Introduction

Exposome-associated morbidity and mortality is a global health concern. Environmental exposures that individuals encounter over their lifetime include air pollution, water pollution, diet, and radiation. The exposome has been linked to heterogeneous negative health effects, and mechanisms remain elusive in many disease states. Studying the exposome provides valuable insights into the interplay between environmental factors and human health1.

Of the various environmental exposures, inhalational exposure has been of great interest, especially in the context of rising levels of global air pollution due to global warming, wildfires, wars, and population growth. Studies investigating the link between environmental exposures and disease have the potential to impact millions globally. Air pollution is associated with 7 million premature deaths annually, and levels have steadily risen over the past few decades2,3. Over half of the world’s population is exposed to levels of air pollution that are substantially above the World Health Organization (WHO) air quality guidelines4. A greater proportion of non-communicable diseases are attributable to environmental exposure in developing countries that utilize industrial production factories, without the most modern emission safeguards and therefore are primary contributors to emissions5.

Particulate matter (PM) exposure, is a global cause of significant aerodigestive morbidity and mortality6,7. Globally, gastroesophageal reflux disease (GERD) prevalence is 10–25%8,9. GERD is the most prevalent gastrointestinal disorder affecting at least 20% of the United States of America (USA) population, and leading to substantial morbidity8,10,11,12. Aerodigestive complications also include Barrett’s Esophagus (BE), and malignancy such as esophageal adenocarcinoma (EAC)13,14,15. The aerodigestive disease can also induce or worsen airway hyperreactivity (AHR) and other forms of obstructive airway disease (OAD). This may be explained by the clearing mechanism of the respiratory system and its proximity to the digestive system at the pharynx leading to gastric reflux being transported into the lungs. However, this is an area of active investigation16. Prior systematic reviews have only focused on single inhalational exposures and/or single diseases/outcomes17,18,19,20,21,22.

Our group has focused on the adverse health effects secondary to the destruction of the World Trade Center (WTC) on September 11, 2001 (9/11). This intense PM exposure of first responders and inhabitants of New York City (NYC) led to heterogeneous end-organ involvement23,24. WTC-PM exposure is associated with OAD and gastroesophageal diseases including GERD and BE25,26,27. Approximately 44% of WTC rescue/recovery workers had developed GERD symptoms by 200528. There is also evidence of comorbid GERD and OAD, as WTC-exposed firefighters with OAD had a 3-fold higher risk of developing GERD27,29. Therefore, due to our interest in a more diverse exposure profile we have also designed our systematic review to focus on heterogeneous gastroesophageal diseases. Specifically, we investigated the associations between the exposures (PM and smoking) with diseases of the gastrointestinal tract (GERD, BE, and malignancy).

Methods

Search Strategy & Identification. Our systematic review adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines30,31. Our Population, Exposure, Outcome (PEO) question was to investigate among adult populations (P), whether there is an association between inhalational exposure (e.g., air pollution, cigarette/tobacco smoke, marijuana smoke, vape/e-cigarette aerosols) (E) and esophageal or gastric disorders/disease (O).

The protocol of our systematic review was registered on PROSPERO, April 29, 2024, and can be accessed at Prospero ID 536,834. A comprehensive search was conducted in PubMed (May 1, 2024) and Web of Science (WoS, August 23, 2024) using predefined MeSH terms related to inhalational exposures (e.g., PM, tobacco smoke, vaping, marijuana) and upper gastrointestinal diseases (e.g., GERD, BE, peptic ulcer disease (PUD), esophagitis, esophageal cancer (ECa), gastric cancer (GCa)). Our search strategy included automatic database filters (full-text, human subjects, English language, publication within the last 10 years) and manual reference-list screening to identify relevant studies.

The following MeSH Terms were searched for using the MeSH Database:

figure aThe alternative text for this image may have been generated using AI.

The complete listing of associated terms that were searched for with each of the above MeSH terms can be found in the MeSH Database. When searching for “stomach diseases” under its associated MeSH term, associated terms Reflux, Gastritis, Peptic Ulcer, Stomach Neoplasms, Zollinger-Ellison syndrome, etc. were included in search results. The reference-list screening was also used.

