Surgery for hip and knee replacement is associated with large blood loss frequently exceeding 700 mL . As a consequence, blood transfusions are frequently required during and after orthopaedic surgery. Apart from the possible serious complications and side-effects, such as ABO blood group incompatibility and human immunodeficiency virus or hepatitis infections, allogeneic transfusions may also have other complications, such as allo-immunization and immune modulation, which may be responsible for the rise in postoperative infections and possibly increased length of hospitalizations that have been reported in transfused patients [2-5].
Minimizing the number of allogeneic blood transfusions by blood management measures is becoming more common. Such measures may include introduction of transfusion protocols that define low transfusion triggers (haemoglobin (Hb) values at which patients may receive an allogeneic blood transfusion), preoperative autologous blood donations, cell saving, haemodilution techniques, tranexamic acid administration or elevation of the preoperative Hb concentration by means of preoperative injections of epoetin alfa [1,6-8]. This last method has proven to increase preoperative Hb in patients with a Hb <13 g dL−1 at screening, as well as to reduce the need for allogeneic blood transfusions and to increase the perioperative Hb concentration [7,9,10]. However, possible changes in consequences of blood transfusions, e.g. infection rate and postoperative recovery, have not been studied after the administration of epoetin alfa in a routine population of orthopaedic patients. We have conducted an international, randomized study to investigate the effects of preoperative administration of epoetin alfa on perioperative Hb concentrations, and its effects on allogeneic perioperative transfusion requirements in mild-to-moderately anaemic patients, scheduled for major elective orthopaedic surgery. In addition, the effect of epoetin alfa on postoperative infection and recovery was recorded.
A prospective, open, randomized phase IV trial was conducted in The Netherlands, France, Germany, Sweden, Belgium and Australia. Patients scheduled for elective major orthopaedic surgery (hip, knee or spine; primary or revision) and with a preoperative Hb concentration between 10 and 13 g dL−1 were included. We excluded patients with clinically relevant diseases and clinically relevant cardiovascular dysfunction, according to the discretion of the investigator. Spine surgery included spinal fusion of at least three segments. Patients were randomized in blocks of nine patients per hospital by telephone operated interactive voice randomization system in a ratio of 1: 2 to receive either no epoetin or epoetin alfa (Eprex®/Erypo®, Ortho Biotech; 40 000 IU subcutaneously once weekly for 3 weeks before surgery and on the day of surgery together with oral iron daily for 3 weeks). This ratio was chosen in order to improve trial acceptability and participation by the patients. Patients in the non-epoetin (control) group could also take iron orally or receive it by intravenous (i.v.) injection, if this was part of the usual standard of care in that hospital. This policy was chosen because many centres include treatment with iron, but it has not been demonstrated that its administration has any effect on transfusion requirements . All patients, irrespective of their group allocation, received blood transfusions when needed. Blood transfusions were only given according to an Hb-based transfusion trigger, as laid down in the hospital transfusion protocol. Blood volume alone was not a reason for transfusion. If no transparent local protocol was available, transfusion with packed cells could only be given during and after surgery if the Hb was <8.0 g dL−1 . Before any blood transfusion, the Hb concentration was recorded.
Hb concentrations, blood transfusions (peroperative and postoperative; type of transfusion; numbers of patients transfused and numbers of units transfused), time to ambulation, time to discharge from hospital, postoperative infections, therapeutic antibiotic use and safety were measured. Evaluations were carried out at study entry, just before surgery, 1 day after surgery, at discharge from hospital and at follow-up (planned at 4-6 weeks after surgery).
A patient was considered to have an infection when one of the following items existed:
- Wound infection: redness, purulent exudate or positive culture of wound fluid.
- Wound abscess: drainage of abscess or spontaneous discharge of pus.
- Abscess or infected haematoma in surgical area or near the implant: positive culture after collection of pus or re-exploration.
- Urinary tract infection: abnormal urine sediment with white blood cells and/or a positive urine culture and/or clinical signs.
- Respiratory tract infection: clinical signs according to the investigator and/or a positive sputum culture leading to treatment with antibiotics.
- Pneumonia: clinical or radiological signs of a pulmonary infiltrate.
- Bacteraemia: typical clinical signs (e.g. fever) and positive blood culture.
These positive clinical signs as well as a raised temperature were regarded as warning signals, but an infection was only documented if a positive culture was found.
