Pleuroparenchymal Fibroelastosis and Serositis as Pleural Complications after Hematopoietic Stem Cell and Lung Transplantation

2.1 Clinical presentation

The clinical presentation of patients with symptomatic serositis after HSCT or lung transplantation depends on the organ involved. Presentation can be with shortness of breath of an acute or sub-acute duration, chest discomfort, abdominal fullness, light headedness, or dizziness due to accumulating fluid in the pleural, pericardial, or peritoneal, spaces [9, 16, 18, 19]. With pleural effusions, patients may also present with cough, that is dry in character, as well as chest pressure or the inability to take a deep breath. Most concerning are patients presenting with the rapid accumulation of fluid in the pericardial space leading to tamponade or in the pleural space causing acute respiratory failure [30, 31]. These patients require immediate evacuation of the fluid for stabilization. However, typically for pleural effusions after HSCT, based, on our experience and the literature, effusions increase over several months before patients develop symptoms and infrequently require emergent drainage for symptom control ([18, 19], and unpublished observations).

2.2 Pathophysiology and potential biomarkers for post-HSCT and lung transplant serositis

The underlying pathophysiology of post-HSCT and lung transplant induced serositis is poorly understood. Due to the rarity of this complication, only limited investigations have been undertaken into the pathogenesis and mechanism of disease either at the time of onset or during progression. At present there are a limited number of clinical markers available to predict those at an increased risk for developing serositis or that are useful to assess the potential for progressive disease and a poor clinical outcome. In contrast to SLE associated serositis, where male sex was identified as an independent risk factor for pleural effusions [28], a strong age or gender predilection for cGvHD-associated serositis has not been described, either in the literature or in our clinical observations ([18, 19] and unpublished observations). This is also true for the type of underlying malignancy requiring HSCT. Our observations further support these findings as neither age, gender, nor type of underlying malignancy were found to be possible risk factors for the development of post-HSCT cGvHD-associated serositis (unpublished observations). However for HSCT, donor type may be important. Seber et al. [32] in their study of post-HSCT serositis reported that 66.6% of unexplained effusions occurred in mismatched and unrelated donors in association with both acute and chronic GvHD. In our group of post-HSCT patients with serositis four patients (33.3%) received transplants from unrelated donors, out of which three had very severe serositis, with one succumbing to the disease (unpublished observations). Our findings therefore support those seen by Seber et al. [32]. Taken together, having an unrelated donor may be a risk factor for the development of serositis after HSCT. In addition, although studies have not described an association between the conditioning regimen and post-HSCT serositis, an increased incidence of acute and chronic GvHD [33] and constrictive pericarditis [34] has been reported in patients receiving TBI as a part of their conditioning regimen. All but one of the patients that we have seen who developed cGvHD pleural effusions after HSCT received TBI (91.6%). Thus, TBI may also be a possible risk factor for the onset serositis in patients with severe cGvHD.

At present there are no serum or pleural/pericardial fluid biomarkers to assist in the diagnosis, risk stratification, or use to target therapy or predict outcomes in patients with serositis after HSCT or lung transplantation. Regarding post-HSCT associate serositis, one study suggested that an acute decrease in serum albumin or increase in circulating monocytes may allow for identification of patients at risk for serositis after HSCT but this has not been followed up upon (19 Leonard). In other reports viral respiratory infections have been suggested to trigger the onset of serositis after HSCT [18]. In our patients we have seen this association in a small number but these findings have not been consistent across our whole patient population (unpublished observations). In contrast, for lung transplant patients that develop late onset pleural effusions, no studies to our knowledge have been undertaken to explore possible biomarkers. Furthermore, no published reports have extensively characterized the pleural or pericardial fluid in post-HSCT with serositis or post lung transplant patients that develop late onset pleural effusions, apart from a description of routine biochemical testing to determine exudative or transudative effusions by Lights criteria and exclusion of infection. A more extensive analysis of pleural or pericardial fluid from serositis patients may point to potential mechanisms underlying the pathogenesis of this complication and allow for the design of targeted therapies and predictors of outcomes.

2.3 Treatment for post-transplant serositis

In regards to treatment for serositis after transplantation, no clinical trials have been published outlining management of serositis after HSCT or lung transplant. At present treatment is focused on the drainage of symptomatic effusions for alleviation of symptoms and determination of other causes. Specifically for post-HSCT serositis associated pleural effusions, augmentation in immunosuppression using corticosteroids, tacrolimus and/or sirolimus has been attempted to prevent fluid re-accumulation [18, 19]. Other modalities tried have included extracorporeal photophoresis, etanercept, and rituximab [18] with variable success. Although an increase in the immunosuppression regimen was successful in a majority of post-HSCT patients with serositis treated by Modi et al., this approach was only effective in a quarter of our patients ([18] and Unpublished observation). For those unresponsive to an increase in immunosuppression, treatment is generally repeated procedures for fluid drainage to obtain symptom control. As these effusions are typically chronic in duration, this approach is sub-optimal to control patients’ symptoms and requires frequent invasive procedures. Pleurodesis, involving the chemical irritation of the pleural space leading to adhesion of the visceral and parietal pleura, has been shown to be effective in a few patients with post-transplant serositis that has been recurrent [35]. The large volume of fluid associated with these pleural effusions, and the rapid re-accumulation that is typically seen, limits the effectiveness of this treatment modality. As stated above, although serositis as a cause for recurrent or non-resolving pleural effusions has not been explicitly described in post lung transplant patients, up to 41% of late onset effusions have been determined to be non-specific and therefore potentially serositis [9]. Detailed management of these non-specific late onset pleural effusions after lung transplant have not been described in the literature. Further research is necessary to determine the risk factors and underlying mechanisms leading to serositis after HSCT and lung transplantation in order to design effective treatments and improve patient outcomes.

