Share this article on:

Emergency Transfusion for Acute Severe Anemia: A Calculated Risk

Weiskopf, Richard B. MD

doi: 10.1213/ANE.0b013e3181f5ba2b
Editorials: Editorials
Chinese Language Editions

From the University of California, San Francisco, San Francisco, California (Professor Emeritus).

Address correspondence and reprint requests to Richard B. Weiskopf, MD, University of California, San Francisco, San Francisco, CA 94143-0648. Address e-mail to

Accepted July 27, 2010

An editorial about a case report is unusual, but no more so than the case reported by Dai et al.1 in this issue of the journal. They report survival, without apparent sequelae, of a 53-year-old man with a hemoglobin concentration as low as 0.7 g/dL (hematocrit 2.2%) for several hours. Despite the presence of a low fibrinogen concentration and an elevated partial thromboplastin time (not discussed here), successful management was facilitated by surgical control and minimal intraoperative blood loss. The reported nadir hemoglobin concentration may be the lowest known during acute anemia associated with survival, the previous apparently being 1.1 g/dL reported by Zollinger et al.2 This raises 2 important questions: (1) How is this possible? and (2) What alternatives are possible when crossmatched erythrocytes are not available?

It is not possible to know accurately the hemoglobin concentration that is associated with mortality or serious morbidity, because prospective experiments in humans with those end points are impossible and data from laboratory animals cannot be extrapolated to humans because of potentially important differences among species. Retrospective analyses of hospital databases3,4 and case reports5 regarding mortality associated with severe acute anemia have suggested that the median value is <5 g/dL. In a recent reexamination of those data, it was estimated that the median hemoglobin concentration associated with anemia-induced mortality is approximately 2.5 g/dL (R. B. Weiskopf, unpublished data, 2010). Cardiovascular disease increases that value3 (also R. B. Weiskopf, unpublished data, 2010). Retrospective examination of other databases relating preoperative hemoglobin concentration to postoperative mortality does not provide useful information because they do not separate those not transfused, nor are they able to account for the rationale for lack of transfusion or provide the hemoglobin concentration at death.68 Adequately powered prospective randomized clinical trials in adults9 and children10 in intensive care units have not found different mortality rates between those transfused with “restrictive” or “liberal” strategies that resulted in hemoglobin concentrations of approximately 8.5 and 10.5 g/dL.

Considering that the human mean fatal hemoglobin concentration is approximately 2.5 g/dL, and that this case report documents a hemoglobin concentration that is lower than previously known during acute anemia, it may seem difficult to understand survival. Nevertheless, there are data to assist in the explanation of this seemingly exceedingly improbable event. The patient's fraction of inspired oxygen (FIO2) was nearly 1.0 throughout the 12 hours of surgery. Similarly, in the case reported by Zollinger et al.,2 FIO2 was 1.0 and PaO2 >400 mm Hg at the time of the nadir hemoglobin concentration. Classic thought is that the amount of oxygen dissolved in plasma (the solubility of oxygen in plasma is 0.0031 mL/dL/mm Hg O2) is too little to be of physiologic consequence. Whereas that may be so during ordinary circumstances with an FIO2 of 0.21, dissolved oxygen can be of substantial benefit during severe anemia, when the FIO2 and PaO2 are high. Hyperoxia reduces mortality of pigs subjected to acute severe anemia and maintained at their critical hemoglobin concentration.11 In healthy humans, breathing oxygen reverses the neurocognitive deficits12 and increased P300 latency (the neurophysiologic correlate)13 induced by severe anemia. High PaO2 in healthy humans decreases the heart rate response to acute severe anemia12 (also J. Feiner, et al., unpublished data, 2010), and oxygen supplementation decreases heart rate after abdominal surgery.14 The physiologic effect of a PaO2 in excess of 400 mm Hg has been estimated to be equivalent to approximately 3 g/dL hemoglobin12 (also J. Feiner, et al., unpublished data, 2010). Thus, the physiologic effect of breathing oxygen when added to the patient's native hemoglobin of 0.7 g/dL produced a heart rate equivalent to nearly 4 g/dL hemoglobin, a value associated with approximately 80% survival (R. B. Weiskopf, unpublished data, 2010). Provision of a high FIO2 can be a useful “bridge” until red cells are available for transfusion.

