Anesth Analg 2004;98:921-926
© 2004 International Anesthesia Research Society
doi: 10.1213/01.ANE.0000106861.92627.BB
CARDIOVASCULAR ANESTHESIA
The Effect of Erythropoietin on Allogeneic Blood Requirement in Patients Undergoing Elective Liver Resection: A Model Simulation
Claude Lentschener, MD*,
Alexandra Gomola, MD*,
Sophie Grabar, MD
,
Olivier Soubrane, MD
,
Bertrand Dousset, MD,
Pierre-Philippe Massault, MD,
Catherine Penhoud*, and
Yves Ozier, MD*
*Departments of Anesthesia and Critical Care,
Biostatistics, and
Surgery, Université Paris V-René Descartes, Hôpital Cochin, Paris, France
Address correspondence and reprint requests to Claude Lentschener, MD, Department of Anesthesia and Critical Care, Hôpital Cochin, 27 rue du Faubourg Saint Jacques, 75679 Paris Cedex 14, France. Address e-mail to claude.lentschener{at}cch.ap-hop-paris.fr
 |
Abstract
|
|---|
We investigated whether recombinant human erythropoietin (rHuEPO) administration would reduce red blood cell (RBC) transfusion requirements in patients undergoing elective liver resection. We retrospectively investigated 200 patients undergoing elective liver resection. Factors likely to predict perioperative RBC transfusion were studied using a logistic regression analysis. A mathematical model was used to simulate RBC transfusion requirements if (a) transfusion thresholds had been predefined at a hemoglobin concentration of 7–8 g/dL, (b) preoperative hemoglobin concentrations had been increased to 15 g/dL by rHuEPO administration in patients with preoperative hemoglobin concentration in the range 10–13 g/dL, and (c) both interventions had been used. A cost/benefit evaluation of rHuEPO administration formed part of this simulation. RBC transfusion was correlated with major and median liver resection, total liver vascular exclusion, and a combined nonhepatic abdominal surgery but was not correlated with a preoperative hemoglobin concentration in the range 10–13 g/dL. Adherence to a small transfusion threshold or rHuEPO administration alone would have resulted in a slight reduction in transfusion requirements and transfusion rates for the whole population. However, the two interventions in combination would have significantly reduced both variables. One-hundred-eighteen patients undergoing median and major liver resection received 92% of RBC transfused. Sixty-six of these 118 patients had preoperative hemoglobin concentrations in the range 10–13 g/dL and could have received rHuEPO before surgery. rHuEPO alone would have avoided the transfusion of 63 RBC packs of 203 in this subgroup and 12 transfused patients of 31 (P = 0.02). rHuEPO administration to these 66 patients would have cost 186,000 Euro. The 63 RBC saved would have cost 10,710 Euro.
IMPLICATIONS: A mathematical model simulation suggests that the routine preoperative administration of erythropoietin to patients scheduled for major and median liver resection presenting with a preoperative hemoglobin concentration in the range 10–13 g/dL could reduce blood transfusion requirements. However, the cost/benefit ratio warrants consideration.
|
Introduction
|
|---|
Perioperative red blood cell (RBC) transfusion depends upon total blood loss caused by the surgical procedure, preoperative patient RBC volume, and the predetermined transfusion threshold taking into account the patient’s clinical condition (1–4). Recombinant human erythropoietin (rHuEPO) is known to enhance erythropoiesis (2). Increasing the preoperative plasma hemoglobin concentration to 15 g/dL by rHuEPO in patients with preoperative hemoglobin concentration in the range 10–13 g/dL increased autologous blood procurement and decreased allogeneic RBC exposure in patients undergoing orthopedic surgery (2). rHuEPO was also proposed to have reduced RBC transfusion requirements in other types of surgical procedures, including colorectal surgery and hepatocellular carcinoma removal (5,6). Despite surgical skill improvement, liver resection still carries the risk of significant blood loss (7). rHuEPO ameliorates and prevents the anemia associated with chemotherapy and cancer (8). rHuEPO is authorized by the French National Health Authorities for the treatment of cancer-associated anemia (8). A frequent prevalence of cancer has been reported in patients undergoing elective liver resection (7). Thus, French regulations permit the administration of prophylactic rHuEPO to a large subgroup of patients scheduled for elective liver resection. We investigated whether rHuEPO would reduce RBC transfusion requirements in patients undergoing elective liver resection.
