JOURNAL HOME CME HOME THIS MONTH PAST ISSUES ETOC COLLECTIONS AUTHORS REVIEWERS EDITORIAL BOARD FEEDBACK RSS HELP
		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (15) Citing Articles via Google Scholar Google Scholar Articles by Massicotte, L. Articles by Roy, A. Search for Related Content PubMed PubMed Citation Articles by Massicotte, L. Articles by Roy, A. Related Collections Cardiovascular Economics and Health Care Research Blood Monitoring (Non-cardiac)

---

EDITORIAL

Transfusion Predictors in Liver Transplant

Luc Massicotte, MD^*, Marie-Pascale Sassine, CPhD^*, Serge Lenis, MD FRCPS^*, and André Roy, MD FRCPS

*Anesthesiology Department and the Department of Surgery, Hepato-Biliary Service, Centre Hospitalier de l’Université de Montréal, Montréal, Canada

Address correspondence and reprint requests to Luc Massicotte, MD, Hôpital St-Luc–CHUM, 1058, St-Denis, Montreal, PQ, Canada, H2X 3J4. Address email to lmassicotte{at}hotmail.com

Abstract

In this study we sought to determine the factors influencing red blood cell (RBC) transfusions and to study the transfusion practice of anesthesiologists during liver transplants. A retrospective study of 206 successive liver transplants was undertaken during a period of 52 mo. Transfused blood products were identified. Twenty variables were analyzed in a univariate fashion. For the multivariate analysis, the cases were divided in 2 subgroups: more than 4 RBC units transfused and 4 or less RBC units transfused. The average number of RBC units transfused during a liver transplant was 2.8 (± 3.5) per patient, 32.0% did not receive any RBC, and 19.4% did not receive any blood products during the transplant. Three variables were related to the number of RBC units transfused: the starting International Normalized Ratio value, the starting platelet count, and the duration of surgery. We found that there was a wide difference in the transfusion practice of the anesthesiologists involved in this series of liver transplants. It was difficult to identify predictive factors for RBC transfusions when the transfusion rate was small and because of the variability in human factors. Plasma transfusion did not decrease the rate of RBC transfusions; sometimes it was the contrary.

IMPLICATIONS: This is a retrospective study of 206 liver transplants over 52 mo to identify the predictive factors of red blood cell transfusions and the anesthesiologists’ transfusion strategies. We conclude that there is a wide difference in transfusion practices among anesthesiologists.

Liver transplantation is a surgical procedure that can lead to massive blood loss and consequently result in transfusion of blood products (1–4). Improvements in surgical and anesthetic technique have decreased the number of blood products transfused (5,6). The Stanford group reported that 24% of their patients did not receive any red blood cell (RBC) transfusion during a liver transplant and that the survival rate at 1 yr decreased in the patient subgroup that received more than 4 RBC units (5). However, the RBC transfusion varies among different transplant institutions from as few as 4.3 U per patient (5) to an extreme of 43 U per patient (1). The predictive factors remain difficult to determine (4); understanding the risk factors for transfusion might lead to strategies to decrease the mortality rate at 1 yr. The literature on liver transplants does not include practice differences among anesthesiologists as a potential factor influencing the need for blood transfusions.

The St. Luc Hospital of the Centre Hospitalier de l’Université de Montréal (CHUM) is a liver transplant center performing more than 50 transplants per year. The purpose of this study was to evaluate the transfusion strategies used by anesthesiologists to identify factors associated with transfusion of blood products during liver transplants.

Methods

With the approval of the CHUM ethics committee, a retrospective study was performed. All the liver transplants performed between January 1998 and April 2002 were included in the study.

Three hepato-biliary surgeons were present during the entire study period. In 1999, a fourth surgeon joined the team. A team of 2 surgeons performed each transplant. Veno-venous bypass was not used, nor were any piggyback techniques. All livers used in the transplants were from cadaveric donors. There were no reduced-size grafts.

Twelve anesthesiologists were involved in the liver transplant team, 6 of whom were involved during the entire study period. A team of 1 anesthesiologist and 1 respiratory therapist managed the anesthesia in each transplant case.

The same preparation for anesthesia was used for every patient. All were fitted with a peripheral IV catheter, a central venous catheter, and pulmonary artery catheter. Electrocardiogram, oxygen saturation, capnometry, invasive and noninvasive arterial blood pressures, pulmonary artery pressure, and central venous pressure (CVP) were all monitored continuously. The main laboratory test used was the arterial blood gas analysis, which also included hemoglobin (Hb), potassium level, and ionized calcium level. All patients had a urinary catheter as well as a fluid warmer and a warming blanket.