As important as the above information may be, it is perhaps more important to know the hemoglobin concentration that is associated with significant morbidity than that associated with mortality, because clinicians would prefer to prevent the former before it results in the latter. The brain seems to be more sensitive to acute anemia than is the heart (R. B. Weiskopf, unpublished data, 2010). Healthy humans have degraded neurocognitive function at a hemoglobin concentration of 6 g/dL12,15 and increased latency of the encephalographic P300 wave13 that is associated with cognitive dysfunction16 and defective information processing.17 Prospective randomized clinical trials have not identified the low hemoglobin concentration that results in an increased morbidity or decreased function compared with that at a higher concentration.9,10,18 The hemoglobin concentrations in those trials were apparently insufficiently low and different from the “liberal” transfusion group to be able to detect any differences.

The patient described by Dai et al.,1 experiencing hemorrhagic shock as the result of an uncontrolled severed axillary artery, was not transfused with erythrocytes for 12 hours because crossmatched compatible blood was not available at that institution. The patient's ABO Rh type was B negative, with an unknown transfusion history, and the hospital blood bank's policy was to withhold Rh-positive blood without knowledge of the absence of anti-D antibodies in the potential recipient. Clinically significant anti-D antibodies do not arise in Rh-negative individuals unless exposed to that antigen by previous transfusion or pregnancy. One may be left with an incorrect impression that under these conditions there are no viable options. Delay in availability of crossmatched compatible erythrocytes is not extremely rare owing to the presence of a single antibody to a donor antigen, or unusual combinations of more than 1 of these. Furthermore, in circumstances of massive hemorrhage in an individual or of mass casualty, the availability of crossmatched compatible erythrocytes may be inadequate. Thus, familiarity with appropriate alternatives can be of benefit to clinicians to enable informed decisions and to facilitate communication with a physician of the blood bank or transfusion service during these circumstances.

For usual transfusions, red cell preparations undergo crossmatching of donor red cells and recipient's serum to ensure absence of immunohematologic incompatibility to avoid ensuing antigen-antibody–mediated hemolysis (although an increasing number of blood banks now only confirm the recipient's ABO type and proceed with type-specific red cells if the antibody screen is negative). In the absence of sufficient time to perform a crossmatch (approximately 30–45 minutes at most institutions) or the availability of crossmatched compatible blood, alternatives are possible.

In the United States (US), the Food and Drug Administration mandates “procedures to expedite transfusion in life-threatening emergencies” of blood and blood components when clinical circumstances so warrant.19 Other organizations that certify blood banks, such as the American Association of Blood Banks and the College of American Pathologists, require that an institution's blood bank have a written policy for such release.a,20 However, specific details are not mandated, and policies and procedures differ among institutions. It is possible to have emergency-release erythrocytes delivered from the institution's blood bank to the operating room within very few minutes.21 Some active trauma centers maintain a supply of type O red cells in the immediate vicinity for even more rapid availability. Operating room personnel may be well advised to become familiar with policies and practices at their institution, and discuss changes to meet clinical needs, as necessary.