3.1 Overview

In addition to serositis, another rare and poorly understood pleural complication after HSCT and lung transplant is pleuroparenchymal fibroelastosis (PPFE). In the most recent review of the idiopathic interstitial pneumonias by the American Thoracic Society in 2013, PPFE was classified as a separate entity [36]. The disease was initially described by Amitani as an upper lobe predominant pleural based disease distinct from apical cap disease and termed pulmonary upper lobe fibrosis (PULF) [37]. It was later renamed pleuroparenchymal fibroelastosis (PPFE) by Frankel in their case series of patients with that name now established to describe the disease [38]. Since its initial description, the number of cases of PPFE continues to increase [36, 37, 38, 39]. At present the diagnosis of PPFE is determined by the distinct radiographic pattern seen on chest X-ray or CT imaging as there is no specific definitive biomarker or test to establish the diagnosis apart from the radiographic description [40, 41]. Establishing a diagnosis of PPFE in post HSCT or lung transplant patients is important to avoid unnecessary or harmful procedures, to assess for the development of complications, and to provide information to patients and family members regarding the typical clinical course and survival. Below we will discuss the incidence, risk factors, clinical presentation, diagnostic work-up, present state of biomarkers, treatment and outcomes for PPFE with a focus on that which occurs after HSCT or lung transplantation.

3.2 Incidence of PPFE

Pleuroparenchymal fibroelastosis after HSCT or lung transplant is an uncommon to rare complication with specific differences in prevalence between the two groups. The overall prevalence of PPFE after HSCT is 0.28–3.3% [42, 43, 44] with a much higher prevalence rate (7.54–15%) seen after lung transplant [42, 45]. The reason for these significant differences in prevalence between the two groups is unknown but may point to specific risk factors or infectious or non-infectious exposure causing PPFE in each group. These differences remain unknown. For both post HSCT and lung transplantation, the onset of PPFE is much later in the course after transplant. Specifically for HSCT, unlike the other long term non-infectious pulmonary complications which occur within 100 days to 2 years after transplant, PPFE occurs much later with a median time to diagnosis of 6.9 years, with a wide time range (0.25–17.9 years) [10, 11, 13, 15, 17, 22, 42, 44, 46, 47]. This wide range in time to onset brings up the question if the PPFE diagnosed at the extreme ends of the time spectrum are the same or are different disease processes. At present it is unclear if an earlier manifestation is a separate and distinct disease process or if the etiologic cause is different than those diagnosed much later with PPFE. No studies to date have identified differences in exposures or potential risk factors between the two groups. For PPFE after lung transplant, the reason for the higher prevalence is unknown. As PPFE after lung transplant is combined in those with the restrictive allograft syndrome (RAS) subtype of chronic lung allograft dysfunction (CLAD), it makes it challenging to isolate it as a specific disease process when reviewing studies to determine individual risk factors or true prevalence [48, 49, 50].

Specifically for post-HSCT PPFE, the disease is mostly confined to those who have undergone allogeneic transplant, although a few cases are seen in patients who received an autologous transplant. [22, 42]. As such, PPFE may be another manifestation of chronic graft versus host disease (cGvHD) and thereby suggesting similar pathophysiology for disease onset. However, as most other organ involvement of cGvHD occurs shortly after 100 days post-transplant and PPFE occurs much later, the underlying cause for PPFE may be unique compared to cGvHD as a whole. Typically post-HSCT patients that develop PPFE have established lung associated cGvHD (i.e. bronchiolitis obliterans syndrome (BOS), based on several studies and the experience at our institution ([22, 43, 46, 47, 51], unpublished observations) with the co-existence of BOS in patients diagnosed with PPFE being high at 47–100% [22, 46, 47, 51]. Taken together these findings suggest a similar underlying etiology and pathophysiology for BOS and PPFE. However mention needs to be made that as the average time to onset for BOS and PPFE after HSCT are different (BOS 1.5 years vs. PPFE 6.9 years) there may still be unique differences in the underlying cause and pathophysiology between the two entities. This is also true for differences seen in the incidence of the two diseases where BOS occurs in 5–10% of post-HSCT patients, but PPFE only occurs in a much smaller group of 0.2–0.3% of all patients undergoing allogeneic HSCT [18, 19, 46, 47]. This suggest the possibility of a “two hit” pathophysiology.

Similar to after HSCT, a similar entity to PPFE has been described in patients who have undergone lung transplantation [45, 52, 53]. However, unlike in post-HSCT patients, PPFE or PPFE-like lesions are grouped into the sub-category of restrictive allograft syndrome (RAS) under the broader post-lung transplant complication of chronic lung allograft dysfunction (CLAD), which includes the classic post-lung transplant chronic rejection entity of bronchiolitis obliterans syndrome (BOS) as well as the newer entity of RAS [48, 49, 50]. This makes the specific investigation of PPFE more challenging in these specific patients. A PPFE like disease in post lung transplant patients was initially described from researchers from the Toronto and Duke lung transplant groups upon a retrospective review of their post lung transplant patients over a 10 year time frame [52, 53]. In their study, the onset of an upper lobe predominant fibrotic lung disease occurred in 1.9% of their patients with a range of onset of 0.41–8 years and a median time to onset of 1.9 years. Similar to as was described above for post HSCT, the range of onset for PPFE after lung transplant is quite wide. In post-lung transplant patients the progression of PPFE appears to be quicker than after HSCT where in one study the calculated upper lobe volume loss was 45% over time as assessed by CT imaging [52]. Development of a pneumothorax was also frequent in this patient group with 71% developing a pneumothorax during follow-up [52]. On pulmonary function testing lung volumes were decreased, as would be expected with a fibrotic lung disease. In addition, in most patients the disease progressed slowly with radiographic findings present for several months prior to the onset of clinical symptoms [52, 53]. Collectively these findings are similar to those in post-HSCT patients that develop PPFE. As mentioned above, since these initial descriptions of PPFE after lung transplant, PPFE or PPFE-like findings after lung transplant have been grouped into the restrictive allograft syndrome, a poorly understood post transplant complication that was officially listed as a complication diagnosis by the International Society of Heart and Lung transplantation (ISHLT) in their update in 2019 [48]. Restrictive allograft syndrome occurs in 30–35% of patients who undergo lung transplantation and is present alone as a form of chronic rejection or is seen co-existing with bronchiolitis obliterans syndrome (BOS) [48, 49, 50, 54]. PPFE is a subset within the group of RAS without clear distinctions of specific groups within RAS identified by ISHLT [48, 49, 54].

