key: cord-0037064-vzzrsqsn
authors: Naidu, C. Sudeep; Sarin, Arti
title: Postoperative Liver Failure
date: 2017-02-08
journal: GI Surgery Annual
DOI: 10.1007/978-981-10-2678-2_3
sha: a55f11db14848d928e1a27944c9e7ba6a8e2de2f
doc_id: 37064
cord_uid: vzzrsqsn

Technical innovations in surgical techniques, anaesthesia, critical care and a spatial understanding of the intra-hepatic anatomy of the liver, have led to an increasing number of liver resections being performed all over the world. However, the number of complications directly attributed to the procedure and leading to inadequate or poor hepatic functional status in the postoperative period remains a matter of concern. There has always been a problem of arriving at a consensus in the definition of the term: postoperative liver failure (PLF). The burgeoning rate of living donor liver transplants, with lives of perfectly healthy donors involved, has mandated a consensual definition, uniform diagnosis and protocol for management of PLF. The absence of a uniform definition has led to poor comparison among various trials. PLF remains a dreaded complication in resection of the liver, with a reported incidence of up to 8 % [1], and mortality rates of up to 30–70 % have been quoted [2]. Several studies have quoted a lower incidence of PLF in eastern countries, but when it occurs the mortality is as high as in the West [3].

In 2005, Balzan et al. [ 5 ] published their results of 775 elective liver resections over a span of 5 years. They focussed on serum bilirubin (SB) and prothrombin time (PT) as important prognostic markers of postoperative liver functional status and proposed the '50-50' criteria for the defi nition of PLF, i.e. the combination of PT >50 % of baseline normal and SB >50 μmol/L on postoperative day (POD) 5 (the '50-50' criteria) was found to be strongly predictive of mortality. The fact that this 'criteria' was a precursor of clinical complications 3-8 days before they appeared lent it a strong base for life-saving interventions. The criticism of this simple calculation has been that it relies only on two laboratory tests and does not factor in the existing clinical status of the patient. The '50-50' rule is limited by the fact that it cannot be applied before POD 5, does not stratify patients and though it predicts death in up to 70 % of patients, it is not based on the clinical severity of the patient.

In 2010, Rahbari et al. [ 2 ] , of the International Study Group of Liver Surgery (ISGLS) defi ned PLF as 'a postoperative acquired deterioration' in the functions of the liver: synthetic, excretory and detoxifying. This deterioration is evident by an increase in international normalized ratio (INR) (There may be a need of clotting factors for abnormal INR) and SB, on or after POD 5 (compared with the values of the preceding day). Biliary obstruction needs to be ruled out as a cause for the deranged SB. They also stratifi ed PLF into three grades: A, B and C, depending on up-scaling of the required level of care.

Traditionally, surgeons have resorted to keeping an eye on deteriorating liver function by charting rising INR (coagulopathy), rising SB (hyperbilirubinaemia) and the advent of hepatic encephalopathy (failure of detoxifi cation) as surrogate markers for the functions of the liver. Other scores such as the Child-Pugh score (CTP) and the Model for end stage liver disease (MELD) have also been used by various authors for defi ning PLF, but a uniform consensus has evaded clinicians due to the multifactorial and diverse aetiology and pathogenesis of PLF.

Following liver resection the patient has multiple pathophysiological mechanisms at work. There is the trauma of surgery, the anaesthetic and haemodynamic changes, the metabolic demands of wound healing and especially in case of the liver: the pathophysiology of ischaemia-reperfusion injury (IRI), liver regeneration and the small for size syndrome (SFSS). Not only do the number of hepatocytes have to be adequate for body homeostasis, they should be functioning optimally and retain their capacity for regeneration. The liver cells suffer mainly from a combination of IRI, hepatic venous congestion and sepsis.

The 'hyperfusion theory' is widely accepted and postulates that the relative spike in sinusoidal perfusion of the decreased cell mass precipitates a vicious cycle. A cascade mechanism of which is a combination of congestion, infl ammation, cholestasis and cell death taking place, preventing the normal function of a hepatocyte: uptake, secretion and excretion [ 6 ] . In addition cell proliferation and regeneration are impeded. A standard liver resection for a liver tumour has to deal with IRI, congestion, portal hypertension and sepsis, while, in transplantation denervation and immunosuppression are added as precipitating factors.

IRI persists even after parenchymal damage during liver resection. After a period of ischaemia the infl ammatory response in the form of the complement cascade is activated. Activated Kupffer cells generate reactive oxygen species (ROS) which cause endothelial cell damage [ 7 ] . Later in the reperfusion phase, these metabolites are swept around leading to a cycle of microvascular injury and microcirculatory changes resulting in apoptosis and cell necrosis with resultant hepatocyte death.

Sepsis may intervene in as high as 50 % of patients after liver resections and may be related to loss of Kupffer cell volume with impaired immune function. The relationship between infection and PLF has not been fully explained [ 8 ] . Patients with ALF are particularly prone to developing sepsis. It has been shown that 73 % of patients with PLF develop postoperative sepsis compared with 18 % of patients without [ 9 ] . Sepsis has a triad of detrimental effects on liver synthetic function, hepatocyte regeneration and apoptosis by inducing a relative hepatic ischaemia due to systemic hypotension. It induces Kupffer cell dysfunction, releases proinfl ammatory cytokines, and diminishes detoxifi cation of liver endotoxins thereby leading to diminished hepatocyte proliferation and regeneration [ 10 ] . Liver surgery by itself may increase the incidence of infection [ 8 ] as major resections are associated with enteric bacterial translocation, which is enhanced by the prolonged infl ow clamping and duration of surgery. A major liver resection involving multiple segments signifi cantly impedes the function of the reticuloendothelial system, which is crucial in immune defence against sepsis.

The precipitating factors for development of PLF can be broadly divided into patient factors, surgery-related factors and postoperative complications and their management (Table 3.1 ) .

Patients with diabetes have a signifi cantly poorer outcome after elective liver resections compared to those who do not have diabetes [ 11 ] . Apart from being at higher risk for infections due to decreased immune tolerance, patients with diabetes may also have a higher incidence of fatty liver with insulin resistance and concomitant poorer functional reserve. Preoperative diabetes mellitus is also an independent predictor of 90-day mortality [ 12 ] .

Patients with cholangitis and active viral hepatitis also do poorly after surgery. In a study, mortality was signifi cantly higher in patients who had resection of hepatocellular carcinoma (HCC) in cirrhosis associated with active hepatitis (8.7 versus 1.5 %; p < 0.05) [ 13 ] . However, increased risk of PLF with raised SB has been controversial [ 14 ] . Cherqui et al. showed that patients with a raised SB level had a morbidity rate of 50 % compared to 15 % in patients with normal SB (p < 0.01). However, the incidence of PLF or mortality did not rise when compared with matched controls [ 14 ] .

