histopathology ofthelung after bone marrow transplantation

J Clin Pathol 1983;36:546-554

Histopathology of the lung after bone marrow

transplantationJP SLOANE,* MH DEPLEDGE,t RL POWLES4 GR MORGENSTERN,t BS TRICKEY,*PJ DADYt

From the *Department ofHistopathology, tLung Function Unit, and tLeukaemia Unit, Royal MarsdenHospital, Downs Road, Sutton, Surrey

SUMMARY The histopathological changes in the lungs of 32 patients who died after bone marrowtransplantation for leukaemia have been studied and compared with those found in 21 patientstreated by conventional chemotherapy. The transplanted patients exhibited a higher incidence ofinterstitial pneumonitis, vascular lesions and viral infections, particularly cytomegalovirus(CMV), although bacterial and fungal diseases were commoner in the non-grafted subjects. Thepathogenesis of interstitial pneumonitis is discussed with specific reference to the possible roles ofirradiation, chemotherapy, viruses and the immunosuppressive drug cyclosporin A. Ten patientsdied of a syndrome characterised clinically by fever, skin rash, fluid retention, uraemia, low serumalbumin concentrations, low central venous pressure and acute pulmonary oedema. Thesepatients exhibited intra-alveolar haemorrhagic fibrinous exudation with or without interstitialchanges. The aetiology of this syndrome is not known but it occurs more frequently in recipientsof mismatched grafts and evidence is presented suggesting that viruses may play a significantcausative role. No lesion was identified that could be directly attributed to Graft-versus-Hostdisease.

The use of bone marrow transplantation in thetreatment of leukaemia has led to a significantimprovement in prognosis.' However, it has alsointroduced a new spectrum of complicated iatro-genic human pathology of which pulmonarycomplications form a major proportion and accountfor a significant number of deaths.The aim of this study was to make a detailed his-

tological appraisal of the various lesions occurring inthe lungs of patients who have died after bone mar-row transplantation for leukaemia. By comparingthese lesions with those seen in conventionallytreated patients, it was hoped to identify thosepathological processes specifically associated withtransplantation. It was also hoped to glean someinformation about the possible aetiological roles ofvarious factors such as radiation, drugs, infectiousagents and Graft-versus-Host (GvH) disease in theproduction of these lesions by looking for correla-tions with various other laboratory and clinical data.At the time of writing allogeneic or syngeneic

marrow transplantation has been performed at the

Accepted for publication 29 December 1982

Royal Marsden Hospital on 173 patients, of whom75 have died. Necropsy material is available on thelungs of 32 of these patients and forms the basis ofthis study.

Material and methods

Full necropsies were performed on 29 patients; thelungs were weighed, inflated with formol saline andcut into slices 1 cm thick. An average of five blocksper case were taken from any suspicious areas. Inthe remaining three cases, only post-mortem needlespecimens were available.The lungs from an unselected consecutive series

of 21 ungrafted patients who had died of acutemyeloid leukaemia were similarly examined forcomparison.

Paraffin-embedded sections (3-4 ,um) werestained with haematoxylin and eosin (H&E), Gramstain for bacteria, Grocott for fungi and Pneumocys-tis carinii, Perls' for haemosiderin, Picro-Mallory forfibrous tissue and fibrin, and Verhoeff s stain forelastin. Indirect immunocytochemical stains forepithelial membrane antigen (EMA) were also car-ried out as described previously.2 EMA is a large

546

copyright. on D

ecember 27, 2021 by guest. P

rotected byhttp://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.36.5.546 on 1 May 1983. D

ownloaded from

http://jcp.bmj.com/

Histopathology of the lung after bone marrow transplantation

molecular weight glycoconjugate which can bedetected in paraffin-embedded sections by conven-tional antisera raised to milk-fat globule mem-branes. It is confined to, but widely distributed in,epithelial cells usually only on the surface mem-branes. In the lung there is staining of the surfacemembranes both of type I and type II pneumocytesas well as the luminal membranes of many of thebronchial epithelial cells. All other cell types arenegative. The stain is thus of value in distinguishingtype 1I pneumocytes from macrophages, in assessingthe thickness of alveolar walls in areas of exudationand, in some cases, assessing the integrity of type Icells where they are too attenuated for resolution bylight microscopy (Fig. 1).The age of the transplanted patients ranged from

