In order to confirm that re-intubation can be a risk factor of nosocomial pneumonia in mechanically ventilated patients, a case-control study was performed. Forty consecutive patients needing re-intubation were selected as cases. Each case was paired with a matched control for the previous duration of mechanical ventilation (+/- 2 d). Nineteen (47%) of the cases developed pneumonia after re-intubation compared with 4 (10%) of the controls (odds ratio [OR] = 8.5; 95% confidence interval [CI] 1.7 to 105.9; p = 0.0007). After adjusting for age, sex, and presence of prior bronchoscopy, the conditional logistic regression analysis demonstrated that re-intubation was the only significant factor related to the development of pneumonia (OR: 5.94; 95% CI 1.27 to 22.71; p = 0.023). Sixteen (73%) of the 22 patients lying semirecumbent during the interval between extubation and re-intubation developed nosocomial pneumonia versus three (16%) of the 18 in supine position (p = 0.001). These results indicate that semirecumbency during the period between extubation and re-intubation may play a role in nosocomial pneumonia development in patients who need re-intubation. Total intensive care unit stay (19.4 +/- 10 versus 13.9 +/- 11.9 days, p = 0.0008) and crude mortality (35 versus 20%, p = 0.14) were also higher in re-intubated patients when compared with controls. We conclude that re-intubation is a risk factor for ventilator-associated pneumonia and might be avoided in a substantial number of cases.