How to minimize blood loss during liver surgery in patients with cirrhosis

Abstract

Patients with liver disease frequently have substantial changes in their haemostatic system. This is reflected in abnormal test results on routine coagulation screening assays such as the prothrombin time (PT), activated thromboplastin time (APTT) and platelet count. Traditionally, attempts were made to correct abnormalities in the haemostatic system as measured by routine coagulation assays prior to invasive procedures by infusion of platelets or fresh frozen plasma (FFP). Recent laboratory and clinical data have indicated that the haemostatic reserve in cirrhotic patients is relatively well maintained although the coagulation screening assays suggest otherwise. Pre-procedural correction of coagulation tests with blood products may therefore not be necessary, and may even have harmful side-effects. In particular, fluid overload resulting in exacerbation of portal hypertension by infusion of blood products may in fact promote bleeding. In recent years, it has become clear that reduction of the central and portal venous pressure by fluid restriction and avoidance of blood product transfusion is a beneficial strategy in minimizing bleeding during liver surgery in cirrhotic patients. Some investigators have even taken this a step further and suggested pre-procedural phlebotomy in liver transplant recipients. The aim of this review is to provide an overview of recent studies and developments which have changed our understanding of the clinical relevance of abnormal coagulation tests in patients with cirrhosis, and which have contributed to a reduction in blood loss and transfusion requirements when liver surgery is needed in these patients.

Introduction

Extensive blood loss is still a major concern in major surgical procedures involving the liver, especially during partial liver resection and orthotopic liver transplantation (OLT). Although blood loss in patients undergoing liver surgery has decreased substantially during the last decade, (excessive) blood loss and transfusion requirements are still a concern. Peri-operative blood loss and transfusion of blood products are associated with increased rates of morbidity and mortality, both in non-cirrhotic and cirrhotic patients.^1–4 Main causes of blood loss are thought to consist of surgical and patient-related factors and impaired haemostasis. In general, surgical factors may contribute to 75–90% of intra-operative and early post-operative bleeding.⁵ In liver transplant surgery, additional factors related to peri-operative blood loss are severity of liver disease, concomitant renal failure, length of the cold ischaemia time and the type of surgical technique used (non-cava sparing versus cava sparing or piggy back technique).^6,7 Additionally, in patients with a poor liver function, abnormalities in the haemostatic system are thought to contribute significantly to peri-operative bleeding. Nevertheless, a substantial proportion of cirrhotic patients with poor haemostatic function can nowadays undergo major invasive procedures, such as liver transplantation, without the requirement for any blood transfusion.^1,3,8 This may imply that haemostatic alterations are not as clinically relevant in determining peri-operative bleeding in cirrhotic patients as previously assumed. Support for a minor role of dysbalanced coagulation in peri-operative blood loss has emerged from laboratory studies which have indicated that defects in pro-haemostatic pathways in patients with liver disease are compensated for, at least in part, by concomitant deficiencies in anti-haemostatic systems.^9–12 As routine tests of coagulation such as the prothrombin time (PT) are only sensitive for procoagulant systems, prolonged PT values in cirrhotics are often misinterpreted as representing a bleeding tendency.^13,14 Indeed, routine coagulation tests such as the PT are poor predictors of blood loss during liver transplantation.^15–17

Traditionally, attempts were made to reverse abnormal coagulation tests in the cirrhotic patient prior to a surgical procedure by the administration of blood products such as fresh frozen plasma (FFP). However, it is debatable whether correction of abnormal test results is really necessary, as these tests do not necessarily reflect in vivo haemostasis.^11,18 Moreover, transfusion of blood products is associated with substantial side-effects, including fluid overload. In recent years, it has become clear that maintenance of a low central venous pressure (CVP) and even reduction of CVP by phlebotomy is a beneficial strategy in minimizing blood loss during liver resection or liver transplantation.^8,19 Maintenance of a low CVP requires a restrictive use of fluids and avoidance of transfusion of (large volume) blood products, which implicates that routine correction of abnormal coagulation tests may not be essential to minimize intra-operative blood loss during surgical liver surgery.

This paper provides an evaluation of clinical and laboratory data used to guide peri-operative strategies directed at a reduction in blood loss in cirrhotic patients undergoing surgery of the liver. In particular, we will provide arguments that contraction of volume is a better treatment to minimize bleeding during invasive liver procedures than a routine correction of abnormal coagulation tests by transfusion of (large volume) blood products.

