Introduction

ARDS is a prevalent clinical syndrome, characterized by acute respiratory failure resulting from diffuse pulmonary inflammation and edema. Characterized by hypoxemia and evident changes in lung imaging via chest X-rays or CT scans, ARDS is associated with a high incidence and mortality [1]. ARDS constitutes a significant health burden globally, accounting for 10.4% of ICU admissions, and approximately 23% of these patients necessitate mechanical ventilation [2]. Despite numerous advancements in treatment, the mortality rate of ARDS remains around 30% as of 2021 [3]. Consequently, emphasis has shifted toward ARDS prevention, highlighting the need to identify and manage modifiable risk factors to decrease its incidence.

The clinical precursors to ARDS encompass a range of clinical conditions including sepsis, pneumonia, aspiration, trauma, surgical procedures, pancreatitis, transfusions, and inhalation injuries from smoke or toxic gases [4]. Despite this understanding, there is still a limited grasp of the environmental factors contributing to ARDS, highlighting the need for additional research in this area [5]. A pivotal 1996 cohort study initially linked smoking with an elevated risk of ARDS [6]. Subsequent research on various patient demographics has underscored this connection, though with inconsistent findings. A 2014 narrative review identified smoking as a key environmental risk factor for ARDS after analyzing three studies [6,7,8,9]. However, robust epidemiological evidence linking smoking to ARDS risk remains sparse [10]. Our systematic review and meta-analysis of observational studies aimed to consolidate recent findings, elucidating the relationship between smoking and ARDS risk.

Methods

Protocol and registration

This systematic review and meta-analysis adhered to the PRISMA guidelines (http://www.prisma-statement.org/) and was registered on PROSPERO (CRD42023483876). (Additional file 1: Table S1).

Search strategy

We conducted a comprehensive search of PubMed, EMBASE, the Cochrane Library, and Web of Science for studies published between 2000 and 2023 without language restrictions. The search terms used included all possible combinations of ARDS and smoking. A copy of our search strategy could be found in our online Supplementary Materials (Additional file 1: Table S2). We screened the reference lists of relevant systematic reviews and meta-analyses to ensure that we did not miss any additional articles.

Study selection criteria

Participants

Patients aged 18 or older with clinical risk factors for ARDS were included.

Exposure

We included studies about assessing smoking via clinical records or laboratory tests. The smoking history (including smokers, nonsmokers, current smokers, former smokers, and never smokers) was extracted from clinical records. The laboratory tests included the detection of tobacco product metabolites, such as serum cotinine (nonsmokers < 0.02 ng/ml, passive smokers 0.02–3.08 ng/ml, active smokers > 3.08 ng/ml) and urine NNAL (nonsmokers < 0.25 pg/mg, passive smokers 0.25–47 pg/mg, active smokers > 47 pg/mg) [11].

Outcomes

Studies with outcomes for ARDS were included. Studies limited to other diagnoses were excluded.

Study design

We included observational cohort and case‒control studies; excluded reviews, animal and cell studies, abstracts, comments, case reports, and cross-sectional studies. In case of duplicated data, the study with more complete and detailed data was chosen. We also excluded studies in which a 2 × 2 table between exposure to smoking and the outcome of ARDS could not be constructed. Two independent reviewers (LJZ and JHX) screened the titles and abstracts, and the full texts of the studies that met the inclusion criteria were obtained. Any disagreements encountered during the screening process were resolved through discussion and, if necessary, with the assistance of a third reviewer (YL).

Data extraction

The data were independently extracted by two reviewers (LJZ and JHX) with electronic spreadsheets. The extracted information included the first author's name, publication year, patient data source, ARDS diagnostic criteria, sample size, clinical risk factors for ARDS, methods of obtaining smoking status, adjustments for confounding factors, smoking status, incidence rates of ARDS, age, injury severity scores (ISSs), and mortality rates. Smoking status was categorized into two groups: smokers (experimental group) and nonsmokers (control group). In the analysis, current smokers, former smokers, active smokers, and passive smokers were all categorized into the smoking group. The method for unknown non-normal distributions was applied to transform the median and quartile into the mean and standard deviation (SD), respectively. Graphical data from the original studies were extracted by using WebPlotDigitizer, a semiautomated tool that extracts underlying digital data by reverse engineering a visual image of the data. The extraction forms were compared, and disagreements were resolved firstly by discussion or with a third reviewer (YL) if a consensus could not be reached.

