pulmonary infiltrates in diagnosis fluid · identified by cytology. twoepisodes were caused by...

J Clin Pathol 1984;37:390-397

Pulmonary infiltrates in immunocompromisedpatients: diagnosis by cytological examination ofbronchoalveolar lavage fluidJENNIFER A YOUNG, JM HOPKIN, WP CUTHBERTSON

From the Departments ofPathology and Medicine, Queen Elizabeth Hospital and University ofBirmingham,Birmingham

SUMMARY Thirty pulmonary infiltrates in 26 patients were investigated by bronchoalveolar lav-age. Sixteen of the patients were on therapeutic immunosuppression for renal disease or trans-plant and 10 had leukaemia, lymphoma, or allied conditions. A rapid specific diagnosis was madein 21 (70%) episodes by cytological examination of the fluid and in 28 (93%) by a combination ofcytology and microbiology. No complications from haemorrhage or pneumothorax ensued.Pneumonia due to Pneumocystis carinii was the most common diagnosis (27%), but opportunisticinfections from cytomegalovirus, candida, aspergillus, zygomycetes, and acid fast bacilli were alsoidentified by cytology. Two episodes were caused by occult pulmonary haemorrhage and fivepatients had malignant infiltration of the lung from leukaemia, myeloma, Hodgkin's disease, andlymphoplasmacytoid lymphoma. In two of these there was also evidence of infection. In sevencases with non-diagnostic cytology infections due to Staphylococcus aureus, Pseudomonasaeruginosa, pneumococcus, micrococcus, and Aspergillus fumigatus were identified on culture. Intwo patients (7%) no specific diagnosis was established by lavage: one had serological evidence oflegionella infection and the second had P aeruginosa septicaemia. Twelve (75%) of the renalpatients and six (60%) of those with leukaemia, lymphoma, and allied conditions recovered.

Pulmonary infiltrates in immunocompromisedpatients are common and have a varied aetiologyand a high mortality.' 2 Opportunistic infection is acommon cause but must be distinguished from otherconditions such as occult haemorrhage, drug orradiation toxicity, malignant disease, and non-specific interstitial pneumonitis. In addition, ifoptimum treatment is to be given the pathogenicorganism must be identified. Non-invasive investiga-tions are seldom helpful,34 and the diagnostic accu-racy of biopsy, whether transbronchial or open lung,varies considerably.5 Also some doctors are reluc-tant to advise biopsy in severely ill immunocom-promised patients.6 Fine needle aspiration of thelung causes fewer serious complications than biopsybut is associated with the same hazards of haemor-rhage and pneumothorax.'The technique of bronchoalveolar lavage has been

developed for the investigation of interstitial lungdisease.89 Published reports show it to be a safe'0

Accepted for publication 24 November 1983

procedure which obtains plentiful cellular materialfrom the lungs." The purpose of the present studywas to assess the value of cytological examination ofbronchoalveolar lavage fluid in the investigation ofimmunocompromised patients with pulmonaryinfiltrates of unknown aetiology.

Material and methods

Thirty pulmonary infiltrates in 26 patients wereinvestigated by bronchoalveolar lavage. Thepatients comprised two groups. Sixteen were fromthe Department of Nephrology and were receivingtherapeutic immunosuppression, 11 after renaltransplant and five for kidney disease (rapidly pro-gressive glomerulonephritis, systemic lupuserythematosus, and Goodpasture's syndrome). Tenpatients were from the Departments of Haematologyand Oncology. These included five patients withleukaemia, one of whom had received a bone mar-row transplant, and two patients with aplasticanaemia, one of whom had also undergone bonemarrow transplant. There was one patient with

