Because red blood cell (RBC) membranes are only slightly permeable to potassium, their movement is largely dependent on energy-dependent transport mechanisms (glycolytically derived adenosine triphosphate). During storage, RBC membranes age, adenosine triphosphate synthesis, and potassium pumping decrease,1 and intracellular potassium leaks into the supernatant. The supernatant of stored RBC units may contain more than 60 mEq/L of potassium.2 However, hyperkalemia related to transfusion depends not only on the potassium concentration in the RBC unit, but also on volume and rate of RBC administration.
A number of reports describing transfusion-related hyperkalemia and cardiac arrest have been published.3–7 In elective cases, hyperkalemia from RBC transfusion may be prevented through washing of stored blood units before administration.8 However, in most instances, rapid transfusion of large volumes takes place during emergencies, which are less amenable to preventative techniques. Therefore, early recognition and prompt pharmacologic intervention may be facilitated by greater awareness and anticipation. It is critical to develop a better understanding of the conditions that lead to hyperkalemia during RBC transfusion. In the present report, we examine the clinical scenarios and outcomes of 16 patients who suffered cardiac arrest after developing hyperkalemia during RBC transfusion.
After Mayo Clinic IRB (Rochester, MN) approval, we retrospectively examined the medical records of patients who developed intraoperative cardiac arrest associated with hyperkalemia during RBC transfusion. Cases were identified by electronic search of the Anesthesia Performance Improvement Database from November 1, 1988, to December 31, 2006. This database was introduced on the November 1, 1988, and serves as a repository of all critical incidents, including perioperative cardiac arrests and deaths.
We defined transfusion-associated hyperkalemic cardiac arrest as an intraoperative arrest requiring cardiopulmonary resuscitation during administration of either large volume or fast RBC transfusion. Our criteria for defining transfusion-associated hyperkalemic cardiac arrest included arrest temporally related to large or fast RBC transfusion, laboratory evidence of hyperkalemia (≥5.5 mEq/L), and/or an anesthesiologist's note that the arrest was due to hyperkalemia. Large or fast RBC transfusion occurred when several RBC units were given in sequence, multiple lines were used to simultaneously transfuse RBC, and/or commercial pressure devices were used to accelerate transfusion. We analyzed only cases that occurred under continuous care of the anesthesia team, i.e., those in the operating room or in the postanesthesia care unit. Using “death” and “cardiac arrest” as keywords, our initial search of the Performance Improvement database identified 1420 patients. This group was further searched with specific key words: external defibrillation, chest compressions, change in cardiac status, arrhythmia requiring intervention, hypothermia, rapid transfusion pump, RBC transfusion, and hyperkalemia. We identified 222 cases, and all charts were reviewed by a single reviewer (S.J.F.). We excluded all events that happened during cardiopulmonary bypass operations, liver transplant operations, and patients who required extracorporeal membrane oxygenation. Physician documentation of the arrests was also scrutinized for statements relating to evidence of electrocardiographic (ECG) patterns consistent with hyperkalemia during RBC transfusion. The charts of 16 patients that fit our search criteria were independently reviewed by the three authors (H.M.S., J.A., S.J.F.), and any inconsistencies were resolved by the consensus of the senior authors (J.S., H.M.S.).
The following information was collected for each patient: age, gender, ASA physical status, type and urgency of surgery (elective versus emergency). Relevant comorbid conditions were also recorded including kidney disease (creatinine ≥2.0 mg/dL), coronary artery disease, congestive heart failure, and diabetes mellitus. Laboratory data collected included serum potassium concentrations (before surgery, last value before cardiac arrest, and the highest level during cardiac arrest), ionized calcium levels (before cardiac arrest or calcium administration), hemoglobin concentration, blood glucose concentrations, and pH (last measurements immediately before cardiac arrest). RBC transfusion was calculated as number of RBC units before the cardiac arrest, and total RBC units for the entire surgery. The transfusion of any other non-RBC products (platelets, fresh frozen plasma, cryoprecipitate) are not reported. Additional data included lowest intraoperative body temperature and ECG presentation of the cardiac arrest (ventricular fibrillation, asystole, pulseless electrical activity, wide complex QRS). Resuscitation measures were divided into mechanical (chest compressions, external or internal cardiac massage, defibrillation, pacemaker placement) and pharmacological (atropine, epinephrine, bicarbonate). Data were collected for specific management of recognized or presumed hyperkalemia, including calcium, insulin, bicarbonate, glucose, albuterol, and hyperventilation. We also collected information regarding administered RBC units: leukoreduction, irradiation status, and type of storage solution. Outcomes in terms of operating room survival and in-hospital survival were also recorded. Only descriptive statistics (mean and standard deviation [sd], median and range) were used.
