**2. Methods**

The study was conducted from 1 January 2009 to 30 June 2011. The description of data of EACs is presented in our previous work [15]. The values of daily NAO indices were obtained from the National Oceanic and Atmospheric Administration (NOAA) database ftp://ftp.cpc.ncep.noaa.gov/cwlinks/. We used the NAOI on the day of the call and on two previous days. As the quintiles of the daily NAOI were −1.01, −0.49, −0.06 and 0.43 during the study period and the non-linear impact of the NAO is probable, we used the NAOI as categorical variable with categories: ≤−1, (−1, −0.5], (−0.5, 0], (0, 0.5] and >0.5. As confounders, we used the non-linear variables of air temperature (T, ◦C), wind speed (WS, kt), barometric pressure (BP, hPA) and relative humidity (RH, %), low and active-stormy geomagnetic activity levels, high-speed solar wind (HSSW), daily concentrations of ozone and PM10 and the highest 8-h moving average of CO concentration. These variables are described in more detail in previous publications [15,16].
