Anemia: Perioperative Risk and Treatment Opportunity

The World Health Organization (Geneva, Switzerland) has defined anemia as a hemoglobin concentration less than 13 g · dl^–1 in men and 12 g · dl^–1 in women.¹⁴  The incidence of erythrocyte transfusion has been demonstrated to increase below a preoperative hemoglobin concentration of 13 g · dl^–1.¹⁶  In a recent review by Warner et al., the authors suggest utilizing a definition of a hemoglobin concentration less than 13 g · dl^–1 for both sexes¹⁵  given that anemia increases the risk of erythrocyte transfusion and associated adverse outcomes in perioperative patients.^14,16,17 

In an attempt to better classify the risk of anemia exposure, clinical studies have focused on the impact of (1) preoperative anemia (chronic or preexisting anemia),^3–6  (2) acute intraoperative anemia^7–10,18,19  (usually associated with acute blood loss and fluid resuscitation), and (3) postoperative anemia,^11–13  which is strongly influenced by the combined impact of preoperative and intraoperative anemia. Studies have also assessed the risk of anemia in patients undergoing both noncardiac and cardiac surgery. Evidence supports that the relative percent change in hemoglobin concentration, or “delta hemoglobin concentration,” is also an important predictor of adverse outcomes.^20,21  Finally, physiologic adaptations to chronic anemia, including an increase in 2,3-diphosphoglycerate, among other mechanisms, may be protective, as the risk of mortality associated with chronic anemia appears to be lower than for patients with acute anemia.^14,22 

Understanding the mechanism(s) of organ injury and death associated with anemia is central to developing, and applying, effective treatments in an attempt to prevent these adverse outcomes. For the purposes of this review, we will evaluate the evidence assessing whether anemia, of any timing or etiology, presents an increased risk for adverse outcomes. Evidence that a common mechanism, such as limited oxygen delivery to tissue (i.e., tissue hypoxia), may contribute to adverse outcomes in anemic patients has been supported by data derived from models of experimental anemia.^23–27  These studies support the hypothesis that anemia-induced tissue hypoxia may contribute to the pathophysiology associated with clinical anemia (fig. 1).²⁸  Identifying such a unifying mechanism would help to characterize the overlapping risk that appears to be associated with all types of anemia. This approach includes the assessment that anemia may represent an independent causal risk factor for organ injury and mortality, but does not exclude the possibility that anemia may also be a “biomarker” of severity of patient comorbidity, which could also influence patient outcome.¹⁴ 

Data from large retrospective databases and prospective cohort studies demonstrate that about 30% of patients presenting for cardiac and noncardiac surgery are anemic.^3,4,6,17,29  In these studies, both “chronic” preoperative and “acute” intraoperative anemia have been associated with increased organ injury and mortality. The presence of chronic preoperative and acute intraoperative anemia is a strong predictor of postoperative mortality after both cardiac^4,5,8,29  and noncardiac surgery.^3,6,12,13,17  When considering the mechanism of organ injury, the impact of acute anemic-tissue hypoxia (cellular energetic failure) and the associated responses including renal tissue hypoxia, increased cardiac output, and myocardial work, and the increase in cerebral blood flow associated with anemia may all contribute to optimized organ perfusion but also may contribute to the observed phenotypes of organ injury and increased mortality (fig. 1).

Analyses of large retrospective databases have identified that preoperative anemia is associated with an increased risk of acute kidney injury in patients undergoing cardiac and noncardiac surgery.^4,5,17  Acute hemodilution on cardiopulmonary bypass has also been associated with acute kidney injury and dialysis dependence after surgery in a manner that is proportional to the degree of acute hemodilution.^8,18 

Preoperative and acute intraoperative anemia have been associated with stroke in a manner that is proportional to the degree of preoperative anemia and the nadir hemoglobin concentration that occurred during surgery.^4,5,11,19 

Analysis of the impact of acute intraoperative anemia has demonstrated an increased risk of myocardial infarction with decreasing hemoglobin levels.^10,22  More recent retrospective analyses of data from large prospective randomized controlled trials and cohort studies have identified that intraoperative and postoperative anemia are both associated with elevated troponin levels, myocardial injury. and mortality after cardiac^5,8  and noncardiac surgery.^6,10,12 

