Liver Disease Case 2

Answer the questions when all the preceding data has been reviewed

History & Examination

A 35 year old alcoholic presents to ED/ER. He has been drinking one bottle of vodka per day for some weeks/months. He is unkempt, emaciated and looks much older than his years. He has never attended the West Hampton or any other hospital in the past. A former private in the British army, he is now unemployed and living alone in a bedsit on welfare. He has been repeatedly vomiting in recent days and has some minor right upper quadrant pain. He had ignored several episodes of melena in the weeks leading up to this admission and had continued drinking. He has no family in London, both of his parents are dead. He would not have attended at all had it not been for the intervention of a neighbour who had called an ambulance having found him slumped on the stairs of their block.

He is cooperative, quiet and although tremulous, no asterixis can be demonstrated. A 5 cm hepatomegaly is present and the liver is mildly tender on palpation. He is jaundiced. No ascites or splenomegaly are detected. He has no spider naevi.

On rectal examination normal formed stool is present.


Soldier CSI

ECG: could not be found. SHO recorded in the notes that it showed sinus tachycardia rate 110 bpm. Repeat ECG is requested.

CXR report: no abnormality identified


sol obs CSi


Q1. What is the differential diagnosis at presentation?

Q2. What additional tests should be requested at presentation?

Q3. Is there a case for liver biopsy at presentation?

Q4. Why is he hypokalemic?

Q5. Given the diagnosis in the answer to Q1, What are the key points in the management of this condition?


Q1. What is the differential diagnosis at presentation? 

1. Alcoholic hepatitis v  decompensated cirrhosis.

The patient is jaundiced on a background history of heavy alcohol intake. There is clear evidence of hepatic dysfunction (high INR in the absence of anticoagulants and low serum albumin). The findings on abdominal examination are limited to tenderness over an enlarged liver. Taken together this picture is consistent with a clinical diagnosis of alcoholic hepatitis. Note that the liver enzymes are only marginally elevated and the hepatitic and bilary enzymes (link) are elevated to a comparable extent. This is typical of alcoholic hepatitis. Indeed, if the liver enzymes are grossly elevated (eg Case 1), we should at least consider alternative diagnoses characterised by acute hepatocyte destruction, for example acetminophen (paracetamol) overdose or viral hepatitis. The elevated WCC with neutrophilia and a low grade pyrexia are also typical findings in patients with alcoholic hepatitis. Although typical, they are of course, nonspecific.

Alcoholic hepatitis is a clinical syndrome characterised by inflammation of the liver secondary to excessive alcohol intake. It is usually observed in individuals with a history of excessive alcohol intake over a prolonged period.

Our case is typical, demonstrating 1) prolonged excessive alcohol intake 2) a low grade pyrexia 3) deranged LFTs with jaundice 4) hepatomegaly with a degree of tenderness on palpation and also 5) the presence of a neutrophilia, a typical feature (but not mandatory for the diagnosis). The syndrome is characterised by a variable degree of overt liver dysfunction. Alcoholic hepatitis is a clinical diagnosis which we sometimes get wrong. We must be very open to revising our diagnosis: alternative diagnoses particularly infection must be borne in mind. You get the idea, we make the diagnosis and then keep it under review.

It is important to try and detect the presence of underlying cirrhosis in such a patient. Acute decompensation of underlying cirrhosis with all its possible causes is an important differential in this case. However, remember not all patients with liver disease are cirrhotic! Even a patient with stigmata of chronic liver disease may not have underlying cirrhosis. Also, as we saw in our first case in this section, you don’t have to cirrhotic to manifest evidence of hepatic dysfunction.  Cirrhosis is a specific histological diagnosis and to make that diagnosis definitively we must show nodular regeneration with fibrosis and architectural distortion of the liver tissue under the microscope. We can of course infer (guess at) the presence of underlying cirrhosis on clinical assessment. The architectural distortion results in portal hypertension which in turn can result in a number of overt clinical findings, not specific for, but in the context described here, suggestive of, underlying cirrhosis. These signs include the presence of ascites, splenomegaly and grossly distended periumbilical veins (caput medusae). Our patient has none of these signs. We did not think he had cirrhosis. He has developed an acute alcoholic hepatitis probably on a background of chronic alcoholic liver damage. Could we be wrong? Definitely.

2. Probable GI bleed at some point in recent weeks

Q2. What additional tests should be requested at presentation? 

