Risk factors for nosocomial pneumonia in the elderly

Pneumonia is now the second most frequent hospital-acquired infection in acute-care facilities, and recent studies indicate that the incidence rate for nosocomial pneumonia in long-term-care facilities is of similar magnitude. The mortality rate is high, especially in the elderly. With prevention of this complication as an overall priority, this study was designed to determine the risk factors associated with nosocomial pneumonia in the elderly in both acute-care and long-term-care settings.

An epidemiologic case-control study was undertaken to compare patients or residents who developed radiographically confirmed pneumonia with control subjects who did not have and did not develop respiratory infection. Thirty-three cases were identified in the acute-care setting during the 18-week period, and 27 cases were identified in the long-term-care setting. Two matched controls were chosen for each case. Data collection involved review of the medical record and verification by medical personnel while the cases or controls were still institutionalized. Risk factor variables were analyzed using an odds ratio and 95% confidence interval calculation for matched triplets, and chi-square analysis. Selected risk factors were entered into a backward stepwise logistic regression to determine the best combination of risk factors for each setting.

In the acute-care setting, current neurologic disease, current renal disease, deteriorating health, altered level of consciousness, disorientation, dependent bathing, dependent bowel function, dependent feeding, aspiration, difficulty with oropharyngeal secretions, and presence of a nasogastric tube were significant risk factors. In the long-term-care setting, deteriorating health, malnourishment, recent weight change, altered level of consciousness, disorientation, aspiration, difficulty with oropharyngeal secretions, suctioning, presence of a nasogastric or gastric tube, upper respiratory infection, inhalation therapy, increased confusion, and increased agitation were determined to be significant. Current pulmonary disease, previous infection, and antibiotic therapy were found not to be associated with the onset of nosocomial pneumonia. Through logistic regression, the best models for prediction of nosocomial pneumonia in the elderly were identified. In the acute-care setting, difficulty with oropharyngeal secretions and presence of a nasogastric tube were the best predictors. In the long-term-care setting, difficulty with oropharyngeal secretions, deteriorating health, and occurrence of an unusual event were the best combination of predictors.

These data confirm prior findi