ACTA CYTOLOGICA 0001-5547/10/5403-0296/$21.00/0 © The International Academy of Cytology296

Objective

To study the role of cytology and polymerase chain reaction(PCR) for the detection of Pneumocystis carinii pneumo-nia (PCP) in bronchoalveolar lavage fluid (BALF).

Study Design

In a prospective observationalstudy, a BALF pellet from 55patients (35 immunosup-pressed patients clinically sus-picious for PCP and 20 im-munocompetent individualsnot suspicious for PCP) weresubjected to cytology examina-tion as well as PCR for PCPusing the major surface glyco-protein gene. Additionally,DNA was extracted from destained cytology slide scrapingsfrom 5 cases each positive for PCR and negative for PCP onBALF, respectively, and was further subjected to PCR.

Results

Of 35 immunosuppressed patients, 2 were positive on mi-

croscopy for PCP, while 5 (including 2 microscopy positive)were positive by PCR from BALF. Four of the 5 positivecases showed predominantly alveolar macrophages, whileone showed 40% alveolar macrophages, 30% lymphocytesand 20% neutrophils upon cellular analysis of BALF. Onecase had coinfection with Cryptococcus. Destained smear

scrapings of PCR-positiveslides (n = 5) were all success-fully amplified for PCP majorsurface glycoprotein PCR afterDNA extraction, while controlslides (n = 5) did not show am-plification.

Conclusion

BALF is considered the sampleof choice for the diagnosis of

PCP. In our study, cases with PCP showed mainly alveolarmacrophages on BALF examination by cytology and no ormild inflammatory cells. Thus, inflammatory cell popula-tion found on cytology may provide direction for PCR. Two-step screening of BALF using cytology followed by PCR fromslides is possible and may overcome the limitations of BALF

An inflammatory cell population ofBAL fluid is an important aspect of

cytologic evaluation and can be usedto target samples for PCR and special

stain analysis.

Role of Cytology and Polymerase ChainReaction Based Detection of Pneumocystisjirovecii Infection in Bronchoalveolar LavageFluid

Rashmi Gupta, M.Sc., Venkateswaran K. Iyer, M.D., Bijay Ranjan Mirdha, M.D.,Randeep Guleria, M.D., Lalit Kumar, M.D., and Sanjay Kumar Agarwal, M.D.

From the Departments of Microbiology, Pathology, Medicine, Medical Oncology, and Nephrology and Transplantation, All India Institute ofMedical Sciences, New Delhi, India.

Ms. Gupta is Ph.D. Student Candidate, Department of Microbiology

Dr. Iyer is Associate Professor, Department of Pathology.

Dr. Mirdha is Additional Professor, Department of Microbiology.

Dr. Guleria is Professor, Department of Medicine.

Dr. Kumar is Additional Professor, Department of Medical Oncology.

Dr. Agarwal is Professor, Department of Nephrology and Transplantation.

Address correspondence to: Bijay Ranjan Mirdha, M.D., Department of Microbiology, All India Institute of Medical Sciences, New Delhi-110029, India (mirdhabr@hotmail.com).

Financial Disclosure: The authors have no connection to any companies or products mentioned in this article.

Received for publication January 20, 2009.

Accepted for publication March 24, 2009.

Molecular Techniques

cytology alone for the diagnosis of PCP. (Acta Cytol2010;54:296–302)

Keywords: bronchoalveolar lavage fluid, Pneumocystispneumonia, polymerase chain reaction.

Pneumocystis carinii pneumonia (PCP) has long beenrecognized as a cause of morbidity and mortality inimmunocompromised populations, particularly thosewith human immunodeficiency virus (HIV) infection.It is caused by a fungus, Pneumocystis jirovecii (former-