For this review, we have defined environmental exposure to include PM2.5, PM10, tobacco/cigarette smoke, vape/e-cigarette aerosols, and marijuana/cannabinoid inhalation. We have defined esophageal and gastric disease to include the following: GERD, BE, PUD, esophagitis, gastritis, ECa, and GCa.

Study criteria

Studies were included if they (1) discussed the environmental/occupational exposure to inhalants, specifically, PM2.5, PM10, tobacco/cigarette smoke, marijuana smoke, and/or vape/e-cigarette vapor, (2) evaluated effects of exposures on esophageal or gastric diseases, (3) performed on adult human subjects, (5) were written in English, and (6) were published within the last 10 years. Studies were excluded if they (1) were not original research, (2) consisted of translational research, (3) were case reports or series, (4) were conference abstracts, or (5) were conducted on a pediatric population.

Data extraction

Articles were reviewed and data regarding study design, patient characteristics, sample size, exposures, and outcomes were extracted. Results from database searches were filtered for full-text articles, human subjects, English language, and publication date and imported into Endnote X9. Original research papers were reviewed (title, abstract, and full text) to ascertain eligibility. We examined references cited in the relevant articles. All results were screened by Daniel Hyun Kim, Aida Fallah Zadeh, Tara Mahmoodi, and Sanjiti Podury and further independently evaluated by Anna Nolan. Disagreements were resolved by consensus.

Risk of bias (RoB) assessment

Systematic review inherent biases (selection, detection, performance, and reporting) were addressed through the study design/search algorithm. Selection bias was addressed by having pre-determined inclusion criteria, exclusion criteria, and distinct definitions. Detection and performance bias were addressed by having at least two rounds of screening individually performed by Daniel Hyun Kim and Sanjiti Podury. Reporting bias was minimized by using PubMed and WoS search filters for peer-reviewed published articles of human subjects written in English and removing duplicates.

The Newcastle-Ottawa Scale (NOS), a domain-based approach was used to assess the degree of bias32,33. Scales adapted for case-control and cross-sectional studies were used. Total scores obtained by the scale were converted to Agency for Healthcare Research and Quality (AHRQ) standards or as done in previous studies to reflect the quality of each paper: low-risk studies were concordant in all domains (green); studies with at least one unclear or high-risk domain were considered as unclear or high risk of bias studies (yellow or red), respectively34,35,36,37. Briefly, cohort studies were assessed for three key domains of interest: (1) Assessment of Outcomes. (2) Comparability and (3) Selection. Case Control Studies included were assessed for three key domains of interest (1) Selection (2) Comparability and (3) Exposure. Finally, Cross-Sectional Studies were assessed for the key domains of (1) Selection (2) Comparability and (3) Outcome, Supplemental Table 7 A-C (details for each of these criteria may be found in the footnote of each table).

Meta-analysis

Meta-analysis was performed (MetaAnalysisOnline.com). The platform supports various models for data analysis, including fixed-effects and random-effects models, depending on the heterogeneity of the included studies. For each study, adjusted odds ratio (aOR) and 95% confidence intervals (CI) were extracted. If a study had no aORs, it was excluded from the meta-analysis.

Heterogeneity across studies was assessed using the I² statistic to determine the appropriate model for analysis and the random-effects model was applied to account for variability between studies. A forest plot was generated for each outcome, to visually represent the individual study effect sizes and the overall pooled effect. Sensitivity analysis was also conducted to assess the strength of the findings by evaluating the impact of each study on the overall effect size. The effect of studies with high weight (%) and with a large effect size were studied.

Ethics approval

This study does not require ethics approval as it involves a review of publicly available research and utilized anonymized original data.

