Prevention of ventilator-associated pneumonia and ventilator-associated conditions: a randomized controlled trial with subglottic secretion suctioning.

Abstract

Ventilator-associated pneumonia diagnosis remains a debatable topic. New definitions of ventilator-associated conditions involving worsening oxygenation have been recently proposed to make surveillance of events possibly linked to ventilator-associated pneumonia as objective as possible. The objective of the study was to confirm the effect of subglottic secretion suctioning on ventilator-associated pneumonia prevalence and to assess its concomitant impact on ventilator-associated conditions and antibiotic use. Randomized controlled clinical trial conducted in five ICUs of the same hospital. Three hundred fifty-two adult patients intubated with a tracheal tube allowing subglottic secretion suctioning were randomly assigned to undergo suctioning (n = 170, group 1) or not (n = 182, group 2). During ventilation, microbiologically confirmed ventilator-associated pneumonia occurred in 15 patients (8.8%) of group 1 and 32 patients (17.6%) of group 2 (p = 0.018). In terms of ventilatory days, ventilator-associated pneumonia rates were 9.6 of 1,000 ventilatory days and 19.8 of 1,000 ventilatory days, respectively (p = 0.0076). Ventilator-associated condition prevalence was 21.8% in group 1 and 22.5% in group 2 (p = 0.84). Among the 47 patients with ventilator-associated pneumonia, 25 (58.2%) experienced a ventilator-associated condition. Neither length of ICU stay nor mortality differed between groups; only ventilator-associated condition was associated with increased mortality. The total number of antibiotic days was 1,696 in group 1, representing 61.6% of the 2,754 ICU days, and 1,965 in group 2, representing 68.5% of the 2,868 ICU days (p < 0.0001). Subglottic secretion suctioning resulted in a significant reduction of ventilator-associated pneumonia prevalence associated with a significant decrease in antibiotic use. By contrast, ventilator-associated condition occurrence did not differ between groups and appeared more related to other medical features than ventilator-associated pneumonia.