Decoding the complexity: mechanistic insights into comorbidities in idiopathic pulmonary fibrosis

Graphical abstract

In this review, we provide a general overview of the relationship between idiopathic pulmonary fibrosis and its comorbidities, emphasising the pathogenic mechanisms, potential shared pathobiology and, where relevant, causal insights from Mendelian randomisation studies.

Abstract

The complex pathogenic relationships between idiopathic pulmonary fibrosis (IPF) and its usually associated comorbidities remain poorly understood. While evidence suggests that some comorbidities may directly influence the development or progression of IPF, or vice versa, whether these associations are causal or arise independently due to shared risk factors, such as ageing, smoking, lifestyle and genetic susceptibility, is still uncertain. Some comorbidities, such as metabolic syndromes, gastro-oesophageal reflux disease and obstructive sleep apnoea, precede the development of IPF. In contrast, others, such as pulmonary hypertension and lung cancer, often become apparent after IPF onset or during its progression. These timing patterns suggest a directional relationship in their associations. The issue is further complicated by the fact that patients often have multiple comorbidities, which may interact and exacerbate one another, creating a vicious cycle. To clarify these correlations, some studies have used causal inference methods (e.g. Mendelian randomisation) and exploration of underlying mechanisms; however, these efforts have not yet generated conclusive insights. In this review, we provide a general overview of the relationship between IPF and its comorbidities, emphasising the pathogenic mechanisms underlying each comorbidity, potential shared pathobiology with IPF and, when available, causal insights from Mendelian randomisation studies.

Shareable abstract

Understanding the connections between IPF and its comorbidities is essential and could help to identify factors that enhance early diagnosis and improve patient outcomes https://bit.ly/4izn3NT

Introduction

Comorbidities are frequent in patients with idiopathic pulmonary fibrosis (IPF) and have a marked influence in the clinical course and outcome of the disease [1]. Thus, their inclusion, mainly pulmonary hypertension (PH) and lung cancer, may help to predict survival beyond basic clinical and physiological parameters [2].

Some of the comorbidities seem to be the consequence of the fibrotic destruction of the lung architecture, such as PH, lung cancer and coronary artery disease (CAD), while others usually precede the initiation of IPF, and likely may contribute to the progression of this disease, such as obesity and type 2 diabetes mellitus (T2DM), hypothyroidism, sarcopenia, gastro-oesophageal reflux (GORD), and obstructive sleep apnoea (OSA).

PH is a prominent complication of IPF and is the result of multifactorial pathogenic mechanisms. The prevalence is variable, ranging from 8–15% in the early stages to >60% in the later stages [3].

Likewise, patients with IPF exhibit an increased risk of developing lung cancer even after adjusting for age, smoking and sex [4–7]. The estimated prevalence has been reported as 13–14% in two meta-analyses [7, 8]. Somatic mutations, common genetic patterns and epigenetic changes impacting epithelial cells and fibroblasts are likely involved. Some studies suggest that IPF may be an independent risk factor for CAD, even after multivariate analysis accounting for the more commonly recognised CAD risk factors, such as ageing and smoking [9–16].

Several metabolic disorders have been found associated to IPF, which may contribute with different mechanisms, including shortening of telomeres and mitochondrial dysfunction, to the development or progression of the disease. Thus, the frequency of obesity among patients with IPF is approximately two-fold higher compared with normal subjects [17], and a higher prevalence of T2DM is also observed in these patients [18]. A decade ago, it was revealed for the first time that the frequency of hypothyroidism was significantly higher in patients with IPF compared with matched controls with COPD and markedly higher than its estimated prevalence in the general population [19].

GORD is frequent in IPF and it has been associated as an independent risk factor for IPF in several observational studies [20, 21]. It is hypothesised to contribute to IPF progression through microaspiration. However, some data suggest that it may be a protective factor [2].

OSA, and its consequent intermittent oxygen desaturation, is a prevalent coexisting condition among individuals with IPF and is associated with poor outcome. Any kind of sleep disorder (primarily OSA) has been reported in 89% of patients [22].

Some of these comorbidities share common risk factors, such as ageing, smoking and shortening of telomeres, but differ in biopathological pathways.

Also, it remains unclear whether these comorbidities have a causal effect on IPF or if IPF contributes to their development (causal relationship). Alternatively, they may develop independently due to shared risk factors, such as ageing, smoking, lifestyle and genetic susceptibility, suggesting an association where a third variable may influence both conditions. In this scenario, one disease may (or may not) indirectly contribute to the progression and outcome of the other.

