Increased cardiomyopathy risk after viral infection evidence from a nationwide cohort in Taiwan

Viral infections have been linked to myocarditis and may contribute to the development of cardiomyopathy. However, large-scale population-based evidence remains limited. This study aimed to investigate the association between viral infections and the subsequent risk of cardiomyopathy using Taiwan's Longitudinal Generation Tracking Database (LGTD). We conducted a nationwide, retrospective cohort study using Taiwan's National Health Insurance Research Database/LGTD. Individuals with a first recorded viral infection were matched to unexposed controls using 1:4 propensity scores incorporating demographics, comorbidities, medications, healthcare utilization, and index year. Incident cardiomyopathy was ascertained from ICD-9-CM diagnosis codes in claims data, without clinical adjudication or imaging confirmation. Risks were estimated using Cox proportional hazards models to obtain adjusted hazard ratios (aHRs) and 95% confidence intervals (CIs). Robustness was assessed via Fine-Gray competing-risk models for death, proportional-hazards diagnostics, alternative exposure/lag definitions, and prespecified subgroup analyses by age, sex, comorbidity burden, and virus categories. Over a mean follow-up of approximately 10 years, the incidence rate of cardiomyopathy was higher in the viral infection group (98.42 per 100,000 person-years) than in the control group (45.69 per 100,000 person-years). Viral infection was significantly associated with increased risk of cardiomyopathy (aHR = 2.915, 95% CI 1.177-4.828, p < 0.001). Subgroup analyses showed consistent risk elevation across sex (male aHR = 2.775; female aHR = 3.097), age groups, income levels, seasons, and urbanization levels (all p < 0.05). Among viral subtypes, viral hepatitis (aHR = 3.435), influenza (aHR = 3.002), and viral pneumonia (aHR = 3.091) were most strongly associated with cardiomyopathy. In this population-based cohort, viral infection was associated with increased long-term risk of cardiomyopathy. Given that outcomes were identified from administrative codes without clinical or imaging validation, causal inference is not warranted; however, the findings support post-infection cardiovascular surveillance and prevention strategies. Further studies integrating adjudicated clinical data, imaging, and biomarkers are needed to clarify mechanisms and refine risk stratification.