What's new?

Air quality in Polish cities is among the worst in Europe. Previous studies reported that air pollution is an important risk factor for cardiovascular diseases, including ischemic heart disease (IHD). This study used the largest dataset to date in Poland to assess the changes in the risk of chronic and acute exacerbations of IHD following exposure to several air pollutants. We found that the risk of hospital admission due to chronic and acute IHD events increased within 3 days of exposure to particulate matter of an aerodynamic diameter no greater than 10 µm (PM₁₀) and no greater than 2.5 µm (PM_2.5). Moreover, exposure to nitrogen dioxide was associated with an increased risk of chronic events. Our results provide further evidence to substantiate the need to improve air quality in Poland.

Introduction

Ischemic heart disease (IHD) affects 126 million people worldwide and accounts for 9 million deaths annually, with a particularly high disease burden in Eastern European countries.^1-3 The disease is progressive and can become unstable, causing acute coronary syndromes (ACSs), such as myocardial infarction (MI) and heart failure.^4,5 However, most clinical presentations of IHD can be classified as chronic coronary syndromes (CCSs), which are typically associated with chest pain and / or dyspnea on exertion that can be relieved with rest and medications.⁶

The main risk factors for IHD include age, male sex, hypertension, diabetes, dyslipidemia, obesity, and smoking.⁷ Arterial hypertension is one of the key causes of circulatory system diseases in general,⁸ having the greatest impact on the incidence of human deaths worldwide, and air pollution is also among the factors contributing to arterial hypertension. However, air pollution emerges as the most important environmental risk factor for IHD.⁹ Chronic exposure to air pollution increases oxidative stress and inflammation, which lead to the development of IHD risk factors.^9,10 Epidemiologic studies showed that long‑term exposure to particulate and gaseous air pollutants is associated with an increased prevalence of IHD.¹¹ In addition, exposure to air pollution may lead to coronary vasomotor abnormalities and plaque destabilization, causing ACSs and CCSs.^9,12-15 Indeed, meta‑analyses of epidemiologic studies showed that exposure to both particulate matter (PM) of an aerodynamic diameter no greater than 2.5 µg (PM_2.5) and to NO₂ was associated with an increased short‑term risk of IHD events.^16-19 Air pollution can also affect the risk of atrial fibrillation (AF). It is the most common heart rhythm disorder, and during smog incidents, the frequency of AF episodes in the population is noted to increase.²⁰

In Poland, both IHD and air pollution are important public health issues. IHD is the leading cause of morbidity and mortality, accounting for 10% of all deaths.²¹ In Poland, recently reported number of deaths from IHD was 16.9 per 10 000 population,²² whereas air quality in Polish cities is among the worst in Europe: nearly half of the 50 most polluted European cities are in Poland.²³ However, the association between exposure to ambient air pollution and the burden of IHD in Poland has been so far only partially elucidated.^24-27 Therefore, we conducted a time‑stratified case‑crossover study to assess the link between ambient air pollution and the risk of hospital admission due to IHD.

Patients and Methods

Results

Air pollution and admissions

The daily concentrations of all air pollutants were the highest in the winter and the lowest in the summer (Figure 1A–1D). The mean concentrations of all air pollutants were the highest in Kraków and the lowest in Tricity (Table 1).

Table 1. Concentration of air pollutants, meteorological variables, and admission counts in the 3 agglomerations in Poland (2012–2017, daytime measurements)

Characteristic	Kraków (n = 2192)a	Tricity (n = 2192)a	Warsaw (n = 2192)a
Data are presented as median (interquartile range) unless provided otherwise. a Total number of days for which daytime measurements were considered Abbreviations: see Figure 1
PM10, μg/m3	35 (23–57)	15 (11–21)	27 (20–39)
PM2.5, μg/m3	25 (17–44)	10 (7–15)	21 (15–32)
NO2, μg/m3	40 (33–48)	14 (9–19)	34 (27–42)
SO2, μg/m3	5.5 (4–9.3)	2.5 (1.8–3.8)	4.8 (3.1–7.2)
Temperature, °C	10 (3–17)	9 (4–16)	9 (3–17)
Relative humidity, %	78 (68–86)	74 (67–80)	77 (65–87)
Atmospheric pressure, hPa	988 (984–992)	1007 (1002–1012)	1003 (998–1008)
Chronic coronary syndromes, n	17 340	11 683	59 444
Acute coronary syndromes, n	10 062	5420	19 921

The total admission counts for each ICD‑10 code were as follows: I20, 29 044; I21, 33 551; I22, 183; I23, 116; I24, 1553; and I25, 59 423. The monthly admission counts for CCSs and ACSs decreased steadily over the study period (Figure 1E and 1F).