There was no standardized protocol pertaining to the transfusion of blood products. Aprotinin was given to every patient following the Hammersmith protocol (7): 2 million units as a bolus before the time of incision and half a million units per hour until the last vascular anastomosis. A cell-saver device was used for 1 patient. Isovolemic hemodilution technique was not used. Vasopressin, phenylephrine, norepinephrine, epinephrine, dopamine, or ephedrine was used as needed to control arterial blood pressure. The anesthesia technique was not standardized; each anesthesiologist selected his own technique.

Twenty variables (age, sex, weight, height, surgeon, history of previous abdominal surgery, previous liver transplant, Pugh’s score, Model of End-stage Liver Disease (MELD) score, anesthesiologist, surgeon work shift, starting Hb, starting International Normalized Ratio (INR), starting platelet count, CVP, body temperature, duration of surgery, clamping time, duration of cold ischemia and diagnosis) were analyzed in a univariate and multivariate fashion.

Distributions were examined to insure the proper statistical evaluation. Parametric and nonparametric statistical tests were used as appropriate. The dependant variable, units of RBC transfused, was treated as a continuous variable. For the multivariate analysis, the group of patients was split in two groups, using the same variables as the Stanford group (5), >4 RBC transfused, to compare against the low transfusion group (4 RBC). To study the anesthesiologists and surgeons as variables, the least transfused group (4 RBC) was broken down in 3 subgroups: no blood product transfused, plasma transfused only, and 1–4 RBC transfused. For continuing dependant variables (RBC or plasma), we used a Spearman correlation rank test when the independent variables were continuous and a Kruskal-Wallis test in the cases of categorical independent variables. To assess subgroup differences, we used a Pearson ² test. For comparison of more than two groups of RBC or plasma, we used the Kruskal-Wallis test. Stepwise regression and logistic regression exploratory models were used to look at the data in a multivariate fashion. The SPSS 10 and Statview 5 (SAS Institute, Cary, NC) were used.

Results

Demographic, preoperative and perioperative data for the 206 cases are presented in Table 1. The age distribution ranged from 17 to 68 yr. No transplant patients were excluded from the study. Two-hundred-six liver transplants were performed on 193 patients. Ten patients had 2 liver transplants and 1 patient had 4 transplants. Four patients had a liver transplant before the study period.

View larger version (28K): [in this window] [in a new window]   Figure 1. Patient stratification by the number of red blood cells (RBC) transfused.

Univariate analysis of the 20 variables demonstrated that 7 of them influenced the RBC transfusion rate: starting INR value, starting platelet count, duration of surgery, starting Hb value, Pugh’s score, MELD score (Spearman correlation rank test for the first 6 variables) and diagnosis (Kruskal-Wallis test).

Because the transfusion rate did not show a Gaussian distribution (Figure 1), 2 groups, low transfusion rate (4 RBC) and high transfusion rate (>4 RBC), were compared in a multivariate fashion. In this analysis, the starting Hb value, the Pugh’s score, the MELD score and the diagnosis lost their predictive strength. The starting INR value was the most sensitive variable capable of identifying the patients who were going to be transfused (Table 3), but it had a low specificity. One patient with a starting INR value at 4.2 did not receive any blood products.

View this table: [in this window] [in a new window]   Table 3. Summary Table of the Multivariate Analysis (Group 4 RBC Versus >4 RBC)

A large variation was noted in the number of RBC transfused by each anesthesiologist (Fig. 2). There was an even larger variation for plasma transfusions (Fig. 3). Smaller variations were observed for the surgeons (Figs. 4, 5). To study the anesthesiologists and the surgeons as a variable, the low transfused group (4 RBC) was split in 3 subgroups: 40 patients who did not receive any blood product, 26 patients who only received plasma (4.9 U/patient), and 97 patients who received between 1 and 4 RBC units (2.6 ± 1.0 RBC units/patient and 3.6 ± 3.0 plasma units/patient). No anesthesiologist was a predictive variable for transfusion of >4 RBC. However, in cases without transfusion, the anesthesiologist seems to matter. Anesthesiologist #1 performed only 13.6% of the total liver transplants but was involved in 34.6% of "plasma only" transplants (Table 2). Anesthesiologists #1, #8, and #9 did not transfuse any RBC units in approximately half of their cases (Table 2). The mean starting Hb value, starting INR value, starting platelet count, Pugh’s score, MELD score, duration of surgery, threshold for RBC transfusion, and the final Hb were the same for each anesthesiologist (Table 4). Blood losses were larger among anesthesiologists who transfused more. Table 5 shows the distribution of cases in relation to the anesthesiologist.

