OBJECTIVE: To determine the correlation of Gram staining with the culture of bronchoalveolar lavage in patients with clinical features of ventilator-associated pneumonia from two adult trauma and surgical intensive care units. METHODS: We conducted a retrospective study of 252 samples of bronchoalveolar lavage from patients with clinical ventilator-associated pneumonia in trauma or surgical postoperative care. Gram staining samples were classified as Gram-positive cocci and Gram-negative bacilli, all other results being excluded. Cultures from bronchoalveolar lavage were compared with Gram staining results. RESULTS: The correlation between Gram staining and culture from the bronchoalveolar lavage showed a kappa index of 0.27. The sensitivity of Gram staining was 53.9% and the specificity, 80.6%. Considering the identification of Gram-positive cocci against other results (negative and Gram-negative bacilli), the negative predictive value was 94.8%. The evaluation of Gram-negative bacilli against other results (negative and Gram-positive cocci) rendered a sensitivity of 27.1% and a specificity of 95.4%. CONCLUSION: It appears that the negative predictive value for Gram-positive cocci is acceptable, but the sensitivity of Gram staining in the etiology of ventilator-associated pneumonia was not able to identify the microorganism before culture.
OBJECTIVE: To determine the correlation of Gram staining with the culture of bronchoalveolar lavage in patients with clinical features of ventilator-associated pneumonia from two adult trauma and surgical intensive care units. METHODS: We conducted a retrospective study of 252 samples of bronchoalveolar lavage from patients with clinical ventilator-associated pneumonia in trauma or surgical postoperative care. Gram staining samples were classified as Gram-positive cocci and Gram-negative bacilli, all other results being excluded. Cultures from bronchoalveolar lavage were compared with Gram staining results. RESULTS: The correlation between Gram staining and culture from the bronchoalveolar lavage showed a kappa index of 0.27. The sensitivity of Gram staining was 53.9% and the specificity, 80.6%. Considering the identification of Gram-positive cocci against other results (negative and Gram-negative bacilli), the negative predictive value was 94.8%. The evaluation of Gram-negative bacilli against other results (negative and Gram-positive cocci) rendered a sensitivity of 27.1% and a specificity of 95.4%. CONCLUSION: It appears that the negative predictive value for Gram-positive cocci is acceptable, but the sensitivity of Gram staining in the etiology of ventilator-associated pneumonia was not able to identify the microorganism before culture.