All-Cause and Cause-Specific Mortality Associated with Long-Term Exposure to Fine Particulate Matter in Japan: The Ibaraki Prefectural Health Study

Abstract

Aims: Long-term exposure to fine particulate matter (PM_2.5) is causally associated with mortality and cardiovascular disease. However, in terms of cardiovascular cause-specific outcomes, there are fewer studies about stroke than about coronary heart disease, particularly in Asia. Furthermore, there remains uncertainty regarding the PM_2.5-respiratory disease association. We examined whether long-term exposure to PM_2.5 is associated with all-cause, cardiovascular and respiratory disease mortality in Japan.

Methods: We used data of 46,974 participants (19,707 men; 27,267 women), who were enrolled in 2009 and followed up until 2019, in a community-based prospective cohort study (the second cohort of the Ibaraki Prefectural Health Study). We estimated PM_2.5 concentrations using the inverse distance weighing methods based on ambient air monitoring data, and assigned each participant to administrative area level concentrations. A Cox proportional hazard model was applied to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) of mortality.

Results: During the average follow-up of 10 years, we confirmed 2,789 all-cause deaths. All outcomes including stroke mortality did not significantly increase as the PM_2.5 concentration increased. For non-malignant respiratory disease mortality, the multivariable adjusted HR per 1 µg/m³ increase in the PM_2.5 concentration was 1.09 (95% CI = 0.97–1.23).

Conclusions: In this population exposed to PM_2.5 at concentrations of 8.3–13.1 µg/m³, there was no evidence that long-term exposure to PM_2.5 had adverse effects on mortality. Weak evidence of positive association observed for non-malignant respiratory disease mortality needs further studies in other populations.

Introduction

There is mounting evidence that long-term exposure to fine particulate matter (PM_2.5, particles with an aerodynamic diameter of 2.5 µm or less) leads to mortality and morbidity due to cardiovascular disease, but a few in Japan¹⁾. Additionally, fewer studies examined the association between long-term PM_2.5 exposure and stroke than reported an association with coronary heart disease (or myocardial infarction), particularly in Asia²⁾. In Japan, the incidence of stroke is higher than that of coronary heart disease³⁾; thus, stroke prevention is prioritised in primary prevention of cardiovascular disease. Previously, we used cohort data obtained by following 91,808 community-dwelling adults from 1993 to 2010 and reported that concentrations of suspended particulate matter (particles with an aerodynamic diameter of approximately 7 µm or less, median exposure level = 32.7 µg/m³) showed a positive association with both coronary heart disease and stroke mortality⁴⁾. The PM_2.5-monitoring network was developed after the establishment of air quality standards in 2009; therefore, we could not investigate the cardiovascular effects of long-term exposure to PM_2.5 in this previous study.

In addition, there remains uncertainty regarding findings about an association between long-term exposure to PM_2.5 and respiratory disease, and evidence of an association after adjustment for co-pollutants is particularly limited; therefore, the U.S. Environmental Protection Agency refrained from concluding a causal association⁵⁾. According to a review that contributed to the WHO Global Air Quality Guidelines in 2021, long-term exposure to PM_2.5 was associated with a risk of respiratory death in a random-effect meta-analysis, but an 80% prediction interval, which is interpreted as the range of the effect size in a future study⁶⁾, included unity⁷⁾. Furthermore, two recently published Japanese studies reported different results. One study, the Japan Public Health Centre-based Prospective Study (JPHC Study), which followed 87,385 residents from 11 areas between 1990 and 2013, reported no association between long-term exposure to PM_2.5 and respiratory disease mortality⁸⁾. Another study, which followed 76,591 participants who resided in one city between 2006 and 2016, reported that PM_2.5 exposure was positively associated with respiratory disease mortality⁹⁾. Therefore, uncertainty about the association between PM_2.5 exposure with respiratory disease cannot be assuaged without obtaining further quality-assured epidemiological evidence that is controlled for chance, bias, and confounding.

The aim of the present study was to obtain additional evidence on whether long-term exposure to PM_2.5 is associated with all-cause, cardiovascular, and respiratory disease mortality in an Asian country. Following on from our previous study⁴⁾, we used cohort data of participants within limited administrative areas. In Japan, stroke mortality is high in north-east areas and PM_2.5 concentrations are high in south-west areas¹⁰⁾. By limiting administrative areas, we attempted to control the inverse correlation between particulate matter concentrations and stroke mortality/morbidity observed from nationwide data and to minimise the impact of background characteristics that are difficult to individually assess, such as traditional dietary habits (salty preference) and climate, which are strongly related to stroke in Japan.

Methods

Study Population

The Ibaraki Prefectural Health Study (IPHS) is a community-based prospective cohort study launched by the Ibaraki Prefectural Government. Ibaraki Prefecture is located in the northeast of the Kanto region including the Tokyo Metropolitan Area, which is almost at the centre of the Japanese islands (Fig.1). This prefecture has Japan’s second largest lake, Kasumigaura, and a population of about 30 million (population density = 487 people/km², 12^th largest per 47 prefectures in terms of population density) according to the 2010 National Census¹¹⁾. The southern part of the prefecture within commuting distance of Tokyo is more urbanised than the northern area and tends to have higher PM_2.5 concentrations¹²⁾.

Fig.1.

Location of the study area

The first cohort of the IPHS was initiated in 1993 ¹³⁾, and the second cohort was started in 2009 with new participants¹⁴⁾. In this study, we used data from the second cohort, which began in the same year that the air quality standard for PM_2.5 was established. Prior to 2009, ground-level measurement data for PM_2.5 in Japan were very limited. The second cohort included 53,339 residents (National Health Insurance subscribers) aged 40–74 years who resided in the 21 local administrative areas of Ibaraki Prefecture and participated in lifestyle health check-ups, including physical measurement, fasting blood and spot urine tests, and a questionnaire survey. We excluded 6,365 residents who reported a history of heart disease and stroke and/or kidney disease. Finally, 46,974 participants (19,707 men and 27,267 women) were enrolled in the analysis and followed up until 31^st December 2019 (pre-coronavirus pandemic). The protocol for the second cohort of the IPHS was approved by the ethics committees of Ibaraki Prefecture (R2-2) and the University of Tsukuba (1628-4), and written informed consent was obtained from all residents before participation.