Data were also collected regarding supernatant potassium concentrations of banked packed RBC units as part of routine institutional quality assurance testing. Randomly selected RBC units were tested during the past 5 yr and the age of the unit and potassium concentrations (mEq/L) were recorded. Values reported in Figure 1 are unrelated to the RBC units given to patients who had cardiac arrest in the present case series. Normal potassium range at our institution is 3.6–4.8 mEq/L. Potassium concentrations were determined by the Mayo Clinic Central Laboratory using the Roche/Hitachi Modular Analytic System E170, Roche Diagnostics, Indianapolis, IN).
We identified 16 patients, 11 adult and 5 pediatric, in whom RBC transfusion-associated hyperkalemia was found to be a significant factor in precipitating cardiac arrest. The majority of patients underwent three major types of surgeries: cancer, major vascular, and trauma. All were receiving rapid blood transfusion before cardiac arrest. All transfused units were packed RBCs and none received whole blood. Three patients received leukoreduced RBC units and only one patient (2 day old) received irradiated blood (Table 2). All patients received normal saline during RBC administration, although lactated Ringer's solution (with potassium content) was given to some patients before the events. Six (37.5%) patients underwent emergency surgery. Table 1 summarizes the demographics, comorbidities, and other intraoperative and laboratory characteristics of these patients.
Mean preoperative serum potassium of 16 patients was 4.2 ± 0.6 mEq/L (mean ± sd) with a range from 3.1 to 5.1 mEq/L (Table 1, Fig. 1). Mean serum potassium measured immediately before cardiac arrest was 5.1 ± 1.0 mEq/L (n = 16 patients), whereas that measured during or immediately after cardiac arrest was 7.2 ± 1.4 mEq/L (range, 5.9–9.2 mEq/L) (these were the highest potassium measurements recorded). In two cases, the laboratory reported the potassium as >8 mEq/L and this value was used in our calculations, although this clearly under-estimates the true serum concentration.
The volume of RBC units administered before cardiac arrest in all patients ranged between 1 (in a 2.7 kg neonate) and 54. Average hemoglobin at the time of cardiac arrest was 9.6 ± 3.2 g/dL. Nearly all patients were found to be hyperglycemic at the time of the arrest, with mean glucose concentrations of 288 ± 136 mg/dL. Almost all patients were acidotic (pH 7.18 ± 0.12), hypocalcemic, and hypothermic at the time of arrest (Table 1). In all patients, cardiac arrest occurred in the operating room. One patient experienced acute postthoracotomy bleeding in the recovery room and was immediately taken back to the operating room where he received 6 U of RBC and suffered fatal hyperkalemic cardiac arrest with a potassium concentration of 9.2 mEq/L. Fourteen (87.5%) patients received RBC via central venous access. Commercial rapid infusion devices were used in 8 of 11 adult patients (72.7%), and in the remainder, RBC was rapidly administered using pressure bags or syringe pumping. Mean resuscitation efforts lasted 32 min with a range from 2 to 127 min. Four patients survived the initial event to be transferred to the intensive care unit, but subsequently died from multiorgan failure (n = 3) or “intractable hypotension unresponsive to treatment” (n = 1). Only two patients survived to hospital discharge (12.5%) (Table 2). Table 2 shows ECG presentation of cardiac arrest, therapeutic measures, outcome, and type of RBC unit transfused.
Analysis of 74 blood bank RBC units (unrelated to the RBC units given to our patients in the present case series) revealed average potassium concentrations of 27.3 ± 13.5 mEq/L. Potassium levels ranged from 7.3 to 77.2 mEq/L (Fig. 2). Within the first week of storage (0–7 days), average measured potassium was 19.0 ± 7.8 mEq/L (n = 34), during the second week (8–14 days) 31.5 ± 14.1 mEq/L (n = 26), and between 15 and 28 days storage time 39.9 ± 10.3 mEq/L (n = 14).