3.3 Etiology and risk factors for PPFE after HSCT or lung transplantation

The cause of PPFE after HSCT or lung transplant is currently unknown (Table 1). For post-HSCT PPFE two possibilities have specifically been discussed, adverse reaction to chemotherapy or radiation used to treat the underlying disease or during the conditioning regimen prior to undergoing HSCT. Chemotherapeutic agents that have been implicated in causing PPFE include cyclophosphamide, and carmustine, which are alkylating agents [39, 47, 51, 55]. Taking away from these agents as possible causes for PPFE is that the overall incidence of PPFE after HSCT is very low whereas the number of transplant patients receiving alkylating agents is very high. In addition the long duration between the time of receiving alkylating chemotherapy and the onset of PPFE, which averages several years, lessens the potential association. Whole body irradiation, used during the conditioning period prior to HSCT, has been examined as a possible etiology for the later development PPFE but as only a few such treated patients later develop PPFE it is an unlikely causative agent [39, 47]. Similar to use of chemotherapeutic agents, An additional cause for both post HSCT, as well as lung transplant recipients, in developing PPFE are respiratory infections, particularly viral infections that occur after transplant [39, 42, 48, 51, 53, 55]. This makes clinical sense and would explain how PPFE can occur at all times post-transplant but particularly later in the course. All types of infections have been suggested to lead to PPFE after HSCT including viral, bacterial, and atypical pathogens. Interestingly there have been recent descriptions of an association of infection with atypical Mycobacteria infection and the development of PPFE in both HSCT and lung transplant patients [53, 56, 57]. Interestingly, this association has been seen both in non-transplant patients with atypical Mycobacteria infection who go on to develop PPFE and those diagnosed with PPFE that also were found to be infected with atypical Mycobacteria [56, 57]. Two large retrospective studies in those with atypical Mycobacterial infection showed the prevalence of PPFE was 11.9–26.3% [56, 57]. To date, however, the association of atypical Mycobacteria infection and PPFE in post HSCT or lung transplant patients has not been extensively examined with only a few case reports or isolated patient reports describing a co-association. At present no infectious cause for PPFE in post-HSCT or lung transplant patients has been identified. A lack of an animal model for PPFE limits the future exploration of a potential link between pulmonary infection and the later development of PPFE. Published cases that have implicated infections as a cause for PPFE in those have undergone HSCT have not provided details of infection type, nor the timing and duration of the infection prior to the development of PPFE [39, 42]. As respiratory infections are common in all patients, and especially after HSCT due to an immunocompromised state, this deflects that respiratory infections are a cause for PPFE after HSCT as PPFE is very rare. In addition, in the studies suggesting a connection between respiratory infections and PPFE, the timing from a purported respiratory infection and PPFE is long, sometimes measured in years, and there is a lack of details describing the severity of the respiratory infection or the duration of the illness [42, 51].

Similar to post-HSCT PPFE, the etiology for RAS, and in particular PPFE, after lung transplant have not been determined. Etiologic investigation after lung transplant has encompassed a broader examination for the cause of chronic lung rejection after lung transplant (aka CLAD). These have included allograft rejection, humoral immunity, gastroesophogeal reflux disease (GERD), and infections [48, 50]. In addition, on reviewing demographics, no clear risk factors have been identified when examining age, sex, reason for transplant, surgical or post-surgical hospital course or complications, or medications [49, 54]. To date, no specific studies have been undertaken in a group of patients with RAS with predominant features of PPFE to determine specific etiologic factors. For CLAD and RAS as a whole, possible etiologic factors examined have included post-transplant hospital course, pre- and post-transplant immunosuppressive medications, GERD with aspiration, airway irritant exposure, rejection, and infections [48, 50, 53, 58]. The development of acute respiratory failure with the presence of diffuse alveolar damage (DAD) on lung biopsy results greater than 3 months after transplant correlated with the onset of RAS [48, 58, 59]. Similar findings of DAD in patients who have developed RAS was also seen by Ofek et al. [49]. Importantly these findings of DAD were observed in RAS patients who specifically were found to have PPFE with these findings present one or more years prior to the diagnosis of RAS [52, 53]. Taken together, this suggests that acute inflammation, and its effects may be a risk factor for PPFE. In addition, several studies have shown that pro-inflammatory mediators are increased in RAS patients including cytokines (Interleukins 5, 6, 8, CXCL10, and 1β), alarmins, and neutrophil elastase [48, 59, 60, 61, 62, 63, 64, 65, 66, 67]. Further suggesting that acute inflammation and possibly the development of DAD may predispose to the development of PPFE. The potential role of these pro-inflammatory mediators will be discussed below further under Biomarkers. Interestingly, levels of vascular endothelial growth factor (VEGF) is decreased in RAS patients [48]. This finding correlates with the observation of vascular injury on lung biopsies from patients with RAS and PPFE [67] and the decrease in vascular perfusion on ventilation perfusion scans taken on patients with an upper lobe predominant fibrosis after lung transplant [53]. Taken together, this suggests that a vascular injury after lung transplantation may place a patient at a higher risk for the development of RAS and therefore PPFE. The specific role of VEFGF in the development of PPFE after lung transplant has not been specifically examined. In regards to an alloimmune response as the cause of RAS and PPFE after lung transplant several avenues have been investigated. There has been no correlation for the development of RAS/PPFE and the use of immunosuppressive medications after transplant [50]. In contrast, the possibility that the presence of donor associated antibodies and antibody mediated rejection (AMR) and the development of RAS/PPFE have been investigated by several groups. In the study by Todd et al., there was an increased incidence of newly detected DSA that correlated to the onset of RAS [68]. Similar findings were reported by Verleden et al. in their group of patients [69]. These findings potentially fit with the observation of an increase in total immunoglobulin G levels in post-lung transplant patients that developed RAS [48]. The presence of GERD was thought be strongly associated with CLAD, and in particular BOS, based on lung biopsies showing multinucleated giant cells and lipid laden macrophages suggestive of chronic aspiration and clinical improvement in those who underwent fundoplication [48, 50, 53]. However, this has been not thought to definitively be a cause of RAS/PPFE. Finally, post-transplant infectious complications, similar to post-HSCT, have been examined as a potential cause for RAS/PPFE in patients who have undergone lung transplantation but no clear association has emerged. As discussed above for PPFE after HSCT, infection with atypical Mycobacteria has been suggested to be as cause of RAS/PPFE after lung transplantation. In the series by Pakhale et al. 5 patients had cultures positive for atypical Mycobacteria, with only two having positive cultures prior to the presence of radiographic changes [53]. As these patients did not have clinical disease suggestive of an atypical Mycobacterial infection, this was not felt to be an active infection and was not treated and accordingly was not felt to be the definite cause of PPFE.