1. Septic screen. The patient has a raised temperature. Alcoholic hepatitis is a recognised cause of low grade pyrexia and is classically associated with a raised WCC (neutrophilia). However we must rule out other possible causes of fever. We must request a CXR, blood cultures, urinalysis and mid stream urine (MSU) for culture and microscopy.

Michael Ball tells us that ‘love changes everything’. It is, in fact, blood cultures which change everything. As we’ll see in future cases a positive blood culture may leads us down a totally different diagnostic pathway and often saves the day. If a patient is pyrexial, do blood cultures.

2. As an alcoholic he is at risk of other upper GI pathologies. He should have a serum amylase to exclude pancreatitis.

3. A VBG to screen for pH disturbance would be reasonable. He is an alcoholic, is there a possibility of poisoning ?methanol ?antifreeze.

4. Viral serology (Hepatitis B or C?)

5. Plasma glucose. Liver dysfunction is a potential cause of hypoglycemia which we must correct if present.

6. Urine toxicology for drugs of abuse.

7. Liver/abdominal ultrasound. This will give a lot of important information. It will help exclude other causes of fever and right upper quadrant tenderness such as cholecystitis/cholangitis. It will tell us something about the echotexture of the liver, whether there is a possibility of underlying cirrhosis. It will identify ascites, a possible site of infection. This could be important if we are incorrect in assuming that the fever is due to alcoholic hepatitis.

8. OGD. He has had episodes of melena in recent weeks. He is at risk of peptic ulcer disease. He may have undiagnosed cirrhosis, so, oesophageal varices are a possibility. He should have an OGD early in this admission. How early? Well that depends on where you work. He is young and appears to be heamodynamically stable so within the first 24 hours of admission would appear reasonable.

9. At some point during his admission when he is stabilised we will carry out a CT Liver. This may help guide a decision towards the need or otherwise for a liver biopsy to exclude cirrhosis.

Q3. Is there a case for liver biopsy at presentation? 


Alcoholic hepatitis is a clinical syndrome. There is considerable variability in clinical presentation and it is often misdiagnosed. Some would argue that definitive diagnosis depends on histological assessment of the liver tissue. There are well established histological criteria for diagnosis 1) Infiltration of liver lobules, the dominant cell type being neutrophils. Neutrophil infiltration is unusual in other causes of liver inflammation. 2) Hepatocellular necrosis. 3) fibrosis and sclerosis and 4) accumulation of Mallory bodies in the hepatocyte cytoplasm.

The neutrophil seems to play a central role in the liver damage of alcoholic hepatitis, however, exactly how the neutrophil infiltrate damages the liver is unknown. The usual suspects, reactive oxygen species or proteases have been implicated but nobody really knows.

There is some evidence that the diagnosis of alcoholic hepatitis may be safely made clinically if certain criteria are met. However, the exact approach will vary from unit to unit.

Q4. Why is he hypokalemic? 

There are a number of reasons. He has been vomiting. Vomiting is a potent cause of hypokalemia as secondary hyperaldosteronism triggers potassium loss in the urine. Also, maintenance of normal potassium levels in the body is critically dependent on adequate oral intake. He is emaciated and, as is often the case with alcoholics, malnourished.

Q5. Given the diagnosis in the answer to Q1, What are the key points in the management of this condition? 

We have gone for a diagnosis of alcoholic hepatitis. Remember, this is a tricky diagnosis. Formal diagnosis would require a liver biopsy. This is hazardous and not usually carried out in these situations. We have made a clinical diagnosis based on the history of heavy long term alcohol abuse combined with the clinical findings (tender hepatomegaly, tachycardia, jaundice and fever: nonspecific but typical finding) and laboratory features (raised bilirubin, mild transaminitis on the LFTs).

The management of alcoholic hepatitis has altered in recent times. There used to be a vogue for treating alcoholic hepatitis with oral steroids and Pentoxifylline. The use of oral steroids was reserved for severe cases based on a Maddrey discriminant function (MDF) score of > or = 32. The MDF is calculated using an equation based on the INR and serum bilirubin concentration. Recently, clinical trials have shown no statistically significant benefit to the use of steroids in acute alcoholic hepatitis (STOPAH trial) vindicating years of scepticism. Despite this evidence, several internet sources still recommend steroid use in this setting with the appropriate MDF. You should consult your own local guidelines on this issue. The STOPAH trial also showed the treatment with pentoxifyllin has no effect on survival in alcoholic hepatitis.

