El Niño–Southern Oscillation and unusual severe acute respiratory infection in Colombia, 2011–2019: a national time-series analysis using distributed lag non-linear models

Background: The El Niño–Southern Oscillation (ENSO) is the main driver of inter-annual climate variability in Colombia, but its relationship with severe respiratory infections has not been quantified at national level. We estimated the association between ENSO intensity and the incidence of unusual severe acute respiratory infection (SARI), a sentinel surveillance event, and examined whether it varied by lag, region and departmental poverty. Methods: We conducted an ecological time-series study of monthly unusual SARI cases notified to the national surveillance system (SIVIGILA, event 348) from 2011 to 2019. ENSO intensity was measured with the Oceanic Niño Index (ONI). We used quasi-Poisson distributed lag non-linear models with lags of 0 to 12 months, adjusting for seasonality and long-term trend, with population as offset. We fitted models nationally and by region and Multidimensional Poverty Index tertile, and ran prespecified sensitivity analyses. Results: A total of 12,984 cases were notified (31.0 per million inhabitants per year). Nationally, ONI was not associated with cumulative incidence (relative risk [RR] per 1 °C 0.95, 95% CI 0.78–1.15). Higher ONI was followed by fewer cases at lags of 0 to 4 months and more cases at 8 to 12 months. In the Caribbean and Pacific regions, higher ONI was associated with lower incidence (RR 0.72, 0.56–0.93, and 0.65, 0.53–0.80), indicating higher incidence after La Niña conditions. The national estimate was sensitive to the flexibility of the long-term trend. Conclusions: ENSO was not a national predictor of unusual SARI in Colombia, but La Niña was followed by higher incidence in the Caribbean and Pacific regions. Regional La Niña alerts could support respiratory preparedness, and future analyses should account for differences in surveillance capacity.