It is important to consider that while randomised controlled trials are regarded as the gold standard for evaluating causality, their feasibility, cost and ethical concerns limit their use. In this context, Mendelian randomisation has emerged as an essential tool for drawing consistent causal extrapolations by utilising genetic variants, often single nucleotide polymorphisms, obtained from large-scale, publicly available, genome-wide association studies (GWAS) to estimate correlations and deduce potential causal relationships between phenotypes or diseases [23, 24]. Mendelian randomisation is increasingly being used due to its ability to address a major limitation of evidence from observational studies, i.e. unmeasured confounding. Their findings could help prioritise clinical trials or drug development and inform clinical or public health decision making. However, like any analytical method, Mendelian randomisation relies on certain assumptions, and the validity of these assumptions must be carefully evaluated [25]. Horizontal pleiotropy represents an important problem in which the exclusion restriction assumption may be violated, i.e. when a genetic variant influences the outcome (directly or indirectly) through other traits by pathways other than the risk factor alone [26].

Interestingly, several recent bidirectional Mendelian randomisation analyses, approaching possible causal associations between IPF and comorbidities, shed some light on distinguishing between association and causality. In this review, we provide a general overview of the relationship between IPF and its comorbidities, emphasising the pathogenic mechanisms underlying each comorbidity, potential shared pathobiology with IPF and, where relevant, causal insights from Mendelian randomisation studies.

Search strategy and selection criteria

We conducted a search on PubMed for articles published in English between January 2000 and November 2024, using the search terms “idiopathic pulmonary fibrosis”, “IPF” and comorbidities, prioritising studies that provided a more in-depth analysis of pathogenic mechanisms or employed Mendelian randomisation. We specifically looked at IPF in relation to each of the nine selected comorbidities: “diabetes mellitus”, “pulmonary hypertension”, “gastro-oesophageal reflux disease”, “obstructive sleep apnoea”, “lung cancer”, “coronary artery disease”, “obesity”, “hypothyroidism” and “sarcopenia”. Additionally, we combined these terms with “pathogenic mechanisms”, “pathogenesis” and “Mendelian randomisation”. The final list of references was chosen based on their relevance to the focus of this review.

IPF as a potential risk factor for comorbidities

Strong evidence suggests that IPF may influence the development of three diseases: PH, lung cancer and CAD (figure 1).

FIGURE 1.

Graphical representation of some of the mechanisms implicating idiopathic pulmonary fibrosis (IPF) as a potential risk factor for pulmonary hypertension, lung cancer and coronary artery disease. The dashed arrow indicates a likely weaker association.

Pulmonary hypertension

PH is characterised by vascular remodelling of distal pulmonary arteries, formation of occlusive intimal lesions and increased vessel wall thickness. Extracellular matrix (ECM) accumulation, loss of vascular elasticity and vessel stiffening finally evolve into impaired right ventricular function [27, 28].

Although the development of PH in IPF is usually associated with lung fibrotic destruction and hypoxic vasoconstriction, recent evidence indicates that other mechanisms, including the cross-talk among multiple cell types, may also contribute to pulmonary arterial remodelling and subsequent hypertension.

While the possible correlation between lung fibrosis severity and PH is unclear, a recent approach using quantitative computed tomography imaging and pulmonary vascular reconstruction suggests a strong relationship between the extent of pulmonary fibrosis and PH [29, 30]. These data, combined with the fact that PH usually occurs during the development and progression of IPF, suggest that there is (at least partially) a unidirectional causal relationship.

Endothelial dysfunction, endothelial-to-mesenchymal transition and inflammation, as well as fibroblast and pulmonary artery smooth muscle proliferation and transition to myofibroblasts, have been revealed in PH [31–33]. Supporting the role of the endothelium, a recent detailed study in fibrotic lung disorders, including IPF and almost all with PH, demonstrated that endothelial cells are phenotypically abnormal and functionally impaired [34]. Thus, markers of dysregulation and activation were observed in different endothelial subpopulations, including general capillary endothelial cells and aerocytes.

Sustained endothelial cell activation enhances an inflammatory response, and there is evidence that the dysregulation of several types of inflammatory cells, such as macrophages, mast cells and myeloid-derived suppressor cells, promotes vascular remodelling [35–38]. Also, patients with end-stage parenchymal lung fibrosis, including IPF, and PH exhibit a deficiency in natural killer T-cells and dysfunction of the natural killer T-cell–STAT1–CXCL9 axis. Remarkably, restoring the functionality of this axis significantly reduced collagen deposition and vascular remodelling [39].

Supporting the complexity of the cell interactions in the pathogenesis of PH, it has been recently found that recombinant human interleukin (rhIL)-11 and soluble rhIL-11 receptor α promote fibrocyte migration, endothelial cell adhesion and myofibroblast transition. Targeting IL-11 reduces fibrocyte lung accumulation in animal models of PH associated with lung fibrosis [40]. Additionally, it has been revealed that hypoxia-inducible factor (HIF)-2α plays a central role in driving pericytes, which contribute to distal vascular remodelling and muscularisation [41]. Therefore, the dysregulation of multiple resident and inflammatory cells contributes to PH.

Recently, the involvement of active transforming growth factor (TGF)-β/Smad-2/3-dependent and β-catenin-dependent Wnt signalling pathways was described in patients with IPF, which can drive endothelial-to-mesenchymal transition, ECM deposition and pulmonary arterial remodelling [42].

