Respiratory Medicine (2014) 108, 1525e1533

CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector

Available online at www.sciencedirect.com

ScienceDirect

journal homepage: www.elsevier .com/locate/rmed

Late-onset noninfectious interstitial lungdisease after allogeneic hematopoietic stemcell transplantation

Frederic Schlemmer a, Sylvie Chevret b,c, Gwenael Lorillon a,Cedric De Bazelaire d, Regis Peffault de Latour e,Veronique Meignin f, Mauricette Michallet g, Eric Hermet h,Benjamin Wyplosz i, Veronique Houdouin j,Sylvain Marchand-Adam k, Gerard Socie e, Abdellatif Tazi a,b,Anne Bergeron a,b,*

a Univ Paris Diderot, Sorbonne Paris-Cite, AP-HP, Hopital Saint-Louis, Service de Pneumologie,F-75010 Paris, Franceb Biostatistics and Clinical Epidemiology Research Team (ECSTRA), UMR 1153 INSERM,Univ Paris Diderot, Sorbonne Paris Cite, Francec Univ Paris Diderot, Sorbonne Paris-Cite, Departement de Biostatistique et Informatique Medicale,AP-HP, Hopital Saint Louis, Paris, Franced Univ Paris Diderot, Sorbonne Paris-Cite, Service de Radiologie, AP-HP, Hopital Saint Louis, Paris,Francee Univ Paris Diderot, Sorbonne Paris-Cite, Hematologie-Greffe, APHP, Hopital Saint-Louis, Paris,Francef Univ Paris Diderot, Sorbonne Paris-Cite, Pathologie, APHP, Hopital Saint-Louis, Paris, Franceg Centre Hospitalier Lyon Sud, Hematologie, Hospices Civils de Lyon, Franceh Service de Therapie Cellulaire et d’hematologie Clinique Adulte, Universite d’Auvergne CREaT e EA7283, INSERM CIC-501, CHU Clermont-Ferrand Hopital Estaing, Francei Service des Maladies Infectieuses et Tropicales, Hopital de Bicetre, Le Kremlin-Bicetre, Francej Univ Paris Diderot, Sorbonne Paris-Cite, Service des Maladies Digestives et Respiratoires de l’Enfant,AP-HP, Hopital Robert Debre, Paris, Francek Univ Francois Rabelais, INSERM UMR 1100, Centre d’Etude des Pathologies Respiratoires,CHRU de Tours, Service de Pneumologie, F-37000 Tours, France

Received 18 July 2014; accepted 8 September 2014Available online 16 September 2014

* Corresponding author. Service de Pneumologie, Hopital Saint-Louis, 1, avenue Claude Vellefaux, 75475 Paris cedex 10, France. Tel.: þ331 42 49 41 66; fax: þ33 1 42 49 93 95.

E-mail address: anne.bergeron-lafaurie@sls.aphp.fr (A. Bergeron).

http://dx.doi.org/10.1016/j.rmed.2014.09.0060954-6111/ª 2014 Elsevier Ltd. All rights reserved.

1526 F. Schlemmer et al.

KEYWORDSInterstitial lungdisease;Bone marrowtransplantation;Organizingpneumonia;Nonspecificinterstitial lungpneumonia;Diffuse alveolardamage;Lymphoid interstitialpneumonia

Abbreviations

AML acute myeloid leukemALL acute lymphoid leukemBAL bronchoalveolar lavagCT computed tomographyDAD diffuse alveolar damagDLCO diffusion capacity forFEV1 forced expiratory voluFVC forced vital capacityGVHD graft-versus-host diseaHRCT high-resolution compuHSCT hematopoietic stem cILD interstitial lung diseasIPS idiopathic pneumoniaIQR inter-quartile rangeLIP lymphoid interstitial pLONIPC late-onset non-infectio

complicationPFT pulmonary function teNYHA New York Heart AssocNSIP non-specific interstitiaOLD obstructive lung diseaOP organizing pneumoniaPBSC peripheral blood stemRLD restrictive lung diseasSLB surgical lung biopsyTBI total body irradiationTLC total lung capacity

Summary

Background: Various late-onset noninfectious pulmonary complications may occur after alloge-neic hematopoietic stem cell transplantation (HSCT). Interstitial lung diseases (ILD) are oftenoverlooked, and few data are available.Methods: We retrospectively analyzed the clinical features, pulmonary function tests, radio-logical features and outcomes of allogeneic HSCT recipients who were diagnosed with a nonin-fectious ILD and were managed in our center between 2001 and 2010.Results: Forty patients were analyzed. The median time from transplant to ILD was 11.3months. The donor hematopoietic stem cell source was peripheral blood stem cells in 75% ofthe cases. Seventy percent of the patients had extra-thoracic chronic graft versus host diseaseat ILD diagnosis. We identified two lung computed tomography (CT) scan patterns according tothe predominance of ground glass opacities or alveolar consolidations. Restrictive ventilatorydefect was the main pulmonary function pattern. Lung histology was available for seven pa-tients and showed diffuse alveolar damage, non-specific interstitial pneumonia, organizingpneumonia or lymphoid interstitial pneumonia. Thirty-five patients (87.5%) were treated withsystemic steroids. Thirteen patients died (32.5%), 10 of respiratory failure. The median survivalrate at 24 months was 61%.Conclusion: This study highlights the existence of noninfectious post-allogeneic HSCT ILD andprovides new insights into the characteristics of these illnesses.ª 2014 Elsevier Ltd. All rights reserved.

