diversity of circulating strains in Asia.

Analysis shows that sapovirus affects people regardless of a country’s wealth. The results across different income groups did not differ significantly (P = 0.583). While lower-middle-income countries had a slightly higher estimate of 2.17, the broad CIs suggest this is likely due to chance. Interestingly, data from high-income countries was much more consistent (I² = 33.6%) than data from middle-income regions. This indicates that while the virus is a global issue, differences in testing or reporting might cause more varied results in developing areas. Overall, these findings confirm that sapovirus is a universal health concern that transcends economic boundaries.

The impact of sapovirus has remained remarkably stable over the last two decades. There was no statistical difference between studies published in the early 2000s and those published recently (P = 0.416). Even as laboratory tests became much more sensitive and accurate, the reported prevalence did not change significantly. The consistently high heterogeneity across all years suggests that the differences in findings are not caused by when a study was done. Instead, these variations likely come from differences in local populations or geography. This stability shows that sapovirus continues to be a persistent public health challenge that has not diminished over time.

A notable trend appeared when looking at the size of the studies. Smaller studies with fewer than 500 participants reported a much higher effect size (2.55) than larger studies with over 1000 participants (1.30). This difference reached marginal significance (P = 0.0646). Such a pattern often suggests a “small-study effect”, where smaller trials might focus on specific outbreaks or high-risk groups. Larger studies usually provide a more accurate and conservative view of the general population. As the sample size increased, the results became more consistent and the heterogeneity decreased. This suggests that larger investigations offer a more reliable baseline for understanding the true reach of sapovirus.

Sapovirus co-infection with other enteric pathogens such as norovirus, rotavirus and enteropathogenic Escherichia coli is common in paediatric patients in Asia, and is associated with more severe disease, including prolonged diarrhoea and increased risk of dehydration. These results are likely due to combined damage to the intestinal epithelium and a weakened host immune response due to pathogen interactions. Further prospective studies in paediatric populations in Asia are needed to determine the prevalence of sapovirus co-infection and synergies with other enteric pathogens.

There is also a need for longitudinal studies that may help assess sapovirus seasonality, transmission dynamics, and immune response profiles. The evaluation of co-infections with other enteric pathogens may clarify sapovirus’ role in the aggravation of diarrheal illnesses’ severity.

Strengths and limitations: This review followed PRISMA guidance and applied a transparent, reproducible search strategy. Thirty-nine studies with a large cumulative sample size were included, and methodological quality was evaluated using a standardized appraisal tool. Pooled estimates were generated with a random-effects model, supported by sensitivity and subgroup analyses that indicated result stability. Publication bias assessment did not suggest marked asymmetry. Residual heterogeneity across studies remained substantial and likely reflects differences in geography, climate, and study period. Most included investigations were hospital based, which limits extrapolation to community settings. Several countries were represented by only a single study, so country-specific prevalence estimates should be interpreted with caution. Variability in diagnostic approach may be another limitation; although RT-PCR is a principal method in viral diagnostics, variation in primer sets, gene targets, protocols, and assay sensitivity, along with occasional combination with serological tests, may have affected detection rates and constrained assay-specific subgroup analyses. Screening was restricted to English-language original articles, so relevant non-English reports may not have been captured.

CONCLUSION

Our analysis estimates the pooled sapovirus prevalence in paediatric patients admitted to hospital in Asia at 1.73 %. Although lower than the reported prevalence for rotavirus and norovirus, this prevalence reflects a permanent contribution to paediatric gastroenteritis and not a negligible incidence. Subgroup analyses show comparable estimates across income categories of the World Bank and no significant time differences over more than two decades, suggesting that the spread of the sapovirus is persistent in different economic environments. Children under five years of age remain the most affected group, which is consistent with their increased susceptibility to dehydration and hospitalisation. Analysis based on sample size suggests that smaller studies tend to report higher prevalence estimates, which points to possible effects of small studies and highlights the need for larger, methodologically standardized studies to refine the baseline estimates. Despite its consistent detection, human sapovirus is rarely included in routine diagnostic algorithms, which may contribute to its under-reporting in clinical and surveillance data. The stability of prevalence over time and income levels supports its importance as an endemic enteric pathogen and not as an episodic cause of the disease. The inclusion of sapovirus in standard molecular diagnostic panels, together with improved hygiene and sanitation, would strengthen surveillance and better characterise its clinical and epidemiological impact, especially in resource constrained environments. Recognition of sapovirus as a significant contributor to paediatric gastroenteritis is essential for informed public health planning, appropriate allocation of resources. Future research should prioritize longitudinal studies to capture temporal trends, community-based surveys to determine true disease burden beyond hospitalized cases, and molecular genotyping to monitor viral evolution and transmission dynamics. These efforts are critical for developing targeted prevention strategies and reducing the global burden of sapovirus-associated gastroenteritis in vulnerable pediatric populations.

ACKNOWLEDGEMENTS

The authors are thankful to the Indian Council of Medical Research, New Delhi, India for providing the necessary resources and support that made this work possible.