Multifactorial analyses in the diagnosis of pneumonia arising in the surgical intensive care unit

Abstract

The diagnosis of ventilator-associated pneumonia in the surgical intensive care unit continues to be problematic. The majority of intensive care units use clinical criteria based on chest x-ray; fever; leukocytosis; alterations in the pulse oximeter observations; the need to alter modes and amounts of ventilatory support; and more specific microbiologic studies, such as appropriate sputum, Gram stain, and culture to identify pneumonia. Diagnosing pneumonia based on clinical criteria alone is often difficult and inaccurate, which may lead to inappropriate use and choice of antibiotics. Invasive diagnostic techniques, such as protected specimen brush and bronchoalveolar lavage, provide an important microbiologic diagnosis. However, the cost and inconvenience limit broad usage. Furthermore, those results that return positive are often too late to dictate the need for, or direction of, therapy. Our use of a “pneumonia grid” may help identify patients likely to have a poor outcome. Until a readily available and cost-effective diagnostic study for pneumonia is developed, clinical criteria remain vital in routine practice.