|Year : 2019 | Volume
| Issue : 2 | Page : 119-125
Predicting packed red blood cell transfusion in living donor liver transplantation: A retrospective analysis
Shweta A Singh1, Kelika Prakash2, Sandeep Sharma2, An Anil2, Viniyendra Pamecha3, Guresh Kumar4, Ajeet Bhadoria5
1 Director and Head Anaesthesia and Critical Care at Centre of Liver and Bilary Sciences, Max Super Speciality Hospital, Saket, Formerly Additional Professor Anesthesiology at ILBS, New Delhi, India
2 Department of Anesthesiology, ILBS, New Delhi, India
3 Department of Transplant and HPB Surgery, ILBS, New Delhi, India
4 Department of Statistics, ILBS, New Delhi, India
5 Department of Epidemiology, ILBS, New Delhi, India
|Date of Web Publication||11-Feb-2019|
Dr. Shweta A Singh
Director and Head Anaesthesia and Critical Care at Centre of Liver and Bilary Sciences, Max Saket Super Speciality Hospital, Saket, New Delhi
Source of Support: None, Conflict of Interest: None
Background and Aims: Blood transfusion is unpredictable in liver transplantation and is associated with increased patient morbidity, mortality and cost. This retrospective analysis was conducted to detect factors which could predict intraoperative transfusion of more than four units of packed red blood cells (PRBCs) during elective living donor liver transplantation (LDLT). Methods: This was a single-centre retrospective study. Demographic, clinical and intraoperative data of 258 adult patients who underwent LDLT from March 2009 to January 2015 were analysed. Univariate and multivariate regression model was used to identify factors responsible for transfusion of more than four PRBCs (defined as massive transfusion [MT]). Results: On univariate regression analysis, preoperative factors like aetiology of liver disease, hypertension, history of spontaneous bacterial peritonitis, low haemoglobin and fibrinogen, high serum bilirubin, high blood urea and creatinine, high model for end-stage liver disease score, portal venous thrombosis, increased duration of surgery and anhepatic phase as well as increased use of other blood products were found to be significantly associated with MT. Multivariate logistic regression analysis revealed that the only independent factor associated with MT was the number of units of fresh frozen plasma transfused (odds ratio = 1.54 [95% CI (1.12–2.12)]). Conclusion: Many factors are responsible for the need for transfusion during LDLT. Preoperative factors alone do not accurately and consistently predict the need for MT as in our study. It is important to be prepared for need for MT during each transplant.
Keywords: Blood transfusion, liver transplantation, living donors
|How to cite this article:|
Singh SA, Prakash K, Sharma S, Anil A, Pamecha V, Kumar G, Bhadoria A. Predicting packed red blood cell transfusion in living donor liver transplantation: A retrospective analysis. Indian J Anaesth 2019;63:119-25
|How to cite this URL:|
Singh SA, Prakash K, Sharma S, Anil A, Pamecha V, Kumar G, Bhadoria A. Predicting packed red blood cell transfusion in living donor liver transplantation: A retrospective analysis. Indian J Anaesth [serial online] 2019 [cited 2020 Jul 4];63:119-25. Available from: http://www.ijaweb.org/text.asp?2019/63/2/119/251967
| Introduction|| |
Massive transfusion (MT) during liver transplantation results in significant increase in morbidity, mortality and cost of treatment.,,,,, Factors associated with increased blood transfusion include pre-existing coagulopathy,, previous abdominal surgery, low haemoglobin, high model for end-stage liver disease (MELD) score, duration of surgery and portal venous thrombosis (PVT). These factors, however, do not predict intraoperative transfusion consistently and studies have been mostly conducted in a deceased donor setting., The aim of this study was to detect factors which predict transfusion of more than four units of packed red blood cells (PRBCs) during living donor liver transplantation (LDLT).
| Methods|| |
This was a single-centre retrospective study. After Institutional Ethics Committee approval, data of patients more than 18 years of age who underwent elective LDLT for chronic liver disease (CLD) between March 2009 and January 2015 were retrospectively analysed. Patients who received a right lobe graft with middle hepatic vein (MHV) were included in the study. All patients were operated by the same surgical unit using a standardised technique. Portocaval shunt was made during anhepatic phase and piggy-back technique was used during implantation instead of venovenous bypass.