The time to ambulation was defined as the number of days between surgery and the first day that the patient was able to get out of bed and walk around in the room, with or without support. The time to discharge was defined as the number of days between surgery and discharge from the hospital where the surgery was performed. This implies that there was no correction for hospitalization for social reasons or for early recovery protocols. We assumed that randomization should correct for these phenomena.
An intention-to-treat (ITT) analysis was performed in all patients who were included in the trial and had a study evaluation on the day before surgery. The on-treatment population was defined as those patients included in the trial who underwent surgery. Statistical analysis was two-tailed and with α = 0.05 using Wilcoxon's two-sample, Fisher's exact and Pearson's χ2-tests.
Apart from differences between both treatment groups, differences between transfused and non-transfused patients were also tested. In order to exclude the influence of variations by country (heterogeneity in blood saving methods, anaesthesia methods, standard hospitalization periods, etc.) a large and homogeneous part of the study population was analysed separately for postoperative infections and recovery data. This population consisted of primary hip replacement arthroplasty patients from one country, The Netherlands (n = 431).
The study was approved by Hospital Institutional Review Boards and the Local Ethics Committees of each hospital. All subjects gave informed consent prior to study entry.
In the study 733 patients were enrolled: 487 in the epoetin group and 246 in the control group. The ITT population amounted to 704 patients (467 epoetin; 237 control) (Table 1). Patients for whom the operation was postponed for more than 10 days were excluded and this brought the actual surgery population to 695 (460 epoetin; 235 control).
There were no differences between the treatment groups regarding age, height, weight, blood pressure (BP), gender, type of surgery, type of anaesthesia and percentage of patients with rheumatoid arthritis or patients possibly having infections (Table 2).
Table 3 shows types of surgery and types of anaesthesia. These did not differ between groups. In most patients (69%) spinal anaesthesia was used; others had general anaesthesia (27%) or a combination (2%). Only 2% were given local anaesthesia. In the control group 76% of the patients received oral iron therapy before surgery and in the epoetin group this was 97%.
Hb values at screening were 12.2 ± 0.7 g dL−1 in the control group and 12.3 ± 0.7 g dL−1 in the epoetin group (mean ± SD). Hb increased to 14.3 ± 1.2 g dL−1 (+2.1) on the day of surgery in the epoetin group, but did not increase in the control group (+0.1). Except for the screening visit, Hb values were different between treatment groups at each time point (P < 0.05; Wilcoxon two-sample test). On the day after surgery, Hb fell in both groups: to 11.4 ± 1.4 g dL−1 in the epoetin group and to 9.7 ± 1.2 g dL−1 in the control group. Hb increased to 12.3 ± 1.0 g dL−1 and 11.9 ± 0.9 g dL−1 at follow-up (4-6 weeks after surgery) in the epoetin and control group, respectively (Fig. 1).
In the epoetin group 12% of patients received a blood transfusion on at least one occasion and in the control group this was 46% (P < 0.05; Fisher's exact test; Fig. 2). In most cases, these were allogeneic transfusions: 9% of epoetin patients and 37% of con-trol patients received only allogeneic transfusions (P < 0.05; Fisher's exact test). Three percent of epoetin patients and 9% of control patients received autologous transfusions (only autologous or mixed transfusions) (P < 0.05; Fisher's exact test). The composition of the transfusions (autologous blood, allogeneic packed cells, allogeneic whole blood and mixed transfusions) was not different between treatment groups: 73% and 78% in the epoetin and control groups, respectively, were allogeneic transfusions. For autologous blood, this was 23% and 16%, respectively (Fig. 3).
Differences between transfusion requirements were significant in all types of surgery (Table 4; P < 0.001 in all types of surgery; Fisher's exact test).
The total quantity of blood transfused in both treatment groups was similar. The number of transfusions per transfused patient was not different between treatments (1.25 ± 0.51 and 1.42 ± 0.70 for epoetin and control, respectively) (P = 0.141; Wilcoxon two-sample test). The number of units transfused per transfused patient was 2.36 ± 1.95 and 2.41 ± 1.24, respectively (P = 0.126; Wilcoxon two-sample test). Leucocyte filtered blood was used in only 6% of the transfusions; 94% were buffy coat-depleted blood transfusions.
The Hb concentration just before transfusion did not differ between the epoetin and the control patients: 8.6 ± 1.2 g dL−1 for epoetin patients and 8.4 ± 0.9 g dL−1 for control patients; overall 8.5 ± 1.0 g dL−1.