ly known as P carinii).1 Host defense against P jiroveciidepends mainly on cellular immunity and depletion ofCD4+ T lymphocytes observed in patients with ac-quired immunodeficiency syndrome results in fatalpneumonia. The hallmark of PCP is profound hypox-emia and abnormal chest radiography findings withnonproductive cough. Cytologic examination of sam-ples from the respiratory tract has been one of themainstays in the laboratory diagnosis of PCP. P jiro-vecii is considered to adhere tightly to the alveolar wallwith its pseudopodia, and it is therefore difficult to ob-tain high-quality sputum for precise diagnosis. Fiber-optic bronchoscopy and analysis of bronchoalveolarlavage fluid (BALF) is efficient in evaluating this con-dition. As in vitro culture of P jirovecii is often not pos-sible, the diagnostic identification of the organism de-pends primarily on direct demonstration of it inappropriate respiratory clinical samples using specialstains, such as Giemsa and Grocott’s silver methan-amine (GMS).2 P jirovecii infection in immunocom-promised hosts sometimes results in rapid and fatalpneumonia, and an early, definitive diagnosis is noteasy.

Polymerase chain reaction (PCR) technology wasfirst applied for the diagnosis of PCP by Wakefield etal3 to improve both sensitivity and specificity. To fur-ther improve the sensitivity of detection, the nestedPCR approach was developed.4 In view of the highsensitivity of the test and possibility of false positiveresults in PCR, real-time PCR is considered the goldstandard for diagnosis; however, it is expensive andnot a feasible option in many developing countrieshaving a high incidence of HIV infection.

297

Detection of Pneumocystis jirovech Infection

Volume 54 Number 3 May–June 2010 ACTA CYTOLOGICA

In our institution, we offer both routine cytologyand conventional PCR on BALF samples for diagno-sis of PCP. The role of cytology has in the past beenrestricted to identification of the infective organism.5We have also described improved detection of P jiro-vecii in respiratory clinical samples from immunocom-promised patients by using PCR.6 The results of PCRand GMS staining of some of the samples shown inthis paper are identical to the BALF data reported inour previous paper. In our study it was observed thatamong the 143 various kinds of respiratory clinicalsamples, 5 were positive by GMS staining and 18 werepositive by PCR assay. Out of 143, 64 were BALFsamples (44%), of which 14 were positive by PCR and5 (n = 5) by GMS staining. Thus, the sensitivity andspecificity were 100% and 99%, respectively, for PCRand 30.7% and 100% for microscopy by GMS stain-ing.6

Cytomorphologic analysis of the inflammatory cellsin BALF and their possible role in a targeted applica-tion of PCR has not been studied. However, differentcell populations of BALF have been analyzed in pa-tients with PCP. In one of the studies, the number ofmacrophages, lymphocytes and neutrophils in BALFdiffered significantly in individuals with PCP in con-trast to patients without PCP.7 It has also been shownthat resolution of neutrophil alveolitis was associatedwith improvement in symptoms. Elevated levels ofneutrophils and lymphocyte counts in BALF cell pop-ulation have been reported in patients with AIDS withPCP, possibly leading to fatal outcomes.7,8 In view ofthis, our experience in the analysis of inflammatorycell populations in BALF cytology, PCR detection ofP jirovecii from cytology slide scrapings and its use instepwise application of molecular diagnostic modali-ties of PCP is presented in this paper.

Materials and Methods

A prospective study was conducted at the departmentsof microbiology, pathology, medicine, medical oncol-ogy and nephrology at the All India Institute of Med-ical Sciences. A total of 55 BALF samples obtainedfrom the same number of patients were included in thestudy. Of these 55, there were 35 patients with variouskinds of immunosuppression and there was a strongclinical suspicion of PCP. Twenty immunocompetentpatients with respiratory disorders other than PCPserved as controls. These 20 cases included patientswho underwent bronchoscopy either due to com-plaints of recurrent cough or who had a past history oftuberculosis or recently detected carcinoma withoutany kind of obvious or apparent immunocompromis-ing condition.