The anaesthesia protocol was also uniform. General anaesthesia was induced using intravenous fentanyl and propofol. Atracurium was used as the non-depolarising muscle relaxant. Rapid sequence intubation using rocuronium was reserved for patients with refractory ascites. Anaesthesia was maintained with isoflurane (MAC 0.8–1) along with iv infusions of fentanyl and atracurium. Fluid and inotrope administration were guided by pulse contour-based cardiac output monitoring using FloTrac™. The target mean arterial pressure was 65 mmHg and stroke volume variation (SVV) <12%. Methylprednisolone was administered to all patients prior to reperfusion intravenous boluses of phenylephrine and/or adrenaline was used to manage the post reperfusion syndrome. All patients were shifted sedated on mechanical ventilation to transplant ICU for further stabilisation, monitoring and management.
A uniform protocol was followed for transfusion of blood products. Leucodepleted PRBCs were transfused to maintain a target haematocrit of 24. Thromboelastography (TEG™) was used to guide transfusion of fresh frozen plasma (FFP), platelets and cryoprecipitate in the presence of oozing as per the departmental protocol. About 2 units of FFP were transfused if the R on TEG was >9 but <14 min while 4 units were transfused if R >14 min. If the angle was <45°, 4–6 units of cryoprecipitate were transfused. For an MA <55 mm, 1 unit of single donor platelet concentrate was transfused.
Data collected included patient demographics, aetiology of liver disease, history of decompensation, previous abdominal surgery, presence of PVT, spontaneous bacterial peritonitis (SBP), MELD score and preoperative laboratory parameters. Intraoperative details collected included total anhepatic time, cold and warm ischaemia times (WIT) and intraoperative blood products transfused.
Patients were grouped into two groups based on the number of units of PRBCs transfused intraoperatively. Patients who received ≤4 units of PRBC were allocated to group A, whereas those who received >4 units of PRBC were allocated to group B. Data have been analysed using SPSS version 22. Numerical data are presented as median with range of minimum and maximum values or as mean with standard deviation as applicable. Categorical data are expressed in number with the corresponding percentage. Chi-square test or Fisher exact test was used for categorical variables as applicable. Normally distributed continuous variables were analysed using the Student t-test, whereas Mann–Whitney U test was used for non-parametric data. Univariate and multivariate logistic regression analysis was performed to determine the significance and odds ratio for various preoperative factors and intraoperative parameters which could predict intraoperative PRBCs transfusion. The results are presented as odds ratio with 95% confidence interval for all parameters. P value <0.05 was considered significant.
| Results|| |
A total of 258 adult patients underwent LDLT (right lobe) for CLD during this period. About 79 patients received less than 4 units of PRBC transfusion and were thus placed in group A, and 179 out of 258 patients received more than 4 units of PRBC transfusion and were thus placed in group B. In our series, the graft-to-recipient weight ratio (GRWR) in both the groups varied between 0.8 and 1.1. [Table 1] shows the demographic and clinical factors analysed with their corresponding P values. The age and BMI were comparable between the two groups. In both groups A and B, there were more male patients as compared to female patients. In both groups, the most common aetiology was alcohol (44.3% in group A vs. 59.88% in group B) followed by viral hepatitis (hepatitis B and C). The prevalence of diabetes mellitus, hypertension and AKI was similar in the two groups. Significantly greater number of patients had a history of SBP and HRS in group B when compared to group A (27.4% and 19.2% in group B vs. 10.1% and 12.7% in group A, respectively). The difference in the mean MELD scores between patients in group A and B was around 3 and the MELD Na was comparable in both groups. Laboratory investigations for both groups are shown in [Table 2]. Among laboratory investigations, significantly higher serum bilirubin, blood urea and serum creatinine values were seen in group B. The mean haemoglobin was also around 2 g/dl lower in group B which was statistically significant (10.06 ± 1.85 mg/dl in group A and 8.74 ± 0.1.5 mg/dl in group B, P= 0.008). Baseline TEG parameters, that is, the median R time, K time, alpha angle and MA, are shown in [Table 2]. The median R time in group B was 8.60 (3.9, 22.7) min which was significantly higher than that in group A [7.5 (1.6, 15.2) min]. MA was 45.65 (12.9, 69.9) mm in group B which was significantly lower than that of group A. Both the K time and the alpha angle were comparable between the two groups. Only two patients had fibrinolysis in group A and three in group B. Intraoperatively, patients in group B had a significantly longer duration of anhepatic phase and a longer surgical time [Table 2]. The cold ischaemia time was 90 (20–260) min in group A and 95 (30–380) min in group B and the WIT was 42.84 ± 12.1 min in group A and 46.62 ± 12.88 min in group B [Table 2]. The median number of FFPs, cryoprecipitates and SDPCs transfused was higher in group B when compared to group A.