390

on July 11, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.37.4.390 on 1 A

pril 1984. Dow

nloaded from

http://jcp.bmj.com/

Pulmonary infiltrates in immunocompromised patients

non-Hodgkin's lymphoma and one with Hodgkin'sdisease. The remaining patient was an elderlywoman who was admitted as an emergency withsevere respiratory distress. She was found to havedysglobulinaemia, the cause of which had not beenestablished at the time of bronchoscopy.The patients who had undergone renal or bone

marrow transplant were all receiving either cyclo-sporin A or prednisolone and azathioprine. Theother renal patients were taking prednisolone andcyclophosphamide, and the patients with leukaemiaand lymphoma were on a range of therapeutic regi-mens which included steroids, cytotoxic drugs, andradiotherapy. Both patients who had received bonemarrow transplants had undergone total bodyirradiation. The indications for bronchoalveolar lav-age were severe illness, with cough or dyspnoea, andabnormal chest x ray film in patients in whom nospecific microbial diagnosis had resulted fromexamination of sputum (if available) and blood. Thepatient with Hodgkin's disease had already had apercutaneous fine needle aspiration of the lung,which was non-diagnostic both cytologically and onculture.The bronchoalveolar lavage were collected during

fibreoptic bronchoscopy by the method previouslydescribed.'2 The washings were obtained from themiddle lobe or lingula in patients with diffuseradiological infiltrates and as appropriate where theabnormality was localised. Two or three 20 ml ali-quots of sterile saline were instilled and aspirated,roughly half of the volume being returned. A sampleof fluid was submitted to the Department of Mic-robiology and the remaining material sent forimmediate cytological processing.The lavage fluid was centrifuged for 5 min at 3000

rpm and smears were prepared from the deposit.These were stained by the Papanicolaou, May-Grunwald-Giemsa, Ziehl-Neelsen, and Grocottmethenamine silver methods. When the pigmentladen macrophages were numerous a Perls' stain forhaemosiderin was included. If organisms suggestiveof nocardia or actinomyces were seen a Gram stainwas also carried out. The specimen from the patientwith dysglobulinaemia was examined for immuno-

Table 1 Results ofbronchoalveolar lavage in 30 episodes

Result Cytology Microbiology Combined

Diagnostic 21* 13* 28Non-diagnostic

Non-specificabnormality 3 1

No abnormality 6 17 1Total 30 30 30Sensitivity 70% 42% 93%

*Both diagnostic in six episodes.

391

globulins (IgG, IgA, IgM, and kappa and lambdalight chains) by the unlabelled antibodyperoxidase-antiperoxidase method.'3 Antisera wereobtained commercially from Mercia-Brocades Ltd,West Byfleet, Surrey, England.The cytology was all reported by one author

(JAY). The fluid was examined for the presence ofabnormal cells, haemosiderin laden macrophages,acid fast bacilli, and evidence of viral, parasitic, ormycotic infection. Apart from acid fast bacilli,nocardia, and actinomyces, bacteria were not lookedfor in the cytological specimens.

Results

The number of specific diagnoses and the sensitivityof cytology, microbiology, and the combined techni-ques are shown in Table 1. On six occasions aspecific diagnosis was made by both cytology andmicrobiology. Four times this was the same, but ontwo occasions different aetiological factors wereidentified. In nine patients the cytology was non-diagnostic, although three displayed non-specificcellular abnormality. Tables 2 and 3 show thefindings in the two groups of patients in more detailtogether with the primary medical condition andeventual outcome.A definitive diagnosis was made in 21 (70%) of

the 30 episodes by cytology. These included 15opportunistic infections, five cases of malignantinfiltration of the lung (in one patient there wascytological evidence of both infection and malig-nancy), and two episodes of occult pulmonaryhaemorrhage. In seven patients in whom cytologywas non-diagnostic a specific diagnosis resulted fromculture, increasing the overall sensitivity to 93% (28episodes). In two of the patients who had undergonerenal transplant the overall bronchoalveolar lavageresults were non-diagnostic. In one of these aretrospective diagnosis of legionella infection wasmade on serology. The second patient was severelyill with graft rejection and ascites in addition torespiratory distress. The bronchoalveolar lavage washypercellular and contained, in addition to numer-ous macrophages, mixed inflammatory cells andlarge single bizarre cells, which were interpreted ashyperplastic alveolar lining cells. Some of the mac-rophages contained haemosiderin. These changeswere considered consistent with interstitialpneumonitis but no cause was identified. Ferrugin-ous bodies were also found in the fluid, which raisedthe possibility of pre-existing fibrosis, although therewas no known history of exposure to asbestos.This patient did not recover. At necropsy areas ofbronchopneumonia coexisted with interstitialpneumonitis. There was oedema, fibrosis, and thick-