Results

Literature search

Our PubMed and WoS searches identified N = 764 and N = 1,036 studies, respectively, Fig. 1. After the removal of 222 and 102 duplicates from our PubMed and WoS searches, respectively, 542 PubMed articles and 934 WoS articles were screened. Following the application of inclusion criteria, 216 articles from PubMed and 626 articles from WoS were excluded and 326 PubMed articles and 308 WoS articles were assessed for eligibility based on exclusion criteria. Application of exclusion criteria involved the removal of 238 (141 from PubMed, 97 from WoS) non-original research articles, 211 (85 from PubMed, 126 from WoS) translational studies, 51 (27 from PubMed, 24 from WoS) case reports/series, and 14 (9 from PubMed, 5 from WoS) pediatric studies for a total of 514 (262 from PubMed, 252 from WoS) articles. 64 original PubMed and 57 original WoS research articles were considered eligible. After the removal of 10 duplicates between the two database searches, N = 111 studies were included in this review, Table 1. Data from screening and extraction are available, Supplemental Tables 1–6.

RoB using NOS was assessed in cohort (N = 39), case-control (N = 39), and cross-sectional studies (N = 29), Supplemental Table 7. Two case-crossover studies, an ecological study, and a time-series study were unable to be assessed for RoB as the NOS and our adaptations did not cover these types of studies. Scores obtained from the NOS were adapted as in previously published studies to reflect the quality of each paper34. Cutoffs for each risk of bias assessment depending on article type can be found within the footnote of Supplemental Table 7. Among cohort studies, N = 33 articles were of good quality, N = 1 of fair quality, and N = 5 of poor quality. Among case-control studies, N = 22 were of good.

Table 1 Study characteristics (N = 111).
Full size table

Quality, N = 8 of fair quality, and N = 9 of poor quality. Among cross-sectional studies, N = 23 were of good quality, N = 4 of satisfactory quality, and N = 2 of unsatisfactory quality.

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Study design per preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines. PRISMA is an evidence-based minimum set of items for reporting in systematic reviews and meta-analyses30,31.

Study characteristics

The populations of patients with esophageal or gastric disease included those afflicted with esophagitis (n = 8), BE (n = 8), ECa (n = 24), GERD (n = 11), PUD (n = 9), and GCa (n = 16). Studies that focused on any other outcomes did not meet exclusion/inclusion criteria. The investigated exposures were smoking, waterpipe smoking, and PM2.5/PM10 exposure. While there were no studies that focused on marijuana smoking or vaping/e-cigarettes that met our inclusion/exclusion criteria we know from the literature that the use of cannabinoids and vaping are linked to the development of gastrointestinal disorders149. One study investigated the role of exposure to second-hand smoke, in addition to direct cigarette smoke exposure70. Most studies produced an odds ratio (OR), risk ratio (RR), correlation coefficient (CC), or hazard ratio (HR) to measure each of the risks associated with their respective exposures for a particular outcome, which are summarized in Fig. 2 (see raw data in Supplemental Table 8 A–E). Among those studies, some reported using adjusted models in their analyses. Additionally, other studies focused on the percent presentation of risk factors;40,53 risk by measuring the increase in incidence of the respective disease;73,92,93,94 the differences in mortality with respect to magnitude of exposure95and utilized a novel predictive model to identify risk factors, Table 198.

Esophagitis

Current tobacco use was identified as a significant risk factor for reflux esophagitis (RE)41,44,58,65,115,122,134. Some studies focused on specific groups of patients and found that smoking was associated with RE among COPD patients124 and liver cirrhosis patients130. When studying gender-specific differences between smoking and risk of RE, Kim et al. found that smoking led to greater risks of RE among women compared to men51, whereas Wang et al. found that smoking led to greater risks of RE among men141. Lee et al. identified smoking as a significant risk factor for asymptomatic erosive esophagitis (EE)54. Associations between smoking and eosinophilic esophagitis (EOE) were also investigated. One study found that those with EOE were significantly less likely to have ever smoked cigarettes compared to non-EOE controls, but smoking was not significantly associated with an increased risk of EOE53.

Our meta-analysis of 8 studies41,53,54,65,115,124,130,141 revealed that inhalational exposures were significantly associated with an increased risk of esophagitis with a pooled estimate of 1.32 (95% CI 1.06–1.65; I2 = 86%), Fig. 3A. In our sensitivity analysis, we excluded one study53 that had high heterogeneity; however, the analysis results revealed no significant differences with a pooled estimate of 1.43 (95% CI 1.15–1.78; I2 = 86%), as shown in Supplementary Fig. 1 (1.3A vs 1.3A’).