iaia

e

ecarbon monoxideme in 1 s

seted tomographyell transplantationesyndrome

neumoniaus pulmonary

stiationl pneumoniase

cellse

Background

Late onset noninfectious pulmonary complications(LONIPCs) might follow allogeneic hematopoietic stem cell

transplantation (HSCT) and have a significant effect onpatient outcomes [1e3]. Previous publications havefocused on bronchiolitis obliterans (BO), which is the mostcommon LONIPC [4e6]. Some epidemiological retrospectivestudies [2,7e9] have reported interstitial lung diseases(ILD) among LONIPCs. Although individual cases of postallogeneic HSCT ILD have been published [10e13], no serieshave described the spectrum of these ILD using an overallpulmonologic approach. Post HSCT noninfectious ILD arelikely misdiagnosed and overlooked. Conversely to BO [6],ILD is characterized by the presence of diffuse lungparenchymal opacities on a CT scan. In the current prac-tice, an infectious cause is first considered. Concomitantly,the patient might be diagnosed with an extra thoracic graftversus host disease (GVHD) and treated with corticoste-roids. In this context, these patients are usually treatedwith a combined treatment associating empirical anti-microbiological drugs and steroids that may explain whythese noninfectious ILD are misdiagnosed.

The process of achieving a multidisciplinary diagnosis ina patient with ILD requires close communication betweenclinician, radiologist, and when appropriate, pathologist[14]. The multidisciplinary approach usually allows todetermine whether the noninvasive approach based onclinical, radiological, lung function, bronchoalveolar lavage(BAL) and laboratory findings is informative enough or if alung biopsy is needed [14]. Besides idiopathic ILD, severalmedical settings such as exposure or collagen vasculardiseases have been shown to cause ILD. In these contexts,knowledge of the clinical characteristics of these ILD hasled to an improvement in their management with a dra-matic decrease in the indications for lung biopsies [14e19].In the setting of allogeneic HSCT, two pathological studiesidentified 6 cases of diffuse alveolar damage (DAD) 8 casesof organizing pneumonia (OP), one case of lymphoidinterstitial pneumonia (LIP) and 13 cases of non-classified

Interstitial lung disease after stem cell transplant 1527

ILD [20,21]. In HSCT recipients, ILD management based onsurgical lung biopsies is, however, questionable: first, alung biopsy is an invasive and high-risk procedure in thissetting [22,23]; second, no correlation between patholog-ical findings and prognosis exists for post allogeneic HSCTILD that would allow to guide the treatment; and third, theefficient microbiological tools applied to samples obtainedfrom less invasive techniques such as BAL or nasal aspirateoften help differentiate ILD from infectious pneumonia[24e26].

Because post HSCT ILD are infrequent, it was importantto retrospectively analyze the patients diagnosed with anILD to describe their clinical presentations based on thepulmonary investigation techniques used in clinical practiceand to analyze the outcomes of these patients.

Patients and methods

This retrospective study was approved by the institutionalreview board of the French learned society for respiratorymedicine CEPRO 2011-055. All of the allogeneic HSCT re-cipients diagnosed with a noninfectious ILD in our centerbetween 2001 and 2010 were eligible for this study. Inaddition, patients who underwent an allogeneic HSCT inanother French bone marrow transplantation departmentand who were referred to our respiratory diseases depart-ment for the management of an ILD were included. Wecollected data from individual medical records.

Diagnosis of noninfectious ILD

A noninfectious ILD was diagnosed when infiltrative opaci-ties were present on HRCT and 1) no pathogens wereidentified in the respiratory samples (BAL and/or nasalaspirate and/or sputum); and/or 2) no clinical or radiolog-ical improvements were observed despite broad antimi-crobial treatment; and/or 3) clinical and radiologicalimprovements occurred after initiating or increasing theimmunosuppressive treatment; and/or 4) no pathogenswere found on the lung biopsy (if it was performed).

