Management of acute liver failure

Management of acute liver failure seldom makes an appearance in the college exams, which makes some sense because there are only a few genuinely helpful strategies to deploy, and most of these are related either to reversal of some sort of toxicity (eg. paracetamol overdose) or the management of hepatic encephalopathy. 

The major influences on these lists of management options were the following papers:

• The position statement by Stravitz et al (2007)
• Discussion of  "quadruple-H therapy" by Warrilow and Bellomo (2014).
• Management strategies by Warrilow and Bellomo (from 2015)
• AASLD statement (2011)
• AGA guidelines (2017) - mainly for various diagnostic recommendations
• EASL guidelines (2017) - though they tend to fall into the same trap as the Surviving Sepsis guidelines, i.e. in their discussion of management for liver failure the authors appear determined to also summarise all existing ICU literature on the management of sepsis, ARDS, fluid therapy and nutrition in critical care.

Specific management

Usually we have very little to offer these people.

Obviously, the withdrawal of the hepatotoxic drug is a major specific step for the management of drug-induced hepatitis, but for the majority of other causes, there is no specific treatment -merely "wait-and-see". Ultimately, the last option is liver transplantation.

Broadly, the goals of therapy are:

• Identify causes which might be potentially reversible
• Maximise the chances of hepatic regeneration by supporting organ function
• Identify the patients who are unlikely to survive supportive management (these need to be referred for urgent transplantation)

There are only a handful of conditions which have reliable treatment options beyond normal ICU support.

In brief summary:

Beyond this, all treatment is "supportive", i.e. there is nothing to be done for the primary cause of liver failure. One can merely stand by and correct all the various abnormalities which take place in the organism. The real definitive solution is transplantation, The EASL guidelines (2017) claim a survival of ~ 80% from this procedure.

Supportive management

Management plans should include the following points: 

1.  Intubate the patient for airway protection, as they will be obtunded (and to control the CO₂). The patient with fulminant liver failure is very likely to require intubation at some stage.
2. Hyperventilation to low-normal PaCO2 (35 mmHg) is a part of HHHH therapy, and is mainly directed to prevent the cerebral vasodilation and intracranial hypertension which is associated with acute liver failure.
3. Vasopressor support: Maintain haemodynamic stability using noradrenaline preferentially. These patients will vasodilate extensively. It will be important to avoid fluid overload because that tends to impair hepatic venous outflow.
4. Sedation with short-acting drugs. Use propofol instead of benzodiazepines, and avoid long-acting opiates.
   Consider an ICP monitor. This offers you a means of monitoring the development of massive and undetected ICP fluctuations, presumably because you will react to them with more HHHH therapy. It is surprisingly safe - Rajajee et al (2017) trialled a protocol where all their acute liver patients got an ICP monitor and found that only one of the 24 had an intracranial haemorrhage (which was apparently asymptomatic and had "an excellent outcome"). All patients were given some Factor VIIa no more than 1hr prior to the procedure, and they all used delicate little intraparenchymal monitors rather than EVDs.
   Management of raised ICP for these patients does not differ significantly from what you'd normally do for stroke or TBI. If the ICP is uncontrollable by normal means, various extreme authors have suggested various extreme measures. Ringe et al (1988) found that the dying liver was causing more harm than good, and removed it entirely ("we considered it a potential advantage to tolerate a prolonged anhepatic period before implantation of a functioning allograft than to leave the necrotic liver in situ")
5. Hypothermia to a temperature of 32-33° - 
   Hypernatremia (to control ICP, with hypertonic saline) to achieve a sodium of 145-155 mmol/L
   Replace phosphate: hypophosphataemia tends to develop in the recovery phase, as hepatocytes reproduce vigorously. As a major intracellular anion, phosphate will be sucked up into the rapidly growing cells. Also you need it to make use of all that glucose you are infusing. In summary, give phosphate.
6. Haemodiafiltration - continuously - to remove ammonia. This prevents acute cerebral oedema. In hyperacute liver failure cerebral oedema may actually develop before hepatic encephalopathy or other major complications of liver failure. Don't use citrate.
   Fluid resuscitation with crystalloid - keeping in mind that fluid overload is undesirable. EASL guidelines (2017) recommend something buffered with acetate.
7. Give 10-50% dextrose as infusion. Anticipate hypoglycaemia and increased resting energy expenditure. Caloric requirements are  increased by 18 to 30% compared with normal controls (Schneeweiss, 1992) and there does not seem to be any special increase the the metabolism of any specific macronutrient group. The authors of that 1992 study estimated that normal hepatic glucose release  rate is about 8μmol/kg/min, and so one should aim to duplicate this with exogenous glucose in the anhepatic patient. That ends up being about 6-7g/hr of dextrose for a 70kg patient, or approximately 125-150ml/hr of 5% dextrose.
   Think about enteral lactulose for management of hepatic encephalopathy, though it is somewhat controversial. Lactulose in acute liver failure is only described as "possibly helpful" by the AASLD statement (Polson et al, 2005) and is not recommended by Stravitz et al (2007). There is no evidence that it improves survival in acute liver failure, and it may distend the gut, causing aspiration and obscuring the operative field during the liver transplant. Bernal & Wendon (2013) described it as dangerous, whereas others defended its use in resource-poor environments where liver transplantation will not be an option.
   Ensure PPIs are administered. Don't give them any excuse to have a GI bleed.
   Drain the tense ascites. It behaves like a gravid uterus, from a haemodynamic standpoint. If there is tense ascites, draining it could potentially improve venous return and haemodynamics.
8. Anticipate coagulopathy.
   Administer Vitamin K empirically, however little that is expected to have. Apparently the correct dose is 10mg. Pereira et al (2005) gave this dose to their patients and found that 27% had some sort of subclinical Vitamin K deficiency. Hard to say whether this had any positive effect on their INRs, let alone survival
   Consider blood products, but view complete correction as unobtainable.
   Remember: in spite of apparent "numerical" coagulopathy, a hypercoagulable state develops.
9. Vigilant surveillance for sepsis: they are prone to it, and it makes the encephalopathy worse. The EASL guidelines (2017) recommend daily surveillance cultures.

As raised by some of the readers, lactulose in acute liver failure is something of a polarising issue here.