Abstract
ABSTRACT. Treatment of the anemia of chronic renal failure with exogenous recombinant human erythropoietin (rHuEpo) is well established. The objective of this randomized clinical trial was to evaluate an anemia management team protocol in hemodialysis patients, using subcutaneous rHuEpo and intravenous iron. A total of 215 patients were randomized to either usual care or the protocol. The primary outcome was the proportion of patient hemoglobin (Hgb) values between 11.0 and 12.5 g/dl over the final 8 wk. The study was halted after 240 d because of an institutional change to intravenous rHuEpo. The proportion of Hgb values in the target range increased from 47.4% to 62.8% overall (P = 0.001); there was no difference between treatment groups. The proportion of baseline Hgb values between 11.0 and 12.5 g/dl increased from 44.6% in patients who had enrolled within the first 3 mo of study inception to 75.0% in those who started later (P = 0.017), suggesting a Hawthorne effect. A nonsignificant decrease in rHuEpo dose was observed in the protocol group; subgroup analysis in patients who were enrolled for at least 5 mo demonstrated a reduction in the rHuEpo dose of 2788 units/wk in the protocol group (P < 0.05), independent of intravenous iron dose. Multivariate analysis demonstrated that a higher transferrin saturation and albumin and protocol group assignment were associated with a lower final rHuEpo dose. This study demonstrated that a protocolized approach to anemia management in hemodialysis patients results in comparable Hgb levels and may reduce rHuEpo requirements, independent of iron use. E-mail: brimbles@mcmaster.ca
Treatment of the anemia of chronic renal failure with exogenous erythropoietin (rHuEpo) is a well-established practice that has led to improvements in a variety of clinical parameters, including exercise tolerance (1,2⇓), quality of life (3,4⇓), and decreased mortality (5). Adjunctive therapy with intravenous iron has been shown to improve responsiveness to rHuEpo (6–12⇓⇓⇓⇓⇓⇓); guidelines for its use have been updated recently (13,14⇓). Guidelines for the initial dosing, monitoring, and subsequent adjustments for rHuEpo have also been described (13,14⇓).
The optimal hemoglobin (Hgb) for patients with ESRD is not clear. Data exist to suggest that normalization of Hgb may be harmful in some patients (15); thus, many experts recommend a target value between 11.0 and 12.0 g/dl (13,14⇓). Despite this recommended level, a substantial proportion of patients remain below this target level; this may be related to underutilization of rHuEpo, unrecognized iron deficiency, or rHuEpo resistance (13).
Initial clinical trials of rHuEpo used simple dosing algorithms to bring Hgb values of patients up to a desired level (16–18⇓⇓); however, maintenance dosing of rHuEpo was not well studied. A study evaluating the effects of rHuEpo on left ventricular geometry used an algorithm in an attempt to standardize rHuEpo dosing across centers; however, the details of the algorithm were not described, and its effectiveness was not evaluated (19). A recent study demonstrated that utilization of an intravenous rHuEpo dosing algorithm in a small cohort of hemodialysis patients led to a greater number of patients achieving target Hgb levels (20). To our knowledge, no randomized, controlled trial has evaluated the efficacy of an anemia management team protocol. The aim of this randomized, controlled trial was to evaluate protocolized subcutaneous rHuEpo and intravenous iron changes by an anemia management team in a representative hemodialysis population.