10-46 yr (mean 28 yr). The primary diagnosis wasacute myeloid leukaemia in 26 cases, chronicgranulocytic leukaemia in three and acute lympho-blastic leukaemia in three. The time interval betweentransplantation and death varied between 2 and120 wk (mean 27 wk). Pretransplant conditioningwas achieved using cyclophosphamide (60 mg/kg x2) and total body irradiation (TBI) in all patients.Irradiation was given as a single fraction with amaximum lung dose of 9*10-12*95 Gy (mean10.13). The dose rate was 0*028 Gy/min in 30patients and 0*08 Gy/min in two. Four patients alsoreceived a second graft from their original donorsfollowing relapse and were conditioned with cyc-lophosphamide (300 mg/M2) and melphalan (180-240 mg/M2). Twenty-eight patients received cyclo-sporin A; in two of these, the drug was giventherapeutically to treat established GvH disease but

. ¶.*'.

*

in the remainder, it was given prophylactically, gen-erally at a dose of 12*5 mg/kg/day starting one daybefore grafting. All patients received marrow fromfamily member donors. In 20 (including one identi-cal twin), the graft was matched on HLA typing andmixed lymphocyte reaction (MLR). In the other 12mismatched grafts there was a varying degree ofHLA and/or MLR incompatibility.

Results

Details of the transplanted patients are summarisedin Table 1.

PATHOLOGICAL FEATURESThe lung lesions fell into seven major categories,namely interstitial pneumonitis, acute haemorrhagicpulmonary oedema, pneumonitis due to bacteria,fungi or pneumocystis, leukaemic infiltration andvascular abnormalities. The incidence of theselesions in the transplant recipients is compared withthat in the 21 non-transplanted patients in Table 2.

INTERSTITIAL PNEUMONITISThis was seen in 16 (50%) of the 32 transplantpatients and was characterised by thickening of thealveolar walls by oedema, fibrosis and mononuclearcell infiltration usually of minor to moderate sever-ity. The alveoli in the affected zones were lined byhyaline membranes and many contained haemor-rhagic fibrinous exudate (Fig. 2). The hyaline mem-branes showed very variable and sometimes nega-tive staining for fibrin. They often contained finelygranular EMA-positive material and this together

r' ('r"l

, e% Fig. 1 Lung exhibiting interstitialpneumonitis stained for EMA.Note positive type II pneumocytesand negatve macrophages. Thealveolar walls are thickned (largelyby fibrosis) and there is a lack ofcontinuity of the epithelium at thebottom right ofthe pictureassociated with intra-alveolarshedding and phagocytosis ofgranular EMA-positive material(arrow). Haemalum counterstainxSOO

547

IV ii

copyright. on D



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/

Sloane, Depledge, Powles, Morgenstern, Trickey, DadyTable 1 Summary of clinical and pathological details of32 patients who have died after bone marrow transplantationPatient Diagnosis TBI Histocompatibility Survivalnumber and status fromSex on Max Dose firstAge grafting lung rate (second)(yr) dose (Gy/min) graft

(Gy) (wk)5-M-21 AML rel 9-50 0-028 M 248-M-10 ALL 2nd CR 9-50 0-028 Syn 1213-M-21 ALL 2nd CR 9-50 0-028 M 4123-M-10 ALL 3rd CR 9-28 0-028 M 1328-M-21 AML 2nd CR 9-10 0-028 M 732-F-33 AML 1st CR 9-75 0-028 M 534-M-34 AML 1st CR 9-44 0-028 M 3952-M-46 AML 1st CR 9 44 0-028 M 1135-F-27 AML 1st CR 9-05 0-028 M 5848-F-36 AML 2nd CR 9-33 0-028 M 2859-M-45 AML rel 9-55 0-028 MM 764-F-21 AML 2nd CR 9-37 0-028 MM 450-M-37 AML 1st CR 9-65 0-028 MM 4274-M-21 AML 1st CR 10-37 0-028 MM 251-M-29 AML 1st CR 9-52 0-028 M 63 (8)89-M-37 AML 1st CR 10-44 0-028 M 7100-F-28 CGL 10-41 0-028 M 679-M-31 CGL 10-12 0-028 M 3161-M-17 AML 1st CR 9-38 0-028 M 79 (16)43-F-13 AML rel 10-00 0-028 M 120 (53)

