Secondary Logo

Journal Logo

Outcomes in Patients Undergoing Cardiac Surgery Who Decline Transfusion and Received Erythropoietin Compared to Patients Who Did Not: A Matched Cohort Study

Duce, Lorent MD*; Cooter, Mary L. MS; McCartney, Sharon L. MD; Lombard, Frederick W. MD†,‡; Guinn, Nicole R. MD

doi: 10.1213/ANE.0000000000002418
Blood Management: Original Clinical Research Report

BACKGROUND: Erythropoiesis-stimulating agents, such as erythropoietin (EPO), can be used to treat preoperative anemia. Some studies suggest an increased risk of mortality and thrombotic events, and use in cardiovascular surgery remains off-label. This study compares outcomes in cardiac surgery patients declining blood transfusion who received EPO with a matched cohort who did not.

METHODS: After institutional review board approval, we conducted a retrospective review of all patients who decline blood transfusion who underwent cardiac surgery and received EPO between January 1, 2004, and June 15, 2015, at a single institution. Control patients who did not receive EPO and were not transfused allogeneic red blood cells perioperatively were identified during the same period. Two controls were matched to each EPO patient using an optimal matching algorithm based on age, date of surgery, gender, operative procedure, and surgeon. The European System for Cardiac Operative Risk Evaluation (EuroSCORE) and baseline characteristics remaining unbalanced in the matched cohorts were controlled for in assessing patient outcomes. The primary outcome was a composite of mortality and thrombotic events, and secondary outcomes included change in hemoglobin (Hb) from baseline to discharge, acute kidney injury (AKI), sternal wound infection, atrial fibrillation, time to extubation, intensive care unit, and hospital length of stay (LOS).

RESULTS: Fifty-three patients who decline transfusion and received EPO were compared to 106 optimally matched control patients who did not receive EPO or red blood cell transfusion in the perioperative period. The median additive EuroSCORE was similar between the EPO and control group [6 (4, 9) vs 5 (3, 7), respectively; P = .39]. There was no difference in the primary outcome (P = .12) and mortality was zero in both groups. The EPO group had a higher mean preoperative Hb (13.91 g/dL vs 13.31; P = .02) and a smaller change in Hb from baseline (−2.65 vs −3.60; P = .001). The incidence of AKI (47.17% vs 41.51%; P = .49) was similar and there was no significant difference in all other outcomes, including time to extubation, hospital LOS, or intensive care unit LOS.

CONCLUSIONS: In this retrospective matched cohort study of patients declining transfusion and receiving EPO matched to control patients, there were no clinically meaningful differences in the outcomes.

From the *Department of Anesthesiology, University of Florida College of Medicine, Gainesville, Florida

Department of Anesthesiology, Duke University, Durham, North Carolina

Department of Anesthesiology, Vanderbilt University Medical Center, Nashville, Tennessee.

Published ahead of print August 29, 2017.

Accepted for publication July 12, 2017.

Funding: N.R.G. received compensation for travel related to consulting from HBO2 Therapeutics.

The authors declare no conflicts of interest.

Reprints will not be available from the authors.

Address correspondence to Nicole R. Guinn, MD, Department of Anesthesiology, Duke University Medical Center, Box 3094, Durham, NC 27710. Address e-mail to

Patients of the Jehovah’s Witness (JW) faith undergoing cardiovascular surgery present a unique challenge to the health care team: Historically, this patient population has declined blood transfusion, even in life-threatening situations. The Watchtower Bible and Tract Society, the governing body of the JW faith, has banned acceptance of blood transfusions based on several Biblical passages.1 Specifically, JWs will not accept whole blood or its primary components, such as packed red blood cells, plasma, or platelets. The acceptance of minor fractions, such as cryoprecipitate, recombinant clotting factors, and albumin, as well as the use of procedures involving the patient’s own blood, such as cell salvage and hemodilution, is left to the individual believer.1 However, JW patients will not accept autologous blood if it is separated from the patient’s circulation, and therefore predeposited autologous blood is not an option.1