3.4 Clinical presentation

The clinical presentation of PPFE after HSCT or RAS/PPFE after lung transplant is similar to cases from other etiologies or those considered to be idiopathic (Table 2). Presentations are of progressive shortness of breath and cough [22, 43, 45, 47, 52,53]. In the majority of cases the cough is non-productive with the shortness of breath slowly progressive over time with the diagnosis of PPFE commonly made several years after the onset of symptoms, with the an average time of 2.5 years, although the time from symptom onset to diagnosis is broad (0.5–6 years) [43, 52, 53, 54, 70]. As such a low level of suspicion for PPFE or Restrictive allograft syndrome (RAS)/PPFE must be maintained to ensure close clinical follow-up for diagnosis. Earlier identification may allow for quicker intervention or prevention of disease progression. Patients frequently have chest pain that is pleuritic in nature but this is not seen in all patients with PPFE [22, 39, 70]. In contrast, extrathoracic symptoms are unusual in patients with post-HSCT PPFE or RAS/PPFE after lung transplant. Patients typically have a history of weight loss with a low body mass index (BMI) with the low BMI used as one diagnostic criteria for PPFE and which can help in differentiating PPFE from the other types of IIP [43, 45, 70]. With an upper lobe predominance for PPFE causing fibrosis in the upper lobes with resultant hilar retraction, patients with PPFE develop a flattened chest or platythorax (Figure 1) [43, 45, 70]. Platythorax severity is calculated via the longitudinal versus transverse ratio at the level of the 6th thoracic vertebra [43]. On lung examination, the presence of crackles are less common in patients with PPFE, compared to those with idiopathic pulmonary fibrosis (IPF) or non-specific interstitial pneumonitis (NSIP),with crackles only heard in patients with PPFE who have co-existing lower lobe involvement, this typically being UIP [45]. In those patients with RAS/PPFE after lung transplant, an overall diminishment of breath sounds or wheezing may been heard in patients with RAS/PPFE due to the concomitant presence of BOS [52, 53]. Nail clubbing is less frequently seen in PPFE or RAS/PPFE after lung transplant, compared to that in other chronic lung diseases, with it being present in only 4–25% of patients with PPFE after HSCT [39].

A diagnosis of PPFE is commonly suspected based on the unique findings observed on chest imaging as the presenting symptoms described above are non-specific. Imaging reveals upper lobes pleural thickening with subpleural fibrosis consisting of characteristic bands of fibrous tissue extending from the pleura to the hilum, with the upper lobe fibrosis resulting in the upward retraction of the hilum (Figure 2) [43, 45, 49, 52, 53, 71, 72, 73]. Pneumothoraces may also be seen on chest imaging at the time of presentation or during the disease course [39, 40, 51, 52, 53, 70] with pneumothoraces common in these patients occurring in 25–75% of PPFE patients after HSCT during their illness [39, 40, 51]. This is also true in post-lung transplant patients with PPFE. In the initial description of an upper lobe fibrotic ling disease in lung transplant patients by Konen, et al. 71% of their patients had a pneumothorax [52]. Patients with PPFE are at a particular risk of developing a pneumothorax after undergoing a surgical lung biopsy for diagnosis with these commonly persisting for prolonged periods of time after the procedure [39, 45]. In contrast, pleural effusions are rare in patients with PPFE or RAS/PPFE in the absence of cardiac or renal dysfunction or volume overload. Similar to chest x-ray imaging, chest CT scans more clearly reveal the pleural thickening and fibrotic bands extending from the pleura to the hilum (Figure 3). Although predominately an upper lobe process, post HSCT and lung transplant RAS/PPFE does less frequently involve the lower lobes and when involved a third of the patients will have an equal distribution between the upper and lower lobes, particularly later in the course of the disease [41, 45, 46, 53, 74]. The pattern on imaging in the lower lobes is variable compared to that in the upper lobes and can show the typical pleural thickening with fibrous bands extending toward the hilum as seen with upper lobe involvement, but may also show interstitial thickening indicative of interstitial lung disease [40, 45, 46, 47, 52, 53, 71]. Pathologically these findings from the lower lobes may be consistent with PPFE but may also show UIP, NSIP or even bronchiolitis obliterans with organizing pneumonia [41, 45, 46, 47, 52, 53, 71]. This occurs in 32–75% of all patients with PPFE after HSCT and lung transplant but may also be seen in the idiopathic forms of PPFE [31, 45, 46, 47, 52, 53, 71]. Differentiating the presence of UIP (clinically idiopathic pulmonary fibrosis) or NSIP co-existing with PPFE can be challenging. In cases of NSIP and IPF upper lobe findings of fibrotic bands and subpleural thickening on imaging would be atypical and therefore not suggestive of either UIP or NSIP. The only way to definitively diagnose a concomitant interstitial lung disease in the setting of PPFE would be an open lung biopsy. However, as stated above, an open lung biopsy is associated with a significant risk of a prolonged pneumothorax with PPFE and accordingly lung biopsy should be avoided in these patients. The use biomarkers, as will be discussed later, may be clinically helpful in differentiating concomitant interstitial lung disease and PPFE. Although this has not been fully explored, a non-invasive test to diagnose PPFE and exclude other interstitial lung diseases would be of great assistance. Finally, the significance of finding a co-existing interstitial lung disease in patients with post HSCT PPFE or RAS/PPFE after lung transplant, as it relates to etiology and clinical outcomes, is at present not fully known but the presence of lower lobe involvement may influence the clinical course and overall prognosis. The finding of ILD in post lung transplant patients would raise concerns of possible recurrence of the primary disease that led to lung transplantation.

3.5 Pulmonary physiology

Findings on pulmonary function testing performed on patients with PPFE show a significant decline (typically below 60%) in the forced vital capacity (FVC) and an increase in the residual volume to total lung capacity (RV/TLC) ratio [39, 43, 45, 47, 70, 74] with the RV/TLC ratio inversely correlated with the FVC [74]. This RV/TLC ratio of >115% of normal can be useful in differentiating PPFE from the other idiopathic interstitial pneumonias, particularly in those with lower lobe involvement and concerns for a concomitant ILD, as lung biopsy is prohibitive in these patients, as described above [39, 75]. The diffusion capacity is also decreased in PPFE but to a lesser extent than is seen in idiopathic pulmonary fibrosis [39, 45]. Caution in using a decline in FVC in patients who have undergone lung transplant as those patients with RAS/PPFE also commonly have concomitant bronchiolitis obliterans syndrome and therefore mixed disease [48, 49]. Co-existing BOS and RAS occurs in several patients with CLAD after lung transplant with several patients developing BOS followed by RAS [48]. In such a situation the FVC will be decreased with also a decline in the FEV-1/FVC ratio for BOS, whereas in RAS/PPFE the FEV-1/FVC will be increased. In the situation of a decline in FVC but a normal FEV-1/FVC ratio, concomitant BOS and RAS/PPFE must be considered and requires clinical expertise. No detailed algorithm using data from pulmonary function testing has been published to identify patients with combined BOS and RAS/PPFE.