Clinical characterization of ILD

As is conventional in pulmonology practice, ILD werecharacterized by clinical symptoms, high-resolutioncomputed tomography (HRCT), pulmonary function testing(PFT), BAL analysis, and lung histology. HRCT scans per-formed at the time of ILD diagnosis and during follow-upwere reviewed by an experienced radiologist (CDB) and twopulmonologists (FS and AB). The following features werenoted: alveolar consolidations, ground glass opacities,septal thickening, subpleural reticulations, honeycombing,bronchiectasis, bronchial thickening and pleural abnor-malities. To assess lesion distribution, each lung wasdivided into three areas (upper/medial/lower) from thelung apices to the domes of the diaphragm. Conclusionswere reached by consensus. The clinical data, PFT and BALfindings were collected. An extensive search for viruses (asdescribed in the e-appendix), bacteria (direct examinationand culture) and fungi (immunofluorescence and molecularbiology for Pneumocystis jiroveci; direct examination and

culture for other species) was performed on the BAL. Whenbronchoscopy could not be performed, bacteria and fungiwere searched for in the sputum, P. jiroveci was searchedfor in induced sputum samples and respiratory viruses weresearched for in nasal aspirates (e-appendix). The total anddifferential cell counts in the BAL were analyzed. Pulmo-nary function tests were performed using a body plethys-mograph (Jaeger Masterscreen Body; Jaeger, GmBH;Wurzburg, Germany). The diffusing capacity of carbonmonoxide (DLCO) was measured using the single breathmethod, and the results were adjusted to the last availablehemoglobin level. Predictive values were determined aspreviously described [27]. Obstruction was defined as anFEV1/FVC ratio <0.7 [28], and restriction was defined as aTLC < 80% of the predicted value. A DLCO <80% of thepredicted value was considered abnormal. An experiencedpathologist (VM) analyzed the lung tissue specimen frompatients who underwent a surgical lung biopsy. The diag-nosis and severity of GVHD were reported based on theclinical grading scores [29,30].

Statistical analyses

Descriptive statistics are presented as the median andinter-quartile range (IQR) [IQR: 25 to 75th percentiles] forcontinuous parameters and as the number and percentagefor non-continuous and qualitative parameters. Survivalafter ILD diagnosis and ILD incidence were estimated usingthe KaplaneMeier method. The prognostic analysis of post-ILD survival was univariate, based on a log rank test fornon-continuous parameters or logistic regression with a Coxmodel for continuous parameters. Given the sample sizeand number of observed events, no multivariate analyseswere performed. Every statistical analysis had a bilateralformulation, and P < 0.05 was considered significant.

Results

Patient characteristics

Between 2001 and 2010, 1277 allogeneic HSCTs were per-formed in our center, including 35% using a myeloablativeconditioning regimen, and 65% using a nonmyeloablativeconditioning regimen with stem cell sources as followed:60% peripheral blood stem cells, 25% bone marrow and 15%cord blood.

Thirty-one allogeneic HSCT recipients were identified ashaving a noninfectious ILD (2.4%). Three patients from aprevious study were included [11]. Nine additional patients,referred to our department because of an ILD during thestudy period, were included. Sixteen patients were diag-nosed between 2001 and 2007 and 24 patients between2008. The clinical characteristics of the patients are sum-marized in Table 1. All ILD occurred within the first threeyears after allogeneic HSCT. The median time from trans-plantation to ILD was 11.3 months [IQR: 5.9e19.2]. Most ofthe patients had received prior chemotherapy, and 30% hadexperienced a prior autologous transplantation. The donorhematopoietic stem cell sources were peripheral bloodstem cells in 75% of the cases, and the conditioning regimenwas non-myeloablative in 57.5% of the cases.

Table 1 Clinical characteristics of the 40 patients at ILDdiagnosis.

Clinical variables Median [IQR];N (%)

Age at transplant, years 40.5 [27.5e51.5]Male 28 (70)Smoking history 17 (43)Age at ILD, years 40.6 [28.5; 51.8]Time from transplantation

to ILD, months11.3 [5.9e19.2]

History of an infectiouspneumonia in the timebetween the transplant and ILD

11 (27.5)

Hematological diseasesAL 15 (37.5)CML 2 (5)Lymphoma 13 (32.5)Myeloma 2 (5)Othersa 8 (20)Status of disease at time of transplantComplete response 28 (70)Relapse 2 (5)Chronic phase/partial response 2 (5)NA 8 (20)Stem cell sourcePBSCs 30 (75)Bone marrow 9 (22.5)Cord blood 1 (2.5)Donor HLA statusRelated 23 (57.5)Unrelatedb 17 (42.5)Conditioning regimenNon-myeloablative 23 (57.5)Busulfan-based 10 (25)TBI 25 (62.5)GVHD prophylaxisCyclosporine alone 6 (15)Cyclosporine-mycophenolate mofetil 24 (60)Cyclosporine-methotrexate 10 (25)Acute GVHD 22 (55)Grade �2 14 (35)Chronic GVHD 28 (70)Limited 21 (52.5)Extensive 7 (17.5)

ILD, interstitial lung disease; AL, acute leukemia; CML, chronicmyeloid leukemia; NA, not available; HLA, human leukocyteantigen; PBSCs, peripheral blood stem cells; TBI, total bodyirradiation; GVHD, graft-versus-host disease.a Myelodysplasic syndrome, myeloproliferative disease other

than CML, aplastic anemia.b 10/10 and 9/10 allelic unrelated donors and cord blood

transplant.

