PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Kidney Int. Author manuscript; available in PMC Apr 28, 2013.
Published in final edited form as:
PMCID: PMC3637948
NIHMSID: NIHMS453946
High doses of epoetin do not lower mortality and cardiovascular risk among elderly hemodialysis patients with diabetes
Yi Zhang,1 Mae Thamer,1 James S. Kaufman,2 Dennis J. Cotter,1 and Miguel A. Hernán3
1Medical Technology and Practice Patterns Institute, Bethesda, Maryland, USA
2VA Boston Healthcare System, Boston, Massachusetts, USA
3Harvard School of Public Health, Boston, Massachusetts, USA
Correspondence: Dennis J. Cotter, Medical Technology and Practice Patterns Institute, 4733 Bethesda Avenue, Suite 510, Bethesda, Maryland 20814, USA. dcott/at/mtppi.org
Background
Whether exposure to high erythropoiesis-stimulating agents (ESA) doses results in adverse cardiovascular outcomes in hemodialysis patients based on the presence of diabetes has been suggested by an earlier randomized trial but remains inconclusive.
Methods
Data from the US Renal Data System were used to identify 35,593 Medicare patients on hemodialysis: 19,034 (53%) were diabetic. We fit a pooled logistic model to estimate the probability of mortality and a composite cardiovascular end point at each month using inverse probability (IP) weighting to adjust for time-dependent confounding by indication. IP weights were estimated separately for diabetic and nondiabetic cohorts.
Findings
The adjusted 9-month mortality risk difference between an ESA dose of 45,000 U/wk and 15,000 U/wk was 13% among diabetics and 5% among nondiabetics (p value = 0.002). Compared with a dose of 20,000 – 30,000 U/wk, the hazard ratio of mortality was 1.32 (95% CI 1.11, 1.58) and the hazard ratio of a composite endpoint of death and cardiovascular events was 1.26 (95% CI 1.07, 1.50) among diabetic patients exposed to the highest ESA doses (>40,000 U/wk). The corresponding hazard ratios in non-diabetic patients were 1.06 (95% CI 0.88, 1.28) and 1.10 (95% CI 0.92, 1.32), respectively.
Interpretation
Using higher ESA doses as usually necessary to achieve higher hemoglobins appears to be not beneficial, and possibly harmful, to diabetic patients. Little effect of dose was found in nondiabetic patients. Our study findings support the FDA’s recent advisories recommending that the lowest possible ESA dose be used when treating hemodialysis patients.
Keywords: ESA, epoetin, mortality, cardiovascular outcomes, diabetes, hemodialysis, marginal structural modeling, inverse probability weighting
Chronic kidney disease (CKD) and end-stage renal disease (ESRD) hemodialysis patients assigned to normal hematocrit (Hct) targets receive high doses of erythropoiesis-stimulating agents (ESA; also referred to as epoetin, epoetin alfa, or darbepeotin). Randomized trials have shown increased mortality1 or no beneficial clinical effects2 among those assigned to normal Hct targets, and observational studies3,4 have found no survival benefits among hemodialysis patients receiving high ESA doses.
Nearly half of all CKD and dialysis patients have co-morbid diabetes. Compared with CKD patients without diabetes, those with diabetes generally receive higher ESA doses, attain lower hematocrit levels,5,6,7 have a higher mortality rate, and experience more cardiovascular events.8 A recent randomized trial (TREAT) was restricted to diabetic CKD predialysis patients9 and found no cardiovascular or renal benefits and an increased risk for stroke for those assigned to darbepoetin treatment targeting hemoglobin levels >13 g/dl compared to those assigned placebo treatment with rescue treatment with darbepoetin for hemoglobin levels <9 g/dl. It is therefore possible that the apparent lack of a beneficial effect for those targeted to higher hemoglobins in all previous RCTs as well as in observational studies could be due to the mixing of diabetic and nondiabetic patients in their study groups.
Whether the adverse effects of targeting normal hemoglobins is due to the hemoglobin level or other factors remains controversial. In order to achieve higher hemoglobin levels most patients require higher ESA doses. The adverse effects of higher hemoglobin targets might be due to the higher ESA doses, particularly in patients who might be ESA hyporesponsive. In fact a recent post hoc analysis of the TREAT study found that patients who had poor initial hematopoietic response to darbepoetin were at greater risk for cardiovascular adverse events and death. The authors could not determine whether the initial poor response identified a patient at greater risk or whether the increased risk was attributable to the higher doses of darbepoetin they received. To determine whether there were adverse consequences of high ESA doses and whether these consequences were specific for diabetic patients, we designed an observational study to estimate the effect of ESA exposure (epoetin alfa) on mortality and cardiovascular outcomes among elderly hemodialysis patients with and without diabetes.