Highly complex cardiovascular procedures often involve significant blood loss in the perioperative period, leading to the need for blood transfusion in up to 67% of patients.2 Transfusions become more imminent in those patients with baseline anemia, a common comorbidity in the cardiac surgical population. Interestingly, both mild preoperative anemia and transfusion of allogeneic blood products have been associated with increased risks of postoperative complications.3–8 Given the frequent need for blood product transfusion, only 109 specialized centers in the United States offer cardiac surgery to the JW population,9 which could potentially service almost 2.5 million believers.10

Since renowned cardiac surgeon Dr Denton Cooley started performing surgery on JWs in the 1960s,11 new approaches, including minimally invasive techniques, and preoperative optimization, such as intravenous iron infusions and erythropoietin (EPO) to treat preoperative anemia, have led to improved outcomes. EPO is a naturally occurring hormone and the primary regulator of erythropoiesis,12 produced in response to tissue hypoxia or severe hemorrhagic stress. EPO leads to increased erythropoiesis from the bone marrow in order to maintain oxygen delivery to vital organs.12,13 EPO for clinical use is produced by recombinant DNA technology and was first approved by the Food and Drug Administration (FDA) in 1989 for use in patients with chronic renal failure.14 The FDA has extended approval of EPO for the treatment of anemia resulting from a number of causes, including preoperative anemia, in patients undergoing elective, noncardiac, nonvascular surgery.15 However, use in cardiac surgery remains off-label due to concerns of increased risk of thrombosis and mortality. Nonetheless, EPO is commonly used in JW patients undergoing cardiac surgery, as the benefit of treating preoperative anemia may outweigh the risk of thrombosis in this patient population. To date, there are no published studies comparing outcomes between JW patients who received EPO for cardiac surgery with a matched control cohort who did not. This study aims to establish whether JW patients undergoing cardiac surgery who were treated with EPO for anemia had greater mortality or morbidity than patients who did not receive EPO during their perioperative experience.

Back to Top | Article Outline


This manuscript adheres to the applicable EQUATOR Network guidelines. After institutional review board (IRB) review and with waiver of informed consent, we retrospectively identified all patients declining transfusion and undergoing cardiovascular surgery from January 1, 2004 to June 15, 2015, at Duke University Medical Center using the Center for Blood Conversation (CBC) database. From this cohort, patients who received perioperative EPO to treat anemia were selected for inclusion in the study. Potential control patients who did not receive EPO or allogeneic red blood cell transfusion in the perioperative period—from day of surgery to 30 days postoperatively—were selected from a separate database of more than 3000 patients undergoing cardiac surgery at the same single institution during the study period. Using a SAS macro for optimal matching, JW patients who received EPO were matched 1:2 to controls based on patient age and date of surgery, with exact matches on gender, operative procedure, and surgeon.16

The CBC is consulted for all patients who decline blood products and are being considered for surgery or have been admitted to the hospital. The function of the CBC is to first confirm the patient’s wishes, then elicit which, if any, products or procedures they will accept. The patients will then sign a consent detailing the refusal or acceptance of blood products and fractions. Finally, the CBC will make recommendations to the medical and surgical teams regarding preparation for invasive procedures and assisting in selecting appropriately trained perioperative team members.

EPO is part of the perioperative optimization protocol recommended by the CBC. Details of this protocol, including dosing of EPO and iron and use of cell salvage and acute normovolemic hemodilution (ANH), have been published previously,17 but generally, EPO is given as 40,000 units (approximately 600 units/kg) subcutaneously weekly for 3–4 weeks as needed to achieve goal hemoglobin (Hb) >14 g/dL. If patients have adequate iron stores (ferritin >100 ng/mL and iron saturation >20%), they are prescribed oral iron. Patients with inadequate iron stores to support erythropoiesis or with intolerance to oral iron are administered intravenous iron.

Outcomes were assessed by retrospective chart review. The primary outcome was a composite of 30-day mortality or thrombotic event (myocardial infarction, stroke, or pulmonary embolism). Secondary outcomes assessed include change in Hb and creatinine, acute kidney injury (AKI), sternal wound infection, atrial fibrillation, myocardial infarction, pulmonary embolism, pneumonia, and time to extubation, in addition to intensive care unit and hospital length of stay between the 2 groups. AKI was defined by the Kidney Disease Improving Global Outcomes classification, modified due to lack of urine output data. The presence of atrial fibrillation was defined as description of new-onset atrial fibrillation occurring perioperatively as described in the discharge summary.

