Secondary Logo

Journal Logo

Postoperative Anemia: A Sign of Treatment Failure

Shander, Aryeh MD, FCCM, FCCP*†‡§; Roy, Raymond C. MD, PhD

doi: 10.1213/ANE.0000000000001272
Editorials: Editorial

From the *Department of Anesthesiology, Critical Care and Hyperbaric Medicine, Englewood Hospital and Medical Center, Englewood, New Jersey; Department of Anesthesiology, Medicine, and §Surgery, Icahn School of Medicine at Mount Sinai, New York, New York; and Department of Anesthesiology, Wake Forest School of Medicine, Winston-Salem, North Carolina.

Accepted for publication February 9, 2016.

Funding: None.

Conflict of Interest: See Disclosures at the end of the article.

Reprints will not be available from the authors.

Address correspondence to Aryeh Shander, MD, FCCM, FCCP, Department of Anesthesiology, Critical Care and Hyperbaric Medicine, Englewood Hospital and Medical Center, 350 Engle St., Englewood, NJ 07631. Address e-mail to

The circulatory and hematopoietic systems have evolved to bridge the gap between the outside world where resources are plentiful and the immediate vicinity of the cells. Any condition compromising this long and tightly regulated chain of supply can result in significant harm. Anemia, in which the oxygen-carrying capacity of blood and its rheological characteristics are compromised, is one such condition.1

The evidence of the harms of anemia is accumulating. Based on the data from over 39,000 patients prospectively collected by the European Surgical Outcomes Study, preoperative anemia is associated with postoperative mortality and morbidity that becomes progressively worse with the severity of the anemia.2 In this issue of Anesthesia & Analgesia, Choi et al.3 retrospectively observed that in a population of 2467 hip arthroplasties (with an average preoperative hemoglobin level of 12.7 g/dL), postoperative anemia (defined as a hemoglobin level <10 g/dL within the 48-hour postoperative period) was associated with a 2-fold increased risk of postoperative acute kidney injury (AKI). Given this risk, and the fact that as anesthesiologists, we have opportunities to address anemia as a part of the preoperative preparation or during intraoperative management, one has to ask whether we are failing to seize those opportunities.

To answer this question, we first need to understand anemia better. Nature does not necessarily comply with the human desire to categorize using often-arbitrary criteria. Anemia is a good example, as the search for the physiologic hemoglobin threshold to define anemia remains unfulfilled. Early in the 1940s and based on their anecdotal clinical experience, Adams and Lundy4 proposed that hemoglobin of 10 g/dL was optimal in the perioperative setting. Despite an absence of evidence, the medical community accepted this suggestion and used it extensively to make clinical decisions. In 1968, the World Health Organization (WHO) published its definition of anemia across genders and ages; namely hemoglobin <12 g/dL in nonpregnant women and hemoglobin <13 g/dL in men, all 15 years or older and all measured at sea level.5 Subsequently, the WHO further defined 3 levels of anemia according to hemoglobin concentrations as mild (hemoglobin 11–11.9 g/dL in adult females and 11–12.9 g/dL in adult males), moderate (hemoglobin 8–10.9 g/dL in adult females or males), and severe (hemoglobin <8 g/dL in either sex).5 These ranges and definitions are largely rooted in epidemiologic studies of the (normal) distribution of hemoglobin levels across various populations, with tails of the distribution histograms arbitrarily flagged as “abnormal.” Although this approach does well when large populations are considered,5,6 these epidemiologic definitions may not apply to individual patients.

Controversy exists among hematologists regarding where the “true” hemoglobin cutoff is for the definition of anemia. One simple reason is that the concentration of hemoglobin can fluctuate widely with differing levels of intravascular volume and within a fairly short period of time. A common scenario during surgery is when anesthetic agents lower blood pressure because of venodilation or cardiac depression and prompt the administration of crystalloid solutions that can acutely change hemoglobin concentrations via a dilutional mechanism. Moreover, methods used to measure hemoglobin concentration have an inherent margin of error, which can be as much as 1 g/dL.7 These revelations are not usually considered by many clinicians who may be relying on a single hemoglobin number to define anemia and, more importantly, to decide whether or not to give allogeneic blood.