Materials and Methods
Study Population and Randomization
This study was a single-center, randomized, controlled trial in which patients were allocated either to protocolized anemia management or to usual care by a primary nephrologist and nurse clinician. The planned study duration was 8 mo. Patients had to meet the following inclusion criteria: (1) on maintenance hemodialysis at least 3 mo, (2) receiving subcutaneous rHuEpo at the time of enrollment, and (3) primary nephrologist agreeable to Hgb target. Patients were excluded for any of the following: (1) inability to provide informed consent; (2) anticipated death, transfer, or transplant within the next 8 mo; (3) allergy to intravenous iron sucrose; (4) current participation in an anemia study; or (5) under the care of any of the study authors. Consenting patients were allocated, using a concealed randomization scheme (blocks of four), to protocolized anemia management or usual care, stratified by primary nephrologist. The nephrologists and patients participating in the study were blinded to the protocol specifics except for the target Hgb range but were aware of patient assignment. All study participants were on hemodialysis at St. Joseph’s Healthcare in Hamilton, ON, Canada. Patients received dialysis thrice weekly using either a Fresenius (2008K) or Baxter (System 1000 or Tina) hemodialysis machine with high-flux biocompatible membranes. The study was approved by the hospital ethics review board. All patients or their substitute decision makers gave written informed consent.
Anemia Protocol
A preliminary anemia management algorithm was developed by one of us (S.B.) in consultation with the other authors, review of clinical practice guidelines (13,14⇓), and the product monograph (Eprex; Jansen-Ortho). The algorithm was then pretested in 15 hemodialysis patients for 2 mo before the protocol was finalized. At the time of the study, prefilled rHuEpo syringes (1,000, 2,000, 3,000, 4,000, and 10,000 units) were available at the study center. Patients who required 8,000 units two or three times per week could be prescribed this dose using aliquots from a 20,000-unit multidose vial; otherwise, usage of the multidose vials was not promoted in the study center during this time period. The maximum dose prescribed in the center is 30,000 units weekly.
Patients who were assigned to the protocol arm could be prescribed rHuEpo and iron only by the anemia management team, consisting of a research nephrologist (S.B.) and a nurse (P.M.), using the designed protocol (Figure 1). The research nurse obtained results each week and, on the basis of the algorithm, recorded changes for the research physician to review. In the protocol arm, rHuEpo dose changes were made every 4 wk on the basis of the most current Hgb value—more frequent changes were made only when the Hgb value or the rate of Hgb change was outside the preset safety parameters (shown at the top of Figure 1). As an example, a patient on 10,000 units of rHuEpo weekly with an Hgb value <11.0 g/dl, a ferritin value >100 μg/L, and a transferrin saturation (TSAT) >20% would have his or her rHuEpo dose increased by one step to 4000 units thrice weekly.
Figure 1. Anemia management algorithm. The most recent hemoglobin (Hgb) value was reviewed every 2 wk; a change was made by at the physician’s discretion if any of the safety parameters were breached. Otherwise, Hgb values were assessed every 4 wk for a potential change in recombinant human erythropoietin (rHuEpo) dose or intravenous iron administration, considering the most recent transferrin saturation (TSAT) and ferritin values. When a change in rHuEpo dose was indicated, a one-step change was made, based on the dosing schedule shown.
The desired Hgb range was set at 11.0 to 12.5 g/dl by consensus among the authors; nephrologists whose patients were participating in the control arm of the study were asked to aim for this target range. Otherwise, nephrologists were asked to continue anemia management in their patients as they saw fit. Updated electronic reports of Hgb, ferritin, and TSAT values were sent weekly to each of the primary nephrologists for all patients who were receiving dialysis at the center to ensure equal access to test results. Hgb values were obtained before the midweek dialysis session every 2 wk, and ferritin and TSAT values were obtained monthly. Serum albumin levels and the urea reduction ratio were obtained monthly and parathyroid hormone and C-reactive protein (CRP) levels were obtained every 3 mo as part of routine practice. Patients who required hospitalization primarily for medical reasons were temporarily withdrawn from the study for safety reasons as the protocol was designed and pretested in patients without intervening acute medical conditions that might lead to unpredictable changes in Hgb values. Once patients were discharged or were considered medically stable, they then resumed participation in the study.