The results of univariate analysis performed by logistic regression are described below:
It was seen that a higher male-to-female ratio was associated with higher odds of transfusion of more than four PRBCs [Table 3]. Out of all the causes of CLD, alcoholic or viral aetiologies had a significantly higher odds ratio. Alcohol induced and viral aetiology predicted the need for MT with an OR of 3.87 (1.36–10.99) and 3.5 (1.11–10.02), respectively. A history of hypertension [OR 4.47 (0.95–21.05)] was also associated with greater PRBC requirement.
On univariate analysis, the presence of SBP was a predictive factor of increased blood transfusion with an odds ratio of 3.34 (1.42–7.9) [Table 1]. Only 1 patient out of 79 patients in group A had PVT as against 7 out of 179 in group B [Table 1]. High MELD score and the presence of PVT were predictors of MT with an OR of 1.06 (1.01–1.12) and 5.9 (0.7–49.76), respectively. MELD Na, history of variceal bleed, prior abdominal surgery, hepatic encephalopathy, hydrothorax, hepatocellular carcinoma and hepatic vein portal gradient did not predict MT [Table 3].
The results of univariate analysis of laboratory investigations are shown in [Table 4]. On univariate analysis, low haemoglobin [OR 0.63 (0.51–0.77)] and high serum bilirubin levels [OR 1.05 (1.009–1.0840)] emerged as predictors of MT. In addition, significant predictors were high blood urea and low serum fibrinogen levels. Interestingly, deranged INR and a low platelet count or serum albumin levels did not predict MT. Increase in R time was associated with a 23% higher chance of transfusion of >4 PRBCs. Higher MA was associated (P = 0.007) with transfusion of less than 4 PRBCs as was a larger alpha angle (P = 0.042). K time was not significant on univariate analysis.
Among surgical factors, the duration of anhepatic phase as well as that of total surgery was associated with more transfusion. It was seen that patients who had received more platelet, FFP and cryoprecipitate transfusions also had more PRBC transfusion [Table 4].
Forward stepwise multivariate logistic regression analysis was applied to the predictors found to be significant on univariate analysis. This helped identify independent predictors of MT. It was seen that increased transfusion of FFP was associated with MT with an odds ratio of 1.55 [95% CI (1.26–1.90), P= 0.001] and the predictive value of this model was found to be 80.5%.
| Discussion|| |
Intraoperative blood transfusion is associated with increased morbidity (acute kidney injury, sepsis) as well as mortality.,,, Moreover, arranging blood products leads to increased utilisation of blood bank resources and finances. In this regard, identifying predictors of MT is an area of continuous research. Considering 1-year graft survival is decreased in recipients receiving more than four units of blood transfusion, we conducted this retrospective analysis to identify factors which could predict the need for transfusion of more than four PRBCs intraoperatively. Forty-three perioperative factors of patients who underwent LDLT over 5 and half years at a single centre in India were analysed. Of these alcohol- and viral-related CLD, a history of SBP, hepatic encephalopathy, hypertension, a high MELD score, PVT, low haemoglobin, high bilirubin, high blood urea and creatinine, low fibrinogen, increased duration of surgery, anhepatic time and greater utilisation of FFP and other blood products were significant predictors of MT on univariate analysis.
There is a lot of heterogeneity with regard to the transfusion trigger and the number of units of PRBC transfusion analysed.,,,,, For example, Roullet et al. had analysed the predictors for more than 1 volume of blood loss and Steib et al. analysed the need for more than 12 units of RBC transfusion. Similarly the transfusion trigger ranged between 6 and 10 g/dl in various studies., It was seen in the present study that 69.38% patients received more than 4 units of PRBC transfusions. The average haemoglobin level (Hb) was 10.06 ± 1.85 g/dl in group A and 8.74 ± 1.5 g/dl in group B. Each institution has its own transfusion protocol. At our centre, transfusion of blood products was based on the presence of clinical ooze in the surgical field guided by TEG™ and the threshold for PRBC transfusion was an Hb of 8 g/dl. On analysis, haemoglobin was not found to be a predictor for MT. This was perhaps because the average Hb of our patients was higher than our transfusion trigger.
Since patients with CLD have a rebalanced status of coagulopathy, using algorithms based on viscoelastic tests (i.e., TEG and ROTEM) rather than conventional tests (i.e., platelet count and INR) have been found to be associated with lower intraoperative PRBC transfusions.,, Though Esmat Gamil et al. found INR to be a predictive factor for transfusion of more than 6 units of PRBCs, our analysis in concordance with many other retrospective analysis, did not find a prolonged INR value to be predictive of MT. On univariate analysis of baseline TEG, deranged R time, MA as well as alpha angle were associated with greater PRBC transfusion as has been shown earlier.
