Association of Fine Particulate Matter Exposure With Bystander-Witnessed Out-of-Hospital Cardiac Arrest of Cardiac Origin in Japan

This case-control study assesses exposure to ambient air pollution, presenting cardiac arrest rhythms, initial resuscitation efforts, and patient and onset factors in individuals who experienced out-of-hospital cardiac arrests in Japan.


Introduction
Out-of-hospital cardiac arrests (OHCAs) are a major public health concern and a leading cause of death worldwide.More than 350 000 individuals in North America and 275 000 individuals in Europe experience OHCAs annually, and the mean post-cardiac arrest survival rate remains at approximately 10%. 1,2The number of people with OHCA has been increasing in Japan, with the most recent reported figure at 110 000, 3 although the survival rate is comparatively lower than that in Western countries. 4In the United States, the 2010 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care outlined comprehensive clinical decision-making rules for the termination of resuscitative efforts and allowed such management before transporting individuals to a hospital if the prescribed parameters were met. 5 Conversely, in Japan, emergency medical services (EMS) responders cannot make the decision to terminate resuscitation and must continue life-saving efforts until either the return of spontaneous circulation or arrival at the hospital. 4Thus, most people who experience OHCA in Japan and are initially treated by EMS personnel are transported to a hospital and are therefore subsequently registered in the All-Japan Utstein Registry.Despite either country's approach, survival rates have not improved, partially because of the development of nonshockable rather than shockable rhythms as the initial cardiac arrest rhythm. 3,4,6posure to ambient air pollution increases morbidity and mortality and has been recognized as a leading contributor to the global disease burden. 7Air pollution exacerbates existing heart conditions and plays a role in disease pathogenesis, with evidence of adverse effects being stronger for particulate matter than for gaseous pollutants. 8Particulate matter with a diameter of 2.5 μm or smaller (PM 2.5 ) is composed of elemental carbon, transition metals, complex organic molecules, sulfate, and nitrate. 8It can permeate lung alveoli, can enter the bloodstream, and is further absorbed by phagocytes on lung surfaces.Alveolar epithelial cells, in turn, generate oxygen radicals that may trigger inflammatory responses. 8This additive inflammatory effect of particulate pollutants may be associated with aggravated existing inflammatory lung diseases and the progression, destabilization, or rupture of atherosclerotic plaques, precipitating acute coronary syndrome. 8Exposure to PM 2.5 is associated with not only increased hospitalization but also a higher risk of death from respiratory and cardiovascular causes. 9th nationwide and multicountry studies have demonstrated that an increase in PM 2.5 exposure is associated with mortality. 10A recent systematic review and meta-analysis found that short-term exposure to PM 2.5 was associated with elevated OHCA risk. 11However, all of the studies included in this meta-analysis were restricted to single cities and involved relatively small numbers of patients who had OHCA. 11 conducted a nationwide case-crossover analysis to examine the association between shortterm exposure to PM 2.5 and bystander-witnessed OHCAs of cardiac origin using the All-Japan Utstein Registry, which contains information on all patients with OHCA.In addition, we investigated the differences in the distribution of initial cardiac arrest rhythms in OHCA among patients with exposure to PM 2.5 .

Data Source, Study Area, and EMS System
The All-Japan Utstein Registry is a prospective, nationwide, population-based database for OHCA with Utstein-style data collection. 12It was established on January 1, 2005; has been maintained by the Fire and Disaster Management Agency; and has been described in detail elsewhere. 6All fire stations with dispatch centers and collaborating medical institutions contribute to the registry.Data from this registry were provided to the Subcommittee on Resuscitation Science of the Japanese Circulation Society in accordance with governmental legal procedures.This study received approval from the Ethics Committee of Kawasaki Medical School, which waived the requirement for patient written informed consent because only deidentified data were used.We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
Japan has an area of approximately 378 000 km 2 , including both urban and rural communities across 47 prefectures from Hokkaido to Okinawa (eTable 1 and eFigure 1 in the Supplement). 13In 2018, the total population was approximately 126 million, 14 and the 728 municipally governed fire stations with dispatch centers throughout the country followed uniform, guideline-based resuscitation protocols. 15Japanese EMS personnel are not authorized to terminate resuscitation efforts; most patients who experience an OHCA are transported to the nearest hospital, and their information is entered into the registry.These cases were included in this present study; cases of decapitation, incineration, decomposition, rigor mortis, and dependent cyanosis were excluded. 3,6