Analysis of large clinical studies provides some insights into the magnitude of risk associated with anemia. A retrospective analysis of 227,425 patients undergoing noncardiac surgery, based on data from the American College of Surgeons (Chicago, Illinois) National Surgical Quality Improvement Program database, demonstrated that 69,229 (approximately 30%) of these patients were anemic.¹⁷  Patients with anemia had an increased adjusted odds ratio for mortality (odds ratio, 1.42 [95% CI, 1.31 to 1.54]), cardiac injury (odds ratio, 1.45 [95% CI, 1.29 to 1.65]), respiratory dysfunction (odds ratio, 1.33 [95% CI, 1.26 to 1.41]), and renal injury (odds ratio, 1.37 [95% CI, 1.23 to 1.53]).¹⁷  Results from the International Surgical Outcomes Study database demonstrated that 11,675 (approximately 30%) of 38,770 patients had anemia. Patients diagnosed with moderate anemia (8.0 to 10.9 g · dl^–1) had an increased risk of in hospital mortality (odds ratio, 2.70 [95% CI, 1.88 to 3.87]) and myocardial infarction (odds ratio, 1.58 [95% CI, 1.00 to 2.49]).³  Anemia also increased risk of mortality and morbidity after cardiac surgery.^4,5,29  Klein et al. described an analysis of 19,033 patients undergoing cardiac surgery of whom 5,895 (31%) had a diagnosis of anemia. Within this cohort, preoperative anemia was associated with increased mortality (odds ratio, 1.42 [95% CI, 1.18 to 1.71]).²⁹  Karkouti et al. published similar data indicating that about 26% of 3,500 patients with anemia experienced an increased risk of a composite outcome of in-hospital death, stroke, and acute kidney injury (odds ratio, 1.8 [95% CI, 1.2 to 2.7]).⁴  Studies that stratify the degree of anemia demonstrate that the magnitude of the risk tends to progress with the severity of anemia from mild to moderate and severe.^3–5,17 

The relationship between the degree of acute anemia and risk of adverse outcomes has been clearly described in (1) patients undergoing noncardiac surgery who refuse blood transfusion^{7,9,14,22,30,31}  and (2) anemic patients who experience acute hemodilution on cardiopulmonary bypass.^4,5  For patients who refuse erythrocyte transfusion,^22,30,31  the severity of acute anemia predicts mortality such that each 1 g · dl^–1 decrease in hemoglobin concentration predicts an increased risk of mortality (odds ratio, 1.55 [95% CI, 1.25 to 1.91]).^7,22  The estimated hemoglobin concentration at which 50% mortality occurs is near 3 g · dl^–1, coincident with the hemoglobin concentration below which oxygen metabolism becomes supply-dependent.³¹  Time to death is accelerated in patients who experience critically low hemoglobin levels,⁷  and 100% mortality is predicted for patients with extremely low hemoglobin levels.^7,9,14,31 

In patients with severe acute anemia, treatment with hemoglobin-based oxygen carriers has been attempted to offset the high risk of mortality at very low hemoglobin levels.^9,31  Under these conditions, treatment with hemoglobin-based oxygen carriers may reduce the incidence of mortality (hazard ratio, 0.42 [95% CI, 0.25 to 0.86]).^9,31  However, these agents have not received regulatory approval in North America, and the potential benefits of utilizing hemoglobin-based oxygen carriers to treat severe life-threatening anemia must be weighed against data from clinical trials suggesting that some hemoglobin-based oxygen carriers may increase the risk of cardiac injury and mortality in less severely anemic patients.³² 

Further evidence supporting a possible causal link between anemia, organ injury, and mortality is provided by analyses of the impact of acute hemodilutional anemia in surgical patients undergoing acute hemodilution during cardiopulmonary bypass. In patients undergoing cardiac surgery, the lowest hemoglobin concentration or hematocrit on bypass has been associated acute kidney injury,^8,18  myocardial infarction,⁸  and stroke.¹⁹  In each case, the incidence of injury is proportional to the degree of acute anemia. Thus, a relationship has been defined between the degree of acute anemia, acute kidney injury, stroke and myocardial injury, and mortality, raising the possibility that severe acute anemia limits tissue oxygen delivery and results in organ injury and mortality (fig. 1).

