International Journal of Cardiology 177 (2014) 436–441
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Acute effect of ambient air pollution on heart failure in Guangzhou, China☆ Changyuan Yang a,b,1, Ailan Chen c,1, Renjie Chen a,b,1, Yongqing Qi d, Jianjun Ye d, Shuangming Li d, Wanglin Li e, Zijing Liang f, Qing Liang f, Duanqiang Guo f, Haidong Kan a,b,⁎, Xinyu Chen g,h,i,⁎⁎ a
School of Public Health, Key Lab of Public Health Safety of the Ministry of Education, Fudan University, Shanghai 200032, China Key Lab of Health Technology Assessment of the Ministry of Health, Fudan University, Shanghai 200032, China Department of Cardiology, The First Affiliated Hospital of Guangzhou Medical University, Guangzhou 510120, China d Guangzhou First Aid Medical service Control Center, Guangzhou 510630, China e Department of Gastrointestinal Surgery, Affiliated Guangzhou First Municipal People's Hospital, Guangzhou Medical University, Guangzhou 21018, China f Department of Emergency, The First Affiliated Hospital of Guangzhou Medical University, Guangzhou 510120, China g State Key Laboratory of Respiratory Disease, Guangzhou Institute of Respiratory Disease, The First Affiliated Hospital, Guangzhou Medical University, Guangzhou 510182, China h Guangzhou Hoffmann Institute of Immunology, School of Basic Sciences, Guangzhou Medical University, Guangzhou 510182, China i Department of Pathogenic Biology, Guangzhou Medical University, Guangzhou 510182, China b c
a r t i c l e
i n f o
Article history: Received 7 February 2014 Received in revised form 2 September 2014 Accepted 15 September 2014 Available online 13 October 2014 Keywords: Air pollution Heart failure Emergency ambulance dispatches Time-series study Emergency medicine
a b s t r a c t Background: Heart failure (HF) is a global public health problem of increasing importance. The association between acute exposure to air pollution and HF has been well established in developed countries, but little evidence was available in developing countries where air pollution levels were much higher. We conducted a time-series study to investigate the short-term association between air pollution and overall emergency ambulance dispatches (EAD) due to HF in Guangzhou, China. Methods: Daily data of EAD due to HF from 1 January 2008 to 31 December 2012 were obtained from Guangzhou Emergency Center. We applied the over-dispersed Poisson generalized addictive model to analyze the associations after controlling for the seasonality, day of the week and weather conditions. Results: We identified a total of 3375 EAD for HF. A 10-μg/m3 increase in the present-day concentrations of particulate matter with an aerodynamic diameter of less than 10 μm, sulfur dioxide and nitrogen dioxide corresponded to increases of 3.54% [95% confidence interval (CI): 1.35%, 5.74%], 5.29% (95% CI: 2.28%, 8.30%) and 4.34% (95% CI: 1.71%, 6.97%) in daily EAD for HF, respectively. The effects of air pollution on acute HF were restricted on the concurrent day and in the cool seasons. Conclusions: Our results provided the first population-based evidence in Mainland China that outdoor air pollution could trigger the exacerbation of HF. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Heart failure (HF) is a common, costly, disabling, and potentially deadly condition. HF is one of the leading causes of hospitalization and deaths in people older than 65 [1]. It was estimated that more than 23
☆ All authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. ⁎ Corresponding author. ⁎⁎ Correspondence to: X. Chen, State Key Laboratory of Respiratory Disease, Guangzhou Institute of Respiratory Disease, The First Affiliated Hospital, Guangzhou Medical University, Guangzhou 510182, China. E-mail addresses: [email protected] (H. Kan), [email protected] (X. Chen). 1 These authors contributed equally to this work.
http://dx.doi.org/10.1016/j.ijcard.2014.09.003 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
million people have suffered from HF worldwide [2]. HF has become a global public health problem of increasing importance, because its prevalence and incidence were increasing due to the extended life span, as well as elevated prevalence of risk factors and improved survival rates from other cardiovascular disease [3]. Many risk factors have been related to HF, including ischemic heart disease, cigarette smoking, hypertension, obesity, diabetes, valvular heart disease, cardiomyopathy and obstructive sleep apnea [4]. In recent decades, the associations of ambient air pollution and adverse cardiovascular outcomes have been well established in terms of both epidemiological and toxicological studies [5–7]. Outdoor air pollution has been identified as a potential trigger of acute HF [6]. The effect was more evident in patients with pre-existing chronic heart congestion, hypertension, and arrhythmia [8]. However, most studies were performed in developed countries; thus evidence needed to be replicated and
C. Yang et al. / International Journal of Cardiology 177 (2014) 436–441
confirmed in developing countries where air pollution levels and prevalence of HF and its risk factors might differ much from the developed countries. China, as the largest developing country, may face the worst air pollution problems in the world. However, no prior evidence exists concerning the effect of air pollution on HF. In this study, we conducted a time-series analysis to investigate the short-term association between air pollution and emergency ambulance dispatches (EAD) due to HF in Guangzhou, China. 2. Methods 2.1. Data collection Guangzhou is the third largest Chinese city and largest city in southern China. In 2010, the city had a permanent population of 12.78 million. Guangzhou has a humid subtropical climate. Summers are wet with high temperatures, high humidity and a high heat index. Winters are mild and comparatively dry. Guangzhou is the main manufacturing hub of the Pearl River Delta, one of China's leading commercial and manufacturing regions. The economic development in Guangzhou is accompanied by severe air pollution problem that might pose harm to people's health [9]. 2.1.1. Health data EAD is sensitive to the acute attack of health events, and thus was used in this study. Daily data of EAD due to HF from 1 January 2008 to 31 December 2012 were obtained from Guangzhou Emergency Center. This center has established the dynamic dispatching system covering the overall urban areas of Guangzhou where more than 7 million permanent people reside. It owns almost 200 ambulance cars which are all well-equipped with the resuscitation. The emergency system is responsible for the routine emergency ambulance service for urban residents and pre-hospital resuscitation. According to the rules of Guangzhou Emergency Center, the ambulance cars must arrive at the scene within 30 min whenever the emergency call is made (even during the midnight). The determination of HF diagnosis was made by physicians in Emergency Departments of hospitals based on the first presentation of patients at the emergency cars. Then, the HF-related EAD were aggregated on a daily basis. 2.1.2. Environmental data We obtained daily data on air pollutants from Guangzhou Environmental Monitoring Center during the same study period, which has 9 state-controlling stations in the urban areas of Guangzhou. The daily (24-hour) average concentrations of particulate matter with an aerodynamic diameter of less than 10 μm (PM10), sulfur dioxide (SO2) and nitrogen dioxide (NO2) were measured using the method of tapered element oscillating microbalance, ultraviolet fluorescence and chemiluminescence, respectively. According to Chinese government rules, the location of the 9 monitoring stations is mandated not to be in the direct vicinity of traffic or of industrial sources and not to be influenced by local pollution sources, thus also avoiding buildings or housing and large emitters such as coal-, waste-, or oil-burning boilers; furnaces; and incinerators. Therefore, the measurements could reflect the general exposure levels of urban background air pollution. The daily air pollutants' concentrations were averaged from the available monitoring results across various stations. To allow adjustment for the effect of weather conditions on HF, daily weather data (mean temperature and relative humidity) were obtained from Guangzhou Meteorological Bureau [10]. 2.2. Statistical analysis Time-series regression was applied to explore the short-term association between air pollution and EAD due to HF. Time-series analysis is a