Study Data
The United States Renal Data System (USRDS) is the national source for demographic and clinical data regarding ESRD patients and their institutional providers of dialysis treatment. Medicare coverage is provided for 93% of US dialysis patients. The USRDS Medicare claims database includes the monthly hematocrit and epoetin dose administered to Medicare dialysis patients. The “Researcher’s Guide to the USRDS Database”, available from http://www.usrds.org, describes variables, data sources, collection methods, and validation studies. We merged the USRDS standard analytic files for the calendar years 2004-2006 with the USRDS core CD files that contained variables from patient, medical evidence, facility, and physician/supplier data files.
We identified incident hemodialysis patients who were 65 years and older, received their first ESRD service between 2004-2005, received their first outpatient dialysis service within the first 90 days after being certified as ESRD, did not have missing claims data during the 90-day baseline period, and had Medicare as their primary payer. These inclusions ensure that we had the complete Medicare (Parts A and B) claims treatment history for our study group. Patients treated in hospital-based dialysis facilities have more co-morbidities and a higher mortality rate when compared to those treated in free-standing dialysis facilities,10 so we restricted our study population to patients treated in free-standing facilities. We also excluded patients with a history of HIV or cancer as these patients might respond differently to epoetin therapy compared with the ESRD population at large.11,12,13 Patients who received darbepoetin (0.9%) were excluded from the study. Finally, we classified a patient as diabetic if diabetes was reported to be the primary cause of renal failure and/or diabetes was listed as a co-morbid condition on the Medical Evidence Form 2728 which is completed when a patient enrolls in the Medicare ESRD program (Figure 1).
Figure 1
Figure 1
Selection of study population from USRDS data.
Outcome ascertainment
The two endpoints of interest were all-cause mortality and a composite outcome including death and hospitalization for myocardial infarction (MI), stroke or congestive heart failure (CHF). The composite outcome was similar to the one used in two recent ESA randomized trials.1,9 We defined the cardiovascular events included in the composite outcome measure by using the following International Classification of Disease, Ninth Revision, Clinical Modification (ICD-9-CM) codes: MI: codes 410.xx (except 410.x2); CHF: codes 402.x1, 425.xx, 428.xx, 518.4, and 398.91; and stroke: codes 430.xx, 431.xx, 432.xx, 433.xx, and 434.xx. Using these ICD-9 codes found on Medicare hospital claims, we determined the primary reason for hospitalization and, identified those patients who experienced an event of interest.
Definition of follow-up and treatment
Once the initial three months of outpatient dialysis treatment had past, patient follow-up began. The observation period ended nine months later, or at death or at a censoring event whichever happened the earliest. The censoring events included: a change of dialysis modality; receipt of a kidney transplant; at 60 days after change of dialysis provider; or when a 30-day gap in outpatient dialysis services occurred. Patients who had experienced any of these censoring events during the first three months on dialysis were also excluded from the analysis.
For each patient and each month of follow-up, we calculated the cumulative average dose of epoetin as the cumulative dose received since the start of follow-up divided by the time of follow-up. Epoetin administered during hospitalization is not available on Medicare claims data even though it is likely that hemodialysis patients with longer hospital stays would have received some epoetin. As in our previous study,4 we assumed that patients receive epoetin from day-five of hospitalization onward at the same dose level (dose per administration) as they received during their immediate pre-hospitalization outpatient dialysis period. We also conducted a sensitivity analysis where we assumed that the same pre-hospitalization epoetin dose was given throughout the hospital stay.
Statistical Methods
We fit a pooled logistic model to estimate the probability of the outcome as a function of cumulative average log epoetin dose (cubic splines with knots located at 5th, 25th, 75th, and 95th percentiles of dose prior to that month), baseline covariates measured at the start of the follow up, and cubic splines of follow-up (months). Our model also included product (“interaction”) terms between cumulative average log epoetin dose and the month variables.
