Respiratory health and eruptions of the Nyiragongo and Nyamulagira volcanoes in the Democratic Republic of Congo: a time-series analysis

BACKGROUND: Nyamulagira and Nyiragongo are active volcanoes situated close to Goma (North Kivu, Democratic Republic of Congo). These volcanoes are among the most prolific sources of volcanic SO(2) pollution on earth. OBJECTIVE: We investigated the possible spatiotemporal relationships between volcanic degassing represented by eruptive emissions of SO(2) that occurred between 2000 and 2010, and the incidence of acute respiratory symptoms (ARS) in populations living in areas up to more than 100 km from the volcanoes. METHODOLOGY: The total flux of SO(2) emitted during eruptions since 2000 and the average spatial distribution of the volcanic plume (2004–2008) were based on publicly available remote sensing data. The monthly numbers of adults and children reporting acute respiratory symptoms were extracted from health data collected routinely by selected local health centres and hospitals between 2000 and 2010. The monthly numbers of persons with ARS recorded during or after eruptions were compared with those recorded before eruptions, using negative binomial regression models allowing the calculation of incidence rate ratios (IRR) and their 95% confidence intervals. We first compared years with and without eruptions and then considered shorter time-windows (months). RESULTS: In the investigated area, ARS were the second most frequent cause of medical visits (12.2%, n = 3.2 million cases), after malaria (32.3%, n = 8.4 million cases). SO(2) emissions gradually increased 30 to 50 times in 2010 compared to 2002. Taking 1999 as a reference, the IRR for ARS increased three-fold between 2000 [0.9 (0.8, 1.1)] and 2009 [2.8 (2.2, 3.7)]. Although the incidence of ARS appeared to increase after some eruptions, especially in areas close (< 26 km) to the volcanoes, we did not find a consistent temporal association between the yearly incidence of ARS and volcanic eruptions when considering the entire observation period. When we analysed shorter time-windows (6 months in the year preceding an eruption), we observed increased ARS incidences in eruptive months, except in 2010. IRRs were increased for centres situated close to volcanoes (< 26 km) in 2001 and 2002. CONCLUSION: ARS incident cases increased over the years in populations living around the Nyamulagira and Nyiragongo volcanoes, but we found no consistent evidence for an association between the yearly incidence of ARS and volcanic eruptions or the intensity of SO(2) emissions, possibly because of interference with man-made events, including massive population displacements caused by insecurity in the area. Nevertheless, some evidence was found for increased incidence of ARS following eruptions, especially in areas close to volcanoes. Assessing personal, ground level exposure to SO(2) and particulates with adequate controlling for confounding, such as viral and other infections, could clarify the contribution, if any, of volcanic emissions of SO(2) to the high burden of respiratory diseases in this region.