437
regular analytic method to explore the acute health effects of air pollution based on the daily aggregate data, because it can control for both time-invariant and time-varying confounders by design [11]. We applied generalized additive models (GAM) to analyze the data [12]. Because daily EAD typically followed an over-dispersed Poisson distribution, we used the quasi-Poisson regression in the GAM [13]. We introduced several covariates to control for their potential confounding effects: (1) a natural cubic smooth function of calendar time with 7 degrees of freedom (df) per year to exclude unmeasured longterm and seasonal trends longer than 2 months [14]; (2) natural smooth functions of the current-day mean temperature (6 df) and relative humidity (3 df) in order to control for the potential nonlinear confounding effects of weather conditions, (3) an indicator variable for “day of the week”, and (4) an indicator variable for holidays. After the basic model was established, we first used single-pollutant models and introduced a priori in turn each air pollutant's concentrations on the concurrent day (lag 0), because previous studies indicated that current-day air pollution was most closely correlated with acute cardiovascular events including HF [6,7]. In order to explore the lag structure for the air pollution's effects on HF, we used more single lag days (lag 1, 2, 3). Considering the possible temporal misalignment of using exposure on single-day lags, we performed additional analyses using moving average exposure of multiple days, including lag 0–1, 0–2, 0–3. We plotted the concentration–response relationship curves for air pollutants and HF-related EAD using a 3 df for the spline function. Because both air pollution levels and incidence of HF events were known to vary by season, we divided our analyses by the cool period (November to April) and warm (May to October) period, and reduced the df per year from 7 in all year to 4 in the cool-period and warmperiod analyses. In order to check the stability of our results, we conducted 4 sensitivity analyses. First, we built two-pollutant models to examine the stability of effect estimates after adjustment for co-pollutants. Second, given that the health effects of ambient air temperature could last for multiple days, we fit the basic model using longer lags of temperatures, including the moving average of lag 0 and lag 1 day (lag 0–1), the moving average of lag 0 and lag 3 day (lag 0–3), the moving average of lag 0 and lag 7 day (lag 0–7), and the moving average of lag 0 and lag 14 day (lag 0–14). Meanwhile, few literatures have shown that humidity had a significant and lagged confounding effect on the association between air pollution and health outcomes, so we did not evaluate the sensitivity by different lags of humidity. Third, we selected alternative df with 4–10 per year for the smoothness of time trends. Fourth, we analyzed the short-term association between air pollution and daily EAD for injuries that were not biologically related with air pollution using the same basic model. The statistical tests were two-sided, and effects of p b 0.05 were considered as statistically significant. All statistical models were run in R software (version 2.15.3) using the MGCV package. The effects were presented as the percentage of change and 95% confidence intervals (CIs) in daily EAD for HF per 10 μg/m3 increase of pollutant concentrations. 3. Results Table 1 provides the summary of the descriptive statistics in this study. There were 3375 EAD for HF during the study period (1827 days). The average time for ambulance transportation was 15 min. The annual average concentrations were 57.0 μg/m3 for PM10, 31.4 μg/m3 for SO2, and 35.0 μg/m3 for NO2. As shown in S-Fig. 1 in the Supplementary Materials, the daily HF events and air pollutant concentrations were the highest in winter and the lowest in summer. The annual average temperature and relative humidity were 22.25 °C and 73.81%, respectively, reflecting the subtropical climate in Guangzhou. There were only 17 days with missing values for air pollutant measurements during our study period, and no missing values for health data and meteorological variables. These missing days were directly excluded from the dataset since they were rare (0.9% of the total).
438
C. Yang et al. / International Journal of Cardiology 177 (2014) 436–441
Table 1 The summary of descriptive statistics in this study. Variables
Mean
SD
Min
Percentiles 1
Daily EAD due to HF Air pollution (μg/m3) PM10 NO2 SO2 Weather conditions Mean temperature (°C) Relative humidity (%)
0
Max 25
50
75
99
0
1
2
3
6
10
1.9
1.5
57.0 35.0 31.4
21.2 17.6 28.7
7.1 9.9 2.8
15.3 14.0 5.3
42.4 23.3 19.1
55.3 30.1 28.7
68.9 40.8 41.7
119.0 100.0 73.3
173.6 136.8 84.1
22.3 73.8
6.5 13.1
5.1 25.0
6.9 35.0
17.6 66.0
24.0 76.0
27.4 83.0
31.4 95.0
33.5 99.0
Abbreviations: EAD, emergency ambulance dispatches; HF, heart failure; SD, standard deviations; PM10, particulate matter with an aerodynamic diameter of less than 10 μm; SO2, sulfur dioxide, and NO2, nitrogen dioxide.
Generally, there were strong correlations among the 3 air pollutants with the Spearman coefficients ranging from 0.60 to 0.82. Air pollutants were moderately or weakly correlated with temperature and humidity (data not shown). Table 2 presents the effect estimates in single-pollutant models using different lag days. We consistently found the air pollution-HF associations were the most significant and largest on the concurrent day
(lag 0), then decreased substantially and turned to be insignificant on the subsequent days. Estimates using cumulative lag days were similar with those using lag 0 day. A 10-μg/m3 increase in the concentrations of PM10, SO2 and NO2 at lag 0 corresponded to increases of 3.54% (95% CI: 1.35%, 5.74%), 5.29% (95% CI: 2.28%, 8.30%), and 4.34% (95% CI: 1.71%, 6.97%) in daily EAD for HF, respectively. Fig. 1 graphically shows the concentration–response relationships for PM10, SO2 and NO2 with HF events. The curves for PM10 and NO2 appeared to be linearly positive without any thresholds. The SO2's effects increased linearly with its concentration, but turned to be almost stable when it was beyond 50 μg/m3. Table 2 also shows the effect estimates separated by the cool and warm periods. It was clearly seen that the effect estimates in cool period were much higher than in warm period, and significant associations were consistently restricted in the cool period. Similarly, the magnitude of air pollution's effects decreased greatly from lag 0 to lag 3 in cool period. However, there was an apparent increase of the effect size and decrease of the precision of estimation at lag 3 in warm period; then, the effect estimates turned to be null at longer lags (data not shown). Table 3 provides the results of two-pollutant models using exposure at lag 0. The effect estimates of all the 3 pollutants attenuated a little but were still statistically significant when adjusting for co-pollutants. S-Table 1 in the Supplementary Materials shows the results of sensitivity analyses adjusting for temperature by different lag days. The effect estimates remained stable with adjustment for different lags of temperature. S-Table 2 demonstrated that the acute effects of air pollution did not change substantially with the adjustment of smoothness of time using alternative df from 4 to 10 per year. As another sensitivity analysis, we analyzed the association between air pollution and injuries. On average, there were 104 EAD for injuries per day. We found null effects of air pollution on EAD due to injuries, suggesting our main results might be not attributable to a chance of modeling (See S-Table 3). 4. Discussion
Fig. 1. The concentration–response relationship curves for the present-day concentrations of PM10, SO2 and NO2 with HF-related EAD. PM10, particulate matter with an aerodynamic diameter of less than 10 μm; SO2, sulfur dioxide, NO2, nitrogen dioxide; HF, heart failure; and EAD, emergency ambulance dispatches.
