Objectives: The primary aim of this study was to determine whether specific preoperative clinical characteristics were associated with low-volume transfusion in liver transplant recipients. Low-volume transfusion was defined as transfusion of < 2100 mL of packed red blood cells intraoperatively during liver transplant. The ability to accurately predict low-volume transfusion could increase patient safety, decrease complications associated with transfusion, improve blood management, and decrease transplant case cost.

Materials and Methods: Data were retrieved by retrospective chart review of 266 patients who received a liver transplant at the Mayo Clinic (Jacksonville, FL, USA). The primary outcome was low-volume transfusion. Associations of preoperative information with low-volume transfusion were explored using single-variable and multivariable logistic regression models; missing data were imputed with the sample median for continuous data and the most frequent category for categorical variables.

Results: Low-volume transfusion occurred in 23% of first-time liver transplant recipients (62/266 patients; 95% confidence interval, 18%-29%). History of hepatitis C virus infection (P = .048), history of hepatocellular carcinoma (P = .050), short cold ischemia time (P = .006), and low international normalized ratio (P = .002) were independently associated with low-volume transfusion during liver transplant in a multivariable logistic regression model.

Conclusions: Multiple studies have shown increased morbidity and mortality after orthotopic liver transplant when more than 6 U of packed red blood cells are administered within 24 hours of surgical incision. A method to identify low-volume transfusion candidates could help predict patient outcomes, decrease blood handling, and reduce costs. If patients with low-volume transfusion could be identified, fewer blood products would need to be prepared in advance. Although elevated preoperative coagulation parameters decrease the probability of low-volume transfusion, a definitive profile of a low-volume transfusion liver transplant recipient was not established.

Key words : Blood coagulation, International normalized ratio, Patient safety, Retrospective studies

Introduction

Orthotopic liver transplant (OLT) is currently the only known curative therapy for patients with end-stage liver disease.1 Blood transfusions are an expensive but necessary component of OLT because this surgery has historically been characterized by significant and sometimes unpredictable blood loss.^1-5 When the blood bank at our institution is notified that an OLT will occur, they immediately prepare 10 U of packed red blood cells (PRBCs) for the transplant recipient.⁶ Although preparation of 10 U of PRBCs is a common approach in liver transplant centers, there is no standard for blood product preparation, and many centers have developed alternative approaches to massive tranfusion.^3,6

Blood transfusions in the United States have been found to be one of the most overused medical therapies.⁷ As with all medical therapies, inherent risk is present, and the possibility for transfusion reactions and administration errors exists.⁸ Packed red blood cell transfusions have been shown in some cases to lead to severe complications, such as immuno­suppression, sepsis, and acute lung injury.^1,3,5,7-10 It is well documented that a major predictor of survival after OLT is the volume of transfused blood.^1,9,11-14 Massive blood transfusion is commonly defined as transfusion of more than 6 U (> 2100 mL) of PRBCs within 24 hours of skin incision.^9,11,14 Multiple studies have demonstrated that patients who received massive blood transfusions during OLT have longer hospital stays, increased morbidity, and decreased survival rates.^{1,2,5,9,11-14} Although advancements in anesthetic and surgical care have decreased intraoperative PRBC transfusion requirements, many liver transplant patients receive at least 1 U of PRBCs during surgery.^{1,2,5,9,11,13,15} The administration of just 1 U has been associated with a 30% increase in morbidity in certain patient populations.14

Many studies have investigated potential preoperative characteristics that are associated with massive blood loss and therefore necessitate massive blood transfusions in OLT.^{1,4-6,9,12-14,16-19} Factors investigated vary by study and institution but include laboratory values of the recipient and donor, medical history of the organ donor, and recipient preoperative characteristics. Mondanlou and as­sociates found that a transplant recipient’s pre­operative creatinine level of greater than 1.3 mg/dL positively correlated with massive blood transfusion.⁶ The Model for End-Stage Liver Disease score is a known method of classifying disease severity, and higher Model for End-Stage Liver Disease scores may correlate with increased transfusion requirements.^9,13,17 Rana and associates found a positive association between warm ischemia time, bilirubin levels, and PRBC transfusion requirements. This study also showed that bleeding associated with prolonged native hepatectomy was strongly correlated with a patient history of previous major abdominal surgery.²⁰ Araújo and associates designed a study to assess the ability to predict intraoperative PRBC transfusion from preoperative information. They determined that statistical associations exist between certain variables and a need for PRBC transfusion, but the predictive power was minimal due to limited sensitivity.⁴ An algorithm that is able to accurately predict the likelihood of low-volume intraoperative blood loss has yet to be developed.^6,18,19 Such a tool would enable the tailoring of blood product crossmatching, allocation, and administration to each individual patient.^6,14

