Intraoperative red blood cell transfusion in liver Transplantation: Influence on patient outcome, prediction of requirements, and measures to reduce them

Intraoperative Red Blood Cell Transfusion in Liver Transplantation: Influence on Patient Outcome, Prediction of Requirements, and Measures to Reduce Them Emilio Ramos, Antonia Dalmau, Antonio Sabate, Carmen Lama, Laura Llado, Juan Figueras, and Eduardo Jaurrieta Objectives of this study are to quantify the need for blood transfusion during liver transplantation (LT) and confirm the importance of intraoperative blood transfusion as an independent prognostic factor for postoperative outcome. Furthermore, we try to detect useful variables for the preoperative identification of patients likely to require transfusion of packed red blood cell units (PRCUs) and identify measures to reduce transfusion needs. Data were collected prospectively between September 1998 and November 2000. One hundred twenty-two LTs were included in the study. Forty-two patients (34%) did not require transfusion of PRCUs. In multivariate analysis, transfusion of more than three PRCUs was found to be the only significant variable associated with prolonged hospital stay. In addition, excluding perioperative deaths, PRCU transfusion, using a cutoff value of six units, was the only variable to reach statistical significance (P ⴝ .008; risk ratio, 4.93; 95% confidence interval, 15 to 15.9) to predict survival in a multivariate analysis that also included Child’s class and United Network for Organ Sharing (UNOS) classification. Moreover, only preoperative hemoglobin (Hb) level was found to significantly predict the need for transfusion of one or more PCRUs. Finally, only UNOS classification and placement of an intraoperative portacaval shunt were found to be statistically significant to predict the need to transfuse more than six PRCUs. We found the requirement of even a moderate number of blood transfusions is associated with longer hospital stay, and transfusion of more than six PRCUs is associated with diminished survival. Preoperative normalization of Hb levels and placement of an intraoperative portacaval shunt can diminish the number of blood transfusions during LT. (Liver Transpl 2003;9:1320-1327.)

need for transfusions. Intraoperative blood transfusion appeared to be directly related to patient and graft survival.1,2 Moreover, excessive use of blood bank resources for transplantation procedures can compromise other surgical activities. Therefore, minimizing the need for blood product transfusion should be an important objective. We,3 as well as others,4-6 have experienced a progressive reduction in the use of most blood components during the last decade. This trend is caused by several factors, such as earlier patient selection, better preservation of grafts, more experienced surgical and anesthesiology teams, implementation of new techniques,7 and use of drugs to improve hemostasis.8-10 As a result of these developments, LT without blood transfusion is no longer uncommon. However, the ideal situation would be if this became the norm. Three objectives are the basis for our study. The first objective is to quantify transfusion need in LT and, in particular, determine the percentage of cases without red blood cell transfusion. Our second aim is to establish whether intraoperative transfusion is an independent prognostic factor for postoperative outcome. Finally, if this were the case, our final objective is to detect useful variables for the preoperative identification of patients likely to require transfusion of packed red blood cell units (PRCUs). This information would allow the implementation of specific measures in highrisk patients to minimize transfusion requirements.

T

Patients and Methods

raditionally, liver transplantation (LT) has been associated with significant operative blood loss. The simplest way to assess such losses is to quantify the

From the Liver Transplantation Unit, Hospital Universitario de Bellvitge, Barcelona, Spain. Address reprint requests to Emilio Ramos Rubio, MD, Unidad de Trasplante Hepa´tico, Hospital Universitario de Bellvitge, Av Feixa Llarga s/n, L’Hospitalet 08907, Barcelona, Spain. Telephone: 34-93260-7940; FAX: 34-93-260-7603; E-mail: [email protected] Copyright © 2003 by the American Association for the Study of Liver Diseases 1527-6465/03/0912-0012$30.00/0 doi:10.1016/jlts.2003.50204

1320

One hundred twenty-two LTs performed between September 1998 and November 2000 were included in this prospective study. Eleven patients who underwent LT during the same period were excluded because of several conditions that made it advisable to modify perioperative anesthesic management: (1) Budd-Chiari syndrome, (2) acute liver failure, (3) early re-LT (⬍1 month), (4) simultaneous kidney and liver transplantation, and (5) primary familial amyloidotic neuropathy. All data were collected prospectively by the investigators selected for this work. A previous study of the prophylactic use of tranexamic acid (TA) and its lack of risk for arterial thrombosis8 was performed on the same group of patients. Mean