|
Methods
|
|---|
This study was conducted in accordance with the Helsinki declaration, and the protocol was approved by our local ethical committee (Comité Consultatif pour la Protection des Personnes se prêtant aux Recherches Biomédicales de Paris-Cochin). Clinical, biological, operative, blood bank, and anesthetic records of 200 consecutive patients undergoing elective liver resection in a single institution between January 1999 and February 2003 were retrospectively investigated. The following data were recorded: age, weight, height, plasma hemoglobin concentration measured before surgery and on the fifth postoperative day, the total amount of intra- and postoperatively transfused RBC (until the fifth postoperative day), the number of transfused patients, and the type of liver resection. We also investigated whether the underlying liver parenchyma was healthy or not, total vascular exclusion (TVE) or portal triad clamping (PTC) were used or not, hepatectomy was conducted under open or laparoscopic surgery, patients had undergone previous laparotomy, and whether an additional nonhepatic procedure was associated with the liver resection.
General anesthesia and perioperative care were at the discretion of the attending anesthesiologists. A central venous catheter was not consistently inserted; however, liquid input was limited to the minimum until skin closure. Three senior surgeons operated on all patients. Red cell salvaging, acute normovolemic hemodilution, and autologous RBC predonation were not used.
A mathematical model was used to simulate the allogeneic RBC transfusion requirements in three idealized situations (1,9). In the first situation, we hypothesized that the ideal thresholds for RBC transfusion were a hemoglobin concentration of 7 g/dL for healthy patients 64 yr of age or younger and 8 g/dL for patients 65 yr of age or older or with preexisting cardiopulmonary disease. Idealized transfusion thresholds were compared with plasma hemoglobin concentrations measured on the fifth postoperative day. In the second situation, we hypothesized that patients with preoperative hemoglobin concentration in the range 10–13 g/dL received a 4-wk course of rHuEPO (600 U/kg x 4) before surgery to increase the preoperative hemoglobin concentration to 15 g/dL (1,2,9). In the third situation, we hypothesized that both low thresholds for RBC transfusion had been adhered to, and a preoperative course of rHuEPO had been administered.
For the mathematical model, as previously described (1,9):
(a) the simulated RBC transfusion (n) if low transfusion thresholds had been predefined at a hemoglobin concentration of 7 g/dL or 8 g/dL was evaluated as ([RBC actually transfused] - [RBC that would have been saved by respecting transfusion thresholds of hemoglobin 7 g/dL or 8 g/dL]);
(b) the number of RBC saved (n) by adhering to transfusion thresholds of hemoglobin 7 g/dL (hematocrit = 21%) or 8 g/dL (hematocrit = 24%) were evaluated as ([hematocrit (%) on the fifth postoperative day - (21 or 24)]/100 x [(estimated blood volume)]/150);
(c) the number of RBC transfused if preoperative hemoglobin concentration had been increased to 15 g/dL by rHuEPO was evaluated as (RBC actually transfused) - (RBC that would have been saved by increasing hemoglobin concentration to 15 g/dL);
(d) the number of RBC saved (n) by increasing hemoglobin concentration to 15 g/dL (hematocrit = 45%) was determined as ([45 - preoperative ht]/100) x ([estimated blood volume]/150)
(e) the number of RBC transfused (n) if both rHuEPO had been administered and if low transfusion thresholds had been adhered to were: (actually transfused RBC) - (RBC that would have been saved if both methods had been used) (in males: [estimated blood volume (mL)] = [body surface area (m2) x 2530]; in females: [estimated blood volume (mL)] = [body surface area (m2) x 2430]); [body surface area (m2)] = 0.0235 x [height (cm)]0.42246 x [weight (kg)]0.51456).
In France, one RBC pack constitutes 250 mL of 60% hematocrit blood expressed in the present study as 150 mL of 100% hematocrit blood. Regarding the number of RBC packs, decimal units were rounded down to the next whole unit when required.
A multivariate logistic-regression analysis was used to assess whether the dependent variable RBC transfusion (yes/no) was independently correlated to age, weight, the importance of liver resection (minor, median, or major), preoperative plasma hemoglobin concentration (<10 g/dL, 10–13 g/dL, or >13 g/dL), underlying liver parenchyma condition (healthy or not), previous abdominal surgery, associated intraabdominal nonhepatic surgery, laparotomy or laparoscopic surgery, a previous chemotherapy, and intraoperative liver vascular interruption (TVE, PTC, or no vascular interruption). All the variables with a P value <0.20 in a preliminary univariate analysis were entered in the multivariate model. The odds ratios and 95% confidence intervals were calculated. The Kruskal-Wallis test, the 
2 test, and the Mann-Whitney U-test were used as appropriate for comparisons of RBC transfusion requirements and transfusion rates. A P value of 0.05 was significant. The software used were the SAS package 8 (SAS Institute Inc, Cary, NC) for the multivariate analysis and Statview 5 for unpaired comparisons (Deltasoft, Meylan, France).
|
Results
|
|---|
Complete data were obtained from 200 records of the 210 investigated. Physical characteristics and perioperative data are shown in Table 1. In 37 of 62 actually transfused patients, the transfusion thresholds were higher than the hemoglobin concentrations of 7 g/dL and 8 g/dL (Table 1).