109-F-34 AML 1st CR 10-29 0-028 M 14114-F-34 AML 2nd CR 12-95 0-028 MM 9127-F-44 AML rel 10-20 0-028 MM 4132-M-19 AML 2nd CR 10-37 0-028 MM 3134-M-33 AML 1st CR 11-31 0-028 MM 2118-F-20 AML 2nd CR 10-60 0-028 MM 24108-F-27 AML 1st CR 10-71 0-028 M 13117-F-15 AML 1st CR 11-88 0-028 MM 10121-M-44 AML 1st CR 10-78 0-08 MM 22137-F-27 CGL 10-59 0-028 M 683-F-20 AML 2nd CR 10-28 0-028 MM 79 (3)82-M-30 AML 1st CR 9-95 0-08 M 73

Diagnosis and status on grafting:AML = acute myeloid leukaemia; ALL = acute lymphoid leukaemia; CGL = granulocytic leukaemia; rel = leukaemic relapse; CR = complremission.HistocompatibilityM = matched graft; Syn = syngeneic graft; MM = mismatched graft.Cyclosporin AN = cyclosporin A not given. T = cyclosporin A given therapeutically only. P = cyclosporin A given prophylactically.Figures in second column = mean of serum concentrations on days 8, 15 and 29 after grafting in ng/ml.

Table 2 Incidence ofpulmonary lesions in leukaemic patents treated by bone marrow transplantation and conventionalchemotherapy

Lesion 32 patients receiving 21 patients treated bytransplants chemotherapy alone

Interstitial pneumonitis (total) 16 (50%) 2 (10%)-with CMV inclusions . 8(25%) 0-with pneumocyte atypia alone 5 (16%) 0-with normal or hyperplastic pneumocytes 3 (9% 2 (10%)Acute haemorrhagic pulmonary oedema 5 (16% 0Vascular abnormalities 16 (50% 3 (14%Giant cell pneumonitis (measles type) 0 1 (5%)Pneumocystis pneumonitis 2 (6%) 0Bacterial pneumonitis 9 (28%) 11 (52%Fungal pneumonitis 1 (3%) 5 (24%Leukaemic infiltration 2 '6%' 7 (33%Normal 2 (6%) 0

with the lack of EMA staining of the alveolar walls Zones of bacterial bronchopneumonia coexistedunderlying the membranes indicated that they were, with interstitial pneumonitis in three cases andat least in part, derived from degenerate epithelial leukaemic infiltration in one. Here the histologicalcells. assessment was made well away from these areas. In

548

copyright. on D



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/


Histological Cyclosporin A Lung Vascular AcuteGvH histology lesions pulmonarydisease oedema

syndrome

Yes N- Bact Pn* ES NoNo N - IPn with PA* None NoYes N - Fungal Pn, P carinii Pn* None NoNo N- lPn with CMV ES + T NoYes T- IPn None NoYes T- IPn ES + T NoYes P- IPn with CMV,* Bact Pn* T NoYes P 1162 IPn with CMV,* Bact Pn* T NoYes P- IPn with PA, LI None NoNo P- P carinii Pn T NoYes P- IPn with CMV* None YesNo P 889 lPn with PA* T YesYes P 703 Bact Pn* None NoYes P- AHPO* ES + T YesYes P 1197 Bact Pn* T NoYes P 92 lPn with CMV T NoYes P 888 Normal None NoYes P 249 Bact Pn, LI None NoYes P 455 Normal None NoNo P- Bact Pn None NoNo P 1294 IPn with CMV* T NoYes P 1596 lPn with CMV* None NoNo P 1427 Bact Pn* None NoYes P 1066 AHPO* None YesNo P 1234 AHPO* ES + T YesYes P 1190 AHPO* None YesYes P 1861 lPn with PA* T YesNo P 1526 lPn with CMV* T YesYes P 2796 IPn with PA* None YesNo P 2000 lPn,* Bact Pn* T NoNo P 1006 Haemorrhage None NoYes P 282 AHPO* T Yes