Table 2 Lung HRCT findings in patients at diagnosis of ILDafter allogeneic hematopoietic stem cell transplantation.

Imaging finding N (%) of patients(N Z 40)

Alveolar consolidations 30 (75)Bilateral 19 (63)

Ground glass opacities 29 (72)Bilateral 24 (83)

Involvement of >3 lung areasa,b 15 (37)Exclusive upper lung area involvementb 5 (12)Exclusive lower lung area involvementb 9 (22)Reticulation 3 (7)Septal thickening 1 (2)Honeycombing 1 (2)Bronchial thickening 4 (10)Pneumomediastinum 3 (7)a Each lung was divided into three areas: upper, medial and

lower.b Concerns alveolar consolidation and ground glass opacity.

1528 F. Schlemmer et al.

ILD characteristics

Clinical characteristicsThe most frequent clinical symptoms were dyspnea(n Z 35, 87.5%; New York Heart Association (NYHA) � 3:nZ 12, 40%), fever (nZ 20, 50%), and cough (nZ 24, 60%).

The median time from the onset of symptoms to ILD diag-nosis was 15 days [IQR: 2e71].

At the time of ILD diagnosis, 27 patients (67.5%) werereceiving systemic immunosuppressive treatment calci-neurin inhibitors (n Z 12), mycophenolate mofetil(n Z 9), everolimus (n Z 1), etanercept (n Z 1) orimatinib (n Z 1); 16 were receiving a single drug, and 10were receiving two or more drugs. Forty percent of thepatients were receiving systemic steroids (median dose ofprednisone; 25 mg per day, [IQR: 20e40]) for coexistingGVHD.

Overall, 34 patients (85%) had a history of GVHD, eitheracute and/or chronic (Table 1). Chronic GVHD was presentin 28 patients (70%) at the time of ILD diagnosis andinvolved the skin and/or mucosa (42.5%), the gut (20%) orthe liver (17.5%; Table 1). Four more patients developedGVHD after the ILD diagnosis.

No patients had a history suggestive of an ILD associatedwith an inhalational exposure.

Thoracic imagingThe HRCT characteristics of ILD are summarized in Table 2.Two different HRCT patterns were identified according tothe predominance of alveolar consolidations (n Z 22, 55%,Fig. 1A) or ground glass opacities (n Z 18, 45%, Fig. 1B).Other HRCT abnormalities were infrequent. Patients with aground glass opacity pattern or alveolar consolidationpattern did not differ in their clinical characteristics ortime from allogeneic HSCT to ILD diagnosis.

Lung functionThe results of the pulmonary function tests (PFTs) arepresented in Table 3. Lung volumes and DLCO were availablefor 29 and 18 patients, respectively. At ILD diagnosis, 24patients (83%) had a restrictive lung disease (RLD), eightpatients (28%) had an obstructive lung disease (OLD) andfour patients (14%) had a mixed pattern. Only two patients(7%) had normal lung volumes. The DLCO was decreased inall of the patients tested.

Figure 1 Lung HRCT patterns of ILD after allogeneic hematopoietic stem cell transplantation. Two patterns were identified: onewith predominant alveolar opacities (A), and one with predominant ground glass opacities (B).

Interstitial lung disease after stem cell transplant 1529

Bronchoalveolar lavage cell countBAL was performed in 30 patients. The total and differen-tial cell counts are shown in Table 4. Except for three pa-tients (10%) who had normal BAL cell counts, all patientshad an alveolitis. Lymphocytic alveolitis (defined by alymphocyte count �15% of total cells) was the mostfrequent and was found in 20 patients (67%). Among thesepatients, the median percentage of alveolar lymphocyteswas 40% [IQR: 21e56%] and seven patients had a concomi-tant neutrophil count �10% of total cells. Three other pa-tients had a neutrophil count �10% of total cells (median20%, [IQR: 15e33%]). Four patients had an eosinophil count�10% of total cells, including two with an eosinophil count�20% of total cells, suggesting an eosinophilic pneumonia.

Both PFT and BAL results were similar whether patientshad a ground glass opacities HRCT pattern or an alveolarconsolidation pattern.

Lung histologyEleven lung biopsies were available for 10 patients. Fourpatients had CT scan-guided percutaneous lung biopsies,and seven patients had surgical lung biopsies (six withvideo-assisted thoracoscopy). Four of these 10 patients(40%) experienced postoperative morbidity or mortality;two patients required prolonged chest tube aspiration (onerequired another surgery), and two patients died within 15days following surgery (one from respiratory failure and theother from septic shock with multisystem organ failure). A

Table 3 Pulmonary function testing in overall patients and according to the lung HRCT pattern at the time of ILD diagnosis.