The purpose of this study was to determine whether specific preoperative clinical characteristics were associated with low-volume transfusion (LVT), which we defined as transfusion of < 2100 mL of PRBCs intraoperatively. Accurately predicting LVT would increase patient safety, decrease complications associated with transfusion, decrease blood handling, improve blood management, save time, and decrease costs.^{6,8,9,16,17,19} Accurate prediction of intraoperative bleeding would not only benefit patients but would also enable better use and coordination of valuable resources, including crossmatching, PRBC allocation, and decreased transfusion department labor.^{4,6,9,12,13,17-19}

Materials and Methods

The study was conducted according to the guidelines of the Declaration of Helsinki, and the study protocol was approved prior to the beginning of the study by our ethics committee. Our aim was to identify independent preoperative clinical characteristics that could predict LVT during liver transplant. Potential predictors included age, sex, history of hepatitis C virus (HCV) infection, history of hepatocellular carcinoma (HCC), continuous venovenous hemo­dialysis, portal vein thrombosis, previous upper abdominal surgery, transjugular intrahepatic portosystemic shunt, Model for End-Stage Liver Disease score, cold ischemia time, and preoperative laboratory measures (creatinine, prothrombin time, international normalized ratio [INR], activated partial thromboplastin time, fibrinogen, and platelet count).

Study patients and data collection
All patients who received a liver transplant at the Mayo Clinic in Jacksonville, FL, USA from January 1, 2012, through December 31, 2013, were included in this retrospective cohort study. Patients were excluded if they had a multiorgan transplant, had prior liver transplant, or were younger than 18 years old.

The primary outcome measure was LVT. Other variables collected included age, sex, race, body mass index at listing, wait time, primary diagnosis, history of HCV infection, history of HCC, continuous venovenous hemodialysis, portal vein thrombosis, previous abdominal surgery, transjugular intrahepatic portosystemic shunt, Model for End-Stage Liver Disease, donation after cardiac death status (yes or no), and cold ischemia time. In addition, the last laboratory measurement before transplant was obtained for creatinine level, prothrombin time, INR, activated partial thromboplastin time, fibrinogen concentration, and platelet count.

Patient information was retrieved from a liver transplant database that is kept and maintained by the Department of Transplantation at the Mayo Clinic in Jacksonville, FL, USA. Study data were collected and managed using REDCap (Research Electronic Data Capture)21 tools hosted at our institution.

Statistical analyses
Numerical variables were summarized with the sample median, minimum, and maximum, and categorical variables were summarized with number and percentage. The proportion of patients with a LVT was estimated along with an exact binomial 95% confidence interval. Associations of preoperative information with LVT were explored using single-variable and multivariable logistic regression models; missing data were imputed with the sample median for continuous data and the most frequent category for categorical variables. Multivariable models were adjusted for those variables showing the strongest association in single-variable analysis allowing for no more than 1 variable in the model for every 10 patients who had LVT. Because of their skewed distributions, laboratory measures were transformed on the logarithm scale. International normalized ratio, prothrombin time, and activated partial thromboplastin time were strongly correlated with each other; therefore, only INR was included in logistic regression analysis for purposes of data reduction and avoiding collinearity. Odds ratios (OR) with 95% confidence intervals were estimated; ORs for continuous variables were given per 1 standard deviation (SD) increment. P values ≤ .05 were considered statistically significant without adjustment for multiple testing. All analyses were performed using SAS (version 9.3, SAS Institute Inc., Cary, NC, USA) and R statistical software (version 2.14, R Foundation for Statistical Computing, Vienna, Austria).

Results

Characteristics of the 266 liver transplant recipients are described in Table 1. The median age was 60 years (range, 22-75 y), with 184 (69%) who were male. Sixty-two patients (23%) had a LVT during liver transplant (95% confidence interval, 18%-29%). Exploratory associations of preoperative charac­teristics with LVT are presented in Tables 2 and 3. In single-variable logistic regression analysis, LVT was associated with male sex (OR = 2.18; P = .028), HCV (OR = 2.34; P = .004), HCC (OR = 4.81; P < .001), lower cold ischemia time (OR [1 SD decrease] = 1.22; P = .004), lower creatinine (OR [1 SD decrease, log scale] = 1.50; P = 0.014), lower INR (OR [1 SD decrease, log scale] = 3.28; P < .001), higher fibrinogen (OR [1 SD increase, log scale] = 1.80; P < 0.001), and higher platelet count (OR [1 SD increase] = 1.53; P = .005). No other variables were significantly associated with LVT in single-variable analyses (all P ≥ .066).