http://jcp.bmj.com

/J C


pril 1984. Dow

nloaded from

http://jcp.bmj.com/

Table 2 Primary medical condition, results ofbronchoalveolar lavage, and outcome in 16 patients (18 episodes) from theDepartment ofNephrology

Primary medical condition Cytology Microbiology Outcome

SLE P carini: RecoveredSLE P carinii RecoveredGlomerulonephritis i) Cytomegalovirus Recovered

(ii) Zygomycetes Mucor DiedGlomerulonephritis Haemorrhage - RecoveredGoodpasture's syndrome Haemorrhage - RecoveredRenal transplant P carinii - RecoveredRenal transplant P carinii - RecoveredRenal transplant P carinii - RecoveredRenal transplant P carinii - RecoveredRenal transplant (i) P carinii - Recovered

(ii) AFB M tuberculosis RecoveredRenal transplant Candida C albicans RecoveredRenal transplant - S aureus DiedRenal transplant - S aureus and P aeruginosa DiedRenal transplant - Micrococcus RecoveredRenal transplant Non-specific pneumonitis - Died*Renal transplant - Recoveredt

*P aeruginosa septicaemia.tSerological evidence of legionella.SLE = systemic lupus erythematosus; AFB = acid fast bacilli.

Table 3 Primary medicalcondition, resultsofbronchoalveolarlavage, and outcomein 10patients (12 episodes) from theDepartments of Haematology and Oncology

Primary medical condition Cytology Microbiology Outcome

AML Leukaemic infiltration - DiedALL (i) P carinii Recovered

(ii) Candida C albicans RecoveredALL + BMT Degeneration Micrococcus Recovered*CGL Leukaemic infiltration P aeruginosa DiedCLL Aspergillus RecoveredAplastic anaemia (i) - A fumigatus Recovered

(ii) - S aureus and P aeruginosa DiedAplastic anaemia + BMT Hypereosinophilia Pneumococcus Recovered*Hodgkin's disease Infiltration by Hodgkin's disease RecoveredNon-Hodgkin's lymphoma CMV and infiltration by P aeruginosa Died

lymphoplasmacytoid lymphomaHyperglobulinaemia Infiltration by myeloma - Recovered*

*Died within six months.AML = acute myeloid leukaemia; ALL = acute lymphoblastic leukaemia; BMT = bone marrow transplant; CGL = chronic granulocyticleukaemia; CLL = chronic lymphocytic leukaemia; CMV = cytomegalovirus.

ening of the alveolar walls. Pneumocyte hyperplasia,atypia, and desquamation were demonstrated byimmunocytochemical staining for epithelial mem-brane antigen. Pseudomonas aeruginosa was grownfrom the blood culture.The most common finding in patients with a

definitive diagnosis was opportunistic infection dueto Pneumocystis carinii. The eight cases occurred ina cluster shortly after the start of the project. All thepatients recovered after treatment with high dosecotrimoxazole. The lavage fluid contained muchamorphous debris and few alveolar macrophages.Characteristic cystic forms of P carinii containingone or two comma shaped structures together withsmall numbers of collapsed empty cysts wereidentified on the slides stained with Grocott

methenamine silver (Fig. 1). Three of these patientshad high antibody titres for cytomegalovirus infec-tion but no viral inclusions were found in the bron-choalveolar lavage fluid. Two of the eight patientsbecame ill a second time (tuberculosis and can-didiasis) during the study.