In contrast to the findings on pulmonary function testing, chest imaging, and lung histology with sub-pleural fibrosis and decrease in lung volumes, interestingly exercise capacity is not diminished and oxygen saturation remain normal patients with PPFE, even with exertion, at the time of diagnosis [70]. Further studies need to be performed to determine exercise tolerance and future need for oxygen in these patients long-term. There are several longer-term follow-up studies showing that patients with PPFE after HSCT or PPE/RAS after lung transplant develop progressive disease leading to respiratory failure and the need for supplemental oxygen [39, 40, 47, 52, 53]. Typically repeat pulmonary function tests in most patients show a gradual decrease in the FVC with a corresponding increase in the RV/TLC ratio over time [42, 53, 71, 74]. However, in some patients this change occurs more rapid with a increase in their symptoms and findings of desaturation at rest or with activity [75, 76].

3.6 Challenges with establishing a diagnosis

The initial patients described with post-HSCT PPFE, or upper lobe fibrosis after lung transplant, had a diagnosis confirmed via histologic findings by either lung biopsy or at the time of autopsy [22, 38, 46, 52, 53, 71]. However, patients with PPFE are at high risk for pneumothorax after undergoing lung biopsy, with this pneumothorax potentially persisting for some time after surgery [39, 40, 70, 73]. In addition to a pneumothorax as a complication from a lung biopsy there is a risk for the development of acute respiratory failure due to the presence of co-existing ILD, particularly UIP, and the development of an acute UIP exacerbation [40]. Accordingly, confirmation of PPFE by surgical biopsy should be avoided in these patients. Other forms of lung biopsy (e.g. transbronchial or cryobiopsy via bronchoscope) should also be avoided due the risk for pneumothorax [77, 78]. In addition, the yield from these biopsies techniques are much lower with the risk for complications outweighing the potential diagnostic yield [77, 78]. Recently, several studies have outlined protocols to be used for establishing a diagnosis of PPFE after HSCT in leu of obtaining lung tissue. Early studies focused solely on radiographic imaging for the diagnosis of PPFE [40, 71]. In the study by Reddy et al. radiologic diagnostic criteria were outlined to separate patients into definite, consistent with, or inconsistent with the diagnosis of PPFE with these findings then paired with lung tissue histologic that were similarly graded as definite, consistent with, or inconsistent with PPFE [71]. A limitation of the study by Reddy et al. is that lung tissue was required to confirm the diagnosis of PPFE, which is commonly not present due to current guidelines [71]. However, findings of a correlate between radiographic imaging and histology is extremely useful in establishing a radiographic only diagnosis of PPFE, similar to that commonly now used in multidisciplinary group (MDG) conferences for UIP. Another limitation of this study was that no attempts were made to establish a correlation between clinical symptoms or lung physiology and histologic and chest imaging findings in their patients. Such a study would then mirror the usefulness of MDG conference for the diagnosis of UIP. In a later study, Enomoto, et al. described their approach to the clinical diagnosis of PPFE in a multicenter study by only chest CT imaging utilizing a similar grading system as that used by Reddy et al. [40, 71] wherein their study required radiologic progression as a confirmation of PPFE to avoid including patients with the benign process apical cap disease [40, 71]. A limitation of this study is the requirement for radiologic progression to confirm a diagnosis of PPFE as PPFE in most patients slowly progresses thereby potentially missing patients with PPFE and delaying the diagnosis and possibility for closer follow-up or initiation of treatment and involvement of patients in clinical trials. Finally, Watanabe et al. proposed an algorithm for the diagnosis of PPFE in patients that did or did not undergo a surgical lung biopsy (SLB) with patients placed into diagnostic groups based on their imaging and lung histology, when available [79]. In contrast to other studies, the study by Watanabe et al. used clinical signs and symptoms to generate their algorithm [79]. A definite diagnosis of PPFE was made in those with histologic findings consistent of PPFE in the setting of upper lobe findings of subpleural consolidation and fibrosis on radiographic imaging. However, if lower lobe findings were observed on imaging this required discussion by a MDG to establish a diagnosis. As this study also included clinical data, this study was the first to describe distinct diagnostic groups being radiologically probable PPFE and those with both radiologic and physiologic findings that were probable for PPFE, with the latter group needing respiratory symptoms or respiratory symptoms plus the physiologic abnormalities of a low BMI, decrease in FVC, and an increase in the RV/TLC ratio [79]. This algorithm at present seems to be the optimal way to diagnose PPFE as it only requires available radiologic imaging, clinical findings, and pulmonary function results without the need for disease progression to confirm a diagnosis of PPFE. In contrast to PPFE after HSCT or the idiopathic forms, no detailed studies have been published of diagnosing PPFE in RAS patients after lung transplant.

At present the approach as outlined in the study by Watanabe et al. is now considered to be standard for the diagnosis of PPFE [79]. As an extension to this study, similar to the protocols used in the diagnosis of other IIP’s, use of an multidiscipilinary group (MDG) that involves clinicians, radiologists, and pathologists has been suggested as a way to diagnosis PPFE [79, 80]. Use of this approach was outlined in in a review of MDG confirmed cases of PPFE from a nationwide database in Japan examining clinical and radiologic features [80]. We would advocate using the algorithm as outlined by Watanabe et al. in considering a diagnosis of PPFE with inclusion of a MDG to confirm the diagnosis. This will allow for an earlier diagnosis of PPFE and facilitate inclusion of patients in research trials and attempts at new therapeutic strategies. Stressing the importance of using a MDG to diagnose PPFE is that at present, apart from symptoms, lung physiology, and radiographic findings, there are no established body fluid biomarkers to make a diagnosis of PPFE. The lack of such diagnostic biomarkers limits the ability to conclusively make a diagnosis of PPFE and more importantly does not allow for the identification of patients at risk for the rapid progression of PPFE and the corresponding poor prognosis in these patients. The current status of biomarkers in the diagnosis of post-HSCT PPFE and RAS/PPFE after lung transplantation will be discussed below.