Based on the daily data on air pollutant concentrations and EAD due to HF in Guangzhou from 2008 to 2012, this time-series study indicated that outdoor air pollution could trigger the acute exacerbation of HF in the general population of Guangzhou, China. The effects of air pollution on acute HF were restricted on the concurrent day and in the cool period. Our estimates were robust to varying model parameters. Up to our knowledge, this was the first epidemiological study examining the potential detrimental effects of air pollution on HF in China. Many epidemiologic studies have estimated a significant association between air pollution and increased risk of HF, although the health indicators varied from hospitalization to mortality [5]. Recently, Shah and colleagues provided a systematic review of the effects of air pollution on HF [6]. Although the significant heterogeneity of the individual effect estimates, the overall evidence showed a significantly positive association of PM10, SO2, NO2 with HF hospitalization or mortality primarily in high-income nations where the level of air pollution was much
C. Yang et al. / International Journal of Cardiology 177 (2014) 436–441
439
Table 2 Percent change (mean and 95% confidence intervals) of daily emergency ambulance dispatches due to heart failure per 10 μg/m3 increase in pollutant concentrations at different lag days in warm and cool period. Pollutants
Lag days
PM10
0 1 2 3 0–1 0–2 0–3 0 1 2 3 0–1 0–2 0–3 0 1 2 3 0–1 0–2 0–3
SO2
NO2
Whole perioda
Cool periodb
Warm periodb
3.54 (1.35, 5.74)⁎ 2.05 (−0.24, 4.34) 0.37 (−1.63, 2.36) 0.10 (−1.84, 2.03) 4.02 (1.51, 6.52)⁎ 3.03 (0.29, 5.78)⁎ 2.97 (0.09, 5.84)⁎ 5.29 (2.28, 8.30)⁎
4.30 (1.64, 6.95)⁎ 2.35 (−0.17, 4.88) 0.24 (−2.11, 2.59) −0.39 (−2.65, 1.87) 4.91 (1.80, 8.02)⁎ 4.05 (0.74, 7.35)⁎ 3.61 (0.17, 7.06)⁎ 7.17 (3.22, 11.11)⁎
1.10 (−1.84, 4.04) −0.49 (−3.24, 2.25) −0.46 (−3.10, 2.18) 4.44 (0.90, 7.98)⁎ 3.81 (−0.08, 7.70) 4.03 (0.01, 8.05)⁎ 4.34 (1.71, 6.97)⁎
1.44 (−2.34, 5.22) −1.11 (−4.57, 2.34) −1.61 (−4.87, 1.65) 6.28 (1.58, 10.98)⁎ 3.94 (−1.05, 8.93) 3.25 (−1.89, 8.38) 5.31 (2.26, 8.36)⁎
2.56 (−0.06, 5.18) 1.51 (−0.94, 3.96) 0.05 (−2.32, 2.43) 4.68 (1.65, 7.70)⁎ 4.88 (1.51, 8.24)⁎ 3.91 (0.32, 7.50)⁎
2.60 (−0.41, 5.62) 1.27 (−1.52, 4.05) −0.69 (−3.37, 2.00) 5.60 (2.03, 9.16)⁎ 5.59 (1.73, 9.45)⁎
1.44 (−3.39, 6.28) 0.73 (−3.86, 5.32) 0.21 (−4.33, 4.74) 1.77 (−2.76, 6.31) 1.33 (−4.13, 6.79) 1.26 (−4.52, 7.05) 2.31 (−3.68, 8.29) 3.64 (−1.81, 9.10) −0.09 (−5.54, 5.36) 0.42 (−5.00, 5.84) 4.52 (−0.86, 9.90) 2.67 (−3.93, 9.27) 3.16 (−3.96, 10.28) 7.08 (−0.25, 14.41) 2.46 (−4.29, 9.21) 1.83 (−4.88, 8.53) 4.32 (−2.34,10.97) 5.91 (−0.71,12.54) 3.12 (−4.86, 11.09) 5.59 (−2.77, 13.96) 6.95 (−1.67, 15.56)
4.05 (−0.10, 8.20)
Abbreviations: PM10, particulate matter with an aerodynamic diameter of less than 10 μm; SO2, sulfur dioxide, NO2, nitrogen dioxide. Cool period: November to April. Warm period: May to October. Lags 0–1, 0–2, 0–3 refer to moving averages of the concurrent day and preceding 1 day, 2 days and 3 days, respectively. ⁎ p b 0.05. a Generalized additive models adjusting for time trends (7 degrees of freedom per year), day of week, holidays, mean temperature and humidity. b Generalized additive models adjusting for time trends (4 degrees of freedom per year), day of week, holidays, mean temperature and humidity.
lower than China. In Shah's meta-analysis, a 10-μg/m3 increase in the present-day concentrations of PM10, SO2 and NO2 were associated with 1.63% (95% CI: 1.20%, 2.07%), 0.83% (95% CI: 0.47%, 1.18%) and 0.83% (95% CI: 0.61%, 1.05%) increases in HF hospitalizations or mortality [6]. We observed high effect estimates in Guangzhou than previous studies in high-income countries. There were several potential reasons for the heterogeneity. First, this may reflect the fact that China's air pollution was much severer than developed countries. China is one of the rapidly urbanizing countries with bouts of smog and high levels of air pollution. For example, national annual average concentrations of PM2.5 (particulate matter with an aerodynamic diameter less than 2.5 μm) in 2013 were 72 μg/m3, almost 7-fold higher than the guidelines (10 μg/m3) recommended by the World Health Organization [15] and the general levels that were observed in Western Europe and North America. In Guangzhou, the PM2.5 level (52 μg/m3) was lower than the national average, but still 5 fold higher than the WHO guidelines [15]. Although no legal monitoring measurements of PM2.5 were available during the study period from 2008 to 2012, the annual average PM10 concentrations (60.0 μg/m3) were still 3-fold higher than the WHO
Table 3 Percent change (mean and 95% confidence intervals) of daily ambulance dispatches due to HF on lag 0 day in two-pollutant models.
PM10
SO2
NO2
Two-pollutant models
Estimates
– + SO2 + NO2 – + PM10 + NO2 – + PM10 + SO2
3.54 (1.35, 5.74) ⁎ 2.43 (0.02, 4.84) ⁎ 2.62 (0.00, 5.24) ⁎ 5.29 (2.28, 8.30) ⁎ 4.82 (1.26, 8.37)⁎ 4.19 (0.66, 7.72)⁎ 4.34 (1.71, 6.97) ⁎ 3.77 (0.74, 6.80)⁎ 3.05 (0.05, 6.04)⁎
Abbreviations: PM10, particulate matter with an aerodynamic diameter of less than 10 μm; SO2, sulfur dioxide, and NO2, nitrogen dioxide. ⁎ p b 0.05.
guidelines (20 μg/m3). Second, the varying magnitude of misclassification of clinical diagnosis, as well as other factors such as statistical models and population characteristics, may explain the differences between our results and previous studies. The higher estimates for SO2/NO2 than PM10 in this study could be explained by the different scales of day-to-day variation in their concentrations. These effect estimates turned to be quite similar when using the scale of interquartile range rather than unit increase of concentrations (data not shown). As an indicator of morbidity, EAD have been used in environmental epidemiological studies [16–19]. The current study adopted daily HFrelated EAD in a city to reflect its overall incidence of acute HF events. The emergency indicator might be more sensitive to reflect the acute attack of HF in China, where the hospitalization data was more likely to be influenced by the bed space and mortality data can not reflect the nonfatal attack of HF events. Nevertheless, the ambulance service system may selectively exclude home deaths caused by HF events. Season was a well-known factor that could modify the associations between air pollution and health outcomes. In current analysis, we found the effects of air pollution on HF events varied greatly by season. The mean effect estimates during the cool period were 2–3 times higher than during the warm period, and the statistical significance was restricted within the cool seasons. The reasons for seasonal differences were still unclear, but may be related with higher pollution levels and incidence rate of HF events in cool period. We found that the associations between air pollution and HF were statistically significant only using the exposure at lag 0. Similarly, Shah and colleagues found the strongest associations were seen at lag 0, with this effect diminishing at longer lag times in a systematic review [6]. The underlying pathophysiological mechanisms for the very acute effects of air pollution on HF events are still unclear, but several pathways were postulated to might be responsible, including dysfunction of autonomic nervous system, systematic inflammation, oxidative stress, blood coagulation [5]. Some basic contributing conditions of HF, such as hypertension, arrhythmia, cardiac ischemia and infarction, might also be exacerbated by acute exposure to air pollution [7,20,21]. Specifically, short exposure to ambient PM has been shown to be associated with increased levels of blood pressure [22,23] and risks of acute
440
C. Yang et al. / International Journal of Cardiology 177 (2014) 436–441
myocardial infarction [7] and heart diseases attributable to chronic obstructive pulmonary diseases [24], all of which had the potential to induce a HF event. Nevertheless, given the multi-linearity and exposure measurement errors of these pollutants, we cannot separate the specific damage that these pollutants could have on the exacerbation of HF. Understanding the exposure–response relationship was crucial to the formation of public health policies. This study indicated that the risks of the exacerbation of HF could increase linearly corresponding to the acute exposure to outdoor air pollution without any appreciable thresholds. Therefore, further abatement of air pollution could gain proportional benefits in decreasing the burden of HF. In China, HF was a major cause for cardiovascular hospitalizations with a total prevalence of 0.9% in the general populations. Its burden was estimated to increase in response to population aging and increased prevalence of chronic non-communicable diseases and their risk factors [25]. Further, unlike other risk factors, exposure to air pollution was ubiquitous. In consequence, the air pollution-related burden of HF can be an increasingly important public health problem in China where the population size was the largest, air pollution levels were among the highest and its effects may be higher than the high-income countries. Public health interventions might be also informed from our results. For example, patients who suffered from chronic HF and related basic diseases can be taught to stay at home when outdoor air pollution is high, especially in cool seasons. Our study had several strengths. First, to our knowledge, most of previous evidence concerning the air pollution-HF associations was obtained in areas of relatively low air pollution levels; this study was the first to investigate the association between air pollution and HF in Mainland China, where the air pollution levels were much higher than the developed countries. Second, this study was one of the few studies based on emergency data, which might be more sensitive to reflect the acute effects of air pollution on HF events than hospitalization or mortality in China. Third, in comparison to previous studies based on emergency department visits of one or several hospitals [26,27], this study collected the overall EAD data in a city, and thus might be more representative for the incidence of acute HF events in the general populations. Nevertheless, several limitations of our study should also be mentioned. First, we failed to analyze the effects of PM2.5 that may have stronger correlations with HF exacerbations [6], because it was not a criteria pollutant in China during our study period. Second, as most previous time-series studies, exposure measurement error was inevitable in this study, which has been shown to bias estimates downward [28,29]. Third, we failed to perform stratified analyses by ages, sex, and comorbid conditions, because this was a retrospective study and the origin data did not include these individual information. Actually, the absence of stratified analysis by age, sex and comorbidities would not bias our estimates, because they are not confounders in this study; but they can modify the air pollution-HF associations in various subpopulations as effect modifiers. For example, the elderly [30] and those with existing chronic cardiopulmonary diseases [31] may have higher risk of HF associated with air pollution exposure. Fourth, the diagnosis of HF events made by emergency physicians might not be absolutely accurate and completely consistent in various hospitals of Guangzhou; we also failed to validate these diagnoses in this retrospective analysis. Therefore, further studies were still needed to confirm our results. In summary, this time-series study demonstrated that short-term exposure to ambient air pollution was associated with increased EAD due to HF in Guangzhou, China, especially in cool seasons. Our results provided the first population-based evidence in Mainland China that outdoor air pollution could trigger the exacerbation of HF. Conflict of interest The authors declare no potential conflicts of interests.
