Effects of air pollutants on the incidence, progression, and mortality of idiopathic pulmonary fibrosis: a systematic review and meta-analysis

Abstract

Background

The widespread distribution of air pollutants poses a major challenge to global public health, causing a range of health problems, including the incidence and progression of idiopathic pulmonary fibrosis (IPF). This study aims to provide insights into the specific effects of air pollutants on the risk of IPF through a systematic evaluation and meta-analysis approach.

Methods

The present study was conducted through a comprehensive search of four data—Embase, Web of Science, PubMed, and Cochrane Library—from their inception to 25th November 2024, limited to English-language literature. To assess the potential relationship between a wide range of air pollutants and IPF, a random effects model was used to estimate the risk factors in this study. In addition, subgroup analyses of the data according to age, sex, smoking habits, and geographic location were performed, with the aim of exploring the relationship between air pollutants and the risk of IPF in different populations.

Result

A total of 17 papers covering 14 countries were included in this study, totaling 18 studies involving 858,557 participants and 25,968 event occurrences. Our systematic evaluation and meta-analysis showed an increased risk of IPF disease progression for every 5 µg/m³ increase in PM_2.5 (RR = 1.08, 95% CI:1.01,1.15; I² = 63.51%; p = 0.01; 7 studies). The risk of IPF progression was increased for every 10 µg/m³ increase in NO₂ ( RR = 1.32,95% CI:1.16,1.50; I² = 38.59%; p = 0.12; 7 studies). For every 10 µg/m³ increase in O₃, there was an increased risk of IPF progression (RR = 1.19, 95% CI:1.03,1.38; I² = 29.05%; p = 0.24; 4 studies). For every 10 µg/m³ increase in CO increase of 10 µg/m³ was associated with an increased risk of IPF progression (RR = 1.28, 95% CI:1.01,1.63; I² = 20.03%; p = 0.29; 3 studies). For every 10 µg/m³ increase in NO_X, the risk of IPF progression was increased (RR = 1.21, 95% CI:1.11, 1.33; I² = 13.69%; p = 0.31; 3 studies). For every 10 µg/m³ increase in NO₂, there was an increased risk of IPF incidence (RR = 1.67, 95% CI:1.05,2.66; I² = 36.94%; p = 0.20; 3 studies).

Conclusions

This study found that NO₂ and PM_2.5 increase the risk of IPF disease progression, while NO_X, CO, and O₃ also increase this risk, albeit with limited data. In addition, NO₂ increases the risk of IPF occurrence. Therefore, global health policies targeting reductions in air pollutants like NO₂ and PM_2.5 may reduce the risk of the occurrence and progression of IPF, with significant implications for the future prevention and treatment of IPF.

Graphical Abstract

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-025-24158-1.

Introduction

Idiopathic pulmonary fibrosis (IPF) is a chronic, fibrotic, fatal interstitial lung disease of unknown cause, characterized by dyspnoea and progressive deterioration of lung function [1]. In recent years, the incidence of IPF has been increasing, particularly among the elderly, affecting approximately 3 million people worldwide and significantly impacting global public health [2]. The prognosis of IPF is poor and survival is short, with the median survival of IPF patients in the United States aged 65 and older being only 3.8 years [3]. Despite advances in the diagnosis and treatment of IPF, treatment outcomes remain limited. Of the available treatment options, lung transplantation is indicated for only a minority of patients, while the majority of patients rely primarily on antifibrotic therapy and supportive and palliative measures [4, 5]. However, current interventions to delay or prevent the development of IPF are quite limited. The pathogenesis of IPF has not been fully elucidated, but evidence suggests that air pollution is strongly associated with its development [6–8]. However, the specific relationship between air pollution and IPF incidence, progression, and mortality remains unclear. Therefore, it has become increasingly urgent to study the effects of air pollution on the development and progression of IPF disease.

Air pollution is one of the key environmental concerns globally [9]. The World Health Organisation (WHO) listed air pollution as one of the major threats to human health in 2019 [10]. Inhaled air pollutants mainly include PM₁₀, PM_2.5, ultrafine particles (UFPs), NO₂, O₃, CO, and SO₂ [11]. With the rapid development of global industrialization and urbanization, air pollutants are generated through industrial emissions, vehicle exhaust, and forest fires, and enter the human body to induce disease. According to the Global Exposure Mortality Model (GEMM), about 9 million people worldwide die from air pollution every year [12]. Studies have shown that genetic, occupational exposure, and environmental factors have a significant impact on the development of IPF [6, 7]. Long-term exposure to air pollution increases the risk of the development and progression of IPF [13]. There is an association between PM_2.5 and the progression of IPF disease, which can further deteriorate lung function and increase mortality in patients [14]. Patients with IPF have an increased risk of developing lung cancer when they are exposed to high concentrations of NO₂ (≥ 21. 0 ppb) [15]. Air pollutants, such as PM_2.5 [14]、NO₂ [15], and UFPs [16], are considered risk factors associated with the risk of IPF. Therefore, these controllable airborne risk factors are potential targets for intervention to delay or prevent IPF.

Although existing studies have confirmed that air pollutants may be involved in the onset, clinical progression, and adverse prognosis of IPF, there is a lack of research on the relationship between exposure levels of specific pollutants and the risk of IPF onset, progression, and mortality. Therefore, this study conducted a systematic evaluation and meta-analysis of existing studies on the association between air pollutants and IPF, aiming to identify potential associations between air pollutant exposure and the risk of IPF onset, progression, and mortality, to provide appropriate information and evidence for future clinicians when discussing the risk of IPF with their patients.

