Within-patient variation of hemoglobin and reticulocytes: implications for evaluating ESA responsiveness in dialysis patients.

INTRODUCTION: Techniques that assess percent reticulocytes (%retics) or hemoglobin (Hb) to detect erythropoiesis-stimulating agents (ESA) use in athletes may be useful in evaluating ESA responsiveness in dialysis patients. However, within-patient variation, appropriate transformation, or the relationship between the blood draw interval length and analyte variation are untested.
METHODS: In a prospective, single-arm trial, we determined Hb and %retics in 30 hemodialysis patients receiving stable ESA doses. Within-patient results were evaluated for variance homogeneity and normality with and without transformation.
RESULTS: Square-root transformation (sqrt) of %retics produced the most constant variance (lowest r-value for correlation between sqrt|normalized residuals| and fitted values: 0.018 highest P-value 0.739) compared with log transformation (r = -0.198, P < 0.001) or no transformation (r = 0.215, P < 0.001) and showed the least departure from normality (highest P-value: 0.002 vs. < 0.001 vs. < 0.001, respectively). Hb results did not improve with transformation. Within-patient variance in both %retics and Hb increased with interval length between lab draws (P < 0.001).
CONCLUSIONS: Initial assessment of anti-doping tool use in dialysis patient anemia management indicates square-root transformation of %retics and adjustment for time between lab draw intervals for both %retics and Hb will be required.