[Value of estimated glomerular filtration rate and serum creatinine for predicting long-term survival in Chinese patients after isolated coronary artery bypass graft surgery].

OBJECTIVE: To compare the predictive value of glomerular filtration rate (GFR) estimated by the Cockcroft-Gault formula or the modification of diet in renal disease (MDRD) equation and serum creatinine for in-hospital and long-term mortality post coronary artery bypass graft surgery (CABG).

METHODS: Clinical data of 5559 consecutive patients undergoing isolated CABG were retrospectively analyzed. The main outcomes were in-hospital mortality and long-term mortality. Estimated GFR was calculated by the Cockcroft-Gault formula and MDRD equation respectively. Receiver-operating characteristic curves and Cox's analysis were used for the comparison.

RESULTS: Follow-up was complete in 5485 patients (97.6%). Analysis of receiver-operating characteristic curves showed that GFR estimated by the Cockcroft-Gault formula had a maximal accuracy for predicting in-hospital mortality (area under the curve: 0.755, P < 0.01). Multivariate logistic analysis and the Cox's analysis results indicated estimated GFR < 60 mlxmin(-1)x1.73 m(-2) base on the Cockcroft-Gault formula was an independent risk factor for in-hospital and long-term mortality (hazard ratio 4.51 for in-hospital mortality, P < 0.01; hazard ratio 1.54 for long-term mortality, P < 0.01), both Cockcroft-Gault formula and MDRD equation were superior to serum creatinine for predicting in-hospital and long-term mortality post CABG.

CONCLUSION: GFR estimated by the Cockcroft-Gault formula was superior to GFR estimated by the MDRD equation for predicting in-hospital mortality, and estimated GFR was superior to serum creatinine for predicting in-hospital and long-term mortality.