Baseline Survey

In the baseline survey, blood samples with or without fasting were collected, and triglyceride, glucose, haemoglobin A_1c, and alanine aminotransferase were measured. A protein dipstick test was performed using spot urine. Blood pressure was measured using an automatic sphygmomanometer in the right arm of seated participants after resting for at least 5 minutes. Body mass index (BMI) was calculated as weight in light clothing (kg) divided by the square of height in stockinged feet (m²). A self-administered questionnaire was used to collect information on smoking status, alcohol drinking, use of antihypertensive, antihyperlipidaemic, and antihyperglycaemic drugs, leisure-time physical activities, and educational background. Hypertension was defined as use of antihypertensive drugs, systolic blood pressure ≥ 140 mmHg, and/or diastolic blood pressure ≥ 90 mmHg. Hyperlipidaemia was defined as use of antihyperlipidaemic drugs, fasting triglyceride concentration ≥ 1.7 mmol/L¹⁵⁾, and/or high-density lipoprotein cholesterol concentration ≥ 1.0 mmol/L¹⁵⁾. Diabetes was defined as use of antihyperglycaemic drugs, fasting blood glucose concentration ≥ 7.0 mmol/L, and/or haemoglobin A_1c level (National Glycohaemoglobin Standardization Program) ≥ 6.5%.

Environmental Data

We obtained measurement data of PM_2.5 at ambient air monitoring stations from the Environmental Observatory Air Pollution Continuous Monitoring Data provided by National Institute for Environmental Studies. We then estimated background PM_2.5 concentrations using inverse distance weighing methods based on neighbourhood ambient air monitoring data (within a 20 km radius from the centre point of a 1 × 1 km grid surface) for each 1 × 1 km grid surface. This approach is similar to the method used in our previous study to measure suspended particulate matter concentrations as an index of particulate matter⁴⁾. PM_2.5 exposure measurement used in this study is not time-variant. Specifically, annual mean concentrations of PM_2.5 in 2016 and 2017 (which is the first two years since PM_2.5 monitoring network started in Ibaraki) was used to represent long-term exposure to PM_2.5 during this study period (2009-2019). This is because the PM_2.5 measurement network in Ibaraki Prefecture was established in 2016. Although PM_2.5 concentrations gradually decreased each year during the study period, low and high trends in the north and south, respectively, were maintained. Co-pollutant concentrations, including O₃ and NO₂, were estimated in the same manner as PM_2.5, and used as possible confounding factors. Although photochemical oxidant, a mixture of O₃ and other secondary oxidants generated by photochemical reactions, is measured in Japan, we described it as O₃, its major component.

We only knew in which administrative area (city, town, or village) each participant resided and did not collect information on their residential address, such as postal code, due to ethical reasons. Therefore, each participant was assigned to the arithmetic average pollutant concentrations in all grids within the administrative area. We also estimated population-weighted average pollutant concentrations of all grids, but the values did not differ from those based on arithmetic average. Data on the ambient temperature in each administrative area were obtained from the Japan Meteorological Agency, and the 2-year (2016 and 2017) mean temperature was also assigned to the participants.

Follow-Up

We performed a detailed review of death certificates and confirmed deaths in this cohort. Information on the date of death or relocation were collected from local governments. We applied to the Ministry of Health, Labour and Welfare for use of data on underlying causes of death and were provided these data by the government. The endpoints were deaths from all-causes (codes A00-R99 in the International Classification of Diseases and Related Health Problems (ICD)-10^th revision)_, cardiovascular disease (I00-99), non-malignant respiratory disease (J00-99), and lung cancer (C34). Deaths from cardiovascular disease were subdivided into those from coronary heart disease (I20-25) and stroke (I60-69). Deaths from respiratory disease were subdivided into those from respiratory tract infections and pneumonia (J1-22) and chronic obstructive pulmonary disease (COPD) and others (J23-99).

Statistical Analysis

The number of person-years during the follow-up period was calculated for each participant from the date of the baseline survey to the date of death, emigration from Ibaraki Prefecture, or the end of the follow-up period (31^st December 2019), whichever occurred first. We applied a Cox proportional hazard model to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) of all-cause and cause-specific mortality according to the quartiles of PM_2.5 concentrations, using the first quartile as the reference. Assuming a linear association similar to earlier studies¹⁾, we also estimated the HRs of mortality per 1 µg/m³ increase in the PM_2.5 concentration. We checked the proportional hazard assumption using scaled Schoenfeld residuals and confirmed there was no violation of proportionality. We treated three medical administrative areas (Mito, Tsuchiura-Tsukuba, and others) as strata to allow for a different baseline hazard in each area. For all-cause and cardiovascular disease mortality, we adjusted for factors closely associated with cardiovascular disease, including age, sex, smoking status (no, ＜20, and ≥ 20 cigarettes/day), alcohol drinking (no and yes), BMI (＜18.5, 18.5–24.9, 25.0–29.9, and ≥ 30.0 kg/m²), hypertension (no and yes), hyperlipidaemia (no and yes), diabetes (no and yes), alanine aminotransferase concentration (＜30 and ≥ 30 U/L), and proteinuria (no and yes). For non-malignant respiratory disease and lung cancer mortality, we adjusted for respiratory-related factors, including age, sex, smoking status, alcohol drinking, BMI, and diabetes. In sensitivity analysis, we additionally adjusted for leisure-time physical activities, educational background, the O₃ concentration, the NO₂ concentration, and temperature. To exclude the possibility of reverse causality, we estimated HRs after excluding mortality cases within 3 years from baseline. We defined exposures as average PM_2.5 concentrations between 2016 and 2017; therefore, we also analysed mortality cases after 2016 only to ensure the temporal association. We further conducted stratified analysis to explore the vulnerable population to PM_2.5 exposure. Effect modification for the stratified factors was tested using cross-product terms of exposure categories and these factors. All analyses were performed using STATA 15 (Stata Corporation, College Station, TX, USA).