Fig. 2
Fig. 2The alternative text for this image may have been generated using AI.
Full size image

Overview of data synthesis: (A) Summary of odds ratios for esophageal diseases (esophagitis, BE, ECa), (B) Summary of odds ratios for gastric diseases (GERD, PUD, GCa), (C) Summary of risk ratios, (D) Summary of correlation coefficients, (E) Summary of hazard ratios.

Fig. 3
Fig. 3The alternative text for this image may have been generated using AI.
Full size image

Meta-analysis of the association between inhalational exposures and upper gastrointestinal diseases: (A) esophagitis, (B) GERD, (C) BE, (D) PUD, (E) ECa, (F) GCa.

Gastroesophageal reflux disease

Multiple studies identified smoking as a risk factor for GERD43,45,57,59,72,102,105. Kim et al. interestingly found that former smoking was significantly associated with risk of GERD, while current smoking was not significantly associated50. When investigating gender-specific differences in the effects of smoking on the risk of GERD, Kim et al. found that smoking increased risks in both men and women51. One study investigated the effects of waterpipe smoking in addition to traditional cigarette smoking on the risk of GERD. Etemadi et al. found that waterpipe smoking was most strongly associated with “severe and frequent” reflux, and the prevalence of the disease was associated with waterpipe use and duration. In addition, they found that cigarette smoking was a significant risk factor for any form of reflux among men87. Additionally, Wang et al. identified smoking and male gender as significant risk factors for reflux esophagitis, a subtype of GERD141. Similarly, a significant association between smoking and increased risk of RE has been reported in multiple studies122,124,130,134While other researches has found no significant association between GERD and smoking103. Almadi et al. observed a higher prevalence of GERD among smokers than non-smokers, but found no significant difference38. Wang et al. also did not find any association between cigarette smoking and risk of GERD79. Seo et al. developed a prediction model that was significantly able to predict GERD-related medical utilization in the South Korean population and identified PM2.5 as a risk factor for GERD98.

Our meta-analysis of 7 studies43,45,72,102,133,138,140 revealed that inhalational exposures were significantly associated with increased risk of GERD with a pooled estimate of 1.71 (95% CI 1.14–2.55; I2 = 94%), Fig. 3B In our sensitivity analysis, we excluded one study102 that had high heterogeneity; the analysis results revealed a reduction of I2 after removing Ahmed 2020 showing differences with a pooled estimate of 1.34 (95% CI 1.06 to 1.68; I2 = 78%), as shown in Supplementary Fig. 1 (1.3B vs 1. 3B′).

Barrett’s esophagus

Smoking was identified as a risk factor for BE41,58,142. Schmidt et al. found that BE cases were significantly more likely to smoke73. Navab et al. found a positive correlation between current and prior tobacco use and BE63. Etemadi et al. found associations between smoking and BE that were independent of intensity, age at initiation, and GERD, but dependent on duration and years since cessation87. Other studies, however, produced conflicting results: some studies found that current and former smoking were not significantly associated with BE43,84. Additionally, Arroyo-Martínez et al. found that cigarette smoking showed no significant association with any form of the progression of BE106.

Our meta-analysis of 4 studies revealed that inhalational exposures were not significantly associated with risk of BE with a pooled estimate of 0.93 (95% CI 0.65–1.34; I2 = 76%), Fig. 3C. Sensitivity analysis was not performed since study weight and effect size were not concerning.

Peptic ulcer disease

Ghanadi et al. found that smoking history was significantly associated with PUD116. Similarly, Chuang et al. also identified current tobacco use as a significant risk factor for PUD and that higher cumulative amounts of tobacco use were at higher risk for PUD41. Further, Begovic et al. found that more than half of ulcer patients enrolled in their study were smokers, and this difference was significant when compared to those who were non-smokers40. Levenstein et al. observed that age-, gender-, and socioeconomic status-adjusted associations were significant for smoking55. Park et al. investigated the role of changes in smoking status in the risk of gastroduodenal ulcer68. They observed that changes in smoking status, particularly from never smoker or former smoker to current smoker, had relatively higher HRs than other groups. When comparing smoking amount levels, they found that smokers who smoked > 20 pack-years had a significantly higher risk of PUD.