From the genetic point of view, in a large transcriptomic study, several gene modules were identified as associated with pulmonary vascular disease severity in pulmonary fibrosis [43]. A distinct transcriptomic module (292 genes) positively correlated with PH associated with pulmonary fibrosis (mostly IPF). Interestingly, this module showed the most substantial enrichment with the idiopathic PH gene signature, indicating shared gene expression patterns. An important finding of this study was the downregulation of oxidative phosphorylation in modules associated with higher PH, supporting the evidence that endothelial cells exhibit a metabolic change, preferentially shifting from oxidative phosphorylation to glycolysis [38].

Taken together, all these findings support the idea that in PH associated with IPF, there is a detrimental cycle driven by the ongoing activation of endothelial, immune, mesenchymal and inflammatory cells. This cycle, which is at least partly linked to the degree of fibrosis likely through cell–ECM cross-talk, contributes to the abnormal interactions that trigger and sustain vascular stiffening.

Lung cancer

Pulmonary fibrosis meets the five criteria proposed by the National Cancer Institute to be classified as a pre-cancerous disease, and similarities in pathobiological pathways have been described [44, 45]. The most frequent type is squamous cell carcinoma, usually located in the peripheral lung adjacent to usual interstitial pneumonia lesions [4, 7].

Pathogenic mechanisms

The epithelial connection

The existence of an aberrant basaloid epithelial cell was recently revealed in IPF lungs [46]. These cells express markers of epithelial-to-mesenchymal transition and senescence-related genes, and exhibit the highest expression of genes associated with the pathogenesis of IPF, such as SOX9 (sex-determining region Y-related 9), GDF15 (growth differentiation factor 15), MMP7 (matrix metalloprotease 7) and EPHB2 (ephrin receptor B2). Interestingly, many of these genes may have a role in carcinogenesis. For example, SOX9 is usually upregulated in various cancers, including non-small cell lung cancer (NSCLC) [47]. Likewise, high expression of MMP7 is an independent predictor of a high incidence of spread through airspaces of adenocarcinoma [48], and recent studies have increasingly linked EPHB2 dysregulation to cancer progression [49].

Therefore, basaloid epithelial cells may be implicated in the pathogenesis of lung cancer, but it is uncertain if some may be the origin of the malignant transformation. More likely, neoplastic transformation may occur in the epithelium re-covering the honeycomb lesions, where lung cancer usually initiates. Basal cells are characterised by the presence of p63⁺ KRT5⁺ (keratin 5) expression, and it is well known that both molecules are markers of squamous cell carcinoma [50].

The fibroblast connection

Like IPF, lung cancer is characterised by the expansion of fibrotic ECM and fibroblasts, known as cancer-associated fibroblasts, which contribute to tumour progression and metastasis. Notably, single-cell RNA sequencing has revealed remarkably similar expression profiles. For instance, two myofibroblast clusters characterised by the expression of LRRC15⁺ (leucine-rich repeat-containing protein 15) and COL3A1⁺ (collagen α-1(III) chain), both of which produce high levels of collagens, and ECM-modifying genes such as CTHRC1 (collagen triple helix repeat-containing protein 1) have been identified in both cancer and fibrosis [51, 52]. Furthermore, senescent fibroblasts in both conditions alter the lung micro-environment in a comparable tumour-promoting way (figure 2) [53, 54]. Recently, it was shown that extracellular vesicles secreted by senescent IPF fibroblasts influence the proliferation, migration and invasion of lung cancer cells, and MMP-1 and protease-activated receptor-1 were identified as crucial factors of the senescence-associated secretory phenotype (SASP) [55].

FIGURE 2.

Hypothetical mechanisms driving carcinogenesis in idiopathic pulmonary fibrosis (IPF). The epithelial cell composition in the distal lungs of individuals with IPF undergoes significant alterations, including an increased presence of airway epithelial cells such as basal cells (p63⁺ KRT5⁺). Additionally, a unique aberrant subtype of epithelial cells, known as basaloid cells (KRT17⁺ KRT5⁻), emerges. These basaloid cells are transcriptionally distinct and express developmental transcription factors and other molecules that may promote carcinogenesis, potentially influencing basal cells and driving malignant transformation. Moreover, the expansion of IPF-associated fibroblasts and myofibroblasts surrounding cancerous lesions plays a crucial role in shaping the tumour micro-environment. Through interactions with other cellular components, these fibroblasts and myofibroblasts (cancer-associated fibroblasts (CAFs)/cancer-associated myofibroblasts (CAmyoFs)) contribute to tumour progression and metastasis, acting as key regulators of cancer dynamics within the fibrotic lung tissue. Also, microvesicles derived by senescent IPF fibroblasts release several molecules such as metalloprotease-1 (MMP-1) and protease-activated receptor-1 (PAR-1) that enhance the proliferation, migration and invasion of lung cancer cells. ECM: extracellular matrix; SASP: senescence-associated secretory phenotype.