Back to Top | Article Outline

Statistical Analysis

Baseline characteristics, such as the European System for Cardiac Operative Risk Evaluation (EuroSCORE) and preoperative lab values, were compared between study cohorts to assess remaining imbalance of potential confounders following cohort matching. We used t tests or Wilcoxon rank sum tests to compare numeric factors between groups, and χ2 or Fisher exact tests to compare categorical factors. Univariate and multivariable generalized estimating equations models were used to study the association between the treated and control groups and patient outcomes, accounting for the matching. When the event rate allowed, multivariable models included terms for EuroSCORE and any covariates found to be unbalanced between the study cohorts and associated with outcome (and thus potentially confounding). Analysis was performed using SAS v.9.4 (SAS Inc, Cary, NC), and statistical significance was assessed at the .05 level.

Back to Top | Article Outline


During the study interval, 58 JW patients who decline transfusion and received EPO perioperatively for cardiovascular surgery were identified. Five patients were then excluded, due to inability to find suitable control matches, leaving a study cohort of 53 JW patients, who were matched to 106 control patients. Of these 53 patients, 43 (81.1%) received EPO preoperatively, 3 (5.7%) received EPO in the postoperative period, and 7 (13.2%) received EPO both preoperatively and postoperatively. The surgeries included coronary artery bypass grafting, valve procedures, combined grafting and valve surgeries, heart transplantations, and thoracic aortic surgery. All patients were operated on by 1 of 6 surgeons, and matched treated and control patients underwent the same procedure by the same surgeon. Preoperative and intraoperative patient characteristics were similar between groups, except preoperative Hb (13.91 vs 13.31 g/dL in EPO versus non-EPO groups, respectively, P = .02) and race (P < .01, Table 1).

Table 1.

Table 1.

Table 2.

Table 2.

Table 3.

Table 3.

Hb at time of initial referral to the Center for Blood Conservation ranged from 10.1 to 16.1 g/dL in the EPO group, with a mean (standard deviation) of 12.8 (1.3). Most patients were treated with oral iron, but 15 patients (28.3%) were treated with intravenous iron. Hb increased a mean (standard deviation) of 1.2 g/dL (1.4) in patients treated with EPO preoperatively. After adjustment for preoperative Hb, nadir Hb during admission was significantly higher in the EPO group than the control group (Table 3; P < .0001).

Cell salvage is set up routinely for cardiac surgery and thus use should not be different between the 2 groups; however, ANH is only performed for patients who do not accept transfusion, and was performed in 43 patients (83%) in the EPO cohort, with a median [Q1, Q3] harvest volume of 1350 mL [1000, 1500]. Additionally, 10 patients (18.9%) received albumin, considered a minor fraction, in the EPO group.

In-hospital mortality was zero in both groups, and there was no difference in the primary outcome, composite of death, or thrombotic event (P = .12, Table 2). The low event rate of the primary outcome (4 total events) prevented the investigation of group differences in a multivariable model; however, the residual confounding left after matching was minimal, and the univariate results indicated no evidence of a significant difference. The 2 groups differed in postoperative Hb change, with the EPO group having a lower Hb loss from baseline to discharge than the non-EPO group (−2.65 vs −3.60; P = .001). Despite having similar preoperative creatinine, the EPO group had a greater creatinine increase from baseline to discharge (0.06 vs −0.04; P = .02), but the incidence of AKI was similar (47.17% vs 41.51%; P = .49). None of the categorical outcomes (incidence of myocardial infarction, stroke, wound infection, pneumonia, pulmonary embolism, or atrial fibrillation) was significantly different between the groups. There were no differences in time to extubation, hospital length of stay, or intensive care unit stay (Table 2). Multivariable generalized estimating equations modeling, adjusting for EuroSCORE (and race when indicated), found similar associations between treatment and change in Hb, nadir Hb, and creatinine, but no association with any other outcome studied (Table 3).