The role of PM exposure as a risk of PUD has been studied. Tsai et al. found that increases in both PM2.5 and PM10 were significantly associated with increased risk of PUD hospitalizations on warm days, but only PM10 was significantly associated with cold days99. Similarly, Wong et al. found that PUD hospitalization was associated with 10 ug/m3 increases in PM2.5. When investigating different types of ulcers, they found that associations with PM2.5 were significant for gastric ulcers, but not for duodenal ulcers91. Wu et al. observed that cumulative RRs for PM2.5 and PM10 showed nearly linear adverse effects100. When looking at gender-adjusted differences, significant associations for men and women were only observed for PM2.5. Quan et al. found that when air pollution exposures were assessed over 3-, 5-, and 7-day averages, pollutants were inversely associated with upper gastrointestinal bleeding (UGIB)96. Yu et al. observed a potential dose-response relationship between quartile concentrations of PM2.5 one month before detection of PUD. Subjects in the highest quartile of PM2.5 exposure displayed significantly higher risk, and the detection of PUD was associated with a 10 ug/m3 in PM2.5101. Our meta-analysis of 5 studies41,96,99,101,116 revealed that inhalational exposures were significantly associated with increased risk of PUD with a pooled estimate of 1.21 (95% CI 1.03–1.43; I2 = 93%), Fig. 3D. Sensitivity analysis was not performed since the study with a different effect size had a low weight.

Esophageal cancer

Smoking as a risk factor for ECa was the focus of several studies62,74,86,104,107,108,120,123,126,127,131,132,136. Other studies focused on esophageal squamous cell carcinoma and also identified smoking as a risk factor and this risk increased with tobacco intensity and smoking duration131but there was no significant difference with respect to macroscopic type of cancer, as smoking showed similarly increased risks for both ulcerative type and medullary type eosinophilic squamous cell carcinoma (ESCC)48,52,61,75,81,82. Jayalekshmi et al. observed higher risks of ESCC for current bidi and cigarette smokers47. In gender association studies, Chen et al. found that there was a significant association between smoking and ECa in men, but not in women110. Etemadi A et al. studied how among tobacco users, metabolites of styrene and xylene were associated with ESCC, showing how specific components of tobacco smoke could contribute to disease112.

Conversely, some studies observed a non-significant relationship between inhalational exposures and ECa41,66,83,137. Arroyo-Martínez et al. found that cigarette smoking showed no significant association with any forms of progression of EAC, but in all forms of progression, smoking showed the highest association in the progression of BE without dysplasia to EAC106. Interestingly, Guo et al. found that current smoking was not significantly associated with any esophageal neoplasms while former smoking was106,118. However, Sheikh M et al. demonstrated that smoking cigarettes was not associated with the risk of ESCC139. Pournaghi et al. found a significant association between hookah smoking and an elevated risk of ESCC. Additionally, the same study identified that individuals who had quit smoking more than 10 times had a higher risk of ESCC137. Furthermore, a study on cumulative tar exposure revealed that individuals in the second tertile had an increased risk of ESCC, although this association became non-significant after adjusting for pack-years due to the high correlation between cumulative tar and pack-years132. Some studies looked at how smoking affected survival for those afflicted with ECa. Spreafico et al. found that smoking conferred worse overall survival in the combined Boston-Toronto Cohort for each 20 pack-year increase77. Others observed how current and former smoking contributed to decreased survival with respect to subtypes, specifically ESCC and EAC80,84. Dighe et al. found that current smoking was significantly associated with mortality of stage I-III ECa111. There were studies, however, that undermined this relationship121.