Deciding among options requires an assessment of the relative risks. The risks of not transfusing (anemia) are summarized above. The risks of erythrocyte components (in addition to the risks attached to all units of blood, independent of compatibility testing, such as potential transmission of viruses, parasites, and bacteria; transfusion-related acute lung injury; immunomodulation; and others, the current estimates of risk of transmission of human immunodeficiency virus or hepatitis C virus are <1 per 1 million units,22 for hepatitis B virus approximately 1 per 300,000 U23; transfusion-related acute lung injury induced by packed red cells has been reported24; however, the risk is unquantified, but low because of the small amount of contained plasma) depend on both the specific component and the recipient. The risks accrue from potential incompatibility between red cell antigens and serum antibodies. In Vietnam, the US forces transfused >100,000 units of “universal donor”25,26 supplied as O-positive whole blood28,29 with low anti-A and anti-B titers29 without a single fatality from transfusion reaction (I calculate the 95% upper confidence limit for a zero incidence as 3.0/100,000 units; and the 99% confidence interval as 4.6/100,000 units).27 However, the vast majority of these were transfused in relatively young males who were unlikely to have been transfused previously. Previous transfusion or pregnancy of the recipient increases the possibility of recipient antibody development and the consequent risk. Data regarding antibodies to red cell antigens in hospitalized patients28 and antigen frequency in the population30 would seem to suggest that for this population an intravascular hemolytic transfusion reaction might occur with an approximate frequency of 1 to 6 per 10,000 transfusions. Although adverse events have been reported after the transfusion of type O whole blood to recipients with other blood types, these have been attributed to high titers of anti-A and/or anti-B antibodies in the donor plasma. Whole blood continues to be used by the military on occasion in combat zones, generally as type specific. Currently, in civilian practice, whole blood is not used for this purpose. “Emergency-release” “blood” currently is most often supplied as type O-negative erythrocytes, although some institutions may issue O-positive red cells instead to conserve their supply of the former (for transfusion to women with potential for childbearing, to prevent them from developing anti-D antibodies that could cause hemolytic disease of the newborn). Inasmuch as packed red cell preparations contain <10% of the plasma of whole blood (Hirschler N. Personal communication. President/CEO, Blood Centers of the Pacific, San Francisco; and Clinical Professor, University of California, San Francisco, 2010), this risk is greatly reduced, and there are no reports of fatal hemolytic reactions after their use as “universal donor.”

Anti-A and anti-B antibodies persist for a variable time after transfusion of type O blood to someone with another ABO type.31 Consequently, it has been recommended to assess these titers in the recipient after transfusion of type O before switching back to the patient's hereditary ABO type when the crossmatch is compatible, although the exact satisfactory titer has not been determined.31

If it is determined that transfusion can wait a few minutes longer, another alternative is possible. If a type and screen (for unexpected antibodies in the recipient's serum) is negative, use of type-specific blood has generally been considered acceptable. Combining examinations of 2 reports from 1 large institution totaling 141,286 crossmatches, the incidence of a clinically significant antibody in a general hospitalized population was 1.2 per 10,000 units transfused,32,33 which I calculate to have a 99.99% upper confidence limit of 2.8 per 10,000 units. Furthermore, all antibodies were weakly positive and would have been unlikely to have caused an immediate hemolytic transfusion reaction, although delayed hemolysis would have been possible. At a different institution, examination of 12,848 type and screens in a single year found 11 recipient antibodies that were not detected (incidence of 8.6 of 10,000 recipients).28 Only 1 of these missed antibodies is associated with clinical hemolytic transfusion reactions (with an antigen of low frequency of approximately 0.0038), resulting in my calculated risk (with 99.99% confidenceb) of a hemolytic transfusion reaction of no more than 2.2 per 1 million transfusions when relying on a type and screen without a crossmatch. Thus, performing an “immediate spin crossmatch” (sometimes called “quick spin”) to verify patient ABO type should allow for provision of red cells within 5 to 10 minutes. However, this does not apply to neonates because the screen may not detect passively transferred ABO immunoglobulin G antibodies.

In the absence of a previously performed type and screen, other remaining options include performing a type and screen, or dispensing with the antibody screen and proceeding with transfusion of type-specific cells. Several reports of relatively small civilian experiences have failed to note a single case of hemolytic transfusion reactions after “emergency” transfusion of type-specific units.29,3436

However, there is significant question whether requesting and transfusing type-specific red cells is safer than type O in emergency situations.29,37 It was noted in the Vietnam experience that the incidents of hemolytic transfusion reactions when using type-specific blood occurred in clusters when mass casualties were treated.25,26,29,31 This has led to the supposition that human error induced by the need for speed and multiple simultaneous procedures and transfusions, misaligning recipients, recipients' blood samples, and donor units resulted in major incompatibilities.29

Different considerations apply to the Rh system. Transfusion of Rh-positive red cells to a patient who is Rh negative, but does not have anti-D antibodies will not result in the rapid destruction of red blood cells unless the patient forms antibodies to the D antigen, which usually takes >4 weeks. Anti-D antibody rarely fixes complement and thus, transfusion of Rh-positive red blood cells to an Rh-negative patient with anti-D antibodies rarely results in intravascular hemolysis.38 Thus, in the absence of other alternatives in an exceptional critical circumstance, after consultation with a physician of the blood bank or transfusion service, it may be acceptable to transfuse Rh-positive red blood cells to a recipient who has anti-D antibodies.