Results

The overall mean 2-year concentrations of PM_2.5 was 10.8 (range: 8.3–13.1) µg/m³ (Table 1). Pearson’s correlation coefficient between PM_2.5 and O₃ concentrations was -0.11, and that between PM_2.5 and NO₂ concentrations was 0.80. The participants’ characteristics at baseline did not substantially differ among the PM_2.5 concentration quartiles (Table 2). The proportion of participants with junior high-school education was slightly higher among those with higher exposure to PM_2.5. However, data on educational background were missing for approximately one-fifth of participants. During the 470,178 person-years of follow-up (average follow-up = 10.0 years), we confirmed 2,789 all-cause deaths, including 521 deaths from cardiovascular disease (125 deaths from coronary heart disease and 198 deaths from stroke), 282 deaths from non-malignant respiratory diseases, and 319 deaths from lung cancer.

Table 1.Average levels of environmental factors in 2016–2017 and Pearson’s correlation coefficient

	Mean (SD)	Median	Range	Pearson’s correlation coefficient
				PM2.5	O3	NO2	Temperature
PM2.5 (μg/m3)	10.8 (1.4)	11.1	8.3–13.1	1
O3 (ppb)	32.3 (1.2)	32.3	31.0–36.0	-0.11	1
NO2 (ppb)	8.7 (2.0)	8.3	5.7–13.1	0.80	-0.07	1
Temperature (℃)	14.4 (0.5)	14.5	12.8–15.4	0.47	0.06	0.56	1

Abbreviations: IQR, interquartile range; NO₂, nitrogen dioxide; O₃, ozone; PM_2.5, fine particulate matter; SD, standard deviation.

Table 2.Baseline characteristics of participants in 2009 according to the PM2.5 concentration quartile

	(PM2.5 concentrations) Total no. of participants a	Quartile1 (8.3–9.6 μg/m3)		Quartile2 (9.7–10.9 μg/m3)		Quartile3 (11.0–12.1 μg/m3)		Quartile4 (12.2–13.1 μg/m3)
		n	(%)	n	(%)	n	(%)	n	(%)
No. of participants	46,974	9,697		13,176		11,395		12,706
Age (years)
40–49	4,511	755	7.8	1,608	12.2	928	8.1	1,220	9.6
50–59	9,213	1,555	16.0	2,990	22.7	1,943	17.1	2,725	21.5
60–69	23,975	5,358	55.3	6,120	46.5	6,259	54.9	6,238	49.1
70–74	9,275	2,029	20.9	2,458	18.7	2,265	19.9	2,523	19.9
Sex
Men	19,707	4,023	41.5	5,589	42.4	4,662	40.9	5,433	42.8
Women	27,267	5,674	58.5	7,587	57.6	6,733	59.1	7,273	57.2
Smoking status
No	39,682	8,265	85.2	11,045	83.8	9,846	86.4	10,526	82.8
Smoking ＜20 cigarettes/day	5,724	1,168	12.0	1,614	12.3	1,239	10.9	1,703	13.4
Smoking ≥ 20 cigarettes/day	1,568	264	2.7	517	3.9	310	2.7	477	3.8
Alcohol drinking
No	25,996	5,454	56.3	7,367	55.9	6,201	54.4	6,974	54.9
Yes	20,976	4,242	43.8	5,809	44.1	5,193	45.6	5,732	45.1
Body mass index (kg/m2)
＜18.5	2,141	421	4.3	572	4.3	612	5.4	536	4.2
18.5–24.9	31,831	6,471	66.7	8,776	66.6	8,029	70.5	8,555	67.3
25.0–29.9	11,618	2,501	25.8	3,410	25.9	2,484	21.8	3,223	25.4
≥ 30.0	1,384	304	3.1	418	3.2	270	2.4	392	3.1
Hypertension
No	33,878	7,052	72.7	9,828	74.6	8,141	71.4	8,857	69.7
Yesb	13,096	2,645	27.3	3,348	25.4	3,254	28.6	3,849	30.3
Systolic blood pressure (mmHg)c	46,974	127 (17)		126 (17)		127 (17)		128 (17)
Diastolic blood pressure (mmHg) c	46,974	77 (11)		76 (11)		77 (11)		77 (11)
Hyperlipidaemia
No	30,078	6,048	62.4	8,408	63.8	7,735	67.9	7,887	62.1
Yesd	16,896	3,649	37.6	4,768	36.2	3,660	32.1	4,819	37.9
Low-density lipoprotein cholesterol (mmol/L) c	46,974	3.2 (0.8)		3.1 (0.8)		3.2 (0.8)		3.2 (0.8)
High-density lipoprotein cholesterol (mmol/L) c	46,974	1.6 (0.4)		1.5 (0.4)		1.6 (0.4)		1.5 (0.4)
Trigriceride (mmol/L) c	46,974	1.4 (0.9)		1.5 (0.9)		1.3 (0.8)		1.5 (1.0)
Diabetes mellitus
No	42,660	8,764	90.4	11,970	90.9	10,373	91.0	11,553	90.9
Yese	4,314	933	9.6	1,206	9.2	1,022	9.0	1,153	9.1
Alanine aminotransferase (U/L)
＜30	41,223	8,449	87.1	11,530	87.5	10,029	88.0	11,215	88.3
≥ 30	5,751	1,248	12.9	1,646	12.5	1,366	12.0	1,491	11.7
Proteinuria
No	45,876	9,462	97.6	12,821	97.4	11,227	98.6	12,366	97.5
Yes	1,056	230	2.4	341	2.6	161	1.4	324	2.5
Leisure-time physical activities
Seldom or never	17,233	3,202	36.2	5,762	48.1	3,510	32.4	4,759	40.8
1-hour or more per week	26,120	5,653	63.8	6,227	51.9	7,324	67.6	6,916	59.2
Educational background
Junior high school	7,318	1,570	18.2	2,319	19.9	1,395	13.2	2,034	29.6
High school	20,975	4,893	56.7	6,821	58.7	5,773	54.6	3,488	50.8
University or higher	9,409	2,164	25.1	2,487	21.4	3,415	32.3	1,343	19.6

^a Subgroup totals do not equal the total number of participants due to missing data.

