1. ¹Medical Technology and Practice Patterns Institute, Bethesda, Maryland, USA
2. ²Renal Section, Veterans Affairs Boston Healthcare Systems and Boston University School of Medicine, Boston, Massachusetts, USA
3. ³Department of Epidemiology, Harvard School of Public Health, Boston, Massachusetts, USA

Correspondence: D Cotter, Medical Technology and Practice Patterns Institute, 4733 Bethesda Avenue, Suite 510, Bethesda, Maryland 20814, USA. E-mail: dcott@mtppi.org

Received 4 May 2007; Revised 2 August 2007; Accepted 28 August 2007; Published online 14 November 2007.

Abstract

Nearly all dialysis patients receive epoetin therapy to treat anemia. Using the United States Renal Data System, we monitored the 14 001 patients aged 65 and older who started dialysis and epoetin treatment in 2003–2004. We estimated the dose–response relationship for the average epoetin dose and hematocrit during a 3-month initiation and subsequent 3-month maintenance phase using a marginal structural model to adjust for measured time-dependent confounding by indication. During the initiation phase, an S-shaped dose–response relationship for average weekly epoetin dose and hematocrit response was found. Average hematocrit levels rose as the epoetin dose was increased from 9000 to approximately 22 500 units per week. At higher doses, the effect of increasing epoetin was minimal with average hematocrit levels plateauing at 38.5% , but this was less evident in the maintenance phase. Among patients who reached this phase, doses required to maintain the hematocrit level were lower than those required to achieve similar hematocrit levels in the initiation phase. The dose–response curve found in our study suggests that published recommendations for starting dose are appropriate, and a starting dose of 7500–15 000 units per week can maintain the hematocrit level in the desired target range of 33–36% .