Respiratory failure (RF) following coronary artery bypass grafting (CABG) has been linked to increased early morbidity and mortality. We sought to determine risk factors for RF and to compare early and long-term outcome.
We studied 3760 consecutive patients who underwent CABG between 1992 and 2002. Patients without RF were compared with those who developed RF postoperatively (intubation and ventilation for a period of 72 hours or more). Long-term survival data (mean follow-up 5.2 years) were obtained from the National Death Index. Groups were compared by Cox proportional hazard models and Kaplan-Meier survival plots. The propensity for RF was determined by logistic regression analysis and each patient with RF was then matched with 5 patients without RF.
One hundred and sixty-two patients (4.3%) developed RF. The independent predictors for RF were increased age, low ejection fraction, shock, congestive heart failure, peripheral vascular disease, chronic obstructive pulmonary disease, smoking and other major complications: stroke, deep sternal wound infection, sepsis, bleeding, renal failure and gastrointestinal complications. After adjustment for all pre-, intra- and post-operative factors the adjusted hazard ratio of long-term mortality for patients with RF was 2.07 (95% confidence interval 1.56-2.75; P<0.001). One hundred and eighteen patients with RF were matched with 590 patients without RF using propensity scores identical to within 1%. Early outcome of matched groups is shown in table and Kaplan-Meier curves are shown in figure. Freedom from all-cause mortality at 5 years was 70.1±2.0% in patients without RF versus 51.7±5.1% in patients with RF (P<0.0001).
We identified preoperative and postoperative risk factors for RF following CABG. Patients with RF had significantly higher early and late mortality when compared with patients without RF.
RF had increased incidence in CABG patients with impaired cardiac or pulmonary function, increased age and major postoperative complications; its effect on patient survival extends far beyond the 30-day and in-hospital mortality time periods. These data suggest the need for a more frequent follow-up among patients with RF. VariableWithout RF (n=590)With RF (n=118)P valueEuroSCORE8.0±3.98.1±3.60.857Length of stay (days)12.8±15.428.2±31.2<0.00130-day mortality, n (%)39 (6.6)20 (16.9)0.001In-hospital mortality, n (%)35 (5.9)25 (21.2)<0.001Intraoperative stroke, n (%)36 (6.1)7 (5.9)0.999Stroke over 24 hours, n (%)17 (2.9)4 (3.4)0.766Postoperative myocardial infarction, n (%)4 (0.7)1 (0.8)0.999Deep sternal wound infection, n (%)10 (1.7)5 (4.2)0.088Bleeding/reoperation, n (%)40 (6.8)9 (7.6)0.694Gastrointestinal complications, n (%)26 (4.4)4 (3.4)0.804Renal failure/dialysis, n (%)14 (2.4)4 (3.4)0.521Sepsis/endocarditis, n (%)13 (2.2)3 (2.5)0.739
I.K. Toumpoulis, None.