This prospective cohort study demonstrated 4 important clinical results. First, converting the route of administration of erythropoietin from subcutaneous to intravenous resulted in a significant increase in dose to maintain a target Hb level in patients followed for 12 months. Second, we were unable to demonstrate a relationship between the loss of RRF and dose requirements, as patients with and without RRF required a significant increase in dose, and RRF at baseline did not predict the net change in erythropoietin dose over time. Third, a significant increase in the dose was required, despite an increase in the percentage of patients receiving the drug 3 times weekly. Fourth, although the majority of patients required an increase in dose during the follow-up period, baseline variables were not associated with the dose.
Some previous studies have shown a decrease in erythropoietin dose when converting in the opposite direction (ie, from intravenous to subcutaneous).3⇓⇓⇓⇓⇓⇓⇓⇓⇓–13 Other studies have not shown a difference.14,15 In a meta-analysis, the average intravenous erythropoietin dose was 161 units/kg/wk and decreased by 30% to 113 units/kg/wk (p < 0.001) when erythropoietin was given by the subcutaneous route.3 In a randomized, unblinded study of 208 hemodialysis patients, the subcutaneous dose was 32% lower than the intravenous dose; however, when the analysis was restricted to subjects who completed a 26 week follow-up, the average dose ± SD in the subcutaneous group (104 ± 60 units/kg/wk) was 27% less than the dose in the intravenous group (142 ± 73 units/kg/wk; p < 0.001).9 In both of these reports, the target hematocrit was less than the current recommendations of 33–36%.16 These studies showed a greater difference in dose than the 20.2% change observed in our study. This may be attributable to our longer follow-up period and higher Hb/hematocrit targets, as well as the observational nature of our study.
Smaller differences between subcutaneous and intravenous erythropoietin doses have previously been demonstrated in observational studies compared with results from randomized, controlled trials. The European Survey on Anemia Management observed a 7.7% increase in the intravenous dose (113.5 units/kg/wk) compared with the subcutaneous dose (104.7 units/kg/wk) to maintain a target Hb level of 11 g/dL.17 This discrepancy in magnitude of difference can be partially explained by the use of specific dosing algorithms, the exclusion of patients with unstable or low Hb or hematocrit levels, and inclusion of patients who are able to give consent in the randomized, controlled studies.18
The association between greater RRF and a higher hematocrit has previously been noted.19 We observed a significant difference in the baseline dose of subcutaneous erythropoietin required in patients with RRF compared with those without RRF (110.4 vs 134.2 units/kg/wk; p = 0.04). However, both groups required a statistically significant increase in dose to maintain the target Hb level at 12 months. The increase in dose was not significantly different between the RRF groups.
Dosing in the majority of our patients was changed from weekly to 3 times weekly at the time of conversion. Previous studies demonstrated an increased dose requirement when erythropoietin was administered once versus 3 times weekly by the intravenous route.4 However, subcutaneous administration can be reduced to once-weekly dosing with maintenance of Hb levels.5 Despite this increase in frequency of dosing with intravenous erythropoietin, we demonstrated a significant increase in the dose.
This study demonstrates the variability in dose among patients when converting from subcutaneous to intravenous erythropoietin, with no association with baseline variables. The percentages of patients requiring an increase, no change, or decrease in dose were 53.5%, 15.6%, and 30.9%, respectively. Similar variability has been reported in a study looking at a conversion in the opposite direction—from intravenous to subcutaneous. In that study, an increase, no change, or a decrease in the dose occurred in 26%, 13%, and 61% of patients, respectively, almost mirroring the results we obtained.5 A multivariate model, including all of the baseline variables, explained less than 4% of the net change in weekly dose, suggesting that other unmeasured variables influenced the dosing change.
There are limitations to our study. First, it represents the experience of a single regional program; however, the population is sufficiently large and the demographics are similar to those reported in the Canadian Organ Replacement Registry, suggesting that our results may be generalizable.20 The one exception is our higher rate of permanent catheters; however, this rate did not change over the period of the study. Second, at the time of the conversion from subcutaneous to intravenous erythropoietin, the frequency of dosing was increased for the majority of patients; however, an increase in the intravenous dose was still demonstrated. Third, we did not employ a standardized anemia management algorithm; rather, anemia was managed by a consistent group of healthcare providers under the supervision of nephrologists. The anemia management team, process of delivery and administration of erythropoietin, as well as Hb targets remained constant over the study period, limiting the bias from an unblinded study. Fourth, we saw a progressive increase in the erythropoietin dose over the 12 months of follow-up. We attribute this to the change in route of administration; however, we did not follow markers of inflammation other than ferritin, nor did we routinely measure vitamin B12 or folate levels.