Accuracy of delayed (24 hours) processing of bronchoalveolar lavage for diagnosing bacterial pneumonia

Pneumonia in the intensive care unit is associated with a high mortality rate. Diagnostic accuracy is mandatory to improve prognosis. However, in many hospitals, samples from the respiratory tract cannot be immediately processed bacteriologically around the clock. This may complicate therapeutic choice based on invasive diagnostic procedures. We evaluated the effect of storing bronchoalveolar lavage fluid at 4 degrees C for 24 hrs on direct examination and culturing for diagnosing pneumonia.

Prospective, paired comparison study.

Intensive care unit in a university hospital.

A total of 93 bronchoalveolar lavages were performed on 66 intensive care unit patients who were suspected to have bacterial pneumonia.

Each sample was divided into two; one half was processed immediately (H0), and the other was processed after refrigeration at 4 degrees C for 24 hrs (H24).

All negative H0 culture samples (n = 31) were also negative for pathogens in H24 samples. Sixty two bronchoalveolar lavage cultures yielded one or more microorganisms, giving a total of 113 microorganisms in one or both samples. The results of positive cultures at H0 and H24 for the culturing diagnostic threshold of 10 colony forming units/mL agreed well (Kappa coefficient, 0.84); agreement was even better (Kappa coefficient, 0.85) when possible contaminants were excluded. The bias calculated as the mean difference between paired culture results was 0.195 +/- 1.31 (Delta log). When considering the accepted threshold of 10 colony forming units/mL, specificity at H24 compared to H0 was excellent (100%), but sensitivity was slightly lower (80%).

Delayed processing of bronchoalveolar lavage sampling is an acceptable alternative when immediate culturing cannot be performed because it enables antibiotic administration.