Liver Transplantation, Vol 9, No 12 (December), 2003: pp 1320-1327

1321

Blood Transfusion in LT

Table 1. Characteristics of Liver Transplant Recipients

Age (yr) Gender (Male/Female; %) Diagnosis Cirrhosis HCC Cholestasis Re-LT Other diagnoses UNOS status 1 (At home) 2 (Occasional admission) 3 (Admitted) Child-Pugh classification A B C Previous abdominal surgery Preoperative Hb (g/L) Postoperative Hb (g/L) Hematocrit (%) Platelets (⫻ 109/L) INR PTT (ratio) Fibrinogen (g/L) Serum creatinine (␮mol/L)

Entire Group (n ⫽ 122)

Patients Without Transfusion (n ⫽ 42)

Patients With Transfusion (n ⫽ 80)

57 ⫾ 9.9 67/33

54 ⫾ 12 69/31

58 ⫾ 8 64/36

48 (39.3) 49 (40.2) 7 (5.7) 7 (5.7) 11 (9)

10 (23.8) 21 (50) 3 (7.1) 1 (2.4) 7 (16.7)

38 (47.5) 28 (35) 4 (5) 6 (7.5) 4 (5)

89 (73) 23 (18.9) 10 (8.2)

39 (93) 3 (7)

50 (62.5) 20 (25) 10 (12.5)

40 (32.8) 57 (46.7) 25 (20.5) 28 (23) 114 ⫾ 22 105 ⫾ 3.4 34 ⫾ 6.5 98 ⫾ 66 1.2 ⫾ 0.2 1.2 ⫾ 0.3 2.8 ⫾ 1.3 89 ⫾ 21

27 (64.3) 14 (33.35) 1 (2.4) 7 (16.7) 130 ⫾ 16 112 ⫾ 4.6 38 ⫾ 4 106 ⫾ 68 1.15 ⫾ 0.17 1.09 ⫾ 0.18 3.12 ⫾ 1 81.9 ⫾ 13.5

13 (16.3) 43 (53.8) 25 (20.5) 21 (26.3) 106 ⫾ 11 101 ⫾ 2.6 32 ⫾ 6 93 ⫾ 65 1.31 ⫾ 0.22 1.29 ⫾ 0.34 2.7 ⫾ 1.5 92.7 ⫾ 24

P .03 .68

.008

.001

.0001 .26 .001 .1 .001 .32 .001 .001 .12 .002

NOTE. Values expressed as mean ⫾ SD or number (percent) unless indicated otherwise. Abbreviation: PTT, partial thromboplastin time.

patient age was 57 ⫾ 9.9 (SD) years, 67% were men, and the most frequent diagnoses were cirrhosis with and without hepatocellular carcinoma (HCC; Table 1). Belzer solution was used as preservation solution in most of the LTs (ViaSpan; Du Pont Pharmaceuticals, Wilmington, DE), whereas Celsior solution (IMTIX, Amstelveen, The Netherlands) was used in only five cases. In our group, anesthesic management during LT is performed using a common protocol previously decided by consensus by all anesthesiologist groups. The anesthetic protocol was the same during the study period. Criteria for replacement of blood components were as follows: administration of PRCUs to maintain hemoglobin (Hb) levels between 80 and 100 g/L, fresh frozen plasma in case of hemorrhage associated with an international normalized ratio (INR) greater than 1.7, platelets to maintain a count greater than 50 ⫻ 109/L, and fibrinogen for plasma levels less than 1 g/L. Hemodilution was avoided, whereas for prophylaxis of fibrinolysis, TA (10 mg/kg/hr) was administered from the start of anesthesia to 2 hours after graft reperfusion. In patients with platelet dysfunction, indicated by a bleeding time longer than 12 minutes or by thromboelastography (maximum amplitude ⬍ 35 mm), desmopressin was administered at the beginning of surgery (0.3 mg/kg in 30 minutes). Clinical fibrinolysis, indicated by diffuse bleeding of non-

