J Clin Pathol 1986;39:945-949

Diagnosing Pneumocystis carinii pneumonia bycytological examination of bronchoalveolar lavagefluid: report of 15 casesJENNIFER A YOUNG,* J W STONE,t R J S MCGONIGLE,4 D ADU,$J MICHAEL:

From the Departments of *Pathology and tMicrobiology, the University of Birmingham, and the IDepartmentof Nephrology, Queen Elizabeth Hospital, Birmingham

SUMMARY Sixty episodes of pneumonia occurring in 53 immunosuppressed patients wereinvestigated by bronchoalveolar lavage. Pneumocystis carinii was diagnosed on 15 (25%) occasions.In all cases the Papanicolaou stained lavage fluid presented a distinctive appearance and containedabundant, often biphasic, staining, "honeycomb" debris, and few alveolar macrophages. TheGrocott methenamine silver technique confirmed the presence of characteristic cystic organisms inthe debris in all 15 instances. Cysts containing internal sporozoites were identified in Gram stainedmaterial only with difficulty. Neither May-Grunwald-Giemsa stain nor fluorescence microscopyunder ultraviolet light were effective for routine investigation.

The importance of Pneumocystis carinii as a lethalopportunistic pathogen in immunocompromisedpatients is now widely recognised.' - Untreatedpneumocystis pneumonia is almost invariably fatal,and survival depends on early specific antimicrobialtreatment.5 Discovering the exact aetiology of respi-ratory distress in immunosuppressed patients is, how-ever, often difficult as the possible causes are manyand the symptoms, clinical signs, and radiographicalfeatures are generally non-specific. Culture of thehuman strain ofP carinii is unsuccessful, and serologyis unreliable in immunodeficient subjects. Definitivediagnosis must be made by morphological identifica-tion of the organism. The value of cytological exam-ination of bronchoalveolar lavage fluid as aninvestigative technique in immunocompromisedpatients with pulmonary infiltrates has been reportedpreviously.6

Material and methods

Over four years bronchoalveolar lavage was per-formed 62 times on 53 immunosuppressed patientsduring 60 episodes of pneumonia. Series I' contained

Accepted for publication 18 March 1986

26 patients (30 examinations) and series II an addi-tional 27 patients (32 examinations). Seven patientsunderwent lavage twice for different episodes ofpulmonary illness, and one patient was examinedthree times during the course of the same episode. Thelavage fluid was collected and processed for cytologyby the methods previously described.67 Specimensfrom series I were stained by Papanicolaou, May-Grunwald-Giemsa, Ziehl-Neelsen, and Grocottmethenamine silver methods, with the addition ofPerls's and Gram stains only if these seemed appro-priate after preliminary microscopy. In series II theGram stain (with carbol fuchsin) was performed onall specimens and the slides examined for evidence ofP carinii, as described by Macher et al.8 The Papa-nicolaou stained slides from series II were also retro-spectively reviewed using fluorescence microscopy,according to the method of Ghali et al.9

Results

A total of 15 cases of P carinii pneumonias were diag-nosed during the four years. The table shows the age,primary medical condition, cytological findings, andoutcome of the illness for each of the patients. Cases1 to 8 are from series I and cases 9 to 15 are fromseries II. Case 15 was examined three times. Apartfrom one early case with acute lymphoblastic leu-

945

copyright. on 27 F

ebruary 2019 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.39.9.945 on 1 S

eptember 1986. D

ownloaded from

Table

Case No Sex Age Primary medical condition Cytologicalfindings Outcome

I M 16 Leukaemia P carinii, Candida Sp Recovered2 M 61 Renal transplant P carinii, Candida Sp Recovered3 F 29 Systemic lupus erythematosus P carinii Recovered4 F 24 Systemic lupus erythematosus P carinii Recovered5 M 51 Renal transplant P carinii Recovered6 M 26 Renal transplant P carinii Recovered7 M 45 Renal transplant P carinii Recovered8 M 43 Renal transplant P carinii Recovered9 M 50 Renal transplant P carinii Died10 F 54 Renal transplant P carinii, Candida Sp RecoveredII F 41 Renal transplant P carinii, cytomegalovirus Died12 F 47 Renal transplant P carinii Recovered13 M 53 Renal transplant P carinii Died14 M 26 Renal transplant P carinii Recovered15 F 54 Renal transplant (i) P carinii, cytomegalovirus