Lagged association of exposure to PM₁₀, PM_2.5, NO₂, and SO₂ with a risk of hospital admission

After adjusting for meteorological variables, the risk of admission due to CCSs increased following the exposure to PM₁₀ (CRR per 10 µg/m³, 1.007; 95% CI, 1.002–1.012), PM_2.5 (CRR per 10 µg/m³, 1.008; 95% CI, 1.001–1.014), and NO₂ (CRR per 10 µg/m³, 1.037; 95% CI, 1.026–1.047). In contrast, the exposure to SO₂ was not associated with the risk for admission due to CCSs (CRR per 10 µg/m³, 0.991; 95% CI, 0.965–1.018). The daily and cumulative RRs for admission due to coronary syndromes and the relationship between pollutant concentrations and the admission risk are shown in Figure 2.

After adjusting for meteorological variables, the risk of admission due to ACSs increased during the 3 days from the exposure to PM₁₀ (CRR per 10 µg/m³, 1.009; 95% CI, 1.002–1.015) and PM_2.5 (CRR per 10 µg/m³, 1.009; 95% CI, 1.001–1.018). The risk increased on the day of exposure to NO₂ and on the first day after exposure to SO₂, but the lower limits of 95% CI for CRRs over 3 days were slightly below the value of no effect, that is, CRR per 10 µg/m³, 1.012; 95% CI, 0.998–1.026 for NO₂ and CRR per 10 µg/m³, 1.032; 95% CI, 0.998–1.067 for SO₂. The daily and cumulative RRs for admission due to ACSs and the relationship between pollutant concentrations and the admission risk are shown in Figure 3.

In single‑pollutant models, the exposure to PM₁₀ and PM_2.5 accounted for 1.9% to 2.9% of admissions due to CCS or ACS, whereas the exposure to NO₂ accounted for 11.7% of admissions due to CCSs (Table 2).

Table 2. Percentage of hospitalizations attributable to air pollution

Variable	Pollutant	Admission percentage (95% CI)
Only significant associations are presented. Abbreviations: see Figure 1
Chronic coronary syndromes	PM10	2.3 (0.7–4.1)
	PM2.5	1.9 (0.2–3.5)
	NO2	11.7 (8.6–14.6)
Acute coronary syndromes	PM10	2.9 (0.7–4.8)
	PM2.5	2.4 (0.1–4.5)

Discussion

In this study, we assessed over 100 000 hospital admissions due to ACSs or CCSs in the 3 largest urban agglomerations in Poland. The analysis showed that exposure to particulate and gaseous air pollutants was associated with an increased short‑term risk of IHD events requiring in‑hospital care. The risk of admission due to CCSs increased significantly after the exposure to PM₁₀, PM_2.5, and NO₂ but not SO₂. The risk of ACSs increased after the exposure to PM₁₀, PM_2.5, NO₂, and SO₂, but this association was significant only for PM. Over the study period, the counts of hospital admissions due to IHD decreased, which is in line with a decreasing incidence of IHD in Poland.³⁵ The diminished correlation between SO₂ concentrations and hospital admissions for ACS and CCS may be attributed to a substantial reduction in SO₂ levels observed approximately a dozen or so years ago. In recent years, SO₂ concentrations have stabilized at very low levels,³⁶ consistently below the World Health Organization guidelines.³⁷ Therefore, they do not significantly impact health deterioration. However, there are geographic differences in SO₂ concentrations, as noted in the cities we examined, with the highest levels in Kraków and the lowest in the Tricity area.