Median [IQR] FEV1 FVC FEV1/FVC%

TLC DLCO

(L) % pred (L) % pred (L) % pred % pred

All patients(n Z 29)

1.805[1.54e2.39]

53[42e72]

2.38[1.82e2.85]

52[42.5e72]

81.9[71.5e87]

4.05[3.53e5.17]

63[58e77]

41[36e47]

Ground glass patterna

(n Z 13)1.675[1.54e2.25]

52.5[42.6e61.2]

2.18[1.82e2.55]

52[43.9e60.5]

81.9[70.6e88.9]

4.04[3.65e5.18]

63.5[60.1e74.2]

43[36e47]

Alveolar consolidationspatterna

(n Z 16)

2.065[1.54e2.51]

53[42e77.5]

2.485[2.03e3.03]

56[42.7e80.2]

81.65[72.5e85.5]

4.09[3.53e4.33]

63[57e79]

39[37e44]

a Lung CT scan pattern; FEV1, forced expiratory volume in 1 s; FVC, forced vital capacity; TLC, total lung capacity; DLCO, diffusingcapacity of carbon oxide; pred, predicted.

1530 F. Schlemmer et al.

definite histological diagnosis was obtained in 7/10 (70%)patients and consisted of DAD (n Z 2), nonspecific inter-stitial pneumonia (NSIP) (n Z 2), OP (n Z 2) and LIP(n Z 1). Patients diagnosed with DAD had either an alveolarconsolidation or a ground glass opacities HRCT pattern. Thetwo patients diagnosed with OP and the patient diagnosedwith LIP had an alveolar consolidations pattern; both of thepatients diagnosed with NSIP had a ground glass opacitiesHRCT pattern.

Treatment and outcome

The median follow-up after diagnosis of ILD was 9.3 months[IQR: 2.1e32]. The patients of both HRCT phenotypes weretreated similarly. Thirty-five patients (87.5%) received sys-temic steroids for the treatment of ILD (introduction orincrease), and 13 of these patients (32.5%) received high-dose intravenous steroids (pulse methylprednisolone5e10 mg/kg for 3 days). One patient received cyclophos-phamide together with pulse methylprednisolone. Fivepatients did not receive steroids at the time ILD was diag-nosed. Among these five patients, one patient finallyreceived high-dose pulsed methylprednisolone one yearafter ILD diagnosis because of progressive lung disease;three patients required steroids 2e6 months after the ILDdiagnosis for extra thoracic GVHD. One patient presentedspontaneous resolution of ILD.

Table 4 Bronchoalveolar lavage cell counts in overallpatients and according to the lung HRCT pattern.

Median [IQR] Total cellcount/mm3

Ma%

Ly%

Neu%

Eo%

All patients(n Z 30)

280[180e460]

54[35e82]

20.5[8e48]

6[2e15]

0[0e3]

Ground glasspatterna

(n Z 15)

250[150e280]

67[25e88]

19[5e48]

5[1e11]

0[0e5]

Alveolarconsolidationspatterna

(n Z 15)

315[225e465]

50[35e77]

23[18e52]

7[2e18]

0[0e2]

a Lung CT scan pattern; Ma, macrophages; Ly, lymphocytes;Neu, neutrophils; Eo, eosinophils.

Thirteen patients died (33%). Ten patients died of res-piratory failure; these patients were receiving systemicsteroids for ILD treatment. The lung histology available fortwo of these patients showed DAD for one and NSIP for theother. The three other patients died of refractory gutcGVHD associated with severe sepsis, septic shock and en-cephalopathy syndrome. The survival rate was estimated at77% [IQR: 64e92%], 70% [IQR: 56e88%] and 61% [IQR:45e82%] at 6, 12 and 24 months, respectively, after ILDdiagnosis (Fig. 2).

Evolution of HRCT and PFT

Although the overall prognosis of ILD was poor, a significantnumber of patients had a favorable outcome regardingHRCT and PFT. We did not find any association between theevolution of both HRCT and PFT and the HRCT scan patternat ILD diagnosis.

Among the survivors, at least two serial HRCT scans wereavailable for 18 patients (alveolar consolidations HRCT

Figure 2 The estimated overall survival distribution after ILDdiagnosis in the study population. The dashed lines representthe 95% CI of survival estimates over time.

Interstitial lung disease after stem cell transplant 1531

pattern (n Z 11), ground glass opacities HRCT pattern(n Z 7)). For these patients, the last available HRCT scanwas performed at a median time of 10.3 months after ILDdiagnosis [IQR: 2e23]. At last follow-up, complete resolu-tion of radiological abnormalities was observed in five pa-tients (alveolar consolidations pattern (n Z 4)) andpersisted in 13 patients (ground glass opacities pattern(n Z 6), alveolar consolidations (n Z 7)). After beingcharacterized as having a particular phenotype by a CT scanat the time of diagnosis, none of the patients evolved to theother phenotype.

All the patients who had PFT at ILD diagnosis had two ormore PFTs during the follow-up with a similar distributionbetween both HRCT pattern; for these patients, the lastavailable PFT was performed at a median time of 7 months[IQR: 3e18] after the first PFT. Fourteen patients (45%)showed greater than 10% improvement of FVC or FEV1

relative to the baseline.