He will need treatment with regular chlordiazepoxide (benzodiazepine) to prevent delerium tremens.

He should be commenced on B vitamin supplements. We need to address his nutritional status with review by the dieticians. We may consider nasogastric feeding when the GI bleed issue is sorted out.

He has been given vitamin K and commenced on an IV proton pump inhibitor for the GI bleed.

As alcoholic hepatitis is associated with low grade pyrexia and as no obvious infection could be identified, this patient was not started on antibiotics. This decision was kept under review.

Inpatient Course

The admitting registrar diagnosed alcoholic hepatitis and gastrointestinal haemorrhage.

Management of alcoholic hepatitis

He is given regular benzodiazepines to prevent Delerium tremens. IV vitamin K is administered. As an alcoholic, he is at risk of various water-soluble vitamin deficiencies and a course of IV B vitamins is prescribed. His hypokalemia was treated with 40 mmol of potassium (K+) diluted in one litre of normal saline administered by IV infusion.

Management of gastrointestinal (GI) bleed

On the basis of his admission FBC result (7 gr/dl) and his recent history of GI bleed, he was transfused 3 units of packed red cells over the night of admission. In view of his INR result, he was given 600 ml (3 units) of fresh frozen plasma. He is commenced on an intravenous proton pump inhibitor. His repeat bloods 24 hours following admission are shown below.

An OGD, carried out 24 hours post-admission, identifies a duodenal ulcer with no visible vessel. There is no evidence of portal gastropathy and no evidence of varices. CT abdomen shows no evidence of splenomegaly and liver echotexture is not suggestive of the presence of underlying cirrhosis.

Investigations (day 2)


Soldier II CSi

In the days following admission, there is an initial period during which his overall condition improves. On day three, however, it is noted on the ward that he is becoming increasingly confused and agitated. His speech has become slurred. The intern is asked to increase his benzodiazepine dose as he is ‘clearly withdrawing’. The young intern is thoughtful and also checks an alcohol level (negative) and carries out a full neurological examination. He thinks that there may be a few beats of clonus affecting the right leg but is unsure of its significance. He increases the benzodiazepine dose.

The next day the patient is disorientated and confused. His speech is clearly slurred and he now has clonus affecting both lower limbs and bilateral upgoing planter responses. It is difficult to assess power due to his disorientated state. An urgent CT brain is reported as normal. The neurologists are asked to review. The cranial nerves are difficult to examine but they feel that, in addition to the findings in the lower limbs, a pseudobulbar palsy may be present.


Q1. What additional imaging is necessary?

Q2. Go over things again. Can you offer a suggestion as to the origin of the patients neurological problems?

Q3. What do you think of the management of his GI bleed at initial presentation?


Q1. What additional imaging is necessary? 

MRI brainstem/cervical cord

CT brain will miss certain CNS pathologies and is of limited value in looking for disease in the brainstem. MRI is often of value in these situations but where are we going to image?

Assuming his neurology is due to focal CNS pathology. In general medicine, we are usually faced with unilateral neurological signs. For example, stroke affects the right side of the brain, the signs appear on the left. When faced with signs on both sides of the body, as in the case here, this implies that the pathology is somewhere central, where the long tracts or cells giving rise to them are close together. The nature of his signs implies upper motor neuron involvement. In our case, if his neurology is due to focal pathology the involvement of cranial nerves would tend to make us think that this pathology lies somewhere in the brainstem. The long tracts are also close together in this site. We requested an MRI of brain, brainstem and cervical cord.

Q2. Go over things again. Can you offer a suggestion as to the origin of the patients neurological problems?

Alcoholics may experience a range of neurological problems at or following admission to hospital. These problems are easily confused with the effects of alcohol withdrawal. Alcoholics are at risk of hepatic encehalopathy, arterial dissections, Wernicke’s encephalopathy and perhaps in older age groups, dementia. Unfortunately, comparison of his blood tests on admission with those one day later raises another possibility. Central pontine myelinolyis (CPM, see article below). Note that his sodium on admission was very low at 109 mmol/l (normal 135 – 145). We frequently see this extreme level of hyponatremia in alcoholics and, perhaps due to the compensatory mechanisms described in our piece on CPM, they are often remarkably tolerant of it. Overnight, however, his hyponatremia has corrected rapidly towards normal (130 mmol/l). This may have been secondary to fluids he had received including the blood products aimed at treating his GI bleed. Blood products have a high sodium content. Such a rapid correction of hyponatremia in alcoholics may be associated with development of the devastating brainstem pathology, CPM.

