Protected specimen brush or bronchoalveolar lavage to diagnose bacterial nosocomial pneumonia in ventilated adults: a meta-analysis

Crit Care Med. 1999 Nov;27(11):2548-60. doi: 10.1097/00003246-199911000-00037.


Objective: We conducted a meta-analysis by using summary receiver operating characteristic curves to compare the diagnostic value for bacterial nosocomial pneumonia of the following: a) quantitative culture (colony-forming units per milliliter or CFU/mL) of respiratory secretions collected with a bronchoscopic protected specimen brush (PSB); b) quantitative culture of a bronchoscopic bronchoalveolar lavage (BAL); and c) the percentage of infected cells (IC) in BAL.

Data sources: All studies published in the English or the French language, through January 1, 1995, on the evaluation of PSB or BAL for the diagnosis of pneumonia were considered for analysis. The relevant literature was identified through computer and reference searching and by experts in the field.

Study selection: A study was included if at least two of three independent readers regarded its purpose as the evaluation of CFU-PSB, CFU-BAL, or IC-BAL for the diagnosis in human beings of bacterial nosocomial pneumonia in ventilated adults and if the study was prospective and published in a peer-reviewed journal.

Data extraction: Three readers reviewed all published articles and decided whether to include each study; consensus was defined as agreement by at least two readers. The authors of each original article included in the meta-analysis were asked to complete a questionnaire in which they were asked to check and to correct the data extracted by one of the independent readers.

Data synthesis: Summary receiver operating characteristic curves were used to compare the efficacy of three diagnostic tests. Eighteen studies on CFU-PSB (795 patients) were included, as well as 11 studies on CFU-BAL (435 patients) and 11 on IC-BAL (766 patients). The accuracy of these tests was not different. However, it seems that administration of previous antibiotics markedly decreased accuracy of CFU-PSB (p = .0002) but not the accuracy of CFU-BAL and that of IC-BAL.

Conclusion: Both PSB and BAL are reliable to diagnose bacterial nosocomial pneumonia. Because CFU-BAL and IC-BAL seemed more resistant to the effects of antibiotics, we recommend BAL rather than PSB if the patient is already receiving antibiotics.

Publication types

  • Comparative Study
  • Meta-Analysis
  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Anti-Bacterial Agents / therapeutic use
  • Bacteria / isolation & purification
  • Bronchoalveolar Lavage Fluid / cytology
  • Bronchoalveolar Lavage Fluid / microbiology
  • Bronchoalveolar Lavage* / standards
  • Cross Infection / diagnosis*
  • Cross Infection / drug therapy
  • Humans
  • Intensive Care Units
  • Lung / microbiology
  • Lung / pathology
  • Middle Aged
  • Mucus / cytology
  • Mucus / microbiology
  • Pneumonia, Bacterial / diagnosis*
  • Pneumonia, Bacterial / drug therapy
  • Prospective Studies
  • ROC Curve
  • Reproducibility of Results
  • Respiration, Artificial*
  • Specimen Handling / instrumentation*
  • Specimen Handling / standards
  • Surveys and Questionnaires


  • Anti-Bacterial Agents