Quality assessment

The quality of the studies was assessed by the Newcastle–Ottawa Scale. High quality was defined as a grade > 6. Both cohort and case‒control studies had a maximum score of 9. Two independent reviewers assessed the quality of the included studies, and any disagreements were resolved through discussion.

Statistical analysis

Data synthesis was conducted with Review Manager software version 5.4 (Cochrane Collaboration). We reported ORs for dichotomous outcomes and MDs for continuous outcomes, with a standard 95% CI. To assess heterogeneity among the studies, we employed the Cochrane Q test and used the I2 statistic to measure the degree of heterogeneity. High heterogeneity was inferred at I2 > 50%, prompting the use of a random-effects model; otherwise, a fixed-effects model was applied. Subgroup analysis focused on clinical risk factors, assessment methods of smoking status, and diagnostic criteria of ARDS. Publication bias was evaluated by using Egger's test, and complemented by visual inspection through funnel plots. We conducted sensitivity analyses using two approaches. Firstly, we excluded individual studies sequentially to assess their individual impact on the overall results. Secondly, in a similar manner, we excluded different subgroups of clinical risk factors to gauge their specific effects on the stability of our findings. Publication bias and sensitivity analyses were conducted with Stata/MP 17.0 (StataCorp LLC, Texas, US). A P value < 0.05 indicated statistical significance.

Results

Literature search process

Database searches and reference lists yielded a total of 3763 articles. After the removal of duplicates, we identified 1598 articles for title and abstract screening, from which we identified 78 articles for full text review. Ultimately, 26 of these articles met the inclusion and exclusion criteria. The process of literature selection is detailed in Fig. 1.

Fig. 1
figure 1

Preferred reporting items for systematic reviews and meta-analyses (PRISMA) chart. The process of literature selection

Full size image

Study characteristics

This review included 26 studies involving a total of 36,995 participants; the results are detailed in Table 1. The majority (25 studies) [10, 12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30, 32,33,34,35,36] were cohort studies, and one was a case‒control study [31]. All the studies were conducted under hospital circumstances. Quality assessment using the Newcastle‒Ottawa Scale revealed that all 26 studies were of medium to high quality. Information on smoking history was extracted from 22 studies to assess the relationship between smoking and ARDS, while four studies assessed patients' exposure to smoking through the detection of tobacco product metabolites. Various ARDS diagnostic criteria were used, including the Berlin definition, the American-European Consensus Conference definition (AECC), the International Classification of Disease-10 standards (ICD-10), the fifth edition of the Indonesian COVID-19 guidelines, and criteria set by the Chinese Medical Association's Critical Care Medicine Branch in 2006.

Table 1 The characteristics of the included studies
Full size table

Meta-analysis

Smoking and ARDS risk

Analysis of 26 studies revealed a significant association between smoking and increased ARDS risk (OR 1.67; 95% CI 1.33–2.08; P < 0.001) (Fig. 2).

Fig. 2
figure 2

ORs of ARDS between smokers and nonsmokers

Full size image

Smoking status assessment methods and ARDS risk

In 22 studies, ARDS risk was evaluated based on smoking history, and the results showed a substantial increase in risk generally similar to that reported in all the studies (OR 1.78; 95% CI 1.38–2.28; P < 0.001). Conversely, four studies utilizing tobacco metabolite detection did not show a notable increase in risk (OR 1.19; 95% CI 0.69–2.05; P = 0.53) (Fig. 3).

Fig. 3
figure 3

a ORs of ARDS between smokers and nonsmokers assessed by smoking history. b ORs of ARDS between smokers and nonsmokers assessed by tobacco metabolites

Full size image

Smoking cessation and ARDS risk

Among the five studies differentiating current smokers from former smokers, smoking cessation did not significantly reduce ARDS risk (OR 1.05; 95% CI 0.61–1.81; P = 0.87) (Fig. 4).

