Few studies have been conducted to investigate the acute health effects of visibility and haze, which may be regarded as proxy indicators of ambient air pollution. We used a distributed lag non-linear model (DLNM) combined with quasi-Poisson regression to estimate the relationship between visibility, haze and mortality in Ningbo, a coastal city of China. We found that the mortality risk of visibility was statistically significant only on the current day, while the risk of haze and PM10 peaked on the second day and could last for three days. When the visibility was less than 10 km, each 1 km decrease of visibility at lag 0 day was associated with a 0.78% (95% CI: 0.22-1.36%) increase in total mortality and a 1.61% (95% CI: 0.39-2.85%) increase in respiratory mortality. The excess risk of haze at lag 0-2 days on total mortality, cardiovascular and respiratory mortality was 7.76% (95% CI: 3.29-12.42%), 7.73% (95% CI: 0.12-15.92%) and 17.77% (95% CI: 7.64-28.86%), respectively. Greater effects of air pollution were observed during the cold season than in the warm season, and the elderly were at higher risk compared to youths. The effects of visibility and haze were attenuated by single pollutants. These findings suggest that visibility and haze could be used as surrogates of air quality where pollutant data are scarce, and strengthen the evidence to develop policy to control air pollution and protect vulnerable populations.
Keywords: ambient air pollution; haze; mortality; visibility; vulnerable populations.