Hyperkalemia associated with cardiac arrest may be a serious complication of massive blood administration.9,10 The main finding of our study is that transfusion-associated hyperkalemic cardiac arrest may develop with rapid RBC administration even with modest transfusion volume (Table 1). Other conditions associated with hemorrhagic shock (acidosis, hyperglycemia)11–13 likely contribute to increasing serum potassium levels. Similarly, hypothermia and hypocalcemia independently increase the risk of potassium cardiotoxicity.14,15 All this is underscored by the low survival rate in patients who experienced cardiac arrest after hyperkalemia during rapid blood administration (12.5%).
Most published cases of transfusion-associated hyperkalemic cardiac arrest have occurred in pediatric patients,2,6,7,16 but little is known about how hyperkalemic cardiac arrests related to transfusion partition in total perioperative cardiac arrests across ages. In our previous study, 223 noncardiac surgery patients experienced perioperative cardiac arrests,17 and 4% (9 of 223) were associated with hyperkalemia during RBC transfusion. In a more recent study of perioperative pediatric cardiac arrests during noncardiac operations, 19.2% (5 of 26) of arrests were associated with hyperkalemia during RBC transfusion.18 These estimates suggest that this etiology of cardiac arrest is more frequently the cause of arrest in pediatric patients compared with adult patients in whom other causes are more prevalent. Smaller circulating blood volumes, immature renal function and potassium handling, and differences in autonomic tone may account for some of the discrepancy in prevalence of cardiac arrests between pediatric and adult patients after rapid RBC administration.2,6
The pathogenesis of hyperkalemia during massive RBC transfusion is complex, and depends on numerous alterations related to hemorrhage (tissue hypoperfusion), RBC unit factors, and factors related to the rate and route of administration (Table 3). Serious physiologic derangement, such as uncontrollable hemorrhage with associated low cardiac output, may slow intracellular distribution of potassium administered with RBC units. It has been shown that the low cardiac output states per se (such as hemorrhage in our study) can be associated with hyperkalemia, providing that the potassium concentration in the rapidly transfused blood (with either packed RBCs or whole blood reconstituted with plasma) exceeds 10 mEq/L.19 We demonstrated that the mean potassium concentration in our banked RBC units is 27.3 mEq/L (10-day-old blood), well above the 10 mEq/L required to produce hyperkalemia in low cardiac output states (Fig. 2). Elevated serum potassium typically normalizes rapidly when the transfusion rate is slowed20; however, intracellular redistribution of potassium depends on adequate circulating blood volume and cardiac output,19 both of which are low in hemorrhagic shock. In addition, management of hemorrhagic shock requires uninterrupted resuscitation; therefore, potassium increases from transfusion continue until hemodynamic stability and effective homeostasis are achieved.
Other mechanisms contributing to the risks of RBC transfusion-induced hyperkalemia or increased potassium cardiotoxicity include hyperglycemia, hypocalcemia, hypothermia, and acidosis. First, surgical stress and shock are associated with hyperglycemia. This acute increase in serum osmolality causes potassium to exit cells.12,13 Second, massive transfusion of citrated blood is associated with hypocalcemia, which predisposes to cardiac membrane instability at lower potassium levels.15 Hypothermia also slows the metabolism of citrate, which exacerbates hypocalcemic states. Third, in hypothermia, the rat myocardium becomes more sensitive to the toxic effects of potassium.14 Finally, almost all our patients had metabolic acidosis before cardiac arrest and this condition significantly contributes to extracellular potassium shift.11 The majority of our patients were hyperglycemic, hypocalcemic, hypothermic, and acidotic (Table 1).
Release of endogenous sources of potassium may also exacerbate hyperkalemia from blood transfusion. For example, massive tissue injury, rhabdomyolysis after trauma,21,22 manipulation of tumor tissues,23 or protracted ischemia during major vascular surgery,4 can all increase serum potassium. The majority of our cases fit into one of these categories: 8 patients (50%) had tumor surgery, 3 (18.8%) had massive motor vehicle trauma, and 3 (18.8%) had major vascular operations. The route and rate of blood administration are important factors that must be considered in the pathogenesis of transfusion-associated hyperkalemic cardiac arrest. Most of the patients in our report received RBCs via central venous access and using high-pressure infusing devices. Any exogenous source of potassium would be carried directly to the right heart through the pulmonary circulation and to the left heart where coronary circulation occurs. Therefore, central venous access may deliver more concentrated potassium loads to the coronary circulation than peripheral venous access, and this could have contributed to cardiac arrest in some of our patients. At the same time, pressure-infusing devices can traumatize RBCs causing additional leak of potassium.24,25 Finally, Linko and Tigerstedt20 demonstrated that hyperkalemia correlated well with the rate of transfusion, and not with the actual amount of blood transfused.