Prognostic analysis

In the univariate analysis, no prognostic factor was identi-fied to be associated with survival (e-Table 1). The patientswith an alveolar consolidation pattern tended to havebetter survival than those with a ground glass opacitypattern (78% vs. 60% at 12 months, 71% vs. 48% at 24months), although the difference did not reach statisticalsignificance (p Z 0.19; e-Fig. 1).

Discussion

This retrospective study is the largest series on allogeneicpost-HSCT non-infectious ILD. Our study highlights the ex-istence of post-allogeneic HSCT non-infectious ILD andprovides new insights into the clinical characteristics of thiscomplication. Previous studies have focused on epidemio-logical data and anecdotal reports. We observed thefollowing: 1) post-allogeneic HSCT ILD typically occurwithin the first 3 years after transplantation and arefrequently associated with extra pulmonary GVHD mani-festations; 2) ILD include different histological entities; 3)two distinct lung HRCT patterns were observed according tothe predominance of ground glass opacities or alveolarconsolidations; 4) lung function restriction and BAL lym-phocytic alveolitis were frequent; and 5) the prognosis wasglobally poor, although in some cases, ILD might improve orresolve with corticosteroids.

In the context of allogeneic HSCT, the occurrence of ILDis often attributed to idiopathic pneumonia syndrome (IPS)or to OP [4,31]. IPS is not a well-defined pulmonary diseaseand includes various conditions [31,32]. The diagnosis ofIPS, which has recently been reviewed, relies on diffuselung opacities on the chest X-ray and the exclusion ofseveral diagnoses, such as lung infection or pulmonaryedema [31]. The median time of onset for IPS is reported tobe 19 days (range, 4e106 days) after HSCT. The intensityand type of HSCT conditioning have been associated withthe development of IPS, which is consistent with datagenerated from mouse models [31]. The ILD described inour study differ in various respects from IPS. ILD occurredlater after transplantation, predominantly in patients who

underwent non-myeloablative conditioning. Only 60% of ourpatients underwent pre-transplant total body irradiation,and none of our patients received carmustine; both ofthese factors are associated with IPS. Thus, the ILD wedescribed do not match the current IPS criteria. Conversely,the time of onset after the HSCT and the clinical, pulmo-nary and radiological characteristics of some of our patientswere similar to the characteristics described in cases ofpost-allogeneic HSCT OP [13,33]. However, based on thelung biopsy results, other ILD entities (NSIP, DAD and LIP)were observed in our patients.

At present, only biopsy-proven BO has been firmlyrecognized as a manifestation of lung GVHD based onepidemiological studies that show a strong association be-tween BO and extra-thoracic chronic GVHD. In severalstudies, OP has also been associated with GVHD [4,13,15]. Afeature of our findings is that almost all patients whodeveloped late onset ILD had a history of acute and/orchronic GVHD. Unfortunately, due to the retrospectivedesign of our study, we could not evaluate the incidence ofGVHD in our population of HSCT recipients who did notdevelop ILD during the study period. Thus, although ourfindings suggest that the ILD spectrum may reflect lungGVHD, we could not formally conclude. However, thepresence of an OLD in some of our patients, suggesting BOassociated with ILD, further strengthens this hypothesis.Although the history of the patients was not suggestive, therole of drug pulmonary toxicity in patients receiving mul-tiple treatments cannot be formally ruled out.

One finding of our study is that, conversely to what isknown for the treatment of BO [6,34], a significant pro-portion of our patients who were treated with steroids hadeither complete resolution of their ILD or a significantimprovement in their PFT. These data emphasize theimportance of identifying these ILD to promptly administersteroids. However, only a randomized study could demon-strate the efficacy of steroids in post allogeneic non-infectious ILD.

The retrospective design and the relatively small num-ber of patients did not allow us to reach firm conclusions onseveral points. First, we could not correlate the radiologicaland clinical diagnosis with the underlying histological dis-ease because of the small number of patients who under-went a surgical lung biopsy. Of note, in our few patientswho underwent a lung biopsy, alveolar consolidations HRCTpattern could be associated with either DAD, OP or LIPwhereas ground glass opacities pattern was associated withNSIP or DAD. Second, during the follow-up period, we didnot report any changes in the HRCT scan patterns. However,follow up HRCT scans were not available for all of the pa-tients. Thus, we could not firmly exclude that both HRCTscan patterns are a spectrum of the same lung diseaseentity. Third, we could not identify any prognostic factors,although the patients with an alveolar consolidationpattern tended to have better survival than the patientswith a ground glass opacities pattern.

In summary, because of the increasing number of long-term HSCT survivors, an understanding of new late onsetnoninfectious complications is important. Herein, wedescribed the characteristics of ILD occurring after alloge-neic HSCT according to the usual approach of ILD. As typi-cally occurs in clinical practices, the majority of these

1532 F. Schlemmer et al.

patients did not undergo lung biopsies. Our data provideinformation regarding the management of these patients.Further studies are needed to clarify the physiopathology,identify the risk factors and optimize the treatment ofthese lung disorders to improve the prognosis.