A better alternative to hemoglobin (or hematocrit) measurements is to use the red cell mass (RCM). RCM is more indicative of the true level of available red cells for oxygen delivery.8 However, RCM measurement is cumbersome and not easily obtained in the clinical environment, leaving the clinician with no options other than reverting to hemoglobin as a surrogate for RCM, which it is not, or searching for other signs and symptoms to deduce whether to transfuse.

Another reason for the lack of consensus is that hemoglobin concentration itself is not a physiologic trigger for erythropoiesis. Rather, erythropoiesis is the end result of complex interactions among many factors including erythropoietin, whose synthesis in the kidney is triggered by hypoxemia, hepcidin, which is synthesized in the liver and controls iron homeostasis and angiotensin II, which serves as a growth factor for erythroid progenitor cells. Angiotensin II is blocked by angiotensin-converting enzyme inhibitors and angiotensin II receptor blockers commonly administered to treat hypertension. Inflammatory cytokines disrupt iron homeostasis by increasing hepcidin synthesis and decreasing responsiveness to erythropoietin. The end result is anemia of chronic disease, also known as anemia of inflammation.9 Thus, not only is hemoglobin an indirect measure of RCM but it also serves as a marker for chronic disease.

The perils of anemia are too many and important to ignore. When defining the levels of anemia from mild to severe, the WHO is quick to point out that even “mild anemia” is a misnomer. In the context of nutritional anemia, iron deficiency is already quite advanced when early signs of a mild anemia emerge, and iron deficiency, even in absence of anemia, is still harmful.5 Across various patient populations, anemia has been associated with increased risk of death, falls, fractures, cardiovascular events, hospital admission, prolonged hospital stay, readmission, poor mobility and physical performance, faster decline in cognitive abilities, declined quality of life, worsening of preexisting kidney disease, and the possibility of AKI.10

Moderate to severe AKI is a devastating condition that should and must be avoided. Patients undergoing complicated cardiac and noncardiac surgery are at higher risk of developing AKI, and those who develop AKI are at increased risk of other complications and death.11,12 Although work is ongoing to establish means of protecting the kidneys and reducing the risk of AKI during surgery,13,14 the incidence remains relatively high, affecting as many as 5% to 30% of patients in some populations.15–17 Anemia is considered a risk factor for AKI, and strategies to prevent and manage perioperative anemia may reduce the incidence of AKI.17

Development of AKI in the perioperative setting is multifactorial, and both patient- and procedure-related factors have been implicated.18 One considerable issue with the study by Choi et al. is their criteria for AKI, although they are the ones created by the Acute Kidney Network. The criteria focus on serum creatinine, a marker of (dys)function, not injury, and on an estimated glomerular filtration rate (GFR) calculated from the serum creatinine, again a measure of function and not necessarily injury. The association between anemia and organ dysfunction has long been established although we do not yet fully understand the complex interactions involved.1 Preclinical data (mostly from animal models) demonstrate the variable impact of anemia on different vital organs and strongly suggest that the kidneys are among the most vulnerable organs to ischemic injury when hemoglobin (more precisely, RCM) is reduced.19 One intriguing question is whether AKI represents irreversible organ damage or is it an indicator of compensatory mechanisms (some reversible and some irreversible) evolved as part of physiologic adaptations to a low oxygen environment.20 It is not rare, for example, for increased creatinine levels to resolve on their own within a few days after surgery.21 However, kidney injury can result in anemia because it leads to decreased erythropoietin synthesis and red cell life span. The cardiorenal anemia syndrome is an example of how these factors act together to reinforce their individual negative impacts and worsen patient outcomes.1,22

It is evident from clinical experience that patients can respond quite differently to the same level of anemia, some presenting with “full-blown” AKI, while others not experiencing any kidney injury at all. The reason can be related to the baseline risk profiles of the individual patients and differences in their ability to activate various compensatory mechanisms, ranging from grossly “mechanical” adaptations to increased oxygen delivery (e.g., increased cardiac output, oxygen extraction, and vasodilation), to subcellular and genetic adaptations regulated by factors such as hypoxia-inducible factors and nitric oxide synthase.23 Returning to the definition of anemia, it seems that we need to not only think about different hemoglobin cutoff points for various organs but perhaps set different cutoff points for each patient based on their specific conditions and profile.9

