Epidemiologic shifts, seasonal reshaping, and exploratory analysis of physician-documented myocardial injury in children with rotavirus infection in Hangzhou, China, 2016–2025: a decade-long single-centre retrospective study

Rotavirus (RV) remains a leading cause of paediatric acute gastroenteritis. COVID-19-associated non-pharmaceutical interventions (NPIs) profoundly altered enteric virus circulation. This single-centre hospital-based retrospective study aimed to describe ten-year epidemiological features among children aged ≤15 years with acute diarrhoea in Hangzhou, including both inpatients and outpatients, with attention to pandemic-related temporal changes, seasonal patterns, and burden of physician-documented clinical myocardial injury. We followed STROBE-RECORD reporting guidelines for electronic-health-record-based observational research. We analysed 71,115 rotavirus antigen test-episodes from children ≤15 years presenting with acute diarrhoea (≥ 3 loose or watery stools within 24 h, episode duration ≤14 days), January 2016-December 2025. Duplicate tests within a single diarrhoeal episode for one individual were excluded; distinct episodes separated by >14 days in the same child were retained. Selected physician-assigned clinical diagnostic labels from electronic medical records were extracted for exploratory secondary analyses, including myocardial-injury diagnoses obtained within 3 days of RV testing. Crude relative risks (RR), modified Poisson regression with robust standard errors, and Firth penalized logistic regression for sparse-event strata were implemented. Negative-binomial regression was applied for formal monthly seasonal inference; cosinor-harmonic regression failed to converge. Myocardial-injury analysis is exploratory because universal standardized cardiac screening was not implemented across the cohort. Sensitivity analyses included a hospitalized-only subset and models incorporating available comorbidity diagnostic tags. Patient-level clustering could not be performed because unique patient identifiers were not retained in the frozen analytical dataset. Electrolyte abnormality labels were unavailable, and cardiac testing frequency could not be quantified; these limitations preclude adjustment for electrolyte disturbances and precise assessment of surveillance bias. Overall RV test-positivity was 16.25% (11,553/71,115). Positivity peaked at age 1 year (23.91%, accounting for 43.31% of all RV-positive test-episodes). Annual RV positivity peaked in 2017 (44.55%), fell to a nadir of 2.80% in 2023, rebounded in 2024 (9.02%), and modestly declined in 2025 (8.51%). Positivity declined stepwise: pre-COVID (2016–2019) 27.55%; mid-COVID (2020–2022) 10.16%; post-COVID (2023–2025) 6.82% (χ 2 =4,697.21, df=2, P <0.001). Age-stratified positivity changes were compatible with shifts in age-specific population susceptibility; larger declines were observed in toddlers, whereas changes were modest in school-aged children (Spearman ρ=− 0.77, P =0.009). Among RV-positive test-episodes, median age increased across periods: pre-COVID 1.0 y (IQR 1–3), mid-COVID 2.0 y (IQR 1–4), post-COVID 3.0 y (IQR 2–5); the proportion of RV-positive children aged ≥2 years rose from 48.2% pre-COVID to 76.5% post-COVID. Negative-binomial regression for monthly aggregated counts showed a significant month-by-period interaction ( P =0.017), consistent with seasonal pattern reshaping: the pre-COVID winter-early-spring peak was attenuated during mid-COVID and shifted toward a spring-centred peak post-COVID. Clinically documented myocardial injury occurred in 1.13% of RV-positive and 0.42% of RV-negative test-episodes; crude RR=2.67 (95% CI 2.16–3.30). Stratified crude RR was 1.51 (95% CI 0.86–2.66, non-significant) pre-COVID; 6.29 (95% CI 4.25–9.32) mid-COVID; and 5.93 (95% CI 4.49–7.83) post-COVID. Period-by-RV-status interaction was statistically significant (Wald χ 2 =11.84, df=2, P =0.003). No physician-documented myocardial injury was recorded among RV-positive infants <1 year old; age-specific myocardial-injury proportions among RV-positive children did not increase monotonically with advancing age. Firth penalized logistic outputs and sensitivity analyses are provided in supplementary tables. One-year-old children had the highest observed RV test-positivity. Observational data are compatible with age-distribution shifts and seasonal pattern reshaping after COVID-19-related NPIs. RV-positivity was associated with elevated risk of physician-diagnosed myocardial injury in mid- and post-COVID periods, whereas no statistically significant association existed pre-pandemic. Findings should be interpreted cautiously given multiple retrospective limitations, including non-standardized cardiac evaluation, inability to quantify cardiac testing frequency, absence of electrolyte abnormality data, potential statistical inflation from repeated test-episodes, non-quantifiable surveillance bias, and incomplete reagent/assay documentation. Increased clinical awareness for potential cardiac complications is suggested for hospitalized RV-positive children. Continuous epidemiological surveillance remains warranted.