View this table:
[in this window]
[in a new window]
|
Table 1. Physical Characteristics and Perioperative Data for 200 Patients Undergoing Elective Liver Resection Between January 1999 and February 2003
|
|
In the logistic-regression analysis, the variable transfusion (yes/no) was not significantly correlated with a preoperative plasma hemoglobin concentration in the range 10–13 g/dL (Table 2). The dependent variable transfusion (yes/no) was significantly correlated to the independent variables median and major liver resection, combined extra-hepatic surgical procedure, TVE, and plasma hemoglobin concentration <10 g/dL (Table 2).
All the following comparisons are made with reference to the actual situation. Adhering to a low transfusion threshold and prophylactic rHuEPO administration to patients with a preoperative hemoglobin concentration in the range 10–13 g/dL would have resulted in a nonsignificant decrease in both RBC transfusion requirements and transfusion rates (Table 3). The two interventions in combination would have resulted in a significant decrease in transfusion rates and in a decrease in RBC transfusion requirements at the upper limit of the significance (Table 3). One-hundred-eighteen patients undergoing median and major liver resection received 298 RBC of the 323 RBC administered during the follow-up period (Table 4). Sixty-six patients of these 118 patients undergoing major and median liver resection had a preoperative plasma hemoglobin concentration in the range 10–13 g/dL (Table 4). Both actual and simulated RBC transfusion requirements and transfusion rates concerning these 66 patients are recorded in Table 4. The adherence to a low transfusion threshold alone would have resulted in a nonsignificant decrease in both RBC transfusion requirements and transfusion rates (Table 4). rHuEPO administration alone and the two interventions in combination would have resulted in a nonsignificant decrease in RBC transfusion requirements and in a significant decrease in transfusion rates (Table 4). The respective costs of (a) rHuEPO administration to all the patients with a preoperative hemoglobin concentration in the range 10–13 g/dL undergoing median and major liver resection and (b) the predicted savings to RBC by rHuEPO administration are also indicated in Table 4.
View this table:
[in this window]
[in a new window]
|
Table 3. Red Blood Cell Transfusion Requirements and Transfused Patients in the Actual Clinical Situation and in Three Simulationsa
|
|
View this table:
[in this window]
[in a new window]
|
Table 4. Patients with Perioperative Hemoglobin in the Range (10–13) g/dL Undergoing Major and Median Liver Resectiona
|
|
|
Discussion
|
|---|
These results show that adhering to a low transfusion threshold combined with increasing the preoperative hemoglobin concentration to 15 g/dL by rHuEPO administration in patients with preoperative hemoglobin concentration in the range 10–13 g/dL is likely to reduce both the RBC transfusion requirements and transfusion rates in patients undergoing elective liver resection. However, the subgroup of patients likely to benefit from the administration of rHuEPO alone remains unclear.