Most patients being planned for liver resections have colorectal liver metastasis and receive chemotherapy with 5-fl uorouracil, oxaliplatin and irinotecan, and newer monoclonal antibodies cituximab and bevacizumab [ 7 ] . Chemotherapy induces marked histopathological changes in the liver parenchyma including fatty liver, chemotherapy-associated steatohepatitis (CASH), or sinusoidal injury (sinusoidal obstruction syndrome, SOS). CASH is pathognomonic of patients treated with irinotecan and is characterized by 'steatosis, lobular infl ammation and ballooning of hepatocytes': also called the 'grey liver syndrome'. SOS is caused by the use of oxaliplatin and the syndrome is called 'blue liver syndrome' because of the characteristic bluish-red hue of the fi rm liver [ 15 , 16 ] . Irinotecan-induced CASH has been shown to be an independent risk factor for postoperative mortality and PLF. Oxaliplatin induced SOS develops after more than 6 cycles and is a risk factor for increased hospital stay and postoperative complications [ 17 , 18 ] .

Hepatic steatosis is a major determinant of postoperative outcomes. Over 20 % of patients undergoing a major liver resection have some degree of steatosis, significant enough to alter postoperative recovery [ 19 ] . Patients with biopsy-proven hepatic steatosis have a higher incidence of PLF (14 %) than patients with healthy livers (4 %) [ 20 ] even though the steatosis was of moderate grade. Belghiti et al. [ 21 ] , reported a series of 478 patients who underwent liver resections, of which 37 patients had steatosis. Steatosis was an independent risk factor for postoperative complications (8 % of patients with steatosis), while only 2 % of patients with a normal histology developed complications. The increasing incidence of CASH, NASH and SOS has sparked interest in the use of preoperative liver biopsy as an assessment of the liver function and structure. Screening of high-risk patients (obese, high body mass index or those on chemotherapy) has been proposed. If a moderate degree of steatosis is established, these patients could benefi t from manipulation of the volume, or surgery could be deferred till the acute changes subside either by treating the underlying aetiology or withdrawing the inciting agent.

Hepatitis B and hepatitis C virus infection-associated HCC develops along with the presence of fi brosis in the liver parenchyma. There is no correlation with fi brosis and liver resection, but when cirrhosis is established by a biopsy, it remains an independent predictor of poor outcomes in terms of overall and tumour-free recurrence. Shen et al. used markers of liver fi brosis preoperatively to predict PLF in patients with HCC undergoing liver resection [ 22 ] . They evaluated preoperative hepatitis B virus (HBV) DNA levels, serum prealbumin (PA), hyaluronic acid (HA) and laminin (LN) levels and correlated these with PLF. A prospective model was used with these four laboratory results and validated in 89 HCC patients with a sensitivity and specifi city of 62 % and 91 %, respectively.

Malnutrition is commonly prevalent among cirrhotics and leads to increased morbidity and mortality [ 23 ] . Malnutrition causes 'disordered mitochondrial function' which alters the immune response and thus reduces the hepatocyte regenerative capacity when exposed to ischaemia [ 24 ] . It is thus essential to objectively evaluate the nutritional status of all patients with liver disease, and to intervene with supplements as indicated.

PLF and postoperative renal dysfunction are independent predictors of 90-day mortality following liver resection but the predictive value for mortality is significantly higher when both systems fail simultaneously. Renal dysfunction following liver surgery may occur because of liver failure and hepatorenal syndrome but also due to hypovolaemia and release of free radicals and pro-infl ammatory mediators during surgery [ 25 ] .

Advanced age is no longer a deterrent to hepatic resections, which is now increasingly performed in elderly people with an acceptable morbidity and mortality [ 26 ] . Advancing age reduces the capacity of the liver to regenerate. The American Society of Anaesthesiology (ASA) and Acute Physiology and Chronic Health Evaluation (APACHE) scores have proved useful in anticipating complications following major liver surgery [ 27 ] .

Assessment of the true functional status of the liver is fraught with complexities and routine blood tests have proved unreliable predictors of PLF. Notwithstanding this, all patients planned for major liver resections should have complete liver biochemistry, a complete blood count and a prothrombin time as baseline investigations [ 10 ] .

Experience in liver surgery/high volume centres show an inverse relationship with outcomes. It has been suggested that patients needing liver resections be referred to centres that perform 10-17 liver resections per year [ 28 ] . Massive estimated blood loss (EBL) remains a key prognostic factor for a safe resection. EBL correlates with the extent of resection and the number of segments resected [ 29 ] and this correlates with the incidence of PLF. Massive EBL during a major liver resection should be anticipated in tumours abutting the inferior vena cava or major hepatic veins, or if there is injury to the middle hepatic vein during resection, and not by patient age, tumour size alone, or type of hepatectomy. Cirrhosis creates a hyperdynamic milieu with increased cardiac output and decreased systemic vascular resistance. The hepatic buffer response of the cirrhotic liver is altered: portal blood fl ow is reduced/shunted as a result of collaterals in portal hypertension, and arterial blood fl ow is impeded because of hepatic fi brosis and sinusoidal resistance. This altered fl ow renders the cirrhotic liver relatively anoxic and less tolerant to hypotension and hypoxia.

An intraoperative blood loss greater than 1000 ml increases the risk of PLF [ 30 ] . This effect may be caused by a massive fl uid shift secondary to EBL causing global infl ammation because of bacterial endotoxins, peripheral vasodilation and pooling in third spaces. Coagulopathy following blood loss with the inability of the liver to catch up, seems to increase the potential for intra-abdominal collections and bacterial infections. Avoiding prolonged hypotension and hypothermia by using judicious and timely transfusion, rapid infusion devices and safe surgical techniques is the key to salvage these patients. In fact, even major blood loss may be tolerated, if adequate efforts have been made to maintain euthermia, perfusion, avoid metabolic acidosis and provide an adequate and timely buffer against the dangerous triad of acidosis, hypothermia and coagulopathy resulting from EBL [ 30 ] . Prolonged operating time also leads to a poorer outcome and is extended in patients with EBL, vascular reconstructions and diffi cult surgery due to adhesions and tumour extensions into unsafe areas.

The functioning liver remnant (FLR) in patients has become a topic of much debate especially with the popularity of living donor liver transplantation. In a liver with normal parenchyma <25 % FLR is associated with a poor outcome, compared to that of patients with ≥25 % FLR [ 31 ] . The risk of PLF is 3 times greater. The FLR and the method of calculating it are even more important in livers with steatosis and fi brosis as in these livers functional reserve is markedly reduced. Patients with histologically proven parenchymal changes of steatosis, fi brosis or cirrhosis, mandate a FLR of up to 40 % [ 32 ] .