Transfusions by country
The percentage and composition of transfusions differed between the participating countries (Fig. 4). One of the major differences between countries is the use of autologous blood transfusions. Autologous blood donation is standard care in France and Germany, but it is rarely used in the other participating countries.
Pre-transfusion Hb varied by country from 7.8 to 9.1 g dL−1. Apart from Germany, all transfusion triggers were above 8.0 g dL−1.
Time to ambulation and time to discharge
On average, patients could walk again after 3.3 days (±2.7 days). Time to ambulation was not different between the two treatment groups, but it was significantly longer in transfused than in non-transfused patients (3.8 ± 4.0 vs. 3.1 ± 2.2 days) (P = 0.004; Wilcoxon two-sample test; Fig. 5). This was found in both the epoetin group (transfused vs. non-transfused 4.2 ± 3.7 vs. 3.2 ± 2.3; P = 0.01) and the control group (transfused vs. non-transfused 3.6 ± 4.1 vs. 2.9 ± 1.7; P = 0.07).
For the homogeneous subgroup of Dutch patients who underwent primary hip surgery the number of days to ambulation was not statistically different from that of the total group: 3.5 ± 2.0 for epoetin and 3.4 ± 1.9 for control (P = 0.354), but also in this group time to ambulation was significantly longer in transfused patients: 4.2 ± 3.1 vs. 3.3 ± 1.6 (P = 0.038; Wilcoxon two-sample test). In this group only 24 of 293 (8.2%) of epoetin patients received transfusions vs. 63 of 148 (42.6%) of control patients.
Patients were discharged from hospital after an average stay of 10.8 days (± 5.5). Again, this parameter was not different between the two treatment groups, but transfused patients stayed significantly longer in hospital than non-transfused patients: 12.9 ± 6.4 vs. 10.2 ± 5.0 days, respectively (P < 0.001; Wilcoxon two-sample test). This was found in both the epoetin group (transfused vs. non-transfused 15.5 ± 7.2 vs. 10.4 ± 5.3; P < 0.001) and the control group (transfused vs. non-transfused 11.5 ± 5.4 vs. 9.4 ± 3.8; P < 0.001).
Dutch primary hip patients stayed on average in the hospital for 9.8 ± 5.0 days. There was no difference between the two treatment groups, but time to discharge was 11.4 ± 6.1 in the transfused patients vs. 9.4 ± 4.6 in the non-transfused patients (P < 0.001; Wilcoxon's two-sample test). In the epoetin group the difference in time to discharge between transfused and non-transfused patients was 5.5 days (15.1 in transfused patients and 9.7 in non-transfused patients), whereas this difference was 1.5 days in the control group (10.0 in transfused patients and 8.6 in non-transfused patients).
In total 9.8% of the patients had one or more postoperative infections, 5.5% were confirmed by a positive culture. Most patients had the first infection in the hospital: 7.2% vs. 2.6% after discharge (with positive culture: 4.3% and 1.2%). Urinary tract infections were the most common in-hospital infections (4.3%; 2.9% with positive culture), followed by wound infections (2.5%; 1.3% with positive culture).
The total percentage of infections was not different between epoetin (9.4%) and control (10.6%). In transfused patients infection rate was 12.9%; in non-transfused patients 8.9% (P = 0.130; Fisher's exact test). However, it was significantly different in the more homogeneous Dutch primary hip patients: 13.8% in transfused vs. 6.8% in non-transfused patients (P = 0.032; Pearson χ2-test).
Overall, there was no significant difference in the infection rate between patients receiving allogeneic or autologous transfusions, with infections in 12.9% of patients receiving only allogeneic transfusions vs. 10.0% in those given only autologous transfusions (Fisher's exact test). In addition, 13.5% of patients took antibiotics for therapeutic use (epoetin: 14.0%; control: 12.6%; transfused patients: 16.9%; non-transfused patients: 13.3%).
No differences were observed in the adverse events frequency between epoetin and control treatments. Three thrombotic events occurred in the population: two in the epoetin group and one in the control group.