Study Population
The OHCA cases registered between January 1, 2005, and December 31, 2016, were included in the study.These patients were those who experienced OHCAs that were witnessed by bystanders and for whom EMS responders initiated resuscitation before hospital transfer.Patients whose cardiac arrest occurred after the arrival of EMS responders, who had unwitnessed OHCA, or who had unidentified witness status were excluded.Those whose cardiac arrest occurred during periods when standardized PM 2.5 data were unavailable (eg, on national holidays) and those with missing PM 2.5 data were also excluded from the analysis.
Cardiac arrest has been defined as the end of cardiac mechanical activity as indicated by the absence of signs of circulation. 12An arrest is presumed to be cardiac in origin unless evidence suggests it is from an external cause (eg, trauma, hanging, drowning, drug overdose, or asphyxiation), a respiratory disease, a cerebrovascular disease, or a malignant tumor, among others.
The physicians in charge (mainly emergency department physicians) who interacted with the EMS personnel are responsible for ascertaining the cause of arrest. 3,6

Data Collection and Quality Control
Data were collected prospectively according to the Utstein templates for resuscitation registries. 12e All-Japan Utstein Registry includes only the name of the prefecture as the place of onset to maintain personal information security.All event times were synchronized with the dispatch center clock. 3,6The times of collapse and administration of public access automated external defibrillators were identified through an interview conducted by the EMS personnel with the bystander witness before the EMS personnel left the scene. 3,6In cases in which the bystander initiated cardiopulmonary resuscitation, chest compression alone and conventional cardiopulmonary resuscitation with rescue breathing were recorded as presence of bystander resuscitation.The initial cardiac arrest rhythm was classified as shockable (ventricular fibrillation and pulseless ventricular tachycardia) or nonshockable (pulseless electric activity and asystole) based on the electrocardiographic data recorded by the automated external defibrillator. 3,6Data forms were completed by EMS personnel in cooperation with the patient's physician in charge, and the information was subsequently entered into the All-Japan Utstein Registry.Forms were logically checked by the computerized system and were verified by the Implementation Working Group for the All-Japan Utstein Registry.If a data form was incomplete, the Fire and Disaster Management Agency returned it to the respective fire station for completion. 3,6

Environmental Data
Hourly measurements of PM 2.5 concentrations were obtained using automated equipment and standard reference methods from the atmospheric environment database of the National Institute for Environmental Studies in Japan.Although each prefecture had 1 or more ambient air pollutionmonitoring station, we incorporated environmental data from only 47 stations that were each located in a distinct prefectural capital, and the measured levels were considered representative of the air quality of that region.We subsequently correlated the PM 2.5 data as applicable to each patient based on the place (prefecture) of OHCA onset.We checked that the PM 2.5 concentrations measured at the applicable station correlated with those assessed at other stations in the same prefecture (mean correlation coefficient, 0.9).We collected hourly measurements of PM 2.5 across a 24-hour period, and we calculated the daily mean concentration values, excluding the days in which more than 4 such values were missing.The median (interquartile range) rate of such missed days in the PM 2.5 data was 1.5% (0.8%-2.3%).
In addition, we measured the daily mean concentration levels of other pollutants, including nitrogen dioxide and sulfur dioxide.Maximum concentrations of ozone were recorded over 8 hourly periods.Data published by the Japan Meteorological Agency were used to evaluate both the daily mean ambient temperature and relative humidity levels.Incidence of influenza was referenced from the database of the National Institute of Infectious Diseases in Japan.Periods of influenza epidemic were defined as weeks in which the number of recorded cases were greater than the 90th percentile of the distribution during the study period.

Statistical Analysis
The case-crossover design (a specific type of case-control study) was used to examine the association between short-term exposure to PM 2.5 and OHCAs of cardiac origin.In a case-crossover study, time-invariant factors, such as age and sex, do not act as confounders because patients are compared with themselves at different periods; that is, a patient's PM 2.5 exposure on the case day was compared with the exposure on control days.Although the case day was defined as the day of the OHCA occurrence, control days were selected using a time-stratified method 16 ; that is, 3 or 4 control days were chosen from the same day of the week, month, and year as the case day.If a patient experienced a cardiac arrest on June 16, 2016, then June 2, 9, 23, and 30 of that year would be assigned as control days.
This time-stratified approach to the referent selection allowed for unbiased estimations using a conditional logistic regression analysis. 16We first applied a prefecture-specific, conditional logistic regression model to estimate the odds ratios (ORs) with 95% CIs for every 10-μg/m 3 increase in PM 2.5 concentrations at lag0-1 (difference in mean PM 2.5 concentrations measured on the case day and 1 day before).The choice of lag0-1 was made before commencing the study and was based on previous evidence. 11,17,18Time-variant factors, including ambient temperature at lag0-1 (a 5-knot natural cubic spline), relative humidity at lag0-1 (a 3-knot cubic spline), and incidence of influenza epidemic, were included in the model.Random-effects meta-analysis was used to obtain prefecturespecific pooled estimates of the associations between PM 2.5 and OHCA.Results were presented as a percentage increase in OHCA incidence, which was calculated as [(OR -1) × 100].Heterogeneity was verified using the I 2 statistic.We repeated this analysis for each type of initial cardiac arrest rhythm (shockable or nonshockable).Furthermore, we constructed a multipollutant model that was adjusted for ozone, nitrogen dioxide, and sulfur dioxide concentrations at lag0-1 to examine the potential confounding effects of different pollutants.We also stratified analyses by age (<75 or Ն75 years) and sex.We performed stratified analyses by season of onset (warm [May to October] or cold [November to April]) and region (East, Central, or West). 10We obtained the initial cardiac arrest rhythm of electrocardiogram