This variance in the impact of anemia on different organs and patients demonstrates a key limitation of the study by Choi et al.3 When studying the impact of anemia on AKI, why should we bind ourselves with an arbitrary hemoglobin threshold? And if we do so, what is the evidence behind using a hemoglobin level of 10 g/dL? Choi et al.’s use of a hemoglobin threshold to create cohorts for comparison is more a nod to convenience than a physiologically relevant parameter. One has to wonder how different the results would have been, if, say, thresholds of 12, 9, or 8 g/dL had been chosen. Of note, animal studies have suggested that the hemoglobin levels at which signs of hypoxia in the kidney emerge are much lower than 10 g/dL and the same is likely to be true for humans.24 Indeed, reports of severely anemic patients who cannot be transfused for various reasons have indicated few complications at hemoglobin levels >6 g/dL.25,26

In addition, Choi et al.3 used their hemoglobin threshold not only to construct the study arms but also to exclude patients who crossed the threshold during the postoperative period. This study thus missed an opportunity to further explore the association between AKI and various hemoglobin levels (as was done for various age groups in the study). By reducing hemoglobin to a dichotomous variable of anemia, the investigators sacrificed a wealth of data harbored below the arbitrary cutoff point. Is the risk of AKI in patients with postoperative hemoglobin of 7 to 8 g/dL similar to those with hemoglobin of 9 to 10 g/dL? Is it possible that a cluster of patients with lower postoperative hemoglobin levels are largely dominating the observed difference? These and many other questions are left unanswered.

When studying the relation between anemia and AKI, another factor to consider is the related consequence of blood transfusion, which independently increases the risk of kidney injury.27 In the study by Choi et al.,3 a hemoglobin trigger of 10 g/dL was used for making decisions regarding packed red blood cell (RBC) transfusions. Subsequently, almost half the patients in both study cohorts received RBC transfusions during the intraoperative or immediate postoperative period. None of the current transfusion guidelines supports RBC transfusions based on a hemoglobin trigger of 10 g/dL,28 and several studies have found a lack of benefit with a “liberal” transfusion strategy based on hemoglobin of 10 g/dL versus a more restrictive approach.3 Some evidence suggests that transfusion at a hemoglobin level of approximately 10 g/dL might even negatively impact oxygen delivery.29,30

The choice of such a high hemoglobin transfusion trigger thus limits the generalizability of the study of Choi et al. as do the relatively high transfusion rates in the study cohorts. Again, opportunities to further explore the potential interaction of transfusion and AKI were missed. The authors reported that transfusions (and the number of transfused RBC units) were not associated with AKI in univariate analysis, and the interaction term of transfusion and hemoglobin level was not identified as a significant predictor of AKI in multivariate analysis. Without much speculation, one might imagine that patients with postoperative hemoglobin >10 g/dL who were transfused were quite distinct from those who were not transfused. The former group likely reached hemoglobin levels <10 g/dL during surgery which alongside other factors triggered a decision to transfuse, whereas the latter group probably did not reach those that reached low hemoglobin levels. In other words, some of the transfused patients ended up in the nonanemic cohort, owing to intraoperative allogeneic blood transfusion. Was the pattern of AKI similar in these versus the nontransfused, nonanemic patients?

Another potential issue in the study of Choi et al. is the degree of baseline renal risk and other potential confounders such as hypotension and perioperative subcellular fluid use. Total volume of fluids and the individual types of fluids, colloids and crystalloids, have all been recognized as risk factors for kidney injury.31 One plausible scenario is that larger volume of IV fluid can result in organ and endothelial injury, while also causing a dilutional anemia and thus, making it seem as if anemia is associated with kidney injury while it is merely a byproduct of the real culprit.

The authors used preoperative serum creatinine levels and estimated GFR to group the patients, but the number of patients with reduced baseline GFR was unusually high for this population (almost 97%). This can be a typo, but the analysis performed with these subgroups is also vague. It appears that similar multivariate analyses were performed in the subgroups, but running a multivariate logistic regression on just 77 patients (with even fewer cases with AKI) may not yield reliable estimates.

