PURPOSE: To investigate whether tracheostomy increases the risk of sternal wound infection (SWI) post cardiac surgery. METHODS: All patients undergoing cardiac surgery via median sternotomy from September 1997 to October 2010 were included in this retrospective observational study. Primary exposure was tracheostomy performed during admission to the cardiac surgical intensive care unit. The primary outcome was SWI during hospital admission. Multivariable logistic regression was used to determine if tracheostomy was an independent predictor of SWI. Restriction and propensity score analyses were then used to assess if tracheostomy is a causal risk factor for SWI. RESULTS: Four hundred and eleven of 18,845 patients (2.2%) were treated with tracheostomy. Incidences of SWI in tracheostomy and non-tracheostomy groups were 19.5% (80/411) and 0.8% (154/18,434), respectively. Using multivariable logistic regression analysis, tracheostomy was found to be an independent predictor of SWI (odds ratio [OR] 2.8; 95% confidence interval [CI] 1.9 to 4.2). In an analysis restricted to respiratory failure patients, tracheostomy was associated with sternal wound infection (OR 3.4; 95% CI 2.4 to 4.9). When the analysis was stratified by the risk of receiving tracheostomy as represented by propensity score (PS), 46 patients (12%) in the intermediate risk category (PS 0.2-0.4) had SWIs (adjusted OR 2.97; 95% CI 1.6 to 5.6), and 52 patients (14%) in the highest risk category (PS > 0.4) had SWIs (OR 1.52; 95% CI 0.85 to 2.87). DISCUSSION: Our single-centre observational study of cardiac surgery patients found tracheostomy to be an independent risk factor for SWI. Our analysis showed a robust association when restricted to patients with respiratory failure and after the population was stratified by the propensity to have a tracheostomy.
PURPOSE: To investigate whether tracheostomy increases the risk of sternal wound infection (SWI) post cardiac surgery. METHODS: All patients undergoing cardiac surgery via median sternotomy from September 1997 to October 2010 were included in this retrospective observational study. Primary exposure was tracheostomy performed during admission to the cardiac surgical intensive care unit. The primary outcome was SWI during hospital admission. Multivariable logistic regression was used to determine if tracheostomy was an independent predictor of SWI. Restriction and propensity score analyses were then used to assess if tracheostomy is a causal risk factor for SWI. RESULTS: Four hundred and eleven of 18,845 patients (2.2%) were treated with tracheostomy. Incidences of SWI in tracheostomy and non-tracheostomy groups were 19.5% (80/411) and 0.8% (154/18,434), respectively. Using multivariable logistic regression analysis, tracheostomy was found to be an independent predictor of SWI (odds ratio [OR] 2.8; 95% confidence interval [CI] 1.9 to 4.2). In an analysis restricted to respiratory failurepatients, tracheostomy was associated with sternal wound infection (OR 3.4; 95% CI 2.4 to 4.9). When the analysis was stratified by the risk of receiving tracheostomy as represented by propensity score (PS), 46 patients (12%) in the intermediate risk category (PS 0.2-0.4) had SWIs (adjusted OR 2.97; 95% CI 1.6 to 5.6), and 52 patients (14%) in the highest risk category (PS > 0.4) had SWIs (OR 1.52; 95% CI 0.85 to 2.87). DISCUSSION: Our single-centre observational study of cardiac surgery patients found tracheostomy to be an independent risk factor for SWI. Our analysis showed a robust association when restricted to patients with respiratory failure and after the population was stratified by the propensity to have a tracheostomy.