Fig. 4
figure 4

ORs of ARDS between current and former smokers

Full size image

Characteristics of the smoking population in ARDS risk studies

Seven studies presented age data, highlighting that smokers in populations with clinical risk factors for ARDS were younger than nonsmokers (MD − 7.15; 95% CI − 11.58 to − 2.72; P = 0.002). Among the trauma cohort studies, three provided ISS data for smoking and nonsmoking populations, indicating that smokers had lower ISSs (MD − 1.95; 95% CI − 3.50 to − 0.39; P = 0.01). Additionally, three studies compared mortality rates between smoking and nonsmoking populations within the same group and revealed no significant difference (MD 0.72; 95% CI 0.37–1.43; P = 0.35) (Fig. 5).

Fig. 5
figure 5

a MDs of age between smokers and nonsmokers. b MDs of the ISSs between smokers and nonsmokers in trauma cohorts. c MDs of mortality between smokers and nonsmokers

Full size image

Subgroup analysis

By analyzing subgroups based on clinical risk factors, the study revealed a notably greater risk of ARDS in patients with sepsis, trauma or burns, and surgery (OR 1.75; 95% CI 1.17–2.62; P = 0.006 vs. OR 1.66; 95% CI 1.15–2.40; P = 0.007 vs. OR 6.37; 95% CI 1.12–36.13; P = 0.04) (Additional file 1: Fig. S1). When the pathogenic factors were divided into pulmonary (COVID-19, other pneumonias) and extrapulmonary (sepsis, trauma burns, surgery, acute pancreatitis, postcerebral hemorrhage) categories, smoking was significantly linked to an increased risk of ARDS from extrapulmonary causes (pulmonary factors OR 1.65; 95% CI 0.87–3.13; P = 0.13 vs. extrapulmonary factors OR 1.85; 95% CI 1.43–2.38; P < 0.001) (Fig. 6).

Fig. 6
figure 6

Subgroup analysis, comparing pulmonary and extrapulmonary risk factors

Full size image

Subgroup analyses based on ARDS diagnostic criteria and assessment methods of smoking status showed no significant difference between subgroups with low heterogeneity (Additional file 1: Figs. S4, S5).

Publication bias

Asymmetry in the funnel plot indicated potential publication bias. However, a trim and fill analysis of studies located on the left side of the funnel plot resulted in an OR of 1.442 (95% CI 1.167–1.782), suggesting that this bias did not substantially impact the overall findings' robustness (Additional file 1: Fig. S2).

Sensitivity analysis

The sensitivity analysis, conducted due to substantial heterogeneity (I2 = 82% > 50%), validated the stability of the meta-analysis results. The analysis showed that removing individual studies and each subgroup of clinical risk factors, independently, did not markedly alter the overall conclusions, thereby ensured the reliability of the combined data (Additional file 1: Fig. S3, Table S3).

Discussion

Adding value to this study, this is the first and most comprehensive meta-analysis focusing on the association between smoking and the risk of ARDS among observational studies. This study critically evaluated various research evidence and examined whether this risk could change with different ARDS pathogenesis factors. This highlighted how smoking significantly elevated the risk of ARDS, categorizing it as a notable environmental risk factor. The analysis particularly underscored the heightened risk in patients involving extrapulmonary pathogenic factors such as sepsis and trauma. Intriguingly, smoking increased ARDS risk even in younger patients with less severe pathogenic injuries. Furthermore, smoking cessation could not substantially alter the elevated risk, suggesting that the adverse impact of smoking on lung health was deeper and more lasting than previously understood.