Although these issues may affect some fine details in the interpretation of the study by Choi et al.,3 the big picture stays the same. Anemia (or reduced oxygen-carrying capacity of blood) is dangerous as it can increase the risk of AKI among many other untoward consequences. The key question to ask is then “What should be done?” For some, a logical response to a low hemoglobin level is RBC transfusion. After reading the study of Choi et al.,3 should we consider transfusion as a “protector” of the kidney from acute injury and give RBC as soon as hemoglobin dips <10 g/dL? We believe the answer is a resounding “No!” No data confirm that transfusion reduces or eliminates the risk of AKI, and as previously discussed, studies have indicated that allogeneic RBC transfusions may not be able to increase oxygen delivery as expected and are associated with AKI and other potential harm.1,29,30 So what are we left with? Anemia is bad for the kidneys and so is the traditional treatment of anemia, transfusion. Do we have other means to avert this conundrum? We believe the answer is a resounding “Yes!”

First and foremost, anemia is a modifiable risk factor. Despite being a global pandemic affecting >2 billion people,32 anemia is generally ignored by health care providers, as well as government health agencies both in the United States and in abroad.10 In the surgical arena, many incorrectly assume that anemia can be safely and effectively treated with a transfusion. Detecting, diagnosing, and treating anemia and its causes before surgery are the more sensible approach.33 Even in cases when not enough time is available to properly treat anemia, efforts can be made to minimize surgical blood loss and ensure that a patient’s limited supply of blood is not further compromised.34 Adoption of conservative and/or monitored fluid replacement during surgery and anesthesia can prevent unnecessary hemodilution and harmful fluid overload.31,35 Vigilant monitoring of blood pressure and the level of anesthesia to avoid hypotension and compromised perfusion can reduce the likelihood of AKI. To protect the kidneys, it is of utmost importance to maintain hemodynamic stability, without transfusion to the extent possible.36 These and other measures can be integrated as part of enhanced recovery after surgery protocols.

Ultimately, however, the choice of treatment for anemia must be dictated by the individual patient’s interests. Clinicians should not hesitate to transfuse when indicated such as during significant intraoperative hemorrhage. But for most patients, an allowable blood loss threshold exists that, if controlled, does not require transfusion.37

Clinicians may find it baffling that anemia is being touted as a serious condition with grave consequences, here AKI, which needs immediate attention, whereas simultaneously they are being asked to withhold the usual treatment of choice for anemia, transfusion, as it also carries risks. Even more confusing is when the question arises whether we are faced with full-blown AKI with long-term sequelae or a process that is either reversible or a marker for decreased function or reserve.21 Clinicians may still find themselves faced with these dilemmas on a daily basis: Should patients who are anemic preoperatively be: (a) rescheduled to allow time for treatments to increase hemoglobin; (b) transfused; or (c) anesthetized in their normovolemic anemia state? When anemic patients become hypotensive intraoperatively, should they be: (a) given hemodiluting nonblood fluids; (b) transfused; (c) given vasopressors or inotropes; or (d) given less anesthesia? When should an anemic patient be transfused and to what extent should the patient be involved in the decision?38 What is the risk/benefit for transfusion in a given patient with these comorbidities for this operation/surgeon? Relevant (and unanswered) questions regarding perioperative anemia do not end here.

With the current heart-warming trend toward more preventive measures, we think the better question to ask is: What can be done to reduce the risk of a patient presenting to the operating room with anemia? Timely screening for anemia during the weeks leading to an elective procedure can help clinicians avoid these last minute questions and not have to choose between bad choices and worse ones.33 The study of Choi et al.3 should not be taken as evidence to support transfusion at a given threshold but rather adds further support for the notion that anemia is not an “innocent bystander” and should not be left untreated.39 Armed with this understanding, it is good to remember that therapeutic options are available if preoperative anemia is identified early. In elective patients, preoperative anemia should be corrected at least to the extent that the allowable loss without transfusion is greater than the average blood loss expected for the surgical procedure. Under these controlled circumstances, postoperative anemia is less likely to occur and we are also less likely to fail our patients.

Back to Top | Article Outline


Name: Aryeh Shander, MD, FCCM, FCCP.

Contribution: This author helped prepare the manuscript.

Attestation: Aryeh Shander approved the final manuscript.

Conflicts of Interest: Aryeh Shander has been a consultant or speaker with honorarium for or received research support from Luitpold, Vifor Pharma, CSL Behring, Masimo, and HBO Therapeutic.

Name: Raymond C. Roy, MD, PhD.

Contribution: This author helped prepare the manuscript.

Attestation: Raymond C. Roy approved the final manuscript.

This manuscript was handled by: Avery Tung, MD.