Our findings are largely in line with previous evidence. The seminal HEAPSS study (Health Effects of Air Pollution among Susceptible Sub‑populations), conducted in several European cities, showed that exposure to PM₁₀, NO₂, and carbon monoxide was associated with a significantly increased risk for same‑day hospital readmission among MI survivors.³⁸ In Bulgaria, hospitalization risk for IHD (I20–I25) increased after the exposure to NO₂ but not to PM₁₀, PM_2.5, and SO₂.³⁹ A nationwide crossover study from China reported that the hospitalization risk due to IHD (I20–I25, about 400 000 cases) increased over 3 days after exposure to PM_2.5 (0.42% per 10 µg/m³), PM₁₀ (0.26% per 10 µg/m³), NO₂ (2.56% per 10 µg/m³), and SO₂ (0.51% per 10 µg/m³).⁴⁰ Likewise, another study from China showed that the risk of death due to acute MI was significantly increased after short‑term exposure to PM₁₀, PM_2.5, NO₂, and SO₂.⁴¹

A study of nearly 140 000 MIs in Japan reported that the risk of hospital admission increased after the exposure to PM_2.5 (odds ratio [OR], 1.06 per 10 µg/m³ over 2 days), but the effect of gaseous pollutants was not investigated.⁴² In Changzhou, China, the exposure to PM_2.5 and PM₁₀, but not to NO₂ and SO₂, was associated with a significantly increased risk of MI.⁴³ However, in another Chinese region, the risk of MI increased significantly after exposure to PM_2.5, PM₁₀, NO₂, and SO₂.⁴⁴ A study in the United States reported that mortality due to chronic IHD was increased in counties with aerosol particle pollution.⁴⁵

A previous study in Poland reported that the exposure to PM_2.5 and PM₁₀ increased the short‑term risk of hospitalization due to cardiovascular diseases (CVDs; I00–I99).⁴⁶ However, the authors did not separate ICD‑10 codes for IHD and did not study gaseous pollutants, such as NO₂ and SO₂. A regional cohort study of 2645 patients in Białystok, Poland, reported that the risk of acute MI (I21) and unstable angina (I20.0) increased after the exposure to NO₂ on the day of exposure (OR, 1.02 per an interquartile‑range increase), but the risk of hospitalization did not change noticeably after exposure to PM₁₀, PM_2.5, and SO₂.⁴⁷ In contrast, a single‑city study from Kraków reported that the risk of acute MI (I21) increased after exposure to PM_2.5 (OR, 1.32 per 10 µg/m³) in all participants, whereas the effect was significant for NO₂ (OR, 1.13, at lags 0 and 1 d) only among patients older than 70 years.⁴⁸ The analysis of nearly 2000 patients with acute MI in the Polish region of Silesia reported that the risk of admission increased significantly on the day of exposure to PM₁₀ (OR, 1.078 per 10 µg/m³) and PM_2.5 (OR, 1.099 per 10 µg/m³) but not after the exposure to SO₂ and NO₂.⁴⁹

There are several potential mechanisms that can explain how air pollutants may exacerbate IHD in the short term.¹⁰ Air pollutants may uncouple the endothelial nitric oxide synthase, leading to the formation of reactive oxygen and nitrogen species. Oxidative stress, in turn, causes vasoconstriction and activates macrophages that release prothrombotic mediators, resulting in thrombosis and platelet activation.^50,51 Further research will likely pinpoint new molecular pathways, which could help elucidate the effects of air pollutants on cardiovascular health.

The limitations of our study need to be acknowledged. As we based it on the National Health Fund database, all analyses were carried out for aggregated daily hospital admissions, which did not allow us to investigate whether the relationship between the exposure to air pollutants and the risk of hospital admission differed in the subgroups of patients (age group, sex, comorbidities). Moreover, we could not identify repeated admissions for the same patients or the patients transferred from other hospitals to reference centers. For the same reason, we could not include admissions for the first time in a given patient. When planning future studies in this area, we will consider a prospective approach, which will allow us to identify individual patients and thus also the medical procedures performed on them. However, this will require a completely different way of obtaining data and significantly increased involvement of the research team.