Methods

Registration of review protocol

We conducted a systematic review and meta-analysis of studies exploring the association between air pollutants and the risk of IPF in accordance with the Preferred Reporting Items for Systematic Evaluation and Meta-Analysis (PRISMA) criteria [17]. This study protocol was registered with the Prospective Registry of International Systematic Reviews (PROSPERO) on 25 November 2024 for our evaluation protocol, registration number CRD42024619049.

Search strategy

We searched electronic databases such as Embase, Web of Science, PubMed, and the Cochrane Library from the establishment of the database to 25 November 2024. The literature search was limited to English-language literature. The search strategy focused on exposure terms(“Air Pollution”, “PM_2.5”, “Particulate Matter”, “PM₁₀”, “NO₂”, “NO_X”, “SO₂”, “O₃”), disease outcomes (“Idiopathic Pulmonary Fibroses”, “Usual Interstitial Pneumonitides”, “Cryptogenic Fibrosing Alveolitis”), and study design (“Cohort Studies”, “Prospective Studies”, “Longitudinal Studies”, “Case-control Studies”, “Cross-sectional Studies”), using medical subject heading and text word combinations. Each subject term used for the search was fully validated by the literature search service provided by the National Library of Medicine (https://www.ncbi.nlm.nih.gov). In addition, we hand-checked references to relevant subject articles to identify other potentially eligible studies. Detailed information on the searches is provided in the Supplementary Materials section titled “Search Strategy”.

Selection criteria

In the systematic evaluation and meta-analysis, we relied on the PECO-S framework to include eligible studies: (1) general population (excluding those with serious diseases, such as cancer, immune system disorders, etc.; age > 18 years; gender and ethnicity were not restricted); (2) exposures to ambient air pollutants (PM_2.5, PM₁₀, NO₂, NO_X, O₃, CO, and SO₂); (3) objective measurements of air pollutant concentrations; (4) IPF-related outcomes (incidence, progression (IPF progression was uniformly defined as meeting ≥ 1 of the following criteria: (1) based on criteria of decreased lung function (FVC and DLco); (2) Criteria based on imaging deterioration; (3) Criteria based on clinical outcomes (acute respiratory failure, etc.)), and mortality); and 5) observational studies (cohort and case-control, etc.).

We excluded studies with the following conditions: (1) special populations (e.g., non-adult and occupational exposure-related participants); (2) no exposure to ambient air pollutants; (3) no effect on IPF-related outcomes; (4) studies for which the full text or data were unavailable; and (5) conference abstracts, letters, reviews, animal experiments, clinical trials, case reports, or reviews (conference abstract, letter, animal experiment, clinical trial research study, case report, or review).

Based on the above criteria, evaluators CL and YYH performed an initial screening of the titles and abstracts of the retrieved literature to include eligible studies. Literature that was not excluded was independently assessed by two other evaluators, JT and BN, by reading the full text. If disagreements arose during the literature screening process, a fifth evaluator, SZ, adjudicated until a consensus was reached.

Data extraction

Evaluators CL and YYH used standardized forms to extract relevant data from the eligible studies, respectively, which mainly included the following information: (1) basic information about the study (including the first author, year of publication, study period, and source of data); (2) details of the study design (including the study location, the number of subjects, the mean age, the proportion of females, the proportion of smokers, BMI, and the type of study design); (3) air pollution exposure details (including types of pollutants and their mean concentrations); (4) outcome definitions (onset, progression, and death from idiopathic pulmonary fibrosis causing effects) and adjusted maximum risk ratios and their 95% confidence intervals (CIs); and (5) covariates (including age, sex, smoking status, and geographic location). In case of unclear information or missing data in an article, the corresponding author of the manuscript will be contacted to resolve the issue. If multiple publications from the same cohort exist, the study will give preference to the study with the most comprehensive and up-to-date information. Subsequently, evaluator XW will conduct quality control of the entire data extraction process; in case of disagreement, evaluator SZ will be involved in the discussion and resolution.

Data synthesis and analysis

To assess the effect of each air pollutant on idiopathic pulmonary fibrosis, we used the ratio of ratios (OR), relative risks (RR), and 95% CI as summary metrics. In addition, we summarised the RRs and 95% CIs for standardized pollutant concentration increments previously identified in the literature: 10 µg/m³ for O₃, PM₁₀, NO₂, and NO_X, 5 µg/m³ for PM_2.5 and 0.1 mg/m³ for CO. Where necessary, ppb was converted to µg/m³, our conversion assumes a constant temperature of 25 °C and standard atmospheric pressure: 1 ppm = 1000 ppb, 1 ppb NO₂ = 1.88 µg/m³, 1 ppb NOx = 1.9125 µg/m³, and 1 ppb O₃ = 1.96 µg/m³ [18, 19] and the equation was:

Where MW represents each air pollutant, and T represents the actual ambient temperature.

We used the DerSimonian-Laird random effects model [20] to analyze the pooled risk ratios and assess the relationship between exposure to air pollutants and the risk of IPF onset, progression, and mortality. Pooled results were expressed as RRs and their 95% CIs for the risk of IPF onset, progression, and mortality for each 10 µg/m³ increase in air concentrations of O₃, PM₁₀, NO₂, and NO_X, for each 5 µg/m³ increase in PM_2.5 concentration and for each 1 mg/m³ increase in CO concentration. If studies assessed associations of two or more outcomes, these were all included in the analysis. Heterogeneity was assessed by Cochran’s Q test, expressed as an I² statistic, and categorized as significant and non-significant. If heterogeneity was significant (I² ≤ 50%), a fixed-effects model was used; if heterogeneity was high (I² > 50%), a random-effects model was used. We performed subgroup analyses based on predetermined study characteristics, such as geographic location, proportion of females, mean age, and proportion of smokers, to explore factors that significantly influenced the heterogeneity of the results of the main analyses. In addition, we performed ‘leave-one-out’ sensitivity analyses to test for the effect of each study on the overall effect by excluding it when the total number of studies was greater than 10. Potential publication bias was analyzed by funnel plots and Egger’s test; if significant bias was found, adjustments were made by the trim-and-fill method. Two-tailed P values less than 0.05 were considered statistically significant.

