Background
Chronic hypersensitivity pneumonitis (cHP), is a complex diffuse parenchymal disease caused by repeated and prolonged exposure to inhaled antigens in susceptible individuals, leading to an inflammatory response in the lungs, and, in some patients, to development of pulmonary fibrosis [
1]. The experimental study on bronchoalveolar lavage fluid (BALF) of patients with hypersensitivity pneumonitis has demonstrated that the alveolar haemostatic balance is shifted in favour of procoagulant and antifibrinolytic activity [
2], similarly to patients with idiopathic pulmonary fibrosis (IPF), the most common progressive fibrotic disorder of unknown aetiology and unfavourable prognosis [
3,
4]. Extensive research over the last two decades have contributed to a better understanding of the pathogenesis of this rare, debilitating disease. In genetically susceptible ageing individuals, the repeated micro injury of the alveolar epithelium has been recognized as the first driver of an altered, aberrant repair process, initiating a complex cascade of events leading to scarring of lung tissue. In the early phase of wound healing process, epithelial and endothelial cells damage result in the activation of coagulation cascade via the tissue factor (TF)-dependent extrinsic pathway. As a result, the balance between coagulation and fibrinolysis is shifted in favour of procoagulant activity, enhanced by increased levels of fibrinolysis inhibitors as plasminogen activator inhibitor 1 and 2 (PAI1 and PAI2), and protein-C inhibitors [
5,
6]. An imbalance between thrombosis and fibrinolysis has been demonstrated in animal models of lung fibrosis and in the alveolar compartment of IPF patients [
7,
8]. Furthermore, several population-based studies of the last decade showed an association between IPF and venous thromboembolism (VTE), indicating an increased risk of prothrombotic state, increased incidence and prevalence of VTE in IPF patients, and ultimately increased mortality [
9‐
13].
The association between chronic hypersensitivity pneumonitis and venous thromboembolism has not been studied yet. The aim of this study was to evaluate the epidemiology of VTE in IPF and cHP patients, diagnosed in our tertiary referral centre and further to evaluate related risk factors. We also examined overall survival of patients with VTE compared to non-VTE group.
Discussion
We found that the incidence rate of symptomatic venous thromboembolism in patients with chronic hypersensitivity pneumonitis was similar to that demonstrated in IPF group (7.1 per 1000 person-years vs 11.8 per 1000 person-years, RR 1.661 95% CI 0.545-6.019, p=0.48, respectively), even though the patients with IPF were older and had more comorbidities. We also showed that there were no significant differences in median, 3- and 5-years survival rate between the patients with and without VTE. To our knowledge, this is the first large comprehensive study on epidemiology of venous thromboembolism performed in well characterized cohort of cHP and IPF.
Venous thromboembolism, comprising deep vein thrombosis and pulmonary embolism, is a common, preventable disease with a substantial morbidity and mortality. An estimated incidence rate of VTE in the general population range from 1.04 to 2.7 per 1000 person-years [
20‐
25]. In the present study we showed that the incidence rate of symptomatic VTE events in patients with cHP was similar to that demonstrated in IPF group and was higher than the rate found in aforementioned population-based studies. Additionally, our results regarding the IPF group are similar to those presented in other studies indicating an increased risk and incidence of PE and DVT in patients with IPF. In large population-based study, Dalleywater et al. [
13] found that people with IPF had higher incidence rates of PE and DVT, compared to the general population (PE 9.3 vs 1.52/1000 person-years, and DVT 4.3 vs 2.11/1000 person-years,
p<0.001, respectively). A similar, large increase in the incidence of DVT (5.9/1000 person-years) was demonstrated in patients with IPF by Hubbard et al. [
9]. We found comparable incidence rates of PE and DVT in both evaluated cohorts IPF and cHP (PE: 7.8 vs 5.7/1000 person-years, RR 1.384 95% CI 0.371-6.282, and DVT 6.9 vs 4.2/1000 person-years, RR 1.615, 95% CI 0.369-9.679, respectively), thought the patients with cHP were younger, had lower BMI and less comorbidities. In recent years, a growing body of evidence indicate that the cHP and IPF may share commonalities in the mechanism involved in their pathogenesis and progression. Pathogenetic mechanism of lung fibrosis seen in IPF and cHP is triggered by repetitive lung parenchyma injuries caused in genetically susceptible individuals by the environmental factors [
26]. Additionally, some genetic variants such as shortened telomeres and the MUC5B promoter variant rs35705950 that contribute to IPF susceptibility, have been linked to cHP as well [
27]. Tissue injury initiates a cascade of events, one of the earliest of which is activation of the clotting system [
5]. There is good evidence that the coagulation cascade is activated in several fibrotic lung diseases, regardless of the factor initiating lung fibrosis, including IPF [
7,
8] and HP [
2].
