Antibiotic Use and the Risk of Pneumonia: 20 Years of Studies, but Where Are We Now?

Since 1986, when Craven et al. [1] published their pioneering multivariate analysis of risk factors for developing ventilator associated pneumonia (VAP), many studies of the issue have been conducted. Some of these have been multi-center studies, and the results have identified different risk factors. However, the greatest confounding factor is that variables reported to be particularly influential by some researchers have been found to be negligible by others or, in some cases, even to protect against disease. After 20 years of study, it is time to assess the current situation. Key areas in VAP research are the effect of the selection of the variables to be explored in the database, the case-mix of the population, and the relationship between antibiotic use and VAP over time. Clinical researchers cannot study all potential risk factors, because a database designed for this purpose would be extremely difficult to analyze. When researchers design a prospective study, variables to be recorded and analyzed should be selected on the basis of clinical experience or on the basis of a review of prior literature. Two studies provide good examples: Kollef et al. [2] reported that transfer from the intensive care unit (ICU) during mechanical ventilation was a risk factor for VAP, and a study by Valles et al. [3] demonstrated that continuous aspiration of subglottic secretions (CASS) was an independent factor that decreased the rate of VAP. However, the study by Kollef et al. [2] did not evaluate the effect of CASS. In turn, study by Valles et al. [3] was not able to explore the effect of transportation out of the ICU, because this variable was not included in the database. In addition, the capacity of retrospective studies to find an association between variables is limited because the database is already closed. One of the limitations of most clinical investigations is that their results cannot be generalized. The first issue to be analyzed is the use of different prophylactic strategies to prevent nosocomial infection. Two recent studies emphasized that the degree of adherence to evidence-based prophylactic measures against VAP may vary widely among health care workers [4, 5]. For instance, limitation of stress-ulcer prophylaxis to high-risk patients—a grade B recommendation—was followed by 66.7% of health care workers, whereas selective digestive decontamination—a grade A recommendation—was applied by only 15.5% of experts. This low level of adherence was independent of the degree of evidence. Barriers to implementation of …