(ii) P carinii(iii) P carinii, cytomegalovirus Died

kaemia all the patients were from the renal unit andreceiving immunosuppressive treatment either forlupus glomerulonephritis or following renal trans-plantation. The 14 cases of P carinii pneumonia in therenal unit occurred in two clusters, each of sixmonths' duration, separated by an interval of twoyears. In all instances the bronchoalveolar lavagefluid presented a characteristic appearance whenstained by the Papanicolaou technique. The moststriking feature was the presence of abundant foamydebris, which displayed considerable depth of focus.At high magnification the debris had a spongy

"honeycomb" texture with numerous small clearcircles surrounded by amorphous material. Somedense aggregates showed biphasic staining (fig 1).The numbers of leucocytes and alveolar macro-

phages were low. As might be expected, paucity ofleucocytes is common in bronchoalveolar lavage fluidfrom immunodeficient subjects but alveolar macro-

phages are generally plentiful. This combination ofhoneycomb debris and reduced macrophages was

seen only in the Papanicolaou stained slides ofpatients who were found to have P carinii infection.Definitive diagnosis, however, requires actualidentification of the cystic forms or trophozoites of Pcarinii, and these are not visualised by the Papa-nicolaou technique. Cysts ofP carinii stain black withGrocott methenamine silver, and on all 15 occasionswere readily found within the aggregates of the foamydebris. (fig 2). They were not identified lying free inthe fluid or within macrophages. Cysts have thin wallsand contain two small "comma shaped" structures.Collapsed cysts have a characteristic "cup like"appearance. Trophozoites do not stain with Grocott.In two cases in series II small numbers of cysts witheight internal sporozoites were clearly seen on theslides stained with Gram (fig 3). In the other cases

they were found only with great difficulty. Diagnosticorganisms could not be identified with certainty in

any of the preparations stained with Man-Grunwald-Giemsa. The Papanicolaou stained slides from seriesII were retrospectively reviewed under ultraviolet illu-mination with a fluorescence microscope. Where thefoamy debris showed a biphasic reaction with thePapanicolaou stain, there was backgroundfluorescence of the orangeophilic stained material.This impeded visualisation of diagnostic cysticorganisms. Occasional cysts, showing faintfluorescence, were identified around the periphery ofsome of the masses of debris, but it was impossible todistinguish confidently these from erythrocytes,which also displayed some fluorescence, withoutreviewing the same field with conventional lightmicroscopy. Ingested dust particles and erythrocyteswithin the small numbers of alveolar macrophagespresent also exhibited fluorescence.

Case 15 was examined three times during thecourse of the same illness, and P carinii was identifiedon each occasion.. In two of these specimens "owl'seye" inclusions (fig 4) indicative of cytomegalovirusinfection were seen in addition to P carinii. Cyto-megalovirus was also diagnosed in case 11, and fungalspores and pseudohyphae consistent with Candida spwere found in the fluid from cases 1, 2, and 10. Two ofthe patients who recovered from pneumocystis pneu-monia became ill a second time; one developedpulmonary tuberculosis and the other invasivecandidiasis.

Discussion

P carinii was first recognised as a cause of epidemicpneumonia during the second world war,10 althoughthe initial description is generally attributed toChagas." Expansion of organ transplant pro-grammes, improved survival in haematological malig-nancies, and since 1981,12 the acquired immunedeficiency syndrome (AIDS) have all contributed to

Young, Stone, McGonigle, Adu, Michael946

copyright. on 27 F

ebruary 2019 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.39.9.945 on 1 S

eptember 1986. D

ownloaded from

Diagnosing pneumocystic carinii pneumonia

Fig I Foamy "honeycomb" debris in lavagefluidfrom Fig 2 Cystic organisms ofP carinii in mass offoamy debris.patient with P carinii pnewnonia. This three dimensional (Grocott.) x 1000.aggregate shows biphasic staihing, afeature often, but notinvariably, present. (Papanicolaou). x 560.

I

d

Fig 3 Intracystic sporozoites ofP carinii (arrow) amongaggregate ofdebris containing weakly stainingfreetrophozoites. (Gram.) x 1000.