The above issues (and the previously higher incidence of transfusion-transmitted infectious disease) have prompted the development of synthetic or semisynthetic oxygen carriers. Recently, the focus has been on potential use when blood is not available. However, a large randomized clinical trial with that goal in mind testing a hemoglobin-based oxygen carrier (HBOC) was not successful, having not met a noninferiority 30-day mortality end point when compared with standard care.39 However, it should be noted that noninferiority margins are generally somewhat arbitrary, that the clinical trial missed the upper confidence limit of noninferiority by <1%, and that the overall mortality did not differ statistically between groups. The failure of that trial and previous HBOC trialsc,40 and the transient increases in serum troponin and lipase in a number of clinical studies41,42 has led some to conclude that these are effects of the entire class of compounds (see refs. 41 and 42). Currently, there are no ongoing clinical trials in the US with HBOCs, although development of at least one HBOC is continuing in Europe.43 Consequently, the potential option of use of HBOCs as a “bridge” until erythrocytes are available will not appear soon in the US. As noted above, a partial bridge, however, may be provided by administration of a high FIO2.

What is the value of having reviewed this information? We still do not know when to transfuse red cells. We have no clinical measures that let us know of impending insufficient oxygenation as anemia progresses. Large studies or clinical trials will never be able to define that point for a specific patient with the specific circumstances at hand. Thus, until better measures are available, clinical judgment is required. Implicit in such judgment is the balance of risks of not transfusing versus transfusing using one of the options available. It is the transfusing clinician who must make that assessment, and choose among options, and, if appropriate, with consultation with a physician from the blood bank or transfusion service. In doing so, it is well to remember that clinical medicine does not include a condition of “no risk”; that even 99.99% confidence is but a probability and allows for a risk, albeit small, of adverse outcome, despite unreasonable societal expectations or demands of absence of medical risk, while not holding themselves to that same impossible standard. The clinician who faces these, at times, difficult assessments and responsibility can request and expect the appropriate component that she or he deems to best satisfy the solution that minimizes the attendant risk. The nature of providing anesthesia is minimization of risk, which at times requires selection among options, all of which entail some risk. To do so implies appropriate knowledge; decisions regarding if and when to transfuse and selecting among transfusion choices fits well with that professional paradigm.

—“Take calculated risks. That is quite different from being rash.”—Gen. George S. Patton, III, letter to Cadet George S. Patton IV, 6 June 1944.

Back to Top | Article Outline


The author has a relationship with or consults for the following companies and organizations that have an interest in erythrocyte transfusion: US Food and Drug Administration, US National Heart, Lung, and Blood Institute/National Institutes of Health, US Army, Sangart Inc., and CaridianBCT. The author was project/corporate VP and Executive Scientific Advisor at Novo Nordisk A/S 2005–2007. The NHLBI/NIH provides partial salary support for the author.

Back to Top | Article Outline


The author thanks Drs. C. Harrison, N. Hirschler, and P. Toy for reviewing the manuscript and for their comments and suggestions.

a College of American Pathologists. Transfusion. Medicine Checklist. Edition 671, 2010. Available at
Cited Here...

b Calculation of confidence intervals of the result of multiplying 2 values, each with its own confidence interval, is not standard. My estimate likely somewhat underestimates the confidence interval, and thus likely overestimates the risk.
Cited Here...

c Hemosol, news release. Available at:; 21 June 2003. Accessed August 4, 2009.
Cited Here...