Certainty of the evidence

This meta-analysis assessed study quality using the Newcastle-Ottawa Scale (NOS), which is applicable to both cohort and case-control studies [21]. The NOS consists of eight items on three dimensions (selection, comparability, and outcome), which are rated on a scale of up to 9 points out of a maximum of 1 for each item. If the item design was controlled for confounders, the score was out of 2. Studies were categorized into low quality (0–3), moderate quality (4–6), and high quality (7–9) based on the score.

Software, data, and code availability

We used Stata software (version 16, 2019, StataCorp LP, TX, USA) for meta-analyses. Statistical tests were two-tailed, and p < 0.05 was considered statistically significant.

Results

Basic information

After the systematic search, we obtained a total of 1254 literatures. After excluding 129 duplicates, we excluded 1051 documents based on title and abstract screening. Subsequently, 74 literatures were screened in full text and finally, 17 literatures met the inclusion criteria and were included in the systematic evaluation (Fig. 1).

Fig. 1.

Flow chart of systematic review and meta-analysis according to PRISMA guidelines. Flow diagram summarises the search strategy and number of studies excluded at each stage

Of the 17 publications, 1 contained 2 studies from different periods [22]. The 18 studies were from 7 in Europe [8, 23–28], 5 in Asia [22, 29–31], 4 in North American states [32–35], 1 in Australia [36], and 1 in a South American state [37], details of which are shown in Table 1. Of these, 17 were cohort studies and 1 was a case-control Study. The studies primarily assessed the effects of a wide range of air pollutants on IPF, including the occurrence (n = 4), progression (n = 10), and death (n = 5) of IPF. A total of 858,557 participants were involved in this study, with 25,968 events occurring. The mean age range of participants was 41.8 to 73.7 years. Air pollutant ranges from 460.4 to 960 ppb for CO, 40.0 to 125.44 µg/m³ for O₃, 12.60 to 80.84 µg/m³ for NO₂, 43.99 to 50.49 µg/m³ for NO_X, 6.8 to 88.9 µg/m³ for PM_2.5, and 16.23 to 67 µg/m³ for PM₁₀. SO₂ was 2.53 ~ 9 ppb. Among the 18 included studies, 14 were rated as high quality, while 4 were rated as moderate quality. For detailed information, please refer to the supplementary materials.

Table 1.

Contextual details of studies included in the systematic review and meta-analysis

