In total, 90, 657 CVD ER visits occurred during the study period, among which 17, 509 (19.3%), 9,565 (10.6%), 3,447 (3.8%), and 29, 112 (32.1%) visits were due to IHD, HRD, HF, and CD, respectively. Table 1 showed the descriptive statistics for ER visits, weather, and air pollution conditions. As shown in Table 1, there were 35 emergency room visits for CVD per day during the study period on average. The daily average number of emergency room visits with IHD, HRD, HF, CD was found to be 7, 4, 2, 11, respectively. Of the total number of cases, 59.5% were young people(༜65 years) and 59.0% were male. The average daily mean temperature was 11.39℃, relative humidity was 50.63%, PM10 was 113.82 µg/m3, SO2 was 22.54 µg/m3 and NO2 was 45.51 µg/m3.
Table 1
Characteristics of ER visits and distributions of meteorological and air pollution variables
Variable | Mean ± SD | Min | P25 | Median | P75 | Max |
Daily visits to emergency rooms |
CVD | 35 ± 14 | 6 | 25 | 33 | 44 | 99 |
IHD | 7 ± 4 | 1 | 4 | 6 | 9 | 31 |
HRD | 4 ± 3 | 0 | 2 | 3 | 5 | 18 |
HF | 2 ± 1 | 0 | 1 | 2 | 2 | 8 |
CD | 11 ± 5 | 1 | 8 | 11 | 14 | 34 |
Sex | | | | | | |
Male | 22 ± 8 | 2 | 14 | 19 | 26 | 62 |
Female | 13 ± 5 | 1 | 10 | 14 | 18 | 50 |
Age | | | | | | |
༜65 | 21 ± 9 | 3 | 15 | 20 | 26 | 67 |
≥ 65 | 14 ± 6 | 1 | 10 | 13 | 18 | 42 |
Air pollutants (µg/m3) |
PM10 | 113.82 ± 81.51 | 16.00 | 70.95 | 100.00 | 137.80 | 1484.54 |
SO2 | 22.54 ± 14.97 | 3.54 | 11.05 | 18.42 | 30.25 | 115.51 |
NO2 | 45.51 ± 17.91 | 6.32 | 33.96 | 44.31 | 54.00 | 146.60 |
Meteorological factors |
Mean temperature (℃) | 11.39 ± 9.85 | -12.30 | 2.35 | 12.89 | 19.89 | 30.40 |
Relative Humidity (%) | 50.63 ± 15.31 | 11.71 | 39.22 | 51.00 | 61.62 | 96.09 |
Figure 2 depicted the three-dimensional plots of the daily mean temperature on total and cause-specific cardiovascular ER visits along lags of 21 days compared to MMT. In general, a non-linear association was found between mean temperature and total CVD, IHD, HRD, HF and CD ER visits, both low and high temperatures were relevant to increased numbers of total and cause-specific cardiovascular ER. As for total CVD, IHD, HRD and CD, both low temperature and high temperature showed immediate harmful effects on the current day (lag 0), and the effects decreased with the extension of lag days. But we found that the association of temperature with HF may have a different lag pattern. For HF, low temperature began to have harmful effect at lag 4, and the harmful effect lasted until lag 21, high temperature had a harmful effect on lag 0, then the corresponding RR decreased from lag 0-lag 14 and then increased markedly after lag 14.
Figure 3 showed the overall effect of ambient temperature on total and cause-specific cardiovascular ER visits over lag 0–21 days. Compared to the MMT, risk of total and subgroups with CVD increased significantly at both high and low temperatures. The effect of temperature on total CVD and CD showed more of an inverted U-shaped curve, whereas on IHD, HRD, and HF showed more of a nearly irregular M-shaped curve.
Figure 4 presented the estimated risks of total and cause-specific cardiovascular ER visits associated with cold (-4.2°C) and hot temperatures (25.4°C) from single lag 0 to lag 21. Delay effects for different cardiovascular ER visits categories were quite different. For example, significant harmful effects of cold temperature were observed in total CVD, IHD, HRD, and CD. The lag intervals with adverse effects were lag 7 ~ lag 18 for total CVD, lag 7 ~ lag 9 and lag 14 ~ lag 21 for IHD, lag 8 ~ lag 17 for HRD, as well as lag 0 and lag 8 ~ lag 11 for CD, suggesting the hysteresis phenomenon of cold temperature was obvious on various CVD ER visits. The significant and strong baleful effects of hot temperatures were observed on the current day (lag 0) for total CVD, IHD, and CD, the relative risks were 1.387 (95% CI: 1.177–1.634), 1.891 (95% CI: 1.297–2.756) and 1.5056 (95% CI: 1.121–2.022), respectively. However, the effects of heat on HRD and HF were not statistically significant. In general, obvious cold effects appeared after a lag of 7–8 days and persisted several days, whereas significant hot effects occurred immediately.