^b Defined as use of antihypertensive drugs, systolic blood pressure ≥ 140 mmHg, and/or diastolic blood pressure ≥ 90 mmHg.

^c Mean (standard deviation)

^d Defined as use of antihyperlipidaemic drugs, fasting triglyceride concentration ≥ 1.7 mmol/L, and/or high-density lipoprotein cholesterol concentration ≥ 1.0 mmol/L.

^e Defined as use of antihyperglycaemic drugs, fasting blood glucose ≥ 7.0 mmol/L, and/or haemoglobin A_1c level ≥ 6.5%.

Table 3 shows the association between long-term exposure to PM_2.5 and all-cause and cause-specific mortality. PM_2.5 exposure did not increase the risk of all-cause mortality (multivariable adjusted HR per 1 µg/m³ increase in the PM_2.5 concentration = 0.96, 95% CI = 0.92–0.99). When we constructed statistical model based on the PM_2.5 concentration quartiles, no certain direction in point estimates of HRs for cardiovascular and lung cancer mortality was presented. In terms of cardiovascular disease mortality, HRs for coronary heart disease and stroke mortality were not elevated.

Table 3.Hazard ratios (HRs) and 95% confidence intervals (CIs) of the associations between exposure to PM2.5 and total and cause-specific mortality in the second cohort of the Ibaraki Prefectural Health Study

(PM2.5 concentration)	Quartile 1 (8.3–9.6 μg/m3)	Quartile 2 (9.7–10.9 μg/m3)	Quartile 3 (11.0–12.1 μg/m3)	Quartile 4 (12.2–13.1 μg/m3)	HR (95% CI) per 1 μg/m3 increase
No. of participants	9,697	13,176	11,395	12,706
Person-years	96,257	131,790	114,205	127,926
Natural death (ICD-10, A-R code)
No. of outcomes	657	787	607	738
Age- and sex-adjusted HR (95% CI)	Reference	0.98 (0.87–1.09)	0.80 (0.71–0.91)	0.90 (0.80–1.02)	0.96 (0.93–0.99)
Multivariable adjusted HR (95% CI) a, b	Reference	0.95 (0.85–1.07)	0.80 (0.70–0.91)	0.87 (0.77–0.99)	0.96 (0.92–0.99)
Cardiovascular disease (I code)
No. of outcomes	120	159	103	139
Age- and sex-adjusted HR (95% CI)	Reference	1.05 (0.81–1.36)	0.72 (0.53–0.98)	0.91 (0.68–1.22)	0.95 (0.87–1.02)
Multivariable adjusted HR (95% CI) a, b	Reference	1.01 (0.78–1.31)	0.72 (0.53–0.98)	0.85 (0.64–1.14)	0.94 (0.87–1.02)
Coronary heart disease(I20-25)
No. of outcomes	28	42	29	26
Age- and sex-adjusted HR (95% CI)	Reference	1.17 (0.70–1.97)	0.98 (0.53–1.82)	0.83 (0.55–1.55)	0.95 (0.80–1.13)
Multivariable adjusted HR (95% CI) a, b	Reference	1.13 (0.67–1.90)	0.98 (0.53–1.81)	0.78 (0.41–1.46)	0.94 (0.80–1.12)
Stroke (I60-69)
No. of outcomes	48	50	40	60
Age- and sex-adjusted HR (95% CI)	Reference	0.86 (0.56–1.33)	0.62 (0.38–1.01)	0.87 (0.55–1.36)	0.92 (0.82–1.05)
Multivariable adjusted HR (95% CI) a, b	Reference	0.84 (0.54–1.30)	0.61 (0.38–0.99)	0.82 (0.52–1.29)	0.92 (0.81–1.04)
Non-malignant respiratory disease (J code)
No. of outcomes	53	86	65	78
Age- and sex-adjusted HR (95% CI)	Reference	1.34 (0.93–1.94)	1.32 (0.86–2.02)	1.48 (0.98–2.24)	1.09 (0.97–1.23)
Multivariable adjusted HR (95% CI) b, c	Reference	1.34 (0.93–1.94)	1.28 (0.83–1.96)	1.48 (0.98–2.24)	1.09 (0.97–1.23)
Respiratory tract infections and pneumonia (J1-22)
No. of outcomes	24	37	27	35
Age- and sex-adjusted HR (95% CI)	Reference	1.18 (0.68–2.03)	1.37 (0.69–2.70)	1.65 (0.86–3.17)	1.12 (0.94–1.35)
Multivariable adjusted HR (95% CI) b, c	Reference	1.18 (0.68–2.03)	1.32 (0.67–2.61)	1.66 (0.86–3.18)	1.13 (0.94–1.35)
COPD and others (J23-99)
No. of outcomes	29	49	38	43
Age- and sex-adjusted HR (95% CI)	Reference	1.49 (0.90–2.47)	1.30 (0.74–2.27)	1.38 (0.80–2.37)	1.07 (0.92–1.24)
Multivariable adjusted HR (95% CI) b, c	Reference	1.50 (0.91–2.48)	1.26 (0.72–2.21)	1.38 (0.80–2.38)	1.07 (0.92–1.25)
Lung cancer (C34)
No. of outcomes	79	81	65	94
Age- and sex-adjusted HR (95% CI)	Reference	0.93 (0.56–1.09)	0.70 (0.48–1.04)	0.94 (0.65–1.35)	1.01 (0.91–1.13)
Multivariable adjusted HR (95% CI) b, c	Reference	0.75 (0.54–1.04)	0.70 (0.47–1.04)	0.88 (0.61–1.27)	1.00 (0.90–1.11)

Abbreviations: COPD, Chronic Obstructive Pulmonary Disease; ICD-10, International Classification of Diseases and Related Health Problems, 10^th Revision.