This study addresses several important questions about blood management in hip and knee arthroplasty in the routine hospital setting. After a drug has been investigated in well-controlled situations, with as little variation as possible, additional naturalistic investigations are important to confirm the usability in daily life, where procedures are used to vary from centre to centre and from doctor to doctor. First we studied the effect of preoperative epoetin alfa on perioperative Hb concentration and on transfusion requirements in daily hospital care. Secondly the implications of this treatment and of transfusions on postoperative recovery time and postoperative infection rate were evaluated.
The observation that preoperative epoetin alfa treatment enhanced peroperative and postoperative Hb concentration and reduced transfusion requirements confirms earlier results [7,10,13]. Furthermore, the observation that iron administration in the control group (taken by 77%) did not increase the Hb values confirms previous observations [11,14].
Epoetin treatment was associated with a lower transfusion rate and transfusion was associated with a significant extension of time to ambulation and time to discharge both within the epoetin and the control groups. However, no differences in time to discharge and infection rate were observed between epoetin and control patients. Possible explanations are that daily life setting in this study may have produced a more heterogeneous population, as patients with several types of surgery, with and without rheumatoid arthritis and from several countries were enrolled. Concerning infection rate, this is very low in orthopaedic surgery, especially for clinically relevant infections.
The results might also have been confounded by a higher complication rate in the group of epoetin patients who received transfusions, as seen by the outcomes (longer time to ambulation, longer hospital stays and the highest number of antibiotic use). The severe reduction in Hb concentration might be explained by surgical complications, causing major haemorrhage. In epoetin-treated patients (with high preoperative Hb concentrations) a more severe blood loss was needed before a transfusion was given. A surgical complication might explain this heavy blood loss and the consequent bad outcome (i.e. delayed mobilization and extended hospital discharge). Unfortunately this assumption cannot be confirmed. This effect is intensified by the very low number of transfused patients in the epoetin group (56 patients; 12%).
The finding that infection rate and hospitalization time are associated with complications (and thus by transfusions) rather than by anaemia was also observed in 410 primary hip revision patients at the Sint Maartenskliniek Hospital, Nijmegen, The Netherlands, in whom hospital procedures regarding surgery techniques, discharge policy, etc., were kept similar. Allogeneic transfusions appeared to have the highest prospective value for longer hospitalization, followed by wound problems, age and operation time. Other factors, such as gender, height, body weight, blood erythrocyte sedimentation rate, C-reactive protein, preoperative albumen, perioperative blood loss and use of gentamicin in the cement were not confounding factors .
A gender ratio of 90% females in this study is consistent with the population consisting mainly of hip surgery patients. Although females are more at risk for transfusion than males, due to their lower Hb concentrations and lower blood volumes, this had no influence on the outcome of this study, as the gender ratio was the same in both treatment groups.
Several studies have been performed to evaluate the impact of blood transfusion on hospitalization time, recovery and infection rate [16-21]. These studies all confirm a relation between allogeneic blood transfusions and increased risk of infections and/or time to discharge, but the absolute data differ considerably between studies. This might be caused by differences in definitions of infection and length of stay. Some studies in surgery only consider wound infections, whereas this study tried to document all infections. Length of hospitalization is a complex parameter, as it is not merely determined by clinical factors. In this study it has been noticed that decisions to discharge were often based on reimbursement issues, availability of home nurses, private family issues, etc.
Nowadays, so-called accelerated stay programmes may influence hospitalization times more than transfusions or complications. In many countries hip and knee surgery are performed in such programmes, which include hospitalization times limited to less than a week. Epoetin alfa treatment may be very important in such programmes, however, as extensive rehabilitation starts already on the first day after surgery with low Hb concentrations. Energy expenditure in patients is elevated after surgery [22,23] and exercise capacity is reduced by 20-25% on the fourth postoperative day (unpublished data Maasland Hospital, Sittard, The Netherlands). Buick and colleagues  found a reduced maximum oxygen (O2) consumption after phlebotomy and an increase in maximum O2 consumption following induced erythrocythaemia. They suggested that O2 transport puts a limit on maximal aerobic capacity. These findings suggest that the postoperative higher Hb concentrations might become more and more important as rehabilitation becomes more strenuous due to reduction in hospitalization. Preoperative epoetin alfa administration may thus become more important, especially in the more compromised patients, as they suffer more from the changes in rehabilitation and lower accepted Hb concentrations.
This study suggests that in the routine daily setting of major orthopaedic surgery, epoetin alfa treatment is an efficient method to decrease perioperative transfusion requirements and to increase perioperative Hb concentration. As the blood transfusion service in all countries is struggling to meet demands, epoetin alfa might help to reduce needs for packed cells. Moreover, preoperative epoetin alfa treatment might also enable reduction in hospitalization time together with new rehabilitation procedures.