Acknowledgment The study was supported by the Public Welfare Research Program of National Health and Family Planning Commission of China (201402022), Guangzhou Department of Science and Technology, China (2013J4500017), Guangdong Natural Science Foundation (S2013010014478, S201101004269), and Foundation of Guangzhou Municipal Health Bureau Scientific Research and Education Management System (201102A213125). Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijcard.2014.09.003. References [1] Zannad F, Agrinier N, Alla F. Heart failure burden and therapy. Europace 2009; 11(Suppl. 5):v1–9. [2] Roger VL. Epidemiology of heart failure. Circ Res 2013;113:646–59. [3] Stewart S, MacIntyre K, Capewell S, McMurray J. Heart failure and the aging population: an increasing burden in the 21st century? Heart 2003;89:49–53. [4] Bui AL, Horwich TB, Fonarow GC. Epidemiology and risk profile of heart failure. Nat Rev Cardiol 2010;8:30–41. [5] Brook RD, Rajagopalan S, Pope CA, et al. Particulate matter air pollution and cardiovascular disease an update to the scientific statement from the American heart association. Circulation 2010;121:2331–78. [6] Shah AS, Langrish JP, Nair H, et al. Global association of air pollution and heart failure: a systematic review and meta-analysis. Lancet 2013;382:1039–48. [7] Mustafic H, Jabre P, Caussin C, et al. Main air pollutants and myocardial infarction: a systematic review and meta-analysis. JAMA 2012;307:713–21. [8] Colais P, Faustini A, Stafoggia M, et al. Particulate air pollution and hospital admissions for cardiac diseases in potentially sensitive subgroups. Epidemiology 2012;23:473–81. [9] Huang D, Xu J, Zhang S. Valuing the health risks of particulate air pollution in the Pearl River Delta, China. Environ Sci Pol 2012;15:38–47. [10] Wilker EH, Yeh G, Wellenius GA, Davis RB, Phillips RS, Mittleman MA. Ambient temperature and biomarkers of heart failure: a repeated measures analysis. Environ Health Perspect 2012;120:1083–7. [11] Nitta H, Yamazaki S, Omori T, Sato T. An introduction to epidemiologic and statistical methods useful in environmental epidemiology. J Epidemiol 2009;20:177–84. [12] Dominici F, Peng RD, Bell ML, et al. Fine particulate air pollution and hospital admission for cardiovascular and respiratory diseases. JAMA 2006;295:1127–34. [13] Wang X, Chen R, Meng X, Geng F, Wang C, Kan H. Associations between fine particle, coarse particle, black carbon and hospital visits in a Chinese city. Sci Total Environ 2013;458:1–6. [14] Peng RD, Dominici F, Louis TA. Model choice in time series studies of air pollution and mortality. J R Stat Soc: Ser A: Appl Stat 2006;169:179–203. [15] Organization W.H. WHO Air quality guidelines for particulate matter, ozone, nitrogen dioxide and sulfur dioxide: global update 2005: summary of risk assessment; 2006. [16] Ng CF, Ueda K, Ono M, Nitta H, Takami A. Characterizing the effect of summer temperature on heatstroke-related emergency ambulance dispatches in the Kanto area of Japan. Int J Biometeorol 2014;58:941–8. [17] Alessandrini E, Zauli Sajani S, Scotto F, Miglio R, Marchesi S, Lauriola P. Emergency ambulance dispatches and apparent temperature: a time series analysis in EmiliaRomagna, Italy. Environ Res 2011;111:1192–200. [18] Straney L, Finn J, Dennekamp M, Bremner A, Tonkin A, Jacobs I. Evaluating the impact of air pollution on the incidence of out-of-hospital cardiac arrest in the Perth Metropolitan Region: 2000–2010. J Epidemiol Community Health 2014;68:6–12. [19] Turner LR, Connell D, Tong S. Exposure to hot and cold temperatures and ambulance attendances in Brisbane, Australia: a time-series study. BMJ 2012;2. [20] Guo Y, Tong S, Zhang Y, Barnett AG, Jia Y, Pan X. The relationship between particulate air pollution and emergency hospital visits for hypertension in Beijing, China. Sci Total Environ 2010;408:4446–50. [21] Bartell SM, Longhurst J, Tjoa T, Sioutas C, Delfino RJ. Particulate air pollution, ambulatory heart rate variability, and cardiac arrhythmia in retirement community residents with coronary artery disease. Environ Health Perspect 2013;121:1135–41. [22] Rich DQ, Freudenberger RS, Ohman-Strickland P, Cho Y, Kipen HM. Right heart pressure increases after acute increases in ambient particulate concentration. Environ Health Perspect 2008;116:1167–71. [23] Auchincloss AH, Roux AVD, Dvonch JT, et al. Associations between recent exposure to ambient fine particulate matter and blood pressure in the Multi-Ethnic Study of Atherosclerosis (MESA). Environ Health Perspect 2008;116:486–91. [24] Meng X, Wang CC, Cao DC, Wong CM, Kan HD. Short-term effect of ambient air pollution on COPD mortality in four Chinese cities. Atmos Environ 2013;77:149–54. [25] Ariely R, Evans K, Mills T. Heart failure in China: a review of the literature. Drugs 2013;73:689–701. [26] Stieb DM, Beveridge RC, Brook JR, et al. Air pollution, aeroallergens and cardiorespiratory emergency department visits in Saint John, Canada. J Expo Anal Environ Epidemiol 2000;10:461–77.
C. Yang et al. / International Journal of Cardiology 177 (2014) 436–441 [27] Stieb DM, Szyszkowicz M, Rowe BH, Leech JA. Air pollution and emergency department visits for cardiac and respiratory conditions: a multi-city time-series analysis. Environ Health 2009;8:25. [28] Strickland MJ, Gass KM, Goldman GT, Mulholland JA. Effects of ambient air pollution measurement error on health effect estimates in time-series studies: a simulationbased analysis. J Expo Sci Environ Epidemiol 2013. http://dx.doi.org/10.1038/jes. 2013.16. [29] Goldman GT, Mulholland JA, Russell AG, et al. Impact of exposure measurement error in air pollution epidemiology: effect of error type in time-series studies. Environ Health 2011;10:61.
441
[30] Pope CA, Renlund DG, Kfoury AG, May HT, Horne BD. Relation of heart failure hospitalization to exposure to fine particulate air pollution. Am J Cardiol 2008;102: 1230–4. [31] Wellenius GA, Bateson TF, Mittleman MA, Schwartz J. Particulate air pollution and the rate of hospitalization for congestive heart failure among medicare beneficiaries in Pittsburgh, Pennsylvania. Am J Epidemiol 2005;161:1030–6.