Because higher epoetin doses are more likely to be prescribed to patients with lower hematocrits who also might be at higher mortality and cardiovascular risk, the estimates from our model needed to be adjusted for the effect of time-dependent confounding by indication.14 As previously described,15 we used inverse probability (IP) weighting to adjust for time-dependent confounding by measured time varying covariates including hematocrit, iron treatment, and hospitalization. When the outcome of interest was death only, we also adjusted the model for hospitalization due to CHF, MI, and stroke.
Each patient received a time-varying weight inversely proportional to the estimated probability of having his/her own observed epoetin dose history, as described elsewhere.4 Briefly, the IP weights were estimated by fitting two nested models: 1) a logistic regression model to estimate each patient’s probability of not receiving epoetin at any given month (7% of the patient-months had zero dose); and 2) a linear regression model to estimate each patient’s density (assumed to be normal) of the log epoetin dose among those patients with nonzero dose in that month. Both models included the baseline covariates and time-varying covariates listed above plus the following product terms: between hematocrit and log epoetin dose; between hematocrit and hospitalization; and between chains and both baseline epoetin dose and time-varying epoetin dose (because chain characteristics are associated with epoetin dosing patterns16). IP weights were also estimated to adjust for potential selection bias due to censoring. Both the epoetin dose and censoring weights were stabilized and the product (of these weights) was used to fit the weighted regression model. The mean of estimated epoetin treatment weights was 1.02 and the 99 percentile was 5.80. The mean of the estimated censoring weights was 0.99 and the 99 percentile was 1.07. We truncated the IP weights to a maximum of 20 and used those observations in the primary analyses. Truncation did not materially affect the point estimates. All models were fit separately for diabetic and nondiabetic patients.
We then applied the weighted logistic model to estimate the survival curve under three hypothetical scenarios: all patients exposed to a cumulative average dose of 1) 15,000 units/week, 2) 30,000 units/week, and 3) 45,000 units/week. These dose levels were selected to approximate low, average and high exposure to epoetin. Point-wise 95% confidence intervals for the survival model were calculated via percentile-based nonparametric bootstrap based on 200 full samples.
For comparison with previous analyses, we also estimated average hazard ratios over the entire follow-up. To do do, we fitted the weighted logistic model for this outcome without product terms between epoetin dose and other variables, and with epoetin dose replaced by categorical variable at five dose levels: 0 - <10,000; 10,000 - <20,000; 20,000 - <30,000; 30,000 - <40,000; and > 40,000 units/week. The reference group, chosen to represent FDA-recommended dosage levels, was 20,000-30,000 U/wk. To assess potential effect modification by diabetic status, we estimated the 9-month survivals separately in diabetics and nondiabetics and used bootstrapping (200 samples) to test whether the differences were equal to zero. All analyses were conducted with SAS 9.1 (SAS Institute, Cary, NC).
Figure 1 depicts the selection process for patients included in the analysis. The analysis for the death only outcome included 35,593 patients, of whom 19,034 (53%) were diabetic. During the follow-up period, 8,238 (23%) of the patients were censored and 7,179 (20%) died. The analysis for the composite outcome included 32,534 patients (after 3,059 patients were excluded for having had a stroke, MI, or CHF during the study baseline period), of whom 17,387 (53%) were diabetic. 6,933 patients (21%) were censored; 8,512 (26%) had a composite event including 4,868 deaths (57%), and 3,644 (43%) cardiovascular events. Of the 3,644 cardiovascular hospitalization events, 2,117 (24.7%) were for CHF, 726 (8.5%) were for stroke, and 801 (9.5%) were for an MI. The event rates for diabetics and nondiabetics were similar for both mortality and composite outcomes (data not shown).
Compared with nondiabetics, diabetics were younger, more likely to be male and black, had a more severe comorbidity burden (higher Charlson score) with almost three times as many cardiovascular comorbidities, and had an increased likelihood of being hospitalized and for longer periods of time (Table 1). Although diabetic patients were more likely to receive predialysis epoetin therapy, they received similar amounts of epoetin during the baseline period and achieved similar hematocrit levels at the end of baseline when compared to nondiabetic patients.
Table 1
Table 1
Characteristics of study population (N = 35,593).
Figure 2 shows the adjusted survival curves under three hypothetical epoetin dose levels: 15,000 U/wk, 30,000 U/wk, and 45,000 U/wk throughout the entire follow-up period and for each outcome separately (Figure 2A shows death only and Figure 2B shows the composite outcome). The survival decreased with increasing doses. For the lowest epoetin doses of 15,000 U/wk, the 9-month risk of death was 20% (28% for composite endpoint) among diabetics and 24% (31%) among nondiabetics.