3.7 Biomarkers

The availability of biomarkers to assist in the diagnosis of PPFE and RAS/PPFE would be extremely helpful, both in establishing a diagnosis and for the identification of novel therapies. Unfortunately, at present, no such biomarker exists for the definitive diagnose PPFE or RAS/PPFE. The use of biomarkers would be helpful to identify PPFE patients at risk for rapid progression in order to target this group for enrollment in therapeutic clinical trials particularly as the rate of progression and mortality in PPFE and RAS/PPFE is quite variable. Biomarker studies for PPFE and RAS/PPFE have explored the use of urinary desmosines, Krebs von den Lungen-6 (KL6), surfactant protein D (SPD), latent transforming growth factor-β binding protein 4 (LTBP-4), lipocalcin-2, matrix metalloproteinase 9 (MMP-9), and cartilage oligomeric matrix protein (COMP) (Table 3) [39, 61, 62, 63, 64, 65, 80, 81, 82, 83, 84, 85, 86, 87]. Early studies examining these biomarkers suggest that they may be helpful in the diagnosis of PPFE and RAS/PPFE and/or their prognosis and clinical outcomes, however further studies are necessary to confirm these preliminary studies. As discussed below, the potential use of biomarkers in PPFE has focused on those with idiopathic PPFE, but hope would be that these results could be extended to patients with PPFE from defined causes, including post_HSCT and RAS/PPFE. At present, no study has examined biomarkers in patients specifically with post-HSCT PPFE. In contrast, the studies described below in patients with RAS are specific to that diagnosis but not specifically to RAS/PPFE.

Desmosines are breakdown products of elastase. As patients with PPFE have active lung fibrosis it was hypothesized that urinary levels of desmosines would be higher in PPFE patients [82]. In addition, as fibrosis rates are higher in patients with PPFE compared to IPF, an elevation in desmosines breakdown products may be useful in differentiating between PPFE and IPF [82]. This possibility was explored in a study measuring urinary desmosine levels in patients with biopsy proven PPFE compared to levels in patients with idiopathic pulmonary fibrosis (IPF), chronic obstructive disease (COPD), or controls [82]. Findings from this study showed that urinary desmosines levels were significantly higher in PPFE patients compared to control patients and that they were also higher in those with chronic obstructive pulmonary disease (COPD). Furthermore desmosines levels were higher in those with PPFE compared to IPF and that levels were able to discriminate between the diagnosis of PPFE and IPF. In contrast, urinary desmosine levels did not correlate with clinical parameters (e.g. BMI), physiology(pulmonary function testing, PaO₂, or 6 minute walk distance), or disease severity [82]. Accordingly, this does allow for the use of urinary desmosines in a multidisciplinary approach to the diagnosis of PPFE at this time. As only a single time point was examined in this study, it is not clear if urinary desmosines levels are useful longitudinally to determine disease progression, morbidity, or prognosis. Measurements of urinary desmosines may be clinically useful at present to differentiate between PPFE and IPF in patients with lower lobe involvement or imaging findings that are suggestive of IPF. Further studies are necessary to determine the role of urinary desmosines in the diagnosis of PPFE and if they are useful in the diagnosis of RAS/PPFE, wherein they have not been examined.

Levels of KL-6 have been shown to be useful in the diagnosis of PPFE, with the ability to differentiate PPFE from other forms of idiopathic interstitial pneumonias (IIP), and have also been beneficial in the diagnosis of RAS/PPFE and the ability to differentiate between BOS and RAS in patients with CLAD [63, 65, 82, 86, 87]. In addition, KL-6 levels also appear to be helpful for prognostics. Levels of KL-6 between those with PPFE and IPF were examined by Oyama et al. and it was found that levels of KL-6 were significantly higher in IPF patients compared to those with PPFE although no potential correlation between KL-6 levels and clinical characteristics or physiologic variables in the PPFE group were explored [82]. In a later study by Ishii et al. ([80] this specific question was addressed. Results from that study showed that PPFE patients with an elevated level of KL-6 had a decreased median survival (24 months shorter) over a 5 year follow-up period than those with KL-6 levels under 600 U/ml [80]. These findings were replicated by Kinoshita et al. wherein patients with KL-6 levels over 550 U/ml had an overall poorer prognosis [83]. Similarly, in studying systemic sclerosis patients with PPFE, d’Alessandro et al. found that patients with a significant decline in their FVC and FEV-1, indicative of clinical progression, also had increasing serum KL-6 levels [81]. In that study KL-6 levels were measured serially over a 6 year time frame which increased the study’s clinical relevance. Although these studies would suggest a role for KL-6 in the diagnosis and management of PPFE, a small study by Sato et al. found no difference in KL-6 levels in PPFE patients [84]. Accordingly, further investigation is necessary to establish a role of KL-6 in the diagnosis or management of idiopathic and secondary PPFE.

Regarding the use of KL-6 in the diagnosis of RAS/PPFE after lung transplantation, levels of KL-6 in the serum and bronchoalveolar lavage fluid have been examined in patients with CLAD to determine both the diagnosis of CLAD and to differentiate those with BOS from those with RAS/PPFE. The initial study by Walter et al. examined if serum KL-6 levels could differentiate post lung transplant patients with BOS versus transplant patients without BOS as well as from normal controls [86]. KL-6 were significantly increased in post lung transplant patients with BOS compared to those without BOS and to normal controls with mean levels over 2 fold higher whereas the levels in post lung transplant patients without BOS did not significantly differ compared to healthy controls. This study was performed prior to the description of RAS as a post lung transplant complication and therefore patients with RAS may have been included in the BOS group and therefore results cannot be directly extended to RAS/PPFE. Other limitations of this study were the small sample size, that only a single time point was examined, and there was no discussion regarding the timing of blood collection as it relates to the time after lung transplantation. Similar findings were observed by Ohshimo et al. when they examined serum KL-6 levels in post lung transplant patients with BOS compared to those without BOS as well as healthy controls [87]. A strength of this study is that serial serum KL-6 levels were obtained at the time patients underwent bronchoscopy for clinical indications. Again, serum KL-6 levels were significantly increased in post lung transplant patients with BOS compared to those without BOS and to healthy controls. Mean KL-6 levels were over two fold higher in BOS patients compared to healthy controls and almost two fold higher when compared to patients without BOS. The study did not discuss if KL-6 levels changed over time in the groups. Importantly this studied compared the KL-6 levels between the two CLAD subtypes, BOS and RAS. The levels of KL-6 were significantly increased in RAS patients compared to those with BOS with mean values over two times higher. KL-6 levels also correlated with the decline in FEV-1in BOS subjects, thereby providing a clinical relevance to their study. Finally, Berastegui et al. asked the question if KL-6 levels in bronchoalveolar lavage fluid would be helpful in diagnosing CLAD, and in particular RAS [63]. Interestingly in this study there was no significant difference in serum KL-6 levels between post lung transplant patients with BOS and those that were stable after transplant and to healthy controls. However, similar to the results described above, serum KL-6 levels in RAS patients were significantly increased compared to lung transplant patients without CLAD and to healthy controls with mean serum KL-6 levels over three times higher than these other groups. In contrast, levels of KL-6 in the bronchoalveolar lavage fluid did not differ between the groups. Similar to the results by Ohshimo et al., serum KL-6 levels correlated with lung function testing, specifically the forced vital capacity (FVC), where higher serum KL-6 levels correlated with a decrease in FVC levels [63, 87].