Cigarette smoke, a complex mixture of gases and particles, exhibits oxidative, proinflammatory, and carcinogenic properties [37]. This composition adversely affects lung function through multiple mechanisms, such as elevating alveolar epithelial permeability, disrupting normal immune responses, and causing vascular endothelial damage, all of which are critical pathological and physiological mechanisms of ARDS [7, 38]. Research indicated that smoking-induced pathological alterations in the alveolar epithelium mirrored those observed in ARDS. Studies in cells and animals revealed that oxidative stress caused by cigarette smoke could negatively affect the integrity of alveolar epithelial cell connections and suppress the expression of alveolar ion channels [39, 40]. This damage could increase the likelihood of severe edema in the lungs when the patient was exposed to ARDS clinical risk factors. Additionally, smoking directly impacted both innate and adaptive immunity, potentially leading to a heightened risk of infections [41]. These effects included impaired mucociliary function, reduced clearance of bacteria by alveolar macrophages, decreased surfactant production, altered T-cell responses, NK cell function suppression, and decreased immunoglobulin levels [38, 42]. Smoking triggered the influx of neutrophils into the tissues surrounding blood vessels. These neutrophils, along with activated macrophages, released inflammatory cytokines and proteolytic enzymes, contributing to the inflammatory response [43]. Additionally, smoking facilitated biofilm formation, enhancing pathogen virulence, which played a critical role in escalating the risk of respiratory infections [44]. These impaired immune defenses and increased infection susceptibility constituted another potential mechanism through which smoking increased the risk of ARDS. Smoking was also known to cause endothelial injury and alter endothelial functions [45]. Oxidative stress caused by cigarette smoke could change the cytoskeletal structure of endothelial cells, leading to dysfunction of the endothelial barrier. In vitro studies showed that cigarette smoke extract exacerbated the endothelial dysfunction caused by lipopolysaccharides [46]. Moreover, smoking could enhance platelet activation, thereby exacerbating endothelial cell damage and triggering thrombosis [47]. The resulting pulmonary artery thrombosis contributed to microvascular filling defects, a key factor in abnormal gas exchange in ARDS patients. The intricate connection between smoking and ARDS still was an area of ongoing investigation. This uncertainty presented challenges in the development of effective treatment targets, highlighting the need for additional foundational research in this area.

As a clinical syndrome, ARDS arises from various causes, leading to a common pathological result: protein-rich pulmonary edema [3]. However, ample evidence suggested that ARDS caused by different etiologies exhibited heterogeneity in its pathological process. Both pulmonary and extrapulmonary risk factors were able to trigger ARDS, with each category influencing the disease's development and characteristics, respectively. Direct pulmonary injury-induced ARDS typically led to extensive damage and inflammation in the alveolar epithelium [48]. In patients with ARDS triggered by extrapulmonary risk factors, the damage extended beyond the lungs, often leading to severe endothelial injury and systemic inflammation. This form of ARDS was closely linked to endothelial dysfunction, where a compromised endothelial barrier allowed systemic inflammation to spread into the lungs. Consequently, this led to heightened alveolar epithelial inflammation and an exaggerated macrophage inflammatory response, playing a pivotal role in the manifestation of ARDS [1]. Taken together, our findings highlighted that smoking predominantly escalated the risk of ARDS stemming from extrapulmonary risk factors. These findings suggested that the impairment of endothelial barrier function, a consequence of smoking, could be a key factor in elevating ARDS risk. Emphasizing the reinforcement of endothelial barrier functionality was pivotal for the prevention of ARDS in patients with extrapulmonary causes, particularly in individuals who smoked. Such a focus could play a crucial role in diminishing the occurrence of ARDS in these susceptible populations.