In multivariable logistic regression analyses (Table 2) adjusting for HCV, HCC, cold ischemia time, INR, and fibrinogen, LVT was associated with HCV (OR = 2.09; P = .048), HCC (OR = 2.09; P = .050), lower cold ischemia time (OR [1 SD decrease]; P = .006), and lower INR (OR [1 SD decrease, log scale] = 2.59; P = .002). Fibrinogen was no longer statistically significant (P = .87). Predicted probabilities of LVT were obtained from a multivariable logistic regression model, including HCV, HCC, cold ischemia time, INR, and fibrinogen (Figure 1).

Discussion

We describe a method of preoperative identification of several risk factors that can be used to identify potential LVT candidates. These variables can assist in guiding blood use strategies when used in conjunction with sound clinical judgment. Current practice dictates type and crossmatch of ample blood products for massive transfusion and correction of coagulopathy. Orthotopic liver transplant remains a formidable surgical operation, and PRBC transfusion volume is a known independent predictor of postoperative outcomes.^{2,5,9,10-12,14,16,22} Modern efforts to minimize PRBC transfusions during OLT may reduce risks to the patient, such as transfusion-related lung injury, immunomodulation, and iatrogenic infection. An LVT risk index tool could result in multiple benefits, including improving patient outcomes, containing transplant costs, and conserving valuable blood bank time and resources.^3,5,8-10

Here, we present data from consecutive liver transplant recipients from a high-volume academic transplant center, with variables easily available in the preoperative period and which could be obtained before surgery at other liver transplant programs for replication. Inconsistencies in transfusion practices exist among individual providers and between institutions.^3,8-10,16,17 Consensus on protocols for PRBC transfusion remains minimal and indivi­dualized.^{3,5,9,10,16,17} Thus, there is a need for evidence-based practice and standardization of care for this highly controversial topic because patients may derive benefits from a restrictive transfusion policy for OLT.^3-5,8-10,16 With modern intraoperative mana­gement, it is now possible to perform OLT without the administration of a single unit of PRBCs.^2,9

The current practice at our institution is to type and crossmatch 10 U of PRBCs, in addition to fresh frozen plasma for each OLT patient regardless of medical history or preoperative laboratory values. This is a time-consuming process for both blood bank personnel and anesthesia providers. When reviewing the existing literature, we found that our practice mirrors the massive blood transfusion protocols of many other OLT programs.⁶ Given the costs of PRBC and type and crossmatch procedures, it is easy to see how quickly the cost of a transplant can increase when the massive blood transfusion protocol is used. Blood product expenses for an OLT surgery may account for up to 10% of total transplant costs, depending on the case and institution. Methods of reducing transfusion can significantly affect the overall cost, thereby benefitting both the patient and the facility.¹⁷

Confirmation and checking of blood products in the operating room occur when anesthesia providers are busy starting the case, inducing the patient, and placing necessary intravenous access devices and monitoring equipment. Time spent on PRBC compatibility confirmation can draw the provider’s attention away from hands-on patient care. This may not be the most beneficial and efficient strategy for this particular patient population, as we typically transfuse an average of 6 U PRBCs per case.^6,8,16

Currently, liver transplant patients receive transfusions at the discretion of the individual provider, according to estimates of intravascular needs during volatile cases with dynamic fluid shifts and acute hemorrhage.^3,8-10,16 Although the decision to use a particular transfusion protocol has traditionally been subjective, this study aimed to provide an objective system to determine transfusion needs that may be useful when combined with clinical judgment. This approach could help delineate which recipients are at low-risk for heavy blood loss and could guide providers in their choices of transfusion protocols and practices, ultimately benefitting the patients. We recognize that a single, universally applicable method to determine LVT candidates may be difficult to develop and that further validation with multicenter prospective analyses is needed.^12,18,19