Back to Top | Article Outline


In this study, we found that patients who decline transfusion and are treated with EPO undergoing cardiovascular surgery have similar postoperative outcomes compared to control patients not treated with EPO who did not receive allogeneic blood transfusion. This finding is important, as it shows that EPO may be a safe and feasible preoperative treatment for anemia in patients undergoing cardiovascular surgery when blood transfusion is not an option.

Although cardiac surgery in JW patients has been performed for decades, advances in pharmacotherapies have made it safer than ever. The introduction of EPO to the market in 1989 brought about new hope in combatting anemia, however, from the very start several safety concerns were raised.18 These concerns, which were reported mainly in the chronic renal failure population, included increased risk of myocardial infarction, stroke, death, and thromboembolic events.18,19 In 2007, the FDA responded by placing a black box warning on EPO, which states that the EPO-stimulating agents increase the risk of death, myocardial infarction, stroke, venous thromboembolism, thrombosis of vascular access, and tumor progression or recurrence,20 and use remains off-label in cardiac and vascular surgery patients due to concerns over risk of thromboembolism.

In both animal studies21 and healthy volunteers,22 even short-term EPO administration increased the overall platelet count and platelet activation. EPO causes an increase in circulating bone marrow progenitor cells, including megakaryocytic progenitor cells,23 and improves platelet function, as indicated by shortened bleeding times and improved platelet aggregation, in patients with uremia.24 Although a study in healthy human volunteers showed no increase in reactivity of platelets 1 hour after EPO infusion, suggesting no effect on the thrombogenicity of circulating platelets, platelet reactivity subsequently increased and reached a peak at day 9 postinfusion, pointing to a thrombogenic effect on the newly synthesized platelets.22 In addition to increased platelet number, there was enhanced reactivity and activation of endothelial cells as evidenced by increased plasma P-selectin and E-selectin.22 This increase in platelet number and reactivity could account for increased thromboembolic effects of EPO. In our study population receiving EPO, 29 of 53 patients were on aspirin therapy preoperatively, and 3 patients were on clopidogrel, although this was held before surgery. Of the patients having a thrombotic complication, 1 patient who had a pulmonary embolus had been on aspirin 81 mg preoperatively, and 1 patient who had a thrombotic stroke had been on aspirin/extended-release dipyradimole (25/200 mg), although it is unclear when it had been stopped. Further studies are needed in the cardiac surgery population to test whether the use of antiplatelet agents could decrease the risk of thrombotic events.

EPO is used commonly in dialysis-dependent patients; however, it remains unclear whether results from studies of patients in chronic renal failure can be extrapolated to the general population, and more specifically those undergoing cardiovascular surgery. There are several key differences between these 2 populations. First, and the most obvious, the baseline health and functional status differ between elective or semielective cardiac surgery patients and those who are dialysis dependent. Second, the dosing regimens vary between short-term use during the perioperative experience and chronic use. Finally, the higher Hb levels achieved with EPO therapy are not sustained over a longer time period due to the acute blood loss that accompanies surgery.13 Despite these concerns, the health care team is faced with a dilemma when offering cardiovascular surgical services to patients of JW faith. Several centers have published their experience with EPO in the JW population with the goal of improving anemia. They concluded that bloodless cardiac surgery could be safely performed with acceptable surgical outcomes25–30 and without increasing cost.31 Our study results support that view.