View larger version (42K): [in this window] [in a new window]   Figure 2. Number of red blood cells (RBC) transfused according to the anesthesiologist. Boxes represent interquartile ranges containing 50% of values. The lines across the boxes indicate median values.

View larger version (45K): [in this window] [in a new window]   Figure 3. Number of plasma units transfused according to the anesthesiologist. Boxes represent interquartile ranges containing 50% of values. The lines across the boxes indicate median values.

View larger version (29K): [in this window] [in a new window]   Figure 4. Number of red blood cells (RBC) transfused according to the surgeon. Boxes represent interquartile ranges containing 50% of values. The lines across the boxes indicate median values.

View larger version (30K): [in this window] [in a new window]   Figure 5. Number of plasma units transfused according to the surgeon. Boxes represent interquartile ranges containing 50% of values. The lines across the boxes indicate median values.

Discussion

On average, 2.8 RBC units per patient were transfused during a liver transplant. This transfusion rate is smaller than reported in previous published literature (1–5). Moreover, in certain studies, diagnoses such as fulminant hepatitis, redo transplant, Budd-Chiari syndrome, hepatorenal syndrome, and renal transplant were eliminated. In our study, all liver transplants were included in the analysis.

As the equipment was not available in our center until recently (8), a cell saver was only used once (retransfusion of 350 mL; hematocrit, 0.7). The isovolemic hemodilution technique was not used to avoid the dilution of coagulation factors. Since 1993, all of our patients received aprotinin as described in the Hammersmith protocol (7,9). No allergic reaction was observed (10). Two patients required reoperation because of a thrombosis of portal vein and hepatic artery (11–13). It was impossible to conclude whether aprotinin was the cause.

The small transfusion rate seems to be attributable to using a lower Hb value as the threshold for transfusing. In their series, Steib et al. (4) were using a Hb value of 100 g/L as their threshold for transfusing. In our center, there is no formal protocol for the transfusion of RBC and plasma. In our series, the anesthesiologists were transfusing when the mean value of the Hb was 68.2 ± 12.8 g/L (the threshold was the same for all the anesthesiologists). As for plasma transfusion, the attitude was more liberal and individual (the starting INR value was also the same for all the anesthesiologists). The threshold for transfusion of RBC was not defined in the Stanford series (5).

Two groups were formed for comparison. The first group represented patients receiving a larger transfusion (43 patients with >4 RBC, closest to the superior quartile), and the second group represented patients receiving a smaller transfusion (163 patients with 4 RBC).

The multivariate analysis showed that there was a relationship between the starting platelet count, duration of surgery, the starting INR value, and the number of RBC units that were transfused (Table 3). The smaller the starting platelet count, the more the patients received RBC. Platelets were seldom administered in our center, as they were not easily available. The duration of the surgery was probably an independent as well as a confounding factor. When the same surgeon was compared to himself, the duration of the surgery became a confounding factor. However, when surgeons were compared, it became an independent factor.

Surgeons were probably the main factor responsible for the small requirement in blood products. As 2 surgeons operated, differences were attenuated (Figs. 4, 5; Table 2). It should be noted that surgeon #3 had been involved with 45% of the patients who did not receive any blood products.

Among anesthesiologists, there was less homogeneity in the transfusion rate of RBC and plasma units. This is probably because the anesthesiologists worked alone (Figs. 2, 3). The difference in the average number of transfused RBC units given once the threshold transfusion was reached was attributable more to the difference in the blood losses. Blood losses were hard to assess with accuracy yet it was more with anesthesiologists who transfused more RBC and more plasma (Table 4).

Anesthesiologist #9 was involved in 25% of the cases that did not receive any transfusion. The final mean Hb level for these 10 patients was 92.2 ± 14.6 g/L. This anesthesiologist only worked twice with surgeon #3. Despite the big difference in transfused RBC, univariate and multivariate analysis did not show the anesthesiologist to be a significant variable for transfusion of RBC, probably because the cohort of patients for each anesthesiologist was too small and there were too few patients in the >4 RBC transfused subgroup. Anesthesiologists appeared more successful in cases without transfusion of RBC; anesthesiologists #1 and #9 performed 13 cases without any RBC transfusion.