HistolopyIPn = interstitial pneumonitis. lPn with PA = interstitial pneumonitis with pneumocyte atypia. IPn with CMV = interstitial pneumonitis withcytomegalovirus inclusions. Bact Pn = bacterial pneumonitis, Fungal Pn = fungal pneumonitis. P carinii Pn = pneumonitis due to pneumocystis.AHPO = acute haemorrhagic pulmonary oedema. LI = leukaemic infiltration.*Asterisk indicates that pulmonary pathology was considered to have contributed to patient's death.Vascular lesionsES = endothelial swelling; T = thrombi.Patients are listed in chronological order of death.

the remaining 12 uncomplicated cases the changeswere randomly distributed throughout the lung;they were patchy but extensive, with lung weightsvarying from 630 to 1260 g (mean 926 g). A small

clear pleural effusion (up to 250 ml) was present inthree of the cases. Apart from CMV inclusions as

described below, no organisms could be demon-strated within the zones of interstitial pneumonitis.Three histological types of interstitial pneumonitiscould be recognised on the basis of the appearanceof the pneumocytes.(a) Interstitial pneumonitis with pneumocyte atypiaand cytomegalovirus (CMV) inclusions was seen ineight (25%) of the 32 transplanted patients but innone of the ungrafted subjects. The pneumocytes(presumably type II cells) were atypical in appear-ance with marked swelling and pleomorphism and a

proportion of them contained CMV inclusions

(Fig. 3a). The CMV-containing cells were identifiedas epithelial from their positive EMA staining.(b) Interstitial pneumonitis with pneumocyte atypiaalone was also restricted to the transplanted groupand found in five patients (16%). The appearance ofpneumocytes was essentially similar to that in group(a) but viral inclusions could not be identified(Fig. 3b).(c) Interstitial pneumonitis with normal or hyper-plastic pneumocytes was seen in three (9%) of thetransplanted and in two (10%) of the ungraftedpatients. Although the pneumocytes were usuallyhyperplastic, the atypical features present in groups(a) and (b) were absent. The two non-graftedpatients had both received unusually large doses ofchemotherapy. In one, acute myeloid leukaemia(treated by conventional chemotherapy) developedsix months after successful chemotherapy for Hodg-

549

copyright. on D



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/

Sloane, Depledge, Powles, Morgenstern, Trickey, Dady

'I"'1111,11W t.jw'r.

-V

.0 ..1 *

Fig. 2 Interstitial pneumonitis showingmononuclear cell infiltration of thealveolar wall, hyaline membranes andintra-alveolar exudation offibrin,erythrocytes and macrophages.Haematoxylin and eosin x 315

04 Aj

t

I- .1.

Fig. 3 (a) A pneumocyte containing a classical CMVinclusion is seen in the centre ofthe field; (b) interstitialpneumonitis with pneumocyte atypia showing an atypicalpneumocyte which resembles that seen in (a) except for thepresence of the intranuclear inclusion. Haematoxylin andeosin x 500

., .

550

001,0% .,F*.:: it I..vI O.'A 4V.T,a

':i,

copyright. on D



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/


kin's disease. In the other, high dose melphalan hadbeen given.3

Acute haemorrhagic pulmonary oedemaThis was seen in five (16%) of the transplant casesand in none of the ungrafted group. It was character-ised by intra-alveolar and, to a lesser extent, intersti-tial exudation of oedema fluid, fibrin and erythro-cytes. Hyaline membranes, pneumocyte changesand interstitial mononuclear cell infiltration andfibrosis were absent. In one case, occasional tiny fociof interstitial pneumonitis with hyperplasticpneumocytes were also present but changes werenot extensive enough to warrant inclusion in theinterstitial pneumonitis group. No organisms couldbe demonstrated in any of these cases.Lung weights were not always very high and var-

ied from 600 to 1250 g (mean 873 g). In one case(an adult male of average size) the lungs weighedonly 580 g and 600 g and yet all patients in thisgroup died an acute lung death as described below.

Vascular abnormalitiesThese took the form of endothelial swelling andthrombi (Fig. 4). The latter were usually confined toarterioles, capillaries and venules although theywere occasionally seen in larger vessels. They werenever extensive with usually only a few lesions persection. They were found in 16 (50%) of the trans-planted patients and in three (14%) of the ungraftedgroup. Of the transplant patients, the lesions werefound in 10/16 (63%) patients with interstitial

.'I

Fig. 4 Pulmonary blood vessel showing endotheli,hyperplasia and thrombosis. Haematoxylin and eos126

pneumonitis, 3/5 patients with acute haemorrhagicpulmonary oedema (60%) and in 3/11 remainingcases (27%). In the three ungrafted cases, thrombiwere seen in one case of interstitial pneumonitis andin two cases of bacterial bronchopneumonia.