Figure 2
Figure 2
(A) Survival outcome probabilities for three selected epoetin dosage regimens: Low dosage (15,000 U/wk), medium dosage (30,000 U/wk), and high dosage (45,000 U/wk), based on the primary analysis which imputes epoetin for patients with a hospital stay (more ...)
The difference in mortality risk between 30,000 units/week and 15,000 units/week was 9% (95% CI 7%, 11%) among diabetics and 5% (95% CI 2%, 7%) among nondiabetics (p-value = 0.04 for heterogeneity between diabetics and nondiabetics). The difference in mortality risk between 45,000 units/week and 15,000 units/week was 13% (95% 10%, 16%) among diabetics compared to 5% (95% CI 2%, 9%) among nondiabetics (p-value = 0.002).
The difference in the risk of a composite endpoint between 30,000 units/week and 15,000 units/week was 6% (95% CI 3%, 8%) among diabetics and 5% (95% CI 2%, 8%) among nondiabetics (p-value = 0.57 for heterogeneity between diabetics and nondiabetics). The difference in mortality risk between 45,000 units/week and 15,000 units/week was 9% (95% 5%, 12%) among diabetics compared to 5% (95% CI 1%, 8%) among nondiabetics (p-value = 0.19).
Table 2 presents the estimated average hazard ratios (HRs) for both mortality and the composite outcome. The estimates were consistent with the findings from the survival curves. Among both diabetics and nondiabetics, lower epoetin dose levels (less than 20,000 U/wk) were associated with lower risk for both mortality and composite outcomes. Compared with a dose of 20,000 – 30,000 U/wk, the hazard ratio of mortality was 1.32 (95% CI 1.11, 1.58) and the hazard ratio of a composite endpoint of death and cardiovascular events was 1.26 (95% CI 1.07, 1.50) among diabetic patients exposed to the highest ESA doses (>40,000 U/wk). The corresponding hazard ratios in non-diabetic patients were 1.06 (95% CI 0.88, 1.28) for mortality and 1.10 (95% CI 0.92, 1.32) for composite end point.
Table 2
Table 2
Cumulative average epoetin dose and hazard ratios (HR) based on inverse probability weighting.
In secondary analyses the estimates were similar when we made different assumptions about the amount of epoetin used during hospitalizations (Appendix Figure 1). Our estimates did not materially change when we used: other summaries of epoetin use (i.e., total cumulative dosage from start of follow-up, recent and past cumulative average dosage) in the logistic model; cubic splines with knots at different locations; IP weights that were not truncated; IP weights were estimated under a gamma-distribution for the log of epoetin dosage; alternative categorizations of hematocrit values; expanded billable service (or claims) gap definition from 30 to 60 days; censoring that did not include change of provider; and censoring criteria that included reduced dialysis sessions.
We found that high ESA doses were associated with a greater risk of death and cardiovascular outcomes in diabetic hemodialysis patients, but not in non-diabetic hemodialysis patients. In diabetic patients, the adjusted 9-month risk of death increased from 20% for a dose of 15,000 units/week to 33% for 45,000 units/week. The corresponding increases in non-diabetic patients were smaller and did not reach the traditional level of statistical significance.
Observational studies have shown that diabetes is a contributory comorbid factor that increases mortality risk among anemic CKD patients.6 Among Type 2 diabetes CKD patients, anemia is associated with an increased risk of cardiovascular events.6,7 At high ESA levels, diabetic hemodialysis patients could be at a higher risk for adverse outcomes when compared to their nondiabetic counterparts due to the increased presence of hypertension in this group.17 Extra-hematopoetic effects of epoetin use, particularly worsening of hypertension, should be explored to determine whether it contributes to an increase in mortality and cardiovascular events among hemodialysis diabetic patients who are exposed to high ESA doses.18
Our findings are consistent with earlier CKD CREATE and CHOIR trials as well as the hemodialysis NHT trial; all showing no survival and/or cardiovascular benefits for those targeted to higher hematocrits and exposed to higher ESA doses. Our findings are also consistent with TREAT trial results, a study that found no increased composite cardiovascular events among diabetic predialysis CKD patients targeted to a hemoglobin level of 13 g/dL. In a post hoc analysis of TREAT, Solomon et al. reported that patients who were initially hyporesponsive to ESA therapy and who subsequently received the higher doses of darbepoetin alfa were at highest risk for adverse outcomes.19 The authors could not distinguish whether their observed effect was due to high ESA doses or other factors that lead to ESA hyporesponsiveness. There are several clinical pathways through which higher epoetin exposures could be detrimental, e.g., hypertension, thrombosis, etc., independent of any effect of achieved hematocrit levels. According to Vaziri18et al. and others,20,21 chronic erythropoeitin administration results in a hematocrit-independent, elevated vasoconstriction-dependent hypertension.
