Review finds norovirus and rotavirus in 18% of pediatric diarrhea cases after rotavirus vaccine introduction

A systematic review and meta-analysis estimated that norovirus or rotavirus was detected in a pooled 18% of children under five with diarrheal infection in lower-middle-income countries, based on 45 eligible studies that included 35,704 children. Rotavirus had a higher pooled prevalence than norovirus, at 24% versus 13%.

The analysis included studies from lower-middle-income countries in which children under five with acute diarrhea were tested using stool specimens. For studies examining rotavirus, the researchers required that data collection begin in or after the year of national rotavirus vaccine introduction. Norovirus studies were not restricted by publication year. Eligible designs included cross-sectional studies, case-control studies and large-scale surveillance studies, with a minimum sample size of 50 participants.

Researchers searched PubMed, Scopus, ScienceDirect, Embase and Web of Science from February 20 to February 28, 2026. They removed duplicates, screened titles and abstracts, reviewed full texts and extracted data with two investigators working independently. The review protocol was registered in PROSPERO as CRD420261349950. The investigators used the Joanna Briggs Institute checklist for studies reporting prevalence; all 45 included reports were rated moderate or high quality, with scores ranging from 55.5% to 100%.

Most of the evidence base consisted of cross-sectional studies, while 11 reports were large-scale surveillance studies. Detection methods also varied: 23 studies used RT-PCR, 19 used both RT-PCR and ELISA, and three used ELISA and multiplex RT-PCR. Individual study sample sizes ranged from 55 to 5,408. India contributed 20 studies, followed by Bangladesh and Ethiopia with four each, and Nigeria with three. Limited representation from several regions and the large number of Indian studies limited geographic representativeness, the authors said.

Using a restricted maximum likelihood random-effects model, the review estimated combined pooled prevalence for norovirus and rotavirus at 18% (95% confidence interval, 15%–21%). The virus-specific estimates were 24% (95% CI, 19.6%–28.3%) for rotavirus and 13% (95% CI, 11%–15%) for norovirus. Norovirus prevalence ranged from 3.12% to 30.50% across studies, while rotavirus prevalence ranged from 17.90% to 46.50%.

Results varied substantially across the included studies. The combined analysis showed very high heterogeneity, with I² reported at approximately 98.7%–98.9% and Cochran’s Q significant at p<0.001. Regional pooled estimates ranged from 11% (95% CI, 9%–14%) in the Middle East to 26% (95% CI, 8%–45%) in Southern Africa. Pooled prevalence was 17% (95% CI, 14%–20%) in cross-sectional studies and 21% (95% CI, 16%–27%) in surveillance studies. By detection method, estimates were 25% for ELISA combined with RT-PCR, 18% for ELISA combined with multiplex PCR and 13% for RT-PCR alone.

Genotype information was available from 43 studies. Among 9,889 samples sequenced, the review identified 1,840 genotypes. For norovirus, genogroup GII predominated, with a pooled prevalence of 76% (95% CI, 61%–91%), followed by GI at 23% (95% CI, 9%–37%). GII.4 was the most frequent norovirus subgenotype, accounting for a pooled 53% of cases (95% CI, 39%–68%). GII.3 was also reported, with 145 detections across included studies.

For rotavirus, the most frequently observed genotype combinations were G3P[8] and G1P[8], with pooled prevalences of 39% (95% CI, 28%–51%) and 17% (95% CI, 10%–25%), respectively. G3P[8] accounted for 1,437 detections, primarily from India but also from Ethiopia and Kenya. G1P[8] accounted for 501 detections from Morocco, Ethiopia and India, while G2P[4] accounted for 489 detections from India, Ethiopia, Honduras and Zambia. The review also recorded 384 mixed-type rotavirus detections, five partially typable detections and 16 non-typable detections.

Several limitations constrain interpretation of the pooled estimates. Most included studies were cross-sectional and therefore could not establish temporal associations. The English-language restriction may have introduced language bias, heterogeneity was substantial, and statistical tests indicated possible small-study effects and publication bias. Although a trim-and-fill analysis left the overall pooled estimate unchanged, a leave-one-out analysis found that removing individual studies did not materially alter it. The authors cautioned that the findings should not be considered conclusive because of these limitations, geographic gaps and the disproportionate contribution of studies from India.

Subject of Research: Prevalence and genotype distribution of norovirus and rotavirus among children under five with diarrhea in lower-middle-income countries

Article Title: Prevalence and genotype distribution of norovirus and rotavirus among children under five with diarrheal infection in lower-middle-income countries after Rotavirus vaccine introduction: A systematic review and meta-analysis

Article References: Ayele, F., Assebe, T., Mandefro, M., Admassu Ayana, D., & Assefa, N. (2026). Prevalence and genotype distribution of norovirus and rotavirus among children under five with diarrheal infection in lower-middle-income countries after Rotavirus vaccine introduction: A systematic review and meta-analysis. PLOS One, 21(10), e0359886. https://doi.org/10.1371/journal.pone.0359886

Image Credits: AI Generated

DOI: 10.1371/journal.pone.0359886

Keywords: norovirus, rotavirus, childhood diarrhea, systematic review, meta-analysis, viral prevalence, genotype distribution, lower-middle-income countries, vaccine introduction, genomic surveillance, Prevalence, genotype