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Acute effect of ambient air pollution on heart failure in Guangzhou, China☆ Changyuan Yang a,b,1, Ailan Chen c,1, Renjie Chen a,b,1, Yongqing Qi d, Jianjun Ye d, Shuangming Li d, Wanglin Li e, Zijing Liang f, Qing Liang f, Duanqiang Guo f, Haidong Kan a,b,⁎, Xinyu Chen g,h,i,⁎⁎ a
School of Public Health, Key Lab of Public Health Safety of the Ministry of Education, Fudan University, Shanghai 200032, China Key Lab of Health Technology Assessment of the Ministry of Health, Fudan University, Shanghai 200032, China Department of Cardiology, The First Affiliated Hospital of Guangzhou Medical University, Guangzhou 510120, China d Guangzhou First Aid Medical service Control Center, Guangzhou 510630, China e Department of Gastrointestinal Surgery, Affiliated Guangzhou First Municipal People's Hospital, Guangzhou Medical University, Guangzhou 21018, China f Department of Emergency, The First Affiliated Hospital of Guangzhou Medical University, Guangzhou 510120, China g State Key Laboratory of Respiratory Disease, Guangzhou Institute of Respiratory Disease, The First Affiliated Hospital, Guangzhou Medical University, Guangzhou 510182, China h Guangzhou Hoffmann Institute of Immunology, School of Basic Sciences, Guangzhou Medical University, Guangzhou 510182, China i Department of Pathogenic Biology, Guangzhou Medical University, Guangzhou 510182, China b c
a r t i c l e
i n f o
Article history: Received 7 February 2014 Received in revised form 2 September 2014 Accepted 15 September 2014 Available online 13 October 2014 Keywords: Air pollution Heart failure Emergency ambulance dispatches Time-series study Emergency medicine
a b s t r a c t Background: Heart failure (HF) is a global public health problem of increasing importance. The association between acute exposure to air pollution and HF has been well established in developed countries, but little evidence was available in developing countries where air pollution levels were much higher. We conducted a time-series study to investigate the short-term association between air pollution and overall emergency ambulance dispatches (EAD) due to HF in Guangzhou, China. Methods: Daily data of EAD due to HF from 1 January 2008 to 31 December 2012 were obtained from Guangzhou Emergency Center. We applied the over-dispersed Poisson generalized addictive model to analyze the associations after controlling for the seasonality, day of the week and weather conditions. Results: We identified a total of 3375 EAD for HF. A 10-μg/m3 increase in the present-day concentrations of particulate matter with an aerodynamic diameter of less than 10 μm, sulfur dioxide and nitrogen dioxide corresponded to increases of 3.54% [95% confidence interval (CI): 1.35%, 5.74%], 5.29% (95% CI: 2.28%, 8.30%) and 4.34% (95% CI: 1.71%, 6.97%) in daily EAD for HF, respectively. The effects of air pollution on acute HF were restricted on the concurrent day and in the cool seasons. Conclusions: Our results provided the first population-based evidence in Mainland China that outdoor air pollution could trigger the exacerbation of HF. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Heart failure (HF) is a common, costly, disabling, and potentially deadly condition. HF is one of the leading causes of hospitalization and deaths in people older than 65 [1]. It was estimated that more than 23
☆ All authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. ⁎ Corresponding author. ⁎⁎ Correspondence to: X. Chen, State Key Laboratory of Respiratory Disease, Guangzhou Institute of Respiratory Disease, The First Affiliated Hospital, Guangzhou Medical University, Guangzhou 510182, China. E-mail addresses: [email protected] (H. Kan), [email protected] (X. Chen). 1 These authors contributed equally to this work.
http://dx.doi.org/10.1016/j.ijcard.2014.09.003 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
million people have suffered from HF worldwide [2]. HF has become a global public health problem of increasing importance, because its prevalence and incidence were increasing due to the extended life span, as well as elevated prevalence of risk factors and improved survival rates from other cardiovascular disease [3]. Many risk factors have been related to HF, including ischemic heart disease, cigarette smoking, hypertension, obesity, diabetes, valvular heart disease, cardiomyopathy and obstructive sleep apnea [4]. In recent decades, the associations of ambient air pollution and adverse cardiovascular outcomes have been well established in terms of both epidemiological and toxicological studies [5–7]. Outdoor air pollution has been identified as a potential trigger of acute HF [6]. The effect was more evident in patients with pre-existing chronic heart congestion, hypertension, and arrhythmia [8]. However, most studies were performed in developed countries; thus evidence needed to be replicated and
C. Yang et al. / International Journal of Cardiology 177 (2014) 436–441
confirmed in developing countries where air pollution levels and prevalence of HF and its risk factors might differ much from the developed countries. China, as the largest developing country, may face the worst air pollution problems in the world. However, no prior evidence exists concerning the effect of air pollution on HF. In this study, we conducted a time-series analysis to investigate the short-term association between air pollution and emergency ambulance dispatches (EAD) due to HF in Guangzhou, China. 2. Methods 2.1. Data collection Guangzhou is the third largest Chinese city and largest city in southern China. In 2010, the city had a permanent population of 12.78 million. Guangzhou has a humid subtropical climate. Summers are wet with high temperatures, high humidity and a high heat index. Winters are mild and comparatively dry. Guangzhou is the main manufacturing hub of the Pearl River Delta, one of China's leading commercial and manufacturing regions. The economic development in Guangzhou is accompanied by severe air pollution problem that might pose harm to people's health [9]. 2.1.1. Health data EAD is sensitive to the acute attack of health events, and thus was used in this study. Daily data of EAD due to HF from 1 January 2008 to 31 December 2012 were obtained from Guangzhou Emergency Center. This center has established the dynamic dispatching system covering the overall urban areas of Guangzhou where more than 7 million permanent people reside. It owns almost 200 ambulance cars which are all well-equipped with the resuscitation. The emergency system is responsible for the routine emergency ambulance service for urban residents and pre-hospital resuscitation. According to the rules of Guangzhou Emergency Center, the ambulance cars must arrive at the scene within 30 min whenever the emergency call is made (even during the midnight). The determination of HF diagnosis was made by physicians in Emergency Departments of hospitals based on the first presentation of patients at the emergency cars. Then, the HF-related EAD were aggregated on a daily basis. 2.1.2. Environmental data We obtained daily data on air pollutants from Guangzhou Environmental Monitoring Center during the same study period, which has 9 state-controlling stations in the urban areas of Guangzhou. The daily (24-hour) average concentrations of particulate matter with an aerodynamic diameter of less than 10 μm (PM10), sulfur dioxide (SO2) and nitrogen dioxide (NO2) were measured using the method of tapered element oscillating microbalance, ultraviolet fluorescence and chemiluminescence, respectively. According to Chinese government rules, the location of the 9 monitoring stations is mandated not to be in the direct vicinity of traffic or of industrial sources and not to be influenced by local pollution sources, thus also avoiding buildings or housing and large emitters such as coal-, waste-, or oil-burning boilers; furnaces; and incinerators. Therefore, the measurements could reflect the general exposure levels of urban background air pollution. The daily air pollutants' concentrations were averaged from the available monitoring results across various stations. To allow adjustment for the effect of weather conditions on HF, daily weather data (mean temperature and relative humidity) were obtained from Guangzhou Meteorological Bureau [10]. 2.2. Statistical analysis Time-series regression was applied to explore the short-term association between air pollution and EAD due to HF. Time-series analysis is a