surgical origin, was treated with boluses of 250 mg of epsilonaminocaproic acid (EACA; a maximum of two administrations). Both EACA and TA can be useful to treat fibrinolysis. We decided to treat clinical fibrinolysis with EACA instead of TA because there is more experience with it. Kang et al11 established that a bolus dose of 250 to 500 mg of EACA was adequate and safe to treat ongoing bleeding caused by fibrinolysis. In our experience, this treatment has been unusual. Intraoperative autotransfusion was not used in any of the operations. All surgical procedures were supervised by one of the two most experienced surgeons, who assume responsibility for the surgical team. When possible, the vena cava preservation technique was used (113 of 122 LTs; 92.6%), whereas a temporary portacaval shunt was placed in 63 patients (51.6%).7 Eighty patients were part of a recently published prospective randomized study12 assessing the usefulness of a portacaval shunt in hemodynamic stability during LT. For the other 42 patients, the choice for a temporary portacaval shunt depended on the preference of the surgeon or anatomic difficulties that made a shunt technically impossible. Prospective collection of data included variables listed in Table 2, which can be grouped in the following manner: (1) donor-dependent variables, (2) recipient pre-LT variables, (3) intraoperative variables, and (4) data concerning postopera-

1322

Ramos et al

Table 2. Variables Recorded During the Study Period Donor

Recipient

Intraoperative

Postoperative

Age (yr) Gender (male/female) Sodium (mEq/L)

Cava preservation (yes/no) Portacaval shunt (yes/no) Cold Ischemia time (min)

ICU stay (d) Hospital stay (d) Surgical mortality (yes/no)

Degree of steatosis (ⱕ30%/⬎30%)

Age (yr) Diagnosis* Previous abdominal surgery (yes/no) Child’s class (A, B, C)

Patient survival (mo)

ICU stay (d)

UNOS classification*

Episodes of hypotension‡ (yes/no)

Hemostasis study INR PTT (ratio) Platelets (⫻ 109/L) Fibrinogen (g/L) Hematocrit (%) Hb (g/L) Creatinine ⬎140 ␮mol/L

Reperfusion syndrome (yes/no)† Duration of anhepatic phase (min) No. of PRCUs transfused

Need for vasopressors (yes/no) Preservation solution (Belzer/Celsior)

*See Table 1. †Decrease greater than 30% in arterial systolic pressure for more than 1 minute in the first 5 minutes after reperfusion. ‡Systolic pressure less than 60 mm Hg for more than 1 hour.

tive outcome. The variable number of PRCUs was not used quantitatively, but converted to a dichotomous variable by varying the cutoff value based on statistical significance. The purpose of this approach is to identify the number of transfusions that was significant from a prognostic point of view for each aspect of the postoperative course. As in general, patients who receive transplants for HCC have better hepatic function than those with other indications; these patients were grouped in a separate category. Mean ⫾ SD are presented for continuous quantitative variables. In the statistical analysis comparing patients with and without transfusions, Student’s t-test was used for quantitative variables, and Chi-squared test, for qualitative variables. To determine the influence of PRCU transfusion on patient outcome, the following dependent variables were evaluated: postoperative mortality, length of intensive care unit (ICU) stay, length of hospital stay, and survival. For the study of the first three variables, a logistical regression method was used. Length of ICU stay and length of hospital stay were previously converted to dichotomous variables using median stay as a cutoff value in each case. For the analysis involving length of hospital stay, patients who died while hospitalized were excluded. Influence on survival was analyzed using a Cox regression model. For the identification of variables predictive of the need for blood transfusion, a logistical regression model was used, whereas the discriminatory power of this predictive model was assessed by the area under the receiver operating characteristic (ROC) curve. Statistical calculations were performed using SPSS, version 10.0 (SPSS Inc, Chicago, IL).

Results Variables related to recipients and donors are listed in Tables 1 and 3, respectively. Mean length of cold ischemia time was 459 ⫾ 162 minutes. Two of seven patients with second transplants had received the first transplant during the study period (1.6%). Mean ICU stay was 5.5 ⫾ 7 days, with a median stay of 3 days (range, 1 to 43 days), and mean hospital stay was 18.9 ⫾ 16 days, with a median stay of 14 days (range, 4 to 114 days). Eight patients died in the first month after surgery or during their post-LT hospital stay before discharge, giving a postoperative mortality rate of 6.6%. Two patients (1.6%) presented with primary graft failure, and in 12 cases (9.8%), liver function was initially poor, but subsequently recovered. Actuarial survival at 3 years was 79%. Intraoperative Transfusion Needs In 122 patients, a mean of 2.9 ⫾ 2.9 PRCUs was transfused, with a median of 2 PRCUs (range, 0 to 14 PRCUs; Fig. 1). However, in 42 of 122 patients (34%), PRCU transfusion was not necessary. For the remaining 80 patients, a mean of 4.4 ⫾ 2.5 PRCUs (median, 4 PRCUs) was transfused. During the ICU stay, 29.5% of patients were administered PRCUs; 14% of previously nontransfused patients and 37.5% of previously transfused patients (P ⫽ .011).