Fig 4 Binucleate cell (arrow) containing "owl's eye"inclusions characteristic ofcytomegalovirus infection.(Papanicolaou.) x 1000.

947

.dou

copyright. on 27 F

ebruary 2019 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.39.9.945 on 1 S

eptember 1986. D

ownloaded from

948

an increase in immunocompromised subjects at riskand a rise in the number of reported cases of thisformerly rare type of opportunistic pneu-monia.5 13 -18 Experience in Britain, however, is stilllimited -6-18 and diagnosis is considered to be prob-lematic.18 Untreated pneumocystis pneumonia isalmost always fatal, but apart from those with AIDS,about 80% of adults recover after being given co-trimoxozole.19 The long term prognosis for renaltransplantation recipients is generally excellent, pro-viding they do not succumb to infection.20 Althoughchemoprophylaxis has been shown to be effective inpaediatric oncology clinics,21 clinical anxiety remainsconcerning recent reports of adverse reactions to co-trimoxazole in adults,22 23 and toxicity is known tooccur with the alternative administration of pen-tamidine.24 Potential for recovery, therefore, dependson prompt specific treatment early in the course of thedisease5: if undesirable multiple empirical treatmentsare to be avoided rapid definitive diagnosis of P car-min infection is imperative.

Patients with pneumocystis pneumonia presentwith a progressive lung infiltrate of unknownaetiology. Possible causes are many and include bac-terial, fungal and viral infections, pulmonary hae-morrhage, malignant infiltration, and drug or radi-ation toxicity. At present, the only reliable method ofdistinguishing P carinii infection is by mor-phologically identifying the organism in materialobtained from the lung fields. Various different sam-pling techniques have been advocated in publishedstudies. Rosen et at25 reported 100% sensitivity fortransbronchial biopsy in 32 cases of AIDS; otherauthorities prefer percutaneous fine needleaspiration26 or open lung biopsy.1 27 Transbronchialbiopsy and needle aspiration are associated with arisk of haemorrhage or pneumothorax,26 28 and thereare limitations on the use of open lung biopsy,especially in patients with recurrent chest problems.29

Bronchoalveolar lavage, however, has been shownto be a safe procedure,30 21 even in thrombocytopenicpatients and those requiring mechanical ventilatorysupport. In common with other groups we found thatthis was an effective investigative method forrecovering diagnostic material from the lungs ofimmunosuppressed patients with pulmonaryinfiltrates.6 7 31For the diagnosis ofP carinii pneumonia by exami-

nation of bronchoalveolar fluid, three prerequisitesshould be borne in mind. The lavage must be thor-ough enough to obtain true alveolar material: serialspecimens of fluid instilled and aspirated at the samesite are generally required to ensure this, as the firstaliquot usually contains only bronchial debris, andalveolar contents containing P carinii may not be

Young, Stone, McGonigle, Adu, Michael

Secondly, maximum care is needed during the fixationand staining of the recovered cells to obtain the bestpossible morphological detail: in particular, controlslides are necessary for timing the Grocott stain.Lastly, the microscopist examining the specimensshould be familiar with the appearance of P carinii.Without correct appreciation of the morphology ofthe organism, distinction from fungal spores,erythrocytes or debris in macrophages may be aproblem, especially in material stained with Grocott.Spores have much thicker walls and may showbudding. Both fungi and P carinii can, however, bepresent in the same specimen (cases 1, 2, and 10). Spu-rious deposition of silver on erythrocytes or leuco-cytes generally has a speckled appearance and debrisis often angular or shows considerable variation insize.P carinii was confirmed as the cause of the respira-