Previous data warranted the present evaluation. Preoperative anemia has been correlated with an increased incidence of intraoperative RBC transfusion (10). EPO is a glycoprotein that stimulates erythropoiesis (2). Thus, of 208 patients undergoing hip replacement surgery, rHuEPO reduced the transfusion requirement by 50% (11). A preoperative hemoglobin concentration of 13.5 g/dL or less was associated with an optimized blood-sparing effect with rHuEPO (11). At the present time, a 3- to 4-week course of subcutaneous rHuEPO is routinely and safely administered to increase preoperative hemoglobin concentrations from 10–13 g/dL to 15 g/dL in hip replacement surgery in many institutions (3). Similar results have been reported in various types of surgery, such as colorectal surgery and liver resection for hepatocellular carcinoma (5,6). A decrease in hemoglobin concentrations to 7 g/dL or less, even acutely developed, is well tolerated in patients with no preexisting heart and lung diseases (4). Lower transfusion thresholds are presently advocated, resulting in a decrease in RBC transfusion requirements (10). Also, despite improved surgical skill, liver resection still carries the risk of RBC transfusion (7). The only study of rHuEPO conducted in patients undergoing liver resection focused on a small, restricted subgroup of patients undergoing hepatectomy for hepatocellular carcinoma developed on diseased liver and reports transfusion requirements higher than usual (5,7). A frequent prevalence of cancer is reported among patients undergoing elective liver resection (7). Cancer contraindicates intraoperative RBC salvage (12). Adverse effects remain associated with allogeneic RBC transfusion (13). New vectors of transmissible disease may appear (14). Available pharmacological interventions likely to decrease transfusion requirements are associated with life-threatening adverse effects (15). No adverse effects have been reported with rHuEPO administration in the setting of perioperative care (2,3,5,6,11). Therefore, rHuEPO could be an adequate blood-saving intervention in elective liver resection. We hypothesized that increasing preoperative hemoglobin concentrations to 15 g/dL using rHuEPO in patients with preoperative hemoglobin concentrations in the range 10–13 g/dL combined with a low transfusion threshold would constitute efficient blood-sparing interventions in elective liver resection, regardless of type.
The present model suggests a limited blood-sparing effect of rHuEPO alone in patients undergoing elective liver resection. This could be anticipated from the fact that a preoperative hemoglobin concentration in the range 10–13 g/dL was not correlated with RBC transfusion. A hemoglobin concentration <10 g/dL was correlated with RBC transfusion. However, most national health authorities have not given approval for the prophylactic administration of rHuEPO to patients with preoperative hemoglobin concentration <10 g/dL. A slight beneficial effect of rHuEPO on transfusion requirements is also suggested in the subgroup of patients undergoing major and median liver resection. The dramatic decrease in blood loss recently reported during liver resection may account for this weak effect (7,16). In most recent series of patients undergoing liver resection, 90% of patients received five RBC or less (7). Cirrhosis is no longer predictive of increased blood loss (7,16). Some series even report major liver resection without blood transfusion (16). Should it be required to study a population of more than 200 patients to reach statistical significance, a preoperative hemoglobin concentration in the range 10–13 g/dL would be a weak predictive criterion of transfusion requirements in elective liver resection, and a limited number of patients are likely to benefit from rHuEPO.
This study has some limitations. The present model simulation assumes that blood was isovolemically lost. Therefore, as in previous investigations based on this model, idealized transfusion thresholds were compared with the fifth postoperative day plasma hemoglobin concentrations to reduce the bias associated with this simulated clinical situation (1,9). By the fifth postoperative day, perioperative circulating volume alterations resulting from infusions, transfusions, hemodilution, and postaggressive alteration of capillary permeability are restored, and volemia has achieved a balance (1,9,17). Furthermore, we conducted a preliminary retrospective study. However, preliminary studies were often consistent with further prospective studies (18,19). Thus, a historical assessment demonstrated a RBC-sparing effect of preoperative hemodilution in patients undergoing elective liver resection, which was further confirmed by a prospective randomized study (18,19). Also, the optimal transfusion threshold has not been elucidated (20). However, the transfusion thresholds predetermined in the current study are within the usually accepted range in clinical practice and are believed to be large enough to be accepted by most anesthesiologists (20).
We believe that our results are of potential clinical relevance. Because median and major liver resections were correlated to transfusion requirements, we conducted a subgroup analysis of patients undergoing median and major liver resections. Sixty-six of 118 patients undergoing major and median liver resection had preoperative hemoglobin concentrations in the range 10–13 g/dL. In this subgroup of patients, prophylactic rHuEPO alone would have accounted for a blood-sparing effect of 63 RBC packs. One-hundred units rHuEPO costs 1.25 Euro, and one RBC pack costs 170 Euro in French hospitals. A standard dose prophylactic course of rHuEPO, systematically administered to the 66 patients with preoperative hemoglobin concentration in the range 10–13 g/dL undergoing median and major liver resection, would cost 186,000 Euro and would have saved 10,710 Euro in France. Erythropoietin and RBC pack prices vary among countries. However, the large discrepancies recorded in France between prices for erythropoietin and for RBC packs are of major worldwide concern. Moreover, RBC transfusion has become safer (21). Consequently, advocating the routine preoperative administration of rHuEPO to patients scheduled for major and median liver resections presenting with a preoperative hemoglobin concentration in the range 10–13 g/dL is highly questionable at the present time when cost-containment is an important consideration.