1323

Blood Transfusion in LT

Table 3. Donor Characteristics

Age (yr) Gender (M / F) Sodium (mEq / L) Steatosis ⬎30%* (%) ICU stay (d) Episodes of hypotension (%) Need for vasopressors (%) Preservation solution (UW/Celsior) (%) Cold ischemia time (min)

Entire group

Patients Administered Transfusion

Patients Not Administered Transfusion

P

44 ⫾ 18 43 / 79 148 ⫾ 8.8 33 2.7 ⫾ 3 24 30 96 / 4 459 ⫾ 162

42.7 ⫾ 19.8 13 / 29 146 ⫾ 8.8 31 2.8 ⫾ 2.2 17 24 95 / 5 466 ⫾ 153

44.9 ⫾ 16.9 30 / 50 148 ⫾ 8.8 33 2.8 ⫾ 3 27 34 96 / 4 454 ⫾ 167

.54 .5 .19 .86 .7 .29 .27 .78 .7

NOTE. Values expressed as mean ⫾ SD or percent. Abbreviation: UW, University of Wisconsin. *Data available for 43 cases.

Influence of Intraoperative Red Blood Cell Transfusion on Patient Outcome As expected, patients administered blood transfusions had a more advanced stage of liver disease, indicated by a higher Child-Pugh classification, and were more frequently hospitalized at the time of LT than those not administered transfusions. In addition, frequency of HCC was less and mean age was older in this group (Table 1). Moreover, in these patients, hematocrits and Hb levels were significantly lower, and routine coagulation test results were more abnormal. Nonetheless, there were no statistically significant differences between Hb levels at the end of LT. Finally, serum creatinine levels were significantly

greater, although in only five patients did the level exceed 140 g/L. Donor-dependent variables were not significantly different between the two groups (Table 3). ICU Stay In univariate analysis, none of the donor or intraoperative variables and only one recipient variable, United Network for Organ Sharing (UNOS) classification, significantly affected length of ICU stay. Patients who were hospitalized at the time of LT had a longer ICU stay compared with nonhospitalized patients (P ⫽ .034; risk ratio [RR], 5.6; 95% confidence interval [CI], 1.1 to 27.6). The only significant intraoperative variable was number of blood transfusions, using a cutoff value of six PRCUs (P ⫽ .047; RR, 3.1; 95% CI, 1.01 to 9.7). Both these variables lost their significance in multivariate analysis. Hospital Stay

Figure 1. Distribution of number of packed red blood cell units transfused in each patient.

None of the donor variables was found to have a statistically significant effect on hospital stay, although the factor age showed a trend. Among recipient variables, only INR approached significance (P ⫽ .054; RR, 5.8; 95% CI, 0.97 to 34.9). Among intraoperative variables, only number of blood transfusions, using a cutoff value of two PRCUs, significantly prolonged hospital stay (P ⫽ .009; RR, 2.7; 95% CI, 1.2 to 6). Preoperative INR, Child’s class, UNOS classification, and PRCUs transfused were included in a multivariate model. Only PRCUs transfused, using a cutoff value of three PRCUs, was found to significantly affect hospital stay (P ⫽ .032; RR, 3.06; 95% CI, 1.1 to 8.5).

1324

Ramos et al

CI, 1 to 7.6; Fig. 2). In the adjusted multivariate model, which also included Child’s class and UNOS classification, number of PRCUs transfused remained the only significant variable (P ⫽ .008; RR, 4.93; 95% CI, 1.5 to 15.9). Factors to Predict the Need for Intraoperative Red Blood Cell Transfusion

Figure 2. Kaplan-Meier plot comparing survival of patients administered six or fewer or greater than six PRCUs.