Unfortunately, in this case, MRI of the brain and spinal cord confirmed the presence of central pontine myelinolysis. There is no specific therapy for CPM and the degree of recovery, if any, is highly variable occurring over a period of months. Many patients are left with significant neurological deficits. It is also worth noting that although ‘safe’ rates of correction of serum sodium are reported in the literature (Na+ rise in concentration of less than 12 mmol/l in 24 hours), CPM has been reported in cases of sodium correction within these limits. On the rare occasions when we need to directly intervene to correct serum sodium, great care is needed.

Q3. What do you think of the management of his GI bleed at initial presentation?

His CPM may have been triggered by the rapid correction of his serum sodium overnight. Many blood products are hypertonic and the doctors caring for the patient subsequently postulated that the rapid correction of his hyponatremia over the first night post-admission may have occurred secondary to the high sodium content of the blood products administered. In addition, his hypokalemia was corrected by IV infusion of potassium diluted in normal saline (vide infra).

It certainly could be argued that the management of his GI bleed at presentation was overaggressive. With a Hb of 7 gr/dL, he was transfused 3 units of packed red cells. Presumably with a target Hb of 9 – 10 gr/dl. In view of his elevated INR (1.9) in the context of GI bleeding he was also given 600 ml of fresh frozen plasma to reverse his coagulopathy. Of course, in cases of life-threatening gastrointestinal haemorrhage, we have no choice but to transfuse aggressively. However, whether the case discussed here needed such aggressive transfusion is certainly questionable. He had been losing blood from the GI tract episodically in recent weeks (reported episodes of melena). However, there was no evidence of active bleeding at the time of presentation (normal stool on PR) and his blood pressure seemed stable with no postural drop. We can explain his tachycardia on the basis of his alcoholic hepatitis.

In any case, recent evidence suggests that even in the presence of active bleeding, over-transfusion of blood products can have a detrimental effect on patient outcome, promoting re-bleeding and augmenting coagulopathy. It has been suggested that a Hb of >7 gr/dL is an acceptable target Hb in active GI bleeders with no comorbidity. In the presence of significant comorbidity, eg ischaemic heart disease (not present in our case), the target Hb rises to > 9 gr/dL.

A less aggressive approach to raising his haemoglobin accompanied by an early (within 24 hours) endoscopy might have been wise. I suspect that if the doctors had considered that this amount of blood product and fluids could correct his sodium so rapidly it would not have been given.

Also, the need to treat this patient’s hypokalemia at presentation with IV potassium is very questionable. Administration of IV potassium carries risks of arrhythmia due to iatrogenic hyperkalemia. Also, in this case, it necessitated further administration of Na+ as normal saline was used as a diluent. His hypokalemia was relatively minor (3.1 mmol/l) at presentation. His ECG shows sinus tachycardia only which can be accounted for on the basis of his alcoholic hepatitis. Agressive replacement of potassium was probably unnecessary in this case and simply increased his Na+ burden. In general, mild asymptomatic hypokalemia should be managed with oral K+ supplementation.

One thing you learn after years in this job is that everything you do to a patient, even the simplest intervention, carries potential consequences for them.

Subsequent Course


MRI of brainstem and cervical cord confirmed the presence of central pontine myelinolysis (CPM). Within days, he was completely bedbound, unable to move or speak and doubly incontinent. He was being fed via a nasogastric tube. Following an episode of probable aspiration, he experienced a cardiac arrest. Limited attempts at resuscitation were unsuccessful. Post mortem examination confirmed the presence of gastric content aspiration. The MRI diagnosis of CPM was confirmed.

A sister of the patient, at the time living in Gibraltar, was eventually traced and contacted. She wished to meet a member of the medical team to discuss her brothers care. The course of events and reasons for her brother’s death were explained to her in detail, including the questionable interventions which may have triggered the CPM. She understood the difficulties the doctors had faced. She was not interested in finding fault, she had merely wanted to know how and why her brother had died. The meeting ended amicably.