Bronchoscopy was performed on all the patientsthrough the transnasal route using a fiberoptic bron-

Use of PCR on BALF cytology slidescrapings for successful

identification of P jirovecii even onmicroscopy-negative samples has

not been demonstrated before.

choscope (Pentax UK Ltd., Langely, Slough, U.K.).The upper airways and the tracheobronchial tree werevisualized in detail. The bronchoscope was thenwedged into the segmental bronchus (mostly rightmiddle lobe), and a lavage procedure was performedwith 100 mL of saline at room temperature with 25-mL aliquots. Fluid was recovered immediately by gen-tle suction after each aliquot until a volume of 60 mLwas collected in a sterile tube and then passed through2 sheets of gauze to eliminate mucus. First aliquot wasdiscarded; other aliquots were pooled and divided.The BAL fluid thus obtained was then transported tothe laboratory and kept at 4°C until processed.

For cytology, all BALF samples (n = 55) were sub-jected to Shandon Cytospin 4 (Thermo Fisher Scien-tific Inc., Waltham, Massachusetts, U.S.A.) prepara-tions stained for Papanicolaou, May-Grünwald-Giemsa(MGG), Ziehl-Neelsen (ZN) stain for acid-fast bacilliand GMS following standard protocols. Slides wereevaluated at routine examination followed by reevalu-ation by 2 observers. A differential count of cells in theBALF was carried out, enumerating the alveolarmacrophages (AM), lymphocytes, neutrophils andother cells. Subsequently, cases were classified on thebasis of predominant cell types. If neutrophils andlymphocytes were <10%, BAL was considered AMpredominant. If neutrophils or lymphocytes were10% or over, then it was considered predominance ofthat cell type. If AM, lymphocytes and neutrophilswere present in approximately equal numbers, it wasconsidered mixed inflammation. If > 90% of cells wereseen to be neutrophils, then it was considered an acuteinflammatory exudate, neutrophil predominant.9

To carry out a PCR test, all the samples (n = 55)were centrifuged at 4,000 rpm for 10 minutes, and thepellet obtained was resuspended in one fifth of super-natant. For DNA extraction, 200 μL of pellet waslysed in equal volume of a sample specific lysis buffer(50 mM KCl, 15 mM Tris-HCl [pH 8.3] and 0.5%NP-40) containing 500 μg of proteinase K.10 DNAextraction was done using Qiagen DNeasy tissue kit(Qiagen, Valencia, Califnornia, U.S.A.). PCR for themajor surface glycoprotein (MSG) gene of P jiroveciiwas performed by using an equal mixture of oligonu-cleotide primers JKK14 (5′-GAATGCAAATCCT-TACAGACAACAG-3′) and JKK15 (5′-GAATGCAAATCTTTACAGACAACAG-3′) as forward primers.JKK17 (5′-AAATCATGAACGAAATAACCATTGC-3′) was used as reverse primer.11 These primers am-plify a 249-bp region of the highly conserved MSGgene. The standard reaction mixture and PCR proto-col described previously was slightly modified. Thereaction mixture contained 50 mM KCl, 10 mM Tris,1.5 mM MgCl2, 0.25 mM dNTPs, 0.5 μM primersand 1.5U Taq polymerase. PCR was performed by ini-

298

Gupta et al

ACTA CYTOLOGICA Volume 54 Number 3 May–June 2010

tial denaturation at 94°C for 5 minutes followed by 35cycles at 94°C for 1 minute, 65°C for 1 minute and72°C for 1 minute in a ABI 2720, thermocycler (Ap-plied Biosystems, Foster City, California, U.S.A.).The amplified products were visualized by ethidiumbromide staining on a 1.5% agarose gel.

To perform PCR test based on destained cytologyslide scrapings, the MGG-stained slides weredestained, and the cellular contents were scraped com-pletely into Eppendorf tubes and suspended into sam-ple specific lysis buffer, described above.10 DNA wasextracted following the manufacturer’s instructionsusing a DNeasy tissue kit (Qiagen). PCR for the MSGgene as described above was performed with appropri-ate positive and negative controls. For positive con-trol, an MGG slide positive for PCP on cytomorpho-logic evaluation was used. A total of 10 destained slidescrapings were used for DNA extraction and weresubjected to PCR. These included 2 slides that werepositive for PCP on cytology (used as positive con-trols), 3 slides negative on cytology but positive forPCP on BALF PCR and 5 slides negative for PCP onboth cytology and BALF PCR.