^a Adjusted for age, sex, smoking status, alcohol drinking, body mass index, hypertension, hyperlipidaemia, diabetes, alanine aminotransferase concentration, and proteinuria.

^b Medical administrative areas were treated as strata to allow for a different baseline hazard for each stratum.

^c Adjusted for age, sex, smoking status, alcohol drinking, body mass index, and diabetes.

The multivariable adjusted HR of non-malignant respiratory mortality per 1 µg/m³ increase in the PM_2.5 concentration was 1.09 (95% CI = 0.97–1.23). Compared with the first quartile, the multivariable adjusted HR was 1.48 (95% CI = 0.98–2.24) in the fourth quartile (Table 3). This non-significant positive association was observed for respiratory tract infections and pneumonia mortality rather than for COPD and others mortality. After additional adjustment for leisure-time physical activities, educational background, the O₃ concentration, the NO₂ concentration, and temperature, exclusion of 15 participants who died in the first 3 years, and analysis of only the 202 participants who died after 2016, the results did not change essentially (Supplementary Fig.1). The PM_2.5-associated risk of death from non-malignant respiratory disease did not vary according to several characteristics, including age (40–64 and 65–74 years), smoking status (no and yes), alcohol drinking (no and yes), BMI (＜25.0 and ≥ 25.0 kg/m²), and diabetes (no and yes). This association was observed in men (multivariable adjusted HR per 1 µg/m³ increase in the PM_2.5 concentration = 1.17, 95% CI = 1.01–1.34 for men; HR = 0.94, 95% CI = 0.76–1.15 for women), but there was no statistical evidence of effect modification by sex (Fig.2).

Supplementary Fig.1. Sensitivity analysis of the association between PM_2.5 exposure and non-malignant respiratory disease mortality

The multivariable adjusted model included age, sex, smoking status, alcohol drinking, body mass index, hypertension, and diabetes. Medical administrative areas were treated as strata to allow for a different baseline hazard for each stratum.

Fig.2. Stratified analysis of the association between PM_2.5 exposure and non-malignant respiratory disease mortality

Adjustment was made for age, sex, smoking status, alcohol drinking, body mass index and diabetes. Medical administrative areas were treated as strata to allow for a different baseline hazard for each stratum. The error bars indicate 95% confidence intervals. No statistical effect modification by age, sex, smoking status, alcohol drinking, body mass index, and diabetes mellitus was detected using cross-product terms of exposure categories and these factors.

Discussion

Among the Japanese residents exposed to PM_2.5 ranged between 8.3–13.1 µg/m³, we did not observe adverse effects of long-term exposure to PM_2.5 on deaths from all causes, coronary heart disease, stroke, non-malignant respiratory disease, and lung cancer. Although there was no statistically significant association, our results suggested weak evidence of positive association for non-malignant respiratory disease.

After Dockery et al, reported a positive association between long-term exposure to PM_2.5 and mortality in six U.S. cities¹⁶⁾, many studies, mainly in the U.S. and Europe, provided evidence of an excess risk of all-cause and cardiovascular disease mortality related to elevated PM_2.5 concentrations. In a U.S. study that used data of all Medicare beneficiaries aged 65 years or older (approximately 73 million) between 2000 and 2016, PM_2.5 exposure at concentrations between 6 and 12 µg/m³ was linearly associated with all-cause mortality¹⁷⁾. An European multi-centre cohort study analysed approximately 28 million adults aged 30 years or older who resided in areas with PM_2.5 concentrations lower than 25 µg/m³ and found a positive association between PM_2.5 exposure and cardiovascular mortality¹⁸⁾. On the other hand, in a study that used U.S. Medicare data from 2000 to 2008 and focused on data effect modification by race, PM_2.5 exposure did not increase the risk of all-cause and cardiovascular disease mortality in the Asian population¹⁹⁾. Additionally, recent Chinese cohort studies based on exposure to a wide range of PM_2.5 concentrations showed that the PM_2.5-related risk of all-cause and cardiovascular mortality was clear at PM_2.5 concentrations higher than 50 (or 60) µg/m³, but not at those lower than this^20-22). Although these studies in China included few participants who resided in areas with PM_2.5 concentrations of 8–13 µg/m³ (the range of our study), our finding that there was no positive association between PM_2.5 exposure and all-cause and cardiovascular disease mortality does not seem to contradict the previous finding. However, our exposure assessment based on administrative area did not fully capture levels and trends in residential exposure to PM_2.5, leading to the exposure misclassification. This would be to attenuate the strength of the association. Thus, we should keep in mind that our finding comes with the proviso that exposure misclassification is noted.

For cardiovascular disease, increase in PM_2.5 exposure was not associated with higher stroke mortality risk as well as coronary heart disease mortality risk. Our previous study based on the first cohort of the IPHS, which included residents of Ibaraki Prefecture, found that particulate matter exposure elevated, albeit insignificantly, the risks of both haemorrhagic and ischaemic stroke mortality⁴⁾. Furthermore, the JPHC Study reported that long-term exposure to PM_2.5 adversely affected cerebrovascular disease mortality⁸⁾. However, both cohorts started in the 1990s, and PM_2.5 concentrations tended to have a wider range and be higher than in the present study (24.0–54.3 µg/m³ suspended particulate matter for the first cohort of the IPHS; 7.2–17.9 µg/m³ in the JPHC Study). Although we did not examine aetiology-specific stroke mortality, namely haemorrhagic and ischaemic stroke, due to the relatively small number of deaths caused by stroke, the association was reported to not substantially differ between haemorrhagic and ischaemic stroke^{4, 23)}. A limitation of the current study is that our target outcome was stroke mortality, not incidence. Therefore, the number of outcomes was small and outcome misclassification was considerable compared with if the target outcome had been stroke incidence. However, based on the findings of the present study, exposure to PM_2.5 is unlikely to increase the risk of stroke in Japan currently.