This study is the first to compare JW patients who received EPO to a cohort that received neither EPO nor blood products, eliminating concerns of complications arising from receipt of blood product transfusion. Despite this, there are several limitations to this study. This is a retrospective study, and selection bias may exist with regards to which patients were offered surgery; the population may not reflect all JW patients with cardiovascular disease who needed cardiac surgery. Although we were limited to the number of patients included in our database who had undergone cardiac surgery and received perioperative EPO, and thus not powered to detect small differences in outcomes, we used a 2:1 match to increase our control cohort to help overcome this limitation. Additionally, most of the JW patients included in this study accepted ANH, cell salvage, and minor blood fractions such as albumin and cryoprecipitate, as these are commonly accepted in the local JW community. Conversely, while cell salvage and minor blood fractions (albumin) are routinely used in all cardiac surgery patients, intraoperative ANH was not provided to the control cohort. Furthermore, by matching JWs to a cohort that was not transfused, it is likely that the control group was healthier and at lower risk despite similar EuroSCOREs. In particular, those patients receiving EPO postoperatively due to significant postoperative anemia would likely have been transfused had they accepted blood products. As any controls that were transfused were excluded from our study, only those patients without substantial blood loss to require a transfusion were left to match to our EPO group. Thus, there could be significant clinical differences in intraoperative and postoperative treatment of the 2 groups, leading to the decision to use EPO in 1 group, versus the decision to not transfuse in the other. This is a limitation of the design of the study, but as there is no ethical way to deny transfusion to those that would accept it, this was the best way to compare outcomes in those receiving EPO and not transfused to controls. So although the observed rate of the primary outcome was higher in the EPO group, and the lack of statistical significance may have been due to type II error, it may also be the case that the elevated event rate was due to the fact that the sample selection of our control population produced a comparison group that was at a lower risk of the primary outcome to begin with, and that if we were able to study a more risk-comparable group, the event rate difference would be further diminished.

Back to Top | Article Outline


The results of this retrospective matched cohort study support the hypothesis that patients that decline transfusion undergoing cardiovascular surgery and receiving EPO to treat anemia do not have clinically significant differences in outcomes compared to a matched cohort not receiving EPO or blood products. More specifically, as the first study of its kind comparing JW patients who received EPO with a matched cohort who did not receive EPO or blood products, we found no evidence that there is a difference in mortality, myocardial infarction, stroke, or venous thromboembolism. Future prospective multicenter studies with a larger number of patients are needed to confirm these findings.

Back to Top | Article Outline


Name: Lorent Duce, MD.

Contribution: This author helped conceive and design the study, collect, analyze, and interpret the data, and draft the article.

Name: Mary L. Cooter, MS.

Contribution: This author helped design the study, analyze, and interpret the data, and draft the article.

Name: Sharon L. McCartney, MD.

Contribution: This author helped interpret the data and draft the article.

Name: Frederick W. Lombard, MD.

Contribution: This author helped interpret the data and draft the article.

Name: Nicole R. Guinn, MD.

Contribution: This author helped conceive and design the study, collect, analyze, and interpret the data, and draft the article.

This manuscript was handled by: Marisa B. Marques, MD.