All patients who were in the protocol arm that were not on iron supplementation at enrollment were started on oral ferrous gluconate 600 to 900 mg daily. Intravenous iron sucrose (Venofer; Luitpold Pharmaceuticals Inc.) was prescribed when the ferritin was <100 μg/L or the TSAT was <20% and the Hgb was below range. Patients received 100 mg intravenously over each of the next 10 hemodialysis treatments and then every 2 wk thereafter. When subsequent ferritin or TSAT values remained below these values, a repeat loading of 100 mg over the next five hemodialysis sessions was given. When the ferritin or TSAT were >800 μg/L and 50%, respectively, the intravenous iron was held until the values fell below these values. Intravenous iron was then restarted at 50 mg every 2 wk.
Outcomes
The primary outcome of the study was the proportion of patients with Hgb values in target range (11.0 to 12.5 g/dl), assessed by averaging all protocol Hgb values over the final 8 wk of the study. The baseline proportion of Hgb values in target range were also obtained by averaging the Hgb values over the 8 wk immediately preceding patient entry into the study. Secondary outcomes evaluated were weekly rHuEpo dose at completion of the study, proportion of patients below the target range, mean Hgb values, and average intravenous iron dose. Univariate and multivariate predictors of the final rHuEpo dose were also evaluated.
Sample Size and Statistical Analyses
The required sample size, adjusted for stratification by a nephrologist and assuming a 10% dropout rate, was estimated to be 215 patients. This calculation assumed a clinically important difference of the proportion of patients within target Hgb range of 30% between the two groups and using α and β values of 0.05 and 0.2, respectively. Mean and SD for continuous variables were calculated for both treatment groups. Differences between means were analyzed using the t test. Differences in proportions were evaluated using the McNemar χ2 test within patient groups (baseline/follow-up) and logistic regression between patient groups (using the baseline value as a covariate). Differences between final rHuEpo doses in the treatment groups were evaluated using ANCOVA with the baseline rHuEpo dose as a covariate (21). Multivariate correlates with final rHuEpo dose were determined using multiple linear regression on significant (P < 0.10) univariate variables, controlling for treatment assignment. P < 0.05 was considered statistically significant. Data analysis was carried out using the SPSS v. 11.0 (SPSS Inc., Chicago, IL) software package.
Results
Patient Characteristics
A total of 468 patients were screened and 215 patients were randomized (45.9%) between October 2001 and March 2002 (Figure 2). The most common reason for not entering the study was that the patient was under the care of one of the study nephrologists (96 patients; 20.5%). Other major reasons for nonentry were patient refusal to provide consent (53 patients; 11.3%), primary nephrologist refusal to allow patient participation (29 patients; 6.2%), and participation in another erythropoietic hormone study (21 patients; 4.5%). A decision was made at the study institution in July 2002 to switch all hemodialysis patients from subcutaneous to intravenous rHuEpo on the basis of the recommendation of the rHuEpo provider (Eprex, Jansen-Ortho) and the Canadian Health Products and Food Branch. This was based on the observed rare risk of pure red cell aplasia associated with subcutaneous Eprex. This policy prompted early termination of the study in July 2002. A total of 167 patients (77.7%) had completed at least 5 mo of the study; a secondary analysis of this study population was also carried out for differences in the outcomes of interest. Baseline demographics and clinical information were similar in the two groups (Table 1) except for the presence of diabetes mellitus (DM) and the use of arteriovenous fistulae (AVF) (both higher in the control group).
Figure 2. The flow of patients within the study. Reason for patient exclusion and failure to provide consent are included within the text.