Postoperative Mortality In univariate analysis, the only variables found to significantly affect postoperative mortality were UNOS classification (P ⫽ .0001; RR, 18; 95% CI, 3.5 to 90.1) and preoperative fibrinogen level (P ⫽ .038; RR, 1.54; 95% CI, 1.02 to 2.33). Postoperative mortality in hospitalized patients with UNOS status 3 was greater compared with those with UNOS statuses 1 and 2. In the multivariate study, UNOS classification was the only significant variable (P ⫽ .003; RR, 14.43; 95% CI, 2.43 to 85.6). Number of PRCUs transfused was not a significant factor related to postoperative mortality. Survival When perioperative deaths (N ⫽ 8) were excluded, only number of PRCUs transfused, using a cutoff value of six units, was found to be statistically significant for patient survival (P ⫽ .05; RR, 2.76; 95%

In univariate logistical regression analysis with the need for transfusion of at least one PRCU as the dependent variable, significant recipient variables were age, diagnosis, Child’s class, UNOS classification, INR, partial thromboplastin time (PTT), and Hb level, and the significant intraoperative variable was intraoperative portacaval shunt. None of the donor variables had a significant influence on number of intraoperative red blood cell transfusions. Of variables found to be significant in univariate analysis, only preoperative Hb level was significant in multivariate analysis. The effect of an intraoperative portacaval shunt on number of blood transfusions was close to significance (Table 4). High preoperative Hb level was associated with less risk for transfusion. The area under the ROC curve was 0.88 for this predictive model. The cutoff value for Hb level that significantly influenced need for transfusion was 102 g/L. When transfusion of more than six PRCUs was used as a dependent variable in univariate analysis, Child’s class, UNOS classification, INR, PTT, preoperative Hb level, and intraoperative portacaval shunt placement were found to be significant variables. In multivariate analysis, only UNOS classification and use of an intraoperative portacaval shunt were found to be statistically significant factors to predict the administration of more than six PRCU’s (Table 5). Area under the ROC curve was 0.81.

Table 4. Results of Logistic Regression Analysis to Detect Factors Predictive of the Need for PRCU Transfusion in LT Univariant RR (95% CI) Age Diagnosis (other v HCC) Child’s class BvA CvA UNOS classification (2 ⫹ 3 v 1) INR PTT (ratio) Hb preoperative Portacaval shunt

1.047 (1.008-1.088) 3.532 (1.595-7.820) 6.379 (2.606-15.616) 49.61 (6.05-406.33) 7.80 (2.21-27.45) 61.91 (6.76-566.28) 52.26 (5.81-469.79) 0.54 (0.42-0.69) 0.45 (0.21-0.98)

P .018 .002 .0001 .0001 .0001 .001 .0001 .0001 .0001 .045

Multivariant RR (95% CI) 1.045 (0.982-1.113) 1.532 (0.390-6.019) 3.544 (0.880-14.268) 13.88 (0.82-235.16) 2.597 (0.546-12.351) 0.243 (0.005-12.153) 6.791 (0.309-149.058) 0.689 (0.499-0.951) 0.350 (0.118-1.043)

P .16 .54 .12 .075 .068 .23 .47 .22 .024 .06

1325

Blood Transfusion in LT

Table 5. Results of Logistic Regression Analysis to Detect Factors Predictive of the Need for Transfusion of More Than 6 Units PRCUs in LT Univariant RR (95% CI) Diagnosis (other v HCC) Child’s Class BvA CvA UNOS classification (2 ⫹ 3 v 1) INR PTT (ratio) Hb preoperative Portacaval shunt

4.424 (0.961-20.360) 5.458 (0.644-46.224) 21.93 (2.56-187.52) 13.773 (3.363-56.407) 54.850 (4.629-649.927) 4.038 (1.024-15.927) 0.626 (0.473-0.830) 0.240 (0.073-0.785)