The cumulative effects of cold and hot temperature (and 95th percentile of temperature) on total and cause-specific cardiovascular ER visits at lag 0, lag 0–3, lag 0–7, lag 0–14 and lag 0–21 were presented in Table 2. For total CVD, IHD, and HRD, cumulative cold effects were more evident at lag 0–14 and lag 0–21. The cumulative cold effects for the above three diseases increased with longer lag days and reached maximum at lag 0–21, the corresponding relative risk (RR) were 1.446 (95% CI: 1.281–1.632) for total CVD, 2.841 (95% CI: 2.155–3.745) for IHD, and 1.884(95% CI: 1.288–2.754) for HRD, respectively. In contrast, the RR of cold temperature on HF ER visits was only statistically significant (RR = 2.447, 95% CI: 1.364–4.393) at lag 0–21. For CD, the cumulative cold effect at lag 0–21 was the biggest (RR = 1.289, 95% CI: 1.039–1.599) due to cold temperatures. Regarding the effects of hot temperature, hot temperature could significantly increase the total and cause-specific CVD ER visits, but hot effects over lag of different days were quite different. For total CVD, IHD, and CD, the high temperature showed a pronounced hazard effects on the current lag day (lag 0), and the RR value first decreased and then increased with the hysteresis days. The RR associated with hot temperature at lag 0–21 were 1.313 (95% CI: 1.133–1.520) for total CVD, 2.648 (95% CI: 1.891–3.709) for IHD, 2.036 (95% CI: 1.285–3.226) for HRD, 1.519 (95% CI: 1.050–2.197) for HF, and 1.445 (95% CI: 1.113–1.874) for CD. Both the cold and hot effects on total and cause-specific cardiovascular ER visits at lag 0–21 were statistically significant. Furthermore, the cold effects were greater than the heat effects for CVD, IHD and HF, in contrast to HRD and CD, where the cold effects were less than the heat effects.
Table 2
The cumulative relative risks and 95% CI of cold and hot temperatures on total and cause-specific cardiovascular diseases, including IHD, HRD, HF and CD, compared with the MMT
| Lag | CVD | IHD | HRD | HF | CD |
Cold effects | Lag 0 | 1.082(0.982,1.191) | 1.144(0.914,1.430) | 0.980(0.727,1.320) | 0.920(0.688,1.232) | 1.213(1.019,1.444)* |
| Lag 0–3 | 1.012(0.927,1.104) | 1.015(0.831,1.239) | 1.041(0.789,1.373) | 0.843(0.402,1.767) | 1.044(0.891,1.224) |
| Lag 0–7 | 1.038(0.949,1.135) | 1.230(1.002,1.509)* | 1.042(0.787,1.379) | 1.016(0.464,2.223) | 1.117(0.950,1.313) |
| Lag 0–14 | 1.314(1.182,1.462)* | 1.545(1.213,1.968)* | 1.712(1.216,2.408)* | 1.740(0.814,3.717) | 1.359(1.124,1.643)* |
| Lag 0–21 | 1.446(1.281,1.632)* | 2.841(2.155,3.745)* | 1.884(1.288,2.754)* | 2.447(1.364,4.393)* | 1.289(1.039,1.599)* |
Hot effects | Lag 0 | 1.387(1.177,1.634)* | 1.891(1.297,2.756)* | 1.403(0.846,2.327) | 1.041(0.926,1.170) | 1.506(1.121,2.022)* |
| Lag 0–3 | 1.026(0.889,1.185) | 1.114(0.803,1.543) | 1.469(0.938,2.302) | 1.174(0.873,1.580) | 1.032(0.798,1.335) |
| Lag 0–7 | 1.155(0.999,1.336) | 1.891(1.358,2.633)* | 1.597(1.015,2.512)* | 1.354(0.970,1.891) | 1.151(0.888,1.492) |
| Lag 0–14 | 1.349(1.147,1.588)* | 1.656(1.144,2.398)* | 1.950(1.159,3.282)* | 1.459(1.023,2.081)* | 1.521(1.138,2.032)* |
| Lag 0–21 | 1.313(1.133,1.520)* | 2.648(1.891,3.709)* | 2.036(1.285,3.226)* | 1.519(1.050,2.197)* | 1.445(1.113,1.874)* |
Cold effects: The 5th percentile of temperature (-4.2℃); Hot effects: The 95th percentile of temperature (25.4℃); The reference temperatures were MMT: -12.3℃ for total CVD, IHD, HRD and CD, 29.7℃ for HF. |
*:p < 0.05 |
Figure 5 showed the cumulative effects of temperature on total and cause-specific cardiovascular ER visits stratified by gender and age. The discrepancy of RR in the gender group varied among total and cause-specific CVD ER visits. For total CVD and IHD, temperatures related risks of both male and female ER visits increased at low and high temperatures compared to the MMT. The risks of cold and hot temperatures were slightly higher for males with IHD or HF than for females, whereas females with total CVD, HRD or CD were slightly more vulnerable to cold and hot temperatures. For total CVD, IHD, and HRD, positive associations between cold and hot temperatures and patients aged < 65 years were observed. However, the risk for total CVD, IHD, and HF in patients aged 65 years and older only increased at cold temperature. There was no significant ER visits risk associated with low and high temperature was observed in those older than 65 years for HRD and CD.
Sensitivity analyses showed that the estimated results changed a little (Supplementary Fig. 1-Supplementary Fig. 5), when we changed the df (6,8,9) of time, the df (3,5,6) for RH, PM10, SO2, NO2 and the maximum lag days (22–23).