Traditional concepts in the treatment of altered haemostasis prior to liver surgery

Patients with end-stage liver disease suffer from complex haemostatic disturbances, as a result of a reduction in platelet count and platelet function, and a reduction in circulating levels of proteins involved in coagulation and fibrinolysis.²⁰ For years it was generally believed that there is a strong causal relation between abnormalities in the haemostatic system and the bleeding tendency in patients with cirrhosis. These haemostatic differences result in abnormalities in the routine screening tests of coagulation, such as the PT, activated partial thromboplastin time (APTT) and the bleeding time.

Traditionally, attempts were made to reverse the abnormalities of routine haemostatic tests before initiation of liver surgery. In particular, a prolonged PT was treated with infusion of FFP, and a low platelet count was treated with platelet concentrates. This policy, however, may not be effective and has a number of disadvantages. First of all, it has been shown in several studies that abnormal pre-operative coagulation tests are a very poor predictor of intra-operative bleeding.^14–17 Second, correction of a prolonged PT with FFP does not result in complete normalization, and the duration of the reversal is relatively short.²¹ Third, correction of a prolonged pre-operative PT by administration of recombinant factor (F)VIIa, does not result in a reduction of intra-operative bleeding or transfusion requirements in cirrhotic patients undergoing partial liver resection or transplantation.^22–24 Although recombinant FVIIa is a very effective drug in normalizing a prolonged PT in patients with cirrhosis,²⁵ several randomized controlled studies have shown that this does not result in a subsequent reduction of intra-operative bleeding or transfusion requirements, as compared with placebo.^22–24 The only category of pharmacological agents that has shown to be effective in reducing blood loss during liver surgery in patients with cirrhosis is that of antifibrinolytic drugs.²⁶ Hyperfibrinolysis, however, is not reflected by routine coagulation tests and cannot be corrected by infusion of blood products.

Transfusion of large amounts of FFP may in fact be counterproductive as it leads to fluid overload and a subsequent increase of the central and portal venous pressure, which is already elevated in many cirrhotic patients. The increase in portal pressure may result in increased bleeding during abdominal surgical dissection. Moreover, transfusion of blood products is associated with important side-effects. It is well known that infusion of platelets or FFP in general is associated with an increased risk of infectious complications and respiratory complications, such as pulmonary oedema and transfusion-related acute lung injury (TRALI).^27,28 These side-effects may account for the negative relation between intra-operative blood transfusion requirements and post-operative infection rate, morbidity and mortality.^2–4,29 Also in liver transplant patients, platelet transfusions have been associated with increased post-operative mortality, as a result of an increased risk of acute lung injury.³⁰

Altogether, these lines of evidence indicate that correction of a prolonged PT in patients with cirrhosis does not necessarily translate to improved haemostasis or reduced intra-operative blood loss when these patients require liver surgery.

New insights in haemostatic function in cirrhotic patients: a rebalance of the haemostatic system?

Recent data from both laboratory models and clinical studies have suggested that the haemostatic dysbalance in cirrhotic patients may not be as severe as suggested by the routine coagulation tests. Although commonly used tests such as the PT, APTT and the bleeding time are frequently prolonged in patients with liver failure, these tests may not reflect in vivo haemostasis in these patients. Recent studies suggest that the haemostatic abnormalities in patients with liver disease are compensated for, or ‘rebalanced’, as a result of concomitant alterations in both pro- and anti-haemostatic systems (reviewed in ²⁰). The defects in haemostatic pathways that promote bleeding are balanced (in part) by changes in the haemostatic system that actually promotes clot formation (see Table 1). For example, deficiencies in procoagulant proteins may be balanced by deficiencies in anticoagulant proteins. Tripodi and coworkers have shown that in patients with liver disease and a prolonged PT, thrombin generation is indistinguishable from that of healthy volunteers, provided that the thrombin generation test was executed in the presence of an activator of the anticoagulant protein C system.¹¹ This observation exemplified that a prolonged PT does not necessarily translate into defective thrombin generation, as PT is only sensitive for levels of procoagulant proteins. In the more sophisticated test as used by Tripodi et al., the balance between pro- and anticoagulant proteins is tested more accurately.

