* Context.-Induced sputum sampling has an approximate 70% sensitivity for detection of Pneumocystis jiroveci in human immunodeficiency virus (HIV) patients. Bronchoalveolar lavage sampling has greater than 90% sensitivity but is a far more invasive procedure. Therefore, bronchoalveolar lavage testing is often recommended as a follow-up after a negative induced sputum. In HIV-negative patients, the utility of induced sputum testing is still not well defined.
Objective.-To determine whether repeat induced sputum sampling increases diagnostic yield and might thereby reduce the need for follow-up bronchoalveolar lavage sampling. To determine the utility of induced sputum sampling in HIV-negative patients.
Design.-A 2-year retrospective review of the utility of repeat induced sputa testing in patients with previous first and/or second negative induced sputa. Retrospective review of induced sputa detection in HIV-negative patients.
Results.-Repeat testing of induced sputa for Pneumocystic jirovecii did not significantly increase diagnostic yield. Furthermore, in HIV-negative patients, induced sputum testing was diagnostically insensitive.
Conclusions.-Bronchoalveolar lavage testing should be performed initially in HIV-negative patients and after a first negative induced sputum in HIV-positive patients.
(Arch Pathol Lab Med. 2007;131:1582-1584)
Pneumocystis jiroveci (formerly Pneumocystic carinii) causes a severe pneumonia in immunocompromised patients. Direct fluorescent antibody testing is a sensitive and standard method for detection of P jiroveci in clinical samples.1 However, optimal sensitivity requires adequate sampling of lower respiratory secretions. In prior studies in human immunodeficiency virus (HIV)-positive patients, induced sputum and bronchoalveolar lavage (BAL) sampling generally yield sensitivities of roughly 70% and greater than 90%, respectively.2,3 Because induced sputum collection is less invasive, induced sputum has often been favored as the initial specimen in HIV-positive patients.1
In a retrospective review of data from a large tertiary care teaching hospital, 2 issues were addressed. The first was whether repeat induced sputum collection increases diagnostic yield and thereby decreases the need for follow- up BAL procedures in HIV-positive patients. The second was the utility of induced sputum collection in HIVnegative patients, a population in which organism burden is lower4 and in which a diagnostic algorithm optimized for HIV-positive patients might not be similarly applicable.
METHODS
Sample Processing
Induced sputa were collected by respiratory therapists after inhalation of nebulized 3% NaCl. Specimens were deemed insufficient that (1) did not appear to be induced as judged by the microbiology technologist; (2) were of poor quality, that is, did not have any cellular material; or (3) were less than 2 mL. Insuf- ficient specimens were excluded from our analysis. Bronchoalveolar lavage samples were similarly assessed for sufficiency by criteria 2 and 3. For each induced sputum or BAL sample, 2 cytospin preparations were made, stained with the MONOFLUO Pneumocystis carinii IFA Test Kit (Bio-Rad, Hercules, Calif) according to the manufacturer’s instructions, and screened for infection.
Data Acquisition
Electronic medical records were reviewed to ascertain relevant laboratory data and clinical history. This study was approved by our institutional review board and conducted according to institutional guidelines.
RESULTS
Diagnostic Yield of Multiple Induced Sputa
Pneumocystis testing results from samples submitted to our microbiology laboratory during a 2-year period from 2004 to 2005 were reviewed. During this time, 687 patients submitted 666 induced sputum specimens and 393 BAL specimens. Among these patients, 615 had known HIV status.
Data are summarized in the Table. Notably, testing of the first induced sputum in HIV-positive patients demonstrated a positivity rate of 11.8%. In contrast, testing of a second induced sputum after a first negative induced sputum yielded only a single additional positive result or a 3.3% yield, whereas testing a third induced sputum after 2 negative induced sputa yielded no additional positives. This contrasts with a 23.8% yield obtained with BAL after a negative induced sputum or a 7.2-fold relative increase in detection. Therefore, repeat induced sputum testing does not significantly increase diagnostic yield.
In contrast to initial high sensitivity for HIV-positive patients, only a single HIV-negative patient of 211 was diagnosed by induced sputum testing. Furthermore, no additional cases were detected after second or third induced sputum. However, testing of BAL fluid after a negative induced sputum led to a 7.4% diagnostic yield or a 13- fold relative increase in detection. Interestingly, those tested initially by BAL sampling had a lower positivity rate of 1.5%. Therefore, in the HIV-negative patient population, clinicians preferentially chose a much less sensitive method for initial testing, that is, induced sputum sampling, in those at greatest risk for Pneumocystis pneumonia.
Information provided by: Findarticles.com