Pneumocystis pneumonitisThis was seen in two (6%) of the transplant patientsand in none of the ungrafted group. In one casethere was also Aspergillus sp but in the otherPneumocystis carinii was the only infectious agent.

Bacterial pneumonitisBacterial pneumonitis occurred in 9 (28%) of thegrafted patients and in 11 (52%) of those not receiv-ing a transplant. There was coexistent interstitialpneumonitis in three and leukaemic infiltrationin one. The appearances were usually typical ofbacterial bronchopneumonia seen in non-immunosuppressed patients except that the degreeof neutrophil infiltration was often reduced. Gramstaining revealed a variety of organisms.

Fungal pneumonitisThis was seen in only one (3%) of the graftedpatients and in five (24%) of those treated bychemotherapy alone. Three cases were due to Can-dida sp, two to Aspergillus sp and one to Trichospo-ron sp.

Leukaemic infiltrationThis was seen more commonly in patients treatedwith chemotherapy alone (see Table 2).

CLINICOPATHOLOGICAL CORRELATIONS INTRANSPLANTED PATIENTSThere was no difference in mean age in any of thepathological groups. The time interval betweengrafting and death ranged from 4-58 wk (mean15 wk) for patients with interstitial pneumonitis,2-73 wk (mean 21 wk) for patients with acutehaemorrhagic pulmonary oedema, and 4-120 wk(mean 47 wk) for the remainder. Although thedegree of fibrosis and mononuclear cell infiltrationof the alveolar wall varied markedly in cases ofinterstitial pneumonitis, this was not related to thetime after transplantation. The mean dose of radia-tion to the lung was 10-13 Gy for patients with inter-stitial pneumonitis, 10-52 Gy with those with acutehaemorrhagic pulmonary oedema and 9-81 Gy forthe remainder. These differences are not significant.However, both patients in this study who had

t received the higher dose rate of 0*08 Gy/min died ofal acute respiratory failure as described below; one of,ia x them had interstitial pneumonitis (with pneumocyte

atypia) and one acute haemorrhagic pulmonary

551

NL

copyright. on D



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/


oedema.The incidence of previous GvH disease was 63%

in the interstitial pneumonitis group, 80% in theacute haemorrhagic pulmonary oedema group and64% in the remainder. The difference between theacute haemorrhagic pulmonary oedema cases andthe remainder does not reach statistical significance.Thirty-one per cent of patients with interstitialpneumonitis received mismatched grafts, all withatypical pneumocytes with or without CMV inclu-sions. This compares with 80% for patients withacute haemorrhagic pulmonary oedema and 27%for the remainder. The number of patients in theacute haemorrhagic pulmonary oedema group is toosmall to achieve statistical significance. Five of the16 patients with histological interstitial pneumonitisdied of acute pulmonary oedema as described belowand four of these five received mismatched grafts.Of the remaining 11 patients, interstitialpneumonitis was considered to have contributed todeath in six.Serum concentrations of cyclosporin A were

determined by radioimmunoassay at two hours afterthe dose on days 8, 15, and 29 after grafting in 20patients. The mean concentration for nine patientswith interstitial pneumonitis was 1468 ng/ml com-pared to 881 ng/ml for the remainder. This is asignificant difference (p < 0.05).Ten patients died with a syndrome characterised

by fever, skin rash, fluid retention, uraemia, lowserum albumin concentrations, low central venouspressure and pulmonary oedema45 (Table 1). Allpatients had interstitial pneumonitis associated withextensive intra-alveolar exudation, or acutehaemorrhagic pulmonary oedema. There was nosignificant difference between this group and theother patients in mean cyclosporin concentration asdefined above, dose of TBI to the lung, time aftergrafting, or in the incidence of GvH disease. How-ever, the proportion of patients who received mis-matched grafts was 8/10 (80%) compared with 4/22(18%) of the remaining patients who did not diewith this syndrome. This difference is significant(p < 0-001). Only one out of the 11 patients withinterstitial pneumonitis who did not die of acutepulmonary oedema received a mismatched graft.