There was no statistical evidence of an association between exposure to PM_2.5 and non-malignant respiratory disease mortality. The PM_2.5 concentration ranged from 8 to 13 µg/m³; however, it seems that the risk of non-malignant respiratory disease mortality somewhat tended to increase as the PM_2.5 concentration increased. After adjustment for co-pollutants including O₃ and NO₂, this tendency did not substantially change. The European Cohorts Integrated Studies, namely, ESCAPE and ELAPSE, found that long-term exposure to PM_2.5 was not associated with non-malignant respiratory and pneumonia-related mortality^{24, 25)}, whereas a study of U.S. Medicare beneficiaries demonstrated that PM_2.5 exposure had adverse effects on non-malignant respiratory disease mortality, especially pneumonia-related mortality¹⁹⁾. Experimental evidence suggests that air pollutants including PM_2.5 disrupt tight junctions of the alveolar epithelium, induce dysfunction of macrophages, modulate cellular receptors that facilitate access of pathogens, and cause dysbiosis of the lung microbiome²⁶⁾. Although there was also the finding that long-term exposure to PM_2.5 was associated with occurrence of chronic lung disease²⁷⁾, pneumonia patients with COPD display higher mortality than those without COPD²⁸⁾. Hence, it is understandable that the insignificant PM_2.5-respiratory mortality association we observed was noted for respiratory tract infections and pneumonia mortality rather than COPD and others mortality. The present study may have overestimated the risk of respiratory mortality because smoking status at baseline might not be adequate to control for potential confounding by smoking and no adjustment was made for passive smoking and indoor air pollution. Given the differences in airway size and respiratory function according to sex, it is not surprising to observe differences in vulnerability to PM_2.5 exposure between men and women. There are not enough findings to conclude whether men or women are more vulnerable to PM_2.5 exposure⁵⁾.

Despite the lack of information on history of lung cancer, a positive association was not detected after adjustment for risk factors of lung cancer. The results of the Three-prefecture Cohort Study in Japan suggested that PM_2.5 exposure increased the risk of lung cancer mortality²⁹⁾. However, this study started in 1983, when air pollutant concentrations were higher than currently, and PM_2.5 concentrations were calculated by converting suspended particulate matter concentrations. Although another Japanese study in Okayama city also found a positive association between PM_2.5 exposure and lung cancer mortality⁹⁾, the average concentrations of PM_2.5 (14 µg/m³) and co-pollutants such as NO₂ (17 ppb)³⁰⁾ were higher than in our study because the study area was urbanised. The International Agency for Research on Cancer classified particulate matter as carcinogenic to humans³¹⁾ and lung cancer is the leading cause of cancer-related death in Japan³²⁾; therefore, further studies including other risk factors, such as smoking history and diet, should be conducted and the cancer risk of PM_2.5 exposure should be carefully considered among the Japanese population.

The strengths of our study include the use of data from a relatively large cohort in Asia, negligible attrition bias, and coverage of the major risk factors for cardiovascular and respiratory diseases. The target population was community-dwelling adults who resided in a single Japanese prefecture near Tokyo. We minimised differences in difficult-to-measure background characteristics, such as traditional dietary habits and climate, which could be residual confounding. Following on from our previous discussion, we acknowledge several limitations of this study. In this study, we examined the linear association between long-term exposure to PM_2.5 and mortality without applying a complex model, because variation in exposure concentrations was poor due to exposure assessment at the administrative area level. Given that its near-linear association was reported⁷⁾, we considered this analysis was adequate. However, detailed evaluation with regard to the sharp of the concentration-response function for long-term exposure to PM_2.5 has remained an issue for the future studies. Also, exposure assessment was based on average PM_2.5 concentrations between 2016 and 2017, whereas the baseline survey of the present cohort was carried out in 2009. However, it appears that spatial patterns of particulate matter concentrations at the administrative area level persisted since the baseline survey. When we only considered deaths after 2016 to evaluate the temporality of the association, a similar association between PM_2.5 exposure and respiratory disease mortality was observed.

In conclusion, long-term exposure to PM_2.5 was not associated with an apparent excess risk of all-cause and cause-specific mortality in the community-dwelling Japanese population. Weak evidence of increased mortality risk due to non-malignant respiratory disease needs further studies in other populations.

Acknowledgements

The authors thank the staffs of Ibaraki Prefectural Government for their research management. The funding sources played no role in the study’s design; in the collection, statistical analysis, or interpretation of data; in the writing of the manuscript; or in the decision to submit this manuscript for publication. The findings and conclusions of this study are solely the responsibility of the authors and do not represent the official views of the Japanese Government.

Funding

This work was supported by the Ministry of the Environment, Japan (no available number) and Ibaraki Prefectural Government (no available number).

Conflict of Interest

All authors have no conflicts of interest to declare.

