Early prediction of response to recombinant human erythropoietin in patients with the anemia of renal failure by serum transferrin receptor and fibrinogen.

Recombinant human erythropoietin (rHuEpo) has been shown to be effective in correcting the anemia of chronic renal failure, but the dose needed may be variable. The reason for this variation is not known, but several factors could be involved, such as iron deficiency, inflammation, aluminum intoxication, hyperparathyroidism, blood losses, or marrow dysfunction. Treatment with rHuEpo was given intravenously thrice weekly after hemodialysis to 64 consecutive unselected patients with the anemia of chronic renal failure. The starting dose was 50 U/kg/dose, which was increased to 75 and 100 U/kg/dose if no response was observed after 1 and 2 months of treatment. After a minimum follow-up of 6 months, response was evaluated as early (hematocrit [Hct] > or = 30% before 3 months) or late (Hct > or = 30% after 3 months) response, or failure (target Hct not attained). We examined the value of various laboratory parameters (baseline values and early changes) as predictors of response to rHuEpo. The best prediction by pretreatment parameters only was obtained with baseline serum transferrin receptor (TfR) (< or > or = 3,500 ng/mL) and fibrinogen (< or > or = 4 g/L): 100% response rate when both parameters were low, versus only 29% when they were both high, and versus 67% when one was low and the other high. When the 2-week TfR increment was greater than 20%, the response rate was 96%. When TfR increment was less than 20%, the response rate was 100% when baseline TfR and fibrinogen were low, 12% when fibrinogen was elevated, and 62% when fibrinogen was low but baseline TfR high. The predictive value of baseline TfR and fibrinogen and of the 2-week increment of TfR was confirmed by life table analysis and stepwise discriminant analysis. Major reasons for failure or late response were identified and included subclinical inflammation, iron deficiency, functional iron deficiency, marrow disorders, hemolysis, bleeding, and low Epo dose. We conclude that response to rHuEpo can be predicted early by pretreatment fibrinogen and TfR, together with early changes of TfR levels. These prognostic factors illustrate the importance of the early erythropoietic response, subclinical inflammation, and functional iron deficiency. Early recognition of a low probability of response in a given patient could help identify and correct specific causes of treatment failure to hasten clinical improvement and avoid prolonged ineffective use of an expensive medication.