One study in particular, Rafiq R et al. evaluated both smoking and second-hand smoke as a risk factor for ECa, with increased risks associated with either exposure70. Pan et al. conducted two studies on the association between smoking and esophageal precancerous lesions (EPL), finding that smoking more than 20 cigarettes a day or having 40 or more pack-years of cumulative smoking was significantly associated with an elevated risk of EPL67,135. Cigarette smoking was also associated with the development of esophageal neuroendocrine neoplasms (NEN) and the risk increased to 4.7 times for all cases and to 4.0 for pure NEN cases143.

The relationship between PM2.5 exposure and ECa was also assessed. Li et al. observed a significantly positive association between PM2.5 and ECa incidence. When investigating the corresponding lag effects on ECa incidence, they found that a lag effect of 4 years showed the greatest risk for the overall population92. Li et al. examined the modifying effects of urbanization and socioeconomic factors and found a stronger association between PM2.5 and incidence for low urbanization groups, and this association was stronger for females than males93. Some studies identified long-term exposure to various components of PM2.5 such as black carbon, organic carbon, nitrate, sulfates, chlorides, and ammonium to be significantly associated with ECa94,148. Rao et al. found that although spatial distributions of the hospitalization rate of ECa in 2016 were not consistent with that of PM2.5 concentration in the same year, concentrations of PM2.5 in 2003 and 2004 had the strongest correlations with the hospitalization rate in 201697. Sun D et al. observed a linear concentration-response relationship between long-term PM2.5 and ECa90. Huang et al. found that PM10 and NO2 had significant linear correlations with ECa mortality rates147. Our meta-analysis of 20 studies41,61,66,67,70,75,81,82,83,86,107,108,118,123,126,131,132,135,136,143 revealed that inhalational exposures were significantly associated with increased risk of ECa; pooled estimate of 1.83(95% CI1.54–2.18; I2 = 73%), Fig. 3E.

Gastric cancer

As with the previous outcomes, most studies identified smoking as a risk factor for GCa113,117,119,125,127,144. Increased risk of GCa was associated with current cigarette smoking status, longer durations of smoking (at least 20 years) or later starting ages of smoking56. Current smoking was also found to have an increased risk of mortality from stomach cancer60. When assessing changes in smoking status, one study found that those who changed their current status to “smoking” showed an increased risk of GCa, and this risk was the highest in heavier smokers69. One study found that smoking was only significantly associated with single GCa and synchronous multiple gastric cancer in advanced gastric cancer patients76.

Current smoking also showed increased risk for gastric adenocarcinoma and gastric non-cardia adenocarcinoma80. Additionally, another study highlighted that smoking history was significantly associated with an increased risk of gastric adenocarcinoma125. Interestingly, Jayalekshmi et al. found that bidi smoking was significantly associated with GCa risk, but cigarette smoking was not. This risk increased with the number of bidis smoked daily and with the duration of bidi smoking46. Conversely, Chuang YS et al. found that tobacco use was a non-significant risk factor for GCa41. Other studies found that current smoking increased risks of intestinal metaplasia for both men and women. Further, this risk increased with increasing duration and total dose49,78. Chen et al. found that smoking significantly increased the risk of progression from chronic atrophic gastritis to intestinal metaplasia109.

In gender association studies, Chen ZM et al. found that there was a significant association between smoking and esophageal cancer in men, but not in women110.

Some studies investigated the role of waterpipe smoking in GCa risk. Several studies in Vietnam showed that waterpipe smoking was positively associated with GCa risk, but there was no significant interaction between the effects of water pipes and cigarette smoking on GCa risk64,88,89. Additionally, Environmental Tobacco Smoke (ETS), commonly known as secondhand smoke, has also been identified as a potential risk factor for GCa, suggesting that passive exposure to tobacco smoke may contribute to increased risk128.

Still, other studies found no such significant relationship between smoking and GCa114,129. Interestingly, Guo et al. found that current smoking was not significantly associated with any gastric neoplasms while former smoking was118.