Interestingly and paradoxically, lung cancer arises in IPF tissues characterised by short telomeres and epithelial and fibroblast senescence. However, although substantial evidence links longer telomeres to an increased risk of cancer, a comprehensive analysis of telomere length across various cancer types revealed an unexpected pattern: in 70% of cases, cancerous tissues displayed relatively shorter telomeres compared with their matched normal controls [56]. Furthermore, senescent cells may facilitate late-life cancers (antagonistic pleiotropy) by promoting a SASP-mediated micro-environment that supports the malignant progression of oncogenic mutated cells [54, 57]. Additionally, tumours originating from senescent cells may exhibit more aggressive phenotypes than those derived from non-senescent transformed cells [58].

The genetic connection

IPF lungs tend to exhibit higher proportions of somatic mutations in cancer driver genes, with more lung samples showing at least one cancer driver gene mutation than normal lungs [59]. Clonal somatic mutations were enriched in disease-related epithelial genes such as MUC5B (mucin 5B), and interestingly, IPF patients carrying the MUC5B risk variant were more likely to develop somatic mutations [59].

The gene expression profile of patients with IPF and NSCLC was compared with healthy controls [60]. 79 common signature genes were found shared by these two diseases. The most enriched pathway was the peroxisome proliferator-activated receptor signalling pathway, which involves several essential biological functions. MMP1, which has been associated with early-onset lung cancer and also with IPF [61, 62], was the most significantly altered common signature gene shared by IPF and lung cancer. MMP-1 protein expression was stronger in the cancer tissue of patient stage IA associated with IPF than that of NSCLC stage IIIA without IPF, suggesting a role in the initiation of cancer in the IPF lungs. However, this finding requires further investigation [61].

In another study performed by next-generation sequencing in lung cancer with and without IPF, 43 mutations in 13 genes were detected [63]. From them, TP53 and MET (hepatocyte growth factor receptor) mutations were the most frequent. Evaluation of the genomic profiles in matched tumour/normal pairs and next-generation sequencing confirmed TP53 as the most prevalent among the somatically mutated genes detected in the examined IPF-NSCLC cohort [64]. Nuclear accumulation of the mutated TP53 gene has been identified in atypical epithelial lesions and carcinoma in IPF patients [65, 66].

The transcription factor p53 plays a key role in the defence against cancer, blocking the replication of damaged DNA by arresting the cell cycle in the G₁ phase, and its mutations are the most consistent finding linking IPF and lung cancer [67]. Of interest, BRAF (B-Raf proto-oncogene, serine/threonine kinase) mutations, although less frequent, were also found at a significantly higher frequency in IPF-lung cancer patients, far exceeding the known prevalence in lung cancer [64]. FHIT (fragile histidine triad) gene allelic loss has also been observed in metaplastic lesions around honeycomb formations in IPF patients with lung cancer [68].

Three Mendelian randomisation studies utilising gene variants with genome-wide significance have been conducted to investigate the potential link between IPF and lung cancer [69–71]. Two of them found associations between IPF and specific lung cancer subtypes: one suggested that IPF may increase the risk of lung squamous cell carcinoma [70], while the other showed a positive association between IPF and small cell lung cancer [71].

A significant genetic correlation with lung cancer was found using large-scale biobank data and transcriptome-wide association analysis to identify novel IPF risk genes [72]. Although the global genetic correlation between both diseases only ranked middle in correlation strength, it had the most significant proportion (26%) of interrelated regions.

A recent study revealed that the rs9265808 locus in MS4A4A (membrane spanning 4-domains A4A) is a susceptibility locus for progression from IPF to lung cancer. In addition, a two-sample Mendelian randomisation approach indicated a causal relationship between the expression of MS4A4A and the risk of IPF, with elevated expression levels increasing the disease risk [73].

Combining single-cell RNA sequencing and multiple bioinformatics techniques, several connected mechanisms between IPF and adenocarcinoma were recently revealed [74]. Interestingly, 650 differentially expressed genes were shared by both diseases. Biological processes included epithelial cell proliferation, regulation of cell adhesion and those involved in collagen-containing ECM. Epithelial cell-derived signature genes involved in the progression of IPF to adenocarcinoma included TRIM2 (tripartite motif-containing protein 2), S100A14 (S100 calcium-binding protein A14), CYP4B1 (cytochrome P450 family 4 subfamily B member 1), LMO7 (LIM domain only 7) and SFN (strafin) [74].

The epigenetic connection

DNA methylation is a critical epigenetic modification induced by environmental factors, smoking and others that are common risk factors in IPF and lung cancer, and may contribute to the association between both diseases [75–77].

Hypermethylated regions in the Smad4 promoter were detected in patients with IPF-lung cancer (mostly NSCLC), and accordingly, the expression of the gene was significantly lower in tumours from these patients than in tumours from lung cancer without IPF, suggesting that the loss of the growth inhibitory response to TGF-β signalling may contribute to the link between IPF and lung cancer [78].

