For assessing the effect and threshold of PM2.5 on mortality in highly polluted areas and further studying the standard applicability, daily data on meteorological factors, air pollutants, and mortality were obtained in Jinan, China, from 2011 to 2017. A generalized additive model (GAM) and a distributed lag non-linear model (DLNM) were employed to assess the nonlinearity and the hysteresis of associations. We further explored the breakpoints to evaluate the existence of the threshold. The correlation between mortality and PM2.5 was nonlinear. The impact of average PM2.5 on non-accidental mortality (RR = 1.11; 95% CI = 1.06, 1.16), cardiovascular disease (CVD) mortality (RR = 1.17; 95% CI = 1.10, 1.24), and respiratory disease (RD) mortality (RR = 1.17; 95% CI = 1.10, 1.24) reached the highest in the current day (lag 0). The excess risks of PM2.5 at secondary standard level to non-accidental, CVD, and RD mortality are 8.79% (95% CI = 3.84, 13.98), 14.41% (95% CI = 7.79, 21.43), 15.35% (95% CI = 1.76, 30.74), respectively. The saturation points exist in highly polluted areas. Above the saturation points of 247 μg/m3 for non-accidental mortality, 245 μg/m3 for CVD mortality, and 250 μg/m3 for RD mortality, the model of all three relationships presented a harvesting effect. This study underscores the necessity of the ongoing efforts of reducing particulate air pollution and the adjustment of the standards in seriously polluted areas to adapt to regional conditions. At the same time, for highly polluted areas, it is advocated to strengthen personal protection to decrease the saturation point and control the concentration of pollutants as much as possible, which will substantially save more cost that benefits the public.
Keywords: Cardiovascular diseases; Particulate matter; Public health; Respiratory tract diseases.