Back to Top | Article Outline


The authors express their gratitude to Mazyar Javidroozi, MD, PhD, for support in preparation of the manuscript.

Back to Top | Article Outline


1. Shander A, Javidroozi M, Ozawa S, Hare GMWhat is really dangerous: anaemia or transfusion?Br J Anaesth2011107(suppl 1)i4159
2. Baron DM, Hochrieser H, Posch M, Metnitz B, Rhodes A, Moreno RP, Pearse RM, Metnitz P; European Surgical Outcomes Study (EuSOS) group for Trials Groups of European Society of Intensive Care Medicine; European Society of AnaesthesiologyEuropean Surgical Outcomes Study (EuSOS) group for Trials Groups of European Society of Intensive Care Medicine; European Society of AnaesthesiologyPreoperative anaemia is associated with poor clinical outcome in non-cardiac surgery patients.Br J Anaesth201411341623
3. Choi YJ, Kim S, Sim JH, Hahm KPostoperative anemia is associated with acute kidney injury in patients undergoing total hip replacement arthroplasty: a retrospective study.Anesth Analg201612219238
4. Adams RC, Lundy JSAnesthesia in cases of poor surgical risk. Some suggestions for decreasing the risk.Surg Gynecol Obstet19427410119
5. WHOHaemoglobin Concentrations for the Diagnosis of Anaemia and Asssessment of Severity. Vitamin and Mineral Nutrition Information System.2011Geneva, SwitzerlandWorld Health Organization
6. Khusun H, Yip R, Schultink W, Dillon DHWorld Health Organization hemoglobin cut-off points for the detection of anemia are valid for an Indonesian population.J Nutr1999129166974
7. Giraud B, Frasca D, Debaene B, Mimoz OComparison of haemoglobin measurement methods in the operating theatre.Br J Anaesth201311194654
8. Jacob M, Annaheim S, Boutellier U, Hinske C, Rehm M, Breymann C, Krafft AHaematocrit is invalid for estimating red cell volume: a prospective study in male volunteers.Blood Transfus2012104719
9. Boutou AK, Hopkinson NS, Polkey MIAnaemia in chronic obstructive pulmonary disease: an insight into its prevalence and pathophysiology.Clin Sci (Lond)201512828395
10. Shander A, Goodnough LT, Javidroozi M, Auerbach M, Carson J, Ershler WB, Ghiglione M, Glaspy J, Lew IIron deficiency anemia—bridging the knowledge and practice gap.Transfus Med Rev20142815666
11. Rewa O, Bagshaw SMAcute kidney injury-epidemiology, outcomes and economics.Nat Rev Nephrol201410193207
12. Pickering JW, James MT, Palmer SCAcute kidney injury and prognosis after cardiopulmonary bypass: a meta-analysis of cohort studies.Am J Kidney Dis20156528393
13. Dardashti A, Ederoth P, Algotsson L, Brondén B, Grins E, Larsson M, Nozohoor S, Zinko G, Bjursten HErythropoietin and protection of renal function in cardiac surgery (the EPRICS Trial).Anesthesiology201412158290
14. Tie HT, Luo MZ, Lin D, Zhang M, Wan JY, Wu QCErythropoietin administration for prevention of cardiac surgery-associated acute kidney injury: a meta-analysis of randomized controlled trials.Eur J Cardiothorac Surg201548329
15. Cho E, Kim SC, Kim MG, Jo SK, Cho WY, Kim HKThe incidence and risk factors of acute kidney injury after hepatobiliary surgery: a prospective observational study.BMC Nephrol201415169
16. Shaw A, Swaminathan M, Stafford-Smith MCardiac surgery-associated acute kidney injury: putting together the pieces of the puzzle.Nephron Physiol2008109p5560
17. Parida S, Badhe ASCardiac surgery-associated acute kidney injury.J Anesth20132743346
18. Mao H, Katz N, Ariyanon W, Blanca-Martos L, Adýbelli Z, Giuliani A, Danesi TH, Kim JC, Nayak A, Neri M, Virzi GM, Brocca A, Scalzotto E, Salvador L, Ronco CCardiac surgery-associated acute kidney injury.Blood Purif201437(suppl 2)3450
19. Lauscher P, Kertscho H, Schmidt O, Zimmermann R, Rosenberger P, Zacharowski K, Meier JDetermination of organ-specific anemia tolerance.Crit Care Med201341103745
