Determining Optimum Hemoglobin Sampling for Anemia Management from Every-Treatment Data

Background and objectives: Anemia Management Protocols in ESRD call for hemoglobin (Hb) monitoring every 2 to 4 weeks. Short-term Hb variability affects the reliability of Hb measurement and may lead to incorrect dosing of erythropoiesis stimulating agents. We prospectively analyzed short-term Hb variability and quantified the relationship between frequency of Hb monitoring and error in Hb estimation. Design, setting, participants, & measurements: Using the Crit-Line III TQA device, we prospectively observed Hb during each dialysis treatment in 49 ESRD patients and quantified long- and short-term Hb variability. We estimated Hb from data sampled at regular intervals; 8x, 4x, 2x, or 1x per month to establish how well we account for short-term variability at different monitoring intervals. We calculated the Hb estimation error (Hb(err)) as a root mean-squared difference between the observed and estimated Hb and compared it with the measurement error. Results: The most accurate Hb estimation is achieved when monitoring 8x per month (Hb(err) = 0.23 +/- 0.05 g/dl), but it exceeds the accuracy of the measurement device. The estimation error increases to 0.34 +/- 0.07 g/dl when monitoring 4x per month, 0.39 +/- 0.08 g/dl when monitoring 2x a month, and 0.45 +/- 0.09 g/dl when monitoring 1x per month. Estimation error comparable to instrument error information is as follows: 8x per month, 15 patients; 4x per month, 22 patients; 2x per month, 6 patients; 1x per a month, 6 patients. Conclusions: Four times a month is the clinically optimal Hb monitoring frequency for anemia management. Clin J Am Soc Nephrol 5: 1939-1945, 2010. doi: 10.2215/CJN.03540410