Early dosing practices and effectiveness of recombinant human erythropoietin

In a national longitudinal-cohort study of 59,462 end-stage renal disease (ESRD) patients, we examined dosing and effectiveness of erythropoietin (EPO) during the first year of its use in clinical practice (July 1989 through June 1990). In unadjusted and multivariate analyses of Medicare claims data, the mean dose of EPO prescribed was: relatively small and similar for initial and maintenance therapy, 2752 (95% confidence interval 2740 to 2764) and 2668 (95% confidence interval 2654 to 2682) units, respectively; lower when initial therapy was started later (591 units lower in September 1989 and 760 units lower in November 1989 vs. July 1989, P < 0.0001); lower by 135 units during initial therapy and by 116 units during maintenance therapy for females (who weigh less) compared to males (P < 0.001); and lower by 400 units for patients treated in for-profit versus not-for-profit centers. In multivariate analysis: hematocrit response was less and mean maintenance dose was 298 units and 621 units greater for patients whose ESRD was due to multiple myeloma and sickle cell disease, respectively, compared to those with hypertension-related ESRD (P < 0.01); and hematocrit response was logarithmically related to dose [hematocrit = 0.97 In (dose), P < 0.0001]. Forty-four percent of patients had a hematocrit ≥ 30 after four months of therapy. The percent of patients transfused during three month periods before and after therapy decreased from 20% to 5%, respectively (P < 0.0001). Our results suggest that dosing practices were substantially modified to prescription of smaller and more fixed doses over time, due to the interplay of clinical concerns and economic forces. They also suggest that the effectiveness of EPO in increasing hematocrit levels and reducing transfusion use in routine clinical practice was less than anticipated based on the experience in clinical trials in part as a result of dosing practices.