Authors’ contribution

FS: Data acquisition, analysis and interpretation, draftingthe article, final approval of the version to be published.SC: Statistical analysis of the data, drafting the article,final approval of the version to be published. GL: Dataanalysis and interpretation, final approval of the version tobe published. CDB: Data acquisition, final approval of theversion to be published.

RPDL: Data analysis and interpretation, final approval ofthe version to be published. VM: Data acquisition, finalapproval of the version to be published. MM: Data acquisi-tion, final approval of the version to be published. EH: Dataacquisition, final approval of the version to be published.BW: Data acquisition, final approval of the version to bepublished.

VH:Data acquisition, final approval of the version to bepublished. SMA: Data acquisition, final approval of theversion to be published. GS: Data analysis and interpreta-tion, final approval of the version to be published. AT: Studyconception and design, data analysis and interpretation,final approval of the version to be published. AB: Studyconception and design, data acquisition, analysis andinterpretation, drafting the article, final approval of theversion to be published.

Funding

None.

Conflict of interest

None declared.

Acknowledgments

We thank Elisabeth Savariau and Robin Nancel for theirexcellent technical assistance. We thank EGMOS, which isan association of bone marrow transplant recipients.

Appendix A. Supplementary data

Supplementary data related to this article can be found athttp://dx.doi.org/10.1016/j.rmed.2014.09.006.

References

[1] Patriarca F, Skert C, Bonifazi F, Sperotto A, Fili C, Stanzani M,et al. Effect on survival of the development of late-onset non-infectious pulmonary complications after stem cell trans-plantation. Haematologica 2006;91:1268e72.

[2] Palmas A, Tefferi A, Myers JL, Scott JP, Swensen SJ, Chen MG,et al. Late-onset noninfectious pulmonary complications afterallogeneic bone marrow transplantation. Br J Haematol 1998;100:680e7.

[3] Dudek AZ, Mahaseth H, DeFor TE, Weisdorf DJ. Bronchiolitisobliterans in chronic graft-versus-host disease: analysis of riskfactors and treatment outcomes. Biol Blood Marrow Trans-plant 2003;9:657e66.

[4] Filipovich AH, Weisdorf D, Pavletic S, Socie G, Wingard JR,Lee SJ, et al. National Institutes of Health consensus devel-opment project on criteria for clinical trials in chronic graft-versus-host disease: I. Diagnosis and staging working groupreport. Biol Blood Marrow Transplant 2005;11:945e56.

[5] Chien JW, Martin PJ, Gooley TA, Flowers ME, Heckbert SR,Nichols WG, et al. Airflow obstruction after myeloablativeallogeneic hematopoietic stem cell transplantation. Am JRespir Crit Care Med 2003;168:208e14.

[6] Barker AF, Bergeron A, Rom WN, Hertz MI. Obliterative bron-chiolitis. N Engl J Med 2014;370:1820e8.

[7] Patriarca F, Skert C, Sperotto A, Damiani D, Cerno M,Geromin A, et al. Incidence, outcome, and risk factors of late-onset noninfectious pulmonary complications after unrelateddonor stem cell transplantation. Bone Marrow Transplant2004;33:751e8.

[8] Huisman C, van der Straaten HM, Canninga-van Dijk MR,Fijnheer R, Verdonck LF. Pulmonary complications after T-cell-depleted allogeneic stem cell transplantation: low inci-dence and strong association with acute graft-versus-hostdisease. Bone Marrow Transplant 2006;38:561e6.

[9] Sakaida E, Nakaseko C, Harima A, Yokota A, Cho R, Saito Y,et al. Late-onset noninfectious pulmonary complications afterallogeneic stem cell transplantation are significantly associ-ated with chronic graft-versus-host disease and with the graft-versus-leukemia effect. Blood 2003;102:4236e42.

[10] Perreault C, Cousineau S, D’Angelo G, Gyger M, Nepveu F,Boileau J, et al. Lymphoid interstitial pneumonia after allo-geneic bone marrow transplantation. A possible manifestationof chronic graft-versus-host disease. Cancer 1985;55:1e9.

[11] Bergeron A, Bengoufa D, Feuillet S, Meignin V, Peffault deLatour R, Rybojad M, et al. The spectrum of lung involvementin collagen vascular-like diseases following allogeneic he-matopoietic stem cell transplantation: report of 6 cases andreview of the literature. Med Baltim 2011;90:146e57.

[12] Wolff D, Reichenberger F, Steiner B, Kahl C, Leithauser M,Skibbe T, et al. Progressive interstitial fibrosis of the lung insclerodermoid chronic graft-versus-host disease. Bone MarrowTransplant 2002;29:357e60.

[13] Freudenberger TD, Madtes DK, Curtis JR, Cummings P,Storer BE, Hackman RC. Association between acute andchronic graft-versus-host disease and bronchiolitis obliteransorganizing pneumonia in recipients of hematopoietic stem celltransplants. Blood 2003;102:3822e8.