An associated clinical problem is that during rapid RBC administration, laboratory evidence of serum potassium concentrations lags in time with resuscitation efforts. Except for ECG monitoring, providers may have no evidence of increasing potassium levels until cardiac arrest ensues. Increases in serum potassium levels may or may not produce accompanying electrocardiographic changes. Patients can manifest peaked T waves, bradycardia, or changing QRS morphology, at serum potassium levels as low as 5.3 mEq/L, whereas patients with chronic renal failure may show no ECG changes, even at much higher potassium levels.26 Brown et al.19 reported that 6 of 18 cardiac arrests were associated with the plasma potassium levels of 6.0 mEq/L or higher. The potassium values in our patients (7.2 mEq/L) were within a range consistent with cardiac arrest reported by others.19,27–29
The quantity of RBC units associated with hyperkalemic cardiac arrest in our patients varied widely (Table 1). The most plausible explanation is that those who received less blood and experienced hyperkalemia either received RBC units with higher potassium content or had increased endogenous potassium release. Anesthesiologists rarely know the potassium content in administered RBC units and may be unaware of the wide variation of potassium in stored RBC units (Fig. 2). Potassium increases with time in stored RBC units. Our highest quality assurance RBC unit tested contained supernatant potassium of 77.0 mEq/L at 14 days of storage, only one-third through its usable shelf life (42 days). One published case of cardiac arrest during blood transfusion reported a potassium concentration of 120 mEq/L in the blood unit tested after the arrest.30 Also, irradiation of blood, by disrupting the RBC membrane, increases the RBC unit free potassium.31 One of our patients, a 2-day-old newborn, received irradiated blood.
Because of the potential to develop transfusion-associated hyperkalemic cardiac arrest during large RBC transfusion, determination of potassium in both RBC units and patients' blood should be routinely considered. Point-of-care laboratory testing may allow for expedited reporting of potassium concentrations to providers. Other measures aimed at preventing the transfusion of hyperkalemic blood products include the preoperative washing of RBC units by transfusion medicine or intraoperative washing of RBC units using cell salvage equipment. A future option may be the use of in-line potassium absorption filters; however, their use has been rarely reported,32,33 because they are not yet commercially available world-wide. In addition, because of flow limitations of these filters, they may be less suitable for situations where fast RBC transfusion is required.
LIMITATIONS OF THE STUDY
Our study has one important limitation in addition to those associated with any retrospective study. Because neither the shelf-age nor the RBC unit potassium content are known before transfusion, we cannot calculate the potassium load given to the patient for each transfused unit. Therefore, we cannot precisely determine the role of hyperkalemia from transfusion versus that of other associated conditions in the pathogenesis of cardiac arrest in our patients.
In conclusion, when dealing with patients undergoing massive hemorrhage and blood replacement, it is imperative to anticipate the potential for developing hyperkalemia. In addition to potassium load in transfused RBC units, other factors associated with bleeding, such as hypotension with low cardiac output, hypothermia, hypocalcemia, hyperglycemia, and acidosis, can contribute to hyperkalemia and/or increase the potential for potassium cardiotoxicity. Anticipation and early recognition of these aberrations, and their prompt correction may theoretically improve otherwise low survival from transfusion-associated hyperkalemic cardiac arrest.