References

• 1)  World Health Organization. WHO global air quality guidelines: particulate matter (PM_2·5 and PM₁₀), ozone, nitrogen dioxide, sulfur dioxide and carbon monoxide. 2021. Available: https: //www.who.int/publications/i/item/9789240034228 [Accessed 12 Sep 2024]
• 2)  Kulick ER, Kaufman JD and Sack C: Ambient air pollution and stroke: An updated review. Stroke, 2023; 54: 882-893
• 3)  Saito I, Yamagishi K, Kokubo Y, Yatsuya H, Iso H, Sawada N, Inoue M and Tsugane S: Association between mortality and incidence rates of coronary heart disease and stroke: The Japan Public Health Center-based prospective (JPHC) study. Int J Cardiol, 2016; 222: 281-286
• 4)  Takeuchi A, Nishiwaki Y, Okamura T, Milojevic A, Ueda K, Asakura K, Takebayashi T, Hasegawa S, Sairenchi T, Irie F, Ota H and Nitta H: Long-term exposure to particulate matter and mortality from cardiovascular diseases in Japan: The Ibaraki Prefectural Health Study (IPHS). J Atheroscler Thromb, 2021; 28: 230-240
• 5)  The U.S. Environmental Protection Agency. Integrated Science Assessment for Particulate Matter. 2019. Avaiable: https: //www.epa.gov/isa/integrated-science-assessment-isa-particulate-matter [Accessed 12 Sep 2024]
• 6)  Nagashima K, Noma H and Furukawa TA: Prediction intervals for random-effects meta-analysis: A confidence distribution approach. Stat Methods Med Res, 2019; 28: 1689-1702
• 7)  Chen J and Hoek G: Long-term exposure to PM and all-cause and cause-specific mortality: A systematic review and meta-analysis. Environ Int, 2020; 143: 105974
• 8)  Sawada N, Nakaya T, Kashima S, Yorifuji T, Hanibuchi T, Charvat H, Yamaji T, Iwasaki M, Inoue M, Iso H and Tsugane S: Long-term exposure to fine particle matter and all-cause mortality and cause-specific mortality in Japan: the JPHC Study. BMC Public Health, 2022; 22: 466
• 9)  Yorifuji T, Kashima S, Tani Y, Yamakawa J and Doi H: Long-term exposure to fine particulate matter and natural-cause and cause-specific mortality in Japan. Environ Epidemiol, 2019; 3: e051
• 10)  Nishiwaki Y, Michikawa T, Takebayashi T, Nitta H, Iso H, Inoue M and Tsugane S: Long-term exposure to particulate matter in relation to mortality and incidence of cardiovascular disease: The JPHC Study. J Atheroscler Thromb, 2013; 20: 296-309
• 11)  Statistics Bureau, Ministry of Internal Affairs and Communications, Japan. 2010 Population Census. 2010. Available: https: //www.stat.go.jp/english/data/kokusei/2010/summary.html [Accessed 12 Sep 2024]
• 12)  Ibaraki Prefecture Govenment. Annual Report on the Environment 2022, Ibaraki. 2022. Avaiable: https: //www.pref.ibaraki.jp/seikatsukankyo/kansei/kankyo/08hakusho/index.html [Accessed 12 Sep 2024]
• 13)  Sairenchi T, Iso H, Yamagishi K, Irie F, Okubo Y, Gunji J, Muto T and Ota H: Mild retinopathy is a risk factor for cardiovascular mortality in Japanese with and without hypertension; The Ibaraki Prefectural Health Study. Circulation, 2011; 124: 2502-2511
• 14)  Sairenchi T, Iso H, Yamagishi K, Irie F, Nagao M, Umesawa M, Haruyama Y, Kobashi G, Watanabe H and Ota H: Impact and attribute of each obesity-related cardiovascular risk factor in combination with abdominal obesity on total health expenditures in adult Japanese National Health insurance beneficiaries: The Ibaraki Prefectural Health Study. J Epidemiol, 2017; 27: 354-359
• 15)  Okamura T, Tsukamoto K, Arai H, Fujioka Y, Ishigaki Y, Koba S, Ohmura H, Shoji T, Yokote K, Yoshida H, Yoshida M, Deguchi J, Dobashi K, Fujiyoshi A, Hamaguchi H, Hara M, Harada-Shiba M, Hirata T, Iida M, Ikeda Y, Ishibashi S, Kanda H, Kihara S, Kitagawa K, Kodama S, Koseki M, Maezawa Y, Masuda D, Miida T, Miyamoto Y, Nishimura R, Node K, Noguchi M, Ohishi M, Saito I, Sawada S, Sone H, Takemoto M, Wakatsuki A and Yanai H: Japan Atherosclerosis Society (JAS) Guidelines for Prevention of Atherosclerotic Cardiovascular Diseases 2022. J Atheroscler Thromb, 2024; 31: 641-853
• 16)  Dockery DW, Pope CA, 3rd, Xu X, Spengler JD, Ware JH, Fay ME, Ferris BG, Jr. and Speizer FE: An association between air pollution and mortality in six U.S. cities. N Engl J Med, 1993; 329: 1753-1759
• 17)  Josey KP, Delaney SW, Wu X, Nethery RC, DeSouza P, Braun D and Dominici F: Air pollution and mortality at the intersection of race and social class. N Engl J Med, 2023; 388: 1396-1404
• 18)  Stafoggia M, Oftedal B, Chen J, Rodopoulou S, Renzi M, Atkinson RW, Bauwelinck M, Klompmaker JO, Mehta A, Vienneau D, Andersen ZJ, Bellander T, Brandt J, Cesaroni G, de Hoogh K, Fecht D, Gulliver J, Hertel O, Hoffmann B, Hvidtfeldt UA, Jöckel K-H, Jørgensen JT, Katsouyanni K, Ketzel M, Kristoffersen DT, Lager A, Leander K, Liu S, Ljungman PLS, Nagel G, Pershagen G, Peters A, Raaschou-Nielsen O, Rizzuto D, Schramm S, Schwarze PE, Severi G, Sigsgaard T, Strak M, van der Schouw YT, Verschuren M, Weinmayr G, Wolf K, Zitt E, Samoli E, Forastiere F, Brunekreef B, Hoek G and Janssen NAH: Long-term exposure to low ambient air pollution concentrations and mortality among 28 million people: results from seven large European cohorts within the ELAPSE project. Lancet Planetary Health, 2022; 6: e9-e18
