A 5-YEAR STUDY OF SEVERE COMMUNITY-ACQUIRED PNEUMONIA WITH EMPHASIS ON PROGNOSIS IN PATIENTS ADMITTED TO AN INTENSIVE-CARE UNIT

Objectives: To characterize the epidemiology and to determine the prognosis factors in severe community-acquired pneumonia among patients admitted to an intensive care unit. Design: Retrospective clinical study. Setting: Intensive Care and Infectious Diseases Unit of a municipal general hospital of Lille University Medical School. Patients: 299 consecutive patients exhibiting severe community-acquired pneumonia. Measurements and results: On admission to ICU, 149 patients required mechanical ventilation for acute respiratory failure and 44 exhibited septic shock. Pulmonary involvement was bilateral in 71 patients. There were 260 organisms isolated from 197 patients (65.9%), the most frequent being Streptococcus pneumoniae (n = 80), Staphylococcus spp. (n = 57) and Gram-negative bacilli (n = 81). Overall mortality was 28.5% (85 patients). According to univariate analysis, mortality was associated with age over 60 years, anticipated death within 5 years, immunosuppression, shock, mechanical ventilation, bilateral pulmonary involvement, bacteremia, neutrophil count < 3500/mm3, total serum protein level < 45 g/1, serum creatinine > 15 mg/1, non-aspiration pneumonia, ineffective initial therapy and complications. Multivariate analysis selected only 5 factors significantly associated with prognosis: anticipated death within 5 years, shock, bacteremia, non-pneumonia-related complications and ineffective initial therapy. Conclusion: The effectiveness of the initial therapy appears to be the most significant prognosis factor and, as the one and only related to the initial medical intervention, suggests a need for permanent optimization of our antimicrobial strategies.