Other biomarkers examined to diagnose CLAD and to differentiate BOS from RAS include cartilage oligomeric matric protein (COMP), Lipocalcin-2, and matrix metalloproteinase 9 (MMP9) [61, 64, 65, 66]. A retrospective study by Novo et al. examined if serial measurements of COMP could predict the later development of CLAD, including BOS and RAS [66]. Levels of COMP were higher at 3 months prior to the development of RAS, whereas COMP levels were not significantly different when comparing patients with CLAD as a whole to transplant patients that did not develop CLAD [66]. This suggests that COMP may be useful to differentiate RAS from BOS in CLAD patients. Patients with PPFE were not highlighted in this study. Similarly lipocalcin-2 levels were found to be elevated in patients with RAS in comparison to BOS [64]. Staining for lipocalcin 2 was present in distal airways and alveolar walls, where RAS is thought to originate [64]. Metalloproteinase 9 (MMP-9) has been linked to fibrotic diseases and has been studied as a biomarker for IPF. As fibrosis is an underlying cause for the symptoms and pathologic findings in the two types of CLAD, BOS (airway) and RAS (alveoli and interstitium), levels of MMP-9 have been examined to determine if they can diagnose CLAD as well as be a predictor for its occurrence [48, 86, 88, 89]. Beeh et al. examined MMP-9 levels in sputum from lung transplant patients and found that levels were higher in lung transplant patients compared to healthy controls with the highest levels seen in those diagnosed with BOS [88]. Similar findings were observed by Hubner et al. that found the ratio of MMP-9 to tissue inhibitor of MMPs (TIMP-1) was higher in lung transplant patients with BOS [89]. Importantly there was a correlation of sputum MMP-9: TIMP-1 ratios to spirometric values and in particular FEV-1. Finally, Pain et al. examined the use of serum MMP-9 levels as a predictor for the onset of BOS [61]. MMP-9 levels were serially measured for 18 months prior to the onset of CLAD. An increase in MMP-9 levels 12 months prior to the onset of CLAD predicted its occurrence with a stronger association with BOS than with RAS. Again, these studies assessed the levels of MMP-9 in patients with BOS, and not always specifically to RAS, without specific mention of PPFE.

Ongoing elastogenesis has been suggested to correlate with the levels of LTBP-4 and accordingly levels of LTB4 have also been explored as a biomarker for the diagnosis of PPFE [83]. Kinoshita et al. in their study measured LTBP-4 levels in both lung tissue and in serum from patients with either lung biopsy confirmed PPFE or those that were clinically diagnosed PPFE with these levels compared to those from healthy controls. Serum levels of LTBP-4 in PPFE patients were significantly higher than in controls and furthermore were 50% higher compared to patients with IPF [83]. Similar results were found for LTBP-4 levels in the lung, wherein PPFE patients had levels 2 fold higher than those with IPF [83]. Interestingly, levels of LTBP-4 in the serum of IPF patients who had co-existing upper lobe PPFE had higher levels compared to IPF patients without co-existing PPFE. Taken together this suggests that LTBP-4 may be a more specific biomarker for PPFE. This study also examined for a potential correlation between clinical parameters and survival with LTBP-4 levels with findings that a significant decline in FVC or the RV/TLC ratio trended with higher LTBP-4 levels in the serum. Higher LTBP-4 levels were associated with a 20% lower survival rate compared to those with normal LTBP-4 levels [83]. Upon comparison of the relative value between KL-6 and LTBP-4 in determining the diagnosis of PPFE versus IPF, levels of KL-6 are more useful for the diagnosis of IPF wherein LTBP-4 levels are useful in the diagnosis as well as management of PPFE [83]. Similar to the study discussed above of KL-6 levels, a limitation of the study is that only a one time measurement of serum LTBP-4 was assessed. Future studies examining serial LTBF-4 measurements would be helpful to determine if levels of LTB-4 can predict PPFE progression rate and prognosis. To date, LTBF-4 has not been examined as a biomarker in RAS/PPFE after lung transplantation.

Levels of surfactant protein D (SPD) have also been examined in PPFE patients to determine its use as a potential biomarker [40, 84, 85, 90, 91]. Several small studies found that SPD levels are increased in PPFE, however levels of SPD did not correlate to baseline physiologic values or patient demographics. Levels of SPD, in a cluster analysis study designed to identify PPFE prognostic factors, did trend with mortality rates that were highest [90]. Additional studies using larger patient cohorts are needed define a potential role for SPD levels in the diagnosis and management of PPFE and in particular if SPD levels over time identifies those patients with a higher risk for rapid progression of disease or a decrease in survival rate. Additional studies are also needed to examine its potential role in RAS/PPFE.

The above studies examining potential biomarkers in PPFE were mostly done in patients with the idiopathic form of the disease and not those specific to post-HSCT or lung transplantation. Pathologic findings from lung tissue obtained by biopsy and autopsy are similar for patients with an identified etiology for PPFE (post-HSCT patients or lung transplant patients with RAS/PPFE) and idiopathic PPFE suggesting that the above described biomarkers may also be helpful in the diagnosis and management PPFE that occurs after HSCT or in RAS/PPFE after lung transplantation.