Patient Characteristics
A total of 1 423 338 OHCAs were documented in Japan from January The mean (SD) daily PM 2.5 concentration was 13.9 (7.9) μg/m 3 according to the nationwide analysis (eTable 2 in the Supplement).Regionally, the mean (SD) PM 2.5 concentrations were 12.5 (7.4)   μg/m 3 in East Japan, 13.7 (7.8) μg/m 3 in Central Japan, and 16.3 (8.7) μg/m 3 in West Japan (eFigure 1 in the Supplement).Prefecture-specific results for environmental factors are presented in eTable 1 in the Supplement.Pearson correlation coefficients for PM 2.5 concentrations were calculated as r = 0.42 for ozone, r = 0.31 for nitrogen dioxide, and r = 0.44 for sulfur dioxide (eTable 3 in the Supplement).
a Odds ratio for percent increase for every 10-μg/m 3 increase in PM 2.5 at lag0-1.
b Total number of regions was 46, with Okinawa excluded from the West Japan region because of geographic differences (eTable 1 and eFigure 1 in the Supplement).a Indicates ventricular fibrillation and pulseless ventricular tachycardia.
b Indicates pulseless electrical activity and asystole.
c See eTable 1 and eFigure 1 in the Supplement.
We analyzed the respective percentage increases for the association of PM 2.5 exposure with OHCA according to the distribution of presenting cardiac arrest rhythm.Although the initial shockable rhythm (% increase, 0.6; 95% CI, -2.0% to 3.2%) was not associated with the PM 2.5 exposure, the nonshockable rhythm showed statistically significant association (% increase, 1.4; 95% CI, 0.1%-2.7%).In the stratified analyses, the initial nonshockable rhythm in Central Japan (% increase, 3.2; 95% CI, 0.8%-5.8%)and presence of bystander resuscitation (% increase, 2.1; 95% CI, 0.3%-4.0%)were associated with PM 2.5 exposure.However, we did not observe a statistically significant difference in the development of nonshockable rhythm among bystander-witnessed OHCAs of cardiac origin according to age, sex, season of onset, time from collapse to initial ECG, region, or bystander resuscitation (Figure 3).