Postoperative management has an important bearing on outcome. The fi rst 48 h after a major hepatic surgery are crucial for a successful outcome. Metabolic, functional and haemodynamic alterations after hepatic resection are unique to each patient and demand specifi c management protocols. A multidisciplinary team approach is required with goal-directed therapeutic options. It is mandatory to have invasive haemodynamic monitoring, mechanical ventilation, critical parameter monitoring, strict antisepsis measures, metabolic control and optimal nutritional support.

Some degree of coagulopathy is the expected norm after a major hepatic resection and can be assessed by markers such as PT/INR, platelet counts and partial thromboplastin time (PTT). Postoperative coagulopathy peaks 48-96 h postoperatively and can be best monitored by a thromboelastograph (TEG) [ 33 ] . The underlying coagulopathy leads to postoperative haemorrhage and blood collecting in the abdomen, leading to postoperative infection.

Postoperative antibiotic prophylaxis does not control infections [ 34 ] and no effort should be spared in obtaining meticulous haemostasis and following strict infection control protocols.

Preoperative risk assessment involves a thorough evaluation of all the factors mentioned above. A physical examination followed by appropriate clinical tests will identify patients at risk of developing PLF. The patient's liver status, hepatic reserve potential and functional aspect need to be investigated along with the metabolic and haematological derangements, which may lead to PLF.

Pre-existing liver disease can be determined by a thorough clinical history, recording previous blood transfusions, hidden illicit drug use, excessive ethanol use, history of jaundice, family history of familial cholestasis as well as history of adverse drug reactions. In India, it is important to take a history of use of complementary and alternative medicines (CAM) as many of them maybe hepatotoxic [ 35 ] . Physical examination should be able to pick up subtle signs of incipient liver failure and decompensation such as proximal muscle wasting, spider naevi, ascites and other signs.

Though routine liver tests have a low yield, they are of value in liver resections as the major indications for hepatectomy in eastern countries are HCC and in the West metastatic disease arising in either normal or cirrhotic livers. HCCs are usually sequelae of HBV/HCV disease and it is prudent to evaluate for active disease as well as cirrhosis. Though colorectal metastasis can occur in a liver with normal parenchyma, the widespread use of preoperative chemotherapy, as mentioned earlier, mandates a more thorough evaluation of steatohepatitis and fi brosis. In patients with HCC, cirrhosis is present in 64-74 %, but conversely in patients with cirrhosis there is only an 11 % incidence of HCC [ 36 ] . The rest would be divided equally between alcohol-related disease and HBV/HCV, with about 5 % due to metabolic factors. In a cirrhotic patient with HCC, it is the underlying liver parenchymal disease which precludes a safe liver resection and this needs to be addressed. A cirrhotic liver is much less rarely involved with metastasis in comparison to normal livers.

Liver tests may be labelled as 1. Screening tests to indicate the presence of liver disease 2. Diagnostic tests to discern aetiology 3. Quantitative tests to measure functional reserve.

1. Serum bilirubin evaluates conjugation and excretion functions. Total bilirubin is a poorly sensitive and specifi c test for liver disease. Direct bilirubin does not differentiate between extra-and intrahepatic cholestasis. However, it is an important factor of scores such as CTP and MELD for prognostication. Miyagawa et al. [ 37 ] found no signifi cant differences in morbidity and mortality after major hepatectomies in spite of a raised SB. Postoperatively, SB is often increased; however this does not always indicate impending PLF. 2. Serum bile acids evaluate excretion and shunting. Elevated bile acids are a good marker for portosystemic shunting with a sensitivity of over 90 % in detecting cirrhosis in patients with normal transaminases [ 38 ] . 3. Alkaline phosphatise (ALP) evaluates cholestasis. It is also synthesized in the bone and placenta and is elevated in metastasis and thus is not a specifi c hepatic marker. ALP also has a lag curve in acute obstruction. High preoperative alkaline phosphatase level is indicative of metastatic disease and may be associated with increased mortality after major resections [ 39 ] . Concomitant with a raised carcinoembryonic antigen (CEA) level, it should raise a strong suspicion of liver metastasis. 4. Gamma glutamyl transpeptidase (GGT) evaluates cholestasis, enzyme induction, alcohol abuse, renal failure, myocardial infarction, diabetes and pancreatic diseases, and intake of enzyme inducing agents. Elevation of both GGT and ALP indicates a hepatic origin for ALP elevation. 5. Transaminases evaluate necrosis and parenchymal damage. Raised serum levels do not correlate with the extent of parenchymal necrosis and have no prognostic value. Both enzymes are evaluated together as ALT is more liver specifi c and AST more sensitive to changes. After full course chemotherapy, transaminases can be elevated up to 2.5 times of normal. Partial hepatectomy induces only a mild-to-moderate increase in serum enzymes [ 40 ] . 6. Coagulation factors and PT evaluate synthetic functions. PT is a routine investigation and is used in prognostic scores. As its half-life is shorter, it is a better index of the synthetic function than serum albumin, which has a half-life of 20 days. Fibrinogen levels in mild liver disease are normal/slightly elevated but markedly decreased in massive hepatocellular damage. If the prothrombin time is prolonged, a detailed evaluation of the coagulation system is warranted. Massive hepatectomy invariable leads to a fall in platelet counts and a depression of coagulation factors such as factors I, II, V, VII, X and plasminogen [ 40 ] with resultant disseminated intravascular coagulation (DIC). 7. Albumin evaluates synthesis. It is not only a part of the CTP score but is associated with many non-hepatic diseases such as renal disease, and nutritional entities such as protein malnutrition, protein-losing enteropathy and burns. Patients with preoperative albumin <3.0 g/dl are at risk of increased operative morbidity [ 41 ] .

The serum concentrations of the above tests may not be truly refl ective of liver function alone as many extrahepatic causes may also alter the results, e.g. transfusion associated haemolysis, resorption of haematomas, extrahepatic loss of albumin in bowel-related diseases, or altered PT due to a lack of absorption of vitamin K in either resected small bowel or due to absence of bile in the gut, i.e. obstructive jaundice. Moreover the production, excretion and absorption of these factors are varied and laboratory techniques also result in variances. 