Author [citation]	Country	Region	Publication year	Study year	Design	Research center	Data source	Events(N)	Total(N)	Age(average or median, years)*	BMI*	Smoke (%)*	Female (%)*	Outcome	Pollutans	Concentration (average or median, µg/m3)*
Pablo Mariscal-Aguilar	Bolivia	South America	2024	2011–2020	Cohort study	Single center	The Interstitial Lung Disease Unit of La Paz University Hospital	69	69	73.7	26.4	43.48	23.19	Progression of IPF	O3	NA
															CO
															CO
															NO2
															NOX
Qiang Zheng	Australia	Oceania	2023	2012–2019	Cohort study	Multiple centers	Australian IPF Registry (AIPFR)	570	570	70.9	28.9	70	29.6	Progression of IPF	PM2.5	6.8 µg/m3
															NO2	12.60 µg/m3
Xiaojie Wang	UK	Europe	2023	2006–2010	Cohort study	Multiple centers	The UK Biobank (UKB)	2589	402,042	62.25	28.72	68.79	39.32	Incidence of IPF	SO2	2.53 µg/m3
Pablo Mariscal-Aguilar	Spain	Europe	2023	2011–2020	Cohort study	Single center	At the Pulmonary Fibrosis Specialty Clinic of the Hospital Universitario La Paz Pneumology Department	69	69	73.7	26.4	43.48	23.19	Progression of IPF	CO	NA
															NO2
															NOX
Feipeng Cui	Spain	Europe	2023	2006–2010	Cohort study	Multiple centers	The UK Biobank (UKB)	1380	433,738	62.8	28.7	45.1	35.8	Progression of IPF	NO2	26.65 µg/m3
															NOX	43.99 µg/m3
															PM2.5	9.99 µg/m3
															PM10	16.23 µg/m3
Hee-Young Yoon	Korea	Asia	2023	1995–2016	Cohort study	Single center	Asan Medical Centre, Seoul, Republic of Korea	946	946	65.4	24.3	23.89	19.13	Progression of IPF	NO2	43.43 µg/m3
Lirong Liang	China	Asia	2022	2013–2017	Cohort study	Multiple centers	A database compiled by the Beijing Public Health Information Center	11,974	11,974	NA	NA	NA	46.15	Progression of IPF	PM2.5	76.7 µg/m3
				2008–2012											PM2.5	88.9 µg/m3
Na’ama Avitzur	USA	North America	2022	2001–2017	Cohort study	Multiple centers	ILD programme database at the University of California, San Francisco (UCSF)	603	603	70.5	NA	30.02	23.71	Progression of IPF	PM2.5	8.7 µg/m3
															O3	40.0 µg/m3
Ioannis Tomos	Greece	Europe	2021	2013–2019	Cohort study	Single center	2nd Pulmonary Medicine Department, National and Kapodistrian University of Athens, Medical School	118	118	72	NA	73.73	25.42	Progression of IPF	NO2	22.2 µg/m3
															PM2.5	27.2 µg/m3
															PM10	30.5 µg/m3
Pablo Mariscal Aguilar	Spain	Europe	2021	2013–2019	Cohort study	Single center	Department of Pneumology at La Paz University Hospital	52	52	66	NA	55.77	21.15	Mortality of IPF	CO	NA
															NO2	58.18 µg/m3
															PM2.5	11.21 µg/m3
															PM10	21.22 µg/m3
															O3	NA
															SO2	NA
Masahiro Tahara	Japan	Asia	2021	2009–2014	Case-control	Multiple centers	Nationwide cloud-based integrated database	152	352	65	NA	NA	30.26	Incidence of IPF	NO	10.8 ppb
															PM2.5	17 µg/m3
															SO2	2.6 ppb
															NO2	31.40 µg/m3
															NOX	50.49 µg/m3
															CO	460.4 ppb
															O3	54.88 µg/m3
															PM10	20.1 µg/m3
Hee-Young Yoon	Korea	Asia	2021	1995–2016	Cohort study	Single center	At the Asan Medical Center, Seoul, Republic of Korea	1114	1114	65.7	24.1	75.9	19.5	Mortality of IPF	NO2	42.3 µg/m3
Robert Dales	Canada	North America	2020	2001–2012	Cohort study	Multiple centers	The Ministry of Health, the official source of disease statistical data in Chile	3989	3989	NA	NA	NA	58.26	Incidence of IPF	CO	960 ppb
															NO2	80.84 µg/m3
															O3	125.44 µg/m3
															SO2	9 ppb
															PM10	67 mg/m3
															PM2.5	29 mg/m3
Sara Conti	Italy	Europe	2018	2005–2010	Cohort study	Multiple centers	Regional healthcare administrative databases	2093	2093	41.8	NA	NA	50.7	Incidence of IPF	NO2	41 µg/m3
Kerri A Johannson	USA	North America	2018	2014	Cohort study	Single center	At the University of California San Francisco	25	25	73.6	NA	68	16	Progression of IPF	NO2	15.08 µg/m3
															PM2.5	9.01 µg/m3
															PM10	17.8 µg/m3
Christopher J Winterbottom	USA	North America	2018	2007–2013	Cohort study	Single center	A single university ILD referral center	135	175	68	NA	64	25	Mortality of IPF	PM10	18.5 µg/m3
															PM2.5	10.5 µg/m3
Lucile Sesé	France	Europe	2018	2007–2014	Cohort study	Multiple centers	The French cohort COhorte FIbrose (COFI)	40	192	68	27.7	67.5	25	Progression of IPF	O3	NA
															NO2	NA
															PM10	19.46 µg/m3
															PM2.5	26.23 µg/m3
														Mortality of IPF	O3	NA
															NO2	NA
															PM10	19.46 µg/m3
															PM2.5	26.23 µg/m3
Kerri A Johannson	USA	North America	2014	2001–2010	Cohort study	Single center	At Asan Medical Center in Seoul, South Korea	75	436	63.7	NA	65.33	22.67	Progression of IPF	O3	NA
															NO2
															PM10
															CO
															SO2
														Mortality of IPF	O3	NA
															NO2

Not all studies reported gender-specific population sizes Smoking status, BMI, air pollutant concentration, or the mean age. The data is mentioned for all studies that reported it

Association between air pollutants and risk of progression of IPF

Pooling the data from all included studies, our meta-analysis showed that CO, NO₂, NO_X, O₃, and PM_2.5 air pollutants were significantly and positively associated with the risk of IPF disease progression, respectively (Fig. 2; Table 2). However, there was no significant association between SO₂ and PM₁₀ and the risk of IPF disease progression. We observed a significant association between ambient NO₂ and the risk of IPF disease progression. Specifically, a 10 µg/m³ increase in ambient NO₂ is associated with a 1.32 (1.16, 1.50) point change in the progress of IPF (Fig. 2; Table 2). Of the eight included studies, six showed a positive association between NO₂ and the risk of IPF disease progression [24–26, 29, 35, 37]. Studies by Qiang Zheng [36] and Lucile Sesé [8] showed no significant positive correlation between ambient NO₂ and the progression of IPF disease.

Fig. 2.

Forest plot of studies reporting exposure to ambient air pollutants and progression of IPF risk. Random-effects meta-analysis was used to meta-analyse the evidence for an association between exposure to air pollutants in the environment and the risk of disease progression in IPF. Squares indicate relative risk and bars indicate 95% CI for each study. All statistical tests were two-sided

Table 2.

Random effects meta-analysis on the correlation between exposure to air pollutants and the risk of progression of idiopathic pulmonary fibrosis