20. Goldfarb M, Rosenberger C, Abassi Z, Shina A, Zilbersat F, Eckardt KU, Rosen S, Heyman SNAcute-on-chronic renal failure in the rat: functional compensation and hypoxia tolerance.Am J Nephrol2006262233
21. Endre ZHAcute kidney injury: definitions and new paradigms.Adv Chronic Kidney Dis20081521321
22. Lu KJ, Kearney LG, Hare DL, Ord M, Toia D, Jones E, Burrell LM, Srivastava PMCardiorenal anemia syndrome as a prognosticator for death in heart failure.Am J Cardiol2013111118791
23. Tsui AK, Marsden PA, Mazer CD, Adamson SL, Henkelman RM, Ho JJ, Wilson DF, Heximer SP, Connelly KA, Bolz SS, Lidington D, El-Beheiry MH, Dattani ND, Chen KM, Hare GMPriming of hypoxia-inducible factor by neuronal nitric oxide synthase is essential for adaptive responses to severe anemia.Proc Natl Acad Sci U S A2011108175449
24. Hare GMTolerance of anemia: understanding the adaptive physiological mechanisms which promote survival.Transfus Apher Sci201450102
25. Carson JL, Noveck H, Berlin JA, Gould SAMortality and morbidity in patients with very low postoperative Hb levels who decline blood transfusion.Transfusion2002428128
26. Shander A, Javidroozi M, Naqvi S, Aregbeyen O, Caylan M, Demir S, Juhl AAn update on mortality and morbidity in patients with very low postoperative hemoglobin levels who decline blood transfusion (CME).Transfusion201454268895
27. Karkouti KTransfusion and risk of acute kidney injury in cardiac surgery.Br J Anaesth2012109(suppl 1)i2938
28. Shander A, Gross I, Hill S, Javidroozi M, Sledge S; College of American Pathologists; American Society of Anesthesiologists; Society of Thoracic Surgeons and Society of Cardiovascular Anesthesiologists; Society of Critical Care Medicine; Italian Society of Transfusion Medicine and Immunohaematology; American Association of Blood BanksCollege of American Pathologists; American Society of Anesthesiologists; Society of Thoracic Surgeons and Society of Cardiovascular Anesthesiologists; Society of Critical Care Medicine; Italian Society of Transfusion Medicine and Immunohaematology; American Association of Blood BanksA new perspective on best transfusion practices.Blood Transfus201311193202
29. Schlager O, Gschwandtner ME, Willfort-Ehringer A, Kurz M, Mueller M, Koppensteiner R, Heinz GTranscutaneous oxygen tension monitoring in critically ill patients receiving packed red blood cells.J Crit Care201429105762
30. Tsai AG, Friesenecker B, McCarthy M, Sakai H, Intaglietta MPlasma viscosity regulates capillary perfusion during extreme hemodilution in hamster skinfold model.Am J Physiol1998275H217080
31. Prowle JR, Bellomo RFluid administration and the kidney.Curr Opin Crit Care2010163326
32. de Benoist B, McLean E, Egli I, Cogwell MWorldwide Prevalence of Anaemia 1993–2005: WHO Global Database on Anaemia2008Geneva, SwitzerlandWorld Health Organization
33. Goodnough LT, Maniatis A, Earnshaw P, Benoni G, Beris P, Bisbe E, Fergusson DA, Gombotz H, Habler O, Monk TG, Ozier Y, Slappendel R, Szpalski MDetection, evaluation, and management of preoperative anaemia in the elective orthopaedic surgical patient: NATA guidelines.Br J Anaesth20111061322
34. Shander A, Moskowitz DM, Javidroozi MBlood conservation in practice: an overview.Br J Hosp Med (Lond)2009701621
35. Gaffney AM, Sladen RNAcute kidney injury in cardiac surgery.Curr Opin Anaesthesiol201528509
36. Zarbock A, Milles KNovel therapy for renal protection.Curr Opin Anaesthesiol2015284318
37. Gross JBEstimating allowable blood loss: corrected for dilution.Anesthesiology19835827780
38. Toledo PShared decision-making and blood transfusions: is it time to Share More?Anesth Analg201411811513
39. Nissenson AR, Goodnough LT, Dubois RWAnemia: not just an innocent bystander?Arch Intern Med200316314004
Copyright © 2016 International Anesthesia Research Society