Background
Despite recent progress in allogeneic hematopoietic stem cell transplantation (HSCT) for hematological diseases, pulmonary complications have been recognized as one of the major causes of morbidity and mortality after allogeneic HSCT, occurring in about 40–70% of patients, with a high mortality rate [
1‐
3]. Pulmonary complications after HSCT are caused by various noninfectious and infectious causes, and develop in both the early and late phases after transplantation, contingent on the day of development before or after 100 days following HSCT.
Opportunistic infection caused by bacteria, fungus, and viruses represents a major cause of infectious pulmonary complications in the early phase after HSCT [
2,
4]. In contrast, noninfectious pulmonary complications are major causes of morbidity and mortality later than 100 days after allogeneic HSCT, which have been labelled late-onset noninfectious pulmonary complications (LONIPCs) [
5]. LONIPCs classically include bronchiolitis obliterans (BO), cryptogenic organizing pneumonia (COP) (previously called bronchiolitis obliterans organizing pneumonia [BOOP]), diffuse alveolar hemorrhage (DAH), and idiopathic pneumonia syndrome (IPS) [
6,
7].
IPS is a severe pulmonary complication occurring after HSCT, which was originally characterized by symptoms and signs of pneumonia, restrictive pulmonary function test abnormality, and alveolar injury without documented lower respiratory tract infection [
8]. The American Thoracic Society (ATS) Research Statement on IPS provides a comprehensive review of IPS, the definition of which has been updated [
9]. IPS is clinically defined by three criteria: widespread alveolar injury, absence of documented infection, and absence of cardiac, renal, or iatrogenic etiology. The evidence of widespread alveolar injury is based on multilobar infiltrates on chest radiographs or computed tomography (CT), symptoms and signs of pneumonia, and evidence of abnormal pulmonary physiology, including restrictive pulmonary function test abnormality. The time of onset for IPS ranges from 4 to 106 days and diffuse alveolar hemorrhage (DAH) occurs in early post-HSCT, with an indicated median onset time of 12–15 days [
9]. Therefore, it seems that LONIPCs generally do not include IPS, according to the ATS statement.
Pleuroparenchymal fibroelastosis (PPFE), originally reported as an idiopathic disease [
10], has previously been reported as a novel radiological and pathological feature in patients with pulmonary disease in the late phase after HSCT [
11]. However, detailed clinical and radiological features of pulmonary complications after HSCT, especially in the late phase, are still largely unknown.
The primary aim of this retrospective study was to clarify the clinical and radiological features of late-onset severe restrictive lung defect after HSCT.
Discussion
In this retrospective observational study of allogeneic HSCT recipients, we investigated the clinical features of patients with late-onset severe restrictive lung defect (%VC less than 60%) and identified 12 cases with characteristic radiological findings and pulmonary function changes.
We have shown the incidence of cases of late-onset severe restrictive lung defect (%VC less than 60%) was 2.6% in patients who survived more than 100 days after allogeneic HSCT and underwent spirometry (12 of 453 cases). The incidence of cases of late-onset severe restrictive lung defect developing later than 100 days after HSCT, which has previously been reported as IPS, was 3.5% in adult patients [
6]. Another report showed the incidence of cases of late-onset severe restrictive lung defect, which has previously been reported as IPS, as 3.1% in pediatric patients with 1.6% severe restrictive ventilatory defect of %VC less than 60% [
16]. The incidence of cases of late-onset severe restrictive lung defect observed in the present study is comparable to those of previous studies. A distinctive characteristic of our retrospective cohort study is that we observed, not only more than 500 consecutive HSCT cases, but also pulmonary function test data on more than 95% of these cases were available. Although this was a retrospective study in a single center, we believe that the HSCT incidence rate and other data are highly reliable and representative.