437
regular analytic method to explore the acute health effects of air pollution based on the daily aggregate data, because it can control for both time-invariant and time-varying confounders by design [11]. We applied generalized additive models (GAM) to analyze the data [12]. Because daily EAD typically followed an over-dispersed Poisson distribution, we used the quasi-Poisson regression in the GAM [13]. We introduced several covariates to control for their potential confounding effects: (1) a natural cubic smooth function of calendar time with 7 degrees of freedom (df) per year to exclude unmeasured longterm and seasonal trends longer than 2 months [14]; (2) natural smooth functions of the current-day mean temperature (6 df) and relative humidity (3 df) in order to control for the potential nonlinear confounding effects of weather conditions, (3) an indicator variable for “day of the week”, and (4) an indicator variable for holidays. After the basic model was established, we first used single-pollutant models and introduced a priori in turn each air pollutant's concentrations on the concurrent day (lag 0), because previous studies indicated that current-day air pollution was most closely correlated with acute cardiovascular events including HF [6,7]. In order to explore the lag structure for the air pollution's effects on HF, we used more single lag days (lag 1, 2, 3). Considering the possible temporal misalignment of using exposure on single-day lags, we performed additional analyses using moving average exposure of multiple days, including lag 0–1, 0–2, 0–3. We plotted the concentration–response relationship curves for air pollutants and HF-related EAD using a 3 df for the spline function. Because both air pollution levels and incidence of HF events were known to vary by season, we divided our analyses by the cool period (November to April) and warm (May to October) period, and reduced the df per year from 7 in all year to 4 in the cool-period and warmperiod analyses. In order to check the stability of our results, we conducted 4 sensitivity analyses. First, we built two-pollutant models to examine the stability of effect estimates after adjustment for co-pollutants. Second, given that the health effects of ambient air temperature could last for multiple days, we fit the basic model using longer lags of temperatures, including the moving average of lag 0 and lag 1 day (lag 0–1), the moving average of lag 0 and lag 3 day (lag 0–3), the moving average of lag 0 and lag 7 day (lag 0–7), and the moving average of lag 0 and lag 14 day (lag 0–14). Meanwhile, few literatures have shown that humidity had a significant and lagged confounding effect on the association between air pollution and health outcomes, so we did not evaluate the sensitivity by different lags of humidity. Third, we selected alternative df with 4–10 per year for the smoothness of time trends. Fourth, we analyzed the short-term association between air pollution and daily EAD for injuries that were not biologically related with air pollution using the same basic model. The statistical tests were two-sided, and effects of p b 0.05 were considered as statistically significant. All statistical models were run in R software (version 2.15.3) using the MGCV package. The effects were presented as the percentage of change and 95% confidence intervals (CIs) in daily EAD for HF per 10 μg/m3 increase of pollutant concentrations. 3. Results Table 1 provides the summary of the descriptive statistics in this study. There were 3375 EAD for HF during the study period (1827 days). The average time for ambulance transportation was 15 min. The annual average concentrations were 57.0 μg/m3 for PM10, 31.4 μg/m3 for SO2, and 35.0 μg/m3 for NO2. As shown in S-Fig. 1 in the Supplementary Materials, the daily HF events and air pollutant concentrations were the highest in winter and the lowest in summer. The annual average temperature and relative humidity were 22.25 °C and 73.81%, respectively, reflecting the subtropical climate in Guangzhou. There were only 17 days with missing values for air pollutant measurements during our study period, and no missing values for health data and meteorological variables. These missing days were directly excluded from the dataset since they were rare (0.9% of the total).
438
C. Yang et al. / International Journal of Cardiology 177 (2014) 436–441
Table 1 The summary of descriptive statistics in this study. Variables
Mean
SD
Min
Percentiles 1
Daily EAD due to HF Air pollution (μg/m3) PM10 NO2 SO2 Weather conditions Mean temperature (°C) Relative humidity (%)
0
Max 25
50
75
99
0
1
2
3
6
10
1.9
1.5
57.0 35.0 31.4
21.2 17.6 28.7
7.1 9.9 2.8
15.3 14.0 5.3
42.4 23.3 19.1
55.3 30.1 28.7
68.9 40.8 41.7
119.0 100.0 73.3
173.6 136.8 84.1
22.3 73.8
6.5 13.1
5.1 25.0
6.9 35.0
17.6 66.0
24.0 76.0
27.4 83.0
31.4 95.0
33.5 99.0
Abbreviations: EAD, emergency ambulance dispatches; HF, heart failure; SD, standard deviations; PM10, particulate matter with an aerodynamic diameter of less than 10 μm; SO2, sulfur dioxide, and NO2, nitrogen dioxide.
Generally, there were strong correlations among the 3 air pollutants with the Spearman coefficients ranging from 0.60 to 0.82. Air pollutants were moderately or weakly correlated with temperature and humidity (data not shown). Table 2 presents the effect estimates in single-pollutant models using different lag days. We consistently found the air pollution-HF associations were the most significant and largest on the concurrent day
(lag 0), then decreased substantially and turned to be insignificant on the subsequent days. Estimates using cumulative lag days were similar with those using lag 0 day. A 10-μg/m3 increase in the concentrations of PM10, SO2 and NO2 at lag 0 corresponded to increases of 3.54% (95% CI: 1.35%, 5.74%), 5.29% (95% CI: 2.28%, 8.30%), and 4.34% (95% CI: 1.71%, 6.97%) in daily EAD for HF, respectively. Fig. 1 graphically shows the concentration–response relationships for PM10, SO2 and NO2 with HF events. The curves for PM10 and NO2 appeared to be linearly positive without any thresholds. The SO2's effects increased linearly with its concentration, but turned to be almost stable when it was beyond 50 μg/m3. Table 2 also shows the effect estimates separated by the cool and warm periods. It was clearly seen that the effect estimates in cool period were much higher than in warm period, and significant associations were consistently restricted in the cool period. Similarly, the magnitude of air pollution's effects decreased greatly from lag 0 to lag 3 in cool period. However, there was an apparent increase of the effect size and decrease of the precision of estimation at lag 3 in warm period; then, the effect estimates turned to be null at longer lags (data not shown). Table 3 provides the results of two-pollutant models using exposure at lag 0. The effect estimates of all the 3 pollutants attenuated a little but were still statistically significant when adjusting for co-pollutants. S-Table 1 in the Supplementary Materials shows the results of sensitivity analyses adjusting for temperature by different lag days. The effect estimates remained stable with adjustment for different lags of temperature. S-Table 2 demonstrated that the acute effects of air pollution did not change substantially with the adjustment of smoothness of time using alternative df from 4 to 10 per year. As another sensitivity analysis, we analyzed the association between air pollution and injuries. On average, there were 104 EAD for injuries per day. We found null effects of air pollution on EAD due to injuries, suggesting our main results might be not attributable to a chance of modeling (See S-Table 3). 4. Discussion
Fig. 1. The concentration–response relationship curves for the present-day concentrations of PM10, SO2 and NO2 with HF-related EAD. PM10, particulate matter with an aerodynamic diameter of less than 10 μm; SO2, sulfur dioxide, NO2, nitrogen dioxide; HF, heart failure; and EAD, emergency ambulance dispatches.