Discussion Mean number of PRCUs transfused and percentage of patients who did not require blood transfusion compare favorably with those in other recent publications.1,13,14 It should be noted that 34% of patients did not require blood transfusions, whereas 86% of patients required maximally six PRCUs, and maximum volume transfused in a single patient was 14 units. In the series of Cacciarelli et al,4 only 24% were not administered blood transfusions intraoperatively, and 47% of patients were administered five or more PRCUs. Although transfusion volume in the report of Cacciarelli et al4 is greater than ours, 3-year survival is slightly greater (86% v 79%). This difference may be explained by the older age of our patients (49.5 v 57 years). Because our surgical blood loss has not been high, it is unlikely that blood loss will have a significant impact on outcome of our patients. Nevertheless, in our experience, intraoperative blood loss requiring replacement of more than three PRCUs is associated with a longer hospital stay. Moreover, transfusion of more than six PRCUs appears to have a negative effect on survival. These findings are important because after exclusion of surgical mortality, intraoperative blood loss had a more significant impact on hospital stay and long-term survival than either Child’s class or UNOS classification. The relationship betweeen perioperative transfusion and patient evolution also was studied by Cacciarelli et al,4 who set the cutoff value at five 5 PCRUs and did not find a significative relationship between blood cell transfusion and patient survival. Age and previous medical condition were more significative factors. This apparent contradiction probably is because we excluded postoperative mortality when calculating survival. Preoperative patient condition may have more influence during this period, but later, other conditions related to

P .056 .004 .119 .005 .0001 .001 .046 .001 .018

Multivariant RR (95% CI) 1.203 (0.207-7.008) 4.777 (0.348-65.515) 10.08 (0.47-215.30) 8.726 (1.528-49.848) 1.089 (0.027-44.676) 0.867 (0.125-6.007) 0.783 (0.554-1.106) 0.256 (0.063-1)

P .83 .32 .24 .14 .015 .96 .88 .16 .05

transfusion, such as graft ischemia-reperfusion injury or viral transmission, would have greater importance. We found that none of the donor variables influenced mortality, length of stay, patient survival, or amount of intraoperative blood loss. This probably is explained by appropriate graft selection and short ischemia times in the case of suboptimal grafts. In the study by Cacciarelli et al,4 ischemic time is longer (7.6 v ⬃11 hours), and this factor is not analyzed. Several reports5,6,15 stressed the difficulties involved in identifying generally applicable preoperative factors that can reliably predict the need for red blood cell transfusion. Moreover, it is logical to assume that results obtained in these studies must be influenced by differences in surgical and anesthesia protocols, as well as by changes that occur over time. Therefore, transplant programs should assess their results regularly to detect emerging risk factors. In our series, preoperative Hb level appears to be a predictive factor for transfusion requirement of minimally one PRCU, but using a cutoff value of six PRCUs, this factor ceased to be significant. Therefore, in a program with a lower percentage of patients without the need for perioperative blood transfusions, this variable probably would not be significant, and it would be unlikely that correction of preoperative Hb level has a significant effect on postoperative outcome. It should be stressed that the difference in Hb levels between our patients with and without blood transfusions (106 ⫾ 2 v 130 ⫾ 1.6 g/L) almost completely explains the difference in transfusion requirements (4.47 PRCUs). Conversely, in the study by Steib et al,5 in which intraoperative requirements were high, multivariate analysis showed that preoperative Hb level, as well other variables, was predictive of increased transfusion requirements (⬎12 PRCUs). Nonetheless, the validity of this observation is limited because of characteristics

1326

Ramos et al

Table 6. Comparison of Patient Characteristics as a Function of Intraoperative Portacaval Shunt

Age (yr) Preoperative Hb (g/L) INR PTT (ratio) Child’s class (%) A B C UNOS classification (%) At home Occasional admittance Admitted Diagnosis (%) Cirrhosis HCC Cholestasis Other tumors Re-LT Other causes