A more recent study was performed with genome-wide DNA methylation analysis in squamous cell carcinoma with or without IPF as well as non-cancerous lung tissue samples from smokers/non-smokers [79]. Lung cancer was classified into low- and high-methylation epigenotypes by hierarchical clustering analysis. Interestingly, IPF was significantly associated with the low methylation epigenotype, which was also associated with worse prognosis in patients with squamous cell carcinoma with or without IPF.

Studies with non-coding RNAs are scarce. Many aberrantly expressed microRNAs have been independently identified in both diseases, with many showing similar patterns of regulation, either upregulated or downregulated [80]. However, there are no studies in lungs with coexisting IPF and lung cancer.

Collectively, these studies suggest that various genetic and epigenetic mechanisms, likely affecting several molecular processes from epithelial cells and fibroblasts, may connect lung fibrosis to lung cancer; however, many different combinations may contribute to carcinogenesis in IPF.

Coronary artery disease

The mechanisms by which IPF is a risk factor for CAD remain unclear. One possibility is that hypoxaemia and reactive oxygen species may contribute to ischaemia in the subendocardium by impairing oxygen delivery. Oxidised choline glycerophospholipids species serve as endogenous ligands for the macrophage scavenger receptor CD36 [81], and accumulation of macrophages plays a role in the progression of atherosclerosis [82].

However, similar mechanisms are likely occurring in COPD where CAD is significantly lower. Therefore, it can be hypothesised that IPF, through unidentified mechanism(s), enhances the release of proatherogenic mediators.

The possible shared genetic architecture and causal relationship between IPF and 10 prevalent heritable comorbidities, including CAD, was recently reported [69]. Latent causal variable analysis was applied to assess the putative causal relationship between traits [83]. Then, multiple bidirectional Mendelian randomisation approaches were applied to the pairs of traits to reveal possible genetic causality, and a modest possible partial causal relationship was found between IPF and CAD [69].

However, a systematic review and meta-analysis including 62 studies with 310 outcomes and several Mendelian randomisation associations exploring the causal relationships between telomere length and various health-related outcomes found that shorter telomere length is associated with increased risk of IPF as well as increased risk of coronary heart disease [84].

Taken together, these findings suggest that the likelihood that IPF causes CAD is low, and more likely the association may be related to other shared mechanisms, such as shortening of telomeres.

Pre-existing comorbidities and their potential role in the development of IPF

Several diseases may impact the risk and development of IPF (figures 3 and 4).

FIGURE 3.

Pre-existing comorbidities that may influence the development of idiopathic pulmonary fibrosis (IPF). Putative mechanisms underlying the association of diabetes mellitus, obesity and hypothyroidism with IPF are shown. The dashed arrow indicates a weaker strength of association. In addition, these conditions, along with physical inactivity, contribute to the development of sarcopenia, which can influence the progression and outcomes of IPF. Notably, obesity exhibits a paradoxical impact on IPF.

FIGURE 4.

Mechanisms through which gastro-oesophageal reflux disease and obstructive sleep apnoea may contribute to the development of idiopathic pulmonary fibrosis (IPF). There is also some evidence suggesting that gastro-oesophageal reflux disease could be protective. HIF: hypoxia-inducible factor; TGF: transforming growth factor.

Diabetes mellitus

In recent years, it has been shown that IPF is often diagnosed in individuals with T2DM, and several meta-analyses have found a higher prevalence of diabetes among individuals with IPF compared with controls, suggesting a possible association between these two conditions [18, 85]. In sharp contrast, in a retrospective, population-based study using the Centers for Disease Control and Prevention Multiple Cause of Death database that included 26 305 568 deaths in the USA, the risk of pulmonary fibrosis (allegedly primarily IPF) was lower in those with diabetes in all age strata and ethnicities, and both sexes [86].

Association studies have given contradictory results. A recent Mendelian randomisation analysis explored the potential causal relationship between type 1 diabetes mellitus (T1DM) or T2DM and IPF. Curiously, univariate and multivariate analyses demonstrated a causal relationship between T1DM and IPF, but not between T2DM and IPF [87]. Likewise, no causal association was also found in another comparable bidirectional Mendelian randomisation study [88]. Moreover, a very recent two-sample Mendelian randomisation approach, including bidirectional inverse variance weighted random effects, demonstrated that the comorbid relationship between diabetes and IPF is not driven by direct causative effects of either disease [89]. Supporting these findings, in a large-scale GWAS it was found that there was a significant genetic correlation but not causal relationship between IPF and T2DM [69].

Using a different approach, multimodal generative topic modelling of disease-omics data of 6955 human diseases identified a molecular relationship of IPF with several non-respiratory diseases, including diabetes [90]. One signalling node in the inter-organ mechanism of IPF, the calmodulin pathway, was found molecularly linked to diabetes [90].