Classic "owl's eye" inclusions (Fig. 2) indicatingpulmonary cytomegalovirus infection were found inthe macrophages from two patients. One of thesealso had cytological evidence of infiltration of thelung by malignant cells from lymphoplasmacytoidnon-Hodgkin's lymphoma in addition to P aerugin-osa on culture. This patient died suddenly frommyocardial infarction 8 h after bronchoscopy. Thecytological findings were all confirmed at necropsy.The other patient with cytomegalovirus infection,

Young, Hopkin, Cuthbertson392


http://jcp.bmj.com

/J C


pril 1984. Dow

nloaded from

http://jcp.bmj.com/


Fig. 1 Cystic forms of Pneumocystis carinii with foamyexudate in background. Grocott methenamine silver x

1060.

who had rapidly progressive glomerulonephritis,made a slow recovery but died during a secondepisode of infection (mucormycosis).

Mycotic infections were found either alone or inassociation with other abnormalities on six occa-

Fig. 2 Cytomegalovirus infection. Macrophage containingintranuclear inclusion body. Papanicolaou x 825.

Fig. 3 Pseudohyphae and budding spores of Candida sp.

Grocott methenamine silver x 825.

sions. Three were due to candida, two to aspergillus,and one to mucor. Candida generally presents as amixture of yeast cells (blastoconidia) and filament-ous elements, which may be true hyphae or

pseudohyphae. Budding spores originating at pointsof constriction on the filaments are characteristic(Fig. 3). The aspergillus was not seen in the cyto-logical preparations in one of the cases but was

found on culture, and in the second instance septatehyphae with dichotomous branching were seen inthe fluid but there was no fungal growth on culture.Both these patients recovered after anti-fungaltreatment, though one later died from pyogenicinfection. Long hyphae, branching at angles up to90°, were found in one patient and were interpretedas probable zygomycetes (Fig. 4). Speciation fromculture confirmed mucormycosis. The infection was

fatal.The two cases of occult pulmonary haemorrhage

occurred in patients with Goodpasture's syndromeand rapidly progressive crescentic IgA nephritis. Oneach occasion the macrophages, when stained by thePapanicolaou technique, were filled with largegranules of golden brown pigment. Perls' stainconfirmed the presence of haemosiderin.

Malignant infiltration of the lung was diagnosed infive patients. One had acute myeloid leukaemia and

_.:

I...Ift,

393

AL qw.-AM,

:11...: ImIV


http://jcp.bmj.com

/J C


pril 1984. Dow

nloaded from

http://jcp.bmj.com/

394

Fig. 4 Ribbon-like hyphae ofzygomycetes (mucor).Grocott methenamine silver x 330.

another had chronic granulocytic leukaemia, inwhom P aeruginosa was also found. Malignant cellswere identified in the fluid from the patient withlymphoplasmacytoid non-Hodgkin's lymphoma,who also had multiple infections (cytomegalovirus,P aeruginosa). In a girl with mediastinal Hodgkin'sdisease, lymphocytes, Reed-Stemnberg cells, andmononuclear Hodgkin's cells were found in thefluid, confirming pulmonary involvement. Theelderly woman with dysglobulinaemia was clinicallysuspected to have pneumocystis infection. Thecytological specimens showed a monoclonal (IgG,lambda) proliferation of morphologically abnormalplasma cells, including many unusually large Motttype cells (Fig. 5). There was no evidence of infec-tion. The bronchoalveolar lavage diagnosis ofrespiratory distress owing to infiltration of the lungsby malignant cells from a plasma cell tumour wassubstantiated when similar cells were found in thebone marrow. Pulmonary function tests and chestx ray film improved considerably on treatment withmelphalan, but she died from myelomatosis fourmonths later.