Discussion

This study demonstrates that patients dying afterbone marrow transplantation for acute leukaemiaare more liable to have interstitial pneumonitis,acute pulmonary oedema of the permeability type,viral infections and vascular lesions in the lung thanthose dying after treatment with conventionalchemotherapy. The incidence of bacterial and fungal

pneumonitis was higher in non-grafted patients,probably because they are more likely to suffer fromprolonged and severe neutropenia due tochemotherapy or leukaemic relapse.

Clinically manifest interstitial pneumonitis is awell recognised complication of bone marrow trans-plantation which occurs in up to 40% of patients insome centres and may carry a high mortality. Somecases are associated with CMV inclusions, whereasin the remainder, labelled idiopathic, organismscannot be found.

Previous studies from our unit have reported alow clinical incidence of interstitial pneumonitis of10% in 107 patients.6 This difference between clini-cal and necropsy incidences is explained partly bypatient selection. However, a few patients with his-tological evidence of interstitial pneumonitis at nec-ropsy were not diagnosed in life because the changeswere of insufficient severity to be detected by con-ventional clinical methods. Recent lung functionstudies, however, have shown an invariable deterio-ration in function after grafting7 but a lack of biopsydata prevents us from assessing the histologicalchanges in these patients. It is thus clearly importantto distinguish interstitial pneumonitis as diagnosedclinically and histologically.The pathogenesis of interstitial pneumonitis

appears to be multifactorial and irradiation almostcertainly plays a role. Interstitial thickening,thrombi, desquamation, hyperplasia and atypia ofalveolar epithelial cells together with intra-alveolarfibrinous exudation and hyaline membrane forma-tion may all occur after radiotherapy to the lung.8The Seattle group has shown that TBI increases

the risk of interstitial pneumonitis in patients receiv-ing transplants for aplastic anaemia9 and others havedemonstrated a relation between interstitialpneumonitis and the dose rate of TBI.'O 1 Theobservation that both patients in the present studywho received the higher dose rate of 0-08 Gy/mindied with acute pulmonary oedema clearly warrantsfurther study.However, there is much evidence to indicate that

irradiation is by no means the only factor involved.In most histological studies, changes have beendescribed in lungs which have been irradiated in thecourse of treating malignant tumours where muchhigher, fractionated doses are given. Interstitialpneumonitis has, however, been described in asignificant number of patients receiving a single doseof 6-10 Gy to the upper body only but the dose ratehas been much higher (0-4-4.0 Gy/min) than is usedin bone marrow recipients."2 Dose rate has beenshown experimentally to be a crucial factor in thegenesis of pulmonary lesions.'3 In short, the preciseeffect of a single dose of irradiation, uncomplicated

552

copyright. on D



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/


by other factors, on the human lung at the level anddose rate given to the patients in this study is notknown.

Several of our observations are against radiationbeing the sole cause. There appears to be no relationbetween the extent of lung damage, the degree offibrosis or the amount of mononuclear cell infiltra-tion and the time after grafting. There was no rela-tion between interstitial pneumonitis and dose ofTBI, and bronchial epithelial changes such as focalnecrosis and squamous metaplasia which have beenseen in radiation pneumonitis were not observed.Furthermore, Depledge et al have shown that thereis no correlation between the extent of deteriorationin lung function after transplantation and the dose ofTBI.7

Cytotoxic drugs must also be implicated as lungchanges similar to those induced by irradiation havebeen seen after administration of a wide range ofagents including busulphan," bleomycin,'5 melpha-lan,'6 and cyclophosphamide;'7 the last of these is, ofcourse, particularly relevant to this study. Further-more, it has been shown that chemotherapy (includ-ing cyclophosphamide) may enhance lung damageafter thoracic irradiation.'8We found significantly higher serum concentra-

tions of cyclosporin A in those patients who werefound to have interstitial pneumonitis at necropsy.Previous authors have observed that this drug mayaffect small blood vessels producing angio-oedema,'9 or intravascular thrombosis.2021 Increasedalveolar capillary permeability and thrombosisinduced by cyclosporin A may be significant factorsin the genesis of interstitial pneumonitis.Although the transplanted patients in the present

