Predicting outbreaks of dengue fever according to climate

This work enabled statistical models to be drawn up explaining and predicting viral episodes. The Caledonian public health authorities have already integrated these tools into their decision-making strategies and a similar approach in other South Pacific countries is being developed, with a new collaborative regional programme.

Mainly vectored by the mosquito Aedes aegypti, dengue fever is present in one hundred or so countries in the intertropical zone. It affects one hundred million people, causing more or less severe symptoms such as fever, muscular pain, headaches, digestive disorders and even hemorrhaging, causing 25 000 deaths per year throughout the world. No vaccine or specific treatment exists to this day.

The South Pacific region is regularly the seat of dengue fever epidemics, of increasing frequency and amplitude. In New Caledonia, the last epidemic in 2009 affected 8 400 people. It is the second most affected French overseas territory. Like for other vector-borne diseases, the occurrence or not of epidemic outbreaks is the consequence of complex interactions between humans, the virus and vector mosquitoes, influenced by environmental and climatic factors. Until now, these mechanisms have remained poorly understood.

In Noumea, a reliable system of monitoring of dengue fever cases has existed for forty years. In parallel, daily meteorological measurements have been made since the 1950's. IRD scientists and their partners(1) compared these monthly, quarterly and annual meteorological and epidemiological data over 40 years. The existence of these long series enabled a statistical model to be constructed that shows the essential role of the local climate in the dynamic of epidemics in Noumea. The scientists also accurately identified the meteorological conditions that were decisive in the occurrence of epidemics.

From explanation...

In order to analyse 40 years of data, doctors, entomologists, climatologists, mathematicians and forecasters first studied the seasonal variability in the disease, setting apart epidemic from non-epidemic years. All the outbreaks were similarly distributed, with a start in January, a peak between March and May, and ending in July.

In a second phase, researchers analysed the variability in the appearance of dengue fever epidemics from one year to the next. To do so, an "explanatory" model was developed. For the first time, it revealed that the occurrence of epidemics in the greater Noumea area could be explained by the two variables of temperature and relative humidity. The same scenario has been operating over 40 years. If the temperature exceeds 32°C for more than 12 days in January, February and March -- or during the southern summer -- and humidity exceeds 95 % for less than 12 days in January, a wave of dengue fever occurs. And vice versa: when cooler summer temperatures are combined with wetter weather, an epidemic occurs in almost all cases.

... to forecasts.

Again, thanks to the long series of data, a second "predictive" model was drawn up to anticipate epidemics from one year to the next. This time, it is based on maximum temperature and relative humidity values for the months of October to December in the previous year. These results are crucial. Using the system developed, public health authorities can anticipate orders for mosquito repellent and insecticide, organise vector control, optimise treatment systems, etc.

El Niño and dengue fever: no direct link

Larger-scale climate variables such as El Niño / La Niña influence the local climate. The phenomenon occurs every 3 to 7 years. Contrary to belief, the study did not reveal any statistical link between epidemics and La Niña in New Caledonia. These results suggest that the influences of this type of climate variable on dengue fever outbreaks depend on complex geographic specificities.

From an operational point of view, the AeDenPAC programme(2), led by the New Caledonia Pasteur Institute and involving numerous IRD scientists and other partners(3), is under way. It aims at developing similar models and extending research on the entomological level in the South Pacific, in particular in French Polynesia and Fiji. This expansion will gradually feed knowledge of the impact of global warming on the dynamic of dengue fever epidemics in the Pacific.

(1) This work was conducted in partnership with Aix-Marseille University, the Noumea Hospital Centre, the New Caledonia Pasteur Institute, Météo France in New Caledonia, the Department of Social and Sanitary Affairs (DASS) and the Secretariat of the South Pacific (SPC).

(2) Climate / mosquito / epidemic modelling

(3) Institut Pasteur NC, IRD, Institut Louis Malardé, Secretariat of the Pacific Community, ministry for Health in Fidji and Tonga, Centre Hospitalier NC, Otago University-NZ.