Portacaval Shunt

No Portacaval Shunt

P

57.8 ⫾ 8.9 11.6 ⫾ 2.2 1.22 ⫾ 0.3 1.23 ⫾ 0.19

56.22 ⫾ 10.9 11.2 ⫾ 2.2 1.23 ⫾ 0.3 1.29 ⫾ 0.23

.37 .35 .9 .1

36.5 49.2 14.3

28.8 44.1 27.1

77.8 15.9 6.3

67.8 22 10.2

33.3 46 7.9 3.2 6.3 3.2

45.8 33.9 3.4

.45

.13

of the study because it includes a 10-year period and therefore may include many changes in surgical technique and patient management. Even more, patients in that study underwent LT with lower Hb levels than our patients, and finally, differences in Hb levels between patients at each side of the cutoff value only justify a small part of differences in PRCUs transfused.In conclusion, this work is not representative of the current situation of LT. Regarding possible methods of improving Hb levels before surgery, it should be emphasized that experience with the use of erythropoietin has been described for a Jehovah’s Witness patient in need of an LT.16 In view of these results and considering the predictive value of preoperative Hb level, it seems justified to perform a prospective study to show the usefulness of erythropoietin administration during time on the waiting list. The objective would be to achieve an Hb level of approximately 102 g/L. Like previous studies, we could not show a statistically significant relationship between preoperative coagulation parameters and need for intraoperative blood products. This observation supports the principle that blood component use during LT should be based on clinical assessment and a well-established transfusion protocol, rather than laboratory tests.17,18 Because of the low blood cell volume transfused in our patients, we believe use of intraoperative autotransfusion would increase the complexity of LT without adding great advantages. Groups that use intraoperative autotransfusion have reported greater transfusion rates

5.1 11.9

than ours.5,9,10 Even more, it has been published that use of intraoperative methods for blood cell recuperation are only efficient from a cost-benefit point of view at more than 12 transfusion units.19 Nonetheless, its selective use in patients with greater risk for need of a high transfusion volume and who undergo LT with anemia may be useful to minimize intraoperative transfusions. Previous studies have shown that various technical modifications in LT, such as venovenous bypass20 and preservation of the vena cava,21 had a significant influence on transfusion volume. Because the technique involving preservation of the vena cava was used in most of our patients, it is not surprising that this variable was not found to significantly increase transfusion needs. Conversely, a temporary portacaval shunt, a technical modification not adopted by most groups, is associated in our experience with a significant reduction in blood loss, probably as a result of the reduction in portal pressure. In particular, use of a temporary portacaval shunt became a significant factor when a greater cutoff value for the red blood cell transfusion was used, i.e., when there is more intraoperative hemorrhage. Clearly, one of the limitations of this study in terms of assessing the influence of a portacaval shunt on postoperative outcome is lack of randomization. However, no significant differences were observed in characteristics of patients with and without use of a temporary portacaval shunt (Table 6). It was not one of the objectives of this study to show the effect of antifibrinolytic agents on blood loss.

Blood Transfusion in LT

Numerous reports have been published on this subject,10,22,23,24 suggesting a positive effect. Other drugs, such as recombinant factor VIIa9 and conjugated estrogens,13,25 likely will contribute to reduce the need for blood transfusion in LT in the future. In conclusion, in the context of patient selection and treatment protocols, we can avoid red blood cell transfusion in one of every three LT patients. It also can be inferred from our results that even moderate transfusion needs are associated with longer hospital stays, and higher needs result in a negative impact on survival. Finally, we also conclude that preoperative normalization of Hb level and placement of an intraoperative portacaval shunt can diminish the need for red blood cell transfusion during LT.

References 1. Cacciarelli TV, Keeffe EB, Moore DH, Burns W, Busque S, Concepcion W, et al. Effect of intraoperative blood transfusion on patient outcome in hepatic transplantation. Arch.Surg 1999; 134:25-29. 2. Palomo Sanchez JC, Jimenez C, Moreno GE, Garcia I, Palma F, Loinaz C, et al. Effects of intraoperative blood transfusion on postoperative complications and survival after orthotopic liver transplantation. Hepatogastroenterology 1998;45:1026-1033. 3. Jaurrieta E, Casais L, Figueras J, Ramos E, Lama C, Rafecas A, et al. Ana´lisis de 500 trasplantes hepa´ticos en el Hospital de Bellvitge. Med Clin (Barc) 2000;115:521-529, 2000. 4. Cacciarelli TV, Keefe EB, Moore DH, Burns W, Chuljian P, Busque S, et al. Primary liver transplantation without transfusion of red blood cells. Surgery 1996;120:698-704. 5. Steib A, Freys G, Lehmann C, Meyer C, Mahoudeau G. Intraoperative blood losses and transfusion requirements during adult liver transplantation remain difficult to predict. Can J Anaesth 2001;48:1075-1079. 6. Deakin M, Gunson BK, Dunn JA, McMaster P, Tisone G, Warwick J, Buckels JA. Factors influencing blood transfusion during adult liver transplantation. Ann R Coll Surg Engl 1993; 75:339-344. 7. Belghiti J, Noun R, Sauvanet A. Temporary portocaval anastomosis with preservation of caval flow during orthotopic liver transplantation. Am J Surg 1995;169:277-279. 8. Dalmau A, Sabate A, Koo M, Rafecas A, Figueras J, Jaurrieta E. Prophylactic use of tranexamic acid and incidence of arterial thrombosis in liver transplantation. Anesth Analg 2001;93:516 (abstr). 9. Hendriks HG, Meijer K, de Wolf JT, Klompmaker IJ, Porte RJ, de Kam PJ, et al. Reduced transfusion requirements by recombinant factor VIIa in orthotopic liver transplantation: A pilot study. Transplantation 2001;71:402-405. 10. Porte RJ, Molenaar IQ, Begliomini B, Groenland TH, Januszkiewicz A, Lindgren L, et al. Aprotinin and transfusion requirements in orthotopic liver transplantation: A multicentre ran-

