Background: Previous studies on predictors of deep sternal wound infection (DSWI) have identified either respiratory failure or tracheostomy as a risk factor for the occurrence of this complication. This study was conducted to analyze the interaction between these two variables. We hypothesize that respiratory failure and not tracheostomy per se is associated with an increased risk of DSWI.
Methods: We analyzed 2823 patients who underwent cardiac operations through median sternotomy between January 2002 and September 2006. Patients were divided into three groups: respiratory failure with or without tracheostomy (tracheostomy versus nontracheostomy) and patients without respiratory failure. The primary outcome measure was the incidence of DSWI in each group and its predictors.
Results: Postoperative respiratory failure was observed in 252 patients (9%): 144 without tracheostomy (57%) and 108 with tracheostomy (43%). The mean duration of intubation in nontracheostomy patients was 19 +/- 12 days. The mean duration to tracheostomy was 13 +/- 6 days. DSWI occurred in 38 patients (1.3%): patients with no respiratory failure, 1%; patients with respiratory failure, 5.1% (p < 0.001). The incidence of DSWI was similar between tracheostomy (4.6%) and nontracheostomy patients (5.6%, p = 0.5). The mean time to diagnosis of DSWI was 25 +/- 14 days and was similar for all groups. The mean number of days to tracheostomy was 12 +/- 3 days in DSWI patients and 13 +/- 6 in patients without DSWI (p = 0.7). In multivariate analysis, respiratory failure was the strongest predictor of DSWI (odds ratio, 5.2). Tracheostomy was not identified as a predictor of DSWI or hospital mortality.
Conclusions: The incidence of DSWI remains high in patients with respiratory failure. Tracheostomy is not a risk factor for DSWI and serves as a surrogate for respiratory failure. Therefore, considering that early tracheostomy may be beneficial in patients with respiratory insufficiency, a more liberal approach to early tracheostomy may be warranted.