Taken together, these and other case–control studies indicate that T2DM is a relatively common comorbid disease, that is not causally related but that might (hypothetically) participate to the biopathology of IPF. However, the mechanisms altered in T2DM that may contribute to IPF remain unclear but likely involve processes critical for cell regeneration. For example, several cross-sectional studies have shown that telomere length is shorter in people with diabetes than in those without. In a recent meta-analysis of 37 observational studies involving 18 181 participants from 14 countries, it was demonstrated that telomere length is shorter in T2DM [91]. It is well known that patients with IPF are characterised by excessive shortening of telomeres which play a role in the development and progression of the disease [92]. Thus, at least hypothetically, T2DM may contribute to the progression of IPF by affecting the length of telomeres. Interestingly, a retrospective cohort analysis based on US national claims data suggested that metformin therapy in patients with both IPF and T2DM was significantly associated with a reduced risk of hospitalisations and lower all-cause mortality [93].

Likewise, in experimental models of diabetes, the cell DNA repair potential is severely compromised, resulting in the activation of a senescence programme and ultimately fibrosis [94, 95]. In IPF, epithelial cell senescence and the SASP activate a cascade of mediators resulting in a fibrotic response [92]. Finally, non-enzymatic reactions involving advanced glycation end-products (AGEs) occur in diabetes, and excessive AGE-mediated cross-links may also enhance lung fibrosis [96].

The paradox of obesity

Studies of the effect of higher body mass index (BMI) on IPF have given contradictory results regarding risk and progression.

Several Mendelian randomisation analyses have been conducted to evaluate a potential causal relationship between obesity and the risk of IPF. These studies have found that increased obesity-related markers are associated with a higher risk of IPF, providing evidence that a genetically determined increase in BMI is linked to an elevated risk of developing IPF [97–99].

More recently, pervasive genetic correlation between IPF and several comorbidities, including obesity, was also reported [69]. 18 risk single nucleotide polymorphisms were associated with the cross-trait pairs exhibiting significant genetic correlations, including four mucin genes (e.g. MUC5B) that were shared between IPF and BMI.

In sharp contrast, a growing body of evidence has indicated that a higher BMI can confer a protective effect on clinical outcomes. For example, a meta-analysis involving 14 datasets and 2080 patients with IPF demonstrated that higher BMI (≥25 kg·m⁻²) was negatively correlated with the risk of death [100].

Likewise, a systematic review including 36 studies assessing the prognostic role of baseline and temporal changes in BMI in IPF patients found that relatively low BMI values at baseline and/or greater temporal declines in BMI were associated with adverse clinical outcomes [101].

In a very recent large-scale, nationwide, population-based study, the association between BMI and the clinical outcomes of IPF was examined by using a claims database [102]. The results confirmed that a lower BMI is associated with worse prognosis, whereas a higher BMI is related to better prognosis, supporting a non-linear relationship between BMI and mortality in patients with IPF.

Therefore, while higher BMI values seem to be associated with the risk of IPF, these patients have a more favourable prognosis than those with lower BMI values, a phenomenon known as the “obesity paradox”. This finding has also been described in other respiratory conditions such as COPD where overweight or obesity showed a protective effect against mortality [103].

The mechanisms involved in the increased IPF risk and those that reduce mortality are unclear. Regarding the first, increased levels of TGF-β1, a potent profibrotic molecule, have been observed in adipose tissue and contribute to the development of fibrosis in several organs, including the lung [104–107]. Likewise, other critical mediators of inflammation and fibrosis, such as monocyte chemoattractant protein-1 and tumour necrosis factor-α, may also be involved [108].

Regarding the protective effect on outcome, a growing body of evidence indicates a putative role for adiponectin. In an experimental model of bleomycin-induced lung fibrosis, adiponectin attenuated body weight loss, alveolar destruction and accumulation of collagen fibres [109]. In addition, adiponectin significantly reduced TGF-β1-induced proliferation in human lung fibroblasts, and halted and reversed stiffness-induced, profibrotic fibroblast activation [110].

However, the fundamental question remains unanswered: why does obesity enhance the IPF risk, but protect for mortality?

Hypothyroidism

A recent study revealed, by two-sample Mendelian randomisation, a positive causal effect of hypothyroidism on IPF, while bidirectional analyses showed no reverse causation [111]. Suggestive evidence was also observed in another Mendelian randomisation study where the bidirectional causal association between IPF and 22 comorbidities was examined [18].

However, the putative mechanisms involved in the causal effect of hypothyroidism on IPF remain unclear. It is well known that thyroid hormone has cytoprotective and regenerative properties through targeting genes critical to maintenance of cellular homeostasis and bioenergetics [112].

In 2018, using two models of murine lung fibrosis, it was found that thyroid hormone likely exerted antifibrotic effects by restoring mitochondrial function through induction of mitochondrial biogenesis and mitophagy [113]. These beneficial effects were associated with suppression of mitochondria-regulated death pathways, and dependent on intact PPARGC1A (peroxisome proliferator-activated receptor γ coactivator α) and PINK1 (PTEN-induced putative kinase-1) pathways. Dysfunctional mitochondria in alveolar epithelial cells from IPF have been found to be related to a low expression of PINK1, the regulator of mitochondrial homeostasis [114].