P450-mediated N-demethylation of a 14 C-or 13 C-labelled methyl group of aminopyrine is measured. The formed 14 CO 2 or 13 CO 2 are trapped and measured. Merkel et al. [ 42 ] studied ABT and CTP scores and reported that both reliably predict death from liver failure in patients with cirrhosis. ABT had 94 % sensitivity and 88 % specifi city and this was independent of the Child's classifi cation. Though the test is easy to do, it has not become popular due to the need for expensive equipment. 2. Organic anionic dyes assess hepatic perfusion and excretory function. Normal liver cells take up sulphobromophthalein (BSP), conjugate it with glutathione and excrete into the bile. BSP clearance differentiates cirrhotic from noncirrhotic livers and provides the status of hepatic uptake and biliary excretion. However, BSP is metabolized in the liver and has been reported to cause anaphylaxis, which has restricted its use. In contrast ICG is not metabolized in the liver.

The 15-min retention rate for indocyanine green (ICG15) is the most common preoperative test for evaluating hepatic reserve [ 43 ] . When a hepatectomy is done in a patient with a high ICG15 retention, the volume of non-tumorous liver resected must be minimized. Hepatic function is estimated by ICG15 or of its maximal removal rate (ICG-Rmax). The ICG15 should be approximately 10 % in normal persons. The threshold value for a safe major hepatectomy is set at 14 %, although the cut-off of ICG clearance has shown signifi cant reduction in cirrhotic patients who underwent resection and died subsequently. This was most accurate on day 3 following surgery. When ICGR15 exceeds 20 %, a major hepatectomy should be deferred [ 43 ] . Patients with ICGR15 between 14 and 20 % benefi t from volume manipulation to achieve a viable FLR.

Preoperative ICG clearance may predict 30-day hospital mortality in patients with cirrhosis [ 44 ] . With an accuracy of 100 % for prediction of long-term prognosis in both retrospectively and prospectively evaluated cases, Noguchi et al. [ 45 ] reported a ratio of ICG-Rmax relative to the FLR after hepatectomy, which could reliably predict outcome. Mainly researched by Japanese surgeons, ICG clearance has not been popular with other centres. An ICGR15 value of 14 % has been proposed as a cut-off for identifying patients who will have high postoperative morbidity following a major hepatic resection [ 45 ] .

Recently, ICG has been investigated again. Fung et al. [ 46 ] studied liver stiffness (fi brosis) using a fi brosis measuring impedance elastograph. Although ICG extraction is unique to the liver with minimal extrahepatic elimination, the clearance rate is dependent on local and systemic haemodynamics. Any change in hepatic fl ow or systemic perfusion causes variances in ICG rates. Therefore, they correlated liver stiffness with ICGR15 and liver biochemistry, to determine its reliability in predicting postoperative outcomes. Liver stiffness correlated well with ICGR15 in liver resection patients, and predicted early postoperative complications and was recommended, to provide 'better prognostic information for patients undergoing resection.' 3. MEGX test: evaluates microsomal function and is a measure of the formation of the metabolite monoethylglycinexylidide (MEGX) after injection of a bolus of lidocaine and is evidence of the conversion rate of lidocaine by hepatic cytochrome P450. A value ≤25 μg/L is related to PLF in patients with cirrhosis [ 47 ] . 4. Technetium-99m galactosyl human serum albumin (99m Tc-GSA): GSA scintigraphy studies [ 48 ] have been reported to be useful for predicting the functional reserve of the liver and superior to ICG. 99m Tc-GSA is a scintigraphy agent that binds specifi c hepatic receptors, and can be used to assess the functional hepatocyte mass and thus the hepatic functional reserve in various physiological and pathological states. Unlike ICG it is not affected by the haemodynamic status.

Various scoring systems are in vogue to assess the suitability and risk stratifi cation of hepatic resections in patients with cirrhosis. The CTP and MELD score were initially designed for other prognostications, and their validity in predicting PLF has been the objective of many trials. The results are inconsistent [ 1 , 2 , 5 , 49 ] . In general, it is well accepted that a CTP class C patient is not suitable for any liver resection and those in class B are suitable for only minor liver resections [ 49 ] . Schroeder et al. [ 50 ] reported the superiority of CTP over the MELD score in predicting 30-day morbidity and mortality after hepatic resections. However, other studies validate the MELD score as a reliable predictor. A MELD score above 11 in patients with cirrhosis could predict PLF accurately [ 51 ] .

3D CT reconstructions or magnetic resonance imaging (MRI) reconstructions are now used exclusively for volumetric analysis and predicting FLR [ 52 ] . Calculation of FLR however remains a cause of disagreement and various techniques and calculations are in vogue. 3D reconstructions allow delineation of the hepatic veins, congestion volumes, exact tumour localization and facilitates virtual resection planning. However, imaging at present over estimates the FLR, and different formulae are in use, in an attempt to account for this error. The crux of any imaging or formula used is to ensure that the FLR is compatible with a smooth recovery and it is vital to assess the functional status of the FLR. Addition of preoperative hepatobiliary scintigraphy and CT volumetric measurement were performed by Dinant et al. in preoperative patients [ 48 ] to assess the accuracy of risk assessment for postoperative morbidity, liver failure and mortality. They concluded that using hepatobiliary scintigraphy with preoperative measurement of 99mTc-mebrofenin uptake in the FLR, proved more valuable than measurement of the FLR on CT alone in assessing the risk of PLF.

Keeping the pathophysiology in mind, PLF can be prevented by a two-pronged strategy: Protect the parenchyma against damage and increase the parenchymal volume.