A greater disparity was noted with the plasma transfusion rates. This seems to indicate a much more liberal plasma transfusion practice. From the core of anesthesiologists, #2 and #12 transfused the largest amount of plasma, 6.2 U/patient and 5.4 U/patient (Figs. 2, 3), in an effort to correct the INR to diminish the blood loss and decrease the number of RBC transfusions. Paradoxically, their RBC transfusion rate was high at 3.6 U. At the other end of the spectrum, anesthesiologist #9 transfused very little plasma and very little RBC, with respective rates of 2.5 U and 2.0 U. However, none of the analyzed variables can explain how this happened. Distribution of cases was totally at random for the surgeons as well as for anesthesiologists (Table 5). All the anesthesiologists tried to keep the CVP low during liver dissection before clamping the vena cava. Also, efforts were made to keep all patients normothermic (Table 1). Perhaps some anesthesiologists better tolerated some iatrogenic hypotension without intervening too rapidly or maybe transfusing plasma at the beginning of the surgery increased volemia enough to increase bleeding.

Two factors influence blood loss: a mechanical or vascular component and a biochemical one. The latter is evaluated from the INR and the platelet count. In our center, because platelets are rare, we do not give any if the starting platelet count is more than 50 x 10⁹ pl/L. At the beginning of a transplant procedure, levels of coagulation factors I, V, VII, VIII, and the D-dimers were not always available. No attempt was made to correct any deficit. In fact, the coagulation functions were seldom monitored intraoperatively because of the short duration of surgery. Often, laboratory results were available only after the procedure was finished. Transfusions of blood products were managed taking the starting INR into consideration and following the clinical picture.

Of the two factors mentioned, the mechanical factor, controlled by the surgeon, seems to be the more important of the factors influencing the transfusion rate. In our series, there was one patient with a starting INR at 4.2, one with a starting Hb level of 81 g/L, and a third one with a starting platelet count of 12 x 10⁹/L. None of them received any blood product for their liver transplant, presumably because the blood losses were curtailed by the surgical expertise.

The starting Hb value was not predictive of why one subgroup received more RBC than another. In the group of patients who received fewer RBC, there were 97 patients who received between 1 and 4 RBC units. Some of these patients had low initial Hb, which indicated that perhaps this was not a predictive factor for transfusion. It would seem logical that if the initial Hb value were high, the RBC transfusion rate would decrease. Our review does not support this. Any strategy aimed at increasing the initial Hb value, such as using erythropoietin , might therefore not be so useful (14–16).

The severity of liver disease can be evaluated in different ways. The Pugh’s score indicates the severity of cirrhosis, somewhat more subjectively than the MELD score, a more objective indicator for liver failure, used to prioritize liver transplants in the United States.

Predictive factors for transfusion were hard to identify, possibly because there were very few patients in the group who received more than 4 RBC units (43 patients), as our average number of units transfused per patient was only 2.8. Perhaps we should have tried to analyze the factors associated with the performance of liver transplant without RBC transfusion (66 patients), as the Stanford group did (17). However, we did not feel the need to examine these factors, as 75% of our liver transplants are presently performed without transfusion of any blood products and the goal of the study was to determine variables that identify the need for transfusion.

In conclusion, in our series of 206 consecutive liver transplants done at the St Luc Hospital of the CHUM, the average number of transfused RBC units was 2.8 per patient. Regularly, some of our surgeons and anesthesiologists are able to transplant a liver without transfusing any blood products. The anesthesiologist’s and the surgeon’s experience and attitude seem to be more important than the correction of any biochemical variables during the liver transplant. Predictive factors for RBC transfusions remain difficult to define, probably because the human factors are more important. The starting INR is the most sensitive variable, although with low specificity, that will identify which patients will receive RBC. Transfusion of plasma in an effort to correct the INR does not prevent RBC transfusion; sometimes it was the contrary.

Acknowledgments

We would like to thank Mrs. D. Bois, L. Lacombe, and M. Raymond for the clerical support, Mr. G. Jean for the data processing support and Drs. R. Seal, L. Thibeault, C. Richard, and Dr. D. Beaulieu for their advice.

