Fig. 1
Exposure-response effect of average fine particulate matter PM2.5 exposure from 1998 to 2013 on cardiovascular disease mortality (upper panel) and cerebrovascular disease mortality (lower panel) in the whole study population. The curves are based on a restricted cubic spline with a knot placed at the median of the distribution of average PM2.5 exposure and the p-values refer to the test of non-linearity of the exposure-response relationship obtained through a likelihood ratio test comparing the depicted spline-based model with a model including only a linear effect of the variable