Resolution of ventilator-associated pneumonia: Prospective evaluation of the clinical pulmonary infection score as an early clinical predictor of outcome* Carlos M. Luna, MD; Daniel Blanzaco, MD; Michael S. Niederman, MD; Walter Matarucco, MD; Natalio C. Baredes, MD; Pablo Desmery, MD; Fernando Palizas, MD; Guillermo Menga, MD; Fernando Rios, MD; Carlos Apezteguia, MD V entilator-associated pneumonia (VAP) is the most frequent in- fectious complication among patients admitted to the inten- sive care unit (ICU), prolonging ICU length of stay and increasing the risk of death in critically ill patients (1). The so-called at- tributable mortality rate appears to vary with patient population and infecting or- ganism (2). Outcome studies suggest that several factors are related to mortality rate in VAP (3– 6). The clinical pulmonary in- fection score (CPIS) originally was de- scribed to facilitate the clinical diagnosis of VAP, by including several elements to quantify signs of pneumonia (7). The sen- sitivity of the CPIS to detect pneumonia confirmed by bronchoscopic methods is under discussion but was found to be 93% in the original description (7). Some au- thors suggested that CPIS could be used to evaluate the response to therapy or to select patients for whom a short course of antibi- otics could be appropriate (8, 9). One re- cently published paper by Dennesen et al. (10) did address the resolution of infectious variables: temperature, leukocyte count in peripheral blood, and PaO 2 /FIO 2 ratio (but not of CPIS score) in patients with VAP (10). However, we know very little about the time course of pneumonia resolution and about how specific clinical features of pneumonia resolve over time. Therefore, it was our hypothesis that serial measure- ments of the CPIS could be used to define the natural history of pneumonia resolu- tion as we suspected that many patients have clinical resolution sooner than the approximately 2-wk time period for which pneumonia is treated in many instances. Knowing more clearly the evolution of the CPIS, or some of its components, we could identify early in the hospital course of VAP *See also p. 969. From the Pulmonary and Critical Care Divisions (CML, DB, NB), Department of Medicine, Hospital de Clı´nicas “Jose ´ de San Martı´n,” Universidad de Buenos Aires; Pulmonary and Critical Care Division (MN), Winthrop University Hospital, Mineola, NY; Critical Care Service (WM), Sanatorio Otamendi- Miroli, Buenos Aires; Critical Care Service (PD), Sanatorio Mitre, Buenos Aires; Critical Care Service (FP), Clı´nica Bazterrica, Buenos Aires; Critical Care Service (GM), Hospital Marı´a Ferrer, Gobierno auto ´ nomo de la Ciudad de Buenos Aires; Critical Care Division (CA, FR), Policlı´nico Alejandro Posadas, Haedo, Provincia de Buenos Aires. Performed at the Intensive Care Units from the following hospitals: Hospital de Clı´nicas “Jose ´ de San Martı´n,” Sanatorio Otamendi-Miroli, Sanatorio Mitre, Clı´nica Bazterrica, Hospital Marı´a Ferrer (all located in the city of Buenos Aires); and Policlı´nico Alejandro Posadas, Haedo, Provincia de Buenos Aires. Address requests for reprints to: Carlos M. Luna, MD, Acevedo 1070, Banfield, 1828, Buenos Aires, Argentina. E-mail: cymluna@fmed.uba.ar Copyright © 2003 by Lippincott Williams & Wilkins DOI: 10.1097/01.CCM.0000055380.86458.1E Objectives: To prospectively evaluate the performance of the Clinical Pulmonary Infection Score (CPIS) and its components to identify early in the hospital course of ventilator-associated pneu- monia (VAP) which patients are responding to therapy. Design: Prospective, multicenter, in a cohort of mechanically ventilated patients. Setting: The intensive care unit of six hospitals located in the metropolitan area of Buenos Aires, Argentina. Patients: Sixty-three patients, from a cohort of 472 mechani- cally ventilated patients hospitalized for >72 hrs, had clinical evidence of VAP and bacteriologic confirmation by bronchoalveo- lar lavage (BAL) or blood cultures. Interventions: Bronchoscopy with BAL fluid culture and blood cultures after establishing a clinical diagnosis of VAP. All patients received antibiotics, 46 before bronchoscopy and 17 immediately after bronchoscopy. Measurements and Results: CPIS was measured at 3 days before VAP (VAP-3); at the onset of VAP (VAP); and at 3 (VAP3), 5 (VAP5), and 7 (VAP7) days after onset. CPIS rose from VAP-3 to VAP and then fell progressively in the population as a whole (p < .001), and the fall in CPIS was significant in 31 survivors, but not in 32 nonsurvivors. From the individual compo- nents of the CPIS, only the PaO 2 /FIO 2 ratio distinguished survivors from nonsurvivors, beginning at VAP3. When CPIS was <6 at 3 or 5 days after VAP onset, mortality was lower than in the remaining patients (p  .018). These differences also related to the finding that those receiving adequate therapy had a slight fall in CPIS and a significant increase of PaO 2 /FIO 2 at VAP3, whereas those getting inadequate therapy did not. Conclusions: Serial measurements of CPIS can define the clinical course of VAP resolution, identifying those with good outcome as early as day 3, and could possibly be of help to define strategies to shorten the duration of therapy. (Crit Care Med 2003; 31:676 –682) KEY WORDS: ventilator-associated pneumonia; intensive care; infection 676 Crit Care Med 2003 Vol. 31, No. 3