Influenza-associated excess mortality and subtype-specific contributions in a city in Guangxi, China: a generalized additive model analysis, 2022–2025

Influenza poses a significant global disease burden, with impacts extending beyond the respiratory system to increase cardiovascular event risks. However, regional data on influenza-associated excess mortality, especially regarding subtype-specific contributions and high-risk populations, remain insufficient in Guangxi. This study aimed to explore influenza-associated excess all-cause, respiratory, and circulatory mortality and its population characteristics in a city in Guangxi from 2022 to 2025, providing evidence for targeted prevention and control. Weekly mortality data, influenza virological surveillance data (A/H3N2, A/H1N1, B/Victoria), meteorological data, and air pollution data (PM₂. ₅, NO₂) were collected from the 14th week of 2022 to the 13th week of 2025 in a city in Guangxi. A generalized additive model (GAM) with a quasi-Poisson distribution was used to estimate influenza-associated excess mortality, adjusting for temperature, humidity, PM2.5, and NO2. The optimal lag structure was determined using QAIC-based model comparison. Excess mortality was estimated using counterfactual predictions with all influenza terms set to zero. Negative binomial models were fitted as a sensitivity analysis. The optimal lag structure was 4 weeks for all three influenza subtypes (A/H3N2, A/H1N1, and B/Victoria), 0 weeks for temperature and PM2.5, and 3 weeks for humidity and NO2. The model demonstrated substantial explanatory power for all-cause mortality (adjusted R² = 0.7275, deviance explained = 77.9%). During the study period, the all-cause excess mortality rate was highest in 2023–2024 (42.36 per 100,000), followed by 2022–2023 (21.00 per 100,000) and 2024–2025 (15.57 per 100,000). Circulatory deaths constituted the majority of excess mortality across all seasons. The elderly population (≥ 65 years) bore the overwhelming majority of excess deaths (3,002 excess deaths). A/H3N2 showed the strongest association with mortality (IRR = 1.0017, P = 0.002). From 2022 to 2025, influenza-associated excess mortality in a city in Guangxi was characterized by a predominance of circulatory diseases and a concentration of risk among the elderly population, with A/H3N2 being the dominant pathogenic subtype.