To compare percutaneous with surgical tracheostomy using a meta-analysis of studies published from 1960 to 1996.
Publications obtained through a MEDLINE data-base search with a Boolean combination (tracheostomy or tracheotomy) and complications, with constraints for human studies and English language.
Publications addressing all peri- and postoperative complications. Studies limited to specific tracheostomy complications or containing insufficient details were excluded. Two authors independently selected the publications.
A list of relevant surgical variables and complications was compiled. Complications were divided into peri- and postoperative groups and further subclassified into severe, intermediate, and minor groups. Because most studies of percutaneous tracheostomy were published after 1985, surgical tracheostomy studies were divided into two periods: 1960 to 1984 and 1985 to 1996. The articles were analyzed independently by three investigators, and rare discrepancies were resolved through discussion and data reexamination.
Earlier surgical tracheostomy studies (n = 17; patients, 4185) have the highest rates of both peri- (8.5%) and postoperative (33%) complications. Comparison of recent surgical (n = 21; patients, 3512) and percutaneous (n = 27; patients, 1817) tracheostomy trials shows that perioperative complications are more frequent with the percutaneous technique (10% vs. 3%), whereas postoperative complications occur more often with surgical tracheotomy (10% vs. 7%). The bulk of the differences is in minor complications, except perioperative death (0.44% vs. 0.03%) and serious cardiorespiratory events (0.33% vs. 0.06%), which were higher with the percutaneous technique. Heterogeneity analysis of complication rates shows higher heterogeneity in older and surgical trials.
Percutaneous tracheostomy is associated with a higher prevalence of perioperative complications and, especially, perioperative deaths and cardiorespiratory arrests. Postoperative complication rates are higher with surgical tracheostomy. (Crit Care Med 1999; 27:1617-1625)