study showed no increased tendency to die with bac-terial or fungal pneumonia, there was a strikingincrease in the incidence of pulmonary viral infec-tions. The predominant organism in this study, aswell as in those from other centres, was CMV.Whether this represents a specific susceptibility ofthe transplant recipient to this particular virus orwhether CMV is merely easier to detect in histologi-cal sections is not clear. Appelbaum et a122 studied100 recipients of marrow from identical twins andfound no evidence of CMV infection and a markedlyreduced incidence of interstitial pneumonitis eventhough all patients had received cyclophosphamideand TBI prior to grafting. This suggests that CMV iscausally related to interstitial pneumonitis; it alsodemonstrates an important relation between the his-tocompatibility of the graft and the incidence of viralinfections. Appelbaum et al suggested that this maybe due to GvH disease destroying host immune cellsand necessitating the administration of post-transplant immunosuppression. Indeed, Meyers et

a19 have reported that moderate to severe GvH dis-ease is associated with an increased risk of CMVassociated interstitial pneumonitis. In the presentstudy, there was an increased incidence of GvH dis-ease in patients with interstitial pneumonitis but thiswas not related specifically to those associated withCMV.

It has been shown recently that CMV infection inbone marrow transplant recipients is associated withlow or absent CMV specific cytotoxic lymphocyteactivity before the onset of infection and that survi-val depends on being able to mount cytotoxicresponses.23 Some of these cytotoxic cells are T cellswith HLA-restricted activity; that is, they lyseHLA-matched target cells better than mismatchedones. This may be another significant factor in theincreased susceptibility to virus infections of reci-pients of allogeneic transplants compared to thosereceiving syngeneic grafts.

In our study there was a high incidence of mis-matched grafts in those who died of acute pulmo-nary oedema. Some of these patients had interstitialpneumonitis whereas in others the alveolar wallsappeared normal, at least at the light microscopelevel. All patients though, had extensive intra-alveolar exudation of fibrinous haemorrhagic mater-ial. Although endothelial damage may contributesignificantly to the pathogenesis of interstitialpneumonitis, current evidence indicates that alveo-lar epithelial cells form the major barrier betweenthe blood and the alveolar lumen.24 The demonstra-tion, by EMA staining, that CMV is within theepithelial cells, the association of acute pulmonaryoedema with the transplantation of mismatchedmarrow and the associated clinical features of skinrash and fever suggest that viral infections ofpneumocytes may play a significant role in produc-ing high molecular weight intra-alveolar leaks intransplanted patients. Although definite CMV inclu-sions were seen in only 3/11 patients, conventionalhistopathology is a most insensitive method fordetecting viruses. Further studies using electronmi-croscopy, immunohistological techniques with anti-bodies to viral components and in situ nuclei acidhybridisation methods would be of value and arebeing investigated at present.Although GvH disease has occasionally been

reported in the lung, the appearance is that of alymphocytic bronchitis25 which was not observed inany of the patients in this study. None of the lesionswe have described was related to the occurrence ofGvH disease in the skin or other sites and none hasbeen reported as a graft versus host reaction in con-trolled studies on experimental animals.26

The bone marrow transplantation programme is

553

copyright. on D



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/


supported by the Leukaemia Research Fund. Weare indebted to Dr MG Ormerod of the Institute ofCancer Research for supplying antisera to EMA.We would like to thank Mr Philip Court for photo-graphic assistance and Mrs VC Williams for typingthe manuscript.

References

'Powles RL, Morgenstern G, Clink HM, et al. The place of bonemarrow transplantation in acute myelogenous leukaemia.Lancet 1980;i: 1047-50.

2 Sloane JP, Ormerod MG. Distribution of epithelial membraneantigen in normal and neoplastic tissues and its value in diag-nostic tumor pathology. Cancer 1981;47:1786-95.

McElwain TJ, Hedley DW, Burton G, et al. Marrow autotrans-plantation accelerates haematological recovery in patients withmalignant melanoma treated with high-dose melphalan. Br JCancer 1979;40:72-80.

4Powles RL, Morgenstern GR. Allogeneic bone marrow trans-plantation using mismatched family donors. In: White D, ed.Cyclosporin A. Elsevier North-Holland, 1982.