Pollutant	Author	N association estimates	RR (95%CI)	I2(%)	Heterogeneity p-value ^
CO	Pablo Mariscal-Aguilar, 2024	3	1.28 (1.01, 1.63)	20.03	0.29
	Pablo Mariscal-Aguilar, 2023
	Kerri A Johannson, 2014
NO2	Pablo Mariscal-Aguilar, 2024	8	1.32 (1.16, 1.50)	38.59	0.12
	Pablo Mariscal-Aguilar, 2023
	Qiang Zheng, 2023
	Feipeng Cui, 2023
	Hee-Young Yoon, 2023
	Ioannis Tomos, 2021
	Lucile Sesé, 2018
	Kerri A Johannson, 2014
NOX	Pablo Mariscal-Aguilar, 2024	3	1.21 (1.11, 1.33)	13.69	0.31
	Pablo Mariscal-Aguilar, 2023
	Feipeng Cui, 2023
O3	Pablo Mariscal-Aguilar, 2024	4	1.19 (1.03, 1.38)	29.05	0.24
	Na’ama Avitzur, 2022
	Lucile Sesé, 2018
	Kerri A Johannson, 2014
PM10	Feipeng Cui, 2023	4	1.20 (0.93, 1.54)	39.43	0.18
	Ioannis Tomos, 2021
	Lucile Sesé, 2018
	Kerri A Johannson, 2014
PM2.5	Lirong Liang, 2022 A	7	1.08 (1.01, 1.15)	63.51	0.01
	Lirong Liang, 2022B
	Qiang Zheng, 2023
	Feipeng Cui, 2023
	Na’ama Avitzur, 2022
	Ioannis Tomos, 2021
	Lucile Sesé, 2018
SO2	Kerri A Johannson, 2014	1	1.08 (0.85, 1.37)	/	/

NO_X is significantly associated with the risk of disease progression in IPF. Specifically, a 10 µg/m³ increase in ambient NO_X is associated with a 1.21 (1.11, 1.33) point change in the progress of IPF (Fig. 2; Table 2). NO_X was positively associated with the risk of disease progression in IPF in all three included studies [24, 25, 37].

O₃ is significantly associated with the risk of disease progression in IPF. Specifically, a 10 µg/m3 increase in ambient O₃ is associated with a 1.19 (1.03, 1.38) point change in the progress of IPF (Fig. 2; Table 2). Of the four included studies, two showed a positive association between O₃ and the risk of disease progression of IPF [32, 37]. The study by Na’ama Avitzur [32] showed no significant negative association between O₃ and the risk of disease progression of IPF. The study by Kerri A Johannson [35] showed no significant positive association between O₃ and the risk of disease progression of IPF. Significant positive correlation.

CO is significantly associated with the risk of disease progression in IPF. Specifically, a 1 mg/m3 increase in ambient CO is associated with a 1.28 (1.01, 1.63) point change in progress of IPF (Fig. 2; Table 2). Of the three included studies, only one showed a positive association between CO and the risk of IPF disease progression [24]. Studies by Kerri A Johannson [35] and Pablo Mariscal-Aguilar [37] showed no significant positive association between CO and the risk of IPF disease progression.

PM_2.5 is significantly associated with the risk of disease progression in IPF. Specifically, a 5 µg/m³ increase in ambient PM_2.5 is associated with a 1.08 (1.01, 1.15) point change in the progress of IPF (p = 0.01) (Fig. 2; Table 2). Of the six included studies, three showed that PM_2.5 was positively associated with the risk of IPF disease progression [22, 25, 26]. studies by Qiang Zheng [36] and Lucile Sesé [8] showed that PM_2.5 was not significantly positively associated with IPF disease progression. a study by Na’ama Avitzur [32] showed that PM_2.5 was not significantly negatively associated with the risk of IPF disease progression was not significantly negatively correlated.

Association between air pollutants and risk of incidence of IPF

The results of our meta-analysis showed that NO₂ and PM₁₀ air pollutants were significantly and positively associated with the risk of incidence of IPF, respectively (Fig. 3; Table 3). However, CO, O₃, PM_2.5, and SO₂ were not significantly associated with the risk of incidence of IPF; NO and NO_X were not included in the analysis due to insufficient literature. We observed a significant association between ambient NO₂ and the risk of incidence of IPF. Specifically, a 10 µg/m³ increase in ambient NO₂ is associated with a 1.37 (1.05, 2.66) point change in the incidence of IPF (Fig. 3; Table 3). Of the three included studies, two showed a positive association between NO₂ and the risk of incidence of IPF [28, 33]. A study by Masahiro Tahara [30] reported no significant positive association between NO₂ and IPF onset.

Fig. 3.

Forest plot of studies reporting exposure to ambient air pollutants and incidence of IPF risk. Random-effects meta-analysis was used to meta-analyse the evidence for an association between exposure to air pollutants in the environment and the risk of incidence of IPF. Squares indicate relative risk and bars indicate 95% CI for each study. All statistical tests were two-sided

Table 3.

Random effects meta-analysis on the correlation between exposure to air pollutants and the risk of incidence of IPF

Pollutant	Author	N association estimates	RR (95%CI)	I2(%)	Heterogeneity p-value ^
NO2	Masahiro Tahara, 2021	3	1.67 (1.05, 2.66)	36.94	0.2
	Robert Dales, 2020
	Sara Conti, 2018
PM10	Masahiro Tahara, 2021	2	1.29 (1.11, 1.50)	0	0.49
	Robert Dales, 2020

PM₁₀ was significantly associated with the risk of IPF onset. Specifically, a 10 µg/m³ increase in ambient PM₁₀ is associated with a 1.29 (1.11, 1.50) point change in incidence of IPF (Fig. 3; Table 3). Of the 2 included studies, only 1 reported that PM₁₀ was positively associated with the risk of IPF onset [33]. A study by Masahiro Tahara [30] showed that PM₁₀ was not significantly associated with the risk of IPF onset.

Association between air pollutants and risk of mortality of IPF

Our meta-analysis showed no significant association between CO, NO₂, O₃, PM_2.5, PM_10, and SO₂ air pollutants and the risk of death from IPF disease (Fig. 4).

Fig. 4.