Table 1.

Alterations to the haemostatic system in patients with liver disease that contribute to bleeding (left) or counteract bleeding (right)

Changes that impair haemostasis	Changes that promote haemostasis
Thrombocytopenia	Elevated levels of von Willebrand factor
Platelet function defects	Decreased levels of ADAMTS-13
Enhanced production of nitric oxide and prostacyclin	Elevated levels of factor VIII
Low levels of factors II, V, VII, IX, X, and XI	Decreased levels of protein C, protein S, antithrombin, α2-macroglobulin, and heparin cofactor II
Vitamin K deficiency	Low levels of plasminogen
Dysfibrinogenemia
Low levels of α2-antiplasmin, factor XIII, and TAFI
Elevated t-PA levels

Source:Modified from the European Association for the Study of the Liver from Lisman et al.²⁰ with permission.

Similarly, it was shown that the fibrinolytic system may also be rebalanced in patients with cirrhosis⁹ although these findings have been challenged.³¹ Finally, the low platelet count in patients with liver disease may be compensated for by highly elevated levels of von Willebrand factor (VWF).¹⁰

Altogether, laboratory evidence suggests that the patient's cirrhosis has a compensated or ‘rebalanced’ haemostatic system. It has been suggested that this system may function adequately (no bleeding or thrombosis), albeit with more narrow margins, compared with healthy individuals.³² The more narrow margins may explain why clinically overt bleeding complications may occur in these patients when the haemostatic balance is heavily challenged i.e. during episodes of sepsis. Conversely, thrombotic complications such as portal vein thrombosis, deep venous thrombosis and pulmonary embolism may occur in patients with liver disease,^33–37 despite the commonly accepted notion that these patients are ‘auto-anticoagulated’, as suggested by their prolonged coagulation tests.

All these changing insights of the performance of the haemostatic system in patients with cirrhosis have consequences for the interpretation of pre-operative coagulation assays and for the pre- and intra-operative management of cirrhotic patients undergoing surgical procedures. This is supported by clinical data showing a lack of excessive blood loss in many patients with end-stage liver disease during liver transplantation, when abnormal coagulation tests are not routinely corrected with the infusion of blood products, and the subtle haemostatic balance is left untouched as much as possible by a restrictive infusion policy.^1,8 A substantial proportion of patients can nowadays be transplanted without the requirement for any blood transfusion, and we believe that major surgery without the requirement for transfusion would never be possible in a patient with a true coagulopathy. Bleeding episodes that do occur in patients with liver disease are often unrelated to deranged coagulation, and appear to be a consequence of portal hypertension and venous collaterals rather than coagulopathy (i.e. variceal bleeding).

Volume contraction: an attractive strategy to reduce blood loss during surgery in patients with cirrhosis?

In view of the rebalanced haemostatic system, correction of haemostatic abnormalities prior to surgical interventions by, for example, plasma transfusion is not necessary and may even be harmful. As discussed above, overfilling patients with blood products is associated with an increased risk of morbidity and mortality.

Bleeding complications in patients with liver disease may not primarily be related to impaired coagulation. Alterations in haemodynamics and vessel wall function may play a more important role in the bleeding tendency of cirrhotic patients. In particular, the hyperdynamic circulation and the presence of portal hypertension may be important factors in peri-operative bleeding in these patients. Also in the non-surgical patient, the most common bleeding complication, i.e. variceal bleeding, is as a result of vascular abnormalities and increased portal venous pressure.³⁸

Reduction of the systemic and portal venous pressure may be therefore an attractive strategy for reduction of blood loss during liver surgery. Indeed, the application of low CVP during liver surgery has widely been studied.^8,39–42 Low CVP is usually obtained by fluid restriction, forced diuresis, venous vasodilatation or even phlebotomy. Maintenance of a low CVP during partial resection of non-cirrhotic livers is associated with a significant reduction in blood loss.^40–42 However, a low CVP may have serious side-effects including renal failure, but reports on this have been contradictory. Schroeder et al. compared two centres performing liver transplantations.⁴³ One liver transplantation centre adhered to a low CVP policy (<5 mmHg) using fluid restriction, whereas the second centre used a normal CVP (7 to 10 mmHg). Both patient groups were similar in demographics, aetiology of liver disease and surgical technique. Patients in the low CVP group received significantly lower amounts of red blood cells, FFP and platelets compared with patients in the normal CVP group. However, peak serum creatinine (3.2 vs. 1.8 mg/dl, P < 0.01) and the need for dialysis (6.8% vs. 1.2%, P < 0.05) were significantly higher in patients with a low CVP during transplantation. In contrast, Wang et al. found no significant differences in post-operative renal function when comparing two groups of 25 cirrhotic patients with a low CVP (<4 mmHg) or normal CVP while undergoing partial liver resection for hepatocellular carcinoma.⁴⁴