On the other hand, a recent Mendelian randomisation study, used to determine the causal impact of the identified hypothyroidism-driven proteins on IPF outcomes, reported a strong association between elevated CXCL10 levels and increased risk of IPF. Moreover, mediation analysis revealed that CXCL10 accounted for one-third of the impact of hypothyroidism on IPF [115].

A mild increase of serum levels of CXCL10 has been found in IPF patients [116, 117]. Of interest, high serum levels of CXCL10 are associated with autoimmune-related features, a better clinical course and survival in patients with idiopathic interstitial pneumonias, including IPF [118].

Sarcopenia

Sarcopenia, characterised by generalised and progressive skeletal muscle disorder with accelerated loss of muscle mass and physical function, is commonly seen in older adults. Frailty is a syndrome associated with multidimensional loss of function. Various factors contribute to the age-related decline in muscle mass and strength, with physical inactivity, often experienced by patients with IPF, being a significant contributor [119].

Interestingly, some metabolic disturbances related to IPF also contribute to sarcopenia. For example, there is epidemiological evidence that T2DM is associated with rapid loss of skeletal muscle mass and strength in older adults [120]. Similarly, sarcopenia and obesity can coexist and, together, they are linked to greater functional decline and poorer outcomes than either condition on its own [119]. A study involving community-dwelling adults aged >65 years showed that obese individuals with decreased lower extremity muscle strength experienced a higher decline in walking speed and a greater risk of developing mobility incapacity compared with those without obesity and low muscle strength [121].

Additionally, a study employing a two-sample Mendelian randomisation approach to investigate the causal relationships between thyroid dysfunction and sarcopenia-related traits through genetic variation analysis suggested that hyperthyroidism, hypothyroidism and subclinical hyperthyroidism all elevate the risk of sarcopenia [122].

Therefore, several IPF-associated comorbidities and IPF-associated sedentary lifestyle contribute to sarcopenia.

In turn, sarcopenia and frailty negatively impact the progression of IPF, and are significantly associated with disease severity, reduced lung function, diminished quality of life and decreased survival, even after adjusting for age and gender [123–125].

Collectively, these findings indicate that metabolic dysregulation, including diabetes, obesity and hypothyroidism, contributes to sarcopenia, which in turn exacerbates the progression and outcomes of IPF.

Gastro-oesophageal reflux disease

IPF patients had increased reflux episodes, and bile salts and pepsin have been detected in the bronchoalveolar lavage fluid which may stimulate the production of TGF-β by epithelial cells [126–129].

However, whether the high prevalence of GORD in patients with IPF is pathologically associated to the development, progression or exacerbation of the disease over time remains unclear [2, 130]. The putative causal role and the directionality of the relationship between both diseases have been widely debated. In a meta-analysis involving 18 case–control studies and a large number of IPF patients and controls, it was found that GORD and IPF may be related, but this association is most likely confounded, especially by smoking [131]. Supporting this finding, approaching this relationship by Mendelian randomisation, after adjusting for smoking no evidence was found indicating that GORD increases susceptibility to IPF or progression [132].

Along the same line of thought, in another Mendelian randomisation evaluation, it was found that while univariable models showed that GORD increased the risk of IPF, much of this association was likely due to confounding through BMI [133].

In sharp contrast, a bidirectional Mendelian randomisation analysis including 22 comorbidities with genetic data available showed convincing evidence that GORD was causally associated with an increased risk of IPF [88]. For the analysis, the GWAS of IPF was obtained from a study performed across five cohorts, and for replication, data were extracted from the Global Biobank Meta-Analysis. Genetic liability was associated with a higher risk of IPF and the direction was reliable across sensitivity analyses, including pleiotropy-robust methods, and adjustment for smoking and controlling for possible reverse causation [88].

Genetic correlation between IPF and GORD was also found in a recent Mendelian randomisation study [69]. Moreover, significant cross-trait risk gene variants shared between IPF and GORD were identified. The analysis revealed 12 candidate genes with false discovery rate significant associations between IPF and GORD.

The debate in this topic has also included the usefulness (or not) of antireflux and antacid therapy. In a systematic review of the literature no statistically significant effect on disease progression was found for antacid medication or antireflux surgery, although the quality of evidence was judged very low owing to study design and indirectness [134]. Moreover, an unexpected putative protective association of GORD with IPF was also reported [2].

Considering all this information, the fundamental question remains: is GORD a causal risk factor for IPF?

Obstructive sleep apnoea

OSA is a frequent comorbidity in IPF [22]. In meta-analyses with numerous studies enrolling IPF patients it was found that the pooled prevalence of OSA was as high as 76% [135, 136], while the median prevalence of OSA among non-IPF patients for all the ethnic groups included in the last study was 16.4% [136].