Forest plot of studies reporting exposure to ambient air pollutants and mortality of IPF risk. Random-effects meta-analysis was used to meta-analyse the evidence for an association between exposure to air pollutants in the environment and the risk of mortality of IPF. Squares indicate relative risk and bars indicate 95% CI for each study. All statistical tests were two-sided

Subgroup analysis

Table 4 mainly shows the results of the subgroup meta-analysis of NO₂, PM_2.5, and O₃ in relation to the risk of IPF disease progression. When analyzing the association of NO₂ and PM_2.5 with the risk of IPF disease progression, we grouped them according to age (> 70 and ≤ 70), and the results showed that the RR values of the two groups were 1.41 (95% CI: 1.09,1.50; I² = 56.65%) and 1.26 (95% CI: 1.09,1.47; I² = 26.23%), both of which were statistically significance, with lower heterogeneity in the ≤ 70 years group. This suggests that the RR values for NO₂ and risk of disease progression in IPF were comparable in both age groups. In the regional analysis of NO₂, only the European region had statistically significant RR values (RR = 1.30; 95% CI: 1.10,1.53; I² = 28.14%) with a low level of heterogeneity. Gender analysis showed that the proportion of females was < 50% in all cases, and males were significantly associated with the risk of progression of NO₂ and IPF (RR = 1.32; 95%CI: 1.16,1.50; I² = 38.58%). The RR values for smoking ≥ 50% (RR = 1.21; 95%CI:1.04,1.42; I² = 36.63%) and < 50% (RR = 1.49; 95%CI:1.22,1.81; I² = 22.66%) were statistically significant and less heterogeneous in the former.

Table 4.

Random effects meta-analysis of the association between exposure to NO₂, PM_2.5, and O₃ and the risk of progression of IPF

Pollutant	Category	Study Characteristics (Number of association estimates)	Summary RR (95%CI)	I2 (%)	p for heterogeneity
NO2	Age	Age ≤ 70 (4)	1.26 (1.09, 1.47)	26.23	0.25
		Age >70 (4)	1.41 (1.09, 1.82)	56.65	0.07
	Female	Female <50 (8)	1.32 (1.16, 1.50)	38.58	0.12
		Female ≥ 50 (0)	/	/	/
	Smoke	Smoke <50 (4)	1.49 (1.22, 1.81)	22.66	0.27
		Smoke ≥ 50 (4)	1.21 (1.04, 1.42)	36.63	0.19
	Region	Europe (4)	1.30 (1.10, 1.53)	28.14	0.24
		Others (4)	1.36 (1.07, 1.73)	58.45	0.07
	Outcome	Incidence (3)	1.67 (1.05, 2.66)	36.93	0.20
		Mortality (4)	1.02 (0.96, 1.08)	0.00	0.78
		Progression (8)	1.32 (1.16, 1.50)	38.58	0.12
PM2.5	Age	Age ≤ 70 (2)	1.38 (1.12, 1.70)	0.00	0.54
		Age >70 (3)	1.15 (0.69, 1.82)	78.52	0.01
	Female	Female <50 (5)	1.23 (0.90, 1.70)	69.05	0.01
		Female ≥ 50 (0)	/	/	/
	Smoke	Smoke <50 (2)	1.01 (0.55, 1.86)	88.19	0
		Smoke ≥ 50 (3)	1.51 (0.94, 2.44)	48.17	0.15
	Region	Europe (3)	1.49 (1.15, 1.94)	11.34	0.32
		Others (4)	1.05 (1.02, 1.09)	33.64	0.21
	Outcome	Incidence (2)	1.67 (0.87, 3.19)	71.19	0.06
		Mortality (3)	2.03 (0.48, 8.54)	82.18	0.00
		Progression (7)	1.08(1.01, 1.15)	63.51	0.01
O3	Age	Age ≤ 70 (2)	1.28 (0.96, 1.70)	47.15	0.17
		Age >70 (2)	1.14 (1.00, 1.30)	0.00	0.40
	Female	Female <50 (4)	1.19 (1.03, 1.38)	29.05	0.24
		Female ≥ 50 (0)	/	/	/
	Smoke	Smoke <50 (2)	1.14 (1.00, 1.30)	0.00	0.40
		Smoke ≥ 50 (2)	1.28 (0.96, 1.70)	47.15	0.17
	Outcome	Incidence (2)	1.22 (0.94, 1.58)	0	0.35
		Mortality (3)	0.97(0.88, 1.07)	0.00	0.46
		Progression (4)	1.19(1.03, 1.38)	29.05	0.24

In the association analysis between PM_2.5 and the risk of disease progression in IPF, the RR value (RR = 1.15; 95%CI: 0.69,1.82; I² = 78.52%) was not statistically significant for those aged > 70, whereas the RR value (RR = 1.38; 95%CI: 1.12,1.70; I² = 0.00%) was statistically significant for those ≤ 70. No statistically significant RR values were found for smoking percentages ≥ 50% and < 50%, although heterogeneity was lower in the former (I² = 48.17%). The RR values were not statistically significant in any of the regions except for the European region (RR = 1.49; 95%CI: 1.15,1.94; I² = 11.34%) and the Asian region (RR = 1.05; 95%CI: 1.03,1.07; I² = 0.00%) where the RR values of the studies were statistically significantly associated. In addition, in the gender analysis, there was no statistically significant association between the RR values of studies with < 50% females.

In the analysis of the association between O₃ and the risk of disease progression in IPF, the RR value (RR = 1.14; 95%CI:1.00,1.30; I² = 0.00%) was statistically significant in the age > 70 population, whereas it was not statistically significant for ≤ 70. In the analysis of gender, the RR value for the proportion of women < 50% for O₃ (RR = 1.19; 95%CI: 1.03,1.38; I² = 29.05%) was statistically significant. The RR value (RR = 1.14; 95%CI:1.00,1.30; I² = 0.00%) for the proportion of smokers < 50% was statistically significant.