Some investigators have taken the concept of avoiding fluid overload and reducing central and portal venous pressure as maximal as possible to a more extreme strategy and have even proposed pre-operative phlebotomy in patients undergoing liver transplantation.^8,17 Using a combined strategy of fluid restriction and pre-operative phlebotomy in patients undergoing liver transplantation, Massicotte et al. have reported significant reduction in intra-operative blood loss and transfusion requirements, when compared with historical control patients with a more conventional fluid management.^8,17 The median red blood cell transfusion requirement was 0.3 units and 81% of the patients did not require any blood transfusion when applying this strategy. Although such a strategy could be potentially associated with an increased rate of other peri-operative complications, such as renal failure or cardiac ischaemic events, these investigators did not observe an increased need for post-operative renal replacement therapy in a consecutive series of 200 patients.¹⁷

In Groningen, we have also adopted a fluid restriction policy in patients undergoing partial liver resection or liver transplantation, although we do not use extreme measures such as pre-operative phlebotomy. Our policy is not to correct routinely abnormal pre-operative coagulation tests with platelet or plasma transfusion. Furthermore, we accept a haematocrit of 25–30%, maintain a low CVP (<5 mmHg) and are very restrictive in transfusing FFP and platelets peri-operatively.^1,29 When indicated, we would rather use vasopressors and diuretics to maintain adequate perfusion pressure and diuresis than the additional infusion of fluids. Adoption of a fluid restriction policy in our centre has been accompanied by a steady decline in transfusion requirements, and a steady increase in patients who are transplanted without the requirement for any red blood cell (RBC) transfusion. During the past 5 years, around 40% of the adult patients undergoing a first liver transplant did not require a RBC transfusion.¹ Although we have not formally studied whether the reduced transfusion requirements are as a consequence of adoption of a very restrictive transfusion policy, this has generally been the perception within the group of surgeons and anaesthetists. Several other centres have observed a steady decline in FFP and RBC usage over the years, despite the fact that the mean model for end-stage liver disease (MELD) score of patients transplanted steadily increases over time.⁴⁵

Conclusion

Recent data from laboratory models and clinic studies have demonstrated that the coagulation system in cirrhotic patients is not as abnormal as traditionally assumed, but is ‘rebalanced’ by concomitant deficiencies in pro- and anti-coagulant systems. These changing insights in the performance of the haemostatic system in patients with cirrhosis have consequences for the interpretation of pre-operative coagulation tests. Pre-operative coagulation assays do not appear to predict bleeding in cirrhotic patients and pre-procedural correction of abnormal coagulation assays with blood or plasma products may therefore not be as effective as traditionally assumed. In fact, efficacy of pre-procedural correction of haemostasis by administration of blood products in patients with liver disease has never been demonstrated in appropriately designed clinical studies. Conversely, transfusion of blood and plasma products has been associated with important side-effects and may even increase the bleeding tendency by inducing volume overload. Alterations in haemodynamics (i.e. portal venous hypertension) and vessel wall function may play a more important role in the bleeding tendency of patients with end-stage liver disease than coagulation disturbances. Indeed, strategies aiming at a reduction of central and portal venous blood pressure by fluid restriction and avoidance of routine transfusion of blood products have been associated with a reduction in peri-operative blood loss in cirrhotic patients undergoing surgery of the liver.

Volume contraction and a very restrictive blood product transfusion protocol may be more effective in reducing bleeding during major surgical procedures involving the liver than routine correction of abnormal coagulation test (i.e. with large volumes of FFP). Further clinical research should assess possible side-effects of maintaining a low central pressure in cirrhotic patients. Moreover, randomized controlled trials comparing different transfusion and infusion protocols are urgently needed.

Conflict of interest

None declared.