Pathogenic connection

Intermittent hypoxaemia exposure, cyclic hypoxia–reoxygenation and oxidative stress occurring during recurrent apnoeas are major pathogenic elements of OSA, and negatively impact several molecular and cellular processes that may be relevant in the pathogenesis of IPF, which may also display a decrease in the expression of antioxidant genes.

For example, in a large bibliometrics and bioinformatics study, 72 genes associated with oxidative stress were identified as differentially expressed in IPF [137]. By employing protein–protein interaction network analysis and machine learning algorithms, PON2 (paraoxonase 2) and TLR4 (Toll like receptor 4) were significantly reduced in IPF, and were revealed as hub genes. PON2 has antioxidative stress effects, and its depletion is directly associated with the pathogenesis of liver fibrosis [138, 139]. Of interest, PON1 seems to protect mice from lung fibrosis induced by bleomycin [140]. Likewise, decreased expression of nuclear factor erythroid 2-related factor 2 (Nrf2), a critical regulator in the induction of endogenous antioxidant enzymes, was associated with a myofibroblast phenotype in IPF fibroblasts [141]. Also, at the protein level it has been shown that fibrotic areas and fibroblastic foci in IPF lungs are notable for the absence of extracellular superoxide dismutase by immunohistochemistry and Western blotting [142].

Oxidative stress provoked by intermittent hypoxaemia can lead to the subsequent activation of several transcription factors, such as HIF-1 and NF-κB [143]. Stabilisation of HIF-1α may provoke epithelial cell apoptosis and importantly, production of TGF-β1, epithelial-to-mesenchymal transition and fibrotic progression [144–147]. Similarly, localised hypoxia in areas of parenchymal damage induced endoplasmic reticulum stress that was mechanistically linked to lung fibrosis through C/EBP homologous protein in several models of lung fibrosis [148].

Importantly, it has been shown that mice pre-exposed to intermittent hypoxaemia display significant weight loss and aggravated bleomycin-induced lung fibrosis [149]. By contrast, no significant effect of intermittent hypoxaemia on lung fibrosis severity was observed in mice co-challenged with intermittent hypoxaemia and bleomycin, suggesting that presenting OSA before IPF contributes to fibrosis pathogenesis [149].

From the genetic point of view, a genome-wide gene expression array performed in patients with OSA (without comparing with other disease) identified the upregulation of several genes related to apoptosis (BIRC3 (baculoviral IAP repeat-containing 3)), endothelial dysfunction (AMOT (angiomotin)), dynamics of assembly/disassembly of the tight junction barrier (PLEKHA7 (pleckstrin homology domain containing A7)) and editing machinery (ADAR1 (adenosine deaminase RNA-specific)) that may also contribute to the pathogenesis of IPF [150].

Interestingly, however, no significant causal link between OSA and IPF was found in a recent two-sample Mendelian randomisation approach that investigated a putative causal link between OSA and various chronic respiratory diseases, including IPF [151].

Conclusions and future directions

The incidence of several major public health diseases associated with ageing is on the rise, and many of them are significant comorbidities of IPF. These diseases may elevate the risk of mortality, irrespective of IPF severity. The bidirectional relationship between IPF and its comorbidities is complex, likely influenced by genetic and epigenetic factors, and requires additional research to elucidate the underlying mechanisms and mitigate risk factors. Distinguishing causality from association carries significant clinical implications. Demonstrating causality may warrant close monitoring and early intervention for the disease exerting the causal influence, such as IPF on lung cancer. However, the current evidence remains limited, with critical gaps in data making it challenging to provide strong recommendations.

In this context, a thorough investigation of the biological factors that make IPF a potential risk factor for comorbidities such as lung cancer, PH and CAD is essential. Equally important is exploring the role of metabolic dysfunction, GORD, OSA and potentially other conditions in the onset or progression of IPF. Gaining a deeper understanding of these connections could lead to the identification of risk factors or biomarkers that enable earlier diagnosis and improve patient outcomes. Likely, in the future, advances in molecular profiling, next-generation sequencing technologies and improved two-sample bidirectional Mendelian randomisation studies will provide valuable tools to deepen our understanding of the interplay between IPF and its comorbidities. Furthermore, Mendelian randomisation studies are valuable not only for increasing confidence in a target's causal role but also for identifying potential targets for drug development. However, interpreting Mendelian randomisation findings requires caution, and adherence to key assumptions is essential for its proper application [24]. In the future, it will be important to have access to larger, more diverse and higher quality genetic association and epidemiological data, combined with early and improved clinical identification of comorbidities related to IPF. In addition, with the anticipated improvements in statistical methods and the incorporation of artificial intelligence, it will become feasible to find strong evidence for causal associations and implement public health measures that minimise risk factors and enhance screening tools, facilitating a transition from a reactive strategy to a more proactive and preventive model.

Shareable PDF

This PDF extract can be shared freely online.

Shareable PDF ERJ-02418-2024.Shareable ^{(624.2KB, pdf)}