Results

Of the 35 BALF samples (obtained from the immuno-suppressed group), 2 were microscopy positive forPCP on cytologic examination (Figure 1A and B),yielding a positivity rate of 5.7%. On applying PCR toall the BALF, a total of 5 samples (5/35, 14.2%) werepositive by PCR for MSG gene. These 5 cases includ-ed the 2 cases that were positive by microscopy and 3positive by PCR alone. Among the 5 positive cases, 3were post–renal transplant (PRT) recipients, 1 was apatient with Cushing’s syndrome and 1 had systemiclupus erythromatosis (SLE) with lupus nephritis.Among the 20 samples from immunocompetent pa-tients, none was positive by either microscopy orPCR.

On performing differential cell counts in the slidesof the 35 cases, it was observed that 12 had predomi-nant AM count (12/35, 34.2%) (neutrophils and lym-phocytes < 10%), 2 had predominant lymphocytecounts (3/35, 8.5%) (lymphocytes > 10%), 10 had pre-dominant neutrophil counts (10/35, 28.5%) (neu-trophils > 10%), 5 had acute inflammatory exudates(5/35, 14.2%) (> 90% of cells were neutrophils) and 5had mixed cell types (5/35, 14.2%) (Table I). Of the 12cases with predominant AM counts, 1 case was posi-tive by both microscopy and PCR (1/12, 8.3%), and 3were positive by PCR alone (3/12, 25%). Thus, a totalof 4 out of the 12 cases (4/12, 33.3%) with predomi-nant AM counts were positive by PCR. Among all theother cases with either predominant lymphocyte orneutrophil counts, none was positive by microscopy

299

Detection of Pneumocystis jirovech Infection

Volume 54 Number 3 May–June 2010 ACTA CYTOLOGICA

Table I Microscopy and PCR Results Among Cases with Different

Predominant Cell Types in BALF Cytology

Predominant No. of Microscopy PCR +ve

cells in BALF cases +ve for PCP for PCP

AM 12 1 4

L 3 0 0

N 15 0 0

Mixed 5 1 1

Total 35 2 5

L = lymphocyte, N = neutrophils.

and/or PCR. Of the 5 cases with mixed cell type, only1 (1/5, 20%) was positive for PCP by microscopy aswell as PCR. Thus, among the 5 cases positive byBALF PCR, 4 had predominant AM counts (4/5,80%). In the immunocompetent group (n = 20), 9 pa-tients had predominant AM count, 1 had predominantlymphocyte count, 9 had predominant neutrophilsand 1 had mixed inflammation. None of the samples inthis group was positive by any of the techniques. OnZN staining, 4 other cases showed presence of acidfast bacilli out of a total of 55. These included 1 pa-tient from the immunocompromised group (1/35) and3 from the immunocompetent group (3/20). On cyto-logic examination, all 4 showed neutrophil predomi-nance with an acute inflammatory exudate picture. Inaddition, 2 cases from the immunocompromised pa-tient group showed Cryptococcus infection; 1 of themhad concurrent PCP (Figure 1C and D).

It was observed that of the 2 HIV+ patients, bothhad predominant AM count, and none was positive forPCP (Table II). Among the 18 PRT recipients, 8 hadpredominant neutrophil count, 6 had predominantAM counts, 2 had mixed cell type and the remaining 2had predominance of lymphocytes. In the PRT group,

A B

C DFigure 1 (A) Alveolar casts of P jirovecii identified on BALF. Note Cryptococcus (arrow, budding yeasts) in the background. (B) Silver stained

organisms of P jirovecii. (C and D) Coinfection by P jirovecii and Cryptococcus (arrows) (A, Papanicolaou stain, × 400; B–D, Grocott’s silver

methanamine, × 1,000).

