The association between weekly mean temperature and the epidemic of influenza across 122 countries/regions, 2014-2019.

The study examined the association between weekly mean temperature and influenza cases across 122 countries/regions (2014-2019) using a distributed lag non-linear model (DLNM). We analyzed 3145206 cases of overall influenza (Flu-All), with influenza A (Flu-A) and influenza B (Flu-B) accounting for 73.49% and 26.51%, respectively. Within a lag of 2 weeks, Flu-All incidence demonstrated a bimodal temperature relationship, with peak relative risks (RR) of 6.02 (95% CI: 1.92-20.77) at -8 ℃ and 3.08 (95% CI: 1.27-7.49) at 22 ℃. Flu-A exhibited a similar bimodal pattern, with RRs of 3.76 (95% CI: 2.39-5.91) at -8 ℃ and 2.08 (95% CI: 1.55-2.80) at 22 ℃. Flu-B demonstrated a single risk peak at 1 ℃ (RR = 4.48, 95% CI: 1.74-11.55). Subgroup analyses of climate zones revealed variations: tropical zones peaked at 12 ℃ (RR = 1.37, 95% CI: 1.08-1.74), while dry and temperate zones exhibited the highest risk at -5 ℃, with RRs of 4.49 (95% CI: 2.46-7.15) and 5.23 (95% CI: 3.17-8.64), respectively. Cold zones peaked at 1 ℃ (RR = 5.96, 95% CI: 3.76-9.43). Subgroup analyses of influenza transmission zones (ITZs) revealed variations: Africa showed higher risk between 6 ℃-14 ℃, Asia showed higher risk below 3 ℃, and Europe exhibited distinct risks of influenza peaks at -1 ℃ (Eastern), 1 ℃ (Southwest), and -20 ℃ (Northern). Elevated risks above 11 ℃ were identified in the Americas and Oceania. These findings establish a predictive framework for influenza outbreak preparedness by integrating regional temperature patterns with global climate variability.