Community-acquired aspiration pneumonia in intensive care units. Epidemiological and prognosis data.

Over a 9-yr period, among 505 patients exhibiting severe community-acquired pneumonia and admitted into a total of six medical ICUs in the north of France, we collected 116 patients (23%) meeting the usual criteria for aspiration pneumonia. Main medical grounds of ICU admission were respiratory distress in 54 patients and neurological disturbances in 62 patients. The main underlying risk factor for aspiration pneumonia was drug overdose (39%). Mechanical ventilation was required for 73 patients. Initial shock was present in 15 patients. Pulmonary involvement was bilateral in 27 patients. There were 94 aerobic organisms isolated from 70 patients (60%), the most frequent being gram-negative bacilli (n = 38), Staphyloccus spp. (n = 27) and Streptococcus pneumoniae (n = 22). Overall mortality was 22%, but only 11 (11%) deaths were directly or indirectly related to aspiration pneumonia. Stepwise multivariate analysis identified four independent predictors of mortality: ineffective initial antimicrobial therapy (p = 0.0001), positive initial blood culture (p = 0.0001), hospital-acquired lower respiratory tract superinfections (p = 0.0054), and use of inotropic support (p = 0.0078). The importance of prevention of hospital-acquired superinfections and permanent optimization of our antimicrobial strategies warranting efficacy of the initial antimicrobial therapy is underlined.