1. Ischaemic preconditioning: After a brief period of infl ow clamping, reperfusion is allowed, prior to the prolonged infl ow clamping ischaemic insult (10 min of ischaemia and 10 min of reperfusion). It increases tolerance to the subsequent prolonged ischaemia and adenosine 5-triphosphate (ATP) depletion by exposing the parenchyma to brief intervals of anoxia and reperfusion before the fi nal resection. It acts by presumably controlling IRI and retards the complement cascade. This reduces reperfusion injury particularly in steatotic patients. Clavien et al. [ 53 , 54 ] did the initial trials and demonstrated a two-fold reduction in the postoperative serum transaminase levels. A reduced mass of apoptotic cells was noted on histopathology. A randomized trial by Chouker et al. [ 55 ] comparing ischaemic preconditioning and continuous clamping, showed stable cardiovascular haemodynamics, lowering the need for adrenaline/noradrenaline after liver reperfusion. Additionally, a recent Cochrane analysis observed no statistically signifi cant difference in the mortality, liver failure, blood loss or haemodynamic changes [ 56 ] . However, length of hospital stay was signifi cantly lower in the ischaemic preconditioning group. 2. Intermittent vascular clamping: consists of repeated periods of 15 min of infl ow clamping followed by 5-min reperfusion phases. Belghiti et al. [ 57 ] reported that in contrast to the presumption, blood loss was signifi cantly more in the intermittent clamping group. Acute phase liver enzymes and transaminase levels were signifi cantly higher in the continuous portal triad clamping group than in the intermittent portal infl ow clamping group when livers with chronic liver disease were included. Postoperative SB levels in cirrhotics were also signifi cantly higher in the continuous infl ow clamping group compared to the intermittent portal infl ow cohort. They concluded that livers with chronic disease do not tolerate continuous vascular clamping as well as normal livers. 3. Avoiding infl ow clamping: The Cochrane meta-analysis published in 2009 [ 56 ] , based on three randomized trials, revealed statistically insignifi cant decreased blood loss with vascular clamping, when compared with no clamping. Total vascular occlusion is to be avoided unless resection is required at the cavohepatic junction when it cannot be avoided. 4. Hypothermic liver preservation: Interest in decreasing warm ischaemia of transported livers has spawned experiments into isolating the infl ow and perfusing cold preservative into the future liver remnant, which is immersed in crushed ice to maintain the core temperature of the liver at 4 °C. Hypothermic liver preservation when combined with total vascular exclusion attenuates IRI. In situ cold isolation techniques are still in their infancy and remain isolated case reports used in special situations with total vascular exclusion/cardiopulmonary bypass [ 58 ] . 5. Pharmacological preconditioning: It has been reported in a clinical study that the use of isofl urane before clamping the infl ow protected the liver from IRI [ 59 ] . Preconditioning with sevofl urane also signifi cantly reduced postoperative transaminase levels and the overall incidence of postoperative complications was reduced especially in patients with fatty livers. Inhaled nitric oxide has also been cited to 'signifi cantly decreasing the length of hospital stay, improving serum transaminase levels and coagulation times, and reducing the number of apoptotic hepatocytes.' A similar effect has been demonstrated with preoperative administration of 500 mg of methylprednisolone [ 59 ] . During major resections, intraoperative preconditioning with 600 mg of alpha-lipoic acid also reduced markers of hepatic damage by infl ow occlusion.

Lesurtel et al. [ 60 ] have made the following recommendations: with better understanding of intrahepatic anatomy, newer energy devices and maintenance of a low central venous pressure during parenchymal transection, vascular clamping cannot be systematically recommended (level A). Portal infl ow clamping reduces blood loss and use of blood products but does not infl uence morbidity (level A). Among various methods of infl ow exclusion, they support intermittent clamping as better tolerated especially in patients with chronic liver disease (level A). Ischaemic preconditioning has been recommended for steatotic patients (level A). Intermittent clamping is preferred over ischaemic preconditioning in major liver resections and prolonged surgery (level A). Currently no evidence supports or refutes the use of ischaemic preconditioning in donor liver retrievals during living donor transplants.

Portal vein ligation (PVL/PVE) by ligation/embolization : PVL is usually performed by surgical ligation or percutaneously by transhepatic portal vein embolization (PVE). PVE induces apoptosis in the ipsilateral lobe, and hypertrophy and hyperplasia of the contralateral lobe. This increases the functional volume of the FLR, thus obviating hyperperfusion in a SFSS scenario. It is also a precursor phenomenon and predicts the regenerative response in the future remnant. PVE can increase contralateral lobe mass by up to 20 %, with the peak in growth occurring within 2-4 weeks of the procedure [ 61 , 62 ] . The failure of the liver to enlarge after PVE is indicative of patients with impaired regenerative capacity in whom major resection should be avoided [ 62 ] . To prevent a surge in tumour growth due to enhanced differential hepatic artery fl ow to the tumour, local control by ablation/chemotherapy are also added as well as biliary drainage.

Hyperperfusion of a small for size graft is often modulated by splenic infl ow control by ligation/embolization or shunting. PLF is determined by haemodynamic parameters of the hepatic circulation and, specifi cally, by a portal blood fl ow that, when excessive for the volume of the liver parenchyma leads to an infl ow/outfl ow mismatch causing pressure build up in sinusoids with a leaking capillary bed in the liver. Perisinusoidal and periportal haemorrhage occurs within a few minutes in a major hepatic resection as well as after the reperfusion of a SFSS graft. Late effects occur due to hepatic arterial and biliary epithelial hypoxia [ 6 ] .

ALPPS--the 'associating liver partition and portal vein ligation in staged hepatectomy' (ALPPS) strategy is one of the surgical innovations used to manage FLR volumes [ 63 ] . The ALPPS approach is proposed to induce rapid hypertrophy of the FLR in patients with HCC and whose preoperative volume does not allow a safe resection. The procedure entails the combination of in situ splitting of the liver along the Cantlie's line and ligating the portal vein on the side of the tumour. Subsequently the second stage hepatectomy is done. The median FLR volume increase was 18.7 % within 1 week after the fi rst phase and 38.6 % after the second [ 64 , 65 ] . Recently, a number of trials comparing ALLPS and post-PVE liver resections have been published [ 66 -68 ] . ALPPS has shown higher hypertrophy rates compared to PVE/PVL (40-80 % within a week compared to 8-27 %). However, ALPPS has higher sepsis rates (16-64 % of patients) and mortality rates (12-23 %). ALPPS facilitates an early removal of tumours whilst waiting for an adequate FLR. Due to the high morbidity rates there has to be a strict criteria for selecting a patient for ALPPS as PVE has shown comparable FLR hypertrophy rates.

Downsizing tumours with chemotherapy, local ablative techniques and embolization is yet another strategy to gain functional reserve volume when planning resections in livers, which are likely to have a low FLR.

The typical clinical features of PLF parallel the clinical picture of ALF: coagulopathy, raised SB and encephalopathy. In addition renal failure, respiratory compromise, hypotension and features of sepsis may be present. This clinical presentation parallels the presentation of ALF, but is closer to that of subacute liver failure than to that of hyperacute liver failure [ 69 ] . With deteriorating liver function the patient will develop hyperbilirubinaemia and coagulopathy which in particular is a poor prognostic marker [ 70 ] .

If PLF is detected in a patient, it should be scored by the ISGLS system [ 2 ] .

PLF grade A: should be monitored well, but may not require specifi c treatment. PLF grade B: it has to be evaluated if the patient should be placed in an intensive care unit (ICU) PLF grade C: need ICU care Rahman et al. [ 71 ] have cited a daily measurement of serum C-reactive protein (CRP) as an early warning indicator of patients likely to develop PLF. Patients prone to developing PLF had a lower CRP level on POD 1 than patients who did not develop PLF. A prognostic utility of postoperative CRP was a serum CRP <32 g/dl, which was an independent predictor of PLF. Initial treatment of PLF is supportive: ventilatory support, vasopressors, renal replacement therapy and anti-sepsis protocols. Controlling coagulopathy and supporting nutrition are the other mainstays.

