Is modification of diet in renal disease formula similar to Cockcroft-Gault formula to assess renal function in elderly hospitalized patients treated with low-molecular-weight heparin?

Background. Repeated administration of low-molecular-weight heparin (LMWH) to elderly patients with an impaired renal function may lead to an accumulation effect with an increased risk of bleeding. In this setting, Cockcroft-Gault (CG) is the most widely used formula for glomerular filtration rate (GFR) estimation. In hospitalized patients over the age of 70, the six-variable Modification of Diet in Renal Disease (MDRD) formula was compared with the CG formula to detect patients with renal impairment who are at higher risk of bleeding when treated with LMWH. Methods. We combined retrospective data from 366 patients aged 86.2 +/- 6.6 years, treated with LMWHs. CG and MDRD GFR estimates were compared using the Bland-Altman method and the agreement between the two formulae by the kappa coefficient. Results. The mean CG and MDRD estimated GFR were 45.9 +/- 21.9 mL/min and 75.6 +/- 32.6 mL/min/1.73 m(2), respectively, with a mean bias of 29.6 mL/min. The concordance between the formulae to classify patients into stages of kidney disease was very poor (weighted kappa = 0.17): 21.8% patients had severe renal function impairment with the CG formula versus 1.3% with the MDRD formula. In our population, the MDRD thresholds that would correspond to CG estimates of 30 mL/min and 60 mL/min were found at 63 mL/min/1.73 m(2) and 80 mL/min/1.73 m(2), respectively. Conclusions. In elderly patients, GFR estimates using MDRD and CG formulae differ widely and identify different numbers of individuals with kidney disease. Prospective comparative studies are needed to validate these formulae and their different thresholds to better detect elderly patients at higher risk of bleeding when treated with LMWH.