Strengths and weaknesses of established indirect models to detect recombinant human erythropoietin abuse on blood samples collected 48-hr post administration.