In addition, in the gender analysis, the RR value for the association of the risk of disease progression in IPF with CO was RR = 1.47; 95%CI:1.08,1.99; I² = 54.41% for the proportion of females < 50%, and RR = 1.21; 95%CI:1.11,1.33; I² = 13.68% for the association with NO_X, both of which were statistically significant (Fig. 5A). In the association analysis between NO_X and the risk of disease progression in IPF, the RR value for the proportion of smoking < 50% (RR = 1.21; 95%CI:1.11,1.33; I² = 13.68%) was statistically significant (Fig. 5B).

Fig. 5.

Forest plot of studies reporting exposure to CO and NO_X air pollutants and progression of IPF risk. A The effect of CO and NO_X on the risk of IPF progression in subgroups of < 50% female analysis. B The effect of NOX on the risk of IPF progression in subgroups with smoking rates < 50

Publication of bias and sensitivity analyses

We performed publication bias and sensitivity analyses mainly on the relationship between NO₂ and PM_2.5 and the risk of progression in IPF (Fig. 6). The funnel plot and Egger’s test revealed that there was publication bias for NO₂ (P = 0.033 < 0.05), which was statistically significant, whereas there was no publication bias for PM_2.5 (P = 0.275 > 0.05), which was not statistically significant, and the trim-and-fill method did not change the results. The sensitivity analyses did not show contradictory results compared to the main analyses, suggesting that the analyses were robust.

Fig. 6.

Funnel plots to assess publication bias. Publication bias in the pooled associations of (left) NO₂ and (right) PM_2.5 air pollution with progression of IPF risk

Discussions

Principal findings

This study is the first systematic review and meta-analysis examining the association between air pollution and IPF onset, progression, and mortality. We analyzed 17 papers (14 countries, 858,557 participants, 25,968 events) and found consistent evidence linking higher air pollutant levels with IPF risk. PM_2.5 and NO₂ were significantly associated with IPF incidence, while CO, NO_X, and O₃ showed ties to progression. In the development of IPF, we only found a significant positive correlation between NO₂ and IPF incidence. Although PM₁₀ also showed a correlation, the literature is limited. Notably, NO₂ was a risk factor for IPF progression in men, possibly due to higher smoking rates, which exacerbate inflammation and fibrosis. Smoking itself showed a dose-dependent relationship with IPF risk [38, 39], supported by Mendelian randomization evidence [40]. NO₂ is a risk factor for IPF disease progression in the European region, possibly related to higher NO₂ concentrations in Europe [41, 42]. This is consistent with Johnson et al. [35] who reported that acute exacerbation of IPF may be associated with increased NO₂ concentration. In recent years, with increasing air pollution, a study by Yang et al. [43] found that IPF is mainly caused by the senes/cence of alveolar type 2 cells and that PM_2.5 can lead to lung injury and fibrosis by inducing cellular senescence. It has also been found that airborne PM_2.5 particulate matter can induce senescence of alveolar epithelial cells and lung tissues and exacerbate disease progression by promoting inflammatory responses and DNA damage [44]. Our study showed that PM_2.5 was an unfavorable factor for IPF disease progression in people aged ≤ 70 years. Therefore, we hypothesize that the main reason is that PM_2.5 is more likely to induce alveolar type 2 epithelial cell senescence in young and old people and damage DNA to promote IPF progression. Further experimental validation is needed.

Potential mechanisms

From the above studies, we found that the current research on the effects of NO₂ and PM_2.5 on IPF is the most focused. However, whether the association between NO₂ and PM_2.5 and IPF indicates a causal relationship, and the specific mechanisms by which they contribute to the occurrence, progression, and death of IPF, respectively, have not yet been fully clarified. IPF is known to be a chronic, progressive and irreversible interstitial lung disease with a complex pathogenesis involving genetic susceptibility, environmental exposure and abnormal repair response. In recent years, air pollutants such as NO₂ and PM_2.5 have been identified as significant environmental risk factors for IPF. NO₂ and PM_2.5 exacerbate IPF through inflammatory responses, oxidative stress, telomere dysfunction, and DNA damage [45, 46]. As lungs are its direct target, PM_2.5 penetrates alveoli and enters circulation, carrying adsorbed toxins (polycyclic aromatic hydrocarbons, metals, dioxins) [47, 48]. These compounds cause oxidative stress by damaging antioxidant defense, increasing reactive oxygen species (ROS) [49]; induce mitochondrial/DNA damage and promote autophagy/apoptosis of alveolar epithelial cells (AECs) and triggering oxidative stress [50, 51]; activate the fibrotic pathway through the release of inflammatory cytokines [49]. NO₂ inhibits telomerase, accelerates alveolar epithelial aging, and mirrors the telomere shortening seen in IPF and pollution-exposed lungs [52]. High NO₂ exposure increases mortality from chronic bronchitis, COPD, asthma, lung cancer, and IPF [15, 53–55] by generating lipid peroxides that injure airway epithelium, trigger compensatory proliferation and inflammation, and drive epithelial-to-mesenchymal transition, leading to fibrotic tissue replacement [56]. The above experimental evidence suggests that long-term exposure to NO₂ and PM_2.5 can break through the physiological barrier of the respiratory tract and settle in the terminal bronchiolar and alveolar regions. This deposition process triggers a local inflammatory cascade, promoting the production of reactive oxygen species, which in turn causes oxidative damage and DNA damage to type II alveolar epithelial cells. The continuous cell damage and abnormal repair eventually lead to the remodeling of lung tissue structure, forming the characteristic pathological changes of fibrosis. This is consistent with the results of this study, which showed that for every 5 µg/m³ increase in PM_2.5 exposure, or 10 µg/m³ increase in NO₂, the risk of disease progression in IPF increased correspondingly. The cohort study by Cui et al. [25]. also confirmed that long-term exposure to PM_2.5 pollution was significantly positively correlated with the risk of IPF. This suggests that long-term exposure to PM_2.5 and NO₂ can lead to oxidative stress and DNA damage, promoting the occurrence and progression of IPF. However, Further studies from basic medicine to clinical medicine level are needed to thoroughly understand the mechanism of action of NO₂, PM_2.5, and other air pollutants on the occurrence and development of IPF.