References

1. Carton EG, Rettke SR. Perioperative care of the liver transplant patient: part 1. Anesth Analg 1994; 78: 120–33.[Free Full Text]
2. Carton EG, Rettke SR. Perioperative care of the liver transplant patient: part 2. Anesth Analg 1994; 78: 382–99.[Medline]
3. Schroeder RA, Johnson LB, Plotkin JS, et al. Total blood transfusion and mortality after orthotopic liver transplantation. Anesthesiology 1999; 91: 329–30.
4. Steib A, Freys G, Lehmann C, et al. Intraoperative blood losses and transfusion requirements during adult liver transplantation remain difficult to predict. Can J Anaesth 2001; 48: 1075–9.[Abstract/Free Full Text]
5. Cacciarelli TV, Keeffe EB, Moore DH, et al. Effect of intraoperative blood transfusion on patient outcome in hepatic transplantation. Arch Surg 1999; 134: 25–9.[Abstract/Free Full Text]
6. Dalmau A, Sabaté A, Acosta F. Tranexamic acid reduces red cell transfusion better than epsilon-aminocaproic acid or placebo in liver transplantation. Anesth Analg 2000; 91: 29–34.[Abstract/Free Full Text]
7. Royston D, Bidstrup BP, Taylor KM, Sapsford RN. Effect of aprotinin on need for blood transfusion after repeat open-heart surgery. Lancet 1987; 2: 1289–91.[ISI][Medline]
8. Kang Y, Aggarwal S, Virji M. Clinical evaluation of autotransfusion during liver transplantation. Anesth Analg 1991; 72: 94–100.[Abstract/Free Full Text]
9. Findlay JY, Rettke SR, Ereth MH, et al. Aprotinin reduces red blood cell transfusion in orthotopic liver transplantation: a prospective, randomized, double-blind study. Liver Transpl 2001; 7: 802–7.[ISI][Medline]
10. Laxenaire MC, Dewachter P, Pecquet C. Allergic risk of aprotinin. Ann Fr Anesth Reanim 2000; 19: 96–104.[Medline]
11. O’Connor CJ, Roozeboom B, Brown R, Tuman KJ. Pulmonary thromboembolism during liver transplantation: possible association with antifibrinolytic drugs and novel treatment options. Anesth Analg 2000; 91: 296–9.[Abstract/Free Full Text]
12. Molenaar IQ, Porte RJ. Aprotinin and thromboembolism in liver transplantation: is there really a causal effect? Anesth Analg 2002; 94: 1367–8.[Free Full Text]
13. Baubillier E, Cherqui D, Dominique C, et al. A fatal thrombotic complication during liver transplantation after aprotinin administration. Transplantation 1994; 57: 1664–6.[ISI][Medline]
14. Baldry C, Backman SB, Metrakos P, et al. Liver transplantation in a Jehovah’s Witness with ankylosing spondylitis. Can J Anaesth 2000; 47: 642–6.[Abstract/Free Full Text]
15. Snook NJ, O’Beirne HA, Enright S, et al. Use of recombinant human erythropoietin to facilitate liver transplantation in a Jehovah’s witness. Br J anaesth 1996; 76: 740–3.[Abstract/Free Full Text]
16. Ramos HC, Todo S, Kang Y, et al. Liver transplantation without the use of blood products. Arch Surg 1994; 129: 528–32.[Abstract]
17. Cacciarelli TV, Keeffe EB, Moore DH, et al. Primary liver transplantation without transfusion of red blood cells. Surgery 1996; 120: 698–704.[Medline]

This article has been cited by other articles:

				N. Jabbour, S. Gagandeep, H. Shah, R. Mateo, M. Stapfer, Y. Genyk, L. Sher, M. Zwierzchoniewska, R. Selby, and G. Zeger Impact of a Transfusion-Free Program on Non-Jehovah's Witness Patients Undergoing Liver Transplantation Arch Surg, September 1, 2006; 141(9): 913 - 917. [Abstract] [Full Text] [PDF]

				P. E. Frasco, K. A. Poterack, J. G. Hentz, and D. C. Mulligan A Comparison of Transfusion Requirements Between Living Donation and Cadaveric Donation Liver Transplantation: Relationship to Model of End-Stage Liver Disease Score and Baseline Coagulation Status Anesth. Analg., July 1, 2005; 101(1): 30 - 37. [Abstract] [Full Text] [PDF]

				L. Massicotte, S. Lenis, L. Thibeault, M.-P. Sassine, R. F. Seal, and A. Roy Reduction of blood product transfusions during liver transplantation Can J Anesth, May 1, 2005; 52(5): 545 - 546. [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a colleague Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (15) Citing Articles via Google Scholar Google Scholar Articles by Massicotte, L. Articles by Roy, A. Search for Related Content PubMed PubMed Citation Articles by Massicotte, L. Articles by Roy, A. Related Collections Cardiovascular Economics and Health Care Research Blood Monitoring (Non-cardiac)

Anesthesia & Analgesia® is published for the International Anesthesia Research Society® by Lippincott Williams & Wilkins with the assistance of Stanford University Libraries' HighWire Press®. Copyright 2006 by the International Anesthesia Research Society. Online ISSN: 1526-7598   Print ISSN: 0003-2999