tory distress syndrome in 15 (25%) of the 60 episodesthat occurred in the 53 patients. As far as the estab-lishment of alternative diagnoses and the clinicalcourse of the disease on the 45 other occasions couldpredict there were no false negative results from bron-choalveolar lavage in either series I or II. The cyto-logical picture of "honeycomb" debris and few mac-rophages, when seen in Papanicolaou stained lavagefluid, is sufficiently distinctive to permit a rapid pre-liminary diagnosis of pneumocystis pneumonia fromthe initial routine slides, which, of course, can also beexamined for alternative or coexisting pathologicalprocesses.6 Visualisation, however, of cysticorganisms or trophozoites using special staining tech-niques is advisable for confirmatory definitive diagno-sis. In all 15 cases cystic forms ofP carinii were readilyidentified with the Grocott stain. Although the regu-lar Grocott method takes about three hours to per-form, examination of the stained slides requires onlya few moments. In cases of great urgency the rapidmethod32 can be used. P carinii has previously beenrecognised in lung imprints stained with Gram.8 Thisis a very quick stain to perform and is available in alllaboratories, certainly an undoubted advantage onoccasion. Diagnosis by this method, however,entailed examining large amounts of cellular materialand was extremely time consuming (up to an hour atthe microscope), and easily identifiable organismswere found only in two cases. Although the devel-opment of an "eye", with practice, wouldundoubtedly speed up the microscopy, we do not feelthis method is sufficiently reliable to be a satisfactorymeans of confirming P carinji infection. Fluorescencemicroscopy, although requiring no additional stain-ing, requires a special microscope and is also verytedious to perform. We agree with Ghali et al9 thatthe technique is no faster than the Grocott stain when

copyright. on 27 F

ebruary 2019 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.39.9.945 on 1 S

eptember 1986. D

ownloaded from

Diagnosing pneumocystic carinii pneumonia

ence this was not found to be a suitable alternative tosilver staining. Fluorescence microscopy may have aplace in retrospective searches of archival material,but silver techniques can, if necessary, be satis-factorily performed on slides already stained withPapanicolaou. We were unable to see organisms inmaterial stained with Man-Griinwald-Giemsa,although some authorities33 recommend this method.

In conclusion, if facilities for fibreoptic bron-choscopy and cytopathology are available exami-nation of bronchoalveolar lavage fluid is an efficientmethod for the diagnosis of P carinii pneumonia.Bronchoalveolar lavage avoids the risk of hae-morrhage or pneumothorax and achieves extensivealveolar sampling. With the routine Papanicolaoustain the appearance is sufficiently distinctive to per-mit rapid preliminary diagnosis while at the sametime the specimens can be examined for other abnor-malities, such as evidence of cytomegalovirus. Gro-cott methenamine silver staining can be readily per-formed on the lavage material and remains the bestconfirmatory technique.We thank the clinicians from the department ofmedicine, who performed the bronchoalveolarlavage, especially Dr JM Hopkin, who examined allthe patients in series I. We also thank Mr WPCuthbertson for technical help and Mrs Joyce Cheneyfor typing the manuscript.The colour photography was financed by a grant

from the Renal Research Fund of the QueenElizabeth Hospital, for which we are extremely grate-ful.References

I Haselton PS, Curry A. Pneumocystis carinii: the continuingenigma. Thorax 1982;37:481-5.

2 Kay SB. Pneumocystis pneumonia. Br Med J 1983;286:499-500.3 Anonymous. Pneumocystis-an orphan organism? [Editorial]

Lancet 1985;i:676-7.4 Macfarlane JJ, Finch RG. Pneumocystis carinii pneumonia.

Thorax 1985;40:561-70.5 Rubin RH, Wolfson JS, Cosimi AB, Tolkoff-Rubin NE.

Infection in the renal transplant recipient. Am J Med198 1;70:405-1 1.

6 Hopkin JM, Young JA, Turney JH, Adu D, Michael J. Rapiddiagnosis of obscure pneumonia in immunosuppressed renalpatients by cytology of alveolar lavage fluid. Lancet1983;ii:299-301.

7 Young JA, Hopkin JM, Cuthbertson WP. Pulmonary infiltratesin immunocompromised patients: diagnosis by cytologicalexamination of bronchoalveolar lavage fluid. J Clin Pathol1984;37:390-7.

8 Macher AB, Shelhamer J, Maclowry J, Parker M, Masur H.Pneumocystis carinii identified by Gram stain of lung imprints.Ann Intern Med 1983;99:484-5.

9 Ghali VS, Garcia RL, Skolom J. Fluorescence of Pneumocystiscarinii in Papanicolaou smears. Hum Pathol 1984;15:907-9.

10 Gajdusek DS. Pneumocystis carinii-etiological agent of inter-stitial plasma cell pneumonia of premature and young infants.Pediatrics 1957;19:543-65.