This trial was sponsored by Ortho Biotech Europe and P.v.d.A. at the time of the study was an employee of the sponsoring company. Statistics were performed by M. Bassano and M. Borelli, Dimensione Ricerca, Rome, Italy.
Study review board
R. Slappendel (Nijmegen, The Netherlands)
E. Weber (Eindhoven, The Netherlands)
Y. Hémon (Marseille, France)
S. Mähler (Langenau, Germany)
T. Dalén (Umeå, Sweden)
E. Rouwet (Enschede, The Netherlands)
J. van Os (Sittard, The Netherlands)
A. Vosmaer (Rotterdam, The Netherlands)
P. van der Ark (Zeist, The Netherlands)
Australia: S. Crawford, The Prince Charles Hospital, Chermside.
Belgium: M. Mulier, Dienst Orthopedie, Lübeck.
France: V. Chevron-Proust, Clinique De Europe, Rouen; B. Cholley, Hôpital Lariboisière, Paris; J. Debue, Clinique Des Maussins, Paris; F. Dubois, Hôpital De La Côte De Nacre, Caen; L. Dupré, Clinique Cleret, Chambery; D. Envain, Institut Calot, Berck-Sur-Mer; Y. Hémon, Hôpital Ste-Marguerite, Marseille; Y. Nedjar, Clinique Des Lilas, Les Lilas; F. Pamela, Clinique Ambroise Paré, Neuilly-Sur-Seine; N. Rosencher, Hôpital Cochin, Paris.
Germany: J. Biscoping, St. Vincentius Krankenhaus, Karlsruhe; R. Franz, Universitätsklinikum Carl Gustav Carus, Dresden; E. Hille, Allgemeines Krankenhaus Barmbek, Hamburg; P. Koch, St. Franciskus Hospital, Köln; H. Laubenthal, St. Josef-Hospital, Bochum; S. Mähler, Kreiskrankenhaus, Langenau; P. Thümler, St. Vinzenzkrankenhaus, Düsseldorf; L. Zichner, Orthopädische Klinik Friedrichsheim, Frankfurt.
Sweden: T. Dalén, Norrlands Universitetssjukhus, Umeå C. Eriksson, Sundsvalls Sjukhus, Sundsvall.
The Netherlands: J. de Waal Malefijt, Sint Elisabeth Ziekenhuis, Tilburg; T. Euverman, Stichting Christelijk Ziekenhuis Refaja, Stadskanaal; H. Hoekstra, St. Anna Ziekenhuis, Geldrop; P. Houweling, Diakonessenhuis, Utrecht; H. Kerkkamp, St. Joseph Ziekenhuis, Veldhoven; E. Klop, Sint Jans Ziekenhuis, Weert; A. Koopman-Van Gemert, Albert Schweitzer Ziekenhuis, Dordrecht; J. Lahaye, Antonius Ziekenhuis, Sneek; C. Matthijssen, Elkerliek Ziekenhuis, Helmond; J. Megens, Velp Ziekenhuis, Velp; H. Naber, Isala Ziekenhuis, Zwolle; D. Paré, Ziekenhuis Midden-Twente, Hengelo; E. Rouwet, Ziekenhuis Koningin Beatrix, Winterswijk; M. Simon, Medisch Spectrum Twente, Enschede; R. Slappendel, Sint Maartenskliniek, Nijmegen; A. van de Wiel, Ziekenhuis Eemland Locatie Lichtenberg, Amersfoort; F. van der Lely, Catharina Ziekenhuis, Eindhoven; H. van der Vis, Ziekenhuis Hilversum, Hilversum; J. van Os, Maasland Ziekenhuis, Sittard; R. van Seventer, Ziekenhuis De Baronie, Breda; B. Verbiest, IJsselland Ziekenhuis, Capelle aan den IJssel; A. Vosmaer, Ikazia Ziekenhuis, Rotterdam; O. Wajer, Rivierenland Ziekenhuis, Tiel; E. Weber, Sint Maartenskliniek, Nijmegen.