Table 1. Baseline characteristics of patients enrolled in the studya
Outcomes
During the study, 43 patients dropped out prematurely. Patient data were censored at the time of withdrawal from the study; patients who were enrolled for >5 mo before censoring were included in the secondary analysis. There were 20 deaths (9.3%) during the trial (eight in control group versus 12 in protocol group; P = 0.36), eight transplants, and five transfers to other centers. There was no difference in hospitalization rates or transfusion requirements (data not shown). Ten patients were withdrawn at the request of either the patient or the primary physician. The mean Hgb values for the control and protocol groups at study completion were 11.7 (± 0.9 g/dl) and 11.5 (± 1.1 g/dl), respectively; these were not significantly different (P = 0.17). Final Hgb values were similar in a secondary analysis restricted to patients who were enrolled for >5 mo (11.7 versus 11.6 g/dl, respectively; P = 0.82). The overall proportion of patients in the target Hgb range (11.0 to 12.5 g/dl) increased from 47.4% at baseline to 62.8% at the completion of the study (P = 0.001); these proportions increased significantly in both the control (49.1% to 62.0%; P = 0.05) and the protocol (45.8% to 63.6%; P = 0.02) groups; however, there was no significant difference (P = 0.80) between the groups (Figure 3). Similar results were seen in the secondary analysis (Figure 4).
Figure 3. Overview of Hgb target results in all patients. Proportion of patient below, within, or above the targeted range at baseline and completion are shown according to group allocation. *P = 0.001; †P = 0.05; ‡P ≤ 0.02. Comparisons are the final to baseline Hgb values within the patient groups.
Figure 4. Overview of Hgb target results in patients enrolled for >5 mo. Proportion of patients below, within, or above the targeted range at baseline and completion are shown according to group allocation. †P < 0.05; *P < 0.001; ‡P = 0.005. Comparisons are the final to baseline Hgb values within the patient groups.
A nonsignificant decrease in the proportion of patients with Hgb values <11.0 g/dl from baseline to the end of the study was observed overall (P = 0.08); this decrease achieved statistical significance in the secondary analysis of patients who were enrolled for at least 5 mo (32.9% to 21.0%; P = 0.01). There was no difference, however, between the two groups in the proportion of patients with Hgb values <11.0 g/dl at completion of the study in either analysis (Figures 3 and 4⇑). There was a decrease in the proportion of patients above the target Hgb range from baseline (27.1%) to completion of the study (13.1%) in the protocol (P = 0.014) but not in the control group (P = 0.86; Figure 3). This finding persisted in the secondary analysis (P = 0.04; Figure 4).
Comparison of the final rHuEpo dose between the two groups showed a nonsignificant difference between the control (12,676 units/wk) and the protocol (10,925 units/wk) groups (P = 0.16). When only patients who had completed 5 mo of the study were considered, a decrease in rHuEpo dose in the treatment group of 2788 units/wk (P < 0.05) was observed. Neither differences between final TSAT and ferritin levels nor difference in iron utilization, based on the proportion of patients prescribed intravenous iron and the average weekly intravenous iron dose, were observed between groups (Table 2). A modestly higher rHuEpo dosing frequency was observed in the control group when the analysis was limited to patients who had completed at least 5 mo of the study (P = 0.03). The average adjustment in the rHuEpo dose in the two groups was markedly different. Protocol use led to an average (directionless) adjustment in rHuEpo dose of 3138 units versus 6343 units in the control patients (P < 0.001). The difference was even more pronounced when restricted to patients who had completed at least 5 mo of the study (Table 2).
Table 2. Comparison of rHuEpo dose requirements and iron indices
Univariate analysis was performed to identify potential predictors of the final rHuEpo dose required in the study population. An increase in the final serum TSAT and albumin and a decrease in the final serum CRP were associated with a lower rHuEpo dose in the entire study population; however, a higher mean urea reduction ratio and assignment to the protocol group were also associated with a lower rHuEpo dose when the analysis was restricted to patients who were enrolled for >5 mo. Other baseline indicators not predictive of final rHuEpo dose included serum ferritin and parathyroid hormone, gender, age, weight, access type, dialysis vintage, and presence of coronary artery disease or diabetes mellitus. Multivariate analysis identified that group assignment and final serum TSAT and albumin were the only independent predictors of final rHuEpo dose (Table 3). Patient assignment to the protocol was associated with a significant decrease in the final rHuEpo dose of 2388 units (P = 0.04). An increase in the TSAT value of 1% was associated with a decrease in weekly rHuEpo requirements of 201.8 units (P < 0.001), whereas a rise in albumin of 1 g/L was associated with a decrease in the weekly rHuEpo dose of 730.8 units (P < 0.001). When the analysis was restricted to patients who were enrolled for at least 5 mo, similar results were observed (Table 3). The impact of group assignment persisted even when patients who had withdrawn from the study prematurely were included.
