Preoperative renal function and surgical outcomes in patients with acute type A aortic dissection†

There are few data concerning the impact of preoperative renal function, assessed using estimated glomerular filtration rate, on surgical outcomes following acute type A aortic dissection. We investigated the accuracy of estimated glomerular filtration rate (in ml/min/1.73 m(2)) in predicting in-hospital mortality and postoperative renal replacement therapy in such cases.

We reviewed 114 consecutive patients with non-dialysis-dependent renal dysfunction who underwent thoracic aortic surgery for acute type A aortic dissection between 1997 and 2012. Preoperative renal function was categorized as normal (estimated glomerular filtration rate >90; n = 15) or as mild (60-89; n = 39), moderate I (45-59; n = 39), moderate II (30-44; n = 14) or severe (15-29; n = 7) renal dysfunction.

In-hospital mortality was 14.9%. Eighteen (15.8%) of 114 patients required renal replacement therapy. A more severe stage stratified by preoperative estimated glomerular filtration rate levels could effectively predict postoperative renal replacement therapy (area under the receiver operating characteristic curve 0.786). The best cut-off value of estimated glomerular filtration rate for predicting postoperative renal replacement therapy was 60 (sensitivity 95%, specificity 59%). On multiple regression analysis, the independent preoperative and intraoperative risk factors for postoperative renal replacement therapy were estimated glomerular filtration rate (P < 0.0001), coronary ischaemic time (P < 0.01) and total arch replacement (P < 0.01). Cardiopulmonary bypass time was the sole independent risk factor for in-hospital mortality (P < 0.001). On the other hand, among the morbidities, stroke [odds ratio (OR), 8.68; P < 0.01] and postoperative renal replacement therapy (OR, 5.47; P < 0.01) were independent risk factors of in-hospital mortality, according to multiple logistic regression analysis.

Preoperative estimated glomerular filtration rate can effectively predict the need for renal replacement therapy after surgery for acute type A aortic dissection. However, it is not an effective diagnostic tool to predict in-hospital mortality. The complexity of the characteristics of patients who undergo surgical procedures may make prediction of surgical outcomes difficult. Risk models to predict hospital mortality and morbidities are needed to assist clinicians in determining the optimal treatment.