Successful prediction of LVT candidates depends not only on the patient’s preexisting comorbidities and pretransplant disposition but also on the status of their physiologic reserve. It is important to note that intraoperative events may alter LVT candidacy at any point, and antibody status with blood compatibility should be carefully considered to ensure that adequate reserves exist. The fundamental idea behind LVT candidacy is that anesthesia care can be tailored to meet individual patient needs. Each OLT recipient is unique, with a specific set of requirements, and a standard massive blood transfusion protocol may not be appropriate for all recipients. The physiologic and cost-benefit advantages of a limited transfusion practice protocol have been clearly established.^8,16,17 Low-volume transfusion may improve both graft and patient survival, and it is hoped that successful prediction of LVT candidates could be achieved in the near future without compromising patient safety.^{1,2,5,9,11-13}

Blood bank resources are a valuable commodity and should be conserved. Blood products assigned to an OLT patient are not available to other patients in the hospital who may need transfusions, which may result in PRBC shortages. These assigned units may be wasted if returned to the blood bank and not used before the expiration date or if they are accidentally allowed to become too warm in the operating room storage coolers.⁶ The ability to determine which preoperative variables predict LVT may improve use of blood bank resources and local and regional blood allocation.¹²

Limitations of this study include restriction to a single center large liver transplant facility with a relatively small number of patients. In addition, this study was retrospective in nature. Because of the high-risk nature of this challenging operation, a prospective randomized trial may not be possible. Clinical judgment is a crucial, imperative, and irreplaceable component of blood transfusion management strategies.²³ Prudence in risk-benefit analysis is needed in the decision-making stratification of PRBC transfusion practices.²⁴ Studies aimed at predicting transfusion needs in OLT patients may yield conflicting results when using preoperative data. This may be due in part to differing characteristics of regional patient populations and varying transplant center data. Individual centers should strive to create a more efficient use of transfusion services and attempt to conserve valuable resources by evaluating their individual practices.⁶

Results of our study found several factors that were independently associated with LVT, including history of HCV infection and HCC, lower cold ischemia time, and lower INR. Because of the relatively small sample size and exploratory nature of our study, additional larger studies are needed to validate these findings. Predictive power was ultimately found to be limited and inconsistent because certain patients experienced massive blood transfusions even when HCV, HCC, lower cold ischemia time, and lower INR were all present in the same case. The possibility of catastrophic intraoperative blood loss and necessary massive blood transfusion requirements remains a real possibility.²³ Given the potential risk to patients of insufficient blood available during transplant, it is hoped that future studies could assist in developing a prediction tool to estimate the actual volume of blood needed based on patient characteristics and laboratory measures. This would give surgical teams a degree of certainty that a sufficient supply will be available. This study aimed to predict LVT in an effort to minimize the overabundance of blood prepared for OLT surgical cases.¹⁸

The assessment of LVT candidacy based on preoperative clinical characteristics is an interesting area of research that could benefit OLT patients. We have found that preoperative variables remain statistically significant risk factors and can be useful, along with clinical judgement, when a provider is determining which patients are at low risk for hemorrhage. Significant associations between LVT and certain preoperative variables do exist; however, predictive power was limited because correlation does not imply causation.²⁴ Blood loss in OLT patients remains difficult and unpredictable.¹² Given this, clinicians must continue to prepare large quantities of blood before surgery to provide a margin of safety for these patients.²³ This study failed to identify preoperative patient characteristics with a strong enough relation to LVT to change our current practice of blood preparation. Until there is an alternative to autologous blood transfusion, anesthesia providers have no choice but to prepare for massive blood transfusions for every liver transplant.

References:

