The purpose of this study was to determine which pre and/or peri-operative factors predicted the development of pneumonia following surgery with cardiopulmonary bypass for coronary artery bypass grafting with or without valvular or other concomitant cardiothoracic surgery.
Data from July 1996 until December 2002 were retrieved from a single University affiliated, tertiary teaching hospital database. Pneumonia was diagnosed by one of the following: positive cultures of sputum, blood, pleural fluid, empyema fluid, transtracheal fluid or transthoracic fluid; consistent with clinical findings that included chest x-ray evidence of pulmonary infiltrate, elevated temperature, new productive cough or purulent sputum. Predictor selection was guided by previous studies examining pulmonary complications following cardiac surgery. The effect of 16 preoperative and 10 peri-operative variables on pneumonia was examined using direct entry multivariate logistic regression.
Of 5364 cases, 342 (6.4%) patients were diagnosed with pneumonia. Patients were on average 63.5 + 12.6 years of age, with a body mass index of 26.9 + 4.5 kg/m2 and 27.8% (n = 1491) of the sample were female. Significant predictors of pneumonia from the preoperative context were gender (OR 1.3, 95%CI:1.014-1.653, p<.05), New York Heart Association Class III (OR 0.753, 95%CI:0.572-0.992, p<.05) and emergency surgical status (OR 0.645, 95%CI:0.432-0.962, p<.05). From the peri-operative context perfusion time greater than 120 minutes (OR 0.606, 95%CI:0.473-0.777, p<.001), the need for mechanical support (OR 0.380, 95%CI:0.268-0.539, p<.001), peri-operative myocardial infarction (OR 0.302, 95%CI:0.110-0.831, p<.05) and peri-operative stroke (OR 0.281, 95%CI:0.171-0.461, p<.001) predicted postoperative pneumonia. While this model correctly classified 93.6% of cases it only explained 4.9% (R2) of the variation in pneumonia outcome.
Pre and peri-operative factors make a minimal contribution to predicting the risk of postoperative pneumonia in cardiac surgical candidates.
Patient and process factors from the postoperative context require exploration and inclusion in multivariate models in order to develop reliable risk prediction strategies for pneumonia.
Rochelle Wynne, None.