Aging serves as both a risk factor for ARDS and a determinant of the severity of lung injury. However, our study revealed that within the same high-risk group for ARDS, the smoking population could tend to be younger. A multicenter study of ARDS patients exposed to cigarette smoke showed that younger smokers, despite having fewer instances of septic shock and better overall health, had similar levels of lung injury as older nonsmokers did [18, 49]. These findings suggested that younger, otherwise healthy individuals who smoke, were more susceptible to ARDS. The elevated risk of ARDS associated with aging could be attributed to a decrease in the body's defense mechanisms; a greater incidence of diseases such as pneumonia and overactivation of inflammatory pathways such as NF-κB [50]. Research findings indicated that neutrophils from elderly individuals exhibited nontargeted tissue migration, increasing release of primary granules, elevating neutrophil elastase activity, and led to heightened tissue inflammation [51]. These cellular changes contributed to a more inflammatory internal state in older adults, escalating their vulnerability to intense inflammatory reactions in response to lung injuries. A systematic review of animal studies also revealed a correlation between increased age and more severe lung injuries [52]. Basic research demonstrated that cigarette smoke could cause aging-like damage. Human lung fibroblasts exposed to cigarette smoke extract exhibited an aging-associated secretory phenotype, releasing a range of inflammatory cytokines and chemokines, thus fostering a proinflammatory milieu and encouraging immune cell recruitment [41]. In human bronchial epithelial cells exposed to cigarette smoke extract, there was an accumulation of mitochondrial fragments, reduced autophagy, heightened reactive oxygen species (ROS) generation, and cellular aging [53]. Smoke-exposed mice also exhibited similar patterns of impaired autophagy and cellular aging [54]. These findings clearly indicated that starting to smoke at an earlier age could lead to a larger pool of individuals being prone to ARDS. Additionally, our research revealed that cessation of smoking could not markedly lower this risk. This underlined the importance of focusing on early smoking prevention and cessation strategies to decrease ARDS risk, particularly in younger populations.

Measuring cigarette smoke exposure in epidemiological studies is challenging. Cohort studies evaluating the quantitative link between smoking and ARDS often measured tobacco metabolite levels (cotinine, NNAL) to identify exposed individuals [9, 15, 17, 18, 36, 55]. Previous studies indicated that in certain patient groups, such as those with blunt trauma injuries and nonlung sepsis, higher levels of tobacco product metabolites were identified as independent risk factors for ARDS [17, 18]. However, our meta-analysis indicated that the current thresholds for tobacco metabolite detection did not markedly elevate ARDS risk in identified smokers. In contrast, a noticeable increase in ARDS risk was observed in populations identified through inquiries about smoking history. These findings suggested that the existing thresholds for tobacco metabolite screening might be too low, including a large group with minimal exposure to cigarette smoke, which were not substantially affected. Therefore, revising these thresholds could lead to more accurate identification of high-risk ARDS patients in clinical practice.

The studies enrolled in our research did not provide data on e-cigarette vaping. Although e-cigarettes were often perceived as less harmful alternatives to traditional cigarettes, recent studies indicated that they could cause pulmonary inflammation, injury, and other pathological changes associated with ARDS [56, 57]. Similar to the traditional cigarettes, e-cigarette vaping could provoke inflammatory responses and oxidative stress, activating pulmonary macrophages and epithelial cells to release proinflammatory cytokines, leading to damage in alveolar epithelial and endothelial cells [58, 59]. Additionally, e-cigarette vaping might increase the risk of ARDS through unique mechanisms, such as toxic metabolites from flavorings and harmful compounds produced at high temperatures, causing cytotoxic responses and alveolar death [57]. Consequently, e-cigarette vaping might elevate the risk of ARDS. Future clinical research should consider e-cigarette vaping among patients to clarify its relationship with the incidence of ARDS.

This study had several limitations. Notably, our results exhibited high heterogeneity. We did extensive subgroup and sensitivity analyses to explain this issue. The results showed that clinical risk factors, assessment methods of smoking status, and ARDS diagnostic criteria were not the primary contributors to the observed heterogeneity. The inherent heterogeneity might be attributed to the characteristics of the observational studies, which often contained uncontrolled confounding factors. Regrettably, available information was too limited for further exploration. Future large-scale prospective studies were recommended to validate these findings. Additionally, the possibility of publication bias was considered, yet the statistical analysis suggested it did not affect the stability of our results. Moreover, two case–control studies linking smoking with a higher risk of transfusion-related lung injury were not included in our meta-analysis due to incomplete data [60, 61]. Finally, given the limited number of studies, the results should be interpreted with caution. Updated meta-analyses incorporating future research findings were recommended to provide a more comprehensive understanding of the topic.

Conclusion

In conclusion, this meta-analysis contributed further evidence that smoking increases the risk of ARDS.