Other studies investigated the relationship between PM and GCa. Some studies identified long-term exposure to various components of PM2.5 such as black carbon, organic carbon, nitrate, sulfates, chlorides, and ammonium to be significantly associated with GCa94,148. Ethan et al. found that as a single pollutant and as a multipollutant in combination with NO2, PM2.5 was significantly associated with stomach cancer mortality145. Fan et al. found that each 1ug/m3 increment in PM2.5 exposure was significantly associated with an increase in the risk of GCa146. Our meta-analysis of 13 studies41,56,64,71,78,85,86,109,113,114,118,129,144 revealed that inhalational exposures were significantly associated with increased risk of GCa with a pooled estimate of 1.71 (95% CI 1.39–2.10; I2 = 94%), Fig. 3F. Sensitivity analysis was not performed for ECa and GCa there were a relatively large number of studies and therefore the their moderate heterogeneity was to be expected.

Discussion

In this systematic review, we investigated the associations between environmental exposures and diseases of the upper gastrointestinal tract. Through a comprehensive review of the available literature, we identified complex relationships between environmental exposures and upper gastrointestinal diseases. Most of the studies showed that exposures including PM, smoking, and waterpipe use were significantly associated with a higher risk of aerodigestive diseases. Based on the meta-analysis results, inhalational exposures were significantly associated with an increased risk of GERD, PUD, GCa, ECa, and esophagitis.

PM exposure is a global cause of significant pulmonary morbidity and mortality7,13,150,151,152,153,154,155,156,157. Our review supports existing evidence suggesting that exposure to PM may also increase the risk of diseases affecting the upper gastrointestinal tract. Studies included in this review demonstrated links between PM exposure and an increased risk of ECa and PUD, although the underlying mechanisms remain to be fully explained. These findings highlight the importance of considering environmental factors, such as air pollution, in the context of upper gastrointestinal health. PM consists of various harmful compounds that can trigger inflammatory responses, oxidative stress, and DNA damage that contribute to the development of cancer and ulceration. Moreover, studies showed that PM may disrupt the gut microbiota, leading to an increased risk of gastrointestinal inflammation and cancers158.

Cigarette smoking has been recognized as a major risk factor for various cancers, including those of the gastrointestinal tract. Consistent with previous research, our review highlights the detrimental effects of smoking on the upper gastrointestinal tract, with a notable association observed between smoking and an elevated risk of BE, GCa, ECa, and PUD. The carcinogenic effect of smoking is attributable to mutations in critical genes caused by tobacco metabolites and chemicals. Smoking is also associated with progression, aggressiveness, and reduced survival rates of existing gastrointestinal cancers. Smoking may be associated with exacerbation of GERD symptoms due to reducing esophageal sphincter tone and increasing gastric acid production159.

Waterpipe smoking has increased worldwide due to a perception that it is less harmful than cigarette smoking. However, waterpipe smoke contains tobacco and several toxicants that may increase the risk of developing aerodigestive disease, as identified in our review. Numerous carcinogens have been identified in waterpipe smoke including polycyclic aromatic hydrocarbons, volatile aldehydes, and heavy metals, which can cause DNA damage and develop cancer over time160. Moreover, emerging evidence suggests that vaping and marijuana use may also impact gastrointestinal health, although further investigation is warranted to better understand the nature of these associations.

Although this review was unable to identify papers identifying risks of gastrointestinal disease from vaping, e-cigarette/ vaping use has been associated with nonspecific symptoms including nausea, vomiting, diarrhea, and abdominal pain161,162. One case report identified non-respiratory complaints including nausea, vomiting, and fever as associated symptoms of E-cigarette or Vaping Product Use–Associated Lung Injury (EVALI)163. Nausea was also more frequently associated with cannabinoid-based vaping, but was not affected by concurrent smoking and vaping164.

Nitrosamines are potent carcinogens from tobacco products that contribute to esophageal and stomach cancer from smoking165,166. E-cigarettes also contain the same nitrosamines that directly cause DNA damage at relatively lower concentrations, but have been shown to be sufficient in inducing lung cancer and bladder hyperplasia in in vitro and murine translational models167. However the lag time of at least two decades for smokers to develop cancer makes it difficult to ascertain the full carcinogenic threat of vaping in humans.

