*4.1. Limitations*

We acknowledge some limitations in our present study. Firstly, we have not explored inflammatory markers, such as fibrinogen, C-reactive protein or cytokine family, in the Northern compared with Southern Chinese. This will be valuable for confirming the hypothesis of generalized vascular inflammation in AP-induced atherogenic process. Secondly, the concentration and LDL-C happened (by chance) to be lower, but hyperemia

(by ultrasound) to be higher in Northern Chinese. These, however, have not contributed to the worse FMD and IMT results, since these two factors would have been associated with better rather than worse FMD and IMT levels in the Northern Chinese. Thirdly, we have identified FMD, an early atherosclerotic surrogate, is lower in Northern Chinese on univariate analyses, but its relationship with PM2.5 concentration is borderline only on multivariate analyses. FMD in more labile and dynamic compared with carotid IMT measurement, subjected to daily fluctuation of PM2.5 concentration. This was measured once only during the study period. Perhaps more FMD measurements for individual subject over the study period may illuminate the real impact of PM2.5 on FMD. Fourthly, realtime long term PM2.5 measurement is more informative and better than yearly mean PM2.5 estimation for studying its relationship with more labile FMD measurement. This, however, has logistic and economic implications which may not be readily resolved.