• 19)  Wang B, Eum KD, Kazemiparkouhi F, Li C, Manjourides J, Pavlu V and Suh H: The impact of long-term PM_2.5 exposure on specific causes of death: exposure-response curves and effect modification among 53 million U.S. Medicare beneficiaries. Environ Health, 2020; 19: 20
• 20)  Li W, Tian A, Shi Y, Chen B, Ji R, Ge J, Su X, Pu B, Lei L, Ma R, Wang Q, Ban J, Song L, Xu W, Zhang Y, He W, Yang H, Li X, Li T and Li J: Associations of long-term fine particulate matter exposure with all-cause and cause-specific mortality: results from the ChinaHEART project. Lancet Reg Health West Pac, 2023; 41: 100908
• 21)  Liang F, Liu F, Huang K, Yang X, Li J, Xiao Q, Chen J, Liu X, Cao J, Shen C, Yu L, Lu F, Wu X, Wu X, Li Y, Hu D, Huang J, Liu Y, Lu X and Gu D: Long-term exposure to fine particulate matter and cardiovascular disease in China. J Am Coll Cardiol, 2020; 75: 707-717
• 22)  Zhang Y, Yin Z, Li S, Zhang JJ, Sun HZ, Liu K, Shirai K, Hu K, Qiu C, Liu X, Li Y, Zeng Y and Yao Y: Ambient PM_2.5, ozone and mortality in Chinese older adults: A nationwide cohort analysis (2005-2018). J Hazard Mater, 2023; 454: 131539
• 23)  Kulick ER, Eliot MN, Szpiro AA, Coull BA, Tinker LF, Eaton CB, Whitsel EA, Stewart JD, Kaufman JD and Wellenius GA: Long-term exposure to ambient particulate matter and stroke etiology: Results from the Women’s Health Initiative. Environ Res, 2023; 224: 115519
• 24)  Dimakopoulou K, Samoli E, Beelen R, Stafoggia M, Andersen ZJ, Hoffmann B, Fischer P, Nieuwenhuijsen M, Vineis P, Xun W, Hoek G, Raaschou-Nielsen O, Oudin A, Forsberg B, Modig L, Jousilahti P, Lanki T, Turunen A, Oftedal B, Nafstad P, Schwarze PE, Penell J, Fratiglioni L, Andersson N, Pedersen N, Korek M, De Faire U, Eriksen KT, Tjonneland A, Becker T, Wang M, Bueno-de-Mesquita B, Tsai MY, Eeftens M, Peeters PH, Meliefste K, Marcon A, Kramer U, Kuhlbusch TA, Vossoughi M, Key T, de Hoogh K, Hampel R, Peters A, Heinrich J, Weinmayr G, Concin H, Nagel G, Ineichen A, Jacquemin B, Stempfelet M, Vilier A, Ricceri F, Sacerdote C, Pedeli X, Katsoulis M, Trichopoulou A, Brunekreef B and Katsouyanni K: Air pollution and nonmalignant respiratory mortality in 16 cohorts within the ESCAPE project. Am J Respir Crit Care Med, 2014; 189: 684-696
• 25)  Liu S, Lim YH, Chen J, Strak M, Wolf K, Weinmayr G, Rodopolou S, de Hoogh K, Bellander T, Brandt J, Concin H, Zitt E, Fecht D, Forastiere F, Gulliver J, Hertel O, Hoffmann B, Hvidtfeldt UA, Verschuren WMM, Jockel KH, Jorgensen JT, So R, Amini H, Cole-Hunter T, Mehta AJ, Mortensen LH, Ketzel M, Lager A, Leander K, Ljungman P, Severi G, Boutron-Ruault MC, Magnusson PKE, Nagel G, Pershagen G, Peters A, Raaschou-Nielsen O, Rizzuto D, van der Schouw YT, Schramm S, Sorensen M, Stafoggia M, Tjonneland A, Katsouyanni K, Huang W, Samoli E, Brunekreef B, Hoek G and Andersen ZJ: Long-term air pollution exposure and pneumonia-related mortality in a large Pooled European cohort. Am J Respir Crit Care Med, 2022; 205: 1429-1439
• 26)  Monoson A, Schott E, Ard K, Kilburg-Basnyat B, Tighe RM, Pannu S and Gowdy KM: Air pollution and respiratory infections: the past, present, and future. Toxicol Sci, 2023; 192: 3-14
• 27)  Wang X, Chen L, Cai M, Tian F, Zou H, Qian ZM, Zhang Z, Li H, Wang C, Howard SW, Peng Y, Zhang L, Bingheim E, Lin H and Zou Y: Air pollution associated with incidence and progression trajectory of chronic lung diseases: a population-based cohort study. Thorax, 2023; 78: 698-705
• 28)  Restrepo MI, Mortensen EM, Pugh JA and Anzueto A: COPD is associated with increased mortality in patients with community-acquired pneumonia. Eur Respir J, 2006; 28: 346-351
• 29)  Katanoda K, Sobue T, Satoh H, Tajima K, Suzuki T, Nakatsuka H, Takezaki T, Nakayama T, Nitta H, Tanabe K and Tominaga S: An association between long-term exposure to ambient air pollution and mortality from lung cancer and respiratory diseases in Japan. J Epidemiol, 2011; 21: 132-143
• 30)  Yorifuji T and Kashima S: Long-term exposure to nitrogen dioxide and natural-cause and cause-specific mortality in Japan. Sci Total Environ, 2020; 741: 140465
• 31)  Loomis D, Grosse Y, Lauby-Secretan B, El Ghissassi F, Bouvard V, Benbrahim-Tallaa L, Guha N, Baan R, Mattock H, Straif K and International Agency for Research on Cancer Monograph Working Group I: The carcinogenicity of outdoor air pollution. Lancet Oncol, 2013; 14: 1262-1263
• 32)  National Cancer Center Cancer Information Service. Current cancer statistics. 2022. Available: https: //ganjoho.jp/public/index.html [Accessed 12 Sep 2024]