In two cases with bacterial infections on culture,the cytology, although non-diagnostic, showed cellu-lar abnormality. The patient who had undergonebone marrow transplant had grade 2/3 graft versushost disease at the time of bronchoscopy. The lav-

Young, Hopkin, Cuthbertson

age fluid showed severe cellular degeneration espe-cially of the bronchial epithelium. The same featurewas seen in sputum specimens. The degenerationgradually cleared with improvement of graft versushost disease and treatment of infection. The speci-mens obtained from the patient who had received abone marrow transplant for aplastic anaemia alsoshowed abnormality. There were abundanteosinophils and strikingly numerous Charcot-Leyden crystals. Thorough investigation failed toidentify any parasitic or fungal cause for theeosinophilia and there was no history of bronchialasthma in the patient or donor. We concluded thatthe pulmonary findings were part of a generalisedhypereosinophilia associated with considerableregeneration of the marrow after transplantation.This patient recovered from the bacterial(pneumococcus) infection but later developed risingantibody titres to cytomegalovirus and died.

Bronchoalveolar lavage was well tolerated by allpatients. No pulmonary haemorrhage orpneumothorax ensued. Twelve (75%) of the 16renal cases and six (60%) of the 10 from thedepartments of haematology and oncology reco-vered.

Fig. 5 Plasma cells, including large binucleate Mott cell,mixed with macrophages. Lavage fluid firom patient withpulmonary infiltrate due to myelomatosis. May-GruinwaldGiemsa x 825.

1. l%-LN...

w.l.A.

O' .4.00, mL

:-1.

Amw..--, i-.4

IRW"':

:.


http://jcp.bmj.com

/J C


pril 1984. Dow

nloaded from

http://jcp.bmj.com/


Discussion

The development of a pulmonary infiltrate in animmunocompromised patient is an urgent clinicalproblem. Arriving at an exact diagnosis is madedifficult by the frequently non-specific nature of clin-ical and radiological findings, the variety of possiblecauses, and the problem of obtaining adequate rep-resentative material from the lungs of severely illimmunosuppressed patients.'4A survey of the results of open lung biopsy5 found

the mean diagnostic yield to be 69%, but in manystudies4 '5 the percentage of specific diagnoses fallsbelow this figure, though 91% was achieved byRosen et al'6 in the diagnosis ofP carinii. The overallcomplication rate was 11%.5 Limitations of openlung biopsy include the need for general anaesthetic,delay encountered in organisation, and the reluc-tance of doctors to use it in patients with recurrentinfiltrates.'7 In eight series of fibreoptic broncho-scopies the mean diagnostic yield from transbron-chial biopsy was 41%, from bronchial brushing27%, and from a combined approach 47%.5Matthay et al'8 obtained a diagnosis in 84% of theircases by the combined use of biopsy and brushing,but the study contained a high incidence of focalcarcinoma, which may have augmented the sensitiv-ity. The mean incidence of haemorrhage (> 25 ml)and pneumothorax was 7% for each complication inseven series of bronchoscopies.5 Cunningham et al,3however, found that 26% suffered from haemor-rhage and 19% from pneumothorax. Fewer studieson fine needle aspiration of the lung in immunosup-pressed patients have been published, but Sagel etal'9 reported a diagnostic rate of 77% in 31 suchpatients and Castellino and Blank20 achieved a rateof 73% in 108 infectious episodes. The most com-mon complication is pneumothorax, which occurs inabout a quarter of all patients.7 Castellino andBlank20 reported this complication in 28%, of which13% required thoracostomy. The present investiga-tion shows the sensitivity of cytological examinationof bronchoalveolar lavage fluid (70%) to be com-parable to that of published reports on biopsy5 andfine needle aspiration'9 20 in immunosuppressedpatients. In contrast, however, there were no com-plications due to haemorrhage or pneumothorax.There is increasing recognition of the value of

morphological assessment of histological andcytological material for the rapid identification ofunusual pathogens,2'23 especially P carinii and cer-tain genera of opportunistic viruses and fungi. Thiscan be carried out on the same preparations and atthe same time as the evaluation of specimens fornon-infectious causes. This is especially valuable inpatients who are immunodeficient as a result of