Back to Top | Article Outline


1. Dai JQ, Tu WF, Yang Z, Lin RH. Intraoperative management of extreme hemodilution in a patient with a severed axillary artery. Anesth Analg 2010 (in press)
2. Zollinger A, Hager P, Singer T, Friedl HP, Pasch T, Spahn D. Extreme hemodilution due to massive blood loss in tumor surgery. Anesthesiology 1997;87:985–7
3. Carson JL, Duff A, Poses RM, Berlin JA, Spence RK, Trout R, Noveck H, Strom BL. Effect of anaemia and cardiovascular disease on surgical mortality and morbidity. Lancet 1996;348:1055–60
4. Carson JL, Noveck H, Berlin JA, Gould SA. Mortality and morbidity in patients with very low postoperative Hb levels who decline blood transfusion. Transfusion 2002;42:812–8
5. Viele MK, Weiskopf RB. What can we learn about the need for transfusion from patients who refuse blood? The experience with Jehovah's Witnesses. Transfusion 1994;34:396–401
6. Wu W, Rathore S, Wang Y, Radford M, Krumholz H. Blood transfusion in elderly patients with acute myocardial infarction. N Engl J Med 2001;345:1230–6
7. Wu WC, Schifftner TL, Henderson WG, Eaton CB, Poses RM, Uttley G, Sharma SC, Vezeridis M, Khuri SF, Friedmann PD. Preoperative hematocrit levels and postoperative outcomes in older patients undergoing noncardiac surgery. JAMA 2007;297:2481–8
8. Beattie WS, Karkouti K, Wijeysundera DN, Tait G. Risk associated with preoperative anemia in noncardiac surgery: a single-center cohort study. Anesthesiology 2009;110:574–81
9. Hébert PC, Wells G, Blajchman MA, Marshall J, Martin C, Pagliarello G, Tweeddale M, Schweitzer I, Yetisir E. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care. N Engl J Med 1999;340:409–17
10. Lacroix J, Hebert PC, Hutchison JS, Hume HA, Tucci M, Ducruet T, Gauvin F, Collet JP, Toledano BJ, Robillard P, Joffe A, Biarent D, Meert K, Peters MJ. Transfusion strategies for patients in pediatric intensive care units. N Engl J Med 2007;356:1609–19
11. Meier J, Kemming GI, Kisch-Wedel H, Wolkhammer S, Habler OP. Hyperoxic ventilation reduces 6-hour mortality at the critical hemoglobin concentration. Anesthesiology 2004;100:70–6
12. Weiskopf RB, Feiner J, Hopf HW, Viele MK, Watson JJ, Kramer JH, Ho R, Toy P. Oxygen reverses deficits of cognitive function and memory and increased heart rate induced by acute severe isovolemic anemia. Anesthesiology 2002;96:871–7
13. Weiskopf RB, Toy P, Hopf HW, Feiner J, Finlay HE, Takahashi M, Bostrom A, Songster C, Aminoff MJ. Acute isovolemic anemia impairs central processing as determined by P300 latency. Clin Neurophysiol 2005;116:1028–32
14. Rosenberg-Adamsen S, Lie C, Bernhard A, Kehlet H, Rosenberg J. Effect of oxygen treatment on heart rate after abdominal surgery. Anesthesiology 1999;90:380–4
15. Weiskopf RB, Kramer JH, Viele M, Neumann M, Feiner JR, Watson JJ, Hopf HW, Toy P. Acute severe isovolemic anemia impairs cognitive function and memory in humans. Anesthesiology 2000;92:1646–52
16. Goodin DS. Event-related potentials. In: Aminoff MJ ed. Electrodiagnosis in Clinical Neurology. 4th ed. San Francisco: Churchill Livingstone, 1999:569–85
17. Polich J. P300 and clinical patients. Clin Electroencephalogr 2002;33:vi
18. Carson J, Terrin M, Barton F, Aaron R, Greenburg A, Heck D, Magaziner J, Merlino F, Bunce G, McClelland B, Duff A, Noveck H. A pilot randomized trial comparing symptomatic vs. hemoglobin-level-driven red blood cell transfusions following hip fracture. Transfusion 1998;38:522–9
19. Code of Federal Regulations, Title 21. Washington, DC: US Food and Drug Administration, Department of Health and Human Services, 2003
20. American Association of Blood Banks. Standards for Blood Banks and Transfusion Services. 26th ed. Bethesda, MD: American Association of Blood Banks, 2009