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

1327

domised double-blind study. EMSALT Study Group. Lancet 2000;355:1303-1309. Kang Y, Lewis JH, Navalgund A, Russell MW, Bontempo FA, Niren LS, Starzl TE. Epsilon-aminocaproic acid for treatment of fibrinolysis during liver transplantation. Anesthesiology 1987; 66:766-773. Figueras J, Llado L, Ramos E, Jaurrieta E, Rafecas A, Fabregat J, et al. Temporary portocaval shunt during liver transplantation with vena cava preservation. Results of a prospective randomized study. Liver Transpl 2001;7:904-911. Frenette L, Cox J, McArdle P, Eckhoff D, Bynon S. Conjugated estrogen reduces transfusion and coagulation factor requirements in orthotopic liver transplantation. Anesth Analg 1998; 86:1183-1186. Schroeder R, Johnson L, Plotkin J, Kuo P, Klein A. Total blood transfusion and mortality after orthotopic liver transplantation. Anesthesiology 1999;91:329-330. Findlay JY, Rettke SR. Poor prediction of blood transfusion requirements in adult liver transplantations from preoperative variables. J Clin Anesth 2000;12:319-323. Snook NJ, O’Beirne HA, Enright S, Young Y, Bellamy MC. Use of recombinant human erythropoietin to facilitate liver transplantation in a Jehovah’s Witness. Br J Anaesth 1996;76:740743. Reyle-Hahn M, Rossaint R. Coagulation techniques are not important in directing blood product transfusion during liver transplantation. Liver Transpl Surg 1997;3:659-663. Gerlach H, Gosse F, Rossaint R, Bechstein WO, Neuhaus P, Falke KJ. The relevance of perioperative coagulation parameters to indications for blood transfusion. A retrospective analysis of 300 liver transplantations. Anaesthesist 1994;43:168-177. Kemper RR, Menitove JE, Hanto DW. Cost analysis of intraoperative blood salvage during orthotopic liver transplantation. Liver Traspl Surg 1997;3:513-517. Moreno GE, Calleja KJ, Perez-Cerda F, Gomez GM, Jover Navalon JM, Lanoa Garcia JI, et al. Reduction of transfusion needs by the early introduction of venovenous bypass in liver transplant. Rev Esp Enferm Dig 1989;75:1-5. Cherqui D, Lauzet JY, Rotman N, Duvoux C, Dhumeaux D, Julien M, Fagniez PL. Orthotopic liver transplantation with preservation of the caval and portal flows. Technique and results in 62 cases. Transplantation 1994;58:793-796. Ramsay M. The use of antifibrinolytic agents results in a reduction in transfused blood products during liver transplantation. Liver Transpl Surg 1997;3:665-668. Findlay JY, Rettke SR, Ereth MH, Plevak DJ, Krom RA, Kufner RP. Aprotinin reduces red blood cell transfusion in orthotopic liver transplantation: A prospective, randomized, double-blind study. Liver Transpl 2001;7:802-807. Dalmau A, Sabate A, Acosta F, Garcia-Huete L, Koo M, Sansano T, et al. Tranexamic acid reduces red cell transfusion better than epsilon-aminocaproic acid or placebo in liver transplantation. Anesth Analg 2000;91:29-34. Porte RJ, Leebeek FW. Pharmacological strategies to decrease transfusion requirements in patients undergoing surgery. Drugs 2002;62:2193-2211.