Nevertheless, the increased risk of VTE in cHP cannot be explained only by the similarity of both diseases and it is also possible that elements unique to the disease may play a role. HP typically results from an immune-mediated reaction induced by an inhaled antigen leading to predominantly lymphocytic inflammatory pattern with granulomatous inflammation. However, neutrophilic inflammation may play a role early in the disease course and during subsequent fibrosis [
1]. Chronic inflammation, as a weak risk factor for venous thromboembolism, is known to provoke a hypercoagulable state similar to fibrosis. Activated leukocytes are the primary source of procoagulant tissue factor-positive microparticles that might locally activate the coagulation cascade. Moreover, neutrophils produce neutrophil extracellular traps (NETs) composed of DNA, histones, and microbial proteins that promote thrombus formation by providing a scaffold for red blood cells, platelets, and procoagulant molecules [
28].
The relationship between IPF and VTE or prothrombotic state was the subject of several epidemiological studies in the past [
9‐
13]. In large, general population-based study of people with IPF, Hubbard et al. [
9] found an increased risk of having DVT in patients with IPF compared with the general population both before and after the diagnosis of interstitial lung disease was first recorded. Subsequent studies in Denmark and the United States have confirmed these results. Sode et al. [
10] examined the entire Danish population from 1980 through 2007, using national registries to identify patients with interstitial pneumonia and VTE events. They showed that prior history of VTE was associated with an increased risk of developing idiopathic interstitial pneumonia, especially among those never treated with anticoagulants. An elevated risk of thromboembolic disease in patients with IPF was also demonstrated in the study using mortality data from all USA decedents [
11]. The authors identified IPF cases and three groups for comparison purposes: population based-controls, chronic obstructive pulmonary disease (COPD) patients and lung cancer patients. They found that decedents with IPF had a significantly greater risk of VTE than the comparison groups and that those with VTE and pulmonary fibrosis died at a younger age than those with pulmonary fibrosis alone. Moreover, a meta-analysis of 5 studies on the relationship between VTE and IPF showed more than a twofold increase in the risk of VTE in patients with IPF [
29].
Hitherto, the association between VTE and cHP has not been studied yet. After a careful search of the medical databases as PubMed, MEDLINE and Web of Science, we found only one case report regarding the association of acute HP due to Aspergillus hypersensitization and recurrent pulmonary embolism [
30]. The authors hypothesized that the inflammation induced by bioaerosol exposure causing HP was a temporal trigger to venous thrombotic events in their patient. Additionally, the research paper from the tertiary referral centre for ILDs has recently been published concerning the association between various comorbid conditions and survival in cHP [
31]. The authors have found that thromboembolic disorders occurred in 4.7% of patients with cHP and were associated with an improved survival. In our cHP group VTE was identified in 3.3% of cases. In contrary to the study of Wälscher et al., where the prevalence of thromboembolic disorders was assessed, we evaluated the incidence rate of VTE events in our cohorts, and additionally the patients with the known risk factors for thromboembolism as well as being treated with anticoagulation were excluded from the analysis.