300

Gupta et al

ACTA CYTOLOGICA Volume 54 Number 3 May–June 2010

2 cases were positive by microscopy as well as by PCR(2/18, 11.11%), and 1 additional case was positive byPCR alone (1/18, 5.55%). Among these 2 microscop-ically positive cases, 1 had AM predominance, and theother had mixed cell type. The other PRT PCR posi-tive case also showed AM predominance. Thus, total-ly 3 cases were positive by PCR in the PRT group. Ofthe 7 patients with various malignancies on chemo-therapy, 2 had predominant AM count, 1 had pre-dominant lymphocytes and four had a predominanceof neutrophils. None of the samples in this group waspositive for PCP. In 8 patients with various otherkinds of immunosuppression, 3 had predominant neu-trophil counts, 2 had mixed cell type, and 2 had pre-dominant AM counts. In this group also, 2 cases (SLEand Cushing’s syndrome) with predominant AMcounts were positive for PCP by PCR.

When DNA from the 3 microscopically negative (1case each with SLE and Cushing’s syndrome and 1PRT recipient) but BALF PCR positive samples wasextracted from the slides and subjected to PCR, posi-tive results were obtained in all 3 samples (Figure 2A).DNA from 5 negative control slides showed no ampli-fication (Figure 2B).

Discussion

Diagnosis of PCP has been based on various methods,like evaluation of sputum, tracheal aspirate, endo-bronchial brush biopsy and percutaneous needle aspi-ration. BALF with or without transbroncial biopsy isthe procedure of choice. The diagnosis of PCP is tra-ditionally being done by microscopic examination inorder to identify and demonstrate the organism fromclinically relevant specimens, such as sputum, BALFor lung tissue.12 Early studies found that the sensitiv-ity of BALF in the diagnosis of PCP using standardcytologic stains ranges from 86% to 90%.13 Directdemonstration of the organism is often difficult or notpossible because the trophozoite or cyst form of theorganisms is either not present in abundant numbers

or is not uniformly spread out in the alveoli of infect-ed lungs.

A number of different staining techniques havebeen used by several investigators for the diagnosis ofP jirovecii. Trophic forms of the organism can be de-tected with modified Papanicolaou, Wright-Giemsa(WG) or Gram-Weigert stain. Cysts can be stainedwith GMS, toluidine blue O or Calcoflour white.14

Monoclonal antibodies for detecting Pneumocystis havea higher sensitivity in induced sputum samples thanconventional tinctorial stains; however, the differenceis much less in BALF.15 More recently, due to these

Figure 2 (A) PCR for MSG gene of P jirovecii performed on DNA

extracted from slide scrapings. Lane 1: 100-bp DNA ladder, lane 2:

positive control (DNA extracted from microscopy-positive slide),

lanes 3–5: samples (DNA extracted from microscopy-negative

slides, positive for PCP on BALF pellet PCR). (B) Lane 1: negative

sample, lane 2: positive control (DNA extracted from

microscopy-positive slide), lane 3 and 4: samples (DNA extracted

from microscopy-negative and BALF pellet PCR negative slides),

lane 5: 100-bp DNA ladder.

Table II Predominant Cell Types in BAL Fluid in Different Groups of Patients

No. of Microscopy PCR +ve

Category cases AM L N Mixed +ve for PCP For PCP

Immunocompromised group

PRT 18 6 2 8 2 2 3

HIV+ 2 2 0 0 0 0 0

Others 8 2 0 3 3 0 2

Malignancies on chemotherapy 7 2 1 4 0 0 0

Total 35 12 3 15 5 2 5

Immunocompetent group 20 9 1 9 1 0 0

Others = other kinds of immunocompromising conditions, like prolonged steroid use, primary immunodeficiency, chronic kidney disease, SLE, Cushing’s syndrome.

A B

limitations of direct demonstration of Pneumocystis inclinical samples, molecular techniques have been de-veloped for its detection, based on the amplification ofP jirovecii DNA using PCR.