malignant disease. Cytology, however, has little tocontribute to the diagnosis of bacterial infections.Bronchoalveolar lavage provides abundant materialwhich is suitable for microbiology as well as cytol-ogy. By using both techniques on the fluid the sen-sitivity rate was increased from 70% to 93%. Twopatients had evidence of both neoplastic infiltrationof the lung and infection, which emphasises thevalue of the combined approach to diagnosis.Opportunistic infection was diagnosed by cytologyas the cause of the infiltrate in 15 (50%) episodes. Ineight instances (27%) the pathogen was P carinii.An incidence of 32-38% has been found in other

4 24series, confirming the increasing importance ofthis organism as a cause of pneumonia inimmunocompromised hosts. P carinii is oftenassociated with cytomegalovirus.25 Three of thesepatients had serological evidence of cytomegalovirusinfection but no indication of pulmonary disease inthe bronchoalveolar lavage fluid. The specimensfrom patients with P carinii pneumonia containedmuch amorphous exudate and only small numbersof alveolar macrophages, which may perhaps haveresulted in false negative findings. The patients allrecovered after treatment with cotrimoxazole. Thisis unlikely to have occurred if pulmonarycytomegalovirus was a contributory factor in therespiratory disease. In interpreting cytologicalspecimens stained with Grocott methenamine silvercare must be taken not to mistake clusters of fungalspores or artefactual deposition of silver on leuco-cytes and erythryocytes as P carinii. Adequate con-trol material is essential.Many viruses produce no specific cytological

changes. Several which cause opportunistic respira-tory infection are associated with characteristiccytological markers, however, which permit rapidpreliminary identification at genus level. In additionto cytomegalovirus, these include, Herpesvirushominis (Herpes simplex virus), Herpesvirus varicel-lae (varicella zoster), and adenovirus. While somefeatures produced by cytomegalovirus, herpes sim-plex virus, and varicella zoster do show overlap,basophilic Cowdry type A26 intranuclear inclusions,intracytoplasmic inclusions, and cytomegaly charac-terise cytomegalovirus infection. The eosinophilicinclusions, "ground glass" nuclei, and cell fusionwith multinucleation and moulding found in herpessimplex virus and varicella zoster are not a feature ofcytomegalovirus infection. Basophilic inclusionsoccur with adenovirus, but these are associated with"smudge" cells,22 which are not seen withcytomegalovirus.

Proof of the presence of mycotic infection is ide-ally established by culture, but this is slow and thefungus may fail to grow.2327 There are also many

395


http://jcp.bmj.com

/J C


pril 1984. Dow

nloaded from

http://jcp.bmj.com/

396

limiting factors in the use of immunologicallyspecific fluorescent antibody stains.23 Preliminarymorphological identification using empirical stainingmethods can therefore be a rapid and valuable aid inpatient management. Both candida and aspergillus,however, occur in the upper respiratory tract unas-

sociated with lower respiratory tract infection.Fungal elements which are not closely intermingledwith alveolar macrophages should therefore beinterpreted with caution. Speciation of candida,aspergillus, and zygomycetes is not satisfactorywhen based on morphology alone, but identificationat genus or class level is adequate for appropriatetreatment to be instituted. Culture showed the caseof zygomycosis to be due to mucor. Aspergillus wasfound in two patients, once by cytology and once onculture. Both patients responded to antifungaltreatment, confirming the diagnoses and the com-plementary nature of the investigations.

In two patients, one with Goodpasture' s syn-drome and the other with rapidly progressive IgAcrescentic glomerulonephritis, no infection wasidentified, but there was cytological evidence ofoccult pulmonary haemorrhage. Antimicrobialtreatment was withheld, both x ray films cleared,and the patients recovered. Macrophages laden withhaemosiderin may also be found in severe pulmo-nary oedema, but this is clinically diagnosable andresponds to diuretics. In the cases of infectivepneumonia and also in the patient with pneumonitisand pseudomonas septicaemia smaller numbers ofPerls' positive macrophages were present but theydid not approach the numbers seen in the two caseswith pulmonary haemorrhage.