[14] Travis WD, Costabel U, Hansell DM, King Jr TE, Lynch DA,Nicholson AG, et al. An Official American Thoracic Soci-ety/European Respiratory Society Statement: update of theInternational Multidisciplinary classification of the idiopathicinterstitial pneumonias. Am J Respir Crit Care Med 2013;188:733e48.

[15] Daniele RP, Elias JA, Epstein PE, Rossman MD. Bronchoalveolarlavage: role in the pathogenesis, diagnosis, and managementof interstitial lung disease. Ann Intern Med 1985;102:93e108.

[16] Ryu JH, Olson EJ, Midthun DE, Swensen SJ. Diagnosticapproach to the patient with diffuse lung disease. Mayo ClinProc 2002;77:1221e7.

[17] Meyer KC. The role of bronchoalveolar lavage in interstitiallung disease. Clin Chest Med 2004;25:637e49.

[18] Johkoh T. Imaging of idiopathic interstitial pneumonias. ClinChest Med 2008;29:133e47.

Interstitial lung disease after stem cell transplant 1533

[19] Kowal-Bielecka O, Kowal K, Chyczewska E. Utility of bron-choalveolar lavage in evaluation of patients with connectivetissue diseases. Clin Chest Med 2010;31:423e31.

[20] Yousem SA. The histological spectrum of pulmonary graft-versus-host disease in bone marrow transplant recipients.Hum Pathol 1995;26:668e75.

[21] Xu L, Drachenberg C, Tavora F, Burke A. Histologic findings inlung biopsies in patients with suspected graft-versus-hostdisease. Hum Pathol 2013;44:1233e40.

[22] Zihlif M, Khanchandani G, Ahmed HP, Soubani AO. Surgicallung biopsy in patients with hematological malignancy or he-matopoietic stem cell transplantation and unexplained pul-monary infiltrates: improved outcome with specific diagnosis.Am J Hematol 2005;78:94e9.

[23] Wingard JR, Hiemenz JW, Jantz MA. How I manage pulmonarynodular lesions and nodular infiltrates in patients with he-matologic malignancies or undergoing hematopoietic celltransplantation. Blood 2012;120:1791e800.

[24] Azoulay E, Mokart D, Rabbat A, Pene F, Kouatchet A,Bruneel F, et al. Diagnostic bronchoscopy in hematology andoncology patients with acute respiratory failure: prospectivemulticenter data. Crit Care Med 2008;36:100e7.

[25] Azoulay E, Mokart D, Lambert J, Lemiale V, Rabbat A,Kouatchet A, et al. Diagnostic strategy for hematology andoncology patients with acute respiratory failure: randomizedcontrolled trial. Am J Respir Crit Care Med 2010;182:1038e46.

[26] Forslow U, Remberger M, Nordlander A, Mattsson J. Theclinical importance of bronchoalveolar lavage in allogeneicSCT patients with pneumonia. Bone Marrow Transplant 2010;45:945e50.

[27] Standardized lung function testing. Official statement of theEuropeanRespiratory Society. Eur Respir J Suppl 1993;16:1e100.

[28] Pauwels RA, Buist AS, Calverley PM, Jenkins CR, Hurd SS. Globalstrategy for the diagnosis, management, and prevention ofchronic obstructive pulmonary disease. NHLBI/WHO globalinitiative for chronic obstructive lung disease (GOLD) workshopsummary. Am J Respir Crit Care Med 2001;163:1256e76.

[29] Przepiorka D, Weisdorf D, Martin P, et al. 1994 consensusconference on acute GVHD grading. Bone Marrow Transplant1995;15:825e8.

[30] Shulman HM, Sullivan KM, Weiden PL, et al. Chronic graft-versus-host syndrome in man. A long-term clinicopathologicstudy of 20 Seattle patients. Am J Med 1980;69:204e17.

[31] Panoskaltsis-Mortari A, Griese M, Madtes DK, Belperio JA,Haddad IY, Folz RJ, et al. An official American Thoracic So-ciety research statement: noninfectious lung injury after he-matopoietic stem cell transplantation: idiopathic pneumoniasyndrome. Am J Respir Crit Care Med 2011;183:1262e79.

[32] Clark JG, Hansen JA, Hertz MI, Parkman R, Jensen L,Peavy HH. NHLBI workshop summary. Idiopathic pneumoniasyndrome after bone marrow transplantation. Am Rev RespirDis 1993;147:1601e6.

[33] Pipavath SN, Chung JH, Chien JW, Godwin JD. Organizing pneu-monia in recipients of hematopoietic stem cell transplantation:CT features in 16 patients. J Comput Assist Tomogr 2012;36:431e6.

[34] Bergeron A, Godet C, Chevret S, Lorillon G, Peffault deLatour R, de Revel T, et al. Bronchiolitis obliterans syndromeafter allogeneic hematopoietic SCT: phenotypes and prog-nosis. Bone Marrow Transplant 2013;48:819e24.