1. Rosencher N, Kerkkamp H, Macheras G, et al
. Orthopedic surgery transfusion hemoglobin European overview (OSTHEO) study: blood management in elective knee and hip arthroplasty in Europe. Transfusion
2. Marquet RL, Hoynck van Papendrecht MA, Busch OR, Jeekel J. Blood donation leads to a decrease in natural killer cell activity: a study in normal blood donors and cancer patients. Transfusion
3. Landers DF, Hill GE, Wong KC, Fox IJ. Blood transfusion
-induced immunomodulation. Anesth Analg
4. Innerhofer P, Luz G, Spotl L, et al
. Immunologic changes after transfusion of autologous
buffy coat-poor versus white cell-reduced blood to patients undergoing arthroplasty. I. Proliferative T-cell responses and the balance of helper and suppressor T cells. Transfusion
5. Spahn DR. Eliminating blood transfusions. New aspects and perspectives. Anesthesiology
6. Slappendel R, Weber EWG. Blood transfusions and postoperative recovery in orthopaedic surgery. The European Eprex surgery trial. Transfusion Alternatives in Transfusion Medicine
2004; 6 (1)
7. Goldberg MA, McCutchen JW, Jove M, et al
. A safety and efficacy comparison study of two dosing regimens of epoetin alfa
in patients undergoing major orthopedic surgery. Am J Orthop
8. Bierbaum BE, Callaghan JJ, Galante JO, Rubash HE, Tooms RE, Welch RB. An analysis of blood management in patients having a total hip or knee arthroplasty. J Bone Joint Surg Am
9. Adamson J. Perisurgical use of epoetin alfa
in orthopedic surgery patients. Semin Hematol
10. De Andrade JR, Jove M, Landon G, Frei D, Guilfoyle M, Young DC. Baseline hemoglobin as a predictor of risk of transfusion and response to epoetin alfa
in orthopedic surgery patients. Am J Orthop
11. Olijhoek G, Megens JG, Musto P, et al
. Role of oral versus IV iron supplementation in the erythropoietic response to rHuEPO: a randomized, placebo-controlled trial. Transfusion
12. Slappendel R, Dirksen R, Weber EWG, Van der Schaaf DB. An algorithm to reduce allogenic red blood cell transfusions for major orthopedic surgery. Acta Orthop Scand
13. Feagan BG, Wong CJ, Kirkley A, et al
with iron supplementation to prevent allogeneic blood transfusion
in total hip joint arthroplasty. Ann Intern Med
14. Van Iperen CE, Kraaijenhagen RJ, Biesma DH, Beguin Y, Marx JJ, van de Wiel An. Iron metabolism and erythropoiesis after surgery. Br J Surg
15. Weber EWG, Slappendel R, Prins MH, Van der Schaaf DB, Durieux ME. Peroperative blood transfusions and delayed wound healing after hip replacement surgery: effects on hospitalisation. Chapter 5 in Thesis of EWG Weber, Maastricht University, The Netherlands, 2003.
16. Blumberg N. Allogeneic
transfusion and infection: economic and clinical implications. Semin Hematol
17. Houbiers JG, Van de Velde CJ, Van de Watering LM, et al
. Transfusion of red cells is associated with increased incidence of bacterial infection after colorectal surgery: a prospective study. Transfusion
18. Vamvakas EC, Carven JH. Allogeneic blood transfusion
, hospital charges, and length of hospitalization
. A study of 487 consecutive patients undergoing colorectal cancer resection. Arch Pathol Lab Med
19. Carson JL, Altman DG, Duff A, et al
. Risk of bacterial infection associated with allogeneic blood transfusion
among patients undergoing hip fracture repair. Transfusion
20. Innerhofer P, Walleczek C, Luz G, et al
. Transfusion of buffy coat-depleted blood components and risk of postoperative infection in orthopedic patients. Transfusion
21. Sonnenberg FA, Gregory P, Yomtovian R, et al
. The cost-effectiveness of autologous
transfusion revisited: implications of an increased risk of bacterial infection with allogeneic
22. Brandi LS, Oleggini M, Lachi S, et al
. Energy metabolism of surgical patients in the early postoperative period: a reappraisal. Crit Care Med
23. Brandi LS, Bertolini R, Janni A, Gioia A, Angeletti CA. Energy metabolism of thoracic surgery surgical patients in the early post operative period. Chest
24. Buick FJ, Gledhill N, Froese AB, Spriet L, Meyers EC. Effect of induced erythrocythemia on aerobic work capacity. J Appl Physiol