21. Weiskopf RB, Webb M, Stangle D, Klinbergs G, Toy P. A procedure for rapid issue of red cells for emergency use. Arch Pathol Lab Med 2005;129:492–6
22. Zou S, Dorsey KA, Notari EP, Foster GA, Krysztof DE, Musavi F, Dodd RY, Stramer SL. Prevalence, incidence, and residual risk of human immunodeficiency virus and hepatitis C virus infections among United States blood donors since the introduction of nucleic acid testing. Transfusion 2010;50:1495–504
23. Zou S, Stramer SL, Notari EP, Kuhns MC, Krysztof D, Musavi F, Fang CT, Dodd RY. Current incidence and residual risk of hepatitis B infection among blood donors in the United States. Transfusion (in press)
24. Win N, Chapman CE, Bowles KM, Green A, Bradley S, Edmondson D, Wallis JP. How much residual plasma may cause TRALI? Transfus Med 2008;18:276–80
25. Camp FR. Lessons learned applicable to civil disaster: recipient identification and blood transfusion. J Trauma 1975;15:743–4
26. Camp FR, Dawson B. Prevention of injury to multiple casualties requiring resuscitation following blood loss. Military Med 1974;139:893–8
27. Hanley JA, Lippman-Hand A. If nothing goes wrong, is everything all right? Interpreting zero numerators. JAMA 1983;249:1743–5
28. Boral LI, Henry JB. The type and screen: a safe alternative and supplement in selected surgical procedures. Transfusion 1977;17:163–8
29. Barnes A. Transfusion of universal donor and uncrossmatched blood. Bibl Haematol 1980:132–42
30. Westhoff CM, Reid ME. Review: the Kell, Duffy, and Kidd blood group systems. Immunohematology 2004;20:37–49
31. Barnes A Jr, Allen TE. Transfusions subsequent to administration of universal donor blood in Vietnam. JAMA 1968;203:695–7
32. Oberman HA, Barnes BA, Friedman BA. The risk of abbreviating the major crossmatch in urgent or massive transfusion. Transfusion 1978;18:137–41
33. Oberman HA, Barnes BA, Steiner EA. Role of the crossmatch in testing for serologic incompatibility. Transfusion 1982;22:12–6
34. Blumberg N, Bove JR. Un-cross-matched blood for emergency transfusion: one year's experience in a civilian setting. JAMA 1978;240:2057–9
35. Gervin AS, Fischer RP. Resuscitation of trauma patients with type-specific uncrossmatched blood. J Trauma 1984;24:327–31
36. Schwab CW, Civil I, Shayne JP. Saline-expanded group O uncrossmatched packed red blood cells as an initial resuscitation fluid in severe shock. Ann Emerg Med 1986;15:1282–7
37. Goldfinger D. Uncrossmatched blood for emergency transfusion. JAMA 1977;237:1826–7
38. Klein HG, Anstee DJ. Mollison's Blood Transfusion in Clinical Medicine. 11th ed. Oxford, UK: Blackwell Publishing Ltd., 2005
39. Moore EE, Moore FA, Fabian TC, Bernard AC, Fulda GJ, Hoyt DB, Duane TM, Weireter LJ Jr, Gomez GA, Cipolle MD, Rodman GH Jr, Malangoni MA, Hides GA, Omert LA, Gould SA. Human polymerized hemoglobin for the treatment of hemorrhagic shock when blood is unavailable: the USA multicenter trial. J Am Coll Surg 2009;208:1–13
40. Sloan EP, Koenigsberg M, Gens D, Cipolle M, Runge J, Mallory MN, Rodman G Jr. Diaspirin cross-linked hemoglobin (DCLHb) in the treatment of severe traumatic hemorrhagic shock: a randomized controlled efficacy trial. JAMA 1999;282:1857–64
41. Silverman TA, Weiskopf RB. Hemoglobin-based oxygen carriers: current status and future directions. Anesthesiology 2009;111:946–63
42. Weiskopf RB. Hemoglobin-based oxygen carriers: compassionate use and compassionate clinical trials. Anesth Analg 2010;110:659–62
43. Olofsson C, Nygards EB, Ponzer S, Fagrell B, Przybelski R, Keipert PE, Winslow N, Winslow RM. A randomized, single-blind, increasing dose safety trial of an oxygen-carrying plasma expander (Hemospan) administered to orthopaedic surgery patients with spinal anaesthesia. Transfus Med 2008;18:28–39
© 2010 International Anesthesia Research Society