|
References
|
|---|
- Mercuriali F, Inghilleri G. Proposal of an algorithm to help the choice of the best transfusion strategy. Curr Med Res Opin 1996; 13: 465–87.[ISI][Medline]
- Goodnough LT, Rudnick S, Price TH, et al. Increased preoperative collection of autologous blood with recombinant human erythropoietin therapy. N Engl J Med 1989; 321: 1163–8.[Abstract]
- Goldberg MA. Perioperative epoetin alfa increases red blood cell mass and reduces exposure to transfusions: results of randomized clinical trials. Semin Hematol 1997; 34: 41–7.[ISI][Medline]
- Leung JM, Weiskopf RB, Feiner J, et al. Electrocardiographic ST-segment changes during acute, severe isovolemic hemodilution in humans. Anesthesiology 2000; 93: 1004–10.[ISI][Medline]
- Shinozuka N, Koyama I, Arai T, et al. Autologous blood transfusion in patients with hepatocellular carcinoma undergoing hepatectomy. Am J Surg 2000; 179: 42–5.[Medline]
- Qvist N, Boesby S, Wolff B, Hansen CP. Recombinant human erythropoietin and hemoglobin concentration at operation and during the postoperative period: reduced need for blood transfusions in patients undergoing colorectal surgery—prospective double-blind placebo-controlled study. World J Surg 1999; 23: 30–5.[Medline]
- Wei AC, Tung-Ping Poon R, Fan ST, Wong J. Risk factors for perioperative morbidity and mortality after extended hepatectomy for hepatocellular carcinoma. Br J Surg 2003; 90: 33–41.[Medline]
- Beguin Y. A risk-benefit assessment of epoetin in the management of anemia associated with cancer. Drug Saf 1998; 19: 269–82.[Medline]
- Brecher ME, Monk T, Goodnough LT. A standardized method for calculating blood loss. Transfusion 1997; 37: 1070–4.[ISI][Medline]
- Canadian Orthopedic Perioperative Erythropoietin Study Group. Effectiveness of perioperative recombinant human erythropoietin in elective hip replacement. Lancet 1993; 341: 1227–32.[ISI][Medline]
- Bracey AW, Radovancevic R, Riggs SA, et al. Lowering the hemoglobin threshold for transfusion in coronary artery bypass procedures: effect on patient outcome. Transfusion 1999; 39: 1070–7.[ISI][Medline]
- Bryson G, Laupacis A, Wells G. Does acute normovolemic hemodilution reduce perioperative allogeneic transfusion: a meta-analysis—the international study of perioperative transfusion. Anesth Analg 1998; 86: 9–15.[Abstract]
- Flesch BK, Neppert J. Transfusion-related acute lung injury caused by human leucocyte antigen class II antibody. Br J Haematol 2002; 116: 673–6.[Medline]
- Busch MP, Kleinman SH, Nemo GJ. Current and emerging infectious risks of blood transfusions. JAMA 2003; 289: 959–62.[Free Full Text]
- Cohen DM, Norberto J, Cartabuke R, Ryu G. Severe anaphylactic reaction after primary exposure to aprotinin. Ann Thorac Surg 1999; 67: 837–8.[Abstract/Free Full Text]
- Gozzetti G, Mazziotti A, Grazzi GL, et al. Liver resection without blood transfusion. Br J Surg 1995; 82: 1105–10.[ISI][Medline]
- Narushima I, Kita T, Kubo K, et al. Highly enhanced permeability of blood-brain barrier induced by repeated administration of endothelin-1 in dogs and rats. Pharmacol Toxicol 2003; 92: 21–6.[Medline]
- Sejourne P, Poirier A, Meakins JL, et al. Effect of haemodilution on transfusion requirements in liver resection. Lancet 1989; 2: 1380–2.[Medline]
- Matot I, Scheinin O, Jurim O, Eid A. Effectiveness of acute normovolemic hemodilution to minimize allogeneic blood transfusion in major liver resections. Anesthesiology 2002; 97: 794–800.[Medline]
- Capraro L, Nuutinen L, Myllila G. Transfusion thresholds in common elective surgical procedures in Finland. Vox Sang 2000; 78: 96–100.[Medline]
- Vamvakas EC. Meta-analysis of randomized controlled trials investigating the risk of postoperative infection in association with white blood cell-containing allogeneic blood transfusion: the effects of the type of transfused red blood cell product and surgical setting. Transfus Med Rev 2002; 16: 304–14.[ISI][Medline]
Accepted for publication October 28, 2003.
Top
Abstract
Introduction
