Severe pneumonia: BAL fluid colony count helped diagnose.

Clinical bottom line (level 1b)

  1. In patients with suspected bacterial pneumonia, BAL fluid with 10 ³ cfu/ mL colony count or more helped to rule pneumonia in, and BAL fluid less than 10 ³ , helps to rule it out.
Cantral et al: American Journal of Medicine 1993; 95: 601-607
Expires March 2003

The study

Setting: university hospital, USA

225 patients (aged ?, 57% male) undergoing fibreoptic bronchoscopy with bronchoalveolar lavage between February 1989 and September 1989
Patients undergoing bronchoscopy and BAL were usually sedated with a combination of intravenous meperidine and midazolam, with the dose individualised according to need. Intravenous atropine (0.6 mg) was also administered. In intubated patients, topical anaesthesia was obtained with preservative-free lidocaine.
Independent blinded reference standard, applied in all patients from a consecutive appropriate spectrum.
Reference standard:
  • Clinical, radiologic, laboratory and histologic data. Pneumonia was defined as the presence of new or worsening pulmonary infiltrates seen on chest radiography, with or without fever or leukocytosis, and with lack of an alternative diagnosis.
Diagnostic test: Fibreoptic bronchoscopy performed using an Olympus P-10 or equivalent bronchoscope. BAL was performed with sequential instillation and immediate aspiration of five 20 mL aliquots followed by one 10 mL aliquot of sterile normal saline in each of three sites of the bronchial tree. BAL specimens were cultured onto plates containing blood agar, MacConkey's agar and chocolate agar. Plates were examined at 24 and 48 hours.

The evidence


diagnostic test number of patients sensitivity for
pneumonia
(95% CI)		 specificity for
pneumonia
(95% CI)		 LR+ LR-
BAL fluid with 10³ cfu/mL colony count 90.0%
(% to %) 		97.0%
(% to %)
BAL fluid with 10^4 cfu/ mL 85.0%
(% to %) 		98.0%
(% to %)
BAL fluid with 10^5 cfu/ mL 55.0%
(% to %) 		100%
(% to %)
total

Comments

  1. Applicable in limited settings (intensive care)
  2. Antibiotics at the time of BAL reduces the sensitivity of the test, and prolonged intubation reduces the specificity of the test.
  3. Likelihood ratios could not be calculated as numbers were described as the number of times BAL was performed, not the number of patients.

Citation

  1. Cantral DE, Tape TG, Reed EC, et al: Quantitative culture of bronchoalveolar lavage fluid for the diagnosis of bacterial pneumonia. American Journal of Medicine 1993; 95: 601-607
Search Terms: pneumonia and diagnosis in Medline
Contributor: Clare Wotton and Musab Hayatli, November 1999
Reviewer: Mitsuhiro Kamei

Clinical Question.
Patient undergoing fibreoptic bronchoscopy and bronchoalveolar lavage (BAL)
Intervention or Exposure quantitative bacterial culture of BAL
Comparison clinical, radiographic, laboratory and histologic data
Outcome diagnosis of pneumonia