Patients of liver resections are normally monitored closely in the intensive care or high dependency units. It is normal for SB levels and the INR to rise in the fi rst 2-3 days postoperatively. SB concentration above 50 μmol/l (3 mg/dl) or INR greater than 1.7 on or beyond 5 days suggests liver dysfunction. Sepsis is indicated by raised serum lactate. The use of antibiotics in patients suffering from ALF is associated with a signifi cant decrease in sepsis and this may also be of benefi t in patients suffering from PLF [ 72 ] . Overall the management of PLF is along the lines for ALF. Identifying and controlling sepsis is the key to managing PLF [ 73 ] .

Trials have shown that prophylactic antibiotics after liver resection do not lead to a reduction in PLF or sepsis [ 74 ] . ALF management guidelines propose that broadspectrum antibiotics should be administered empirically to patients who progress to grade 3 or 4 hepatic encephalopathy, renal failure and/or worsening systemic infl ammatory response syndrome (SIRS) [ 73 , 75 ] .

Many clinicians strive to provide heptoprotection with N-acetylcysteine (NAC) [ 76 ] . However, no evidence exists that it has any benefi t in ALF. NAC is advocated in the management of paracetamol-induced ALF and its use in non-paracetamol hepatic failure remains controversial. Sporadic papers do mention a benefi t for NAC and it is used empirically for its anti-oxidant role. Early stage non-acetaminophen patients with ALF benefi t from intravenous NAC. Patients with encephalopathy grades 3 or 4 do not benefi t from NAC and will require emergency liver transplantation. NAC is commenced in a loading dose of 150 mg per kg per hour for 1 h followed by 12.5 mg per kg per hour for 4 h and 6.25 mg per kg per hour for the remaining 67 h.

The rest of the management is along supportive care protocols as shown in Table 3 .2 . Hepatocyte transplantation has been used in trials as an effort to rejuvenate existing liver function. Intrahepatic hepatocyte infusion [ 77 ] has been used successfully to treat patients with certain metabolic disorders of the liver. Results in liver failure (ALF and PLF) have been poor due to insuffi cient delivery of viable and sustainable functional hepatocytes.

Though, liver support systems have been available for some years now, their high operational cost and sepsis rates have not improved. These include:

1. Molecular absorbent recirculating system (MARS) 2. Modifi ed fractionated plasma separation and adsorption (Prometheus) 3. Bioartifi cial liver (BAL) and extracorporeal liver assist device (ELAD).

Extracorporeal systems are predominantly sustained on albumin dialysis, and bioartifi cial devices are bioreactors with permeable membranes containing hepatocytes, either synthetic or natural. Very few trials exist in the setting of PLF, with the exception of one case series which showed no signifi cant benefi t [ 78 , 79 ] . They are not currently recommended in the medical management of ALF. Because their actual place in the global fi eld of acute or acute on chronic liver failure remains to be determined, their role in PLF is undefi ned. However, because outcomes in PLF are morbid, it is worthwhile to continue to investigate the benefi cial roles of these devices [ 80 -82 ] . 

The use of a rescue hepatectomy (removal of necrotic portions or segments) in patients suffering from PLF may be of value when faced with a very sick patient. It is based on the concept that the 'necrotic liver' is the source of unknown humoral substances that contribute to SIRS [ 83 ] . The effi cacy of orthotopic liver transplantation (OLT) for PLF has recently been reported [ 84 ] . In this paper, Otsuka et al. did a retrospective review of 435 patients who had a liver resection between 1990 and 2004. Nine of them (2 %) developed PLF of which seven were offered OLT at a mean of 25 days post resection. Indications for resection included malignancies and benign disease. Patients developing PLF had signifi cantly altered biochemical and coagulation parameters manifesting on POD 2 and had the classical triad of coagulopathy, hyperbilirubinaemia and encephalopathy. There was no mortality following OLT, though one patient required a retransplant. The mean survival with and without OLT was 42.2 and 1.4 months, respectively (p = 0.03).

They concluded that all patients ( n = 4) who suffered from PLF but were not considered suitable for liver transplantation, died, while all those undergoing OLT survived ( n = 7). OLT allows salvage of an otherwise fatal condition. However, no defi nitive criteria are available for emergency liver transplantation for PLF.

OLT as a rescue for PLF has been gaining in popularity, governed only by the availability of organs or suitable donors in an emergency. The principles of transplantation should however be adhered to: in most instances the indications for OLT after PLF should be limited to patients who fulfi ll the primary indication for HCC-pre-resection tumour burden within the Milan criteria [ 85 ] . Patients with metastatic disease, beyond Milan criteria, advanced medical and anaesthetic comorbid conditions, and poor functional status, should not be candidates for OLT.

OLT is the only radical surgical remedy that improves survival in patients with end-stage liver disease. However, patients suffering from PLF are rarely eligible for liver transplantation because of tumour characteristics or comorbid conditions.

The incidence of PLF after a major hepatic resection averages 8 %. An abnormal FLR is the main cause for the pathogenesis of PLF. Other reasons, which worsen PLF, are hepatic parenchymal congestion, IRI and postoperative sepsis. These can exist singly or in combination.

Risk factors for the development of PLF are small FLR, blood loss, malnutrition, diabetes mellitus and active liver disease. A comprehensive preoperative assessment includes evaluation of liver volume, anatomy and function. A physical examination followed by appropriate clinical tests will identify patients at risk of developing PLF. The patient's liver status, hepatic reserve potential and functional aspect need to be investigated along with the metabolic and haematological derangements, which may lead to postoperative liver failure. Corrective measures should be applied when-ever possible, as curative treatment options are limited. The risk of PLF is high when FLR is below 25-30 % in livers without cirrhosis and below 40 % in livers with preexisting liver disease. PVE and/or two-stage hepatectomy are options when surgery cannot be deferred. Additional liver damage may be prevented by intermittent clamping techniques, though used with caution in steatotic livers. Management principles are on the lines of management of ALF with support of liver, cardiorespiratory and renal support function. Control of sepsis is an important aspect. Emergency liver transplantation has shown promise as a remedy for PLF.