5Depledge MH, Barrett A, Morgenstern G, Sloane JP, PowlesRL. Pulmonary oedema with leaky endothelia syndrome. ExpHematol 1982;10:113.

6 Barrett A, Depledge MH, Morgenstern GR, Powles RL. intersti-tial pneumonitis following bone marrow transplantation afterlow dose rate total body irradiation. Int J Radiat Oncol BiolPhys (in press).

Depledge MH, Barrett A, Powles RL. Lung function after bonemarrow grafting. Int J Radiat Oncol Biol Phys (in press).

8 Margolis LW, Phillips TL. Whole-lung irradiation for metastatictumor. Radiology 1969;93:1173-9.

Meyers JD, Flournoy N, Thomas ED. Nonbacterial pneumoniaafter allogeneic marrow transplantation: Review of ten years'experience. Rev Infect Dis (in press).

Bortin MM, Kay HEM, Gale RP, Rimm AA. Factors associatedwith interstitial pneumonitis after bone marrow transplanta-tion for acute leukaemia. Lancet 1982;i:437-9.

Barrett A, Depledge MH. Total body irradiation: some factorsaffecting outcome. Exp Hemnatol 1982;10:56-63.

12 Van Dyk J, Keane TJ, Kan S, Rider WD, Fryer CJH. Radiationpneumonitis following large single dose irradiation: A re-

evaluation based on absolute dose to lung. Int J Radiat OncolBiol Phys 1981;7:461-7.

13 Depledge MH, Barrett A. Dose-rate dependence of lung damageafter total body irradiation in mice. Int J Radiat Biol1982;41:325-34.

41 Heard BE, Cooke RA. Busulphan lung. Thorax 1968;23:187.,s Bedrossian CWM, Luna MA, Mackay B, Lichtiger B. Ultrastruc-

ture of pulmonary bleomycin toxicity. Cancer 1973;32:44-51.16 Codling BW, Chakera MH. Pulmonary fibrosis following therapy

with Melphalan for multiple myeloma. J Clin Pathol1972;25:668-73.

"1 Gould VE, Miller J. Sclerosing alveolitis induced by cyclophos-phamide. Am J Pathol 1975;81:513-30.

18 Steel GG, Adams K, Peckham MJ. Lung damage in C57B 1 micefollowing thoracic irradiation: enhancement by chemotherapy.Br J Radiol 1979;52:741-7.

9 Isenberg DA, Snaith ML, Al-Khader AA, Cohen SL, Fisher C,Morrow WJW, Mowbray J. Cyclosporin relieves arthralgia,causes angioedema. N Engl J Med 1980;303:754.

20 Leithner C, Singzinger H, Pohanka E, Schwarz M, KretschmerG, Syre G. Recurrence of haemolytic uraemic syndrome trig-gered by cyclosporin A after renal transplantation. Lancet1982;i: 1470.

21 Shulman H, Striker G, Deeg HJ, Kennedy M, Storb R, ThomasED. Nephrotoxicity of Cyclosporin A after allogeneic marrowtransplantation. N Engl J Med 1981;305:1392-5.

22 Appelbaum FR, Meyers JD, Fefer A, et al. Nonbacterial non-

fungal pneumonia following marrow transplantation in 100identical twins. Transplantation 1982;33:265-8.

23 Quinnan GV, Kirmani N, Rook AH, et al. Cytotoxic T cells incytomegalovirus infection. N Engl J Med 1982;307:6-13.

24 Robin ED. Permeability pulmonary edema. In: Fishman AP,Renkin EM, eds. Pulmonary edema. American PhysiologicalSociety, Bethesda, Maryland: 1979.

25 Berschorner WE, Saral R, Hutchins GM, Tutschka PJ, SantosGW. Lymphocytic bronchitis associated with Graft-versus-Host Disease in recipients of bone marrow transplants. N EnglJ Med 1978;299:1030-6.

26 Woodruff JM, Eltringham JR, Casey HW. Early secondary dis-ease in the rhesus monkey: I. A comparative histopathologicstudy. Lab Invest 1969;20:499-511.

Requests for reprints to: Dr JP Sloane, The RoyalMarsden Hospital, The Haddow Laboratories, CliftonAvenue, Sutton, Surrey SM2 5PX, England.

554

copyright. on D



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/

histopathology ofthelung after bone marrow transplantation

Documents