A number of studies have demonstrated that PCRbased methods are more sensitive (104–106 times) inorder of magnitude as well as being more specific thanstaining methods.16-18 According to a report by Pin-laor et al,19 PCR using a part of the mitochondriallarge subunit rRNA (mtLSU rRNA) gene has shownhigher sensitivity and negative predictive values forBALF (100% each) vs. sputum, samples at 84.62% and98.41%, respectively. Using WG and GMS stain, sen-sitivity values were 64.55% and 63.64% for BALF and15.38% and 61.54% for sputum samples, respectively.The specificity of PCR was 90% and 98.4% for BALFand sputum samples, respectively. WG, GMS and im-munofluorescence assay tests had the same specifici-ties (100%) and positive predictive values for detec-tion of Pneumocystis in BALF and sputum, respectively.Similarly, studies performed by Sing et al20 showedmicroscopy and single round PCR of BALF speci-mens from HIV patients to be 100% sensitive and spe-cific in detecting PCP, whereas nested PCR was 100%sensitive and achieved a specificity of 97.5%. Recent-ly, cloning and sequencing of the MSG gene of P jiro-vecii provided a better gene target for its detection.MSG is encoded by a multicopy gene family, and it isestimated that there are 100 or more copies of MSGgene in the Pneumocystis genome.21 A study by Huanget al11 showed that PCR for the MSG gene was moresensitive in detection of Pneumocystis DNA than morewidely used PCR for the mtLSU rRNA gene.

However, one problem with PCR is that it is fre-quently positive in patients with asymptomatic car-riage with a rate of 2–21%.20-24 Real-time PCR forPneumocystis, described by Helweg-Larsen et al,23 is ofimmense use in distinguishing between colonizationand true infection.25,26

The diagnosis of PCP requires a real-time PCRanalysis of BAL fluid as the best diagnostic modality.However, countries worst affected by HIV may not beable to provide technically demanding moleculartechniques, such as real-time PCR, making it impor-tant to evaluate other investigative modalities. Evalu-ation of differential cell counts on cytology and directdemonstration of the organism by staining followedby PCR provides a better alternative.

We found that patients with PCP had few or no in-flammatory cells. The BALF cytology showed mostlyAM, indicating a normal cytology differential count.Four out of the 5 patients showed AM predominantpicture, while one showed mixed inflammatory celltypes in BAL. Earlier studies on the cell types ob-served on cytology are few.7,27 Most reports describe

301

Detection of Pneumocystis jirovech Infection

Volume 54 Number 3 May–June 2010 ACTA CYTOLOGICA

the detection of the organism with little or no inflam-mation in the background. If a diagnosis of PCP isconfirmed, the presence of neutrophils of > 5% of cellson differential count may have a predictive value onresponse to treatment.28 However, in a group of pa-tients in whom PCP is suspected, the role of back-ground inflammatory cells in diagnosis has not beenemphasized. Our observations suggest that cases withPCP show no or minimal inflammation with predom-inance of AM with < 10% inflammatory cells. Casesshowing acute inflammatory cells—that is, predomi-nance of neutrophils—did not have PCP. Hence, pa-tients showing minimal inflammation on BALF cytol-ogy may benefit from a higher index of suspicion.Since the sensitivity of microscopy is much lower thanmolecular diagnostic approaches, samples negative bymicroscopy but with a mainly alveolar macrophagecount should be subjected to PCR amplification. A 2-step screening of BALF can be done. Routine cytol-ogy followed by PCR from slides can be used as a di-agnostic method for PCP. This can help in focusingdiagnostic facilities to patients most likely to havePCP. In countries where routine cytology is easilyavailable but molecular tests are not easily accessible,such an approach is of value. These samples, despitenot staining positive for P jirovecii, clearly have DNAtargets for PCR amplification in the material scrapedfrom the slides, although the organisms are not visible.This renders the slide material equivalent to DNA ex-traction from fresh BALF pellets and useful in archivalmaterial.

Of the 35 patients in the immunocompromisedgroup, 5 had PCP, 1 with concurrent cryptococcosis.Hence, patients suspected to have PCP on clinicalgrounds may have other infections that can also be simultaneously diagnosed on cytology. Among the 2patients with Cryptococcus infection, 1 had a predomi-nant AM count, while the other had mixed inflamma-tory cell types.