MELD score has been inconsistently found to be a predictor of massive blood transfusion probably because INR value is used to derive MELD score., In the current analysis too, MELD score was found to be significant on univariate analysis but not on multivariate analysis. Similarly, we found that low fibrinogen levels were also significant only on univariate analysis. In a study conducted by Findlay et al., high serum bilirubin was an independent predictor, but on literature review, it is noted that this finding has not been replicated in other studies. In concordance with this, in the present study too, high bilirubin level was not found to be significant on multivariate analysis. Though the presence of PVT (requiring intraoperative thrombectomy) and adhesions due to previous abdominal surgery lead to significant blood loss, the number of patients having PVT or prior abdominal surgery in the current study was too low to draw any statistical inference. In this present study, the duration of surgery significantly predicted the need for greater PRBC transfusion on univariate analysis as reported previously.
Preoperative variables show poor correlations, and in spite of many attempts, predicting intraoperative PRBC transfusion remains difficult.,, In the retrospective analysis conducted by Roullet et al., preoperative haemoglobin, Child class and a history of oesophageal bleeding were risk factors on multivariate analysis. McCluskey et al. developed a risk index based on age, preoperative haemoglobin, platelets, INR, creatinine and albumin to predict blood loss, which was validated by Escoresca et al. However, this study was conducted during the initial stages of the liver transplant programme at their centre. The authors acknowledge that the number of PRBC transfusions required intraoperatively has since reduced which could be a result of improvement in operative skills. Deakin et al. also suggested that the experience of the surgical team has got high influence on intraoperative blood loss and subsequent transfusion.
In this present study, the only independent factor on multivariate analysis associated with MT was the number of units of FFPs transfused. Viscoelastic tests like ROTEM and TEG are commonly used to assess coagulation and guide blood product transfusion. Indeed, in a single-centre study, maximum amplitude less than 47 mm on TEG predicted the need for MT in liver transplantation. In our centre, FFP and other blood product transfusions were guided by TEG in the presence of clinical oozing. On no occasion was FFP transfused as a fixed ratio with PRBC. Perhaps, the use of prothrombin complex concentrates instead of FFP can reduce PRBC transfusion by reducing losses associated with fluid overload. In a recent study published from Mayo Clinic, the authors stated that the predictors of transfusion did not appear to be strong enough to change the practice of preoperative preparation of blood products. This present study in recipients of LDLT also failed to identify factors which could consistently predict the need for MT preoperatively. With growing experience of transfusion, efforts could be made to identify patient characteristics preoperatively who could have a LT with low-volume transfusion.
This study has a few limitations. We did not take into account a change in transfusion trend which could have occurred over 5 years with improvement of surgical and anaesthetic skills. Moreover, the study did not focus on the role of central venous pressure during intraoperative management. In spite of these limitations, it must be noted that unlike most studies which have been conducted in the setting of deceased donor liver transplantation, the present study evaluated predictors during LDLT where a split graft is used. Further research on volume restrictive strategies incorporating FFP substitutes like prothrombin complex concentrates and algorithms targeting higher transfusion triggers for blood products are needed.
| Conclusion|| |
In this study conducted over 5 and a half years in an LDLT setup, it was difficult to identify predictors of MT. Until conclusive data emerge from further research, it continues to be imperative to anticipate blood loss and arrange adequate blood and blood products for each liver transplantation.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Feltracco P, Brezzi ML, Barbieri S, Galligioni H, Milevoj M, Carollo C, et al
. Blood loss, predictors of bleeding, transfusion practice and strategies of blood cell salvaging during liver transplantation. World J Hepatol 2013;5:1-15.
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;134:25-9.
Massicotte L, Sassine MP, Lenis S, Seal RF, Roy A. Survival rate changes with transfusion of blood products during liver transplantation. Can J Anesth 2005;52:148-55.
Ramos E, Dalmau A, Sabate A, Lama C, Llado L, Figueras J, 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;9:1320-7.
Rana A, Petrowsky H, Hong JC, Agopian VG, Kaldas FM, Farmer D, et al
. Blood transfusion requirement during liver transplantation is an important risk factor for mortality. J Am CollSurg 2013;216:902-7.
Biancofiore G, Bindi L, Cellai F, Consani G, Meacci I, Boggi U, et al
. Orthotopic liver transplant. Analysis of costs related to anesthesiologic and intensive care phases. Minerva Anestesiol 1998;64:587-91.
Amitrano L, Guardascione MA, Brancaccio V, Balzano A. Coagulation disorders in liver disease. Semin Liver Dis 2002;22:83-96.