Late-onset severe restrictive lung defect cases in the present study showed the following characteristic radiological findings: pleuroparenchymal thickening with volume loss, predominantly in the upper lobe (PPFE pattern) in 7 of 12 patients. A previous report demonstrated that PPFE, originally reported as idiopathic [
10], had features of late-onset lung involvement after allogeneic HSCT [
11]. Although we have not confirmed PPFE by pathological examination in the present study, CT findings in 7 of 12 patients were consistent with the findings seen in patients with PPFE. The other 5 non-PPFE cases were evaluated, and we found that 3 cases were airway-predominant pattern, and 2 cases were unclassifiable IP pattern. Therefore, late-onset severe restrictive lung defect cases in the present study could be categorized into 3 groups based on radiological features: PPFE pattern IP (
n = 7), airway-predominant pattern with BO (
n = 3), and unclassifiable IP pattern (
n = 2). In the present study, half of the patients had a history of pneumothorax, and 5 patients showed BO. Only one of the 5 patients with BO developed pneumothorax. Conversely, PPFE could have contributed to the onset of pneumothorax as subpleural fibrosis is attributed to recurrent rupturing of bullae [
11]. Indeed, among 6 cases with a history of pneumothorax, 4 patients showed typical PPFE pattern on HRCT. The 2 other cases were both unclassifiable IP, suggesting that pulmonary fibrosis may be a risk factor for pneumothorax in patients with late-onset severe restrictive lung defect.
With regard to clinical features, we believe there are mainly two clinical courses leading to late-onset severe restrictive lung defect: those associated with and those not associated with BO. Late-onset severe restrictive lung defect associated with BO would display mixed ventilatory impairment after exacerbation of obstructive ventilatory impairment due to BO. In contrast, late-onset severe restrictive lung defect not associated with BO directly displays severe restrictive ventilatory impairment induced by subpleural fibrosis and rupture involving pneumothorax.
The present study demonstrated that VC tended to decrease after pneumothorax and pulmonary infection. There were two exceptional cases (cases No. 8 and No. 10) of airway-predominant diseases, of which %VC was mildly increased after the diagnosis of late-onset severe restrictive lung defect. This may be associated with cases that are not progressive PPFE pattern IP but airway-predominant diseases with BO. TLC and DLco were decreased, while RV/TLC ratio was increased in the present study. A previous report demonstrated that idiopathic PPFE showed increased RV/TLC ratio [
17]. This might be due to compensated hyperinflation in the lower lobes, which is not observed in typical idiopathic pulmonary fibrosis. Therefore, increased RV/TLC ratio, in the present study, might also be due to compensated hyperinflation in the lower lobes as well as decreased TLC, especially in patients with PPFE pattern.
It has been reported that extensive chronic GVHD was a risk factor for LONIPCs, which possibly included cases of late-onset severe restrictive lung defect [
5]. Consistent with these previous studies, all patients of late-onset severe restrictive lung defect, in the present study, had developed chronic GVHD. These results suggest that chronic GVHD could be a risk factor for late-onset severe restrictive lung defect cases.
The outcomes of patients with late-onset severe restrictive lung defect in the present study were unfavorable [
9]; only 4 of 12 patients were alive at the time of the analysis with a median month from diagnosis to death of 33.5 months. Except for one case of death due to the relapse of multiple myeloma, all patients succumbed due to systemic or pulmonary infection, which is a novel finding in patients with late-onset severe restrictive lung defect.
There were limitations in the present study; one is that we were not able to obtain pathological specimens. Therefore, we could not pathologically confirm either PPFE or BO. However, the diagnosis of BO does not necessarily require biopsy per the 2014 NIH consensus development project [
18]. Consequently, we do not believe pathological confirmation is essential to diagnose patients with late-onset severe restrictive lung defect, considering the patients’ QOL and the difficulty in performing biopsy. The second is that the range of the timing of the pulmonary function tests may be too wide as pulmonary function tests presented as pre-diagnosis were those before HSCT or those performed after HSCT at a median of 12.6 months after HSCT (range 5.0–40.2 months). The third is that VC levels might not be accurate for identifying the development of late restrictive lung defect, because the patient could have a moderate restrictive lung defect due to early BO. A large, prospective study may contribute to clarifying the late-onset severe restrictive lung defect after HSCT.
We should be more aware of the unique entities after HSCT. Although there are still few reports on these entities, more unrecognized cases are expected to happen in the real clinical setting. Considering from the perspective of HSCT, only 2.4% of HSCT patients proceed to late-onset severe restrictive lung defect, while most of the cases do not proceed to this condition. We speculate that more unknown risk factors related to late-onset severe restrictive lung defect exist. In the future, we should seek more definite risk factors of severe restrictive lung defects and interventions to prevent late-onset severe restrictive lung defect in the early stage post-HSCT by accumulating more cases.