Climate and malaria: modeling non-linear dynamics in the Nordic countries during the 18th and 19th centuries.

Aims: Until the late 19th century, Plasmodium vivax malaria was endemic in most of Europe including in the Nordic countries. In Denmark, Finland, and Sweden, the fluctuations in malaria cases and malaria-attributed deaths are known to have been associated with weather conditions, in particular with mean summer temperature variations. However, to what extent other environmental factors could have increased or decreased the risk of malaria has not previously been evaluated using historical records.

Methods: In this study, we illustrate the non-linear association between different environmental variables (temperature, precipitation, and sea-level variations) and symptom-based malaria (case and death) data, using the quasi-Poisson distributed lag non-linear model. The robustness of the model results was examined through sensitivity analysis.

Results: The modeling results showed that the risk associated with temperature increased by ∼25% in Denmark and by ∼67% in Sweden and Finland, with a mean summer temperature increase from 16°C to 18°C, was highest at 1-2 lagged years. Furthermore, average precipitation could have a noticeable effect on the malaria risk in Sweden and Finland, but this effect was not observed in Denmark. Environmental perturbations associated with extreme sea levels (>99.7th percentile or <0.1th percentile), including subsequent saltwater intrusion, could lead to increasing malaria risk in low-lying coastal areas.

Conclusions: The historical evidence and modeling results suggest that specific weather conditions and extreme events have substantial impacts on malaria in temperate regions.