Our analysis has several limitations. First, potential lack of control for confounding among diabetics might have resulted in exaggerating the risk of high doses. For example, some components of diabetes treatment such as use of insulin, beta blockers due to increased risk of MI, and other drugs, were not available and their impact on outcomes could not be estimated. IP weight adjustment brought the hazard ratios closer to the null; it is possible that further adjustment would further attenuate the estimated risk at elevated dose levels. Second, although we used both diabetes as an underlying cause of ESRD and Form 2728 listing diabetes as a co-morbid condition at the time of ESRD to identify the diabetes cohort, to the extent that these are underreported and misclassification of diabetics exists in the nondiabetes cohort, the differences based on diabetic status reported herein would tend be conservative lower estimates. Third, our study used monthly Medicare claims data that were collected primarily for the purposes of billing for healthcare services. To limit the number of false positives for cardiovascular disease (which were identified by using ICD-9-CM codes in our study), we used only the primary reason (diagnosis) for hospitalization i.e., congestive heart failure, or MI, or stroke. To assess the sensitivity of our estimates to the lack of data on ESA use during hospitalizations, we conducted several analyses under different plausible scenarios.
In conclusion, our findings suggest that usign higher ESA doses as would be needed to achieve higher hemoglobin levels is not beneficial, and is possibly harmful, for diabetic patients. Our results are consistent with FDA’s recent advisories that recommend that the lowest possible ESA dose be used when treating hemodialysis patients.
Acknowledgments
Funding— U.S. National Institutes of Health grants R01-DK066011-01A2 and R01-HL080644-01 and the U.S. Agency for Healthcare Research and Quality grant R21-HS19513-01. The data reported here have been supplied by the U.S. Renal Data System. The interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as the official policy or interpretation of the U.S. government.
Appendix
Figure 1
Figure 1
(A) Survival outcome probabilities for three selected epoetin dosage regimens: Low dosage (15,000 U/wk), medium dosage (30,000 U/wk), and high dosage (45,000 U/wk), based on the secondary analysis which inputes epoetin throughout the duration of the hospital (more ...)
Appendix Table 1
Cumulative average epoetin dose and mortality hazard ratios based on standard models (controlling for baseline variables only).
Death OnlyComposite

Epoetin DoseDiabeticNon-DiabeticDiabeticNon-Diabetic
Units/weekHR95% CIHR95% CIHR95% CIHR95% CI
<10,0000.600.540.680.530.470.600.660.590.730.590.530.66
10,000-<20,0000.720.650.790.750.680.840.770.700.840.800.720.88
20,000-<30,000 (ref group)refrefrefref
30,000-<40,0001.151.021.301.231.071.401.120.991.261.271.121.45
>=40,0001.491.321.681.381.211.571.391.241.561.351.191.53
Appendix Table 2
Cumulative average epoetin dosage and HRs for weighted regression models based on un-truncated weights.