Postoperative liver failure after hepatectomy is a potentially life-threatening complication. Fortunately, even though the rates of liver resection are increasing, the mortality from the procedure is decreasing. The operative mortality at specialized centres varies from 0 to 6 % [ 86 ] . However, the morbidity continues to be high. The defi nition of post-hepatectomy liver failure has not been standardized. Three defi nitions are currently being followed. Two of these (the 50-50 criteria and the International Study Group of Liver Surgery [ISGLS] criteria) have been mentioned by the authors [ 2 , 5 ] . The latter not only diagnoses the condition, it stratifi es the severity into 3 categories. Grade A does not need any change in management strategy, Grade B may be managed without any invasive intervention while Grade C needs alteration in management including invasive intervention. The ISGLS criteria have been validated by one study [ 87 ] . The study detected post-hepatectomy liver failure using this criteria in 11 % of patients-8 % had Grade A, 72 % Grade B and 20 % Grade C. The mortality in these patients was 0 % in grade A, 12 % in grade B and 54 % in grade C. Thus, the grading has been shown to correlate with mortality. However, in 2 separate reports, 41 and 67 % of patients fulfi lling the ISGLS criteria recovered completely highlighting the importance of adequate management rather than just using the defi nition [ 4 , 5 ] .

The third defi nition described by Jarnagin et al. is simple. They defi ne post-hepatectomy liver failure as high bilirubin in the absence of biliary obstruction or leak occurring with ascites and coagulopathy with or without encephalopathy [ 88 ] .

The authors have discussed the possible pathophysiological factors in detail. To this I may add the possibility of hepatic venous outfl ow tract obstruction as described by Lhuaire et al. [ 89 ] The authors documented outfl ow obstruction with contrast enhanced CT, Doppler ultrasound, cavography and assessment of pressure gradient between the hepatic vein and inferior vena cava. They managed the patient successfully with placement of a metallic stent in the left hepatic vein and documented reduction in the hepatic veininferior vena cava pressure gradient. The patient improved thereafter and was discharged. Hepatic venous outfl ow obstruction following hepatectomy has been reported in experimental studies in rats [ 90 ] .

Various risk factors have been indentifi ed and have been discussed elaborately by the authors. This should allow surgeons to select proper patients for extensive resection to avoid post-hepatectomy liver failure. The main issue is the adequacy of the residual liver volume and presence of cirrhosis. While 30 % of residual volume is adequate for extended resections in a normal liver, it is grossly inadequate for a cirrhotic patient in whom at least 40 % of liver volume is required. This brings us to the question of increasing future liver remnant by various strategies. Both portal vein embolization (PVE) and liver partition with portal vein ligation are effective. However, the simplicity of PVE makes it the most used approach. It has the advantage of the tumour biology being assessed during the waiting period of 4-6 weeks. If the tumour progresses especially with chemotherapy the patient should not undergo liver resection. On the question of properly selecting patients for major resection in cirrhosis the prevailing guidelines should be followed. These are Child's A status, platelet count above 100,000/cm, absence of clinically signifi cant portal hypertension, a future liver remnant of 40 %, and the 15 min indocyanine green (ICG) clearance of no more than 15 %. A number of tests are available to assess various functional aspects of the liver but they are cumbersome, not available at most centres and more importantly are not accurate with the exception of 15 min ICG clearance. Even this is not done routinely in the West.

Patients with features of postoperative liver failure by either the '50-50' criteria or the ISGLS criteria have circulatory changes as seen in septic shock such as vasodilatation, increased vascular permeability resulting in accumulation of fl uid in the third space, tachycardia and increased cardiac output leading eventually to hypotension. Coagulapathy too is seen commonly. Altered kidney function usually occurs due to hepatorenal syndrome or acute tubular necrosis because of sustained hypotension causing compromised renal perfusion. With worsening renal function, fl uid accumulates in the periphery or pulmonary bed. This often necessitates renal replacement therapy. With improvement in renal function the liver function also improves. Hepatic encephalopathy is seen more commonly in patients with renal failure because serum ammonia cannot be cleared by either the kidney or the liver. The presence of sepsis is another problem. Hypotension resulting from sepsis is detrimental to liver regeneration essentially due to ischaemia. Endotoxins produced in sepsis interfere with Kupffer cell activation and its function. One should not forget that following hepatic resection there is a depletion of Kupffer cells in the liver. Thus following hepatectomy there is a higher incidence of sepsis, and sepsis interferes with hepatic regeneration [ 91 ] . Therefore, every attempt must be made to avoid sepsis. Execution of the surgical procedure with utmost care avoiding excess blood loss, tissue necrosis, haematoma formation, prolonged ischaemia, etc. should minimize infection.

Management of post-hepatectomy liver failure has been duly addressed. I will emphasize on the fl uid and nutrition therapy. Following hepatectomy the urine output may be low but it is expected. It is not necessary to give bolus fl uid therapy in these patients unless they are grossly oliguric. Cirrhotic patients should probably be given parenteral nutrition in place of dextrose solutions because parenteral preparations have branch chain amino acids, dextrose, medium chain triglycerides, phosphates and vitamins. Usually up to 2 l of such fl uid is necessary. If volume defi cit exists, it should be managed with additional fl uids. Parenteral nutrition in post-hepatectomy liver failure helps liver regeneration and decreases septic complications. The addition of phosphate is extremely important because it is necessary for both energy production and liver regeneration. Moreover the phosphate level goes down considerably after major liver resection [ 92 ] . Since the available phosphate solutions contain potassium, these may need to be stopped if the serum potassium level is high. The phosphate will then need to be replaced orally and the oral forms have sodium. Hence, the serum sodium would need to be monitored.

Posthepatectomy liver failure

Posthepatectomy liver failure: a defi nition and grading by the International Study Group of Liver Surgery (ISGLS)

Indocyanine green retention test avoiding liver failure after hepatectomy for hepatolithiasis

Hepatic insuffi ciency and mortality in 1,059 noncirrhotic patients undergoing major hepatectomy

The '50-50 criteria' on postoperative day 5: An accurate predictor of liver failure and death after hepatectomy

Pathophysiologic observations and histopathologic recognition of the portal hyperperfusion or small-for-size syndrome

Molecular mechanisms of hepatic ischemia-reperfusion injury and preconditioning

Liver dysfunction and sepsis determine operative mortality after liver resection

Advances in critical care management of hepatic failure and insuffi ciency

Liver failure after partial hepatic resection: defi nition, pathophysiology, risk factors and treatment

Diabetes is associated with increased perioperative mortality but equivalent long-term outcome after hepatic resection for colorectal cancer

Post-hepatectomy liver failure

Presence of active hepatitis associated with liver cirrhosis is a risk factor for mortality caused by posthepatectomy liver failure

Major liver resection for carcinoma in jaundiced patients without preoperative biliary drainage

Chemotherapyassociated hepatotoxicity and surgery for colorectal liver metastases

Hepatic complications following preoperative chemotherapy with oxaliplatin or irinotecan for hepatic colorectal metastases

Chemotherapy regimen predicts steatohepatitis and an increase in 90-day mortality after surgery for hepatic colorectal metastases

Severe hepatic sinusoidal obstruction associated with oxaliplatin-based chemotherapy in patients with metastatic colorectal cancer

Impact of steatosis on perioperative outcome following hepatic resection

Steatosis as a risk factor in liver surgery

Seven hundred fortyseven hepatectomies in the 1990s: an update to evaluate the actual risk of liver resection

Prediction of post-operative liver dysfunction by serum markers of liver fi brosis in hepatocellular carcinoma

Nutrition and survival in patients with liver cirrhosis

Perioperative nutritional support in patients undergoing hepatectomy for hepatocellular carcinoma

Kidney failure following liver resection

Frailty as a predictor of surgical outcomes in older patients

Analysis of morbidity and mortality rates in right hepatectomy with the preoperative APACHE II score

Should hepatic resections be performed at high-volume referral centers?