Five patients had malignant infiltration. Thelikelihood of this is influenced by the general state ofthe disease-pulmonary dissemination is unlikelywithout evidence of advancing malignancy else-where. Both patients with leukaemia were inhaematological relapse, the patient with Hodgkin'sdisease had mediastinal node disease, and thepatient with lymphoplasmacytoid lymphoma hadrecurrent lymphadenopathy. The elderly womanwith myelomatosis, who presented as an urgent caseof respiratory distress, also had widely disseminateddisease. Evidence of malignant infiltration of thelung does not, however, preclude the possibility ofcoexisting infection, as was found in two cases in thisseries.Pneumonia accounts for half of the deaths in some

transplant programmes,' but mortality is highest inthe first six months and long term prognosis is gen-erally good if early infection is survived. The poten-tial for recovery is increased by rapid diagnosis andthe institution of prompt specific antimicrobialtreatment.28 This is especially true of infection by P

Young, Hopkin, Cuthbertson

carinii. In the present series the 12 renal patientswho survived all remained alive and well at followup one year later. Six (60%) of the patients withleukaemia, lymphoma, and allied conditions reco-vered. Owing to the nature of their primary disease,however, the long term outlook of this group ismuch less favourable and at six months survival hadfallen to 30%.As a result of this study, we believe that cytologi-

cal examination of bronchoalveolar lavage fluid is asafe, rapid, and helpful diagnostic aid in the man-agement of immunocompromised patients withpulmonary infiltrates.

We thank the physicians from the Departments ofNephrology, Haematology, and Oncology, whoreferred their patients for bronchoscopy and espe-cially Dr D Adu, Dr J Michael, Dr JH Turney, DrBJ Boughton, and Dr IM Franklin. We gratefullyacknowledge the contribution of many members ofthe Department of Microbiology and thank Dr JCrocker, Dr S Ghosh, and Dr J Simon for post-mortem histology, Dr AJ Howie for theimmunocytochemical staining for EMA, and Mrs JCheney for secretarial assistance.

References

'Ramsay PG, Rubin RH, Tolkoff-Rubin NE, Cosmi AB, RussellPS, Greene R. The renal transplant patient with fever andpulmonary infiltrates: etiology, clinical manifestations, andmanagement. Medicine 1980;59:206-22.

2 Pennington JE, Feldman NT. Pulmonary infiltrates and fever inpatients with hematologic malignancy: assessment of trans-bronchial biopsy. Am J Med 1977;62:581-7.

Cunningham JH, Zavala DC, Corry RJ, Keim LW. Trephine airdrill, bronchial brush and fibreoptic transbronchial lung biop-sies in immunosuppressed patients. Am Rev Resp Dis1977;115:213-20.

Singer C, Armstrong D, Rosen PP, Walzer PD, Yu B. Diffusepulmonary infiltrate in immunosuppressed patients: prospec-tive study of 80 cases. Am J Med 1979;66: 110-20.

Matthay RA, Moritz ED. Invasive procedures for diagnosingpulmonary infection: a critical review. Clinics in ChestMedicine 1981;2:3-18.

6 Anonymous. Pulmonary problems of the immunocompromisedpatient (Editorial). Br Med J 1981;282:2077.

7Sinner WN. Complications of percutaneous transthoracic needleaspiration biopsy. Acta Rad (Diag) 1976;17:813-28.

Weinberger SE, Kelman JA, Elson NA, et al. Bronchoalveolarlavage in interstitial lung disease. Ann Intern Med1978;89:459-66.

9Hunninghake CW, Cade JE, Kawanami 0, Ferrans VJ, CrystalRG. Inflammatory and immune processes in the human lung inhealth and disease: evaluation by bronchoalveolar lavage. AmJ Pathol 1979;97:149-206.