Death OnlyComposite

DiabeticNon-DiabeticDiabeticNon-Diabetic
Units/weekHR95% CIHR95% CIHR95% CIHR95% CI
<10,0000.710.580.870.680.520.910.820.690.990.640.510.80
10,000-<20,0000.710.600.830.860.691.070.880.761.020.890.721.10
20,000-<30,000 (ref group)refrefrefref
30,000-<40,0001.210.951.531.110.881.391.251.011.561.140.911.42
>=40,0001.301.081.571.040.841.291.261.051.511.100.901.36
1. Singh AK, Szczech L, Tang KL, et al. CHOIR Investigators. Correction of anemia with epoetin alfa in chronic kidney disease. N Engl J Med. 2006;355:2085–2098. [PubMed]
2. Drüeke TB, Locatelli F, Clyne N, et al. CREATE Investigators. Normalization of hemoglobin level in patients with chronic kidney disease and anemia. N Engl J Med. 2006;355:2071–2084. [PubMed]
3. Zhang Y, Thamer M, Stefanik K, Kaufman J, Cotter DJ. Epoetin requirements predict mortality in hemodialysis patients. Am J Kidney Dis. 2004;44(5):866–876. [PubMed]
4. Zhang Y, Thamer M, Cotter D, Kaufman J, Hernán MA. Estimated effect of epoetin dosage on survival among elderly hemodialysis patients in the United States. Clin J Am Soc Nephrol. 2009;4(3):638–644. [PubMed]
5. Khoshdel A, Carney S, Gillies A, et al. Potential roles of erythropoietin in the management of anaemia and other complications diabetes. Diabetes Obes Metab. 2008;10(1):1–9. [PubMed]
6. Vlagopoulos PT, Tighiouart H, Weiner DE, et al. Anemia as a risk factor for cardiovascular disease and all-cause mortality in diabetes: the impact of chronic kidney disease. J Am Soc Nephrol. 2005;16:3403–10. [PubMed]
7. Tong PCY, Kong APS, So W-Y, et al. Hematocrit, independent of chronic kidney disease, predicts adverse cardiovascular outcomes in Chinese patients with type 2 diabetes. Diabetes Care. 2006;29:2439–2444. [PubMed]
8. U.S. Renal Data System Excerpts From the United States Renal Data System 2005 Annual Data Report. Am J Kidney Dis. 2006;47(Suppl 1):S1. [PubMed]
9. Pfeffer M, Burdmann EA, Cooper DE, et al. A trial of darbepoetin-alfa in type 2 diabetes and chronic kidney disease. N Engl J Med. 2009;361:1–14. [PubMed]
10. U.S. Renal Data System USRDS 2007 Annual Data Report: Atlas of End-stage Renal Diesease in the United States. Natioanl Institutes of Health, National Institute of Diabetes and Digestive and Kidney Disease; Bethesda, MD: 2007. p. 221.
11. Henke M, Laszig R, Rube C, et al. Erythropoietin to treat head and neck cancer patients with anemia undergoing radiotherapy. Lancet. 2003;362:1255–1260. [PubMed]
12. Leyland-Jones B. BEST investigators. Breast cancer trial with erythropoietin terminated unexpectedly. Lancet. 2003;4:459–460. http://oncology.thelancet.com. [PubMed]
13. Ifudu O, Mathew JJ, Mayers JD, et al. Severity of AIDS and response to epoetin in uremia. Am J Kidney Dis. 1997;30:28–35. [PubMed]
14. Hernán MA, Hernandez-Diaz S, Robins JM. A structural approach to selection bias. Epidemiology. 2004;15:615–625. [PubMed]
15. Robins JM, Hernán MA, Brumback B. Marginal structural models and causal inference in epidemiology. Epidemiology. 2000;11:550–560. [PubMed]
16. Thamer M, Zhang Y, Kaufman J, et al. Dialysis facility ownership and epoetin dosing in patients receiving hemodialysis. JAMA. 2007;297:1667–1674. [PubMed]
17. Suh DC, Kim CM, Choi IS, Plauschinat CA, Barone JA. Trends in blood pressure control and treatment among type 2 diabetes with comorbid hypertension in the United States: 1988-2004. J Hypertens. 2009 Sep;27(9):1908–1916. [PubMed]
18. Vaziri ND. Cardiovascular effects of erythropoietin and anemia correction. Current Opinion in Nephrol Hypertension. 2001;10:633–637. [PubMed]
19. Solomon SD, Uno H, Lewis EF, Eckardt KU, Lin J, Burdmann EA, de Zeeuw D, Ivanovich P, Levey AS, Parfrey P, Remuzzi G, Singh AK, Toto R, Huang F, Rossert J, McMurray JJ, Pfeffer MA., Trial to Reduce Cardiovascular Events with Aranesp Therapy (TREAT) Investigators Erythropoietic response and outcomes in kidney disease and type 2 diabetes. N Engl J Med. 2010;363(12):1146–1155. [PubMed]
20. Sasaki N, Ando Y, Kusano E, Asano Y. A case of erythropoietin induced hypertension in a bilaterally nephrectomized patient. ASAIO J. 2003;49:131–135. [PubMed]
21. Sunder-Plassmann G, Horl WH. Effect of erythropoietin on cardiovascular diseases. Am J Kidney Dis. 2001 Oct;38(4 Suppl 1):S20–S25. [PubMed]