1. Ronquist G, Waldenstrom A. Imbalance of plasma membrane ion leak and pump relationship as a new aetiological basis of certain disease states. J Intern Med 2003;254:517–26
2. Hall TL, Barnes A, Miller JR, Bethencourt DM, Nestor L. Neonatal mortality following transfusion of red cells with high plasma potassium levels. Transfusion 1993;33:606–9
3. Horsey PJ. Hyperkalaemia associated with transfusion of plasma reduced blood. Anaesthesia 2000;55:294–5
4. Murthy BV, Waiker HD, Neelakanthan K, Majam Das K. Hyperkalaemia following blood transfusion. Postgrad Med J 1999;75:501–3
5. Davies JR. Hyperkalaemia and rapid blood transfusion. Anaesthesia 2000;55:928–9
6. Chen CH, Hong CL, Kau YC, Lee HL, Chen CK, Shyr MH. Fatal hyperkalemia during rapid and massive blood transfusion in a child undergoing hip surgery–a case report. Acta Anaesthesiol Sin 1999;37:163–6
7. Buntain SG, Pabari M. Massive transfusion and hyperkalaemic cardiac arrest in craniofacial surgery in a child. Anaesth Intensive Care 1999;27:530–3
8. Knichwitz G, Zahl M, Van Aken H, Semjonow A, Booke M. Intraoperative washing of long-stored packed red blood cells by using an autotransfusion device prevents hyperkalemia. Anesth Analg 2002;95:324–5
9. Leveen HH, Pasternack HS, Lustrin I, Shapiro RB, Becker E, Helft AE. Hemorrhage and transfusion as the major cause of cardiac arrest. JAMA 1960;173:770–7
10. Marshall M. Potassium intoxication from blood and plasma transfusions. Anaesthesia 1962;17:145–8
11. Graber M. A model of the hyperkalemia produced by metabolic acidosis. Am J Kidney Dis 1993;22:436–44
12. Goldfarb S, Cox M, Singer I, Goldberg M. Acute hyperkalemia induced by hyperglycemia: hormonal mechanisms. Ann Intern Med 1976;84:426–32
13. Popp D, Achtenberg JF, Cryer PE. Hyperkalemia and hyperglycemic increments in plasma potassium in diabetes mellitus. Arch Intern Med 1980;140:1617–21
14. Sprung J, Cheng EY, Gamulin S, Kampine JP, Bosnjak ZJ. The effect of acute hypothermia and serum potassium concentration on potassium cardiotoxicity in anesthetized rats. Acta Anaesthesiol Scand 1992;36:825–30
15. Denlinger JK, Nahrwold ML, Gibbs PS, Lecky JH. Hypocalcaemia during rapid blood transfusion in anaesthetized man. Br J Anaesth 1976;48:995–1000
16. Ivens D, Camu F. Sudden hyperkalemia during cardiopulmonary bypass with hypothermic cardiac arrest in an infant. J Cardiothorac Vasc Anesth 1996;10:258–60
17. Sprung J, Warner ME, Contreras MG, Schroeder DR, Beighley CM, Wilson GA, Warner DO. Predictors of survival following cardiac arrest in patients undergoing noncardiac surgery: a study of 518,294 patients at a tertiary referral center. Anesthesiology 2003;99:259–69
18. Flick RP, Sprung J, Harrison TE, Gleich SJ, Schroeder DR, Hanson AC, Buenvenida SL, Warner DO. Perioperative cardiac arrests in children between 1988 and 2005 at a tertiary referral center: a study of 92,881 patients. Anesthesiology 2007;106: 226–37
19. Brown KA, Bissonnette B, McIntyre B. Hyperkalaemia during rapid blood transfusion and hypovolaemic cardiac arrest in children. Can J Anaesth 1990;37:747–54
20. Linko K, Tigerstedt I. Hyperpotassemia during massive blood transfusions. Acta Anaesthesiol Scand 1984;28:220–1
21. Erek E, Sever MS, Serdengecti K, Vanholder R, Akoglu E, Yavuz M, Ergin H, Tekce M, Duman N, Lameire N. An overview of morbidity and mortality in patients with acute renal failure due to crush syndrome: the Marmara earthquake experience. Nephrol Dial Transplant 2002;17:33–40
22. Perkins RM, Aboudara MC, Abbott KC, Holcomb JB. Resuscitative hyperkalemia in noncrush trauma: a prospective, observational study. Clin J Am Soc Nephrol 2007;2:313–9
23. Lobe TE, Karkera MS, Custer MD, Shenefelt RE, Douglass EC. Fatal refractory hyperkalemia due to tumor lysis during primary resection for hepatoblastoma. J Pediatr Surg 1990;25: 249–50