'° Strumpf IJ, Feld MK, Cornelius MJ, Keogh BA, Crystal RG.Safety of fibreoptic bronchoalveolar lavage in evaluation ofinterstitial lung disease. Chest 1981;80:268-71.

"Haslam PL, Turton CWG, Heard B, et al. Bronchoalveolar lav-age in pulmonary fibrosis: comparison of cells obtained withlung biopsy and clinical features. Thorax 1980;35:9-18.


http://jcp.bmj.com

/J C


pril 1984. Dow

nloaded from

http://jcp.bmj.com/

Pulmonary infiltrates in immunocompromised patientsHopkin JM, Turney JH, Young JA, Adu D, Micheal J. The rapid

diagnosis of abscure pneumonia in immunosuppressed renalpatients by cytology of alveolar lavage fluid. Lancet 1983;ii:299-301.

Curran RC, Gregory J. Demonstration of immunoglobulin incryostat and paraffin sections of human tonsil byimmunofluorescence and immunoperoxidase techniques. JClin Path 1978;31:974-83.

4 Fanta- CH, Pennington JE. Fever and new lung infiltrates in theimmunocompromised host. Clinics in Chest Medicine1981;2:19-39.

's Rossiter SS, Miller DC, Churg AM, Carrington CB, Mark JB.Open lung biopsy in the immunosuppressed patient. Is it reallybeneficial? J Thorac Cardiovasc Surg 1979;77:338-45.

6 Rosen PP, Martini N, Armstrong D. Pneumocystis cariniipneumonia: diagnosis by lung biopsy. Am J Med1975;58:794-802.

Puksa S, Hutcheon MA, Hyland RH. Usefulness of transbron-chial biopsy in immunosuppressed patients with pulmonaryinfiltrates. Thorax 1983;38:146-50.

18 Matthay RA, Farmer WC, Odero D. Diagnostic fibreoptic bron-choscopy in the immunosuppressed host with pulmonaryinfiltrates. Thorax 1977;32:539-45.

Sagel SS, Ferguson TB, Forrest JV, Roper CL, Weldon CS,Clark RE. Percutaneous transthoracic aspiration needlebiopsy. Ann Thorac Surg 1978;26:399-405.

20 Castellino RA, Blank N. Etiologic diagnosis of focal pulmonaryinfection in immunocompromised patients by fluoroscopicallyguided percutaneous needle aspiration. Radiology 1979;132:563-7.

21 Johnston WW, Frable WJ. Diagnostic respiratory cytopathology.New York: Masson Publishing USA, 1979:96-150.

22 Myerowitz RL. The pathology of opportunistic infections. NewYork: Raven Press, 1983:23-49.

23 Schwarz J. The diagnosis of deep mycoses by morphologicmethods. Hum Pathol 1982;13:519-33.

24 Tenholder MF, Hooper RG. Pulmonary infiltrates in leukaemia.Chest 1980;78:468-73.

25 Haselton PS, Curry A. Pneumocystis carinii: the continuingenigma (Editorial). Thorax 1982;37:481-5.

26 Cowdry EV. The problem of intranuclear inclusions in virusdiseases. Arch Pathol 1934;18:528-42.

27 Pennington JE. Aspergillus pneumonia in hematologic malig-nancy: improvements in diagnosis and therapy. Arch InternMed 1977;137:769-71.

28 Rubin RH, Wolfson JS, Cosimi AB, Tolkoff-Rubin NE. Infec-tion in the renal transplant recipient. Am J Med1981;70:405-1 1.

Requests for reprints to: Dr Jennifer A Young, Depart-ment of Pathology, The Medical School, University ofBirmingham, Birmingham B15 2TJ, England.

397


http://jcp.bmj.com

/J C


pril 1984. Dow

nloaded from

http://jcp.bmj.com/

pulmonary infiltrates in diagnosis fluid · identified by cytology. twoepisodes were caused by...

Documents