Based on the daily data on air pollutant concentrations and EAD due to HF in Guangzhou from 2008 to 2012, this time-series study indicated that outdoor air pollution could trigger the acute exacerbation of HF in the general population of Guangzhou, China. The effects of air pollution on acute HF were restricted on the concurrent day and in the cool period. Our estimates were robust to varying model parameters. Up to our knowledge, this was the first epidemiological study examining the potential detrimental effects of air pollution on HF in China. Many epidemiologic studies have estimated a significant association between air pollution and increased risk of HF, although the health indicators varied from hospitalization to mortality [5]. Recently, Shah and colleagues provided a systematic review of the effects of air pollution on HF [6]. Although the significant heterogeneity of the individual effect estimates, the overall evidence showed a significantly positive association of PM10, SO2, NO2 with HF hospitalization or mortality primarily in high-income nations where the level of air pollution was much
C. Yang et al. / International Journal of Cardiology 177 (2014) 436–441
439
Table 2 Percent change (mean and 95% confidence intervals) of daily emergency ambulance dispatches due to heart failure per 10 μg/m3 increase in pollutant concentrations at different lag days in warm and cool period. Pollutants
Lag days
PM10
0 1 2 3 0–1 0–2 0–3 0 1 2 3 0–1 0–2 0–3 0 1 2 3 0–1 0–2 0–3
SO2
NO2
Whole perioda
Cool periodb
Warm periodb
3.54 (1.35, 5.74)⁎ 2.05 (−0.24, 4.34) 0.37 (−1.63, 2.36) 0.10 (−1.84, 2.03) 4.02 (1.51, 6.52)⁎ 3.03 (0.29, 5.78)⁎ 2.97 (0.09, 5.84)⁎ 5.29 (2.28, 8.30)⁎
4.30 (1.64, 6.95)⁎ 2.35 (−0.17, 4.88) 0.24 (−2.11, 2.59) −0.39 (−2.65, 1.87) 4.91 (1.80, 8.02)⁎ 4.05 (0.74, 7.35)⁎ 3.61 (0.17, 7.06)⁎ 7.17 (3.22, 11.11)⁎
1.10 (−1.84, 4.04) −0.49 (−3.24, 2.25) −0.46 (−3.10, 2.18) 4.44 (0.90, 7.98)⁎ 3.81 (−0.08, 7.70) 4.03 (0.01, 8.05)⁎ 4.34 (1.71, 6.97)⁎
1.44 (−2.34, 5.22) −1.11 (−4.57, 2.34) −1.61 (−4.87, 1.65) 6.28 (1.58, 10.98)⁎ 3.94 (−1.05, 8.93) 3.25 (−1.89, 8.38) 5.31 (2.26, 8.36)⁎
2.56 (−0.06, 5.18) 1.51 (−0.94, 3.96) 0.05 (−2.32, 2.43) 4.68 (1.65, 7.70)⁎ 4.88 (1.51, 8.24)⁎ 3.91 (0.32, 7.50)⁎
2.60 (−0.41, 5.62) 1.27 (−1.52, 4.05) −0.69 (−3.37, 2.00) 5.60 (2.03, 9.16)⁎ 5.59 (1.73, 9.45)⁎
1.44 (−3.39, 6.28) 0.73 (−3.86, 5.32) 0.21 (−4.33, 4.74) 1.77 (−2.76, 6.31) 1.33 (−4.13, 6.79) 1.26 (−4.52, 7.05) 2.31 (−3.68, 8.29) 3.64 (−1.81, 9.10) −0.09 (−5.54, 5.36) 0.42 (−5.00, 5.84) 4.52 (−0.86, 9.90) 2.67 (−3.93, 9.27) 3.16 (−3.96, 10.28) 7.08 (−0.25, 14.41) 2.46 (−4.29, 9.21) 1.83 (−4.88, 8.53) 4.32 (−2.34,10.97) 5.91 (−0.71,12.54) 3.12 (−4.86, 11.09) 5.59 (−2.77, 13.96) 6.95 (−1.67, 15.56)
4.05 (−0.10, 8.20)
Abbreviations: PM10, particulate matter with an aerodynamic diameter of less than 10 μm; SO2, sulfur dioxide, NO2, nitrogen dioxide. Cool period: November to April. Warm period: May to October. Lags 0–1, 0–2, 0–3 refer to moving averages of the concurrent day and preceding 1 day, 2 days and 3 days, respectively. ⁎ p b 0.05. a Generalized additive models adjusting for time trends (7 degrees of freedom per year), day of week, holidays, mean temperature and humidity. b Generalized additive models adjusting for time trends (4 degrees of freedom per year), day of week, holidays, mean temperature and humidity.
lower than China. In Shah's meta-analysis, a 10-μg/m3 increase in the present-day concentrations of PM10, SO2 and NO2 were associated with 1.63% (95% CI: 1.20%, 2.07%), 0.83% (95% CI: 0.47%, 1.18%) and 0.83% (95% CI: 0.61%, 1.05%) increases in HF hospitalizations or mortality [6]. We observed high effect estimates in Guangzhou than previous studies in high-income countries. There were several potential reasons for the heterogeneity. First, this may reflect the fact that China's air pollution was much severer than developed countries. China is one of the rapidly urbanizing countries with bouts of smog and high levels of air pollution. For example, national annual average concentrations of PM2.5 (particulate matter with an aerodynamic diameter less than 2.5 μm) in 2013 were 72 μg/m3, almost 7-fold higher than the guidelines (10 μg/m3) recommended by the World Health Organization [15] and the general levels that were observed in Western Europe and North America. In Guangzhou, the PM2.5 level (52 μg/m3) was lower than the national average, but still 5 fold higher than the WHO guidelines [15]. Although no legal monitoring measurements of PM2.5 were available during the study period from 2008 to 2012, the annual average PM10 concentrations (60.0 μg/m3) were still 3-fold higher than the WHO
Table 3 Percent change (mean and 95% confidence intervals) of daily ambulance dispatches due to HF on lag 0 day in two-pollutant models.
PM10
SO2
NO2
Two-pollutant models
Estimates
– + SO2 + NO2 – + PM10 + NO2 – + PM10 + SO2
3.54 (1.35, 5.74) ⁎ 2.43 (0.02, 4.84) ⁎ 2.62 (0.00, 5.24) ⁎ 5.29 (2.28, 8.30) ⁎ 4.82 (1.26, 8.37)⁎ 4.19 (0.66, 7.72)⁎ 4.34 (1.71, 6.97) ⁎ 3.77 (0.74, 6.80)⁎ 3.05 (0.05, 6.04)⁎
Abbreviations: PM10, particulate matter with an aerodynamic diameter of less than 10 μm; SO2, sulfur dioxide, and NO2, nitrogen dioxide. ⁎ p b 0.05.
guidelines (20 μg/m3). Second, the varying magnitude of misclassification of clinical diagnosis, as well as other factors such as statistical models and population characteristics, may explain the differences between our results and previous studies. The higher estimates for SO2/NO2 than PM10 in this study could be explained by the different scales of day-to-day variation in their concentrations. These effect estimates turned to be quite similar when using the scale of interquartile range rather than unit increase of concentrations (data not shown). As an indicator of morbidity, EAD have been used in environmental epidemiological studies [16–19]. The current study adopted daily HFrelated EAD in a city to reflect its overall incidence of acute HF events. The emergency indicator might be more sensitive to reflect the acute attack of HF in China, where the hospitalization data was more likely to be influenced by the bed space and mortality data can not reflect the nonfatal attack of HF events. Nevertheless, the ambulance service system may selectively exclude home deaths caused by HF events. Season was a well-known factor that could modify the associations between air pollution and health outcomes. In current analysis, we found the effects of air pollution on HF events varied greatly by season. The mean effect estimates during the cool period were 2–3 times higher than during the warm period, and the statistical significance was restricted within the cool seasons. The reasons for seasonal differences were still unclear, but may be related with higher pollution levels and incidence rate of HF events in cool period. We found that the associations between air pollution and HF were statistically significant only using the exposure at lag 0. Similarly, Shah and colleagues found the strongest associations were seen at lag 0, with this effect diminishing at longer lag times in a systematic review [6]. The underlying pathophysiological mechanisms for the very acute effects of air pollution on HF events are still unclear, but several pathways were postulated to might be responsible, including dysfunction of autonomic nervous system, systematic inflammation, oxidative stress, blood coagulation [5]. Some basic contributing conditions of HF, such as hypertension, arrhythmia, cardiac ischemia and infarction, might also be exacerbated by acute exposure to air pollution [7,20,21]. Specifically, short exposure to ambient PM has been shown to be associated with increased levels of blood pressure [22,23] and risks of acute
440
C. Yang et al. / International Journal of Cardiology 177 (2014) 436–441
myocardial infarction [7] and heart diseases attributable to chronic obstructive pulmonary diseases [24], all of which had the potential to induce a HF event. Nevertheless, given the multi-linearity and exposure measurement errors of these pollutants, we cannot separate the specific damage that these pollutants could have on the exacerbation of HF. Understanding the exposure–response relationship was crucial to the formation of public health policies. This study indicated that the risks of the exacerbation of HF could increase linearly corresponding to the acute exposure to outdoor air pollution without any appreciable thresholds. Therefore, further abatement of air pollution could gain proportional benefits in decreasing the burden of HF. In China, HF was a major cause for cardiovascular hospitalizations with a total prevalence of 0.9% in the general populations. Its burden was estimated to increase in response to population aging and increased prevalence of chronic non-communicable diseases and their risk factors [25]. Further, unlike other risk factors, exposure to air pollution was ubiquitous. In consequence, the air pollution-related burden of HF can be an increasingly important public health problem in China where the population size was the largest, air pollution levels were among the highest and its effects may be higher than the high-income countries. Public health interventions might be also informed from our results. For example, patients who suffered from chronic HF and related basic diseases can be taught to stay at home when outdoor air pollution is high, especially in cool seasons. Our study had several strengths. First, to our knowledge, most of previous evidence concerning the air pollution-HF associations was obtained in areas of relatively low air pollution levels; this study was the first to investigate the association between air pollution and HF in Mainland China, where the air pollution levels were much higher than the developed countries. Second, this study was one of the few studies based on emergency data, which might be more sensitive to reflect the acute effects of air pollution on HF events than hospitalization or mortality in China. Third, in comparison to previous studies based on emergency department visits of one or several hospitals [26,27], this study collected the overall EAD data in a city, and thus might be more representative for the incidence of acute HF events in the general populations. Nevertheless, several limitations of our study should also be mentioned. First, we failed to analyze the effects of PM2.5 that may have stronger correlations with HF exacerbations [6], because it was not a criteria pollutant in China during our study period. Second, as most previous time-series studies, exposure measurement error was inevitable in this study, which has been shown to bias estimates downward [28,29]. Third, we failed to perform stratified analyses by ages, sex, and comorbid conditions, because this was a retrospective study and the origin data did not include these individual information. Actually, the absence of stratified analysis by age, sex and comorbidities would not bias our estimates, because they are not confounders in this study; but they can modify the air pollution-HF associations in various subpopulations as effect modifiers. For example, the elderly [30] and those with existing chronic cardiopulmonary diseases [31] may have higher risk of HF associated with air pollution exposure. Fourth, the diagnosis of HF events made by emergency physicians might not be absolutely accurate and completely consistent in various hospitals of Guangzhou; we also failed to validate these diagnoses in this retrospective analysis. Therefore, further studies were still needed to confirm our results. In summary, this time-series study demonstrated that short-term exposure to ambient air pollution was associated with increased EAD due to HF in Guangzhou, China, especially in cool seasons. Our results provided the first population-based evidence in Mainland China that outdoor air pollution could trigger the exacerbation of HF. Conflict of interest The authors declare no potential conflicts of interests.