Table 3. Multivariate predictors of the final rHuEpo dosea
Discussion
This randomized, controlled trial demonstrated that there was no improvement in the ability to achieve target Hgb values in hemodialysis patients by using the described anemia management team protocol. Nonetheless, there was a substantial reduction in rHuEpo dose in patients who were enrolled for >5 mo, independent of the effects of intravenous iron administration and targeted Hgb.
There are several potential explanations for the protocol’s inability to increase the proportion of patients who achieved a desired Hgb target. Early termination of the study may have had an impact on the final results; it seems unlikely, however, that continuation of the study would have led to any measurable difference in the primary outcome. A previous study (20) demonstrated that implementation of an intravenous rHuEpo algorithm led to a measurable improvement in achieving target Hgb values after only 3 mo.
Subsequent revision to the protocol might lead to a greater proportion of patients achieving the desired Hgb range; however, this is difficult to assess. The use of prefilled rHuEpo syringes (as currently funded in Ontario, Canada) led to minor restrictions in our ability to make desired dosage adjustments; utilization of multidose vials or the availability of additional doses in prefilled syringes may have improved the performance of the algorithm. A more obvious explanation for the ineffectiveness of the protocol was the impressive results observed in the control group. The previous nonrandomized study described (20) had demonstrated that before implementation of an anemia management protocol, the proportion of patients who achieved target hematocrit levels (31% to 35%) was only 27% of patients, whereas the proportion of patients below target range (hematocrit <31%) was 46%. This compares with 62.0% achieving a target Hgb and only 21.3% of patients below the target in the control patients in the current study, although there were minor differences in the Hgb targets. Of interest was the improvement in the control group during the course of the study; there was a 26.3% improvement in the proportion of patients who achieved target Hgb values from baseline. This may have been because at the time of enrollment, many of the patients’ Hgb levels had not yet achieved steady state; this seemed unlikely as there was neither a difference in the proportion of patients in target Hgb range nor a difference in the mean Hgb values between the baseline values and those over the preceding 8 wk. Alternatively, nephrologists may have been previously targeting different Hgb values; hence, the improvement in the primary outcome was simply due to the nephrologists’ making the necessary rHuEpo or iron dose adjustments to achieve a new Hgb target. There is some support for this theory: patients’ Hgb values in the control group did tend to be higher at the end of the study compared with baseline. However, the proportion of baseline Hgb values in target range increased from 44.6% in patients who had enrolled within the first 3 mo of the study’s start date to 75.0% in those who started later (P = 0.017), suggesting that initiation of the study led to an overall improvement in anemia care in the hemodialysis unit. Such a phenomenon, known as the Hawthorne effect, has been described previously (22).
Although there was no difference in the primary outcome between the two groups, a difference in the rHuEpo dose requirements in the two groups was observed. This did not seem to be a result of greater iron utilization—both the average weekly iron dose and the indices of iron stores (TSAT and ferritin) were similar in the two groups. In addition, there was no difference in the upper level of rHuEpo used (30,000 units/wk). There did seem to be a reduction in the proportion of patients with Hgb values >12.5 g/dl in the protocol group by the end of the study, which may have led to minor differences in rHuEpo usage. The major difference in the groups contributing to lower rHuEpo requirements in the protocol group seemed to be the incremental nature of rHuEpo dose changes, as demonstrated by the striking difference in magnitude of the average dose adjustments observed in the two groups. Supportive of this hypothesis is that there was a strong relationship between the final rHuEpo dose and the average rHuEpo dose adjustment in patients, even when the latter was normalized to the patients’ final rHuEpo dose (data not shown).