Preoperative and intraoperative anemia are among the strongest risk factors for intraoperative transfusion.^16,29,33  Thus, the risks of anemia and erythrocyte transfusion are tightly linked or “intertwined.”^14,34  Since both anemia and transfusion have been associated with adverse outcomes, it remains difficult to isolate the attributable risk from each exposure.³⁴  This combined risk strongly emphasizes the potential benefit of effective patient blood management and treatment of preoperative anemia to avoid the potential risk of both anemia and transfusion.^14,15,35 

Experimental studies that have investigated the physiologic adaptation of mammals to acute anemia have consistently demonstrated that the oxygen-sensing mechanisms involved are extremely sensitive, complex, and redundant, emphasizing the importance of maintaining adequate oxygen delivery to tissue to maintain survival.^14,24,36,37  These mechanisms are capable of detecting early changes in arterial oxygen content (Cao₂) and then amplifying integrated adaptive responses in a manner that is proportional to the severity reduced Cao₂.^{23,25,26,38,39}  This assumes that effective mechanisms exist to (1) sense changes in Cao₂, (2) generate afferent signals to inform central and/or peripheral processing centers that tissue hypoxia is occurring, and (3) generate efferent responses to activate cardiovascular responses that increase cardiac output, vital organ blood flow, and tissue oxygen delivery to preserve vital organ perfusion, tissue oxygen homeostasis, and cellular function (fig. 1).^{24,26,27,38,39} 

One of the primary cellular responses to inadequate tissue oxygen delivery is to enhance or activate adaptive hypoxic cellular mechanism(s) including those derived from hypoxia-inducible factor, the master genetic regulator of hypoxic cellular responses.³⁷  The importance of oxygen sensing at the cellular level is supported by the ubiquitous presence of hypoxia-inducible factor, and other oxygen sensing mechanisms (mitochondria), that are present in every cell in the human body (fig. 1).^36,37  The importance of hypoxia-inducible factor in maintaining organism homeostasis and survival during acute anemia is demonstrated by the finding that when this system is disrupted, an increase in anemia-induced mortality is observed.²⁶  These translational studies support the hypothesis that hypoxic cellular responses are initially adaptive and indicate the presence of anemia-induced tissue hypoxia. Further clinical studies are required to determine if these “hypoxic biomarkers” can be measured in patients and utilized to predict adverse outcomes, and direct treatment to improve patient care.²⁸ 

The important role of patient blood management as a means of avoiding the negative impact of both anemia and erythrocyte transfusion has been strongly emphasized.^14–16  Prospectively collected data have demonstrated that integrated blood management programs are associated with reduced erythrocyte transfusion and improved patient outcomes.^15,16,35  Published reviews have clearly outlined the diagnostic criteria for iron deficiency and iron restrictive anemia as well as the specific algorithms for treatment, including oral and iv iron and erythroid stimulating agents (table 1).^15,35,40  A recently published network meta-analysis identified that patient blood management resulted in reduction in risk of erythrocyte transfusion (risk ratio, 0.60 [95% CI, 0.57 to 0.63; I² = 77%]) and favored shorter intensive care and hospital length of stay but did not influence mortality or other adverse outcomes.³⁵  These findings support the approach to avoid “anemia neglect” and to apply appropriate resources and initiatives to treat anemia preoperatively, and not to rely entirely on erythrocyte transfusion as a default treatment.¹⁵  The approach to optimal patient blood management requires an appreciation of the potential harm associated with both anemia and erythrocyte transfusion^14,34  as well as allocation of adequate resources to enable effective treatments. To successfully achieve optimal patient blood management practices, clinicians and investigators have assessed different approaches, and timelines, to treating preoperative anemia (fig. 2; table 1).

While oral iron therapy has long been the mainstay of treatment of iron-deficient anemia, the duration of therapy, low bioavailability, and lack of tolerance to oral iron often make this approach inadequate for rapid preoperative treatment of iron-deficient anemia. This has promoted the adoption of intravenous iron as a means of more rapidly replenishing iron stores and improving hemoglobin levels. A recent trial assessing efficacy of iv versus oral iron therapy in blood donors demonstrated that a single dose of iv iron was more effective than oral iron at increasing iron stores as reflected by ferritin levels (median, 105 [interquartile range, 75 to 145] vs. 25 [17 to 34] ng/ml).⁴⁷  In support of this finding, a systematic review has reported evidence that supports that iv iron was more effective than oral iron at reducing erythrocyte transfusion in surgical patients.⁴⁸  The recently published Preoperative Intravenous Iron to Treat Anemia Before Major Abdominal Surgery (PREVENTT) trial assessed the impact of iv iron on anemic patients after abdominal surgery.⁴⁹  While no difference in erythrocyte transfusion was observed, patients in the treatment arm had higher postoperative hemoglobin levels (at 8 weeks and 6 months) and had significantly fewer readmissions to the hospital (6 months).^49,50  Limitations of this trial include the lack of diagnosis of iron deficiency (anemia etiology) and the relatively short duration between treatment and surgery in patients treated with iv iron, which may have negatively impacted the efficacy of iv iron therapy.⁵⁰  The synthesis of these data and studies suggests that iv iron may be a more effective means of restoring iron levels and improving patient outcomes, but that by itself, iv iron preoperatively may not readily reduce erythrocyte transfusion.