1. Esmat Gamil M, Pirenne J, Van Malenstein H, et al. Risk factors for bleeding and clinical implications in patients undergoing liver transplantation. Transplant Proc. 2012;44(9):2857-2860.
   CrossRef - PubMed
2. Ramos E, Dalmau A, Sabate A, et al. Intraoperative red blood cell transfusion in liver transplantation: influence on patient outcome, prediction of requirements, and measures to reduce them. Liver Transpl. 2003;9(12):1320-1327.
   CrossRef - PubMed
3. Chidananda Swamy M. Blood transfusion practices in liver transplantation. Indian J Anaesth. 2014;58(5):647-651.
   CrossRef - PubMed
4. Araújo T, Cordeiro A, Proença P, Perdigoto R, Martins A, Barroso E. Predictive variables affecting transfusion requirements in orthotopic liver transplantation. Transplant Proc. 2010;42(5):1758-1759.
   CrossRef - PubMed
5. Lopez-Plaza I. Transfusion guidelines and liver transplantation: time for consensus. Liver Transpl. 2007;13(12):1630-1632.
   CrossRef - PubMed
6. Modanlou KA, Oliver DA, Grossman BJ. Liver donor's age and recipient's serum creatinine predict blood component use during liver transplantation. Transfusion. 2009;49(12):2645-2651.
   CrossRef - PubMed
7. Goodnough LT, Maggio P, Hadhazy E, et al. Restrictive blood transfusion practices are associated with improved patient outcomes. Transfusion. 2014;54(10 Pt 2):2753-2759.
   CrossRef - PubMed
8. Hannaman MJ, Hevesi ZG. Anesthesia care for liver transplantation. Transplant Rev (Orlando). 2011;25(1):36-43.
   CrossRef - PubMed
9. Clevenger B, Mallett SV. Transfusion and coagulation management in liver transplantation. World J Gastroenterol. 2014;20(20):6146-6158.
   CrossRef - PubMed
10. Benson AB, Burton JR, Austin GL, et al. Differential effects of plasma and red blood cell transfusions on acute lung injury and infection risk following liver transplantation. Liver Transpl. 2011;17(2):149-158.
    CrossRef - PubMed
11. Boin IF, Leonardi MI, Luzo AC, Cardoso AR, Caruy CA, Leonardi LS. Intraoperative massive transfusion decreases survival after liver transplantation. Transplant Proc. 2008;40(3):789-791.
    CrossRef - PubMed
12. Cywinski JB, Alster JM, Miller C, Vogt DP, Parker BM. Prediction of intraoperative transfusion requirements during orthotopic liver transplantation and the influence on postoperative patient survival. Anesth Analg. 2014;118(2):428-437.
    CrossRef - PubMed
13. Varotti G, Santori G, Andorno E, et al. Impact of Model for End-Stage Liver Disease score on transfusion rates in liver transplantation. Transplant Proc. 2013;45(7):2684-2688.
    CrossRef - PubMed
14. Mccluskey SA, Karkouti K, Wijeysundera DN, et al. Derivation of a risk index for the prediction of massive blood transfusion in liver transplantation. Liver Transpl. 2006;12(11):1584-1593.
    CrossRef - PubMed
15. Massicotte L, Denault AY, Beaulieu D, et al. Transfusion rate for 500 consecutive liver transplantations: experience of one liver transplantation center. Transplantation. 2012;93(12):1276-1281.
    CrossRef - PubMed
16. Goldaracena N, Méndez P, Quiñonez E, et al. Liver transplantation without perioperative transfusions single-center experience showing better early outcome and shorter hospital stay. J Transplant. 2013;2013:649209.
    CrossRef -
17. Feltracco P, Brezzi M, Barbieri S, et al. Blood loss, predictors of bleeding, transfusion practice and strategies of blood cell salvaging during liver transplantation. World J Hepatol. 2013;5(1):1-15.
    CrossRef - PubMed
18. Massicotte L, Capitanio U, Beaulieu D, Roy JD, Roy A, Karakiewicz PI. Independent validation of a model predicting the need for packed red blood cell transfusion at liver transplantation. Transplantation. 2009;88(3):386-391.
    CrossRef - PubMed
19. Ghaffaripour S, Mahmoudi H, Khosravi M, et al. Preoperative factors as predictors of blood product transfusion requirements in orthotopic liver transplantation. Prog Transplant. 2011;21(3):254-259.
    CrossRef - PubMed
20. Rana A, Petrowsky H, Hong JC, et al. Blood transfusion requirement during liver transplantation is an important risk factor for mortality. J Am Coll Surg. 2013;216(5):902-907.
    CrossRef - PubMed
21. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)- A metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. 2009;42(2):377-381.
    CrossRef - PubMed
22. Bulatao IG, Heckman MG, Rawal B, et al. Avoiding stay in the intensive care unit after liver transplantation: a score to assign location of care. Am J Transplant. 2014;14(9):2088-2096.
    CrossRef - PubMed
23. Roullet S, Biais M, Millas E, Revel P, Quinart A, Sztark F. Risk factors for bleeding and transfusion during orthotopic liver transplantation. Ann Fr Anesth Reanim. 2011;30(4):349-352.
    CrossRef - PubMed
24. Massicotte L, Sassine M, Lenis S, Seal RF, Roy A. Survival rate changes with transfusion of blood products during liver transplantation. Can J Anesth. 2005;52(2):148-155.
    CrossRef - PubMed