Furthermore, the nicotinic acetylcholine receptor (nAChR), a genetic variant of which is consistently linked to lung cancer in large genetic studies, might mediate carcinogenesis through directly binding nicotine (and nitrosamines) in airway epithelium. This mechanism could provide direct carcinogenesis of nicotine and nicotine metabolites to all cells that express the nAChR, particularly in carriers of the variants that are associated with tobacco smoking and cancer168,169,170,171. Following the idea that inhaled nicotine could produce carcinogenic molecules in human users, an untargeted metabolomics analysis of urine demonstrated a trend of increased carcinogen biomarkers in the samples of a relatively small cohort of vapers (n = 34 vs. n = 45 non-users)172.

Limitations

While this systematic review provides valuable insights into the associations between environmental exposures and upper gastrointestinal diseases, the included studies vary in design, methodology, and population characteristics. Some studies used adjusted models when calculating ORs or HRs (aOR; aHR), which varied in complexity and contributed to high heterogeneity. To address this problem, we used a random-effects model to account for variability between studies. However, the considerable heterogeneity, evident from a high I2may limit the generalizability of the pooled effect sizes. Factors such as differences in study designs, populations, interventions, and outcome measures may have contributed to this variability.

Additionally, many studies are observational thus, limiting causal inference and necessitating further research, including prospective cohort studies and mechanistic investigations. Furthermore, studies were required to be in English, which limited access to data from articles in languages in other languages. While this study included both PubMed and WoS databases for the identification of potentially eligible studies we understand that there are several databases that while complementary may have additional studies. Our risk of bias assessment (NOS) was able to evaluate the majority, but not all studies assessed in this review.

Other limitations revolved around how we defined environmental exposure and aerodigestive disease. Our study defined environmental exposures as air pollution in the form of PM, cigarette/tobacco smoke, marijuana smoke, and vape/e-cigarette aerosols. Due to this, it was not possible to completely cover the entire scope of environmental exposures that afflict society. In addition, our definition of aerodigestive disease focused on diseases of the upper gastrointestinal tract, which comprised esophagitis, GERD, BE, PUD, and esophageal/GCa based on the articles we found. There are likely other aerodigestive diseases that interact strongly with environmental exposures that were not covered by this paper. Due to these definitions and our inclusion/exclusion criteria, we also found no eligible articles that investigated the interactions of marijuana smoke and vape/e-cigarette aerosols with aerodigestive disease.

Future research

Our search was unable to identify human studies that clearly define the carcinogenic potential of non-burning nicotine delivery products such as e-cigarettes. However, this could be due to the very extensive lag time between carcinogen exposure and clinical cancer diagnosis in humans. Future studies could expand our definitions to account for interactions that have not been included in this review, such as those of marijuana smoking, vaping, and those of the lower intestinal tract. Additionally, future studies could include occupational exposures as this study focused on the general population. Such additional exposures include asbestos, synthetic fiber dust, chrysotile dust, nephrite, and potentially harmful elements which are all commonly present in mining or textile industries and developing societies. Such investigations could yield valuable insights for those whose occupation or geographic location puts them at risk for such diseases, as aerodigestive disease is often underrecognized in those populations. In addition, this could identify how specific exposures incite disease in various cohorts.

Conclusion

The implications of these findings are significant from both a public health and clinical perspective. Efforts to reduce exposure to environmental pollutants, such as PM, could potentially mitigate the burden of upper gastrointestinal diseases in affected populations. Similarly, targeted interventions aimed at reducing smoking behavior and promoting smoking cessation may help reduce the incidence of gastrointestinal disease and malignancy. Furthermore, continued research into the potential health effects of emerging trends, such as vaping and marijuana use, is crucial for informing preventive strategies and improving patient outcomes.

This review provides support for the connection between environmental exposures and digestive health, which is especially important considering that those who have been exposed to environmental/occupational inhalants are generally not screened for gastrointestinal disease as part of their exposure. We hope that this review will promote further recognition of the treatment of digestive disease with inhalational exposure.

In conclusion, this systematic review contributes to our understanding of the interplay between exposure to inhalational exposures and diseases of the upper gastrointestinal tract. By analyzing existing evidence and identifying knowledge gaps, this study highlights the need for approaches to address environmental risk factors and promote gastrointestinal health.