The efficacy of erythropoiesis-stimulating agents and iron versus iron alone have been assessed in two recent meta-analyses.^51,52  The data from these analyses support that addition of an erythropoiesis-stimulating agent to iron (oral or iv) is superior to iron therapy alone with respect to erythrocyte transfusion avoidance. Specifically, treatment with combined erythropoiesis-stimulating agent and iron therapy for patients undergoing orthopedic and cardiac surgical procedures resulted in a reduction in the risk of erythrocyte transfusion relative to iron therapy alone (odds ratio, 0.49 [95% CI, 0.32 to 0.76] and 0.51 [0.32 to 0.79], respectively).⁵²  These positive outcomes were not associated with an increased in thrombotic complications including deep vein thrombosis or pulmonary embolism, supporting that these approaches can be safe.^51,52  This clinical data must be balanced with the potential risks of erythropoiesis-stimulating agent therapy as outlined in the revised black box warning from the U.S. Food and Drug Administration (Silver Spring, Maryland), which warns that research has found a higher risk of death, heart disease, and stroke in those who took erythropoiesis-stimulating agents.⁵³ 

Given the challenges of arranging preoperative assessment and treatment and the frequent need for urgent surgery, many studies have assessed the impact of acute treatment of anemia at the time of admission for surgery. Support for a combined pharmaceutical approach including an erythropoiesis-stimulating agent plus nutritional supplementation (iron, B₁₂, folate) is provided by the data from a recent randomized controlled trial in patients undergoing cardiac surgery.⁵⁴  The authors of this study demonstrate that acute therapy near the time of surgery reduced the incidence of erythrocyte transfusion in this patient population at high risk for transfusion (erythrocyte transfusion, median, 0 [interquartile range, 0 to 2] vs. 1 [interquartile range, 0 to 3] units), and treated patients had a higher reticulocyte count and reticulocyte hemoglobin content at 7 days.⁵⁴  A subanalysis of available studies comparing the impact of anemia treatment just before the time of surgery, which include one dose of an erythropoiesis-stimulating agent, suggests that this approach may reduce erythrocyte transfusion (relative risk, 0.80 [95% CI, 0.57 to 1.11]; I² = 80%; P = 0.18) and increased postoperative hemoglobin levels (mean difference, 0.53 [95% CI, –0.05 to 1.12] g · dl^–1; I² = 95%; P = 0.08). The overall small samples sizes and high degree of study heterogeneity suggest that additional adequately powered trials are required to fully assess the efficacy, and safety, of this approach (fig. 2).

While evidence of adverse events associated with erythrocyte transfusion supports continued efforts to promote patient blood management programs, the knowledge that severe acute anemia will lead to increased mortality^7,30  requires us to carefully balance the risks of anemia with the risks of transfusion.¹⁴  With a growing number of trials assessing restrictive versus liberal approaches to erythrocyte transfusion, a frequent analysis or recommendation remains that a restrictive transfusion strategy is noninferior to a more liberal strategy in terms of mortality and major morbidity.^55,56  However, a closer look at this overall approach is required as subanalysis of these studies supports the hypothesis that different transfusion strategies may be appropriate for patient populations who undergo different surgical procedures.⁵⁷  Indeed, recent guidelines for perioperative erythrocyte transfusion have supported the approach that different patient populations have evidence for different erythrocyte transfusion thresholds.⁵⁸  Other hypothesis generating data in support of defining erythrocyte transfusion thresholds in specific patient populations is derived from the Transfusion Requirements in Cardiac Surgery III (TRICS-III) study, the largest clinical trial assessing liberal versus restrictive transfusion in cardiac surgery.⁵⁹  With respect to the primary composite outcomes of death, myocardial infarction, stroke, and new-onset renal failure requiring dialysis at 6 months after surgery, there was strong suggestive evidence for effect modification by age, where a restrictive strategy was favored in patients 75 yr and older (n = 2,327; odds ratio, 0.77 [95% CI, 0.62 to 0.96]), and a liberal strategy was favored in patients younger than 75 yr (n = 2,337; odds ratio, 1.32 [95% CI, 1.07 to 1.64]; P_interaction = 0.001).⁵⁹  A study testing the hypothesis that liberal transfusion therapy may be superior to a restrictive strategy in younger patients undergoing cardiac surgery is currently underway (Transfusion Requirements in Younger Patients Undergoing Cardiac Surgery [TRICS-IV], NCT04754022).