3.8 Treatment, clinical course, and prognosis

To date no pharmacologic treatment for PPFE after HSCT or for RAS/PPFE in lung transplant patients has been identified to improve symptoms or to slow or reverse the progression of disease. In regards to post-HSCT patients, several case series describe patients that were treated with corticosteroids, however such treatment was not beneficial in slowing down the disease process or changing outcomes [39, 42, 45, 51, 70]. Similarly, as for HSCT patients the initiation, or increase in dose, of other immunosuppressants, such as cyclosporin, cyclophosphamide, mycophenolate, sirolimus, or tacrolimus has been tried in lung transplant patients but has not been shown to slow down or reverse the disease [38, 50, 54]. Other adjunctive treatments, including N-acetylcysteine and azithromycin, have been tried but there are no reports of a consistent response to these treatments [39, 42, 70]. That being said, for lung transplant patients azithromycin has been shown to be effective in the treatment of neutrophil reversible allograft dysfunction, a distinct entity apart from RAS with its own pathophysiology, but has not been effective in the treatment of RAS [38, 50]. Photophoresis therapy has been tried in lung transplant patients with RAS/PPFE but similar to other treatments described above has not been shown to be effective, although it may be helpful in those with strictly BOS [92, 93]. Recently, use of nintedanib and pirfenidone has been described in a small series of patients with either idiopathic or secondary PPFE [37, 91] and in case reports after lung transplant in those with RAS/PPFE [94, 95]. In the study by Nasser et al. the decline in FVC was less in those that received nintedanib compared to placebo, however the upper lobe volume loss determined by CT measurements did not differ over time between groups [38, 91]. In the case report by Pluchart et al. regarding the use of nintedanib in a post-lung transplant patient, there was no clinical change after 4 months of treatment with treatment eventually discontinued due to adverse effects [95]. For pirfenidone, in the case report by Sato et al., their patient had stable pulmonary function tests after initiation of pirfenidone, however this patient had mixed PPFE and UIP, as confirmed by open lung biopsy, so it is unclear if the PPFE responded to pirfenidone or just the UIP [94]. In regards to the pharmacologic treatment for PPFE, future studies are necessary to identify biomarkers and pathophysiologic mechanism in order to design novel pharmacologic approaches for the treatment of PPFE.

For those with progressive disease and respiratory failure, the only treatment option at present is lung transplantation [39, 47, 70, 96, 97, 98]. To date only 50 patients have been described that have undergone lung transplantation for PPFE, with most of these patients described in case reports or small case series [96, 97, 98]. Overall these studies describe the unique difficulties in performing lung transplants in PPFE and RAS/PPFE patients including pleural and chest wall adhesions, the flat thorax commonly seen, and the low body mass index (BMI). In the single center study by Faccioli et al. the median survival after lung transplant for PPFE patients was 24.5 months which is significantly decreased compared to the lung transplant patients as a whole who currently have a median survival of 62.5 months [98]. This is in contrast to the study by Shiya et al. that showed that mortality after lung transplant in patients with PPFE did not differ compared to those with ILD [99]. Larger studies are necessary to fully understand the role of lung transplant as a treatment for PPFE as a whole. However, what does seem clear from the literature is that patients with secondary PPFE to do clinically worse after lung transplant compared to those with the idiopathic form with survival times of around only 8 months [98].

The diagnosis of PPFE is associated with a poor prognosis after HSCT with 5 year survival rates estimated at 23.3–58.9%, although the median survival rate is 11.8 years [39, 40, 71, 83]. For patients with RAS/PPFE after lung transplant, the two studies initially describing PPFE in lung transplant patients gave no specific mention of life expectancy in their patients [49, 52]. Although no direct study on mortality after lung transplant in those with RAS with predominately PPFE has been performed, the overall survival after development of RAS is 6–18 months, suggesting a poorer prognosis for PPFE after lung transplant than after HSCT [54]. Interestingly, after either HSCT or lung transplantation two clinical scenarios have been described in the course of PPFE suggesting a common mechanism and pathophysiology. The two courses are either of a slow progression of disease over several months to years or a more precipitous decline, with the former being the clinical course that typically occurs [45, 74, 75, 97]. Currently there are no identified risk factors or clinical features to accurately predict the clinical course of an individual patient, nor are there known precipitants that causes transition from a slow progressive disease and induces a rapid progression or acute decompensation of the disease process. Progression of PPFE is typically manifested by progressive fibrosis and continued upper lobe volume loss, onset of radiographic changes in the mid and lower lung fields, continued weight loss and a decreasing BMI, further flattening of the thoracic cage with impairment of respiratory mechanics, an increase in shortness of breath and dyspnea on exertion, and acute respiratory failure [45, 53, 74, 79, 99]. Physiologically there is a continued decline in the FVC and radiographically pneumothoraces may be observed in addition to the signs of increased fibrosis [45, 52, 74, 79, 99]. This progression in volume loss was well outlined in the case series by Konen et al. where progressive lung changes were seen during serial yearly chest CT reviews over a greater than 4 year time frame [53]. Specifically in regards to radiographic changes, findings of lower lobe involvement and/or the co-existing interstitial lung disease, has been associated with a rapid progression in the disease course as well as shorter survival times [100]. It is unclear if this represents solely a progression of PPFE or the development of a secondary interstitial lung disease. Recently, studies in post HSCT patients have described scoring systems to assist in determining the prognosis of individuals with PPFE. Unfortunately no such scoring system is available for post lung transplant patients with RAS/PPFE after lung transplantation. The gender-age-physiology (GAP) model is a prognostic tool used for idiopathic pulmonary fibrosis patients [80]. This model was evaluated for patients with PPFE after HSCT by Shioya et al. and showed that those with a higher GAP score had a worse prognosis [99]. Kinoshita et al. recently published a PPFE prognostic model using variables shown to be significant in post-HSCT PPFE patients [41]. By examining data from two large patient cohorts with PPFE and using a multivariate Cox regression model they were able to identify 4 variables that predicted mortality in PPFE. Variables found to be important included forced vital capacity, history of pneumothorax, findings of lower lobe interstitial lung disease, and serum KL-6 levels [83]. A point system was derived using these four variables with three patient groups identified predicting survival based on the total score. The above type of studies are necessary but require further development, expansion, and refinement to allow scoring systems to prognosticate patients’ expected outcomes and clinical course and when attempt treatment or enrollment in clinical treatment trials.