Acknowledgment The study was supported by the Public Welfare Research Program of National Health and Family Planning Commission of China (201402022), Guangzhou Department of Science and Technology, China (2013J4500017), Guangdong Natural Science Foundation (S2013010014478, S201101004269), and Foundation of Guangzhou Municipal Health Bureau Scientific Research and Education Management System (201102A213125). Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijcard.2014.09.003. References [1] Zannad F, Agrinier N, Alla F. Heart failure burden and therapy. Europace 2009; 11(Suppl. 5):v1–9. [2] Roger VL. Epidemiology of heart failure. Circ Res 2013;113:646–59. [3] Stewart S, MacIntyre K, Capewell S, McMurray J. Heart failure and the aging population: an increasing burden in the 21st century? Heart 2003;89:49–53. [4] Bui AL, Horwich TB, Fonarow GC. Epidemiology and risk profile of heart failure. Nat Rev Cardiol 2010;8:30–41. [5] Brook RD, Rajagopalan S, Pope CA, et al. Particulate matter air pollution and cardiovascular disease an update to the scientific statement from the American heart association. Circulation 2010;121:2331–78. [6] Shah AS, Langrish JP, Nair H, et al. Global association of air pollution and heart failure: a systematic review and meta-analysis. Lancet 2013;382:1039–48. [7] Mustafic H, Jabre P, Caussin C, et al. Main air pollutants and myocardial infarction: a systematic review and meta-analysis. JAMA 2012;307:713–21. [8] Colais P, Faustini A, Stafoggia M, et al. Particulate air pollution and hospital admissions for cardiac diseases in potentially sensitive subgroups. Epidemiology 2012;23:473–81. [9] Huang D, Xu J, Zhang S. Valuing the health risks of particulate air pollution in the Pearl River Delta, China. Environ Sci Pol 2012;15:38–47. [10] Wilker EH, Yeh G, Wellenius GA, Davis RB, Phillips RS, Mittleman MA. Ambient temperature and biomarkers of heart failure: a repeated measures analysis. Environ Health Perspect 2012;120:1083–7. [11] Nitta H, Yamazaki S, Omori T, Sato T. An introduction to epidemiologic and statistical methods useful in environmental epidemiology. J Epidemiol 2009;20:177–84. [12] Dominici F, Peng RD, Bell ML, et al. Fine particulate air pollution and hospital admission for cardiovascular and respiratory diseases. JAMA 2006;295:1127–34. [13] Wang X, Chen R, Meng X, Geng F, Wang C, Kan H. Associations between fine particle, coarse particle, black carbon and hospital visits in a Chinese city. Sci Total Environ 2013;458:1–6. [14] Peng RD, Dominici F, Louis TA. Model choice in time series studies of air pollution and mortality. J R Stat Soc: Ser A: Appl Stat 2006;169:179–203. [15] Organization W.H. WHO Air quality guidelines for particulate matter, ozone, nitrogen dioxide and sulfur dioxide: global update 2005: summary of risk assessment; 2006. [16] Ng CF, Ueda K, Ono M, Nitta H, Takami A. Characterizing the effect of summer temperature on heatstroke-related emergency ambulance dispatches in the Kanto area of Japan. Int J Biometeorol 2014;58:941–8. [17] Alessandrini E, Zauli Sajani S, Scotto F, Miglio R, Marchesi S, Lauriola P. Emergency ambulance dispatches and apparent temperature: a time series analysis in EmiliaRomagna, Italy. Environ Res 2011;111:1192–200. [18] Straney L, Finn J, Dennekamp M, Bremner A, Tonkin A, Jacobs I. Evaluating the impact of air pollution on the incidence of out-of-hospital cardiac arrest in the Perth Metropolitan Region: 2000–2010. J Epidemiol Community Health 2014;68:6–12. [19] Turner LR, Connell D, Tong S. Exposure to hot and cold temperatures and ambulance attendances in Brisbane, Australia: a time-series study. BMJ 2012;2. [20] Guo Y, Tong S, Zhang Y, Barnett AG, Jia Y, Pan X. The relationship between particulate air pollution and emergency hospital visits for hypertension in Beijing, China. Sci Total Environ 2010;408:4446–50. [21] Bartell SM, Longhurst J, Tjoa T, Sioutas C, Delfino RJ. Particulate air pollution, ambulatory heart rate variability, and cardiac arrhythmia in retirement community residents with coronary artery disease. Environ Health Perspect 2013;121:1135–41. [22] Rich DQ, Freudenberger RS, Ohman-Strickland P, Cho Y, Kipen HM. Right heart pressure increases after acute increases in ambient particulate concentration. Environ Health Perspect 2008;116:1167–71. [23] Auchincloss AH, Roux AVD, Dvonch JT, et al. Associations between recent exposure to ambient fine particulate matter and blood pressure in the Multi-Ethnic Study of Atherosclerosis (MESA). Environ Health Perspect 2008;116:486–91. [24] Meng X, Wang CC, Cao DC, Wong CM, Kan HD. Short-term effect of ambient air pollution on COPD mortality in four Chinese cities. Atmos Environ 2013;77:149–54. [25] Ariely R, Evans K, Mills T. Heart failure in China: a review of the literature. Drugs 2013;73:689–701. [26] Stieb DM, Beveridge RC, Brook JR, et al. Air pollution, aeroallergens and cardiorespiratory emergency department visits in Saint John, Canada. J Expo Anal Environ Epidemiol 2000;10:461–77.
C. Yang et al. / International Journal of Cardiology 177 (2014) 436–441 [27] Stieb DM, Szyszkowicz M, Rowe BH, Leech JA. Air pollution and emergency department visits for cardiac and respiratory conditions: a multi-city time-series analysis. Environ Health 2009;8:25. [28] Strickland MJ, Gass KM, Goldman GT, Mulholland JA. Effects of ambient air pollution measurement error on health effect estimates in time-series studies: a simulationbased analysis. J Expo Sci Environ Epidemiol 2013. http://dx.doi.org/10.1038/jes. 2013.16. [29] Goldman GT, Mulholland JA, Russell AG, et al. Impact of exposure measurement error in air pollution epidemiology: effect of error type in time-series studies. Environ Health 2011;10:61.
441
[30] Pope CA, Renlund DG, Kfoury AG, May HT, Horne BD. Relation of heart failure hospitalization to exposure to fine particulate air pollution. Am J Cardiol 2008;102: 1230–4. [31] Wellenius GA, Bateson TF, Mittleman MA, Schwartz J. Particulate air pollution and the rate of hospitalization for congestive heart failure among medicare beneficiaries in Pittsburgh, Pennsylvania. Am J Epidemiol 2005;161:1030–6.
