Evaluation of continuous aspiration of subglottic secretion in an in vivo study.

OBJECTIVE Continuous aspiration of subglottic secretions (CASS) is believed to lower the incidence of ventilator-associated pneumonia. Animal studies to establish safety and efficacy of CASS have not been conducted. DESIGN Prospective randomized animal study. SETTING Animal-research facility at the U.S. National Institutes of Health. SUBJECTS Twenty-two sheep. INTERVENTIONS Sheep were randomized into three groups. In group C (control), eight sheep were kept prone, intubated with a standard endotracheal tube (ETT), and mechanically ventilated for 72 hrs with head and ETT elevated at an angle of 30 degrees. In group CASS-HU (CASS, head up), seven sheep were managed as group C and intubated with a Hi-Lo Evac, Mallinckrodt ETT (CASS suction kept at < or =20 mm Hg). In group CASS-HD (CASS, head down), seven sheep were kept prone with CASS, and the ETT and trachea were horizontal to promote spontaneous drainage of mucus from the ETT. MEASUREMENTS AND RESULTS The lower respiratory tract in the CASS-HU group was heavily colonized in all seven sheep (median 4.6 x 10(9), range, 1.5 x 10(8) to 7.9 x 10(9) colony-forming units/g), with a reduction of lung bacterial colonization compared with the C group (p = .05). In group CASS-HD, the lower respiratory tract was not colonized in six of seven sheep. One sheep showed low levels of bacterial growth (median, 0; range, 0-2.2 x 10(5)). At autopsy, in all 14 sheep with CASS, we found tracheal mucosal injury of different degrees of severity at the level of the suction port of the ETT. CONCLUSIONS In group CASS-HU, regardless of finding a marginal decrease of the bacterial colonization of the lower airways, there was pervasive trachea-bronchial-lung bacterial colonization. Second, there was minimal, or absent, bacterial colonization when the orientation of the CASS ETT was at, or just below, horizontal. Third, there was widespread injury to tracheal mucosa/submucosa from the use of CASS. Note that results of studies conducted in an animal model are always difficult to extrapolate to the clinical practice due to anatomical and functional differences.