Back to Top | Article Outline


1. Watch Tower Bible & Tract Society. Keep Yourself in God’s Love. 2014:Wallkill, NY: Watch Tower Bible and Tract Society of Pennsylvania Publishers; 215–218.
2. McQuilten ZK, Andrianopoulos N, Wood EM, et al. Transfusion practice varies widely in cardiac surgery: results from a national registry. J Thorac Cardiovasc Surg. 2014;147:1684–1690.e1.
3. Musallam KM, Tamim HM, Richards T, et al. Preoperative anaemia and postoperative outcomes in non-cardiac surgery: a retrospective cohort study. Lancet. 2011;378:1396–1407.
4. Musallam KM, Jamali FR, Rosendaal FR, et al. Preoperative hematocrit concentration and the risk of stroke in patients undergoing isolated coronary-artery bypass grafting. Anemia. 2013;2013:206829.
5. Bhaskar B, Dulhunty J, Mullany DV, Fraser JF. Impact of blood product transfusion on short and long-term survival after cardiac surgery: more evidence. Ann Thorac Surg. 2012;94:460–467.
6. Vamvakas EC, Blajchman MA. Transfusion-related mortality: the ongoing risks of allogeneic blood transfusion and the available strategies for their prevention. Blood. 2009;113:3406–3417.
7. Shaw RE, Johnson CK, Ferrari G, et al. Blood transfusion in cardiac surgery does increase the risk of 5-year mortality: results from a contemporary series of 1714 propensity-matched patients. Transfusion. 2014;54:1106–1113.
8. Galas FR, Almeida JP, Fukushima JT, et al. Blood transfusion in cardiac surgery is a risk factor for increased hospital length of stay in adult patients. J Cardiothorac Surg. 2013;8:54.
9. Patient Blood Management Programs. Society for the Advancement of Blood Management web site. Available at: Accessed March 20, 2017.
10. Memorial Commemoration. 2016 Yearbook of Jehovah’s Witnesses. 2016:Wallkill, NY: Watchtower Bible And Tract Society Of New York, Inc; 117. Available at: Accessed March 20, 2017.
11. Ott DA, Cooley DA. Cardiovascular surgery in Jehovah’s Witnesses. Report of 542 operations without blood transfusion. JAMA. 1977;238:1256–1258.
12. Jelkmann W. Regulation of erythropoietin production. J Physiol. 2011;589:1251–1258.
13. Tran DH, Wong GT, Chee YE, Irwin MG. Effectiveness and safety of erythropoiesis-stimulating agent use in the perioperative period. Expert Opin Biol Ther. 2014;14:51–61.
14. McKoy JM, Stonecash RE, Cournoyer D, et al. Epoetin-associated pure red cell aplasia: past, present, and future considerations. Transfusion. 2008;48:1754–1762.
15. US Food and Drug Administration. PROCRIT® (Epoetin alfa) FOR INJECTION. Available at: Accessed March 20, 2017.
17. McCartney S, Guinn N, Roberson R, Broomer B, White W, Hill S. Jehovah’s Witnesses and cardiac surgery: a single institution’s experience. Transfusion. 2014;54:2745–2752.
18. Besarab A, Bolton WK, Browne JK, et al. The effects of normal as compared with low hematocrit values in patients with cardiac disease who are receiving hemodialysis and epoetin. N Engl J Med. 1998;339:584–590.
19. 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.
20. US Food and Drug Administration. Information for Healthcare Professionals: Erythropoiesis Stimulating Agents (ESA) [Aranesp (darbepoetin), Epogen (epoetin alfa), and Procrit (epoetin alfa)]. Available at: Accessed March 20, 2017.
21. Berridge MV, Fraser JK, Carter JM, Lin FK. Effects of recombinant human erythropoietin on megakaryocytes and on platelet production in the rat. Blood. 1988;72:970–977.
22. Stohlawetz PJ, Dzirlo L, Hergovich N, et al. Effects of erythropoietin on platelet reactivity and thrombopoiesis in humans. Blood. 2000;95:2983–2989.
23. Beguin Y. Erythropoietin and platelet production. Haematologica. 1999;84:541–547.
24. Cases A, Escolar G, Reverter JC, et al. Recombinant human erythropoietin treatment improves platelet function in uremic patients. Kidney Int. 1992;42:668–672.
25. Marinakis S, Van der Linden P, Tortora R, Massaut J, Pierrakos C, Wauthy P. Outcomes from cardiac surgery in Jehovah’s witness patients: experience over twenty-one years. J Cardiothorac Surg. 2016;11:67.
26. Yoo YC, Shim JK, Kim JC, Jo YY, Lee JH, Kwak YL. Effect of single recombinant human erythropoietin injection on transfusion requirements in preoperatively anemic patients undergoing valvular heart surgery. Anesthesiology. 2011;115:929–937.
27. Cladellas M, Farré N, Comín-Colet J, et al. Effects of preoperative intravenous erythropoietin plus iron on outcome in anemic patients after cardiac valve replacement. Am J Cardiol. 2012;110:1021–1026.
28. Lin DM, Lin ES, Tran MH. Efficacy and safety of erythropoietin and intravenous iron in perioperative blood management: a systematic review. Transfus Med Rev. 2013;27:221–234.
29. D’Ambra MN, Gray RJ, Hillman R, et al. Effect of recombinant human erythropoietin on transfusion risk in coronary bypass patients. Ann Thorac Surg. 1997;64:1686–1693.
30. Pattakos G, Koch CG, Brizzio ME, et al. Outcome of patients who refuse transfusion after cardiac surgery: a natural experiment with severe blood conservation. Arch Intern Med. 2012;172:1154–1160.
31. Guinn NR, Roberson RS, White W, et al. Costs and outcomes after cardiac surgery in patients refusing transfusion compared with those who do not: a case-matched study. Transfusion. 2015;55:2791–2798.
Copyright © 2017 International Anesthesia Research Society