Umeå universitets logga

Background: Dengue, transmitted by Aedes mosquitoes, is a major public health problem in Sri Lanka. Weather affects the abundance, feeding patterns, and longevity of Aedes vectors and hence the risk of dengue transmission. We aimed to quantify the effect of weather variability on dengue vector indices in ten Medical Officer of Health (MOH) divisions in Kalutara, Sri Lanka.

Methods: Monthly weather variables (rainfall, temperature, and Oceanic Niño Index [ONI]) and Aedes larval indices in each division in Kalutara were obtained from 2010 to 2018. Using a distributed lag non-linear model and a two-stage hierarchical analysis, we estimated and compared division-level and overall relationships between weather and premise index, Breteau index, and container index.

Findings: From Jan 1, 2010, to Dec 31, 2018, three El Niño events (2010, 2015–16, and 2018) occurred. Increasing monthly cumulative rainfall higher than 200 mm at a lag of 0 months, mean temperatures higher than 31·5°C at a lag of 1–2 months, and El Niño conditions (ie, ONI >0·5) at a lag of 6 months were associated with an increased relative risk of premise index and Breteau index. Container index was found to be less sensitive to temperature and ONI, and rainfall. The associations of rainfall and temperature were rather homogeneous across divisions.

Interpretation: Both temperature and ONI have the potential to serve as predictors of vector activity at a lead time of 1–6 months, while the amount of rainfall could indicate the magnitude of vector prevalence in the same month. This information, along with knowledge of the distribution of breeding sites, is useful for spatial risk prediction and implementation of effective Aedes control interventions. Funding: None.