11 Dutz W. Pneumocystis carinii. Pathol Annu 1970;5:309-41.12 Gottlieb MS, Schroff R, Shanker HM. Pneumocystis carinii

pneumonia and mucosal candidiasis in previously healthyhomosexual men: evidence of a new acquired cellular

949immunodeficiency. N EnglJ Med 1981;301:1425-30.

13 Kuller J, First MR, D'Achiardi R, Munda R. Pneumocystis cari-nii pneumonia in renal transplant recipients. Am J Nephrol1982;2:213-5.

14 Murray JF, Felton CP, Garay SM, et al. Pulmonary compli-cations of the acquired immune deficiency syndrome. N Engi JMed 1984;310:1682-8.

15 Marchevsky A, Rosen MJ, Chrystal G, Kleinerman J. Pulmonarycomplications of the acquired immunodeficiency syndrome: aclinicopathologic study of 70 cases. Hum Pathol 1985;16:659-70.

16 Haselton PS, Curry A, Rankin EM. Pneumocystis carinii pneu-monia: a light microscopical and ultrastructural study. J ClinPathol 1981;34:1 138-46.

17 Swinburn CR, Pozniak AL, Sutherland S, Banks RA, Teall AJ,Johnson NMCI. Early experience and difficulties with bron-choalveolar lavage and transbronchial biopsy in the diagnosisof AIDS associated pneumonia in Britain. Thorax1985;40: 166-70.

18 Ballardie FW, Winearls CG, Cohen J, Carr DH, Rees AJ,Williams G. Pneumocystis carinii pneumonia in renal trans-plant recipients-clinical and radiographic features, diagnosisand complications of treatment. QJ Med 1985;57:729-47.

19 Young LS. Trimethoprim-sulphamethoxazole in the treatmentof adults with pneumonia due to Pneumocystis carinii.Rev Infect Dis 1982;4:608-13.

20 Ramsay PG, Rubin RH, Tolkoff-Rubin NE, Cosmi AB, RussellPS, Greene R. The renal transplant patient with fever and pul-monary infiltrates: etiology, clinical manifestations, and man-agement. Medicine 1980;59:206-22.

21 Hughes WT. Five-year absence of Pneumocystis carinii pneu-monitis in a paediatric oncology centre. J Infect Dis1984;150:305-6.

22 Gordin FM, Simon GL, Wolsy CB, Ills J. Adverse reactions totrimethoprim-sulphamethoxazole in patients with the acquiredimmunodeficiency syndrome. Ann Intern Med 1984;100:495-9.

23 Committee on Safety in Medicines. Deaths associated with co-trimoxazole, ampicillin and trimethoprim. Current Problems1985;15:1.

24 Walzer PD, Perl DP, Krogstad DJ, Rawson PG, Schultz MG.Pneumocystis carinii pneumonia in the United States. AnnIntern Med 1974;80:83-93.

25 Rosen MJ, Tow TWY, Teirstein AS, Chuang MT, MarchevskyA, Bottone J. Diagnosis of pulmonary complications of theacquired immune deficiency syndrome. Thorax 1985;40:571-5.

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

27 Groeve AH, Saul VA, Aki BF. Role of different methods of lungbiopsy in the diagnosis of lung lesions. Am Rev Respir Dis1982;125:601-3.

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

29 Puksa S, Hutcheon MA, Hyland RH. Usefulness of trans-bronchial biopsy in immunosuppressed patients with pul-monary infiltrates. Thorax 1983;38:146-50.

30 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.

31 Stover DE, Zaman MB, Hajdu SI, Lange M, Gold J, ArmstronD. Bronchoalveolar lavage in the diagnosis of diffuse pul-monary infiltrates in the immunosuppressed host. Ann InternMed 1984;101:1-7.

32 Mahan CT, Sale GE. Rapid methenamine silver stain for pneu-mocystis and fungi. Arch Pathol Lab Med 1978;102:351-2.

33 Dunnill MS. Pulmonary pathology. Edinburgh: ChurchillLivingstone, 1982:173-6.

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

copyright. on 27 F

ebruary 2019 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.39.9.945 on 1 S

eptember 1986. D

ownloaded from