The results of our study support the previous evidence of increased risk of VTE in patients with IPF, and additionally provide new data on the increased risk of DVT and PE with cHP, comparable to that of demonstrated in IPF. The possible explanation of these relationships might be, apart from the similar pathogenetic mechanisms resulting in the shift of the haemostasis balance in favour of the procoagulant and antifibrinolytic activity, simply the reduced mobility of patients with IPF and cHP due to respiratory symptoms and reduced exercise capacity. Furthermore, the increased risk of VTE could result from the exposure to glucocorticoids, the medication often applied in the treatment of cHP and in the past used in patients with IPF. The approach to IPF treatment changed in 2012 after the publication of the results of the PANTHER-IPF clinical trial, in which participants treated with the combined three-drug regimen of prednisone, azathioprine, and
N-acetylcysteine, compared to the placebo had higher mortality, more hospitalizations and serious adverse events [
32]. As our retrospective analysis included patients with IPF diagnosed between 2005 and 2012, some of them were treated with glucocorticosteroids and azathioprine. Our analysis revealed that the treatment with systemic steroids and more advanced disease were significant risk factors for VTE in patients with IPF. This was not demonstrated in the cHP group, where the vast majority of patients were treated with systemic steroids. On the other hand, it is worth mentioning that all patients with confirmed VTE in the cHP group were treated with systemic steroids. The association between glucocorticoids and VTE was examined in a nationwide population-based case-control study in the past. The authors found that glucocorticoid users had a dose-dependent increased risk of VTE [
33].
We found that in cHP group the presence of arterial hypertension and pulmonary hypertension significantly increased risk of VTE. In the study evaluating the epidemiology of VTE in large, well-characterized cohort of patients with systemic sclerosis, Johnson et al. [
34] reported that the presence of pulmonary arterial hypertension (PAH) was an independent risk factor for VTE and predictor of mortality in this disorder. Similar as in their study, this risk factor (and the other baseline characteristics we evaluated) was present prior to the occurrence of VTE. This may be of importance in the evaluation of patients with cHP who show the echocardiographic signs of pulmonary hypertension and have symptoms of worsening dyspnoea, where the clinicians should maintain a high index of suspicion for VTE.
Finally, we assessed the impact of VTE on survival concomitantly in the IPF and cHP groups due to the low number of VTE cases. We did not find any significant differences in median, 3- and 5-years survival between those with and without VTE. The exclusion of known risk factors for VTE, such as cancer, thrombophilia and major trauma, and the presence of effective treatment might have had an impact on this outcome.
The significant strength of our investigation is a relatively large size of well-diagnosed individuals with IPF and cHP as well as the confirmation of VTE event by imaging studies. We are confident of the diagnosis of IPF and cHP in our cohort as all patients were re-evaluated in a multidisciplinary team discussion in terms of the currently published guidelines [
14] or whether they would meet the criteria of cHP diagnosis according to the recently published reports [
15,
17]. Moreover, we excluded from the study patients with known potential factors increasing the risk of VTE. Thus, the VTE incidence rate increased to 7.1/1000 person-years in cHP and to 11.8/1000 person-years in IPF compared to 1–2.7/1000 person-years in the general population [
20‐
25], may indicate an association between these interstitial lung diseases and a prothrombotic state.
The aforementioned research papers concerning the link between VTE and ILD were medical registry-based studies [
9‐
11,
13], with the exception for the study by Navaratnam et al. [
12], where the diagnosis of IPF was made by clinicians who saw the patients. However, the diagnosis of VTE was established based on participants reports. Thus, potential limitation of these studies is misclassification of diagnoses as it was based on diagnostic code from the registry in the most of them.
There are some limitations of our study to consider. First and foremost, this was a retrospective, observational, single-center study and selection bias may have affected the outcomes of our research. Second, we reported only symptomatic VTE, as the investigations for VTE were based on symptoms and/or signs. We did not systematically screen for VTE all cases diagnosed with ILD. Additionally, VTE may have occurred in patients who were lost to follow-up. Thus, these factors might be the possible causes of underestimation of the incidence of VTE in patients with IPF and cHP. The authors are aware that the results of the study should be treated with caution due to statistical analyzes on a small group of patients due to the rarity of evaluated diseases. It is worth noting that symptomatic patients occurred with similar frequency in both cohorts, as the CT angiography was performed in comparable percentage of patients in both groups, and this reflects the real-world practice.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.