Discussion
Findings of this study suggest that short-term exposure to particulate pollutants is associated with bystander-witnessed OHCAs of cardiac origin in Japan.Increased OHCA incidence was associated with the mean increase in PM 2.5 concentrations greater than those levels observed on the day before the cardiac arrest.Exposure to PM 2.5 was also associated with nonshockable rhythm as the initial, presenting OHCA rhythm.
Air pollution has been reported to play a role in the development of cardiovascular diseases. 8evious studies have observed an association between elevated pollutant concentration and risk of myocardial infarction 19 and an approximately 1.0% worldwide increase in the mean all-cause mortality for every incremental 10-μg/m 3 short-term exposure to PM 2.5 . 8,10,20 a meta-analysis of 12 studies, the short-term exposure to PM 2.5 was associated with an increased risk of OHCA; however, those studies were performed mainly in a single city. 11Results of a study of more than 20 000 Asian individuals with OHCA corroborate our finding that an increase in Ambient temperature at lag0-1, relative humidity at lag0-1, and incidence of influenza epidemic were included in the model.ECG indicates electrocardiogram; lag0-1 indicates difference in mean PM 2.5 concentrations measured on the case day and 1 day before.
a Odds ratio for percent increase for every 10-μg/m 3 increase in PM 2.5 at lag0-1.
b Total number of regions was 46, with Okinawa excluded from the West Japan region because of geographic differences (eTable 1 and eFigure 1 in the Supplement).
OHCA incidence and risk may be associated with PM 2.5 exposure. 21Nonetheless, the present study has advantages over other studies owing to the larger sample size (nationwide data of more than 100 000 bystander-witnessed OHCAs of cardiac origin were analyzed) and the random-effects meta-analysis used to obtain prefecture-specific pooled estimates of the association between PM 2.5 and OHCA, which may yield higher accuracy.Furthermore, we studied the implications of novel factors, including seasonal and regional differences, for the association between PM 2.5 exposure and OHCAs, along with the types of presenting cardiac arrest rhythms.
As in past studies, we found that older individuals (aged Ն75 years) and men experienced more OHCAs associated with PM 2.5 exposure, and this association may reflect the frequency of occurrence of cardiovascular disease in these populations. 17,21A significant association between PM 2.5 and OHCA of cardiac origin was observed during warm seasons and in Central Japan, which may indicate seasonal and regional variations in the PM 2.5 chemical compositions.Nitrate concentrations have been demonstrated to vary by season 22 and to be low in the summer.Conversely, although sulfate concentrations are high in the summer, 22 levels in Central Japan are lower than those in West Japan, which is susceptible to the effects of long-range transport of air pollutants. 23However, seasonal and regional differences in nitrate and sulfate concentrations did not correlate with the association between PM 2.5 and OHCA of cardiac origin across the applicable seasons and regions.Little evidence is available on the implications of PM 2.5 composition for health in Japan.Therefore, further research is required to assess the seasonal and regional differences of particulate air pollutants and their implications for health.
Interest in the association between air pollution and cardiovascular disease occurrence has been growing, most notably interest in the association between PM 2.5 and cardiac arrest. 85][26] However, a few studies that described an association between asystole or a nonshockable initial OHCA rhythm with an increase in PM 2.5 exposure have received less attention. 21,27In the present study, the occurrence of an initial nonshockable cardiac rhythm was associated with an increase in PM 2.5 concentration.
Although the pathophysiological details of this effect have not been elucidated, individuals with traditional risk factors are likely to be at a higher risk for cardiovascular events because of exposure to air pollutants. 19,20Bystander-witnessed OHCAs of cardiac origin occurring secondary to heterogeneous underlying diseases may be accompanied by higher risks after PM 2.5 exposure, including susceptibility to presenting with nonshockable rhythm.Older age, female sex, absence of bystander resuscitation, and a longer time from collapse to initial ECG interval may be associated with the development of initial nonshockable rhythm. 28This association was also observed in comparisons of the occurrence of shockable rhythms in this study.However, the association between PM 2.5 exposure and the development of nonshockable rhythm may not be modifiable by these factors because of the lack of adequate study power.Therefore, it would be premature to explain why we observed the association between increased PM 2.5 concentration and a presentation with a nonshockable rhythm.Nevertheless, the All-Japan Utstein Registry data have demonstrated the current nationwide status of OHCAs in Japan because almost all Japanese patients who experienced a cardiac arrest were registered. 3,4,6

Limitations
This study has several limitations.First, similar to past epidemiologic environmental studies of the health effects of air pollution, this study demonstrated that the occurrence of exposure misclassification was inevitable given that the results were based on pollutant data measured at a single monitoring station in each prefecture.It was difficult to estimate the biasing effect of a combination of the Berkson error and classical error on our estimates of the association between PM 2.5 and OHCA. 29Second, as in all epidemiologic studies, the data integrity, validity, and ascertainment bias in this study were potential limitations.However, the uniform data collection,

JAMA Network Open | Global Health
Association of Fine Particulate Matter Exposure With Out-of-Hospital Cardiac Arrest in Japan

Figure 2 .
Figure 2. Association Between Exposure to Particulate Matter With Diameter of 2.5-μm or Less (PM 2.5 ) and Total Bystander-Witnessed Out-of-Hospital Cardiac Arrest (OHCAs) of Cardiac Origin

Figure 3 .
Figure 3. Stratified Analyses Comparing Exposure to Particulate Matter With Diameter of 2.5 μm or Less (PM 2.5 ) and Bystander-Witnessed Out-of-Hospital Cardiac Arrest (OHCAs) of Cardiac Origin With an Initial Nonshockable Rhythm Association of Fine Particulate Matter Exposure With Out-of-Hospital Cardiac Arrest in Japan (ECG), presence of bystander resuscitation, and time from collapse to initial ECG for the association between PM 2.5 and OHCA to explore the possibility of differences owing to these factors.All analyses were performed from May 7, 2019, to January 23, 2020, using Stata, version 15 (StataCorp LLC).Analyses used 2-tailed, paired testing.Findings were considered statistically significant at P < .05.

Table 1 .
Characteristics of Bystander-Witnessed Out-of-Hospital Cardiac Arrest of Cardiac Origin in 47 Prefectures in Japan, April 2011 to December 2016 a See eTable 1 and eFigure 1 in the Supplement.bIndicates ventricular fibrillation and pulseless ventricular tachycardia.cIndicates pulseless electrical activity and asystole.PM 2.

Table 2 .
Characteristics of Bystander-Witnessed OHCA of Cardiac Origin by Initial Cardiac Arrest Rhythms Abbreviation: ECG, electrocardiogram.