We also had no influence on possible coding errors. The possibility of misclassification of IHD types in hospital records, as well as the inability to include individual patient data, are significant limitations. Combining the so‑called ACSs (I20, IHD; I21, acute MI), I22 (subsequent MI), I23 (certain current complications following acute MI), I24 (other acute IHD) may raise concerns that during coding for the National Health Fund, the person analyzing the documentation will indicate the better scored ACS procedure as the main diagnosis. However, in our approach to data analysis, we assumed that the physicians coding the procedures (cardiologists, internal medicine specialists) were able to correctly recognize ACS, which required treatments typical of this procedure, for example, coronary angiography, coronary balloon angioplasty, or stent implantation in the coronary vessels. Therefore, we would rather not expect major errors in ACS coding. For chronic IHD (I25), symptoms of decompensation requiring hospital admission may be included in the same group as planned admissions related to the need to optimize chronic treatment or treatment of concomitant circulatory system diseases, such as heart failure, AF, etc. We were not able to clearly determine subgroups in the group of patients qualified as I25, based on the collected data. However, since we used the same method of collecting and analyzing data in each of the considered 3 urban agglomerations, we expect that the risk of a significant error is relatively small. However, to avoid such errors, we intend to conduct a prospective research and group the examined patients according to the medical procedures used during their hospital stay, taking into account all diagnoses that the patient received.

In the case of research similar to ours, based on retrospective data, some errors and shortcomings may also result from the imperfections of available registers, including incorrect initial diagnosis, incomplete data, different methods of data collection, or failure to take into account various confounding variables, such as the impact of comorbidities, infections, risk factors for atherosclerosis, etc. Moreover, neither ours nor other studies constructed this way took into account individual exposure to air pollutants or susceptibility of the studied persons, different chemical composition of the pollutants in individual cities related to differences in emission sources, or the summation of the impact of various other pollutants (including, for example, indoor air pollution, which is not included in the studies of ambient air quality, conducted as part of the State Environmental Monitoring by the Chief Inspectorate of Environmental Protection).

Our study has also some strengths. Most previous studies assessed the relationship between exposure to air pollution and hospitalization risk due to CVDs without dedicated analyses for IHD. Moreover, studies that did assess IHD focused on MI only or aggregated all IHD ICD‑10 codes (I20–I25). Contrary to that, we analyzed ACSs and CCSs separately. Another strength of our study is that the 3 agglomerations we investigated account for nearly 10% of Poland’s population, and the dataset we used is the largest to date in Poland. Moreover, we used a time‑stratified case‑crossover design, which reduced confounding by conditioning data specific to the location and a short time window, that is, the control days were the same days of the week, in the same month, and in the same year as the case days.^29,30 Such a design also allowed us to account for the long‑term downward trend in hospital admissions due to IHD in Poland over the study period. Moreover, we accounted for the effect of meteorological variables, such as temperature, humidity, and precipitation, which may also increase the risk of IHD events.^52-54

As indicated by experts from the World Heart Federation, American College of Cardiology, American Heart Association, and the European Society of Cardiology,⁵⁵ citing the Global Burden of Disease Study,⁵⁶ air pollution can be attributed to 20% of all CVD deaths. Our research indicates that air pollution with PM and nitrogen oxides has a short‑term and negative impact on health by increasing the number of ACSs and CCSs. This fact indicates that air pollution has a greater impact on CVD than an improper diet or lack of physical activity. Therefore, not only a patient‑tailored approach is essential for management of CCSs⁵⁷ but air pollution issues should definitely be included in the national and global strategies for the prevention and treatment of CVD. Experts include the following methods of reducing the negative impact of air pollution on health: 1) informing and educating patients about the harmfulness of air pollution and encouraging them to monitor information about local pollution; 2) avoiding walking and other activities (eg, cycling, rollerblading) along traffic arteries with high traffic volume; 3) planning physical activity in parks and gardens, but not near busy roads and on days when the concentration of pollution increases or conducting physical activity in rooms with an air filtering system on days when air pollution increases; 4) limiting physical activity outdoors, especially by people from risk groups (pregnant women, children, the elderly, and people with chronic circulatory and respiratory diseases); 5) installing ventilation and air filtering systems in apartments, especially in areas with high concentrations of air pollution.

In conclusion, exposure to PM₁₀, PM_2.5, and NO₂ was associated with an increased short‑term risk of hospital admission due to ACSs and CCSs in Poland. Our findings add to the evidence linking air pollution with disease burden in Poland.^33,58,59