Thus, we conclude that an inflammatory cell popu-lation of BAL fluid is an important aspect of cytologicevaluation and can be used to target samples for PCRand special stain analysis. Use of PCR on BALF cytol-ogy slide scrapings for successful identification of Pjirovecii even on microscopy-negative samples has notbeen demonstrated before.

References

1. Stringer JR, Beard CB, Miller RF, Wakefield AE: A new name(Pneumocystis jirovecii) for Pneumocystis from humans. Emerg In-fect Dis 2002;8:891–896

2. Saito K, Nakayamada S, Nakano K, Tokunaga M, Tsujimura S,Nakatsuka K, Adachi T, Tanaka Y: Detection of Pneumocystiscarinii by DNA amplification in patients with connective tissuediseases: Re-evaluation of clinical features of P. carinii pneumo-

302

Gupta et al

ACTA CYTOLOGICA Volume 54 Number 3 May–June 2010

nia in rheumatic diseases. Rheumatology (Oxf) 2004;43:479–485

3. Wakefield AE, Pixley FJ, Banerji S, Sinclair K, Miller RF,Moxon ER, Hopkin JM: Detection of Pneumocystis carinii withDNA amplification. Lancet 1990;336:451–453

4. Rabodonirina M, Raffenot D, Cotte L, Boibieux A, MayençonM, Bayle G, Persat F, Rabatel F, Trepo C, Peyramond D, PiensM: Rapid detection of Pneumocystis carinii in bronchoalveolarlavage specimens from human immunodeficiency virus-infectedpatients: Use of a simple DNA extraction procedure and nestedPCR. J Clin Microbiol 1997;35:2748–2751

5. Dahiya S, Mathur SR, Iyer VK, Kapila K, Verma K: Diagnosisof Pneumocystis pneumonia by bronchoalveolar lavage cytology:Experience at a tertiary care centre in India. Indian J Chest DisAllied Sci 2005;47:259–265

6. Gupta R, Mirdha BR, Guleria R, Mohan A, Agarwal SK,Kumar L, Kabra SK, Samantaray JC: Improved detection ofPneumocystis jirovecii infection in a tertiary care reference hospi-tal in India. Scand J Infect Dis 2007;39:571–576

7. Smith RL, el-Sadr WM, Lewis ML: Correlation of bron-choalveolar lavage cell populations with clinical severity ofPneumocystis carinii pneumonia. Chest 1988;93:60–64

8. Venet A, Dennewald G, Sandron D, Stern M, Jauhert F, Lei-bowitch J: Broncolaveolar lavage in acquired immunodeficien-cy syndrome (lett). Lancet 1983;2:53

9. Tani EM, Schmitt FC, Oliveira ML, Gobetti SM, Decarls RM:Pulmonary cytology in tuberculosis. Acta Cytol 1987;31:460–463

10. Lee CH, Lu JJ, Bartlett MS, Durkin MM, Liu TH, Wang J,Jiang B, Smith JW: Nucleotide sequence variation in Pneumo-cystis carinii strains that infect humans. J Clin Microbiol 1993;31:754–757

11. Huang SN, Fischer SH, O’Shaughnessy E, Gill VJ, Masur H,Kovacs JA: Development of a PCR assay for diagnosis of Pneu-mocystis carinii pneumonia based on amplification of the multi-copy major surface glycoprotein gene family. Diagn MicrobiolInfect Dis 1999;35:27–32

12. Cushion MT, Beck M: Summary of Pneumocystis research pre-sented at the 7th International Workshop on OpportunisticProtists. J Eukaryot Microbiol (suppl) 2001;48:1015–1055

13. Ognibene FP, Shelhamer J, Gill V, Macher AM, Loew D,Parker MM, Gelmann E, Fauci AS, Parrillo JE, Masur H: Thediagnosis of Pneumocystis carinii pneumonia in patients with theacquired immunodeficiency syndrome using sub-segmentalbronchoalveolar lavage. Am Rev Respir Dis 1984;129:929–932

14. Thomas CF, Limper AH: Pneumocystis pneumonia. N Engl JMed 2004;350:2487–2498

15. Armbruster C, Pokiesser L, Hassl A: Diagnosis of Pneumocystispneumonia by BAL in AIDS patients: Comparison of Diff-Quik, fungi flour stain, direct IF test and PCR. Acta Cytol 1995;39:1089–1093

