Pneumocystis carinii in bronchoalveolar lavage fluid and bronchial washings

BMJ 1994; 308 doi: (Published 07 May 1994) Cite this as: BMJ 1994;308:1206
  1. S P Matusiewicz,
  2. R J Fergusson,
  3. A P Greening,
  4. G K Crompton,
  5. S M Burns
  1. Respiratory Unit, Western General Hospital, Edinburgh EH4 2XU
  2. Regional Virus Laboratory, City Hospital, Edinburgh EH10 5SB
  1. Correspondence to: Dr Matusiewicz
  • Accepted 17 December 1993

Antibodies to Pneumocystis carinii develop in early childhood,1 although no associated illness has been identified. The classic concept in that P carinii pneumonia in immunocompromised patients represents a reactivation of dormant childhood infection. Studies with DNA amplification, however, have failed to detect pneumocystis in sputum2 or specimens of lung tissue taken at necropsy3 from immunocompetent patients. Reinfection rather than reactivation may therefore be more relevant. To see whether P carinii exists as a commensal organism of the lower respiratory tract we examined bronchoalveolar lavage fluid and bronchial washings in apparently immunocompetent patients undergoing routine diagnostic (or research) bronchoscopy by using an immunofluorescent antibody test.

Subjects, methods, and results

We studied 220 patients with various clinical diagnoses (table). Patients were considered to have no active disease if investigations, including bronchoscopy, failed to show any lung disease. Patients with known HIV infection or suspected pneumocystis pneumonia or those taking immunosuppressive drugs were excluded. In all, 116 patients had bronchoalveolar lavage, specimens being obtained in aliquots of 10 ml from 240 ml lavages (8x30 ml) of the lingula or middle lobe; 104 patients had bronchial washing, saline being introduced into and aspirated from the lung with disease diagnosed radiologically or seen to be abnormal at bronchoscopy. All specimens were examined for pneumocystis by a virologist (SB) without access to clinical data.

Details of 220 patients examined for presence of Pneumocystis carinii in bronchial washing or bronchoalveolar lavage specimens

View this table:

Cytospin preparations of samples were fixed (50% methanol-acetone) and stained with murine monoclonal antibodies labelled with fluoresce in that react with cysts and trophozoites (Genetic Systems Corporation, SYVA). Samples were considered positive if cysts or trophozoites were seen in three or more high power fields.

Of 220 specimens examined, only five (three lavage specimens and two washings) were positive for P carinii by the fluorescent antibody test. The five positive samples all had cysts and trophozoites in more than three high power fields. All were from patients with small cell lung cancer. No cysts or trophozoites were seen in any fields of samples defined as negative. No patients were thought to have clinical or radiological evidence of P carinii pneumonia at the time of bronchoscopy. Two positive specimens were available for staining by an alternative less sensitive method (toluidine blue O), and both were positive. Four of the five patients were clinically well. Two patients had evidence of macroscopic extrahemithoracic disease at presentation. Two patients had additional causes for impaired immunity - namely, previous cytotoxic treatment and ectopic secretion of andrenocorticotrophic hormone.


We were unable to detect P carinii by fluorescent antibody testing of bronchoalveolar lavage fluid or bronchial washings in 215 of the 220 patients studied and therefore consider that it may not be a commensal organism of the lower respiratory tract. Although the antibody test requires interpretation by a skilled pathologist, it is more sensitive and specific than silver staining. Theoretically, DNA amplification techniques are able to detect one or two cells but have failed to detect organisms in sputum2 or postmortem samples of lung3 in normal subjects. Not all studies, however, have shown any increase in diagnostic sensitivity of DNA amplification over fluorescent antibody testing in bronchoalveolar lavage samples.4 Our findings support the concept that reinfection rather than reactivation leads to disease in susceptible patients. No environmental source has been identified, although infection rates show a seasonal variation. Infection was first identified in debilitated young and elderly subjects. These groups, in addition to patients infected with HIV, may act as a reservoir for pneumocystis.

Subclinical P carinii infection was found in five out of 27 patients with histologically proved small cell lung cancer. This may reflect a local defect in immune function associated with small cell lung cancer, which may have prognostic implications. Previous colonisation with pneumocystis may be clinically relevant for patients who subsequently receive cancer chemotherapy as pneumocystis pneumonia is a recognised and perhaps underreported terminal event. The importance of subclinical colonisation of the lower respiratory tract needs to be studied.

We thank Dr Margaret McIntyre and staff of the Department of Pathology, Western General Hospital, Edinburgh.


View Abstract