Allison MG, Shanholtz CB, Sachdeva A. Hematological issues in liver disease. Crit Care Clin 2016;32:385-96.
Massicotte L, Sassine M-P, Lenis S, Roy A. Transfusion predictors in liver transplant. Anesth Analg 2004;98:1245-51.
Reichert B, Kaltenborn A, Becker T, Schiffer M, Klempnauer J, Schrem H. Massive blood transfusion after the first cut in liver transplantation predicts renal outcome and survival. Langenbecks Arch Surg 2014;399:429-40.
Chung HS, Jung DH, Park CS. Intraoperative predictors of short-term mortality in living donor liver transplantation due to acute liver failure. Transplant Proc 2013;45:236-40.
Cywinski JB, You J, Argalious M, Irefin S, Parker BM, Fung JJ, et al
. Transfusion of older red blood cells is associated with decreased graft survival after orthotopic liver transplantation. Liver Transplant 2013;19:1181-8.
Ozkardesler S, Avkan-Oguz V, Akan M, Unek T, Ozbilgin M, Meseri R, et al
. Effects of blood products on nosocomial infections in liver transplant recipients. Exp Clin Transplant Off J Middle East Soc Organ Transplant 2013;11:530-6.
Esmat Gamil M, Pirenne J, Van Malenstein H, Verhaegen M, Desschans B, Monbaliu D, et al
. Risk factors for bleeding and clinical implications in patients undergoing liver transplantation. Transplant Proc 2012;44:2857-60.
Massicotte L, Beaulieu D, Thibeault L, Roy JD, Marleau D, Lapointe R, et al
. Coagulation defects do not predict blood product requirements during liver transplantation. Transplantation 2008;85:956-62.
Roullet S, Biais M, Millas E, Revel P, Quinart A, Sztark F. Risk factors for bleeding and transfusion during orthotopic liver transplantation. Ann Fr Anaesth Reanim 2011;30:349-52.
Pustavoitau A, Lesley M, Ariyo P, Latif A, Villamayor AJ, Frank SM, et al
. Predictive modeling of massive transfusion requirements during liver transplantation and its potential to reduce utilization of blood bank resources. Anesth Analg 2017;124:1644-52.
Mangus RS, Kinsella SB, Nobari MM, Fridell JA, Vianna RM, Ward ES, et al
. Predictors of blood product use in orthotopic liver transplantation using the piggyback hepatectomy technique. Transplant Proc 2007;39:3207-13.
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 Anesth 2001;48:1075-9.
Ozier Y, Pessione F, Samain E, Courtois F, French Study Group on Blood Transfusion in Liver Transplantation. Institutional variability in transfusion practice for liver transplantation. Anesth Analg 2003;97:671-9.
Findlay JY, Rettke SR. Poor prediction of blood transfusion requirements in adult liver transplantations from preoperative variables. J Clin Anaesth 2000;12:319-23.
McCluskey SA, Karkouti K, Wijeysundera DN, Kakizawa K, Ghannam M, Hamdy A, et al
. Derivation of a risk index for the prediction of massive blood transfusion in liver transplantation. Liver Transplant 2006;12:1584-93.
Escoresca Ortega AM, Mogollón Jiménez MV, Hinojosa Pérez R, FerrándizMillón CM, Salgado Algarrada JC, HerruzoAvilés A, et al
. Application of the McCluskey Index to predict blood product requirements during liver transplantation. Transplant Proc 2008;40:2981-2.
Deakin M, Gunson BK, Dunn JA, McMaster P, Tisone G, Warwick J, et al
. Factors influencing blood transfusion during adult liver transplantation. Ann R Coll Surg Engl 1993;75:339-44.
Das J, Khanna S, Kumar S, Mehta Y. Blood transfusion practices in liver transplantation. Indian J Anaesth 2015;59:266-7.
] [Full text]
Lawson PJ, Moore HB, Moore EE, Stettler GR, Pshak TJ, Kam I, et al
. Preoperative thrombelastography maximum amplitude predicts massive transfusion in liver transplantation. J Surg Res 2017;1:171-5.
Massicotte L, Lenis S, Thibeault L, Sassine MP, Seal RF, Roy A. Effect of low central venous pressure and phlebotomy on blood product transfusion requirements during liver transplantations. Liver Transplant 2006;12:117-23.
Blake BS, Aniskevich S, Thomas CS, Ladlie BL. Preoperative clinical characteristics that identify potential low-volume transfusion candidates among orthotopic liver transplant patients. Exp Clin Transplant 2016;14:405-11.
[Table 1], [Table 2], [Table 3], [Table 4]