Basic science research has demonstrated that anemia activates a number of profound cellular and physiologic adaptations including hypoxic cellular responses, such as hypoxia-inducible factor.^25,26,39  Hypoxia-inducible factor was discovered as the promotor of hypoxia-induced erythropoietin transcription.³⁷  As such, pharmacologically augmenting the hypoxia-inducible factor response in anemic patients may act to correct anemia and protect against hypoxic cellular injury.²⁶  This potential has led to the development of a number of pharmaceuticals targeting the inhibition of the enzyme hypoxia-inducible factor prolyl-hydroxylases, which are primarily responsible for initiating the degradation of hypoxia-inducible factor α. Inhibiting these enzymes serves to augmenting the hypoxia-inducible factor response, including promotion of erythrogenesis and inhibition of hepcidin, a molecule that limits iron transport.^60,61  A recent network meta-analysis assessed the efficacy of a number of oral hypoxia-inducible factor prolyl-hydroxylase inhibitors to determine if they are as effective as erythropoiesis-stimulating agent therapies at treating anemia.⁶²  Two of the largest prospective trials reported the efficacy of roxadustat versus placebo or recombinant erythropoietin, with respect to increasing hemoglobin concentration in anemic patients with renal dysfunction.^60,61  They observe that roxadustat increased hemoglobin levels (erythropoiesis) and reduced hepcidin (leads to increased iron transport) in a manner that was noninferior to recombinant erythropoietin.⁶⁰  The incidence of serious adverse events, 29 (14.2%) versus 10 (10%), suggests the need for future trials to assess safety and their potential role in the perioperative setting.⁶⁰ 

In summary, perioperative anemia is associated with adverse outcomes by complex and incompletely understood mechanisms. At severe levels of anemia, insufficient tissue oxygen delivery likely contributes to organ injury and increased mortality. Clinical and experimental studies have defined the strong adaptive hypoxic cellular responses to acute anemia, which in part explain the intrinsic mechanism(s) of adaptation to anemia, as well as potential mechanisms(s) of anemia-induced morbidity and mortality. Evidence of morbidity and mortality associated with all levels of anemia suggest that ongoing adequately powered clinical trials are needed to define treatment strategies that both improve anemia and prevent adverse clinical outcomes in anemic patients undergoing cardiac and noncardiac surgery.

The authors would like to acknowledge Richard B. Weiskopf, M.D., Professor Emeritus, University of California, San Francisco (San Francisco, California) for his related research and thoughtful review of an earlier version of this manuscript. All figures were created by Iryna Savinova, M.Sc., Department of Human Health and Nutritional Sciences, University of Guelph, Guelph, Ontario, Canada. Fig. 1 was designed using BioRender.com (Toronto, Ontario, Canada). The forest plot analysis assessing the impact of a single dose of erythropoiesis-stimulating agent was performed by Nikhil Mistry, M.Sc. (Institute of Medical Science, University of Toronto, Toronto, Ontario, Canada).

Drs. Hare and Mazer were supported by University of Toronto (Toronto, Ontario, Canada) Merit Awards.

Dr. Hare has received honoraria support from Pfizer (Kirkland, Quebec, Canada) and Johnson & Johnson (Markham, Ontario, Canada). Dr. Mazer has received advisory board honoraria from Amgen (Mississauga, Ontario, Canada), AstraZeneca (Mississauga, Ontario, Canada), Boehringer Ingelheim (Burlington, Ontario, Canada), and Octapharma (Toronto, Ontario, Canada).