Consistent with the findings of prior studies, we observed that PM10 was significantly associated with reduced TP and LF in Wuhan; whereas in Zhuhai, there was no clear dose–response relationship between PM10 and HRV. Meanwhile, the association of PM2.5 with HRV appears to be inconsistent with earlier observational studies in panel of elderly subjects (Pope et al., 2004). However, Folino et al. (2009) found that HRV reductions were not associated with personal 24 h cumulative exposure to PM10 and PM2.5 in patients with myocardial infarction and this 8-CPT-2Me-cAMP finding is in accord with our negative results. Conflicting results from the present study may be due in part to differences in PM sources and compositions, exposure intensity and duration, study design or populations, use of different methods of exposure assessment or other reasons. Previous studies did not always have consistent results about the effects on autonomic functions from exposures to PM. Some studies reported negative exposure–response relationship between PM2.5 and HRV (Cavallari et al., 2007, Wu et al., 2010 and Wu et al., 2013), whereas others have reported positive associations (Magari et al., 2002 and Riediker et al., 2004) or zero association (Sullivan et al., 2005). Heterogeneity in the effect of PM has also been observed in previous multicenter studies, like the ULTRA study (Timonen et al., 2006), which found that the effects of PM on HRV likely vary depending on pollution sources and particulate constituents. Increase in PM2.5 was associated with decrease in HF in Helsinki, but the opposite was true in Erfurt, and no such association was observed in Amsterdam. On the other hand, the results might partly be explained by differences in the health status of the subjects. In the present study, there are obvious differences in characteristics of study population, including smoking, alcohol use, physical activity and history of diabetes between Wuhan and Zhuhai. Various physiological conditions or disease processes can alter autonomic control, and change the HRV. Moreover, baseline HRV status including TP and LF were significantly lower in Wuhan than in Zhuhai. Therefore, energy is possible that effects of particulate air pollution on HRV can be different even among healthy subjects. Few other studies have examined this association in healthy volunteer in different PM exposure levels. The annual averages of API in Wuhan were continually higher than those in Zhuhai, thus the personal concentrations of PM in Wuhan are much higher than those in Zhuhai. It seemed that serious air pollution may be thought to contribute to the effect of PM10 on HRV. Moreover, a better exposure–response relationship in the city-combined than in the city-specific analyses due to the increasing magnitude of exposure and the lager sample numbers or originate from Wuhan environment. Furthermore, most epidemiological studies had a window of PM measurements ranging from 2 h up to 1 day or two before HRV measurements (Pieters et al., 2012) or concurrent measurements of several minutes PM and HRV (Jia et al., 2012), however, our study focused on the concurrent and continuous effects of PM exposure on HRV within 24 h and not allow for evaluation of the lag effects or real-time/transient effects of PM on HRV. Therefore, our results cannot be compared directly with that of previous published studies.