This anemia management team protocol may be associated with significant cost savings. The current cost of rHuEpo at the study center was $0.0086 per unit (U.S.). On the basis of a dose reduction of 2788 units/wk, a cost savings of $1245.98 per patient would be achieved. In the study center alone, which currently has >430 hemodialysis patients, an estimated savings on rHuEpo costs of >$535,000 ($835,000 Canadian funds) would be realized. Such savings would easily cover the cost of one or two additional nurse coordinators if a center chose to centralize anemia management in the hemodialysis unit.
The final results achieved in the protocol group were comparable to those observed in a previous prospective cohort study that used an intravenous rHuEpo algorithm (20). The previous study showed that utilization of an anemia protocol led to an increase in patients with Hgb values in target range from 27% to 61% while reducing the proportion of patients below the target range from 46% to 18%. The respective baseline values of 47.4% in target range and 29.8% below range in this study were significantly better than that observed in the previous study and in an earlier U.S. national survey (23). A more recent survey, however, has shown that these anemia parameters have been improving in dialysis patients over time (24).
Several studies have evaluated the role of various laboratory measures in predicting rHuEpo responsiveness, including albumin (25,26⇓), CRP (25–27⇓⇓), ferritin (25,28⇓), TSAT (25,27,28⇓⇓), and several dialysis adequacy indices (26,28⇓). These studies support the notion that markers of inflammation such as an elevated CRP or reduced serum albumin tend to be associated with higher rHuEpo requirements, whereas greater iron stores and increased dialysis dose are associated with reduced rHuEpo needs. The current study found similar results with the exception that serum ferritin was not an important predictor of final rHuEpo dose. The reason for this is not clear but may be related to differences in patient populations in the different studies. Serum ferritin is known to rise with inflammation, independent of iron stores (29). In addition, this study looked at the correlation of such markers with rHuEpo requirements rather than the utility of these markers in predicting responsiveness to iron therapy, as has been recently described (30).
There are limitations to the present study. The study was terminated early and evaluated the use of subcutaneous Eprex. It is now recognized that subcutaneous administration of Eprex seems to be associated with an increase in the rate of pure red cell aplasia, and Canadian prescribing information for Eprex (December 2002) states, “It is recommended that in hemodialysis patients, the intravenous route of administration should be considered, where feasible.” Nonetheless, it would be reasonable to assume that a similar approach could be used for subcutaneous rHuEpo products not associated with this increased risk. In addition, there is no reason to believe that an algorithmic approach as described in the current study could not be extended to intravenous preparations of rHuEpo such as described previously (20). In addition to simplifying and standardizing the approach to anemia management in the hemodialysis unit, it may reduce overall rHuEpo requirements and costs.
In summary, this study demonstrated that a team-based protocolized approach to anemia management in a representative hemodialysis population could safely achieve target Hgb values in >60% of patients. This approach did not seem to offer any advantage over usual care at the study center; this seemed to be substantially due to an improvement in the results of control patients during the course of the study. Nonetheless, the results obtained in the current study were comparable to those previously described in a nonrandomized study and were associated with a substantial reduction in rHuEpo requirements, independent of intravenous iron use.
Acknowledgments
This trial was supported by a grant from Jansen-Ortho. K.S.B. and C.G.R. have received honoraria from Jansen-Ortho. These data were presented in abstract form at the 2003 Canadian Society of Nephrology meeting.
- © 2003 American Society of Nephrology
References
Jump to section
More in this TOC Section
Cited By...
- Effectiveness of Quality Improvement Strategies for the Management of CKD: A Meta-Analysis
- Comparing Mandated Health Care Reforms: The Affordable Care Act, Accountable Care Organizations, and the Medicare ESRD Program
- The New FDA Labeling for ESA--Implications for Patients and Providers
- Functional Data Analysis Applied to a Randomized Controlled Clinical Trial in Hemodialysis Patients Describes the Variability of Patient Responses in the Control of Renal Anemia
- CHAPTER 1: Hemodialysis Adequacy in Adults