Concordance in critical patients between the equations designed for the calculation of glomerular filtration rate and 24-hour creatinine clearance

Objective: To study the behavior of the different equations used to estimate glomerular filtration rate (GFR) applied to critical care patients compared to the standard method: 24-hour creatinine clearance (24-CrCl). Design: Retrospective analysis of data base from a previous observational prospective study. Setting: Polyvalent ICU in a tertiary Hospital. Population: All adult patients admitted to our Unit during the study who had a bladder catheter inserted. Anuric patients were excluded. Interventions: We measured 24-CrCl and estimated GFR by MDRD, modified Jelliffe (JF), Mayo-Clinic (CM) and Cockroft-Gault (C-G) equations. Variables: To evaluate degree of agreement, we grouped patients regarding 24-CrCl as normal (>70), moderate dysfunction (69-50) or severe renal dysfunction ( < 50 mL/min/1.73 m(2)). Results: 307 patients, aged 54 +/- 18, 69.7% males. Measured 24-CrCl was 109.2 +/- 78.2 mL/min/1.73 m(2) and the estimate one 95.5 +/- 56.7 for JF, 87.4 +/- 53.4 for C-G, 86.9 +/- 55.9 for MDRD and 85.6 +/- 39.9 for CM. The difference was significant (p < 0.001) for all estimates but lower for (13.7 +/- 53.2 mL/min/1.73 m(2)) than C-G (21.9 +/- 58.3), CM (23.6 +/- 59.6) or MDRD (22.3 +/- 60.4). Correlation coefficient was 0.73 for JF, 0.67 C-G or CM and 0.64 for MDRD. The degree of agreement was only fair for all measures (Kappa 0.55 for JF or MDRD, 0.51 for C-G and 0.5 for CM). Conclusions: Modified Jelliffe equation showed higher agreement with 24-CrCl than Cockroft-Gault, MDRD or Mayo-Clinic equations when used in critically ill patients. However, when exact measurement is needed, none of the equations can be considered adequate and in these cases, the CrCl should be calculated. (C) 2009 Elsevier Espana, S.L. and SEMICYUC. All rights reserved.