Comparison with previous studies

A previous meta-analysis, mainly Harari et al. [6] assessing the effects of air pollutants on fibrotic interstitial lung disease, did not provide risk ratios for relevant findings. In contrast to the Harari et al. study, the present study specifically collected and categorized studies on the effects of multiple air pollutants on IPF outcomes, including the onset, progression, and mortality of IPF, and included air pollutants such as NO_X and SO₂ in the exposure-based analysis. As the studies included by Harari et al. were older, relevant new studies from recent years were included in this study for updating. These processes increased the rigor and comprehensiveness of this study and helped to improve the quality of the study.

Notably, our study also applied subgroup analyses to explore the heterogeneity among the included studies, such as a statistically significant higher risk of exacerbation in male IPF patients compared to females under the influence of PM_2.5, NO₂, CO, and NO_X. For O₃, the results were reversed. It is unclear whether men are more sensitive to air pollution than women. However, studies by Bell [57] and Kim [58] showed that females were more susceptible to air pollutants than males. Although the study by Harari et al. provided preliminary evidence for the association of air pollutants with the risk of IPF, the present study accurately differentiated between exposures to each air pollutant by implementing stricter inclusion and exclusion criteria and included more new studies from recent years.

Strengths and limitations

This study is the first meta-analysis of the association between multiple air pollutants and the risk of IPF incidence, progression, and mortality. First, we conducted a comprehensive and systematic search in four databases. Second, we conducted a rigorous and systematic quality assessment of the included studies. We included a total of 18 studies, including 17 cohort studies, with high-quality assessment of the literature, risk of bias assessment, and sensitivity analyses to ensure data credibility and robustness of the results. We conducted subgroup analyses by age, sex, smoking, and geographic location to explore whether associations varied by these clinical characteristics.

This study has several limitations. First, most included studies lacked detailed air pollutant concentration data, preventing dose-response analysis. While observational data suggest associations between pollutants and IPF risk, causal relationships remain unproven—future Mendelian randomization studies could clarify this. Additionally, exposure settings (indoor or outdoor) were rarely specified, though pollutant profiles differ significantly between these environments. Second, socioeconomic factors (income, education), occupational exposures (dust), and comorbidities (COPD, cardiovascular disease) may modify pollutant effects, but standardized reporting of these confounders was absent. We urge future studies to: (1) assess how socioeconomic stratification influences pollution-disease associations; (2) evaluate occupational protections for high-risk groups; and (3) define comorbidity-driven toxicity thresholds. Furthermore, limited study numbers in subgroup analyses reduced robustness against publication bias. The included studies applied heterogeneous diagnostic criteria for defining IPF disease progression. This variability in definitions may introduce heterogeneity in effect sizes, thereby compromising the stability of pooled estimates. Although we attempted to mitigate such variation through subgroup analyses and random-effects models, the results should be interpreted with caution when combining data. Future studies must standardize the definition of disease progression to enhance the generalizability of meta-analysis conclusions. Finally, single-pollutant models may not be able to assess interactions between air pollutants, and future studies should use hybrid models. Notably, no significant link emerged between pollutants and IPF mortality, possibly due to advanced IPF are older, have largely lost their lung function, and are less sensitive to air pollutants; insufficient sample sizes may also have led to meaningless results. Therefore, larger, standardized studies are warranted.

Conclusions

Our findings suggest that high NO₂ and PM_2.5 exposures increase the risk of IPF progression, with associations with NO_X, CO, and O₃ as well, but there is limited data in the literature. In particular, high NO₂ concentrations also increased the risk of IPF development. Therefore, reducing the concentration of air pollutants such as NO₂ and PM_2.5 is potentially relevant for the prevention and treatment of IPF, especially in women and non-smokers with a significant protective effect. Overall, our findings strengthen the evidence for air pollutants as risk factors for IPF and provide a reference for clinicians and health policymakers to develop measures or actions to ameliorate global environmental problems and reduce the risk of IPF.

Authors’ contributions

Cheng Luo: Conceptualization, Formal analysis, Methodology(statistical analysis & extraction), Writing - original draft, Writing - review & editing. Xinhui Wu: Writing - original draft, Software, Methodology(extraction). Shipeng Zhang: Writing - original draft, Methodology(statistical analysis & extraction), Software, Visualization. Junwen Tan: Writing - original draft, Formal analysis, Methodology(extraction). Xingling Song: Writing - original draft. Bo Ning: Writing - original draft. Qi Tang: Writing - original draft. Yuzhi Huo: Writing - original draft. Jiajie Li: Writing - original draft. Yuanhang Ye: Writing - original draft, Formal analysis, Software, Writing - review & editing, Methodology(statistical analysis & extraction). Fei Wang: Conceptualization, Project administration, Supervision, Funding acquisition, Writing - review & editing.

Funding

The research was funded by Joint Innovation Fund of Health Commission of Chengdu and Chengdu University of Traditional Chinese Medicine (2024120973), National Science and Technology Major Project (2020YFC2003104) and National Natural Science Foundation of China (82174347).

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

Ethical approval and informed consent were not applicable. However, the review protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO: CRD42024619049).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.