16. Leigh TR, Gazzard BG, Rowbottom A, Collins JV: Quantita-tive and qualitative comparison of DNA amplification by PCRwith immunofluorescence staining for diagnosis of Pneumocystis

carinii pneumonia. J Clin Pathol 1993;46:140–144

17. Cartwright CP, Nelson NA, Gill VJ: Development and evalua-tion of a rapid and simple procedure for detection of Pneumo-cystis carinii by PCR. J Clin Microbiol 1994;32:1634–1638

18. Leibovitz E, Pollack H, Moore T, Papellas J, Gallo L, Krasin-ski K, Borkowsky W: Comparison of PCR and standard cyto-logical staining for detection of Pneumocystis carinii from respi-ratory specimens from patients with or at high risk for infectionby human immunodeficiency virus. J Clin Microbiol 1995;33:3004–3007

19. Pinlaor S, Mootsikapun P, Pinlaor P, Phunmanee A, PipitgoolV, Sithithaworn P, Chumpia W, Sithithaworn J: PCR diagno-sis of Pneumocystis carinii on sputum and bronchoalveolar lavagesamples in immuno-compromised patients. Parasitol Res 2004;94:213–218

20. Sing A, Roggenkamp A, Autenrieth IB, Heesemann J: Pneumo-cystis carinii carriage in immunocompetent patients with pri-mary pulmonary disorders as detected by single or nested PCR.J Clin Microbiol 1999;37:3409–3410

21. Mei Q, Turner RE, Sorial V, Klivington D, Angus CW, KovacsJA: Characterization of major surface glycoprotein genes ofhuman Pneumocystis carinii and high level expression of a con-served region. Infect Immun 1998;66:4268–4273

22. Nevez G, Raccurt C, Jounieaux V, Dei-Cas E, Mazars E: Pneu-mocystosis versus pulmonary Pneumocystis carinii colonizationin HIV-negative and HIV-positive patients. AIDS 1999;13:535–536

23. Helweg-Larsen J, Jensen JS, Dohn B, Benfield TL, LundgrenB: Detection of Pneumocystis DNA in samples from patients sus-pected of bacterial pneumonia: A case-control study. BMC In-fect Dis 2002;2:28

24. Maskell NA, Waine DJ, Lindley A, Pepperell JC, WakefieldAE, Miller RF, Davies RJ: Asymptomatic carriage of Pneumo-cystis jirovecii in subjects undergoing bronchoscopy: A prospec-tive study. Thorax 2003;58:594–597

25. Flori P, Bellete B, Durand F, Raberin H, Cazorla C, Hafid J,Lucht F, Sung RT: Comparison between real-time PCR, con-ventional PCR and different staining techniques for diagnosingPneumocystis jirovecii pneumonia from bronchoalveolar lavagespecimens. J Med Microbiol 2004;53:603–607

26. Alvarez-Martínez MJ, Miró JM, Valls ME, Moreno A, RivasPV, Solé M, Benito N, Domingo P, Muñoz C, Rivera E, ZarHJ, Wissmann G, Diehl AR, Prolla JC, de Anta MT, Gatell JM,Wilson PE, Meshnick SR, Spanish PCP Working Group: Sen-sitivity and specificity of nested and real-time PCR for the de-tection of Pneumocystis jirovecii in clinical specimens. Diagn Mi-crobiol Infect Dis 2006;56:153–160

27. Jacobs JA, Dieleman MM, Cornelissen EI, Groen EA, Wage-naar SS, Drent M: Bronchoalveolar lavage fluid cytology in pa-tients with Pneumocystis carinii pneumonia. Acta Cytol 2001;45:317–326

28. Sadaghdar H, Huang ZB, Eden E: Correlation of bronchoalve-olar lavage findings to severity of Pneumocystis carinii pneumo-nia in AIDS: Evidence for the development of high-permeabilitypulmonary edema. Chest 1992;102:63–69