Predicting mortality using two renal function estimation methods in hospitalised stroke patients.

The prognostic abilities of the MDRD and Cockcroft-Gault methods for estimating renal function were compared in a cohort study of 1287 patients with acute stroke admitted to a Scottish tertiary care teaching hospital. Using Cox regression analysis corrected for other prognostic variables, both the MDRD and Cockcroft-Gault equations predicted mortality independently of other prognostic factors. A 1 ml/min reduction in GFR as calculated by MDRD was associated with a 1.0% (95% CI: 0.3-1.6) increase in risk of death. A 1 ml/min reduction in creatinine clearance from the Cockcroft-Gault equation was associated with a 1.7% (95% CI: 0.9-2.6) increase in risk of death. The Cockcroft-Gault equation weakly predicted length of stay (r=0.066, p=0.02, Spearman's rank test). In conclusion, both methods independently predict early and late mortality in stroke patients, but the Cockcroft-Gault estimate has greater predictive power in this population.