24. Hansen TG, Sprogoe-Jakobsen U, Pedersen CM, Olsen KS, Kristensen SR. Haemolysis following rapid experimental red blood cell transfusion–an evaluation of two infusion pumps. Acta Anaesthesiol Scand 1998;42:57–62
25. Frey B, Eber S, Weiss M. Changes in red blood cell integrity related to infusion pumps: a comparison of three different pump mechanisms. Pediatr Crit Care Med 2003;4:465–70
26. Yu AS. Atypical electrocardiographic changes in severe hyperkalemia. Am J Cardiol 1996;77:906–8
27. Parshuram CS, Cox PN. Neonatal hyperkalemic-hypocalcemic cardiac arrest associated with initiation of blood-primed continuous venovenous hemofiltration. Pediatr Crit Care Med 2002;3:67–9
28. Parshuram CS, Joffe AR. Prospective study of potassium-associated acute transfusion events in pediatric intensive care. Pediatr Crit Care Med 2003;4:65–8
29. Tigerstedt I, Sivulainen S. Massive bleeding and hyperpotassaemia. A case report. Ann Chir Gynaecol 1982;71:175–7
30. Harris B, Lumadue, J, Luban, NLC, Pollack, M. Transfusion-related hyperkalemic arrest from irradiated packed red blood cells. Transfusion (suppl) 1998;38:S69
31. Agarwal P, Ray VL, Choudhury N, Chaudhary RK. Effect of pre-storage gamma irradiation on red blood cells. Indian J Med Res 2005;122:385–7
32. Inaba S, Nibu K, Takano H, Maeda Y, Uehara K, Oshige T, Yuasa T, Nakashima H. Potassium-adsorption filter for RBC transfusion: a phase III clinical trial. Transfusion 2000;40: 1469–74
33. Nakagawa M, Kubota M, Endo I, Inoue S, Seo N. Use of a K(+)-adsorption filter for the massive transfusion of irradiated red blood cells in a child. Can J Anaesth 2004;51:639–40
34. Bashour TT, Ryan C, Kabbani SS, Crew J. Hypocalcemic acute myocardial failure secondary to rapid transfusion of citrated blood. Am Heart J 1984;108:1040–2
35. Rosa RM, Silva P, Young JB, Landsberg L, Brown RS, Rowe JW, Epstein FH. Adrenergic modulation of extrarenal potassium disposal. N Engl J Med 1980;302:431–4
36. Rudolph R, Boyd CR. Massive transfusion: complications and their management. South Med J 1990;83:1065–70
37. Brown KA, Bissonnette B, MacDonald M, Poon AO. Hyperkalaemia during massive blood transfusion in paediatric craniofacial surgery. Can J Anaesth 1990;37:401–8
38. Batton DG, Maisels MJ, Shulman G. Serum potassium changes following packed red cell transfusions in newborn infants. Transfusion 1983;23:163–4
39. Miller MA, Schlueter AJ. Transfusions via hand-held syringes and small-gauge needles as risk factors for hyperkalemia. Transfusion 2004;44:373–81
40. Holme S. Current issues related to the quality of stored RBCs. Transfus Apher Sci 2005;33:55–61
41. Linko K, Saxelin I. Electrolyte and acid-base disturbances caused by blood transfusions. Acta Anaesthesiol Scand 1986;30:139–44
42. Keidan I, Amir G, Mandel M, Mishali D. The metabolic effects of fresh versus old stored blood in the priming of cardiopulmonary bypass solution for pediatric patients. J Thorac Cardiovasc Surg 2004;127:949–52
43. Weinmann M, Hoffmann W, Rodegerdts E, Bamberg M. Biological effects of ionizing radiation on human blood compounds ex vivo. J Cancer Res Clin Oncol 2000;126:584–8
44. Weiskopf RB, Schnapp S, Rouine-Rapp K, Bostrom A, Toy P. Extracellular potassium concentrations in red blood cell suspensions after irradiation and washing. Transfusion 2005;45: 1295–301
45. Glazier DB, Littledike ET, Evans RD. Electrocardiographic changes in induced hyperkalemia in ponies. Am J Vet Res 1982;43:1934–7
46. Hiatt N, Hiatt J. Hyperkalemia and the electrocardiogram in dogs. Basic Res Cardiol 1988;83:137–40
47. Jameson LC, Popic PM, Harms BA. Hyperkalemic death during use of a high-capacity fluid warmer for massive transfusion. Anesthesiology 1990;73:1050–2
48. Rutledge R, Sheldon GF, Collins ML. Massive transfusion. Crit Care Clin 1986;2:791–805