Barcelona-Clínic Liver Cancer Group. The Barcelona approach: diagnosis, staging, and treatment of hepatocellular carcinoma

Perioperative factors and outcome associated with massive blood loss during major liver resections

Volumetric analysis predicts hepatic dysfunction in patients undergoing major liver resection

Postoperative liver failure after major hepatic resection for hepatocellular carcinoma in the modern era with special reference to remnant liver volume

Thromboelastographic changes in liver and pancreatic cancer surgery: Hypercoagulability, hypocoagulability or normocoagulability?

Evaluation of postoperative antibiotic prophylaxis after liver resection: a randomized controlled trial

Hepatotoxicity due to herbal medicines and dietary supplements. Available at www.uptodate.com/contents/hepatotoxicity-due-to-herbal-medications-anddietary-supplements?source=search_result&search=hepatotoxicity+due+to+herbal+medications &selectedTitle=1~150

Liver resections in cirrhotic patients: a western experience

Changes in serum amylase level following hepatic resection in chronic liver disease

Increased serum bile acids in cirrhosis with normal transaminases

The prognostic value of preoperative alkaline phosphatase for resection of solitary liver metastasis from colorectal carcinoma

The relation of preoperative coagulation fi ndings to diagnosis, blood usage, and survival in adult liver transplantation

Nutritional risk factors in major hepatobiliary surgery

Aminopyrine breath test in the prognostic evaluation of patients with cirrhosis

Indocyanine green clearance as a predictor of successful hepatic resection in cirrhotic patients

Hospital mortality of major hepatectomy for hepatocellular carcinoma associated with cirrhosis

Preoperative estimation of surgical risk of hepatectomy in cirrhotic patients

Use of liver stiffness measurement for liver resection surgery: correlation with indocyanine green clearance testing and postoperative outcome

The lidocaine (MEGX) test as an index of hepatic function: its clinical usefulness in liver surgery

Risk assessment of posthepatectomy liver failure using hepatobiliary scintigraphy and CT volumetry

Prognosis of hepatocellular carcinoma: the BCLC staging classification

Predictive indices of morbidity and mortality after liver resection

Impact of model for end-stage liver disease (MELD) score on prognosis after hepatectomy for hepatocellular carcinoma on cirrhosis

Virtual hepatic resection using three-dimensional reconstruction of helical computed tomography angioportograms

Strategies for safer liver surgery and partial liver transplantation

A prospective randomized study in 100 consecutive patients undergoing major liver resection with versus without ischemic preconditioning

Effects of Pringle manoeuvre and ischaemic preconditioning on haemodynamic stability in patients undergoing elective hepatectomy: a randomized trial

Ischaemic pre-conditioning for elective liver resections performed under vascular occlusion

Portal triad clamping or hepatic vascular exclusion for major liver resection. A controlled study

Hepatocellular carcinoma on cirrhosis complicated with tumoral thrombi extended to the right atrium: results in three cases treated with major hepatectomy and thrombectomy under hypothermic cardiocirculatory arrest and literature review

Pharmacological interventions for ischaemia reperfusion injury in liver resection surgery performed under vascular control

Clamping techniques and protecting strategies in liver surgery

Portal vein embolization in preparation for major hepatic resection: evolution of a new standard of care

Portal vein embolization before right hepatectomy: Prospective clinical trial

Portal vein ligation and partial hepatectomy differentially infl uence growth of intrahepatic metastasis and liver regeneration in mice

Safety and effi cacy of portal vein embolization before planned major or extended hepatectomy: an institutional experience of 358 patients

Portal vein embolization before liver resection: A systematic review

Salvage parenchymal liver transection for patients with insuffi cient volume increase after portal vein occlusion--an extension of the ALPPS approach

ALPPS offers a better chance of complete resection in patients with primarily unresectable liver tumors compared with conventional-staged hepatectomies: Results of a multicenter analysis

Little girl who conquered the

Acute liver failure

Management of post-hepatectomy complications

Prognostic utility of postoperative C-reactive protein for posthepatectomy liver failure

Prospective controlled trial of selective parenteral and enteral antimicrobial regimen in fulminant liver failure

Surviving Sepsis Campaign guidelines for management of severe sepsis and septic shock

Prospective randomized trial of systemic antibiotics in patients undergoing liver resection

Management of acute liver failure

N-acetylcysteine for nonparacetamol drug-induced liver injury: a systematic review

Prospects for stem cell transplantation in the treatment of hepatic disease

Artifi cial liver support at present and in the future

Artifi cial and bioartifi cial support systems for liver failure: A Cochrane Hepato-Biliary group protocol

Technology insight: artifi cial extracorporeal liver support--how does Prometheus compare with MARS?

Application of molecular adsorbent recirculating system in patients with severe liver failure after hepatic resection or transplantation: Initial single-centre experiences

In vivo quantifi cation of liver dialysis: Comparison of albumin dialysis and fractionated plasma separation

Liver derived pro-infl ammatory cytokines may be important in producing intracranial hypertension in acute liver failure

Postresection hepatic failure: Successful treatment with liver transplantation

Management of hepatocellular carcinoma

Defi ning postheptectomy liver insuffi ciency: Where do we stand?

Post operative course and clinical signifi cance of biochemical blood tests following hepatic resection

Improvement in perioperative outcome after hepatic resection: analysis of 1,803 consecutive cases over the past decade

Post-hepatectomy liver failure: should we consider venous outfl ow?

Impact of hepatic vein deprivation on liver regeneration after major hepatectomy

Loss of NF-kB activation in Kupffer cell-depleted mice impairs liver regeneration after partial hepatectomy

The clinical implications of hypophosphataemia following hepatic resection or cryosurgery