Unusually enough in the great city of London, a few days later, one of the doctors involved in the case was boarding the tube at Paddington when he saw the man’s sister sitting across from him in the carriage. They smiled at each other and he sat down beside her to talk. There was a long history of alcohol problems in her family. Her mother had always ‘liked a drink’ but following the death of her young husband while stationed in Northern Ireland, financial difficulties and isolation had got on top of her.  Her drinking got out of hand and her children were taken into care. The brother, who had been very proud of his Dad, had reacted badly. In their teenage years, the siblings had lost touch with one another completely. Despite their difficulties, the woman’s memories of her childhood with her brother were happy ones. In recent days, her thoughts were dominated by the simple memory of the two of them as children playing together on the kitchen floor while their mother did the housework. The doctor helped her with her bags as she left the train at Earls Court. Again, she thanked him for the efforts made on her brothers behalf. They shook hands and wished each other well.

In medicine there is too much to know for any one individual to know. No doctor can ever be ‘bullet proof’. Therefore, even with the best intentions, things will go badly wrong from time to time. Fortunately for us, in these situations we can almost always depend on the kindness and wisdom of strangers.


Central Pontine Myelinolysis

In the 1950s, Adams and colleagues described the development of acute demyelinating lesions in the brain of malnourished alcoholics. The lesions they described tended to focus on the basis pontis, hence the term central pontine myelinolysis (CPM). Demyelination of the motor tracts in the pons in this condition results in quadraparesis associated with upper motor neuron signs (the upgoing planters in our case?). Loss of cranial nerve function results in distressing problems with speech and swallowing. An affected patient may ultimately develop a ‘locked in’ syndrome.

The introduction of routine electrolyte measurements in the 1970s lead to the identification of chronic hyponatremia as a major factor in determining the susceptibility to develop CPM. Furthermore, the clinical suspicion that rapid correction of serum sodium concentration in individuals with chronic hyponatremia was the trigger for this pathology was supported by findings in animal models. Rapid correction of chronic hyponatremia appears to act as a trigger for CPM. Patients in whom an acute fall in serum sodium is corrected appear to be at less risk. This has lead to the hypothesis that, in patients with chronic hyponatremia, neurons are able to raise the osmolality of the ECF surrounding them by extruding osmotically active substances (osmolytes) into the ECF. This compensatory response however places them at risk of profound cellular dehydration in the CNS if a doctor comes along and raises the ECF osmolality further by administering sodium in intravenous fluids.


The figure above summarises the pathogenesis of CPM. Under normal circumstances, the balance of fluid distribution between the intracellular fluid of myelin producing cells in the CNS and the extracellular fluid is regulated by the opposing osmotic pressures of Na+ in the ECF and intracellular fluid (ICF) K+. B. The acute development of a low ECF Na+ concentration (hyponatremia) results in a fall in ECF osmotic pressure and a shift of water (blue arrow) from the ECF to the ICF (cellular oedema). This cellular oedema may result in CNS symptoms. C. However, with time (>48 hours) the cells can extrude osmotically active substances ‘osmolytes’ (red) into their surroundings returning the osmotic pressure differential surrounding the cell to normal and normalising fluid distribution between the two compartments. D. However, medically normalising the serum Na+ when this compensated situation has stabilised, results in a supranormal osmotic pressure surrounding the cell with flow of water from the intra to extracellular fluid compartment (blue arrow). The resultant cellular dehydration is poorly tolerated in the pons with destruction of vital cells.

Compensation will not have taken place within the first 48 hours of the development of hyponatremia. Therefore, correction of hyponatremia in the acute setting is unlikely to trigger CPM. This is why you will hear much talk about the importance of distinguishing acute (<48 hours old) and chronic hyponatremia (present > 48 hours) in terms of planning management.

A few comments on managing hyponatremia

This is an academic discussion for medical students, no guidance regarding the need to, or method of, treating patients with hyponatremia is intended here.


 ‘The road to Hell is paved with good intentions’

Saint Bernard of Clairvaux (c1150)

Hyponatremia may be associated with cerebral dysfunction which may endanger the patient. Hyponatremia may be classified biochemically as mild (130 – 135 mmol/l), moderate (125 – 129 mmol/l) or profound (<125 mmol/l). If a patient’s symptoms are related to hyponatremia, these symptoms may be classified as moderate  (nausea, confusion, headache) or severe (vomiting, somnolence, seizures, coma). In general, the larger the fall in serum sodium the more marked the CNS effects.

These finding are crucially affected by the rate at which the hyponatremia has developed. If the hyponatremia develops acutely the effects on brain function may be marked. If the hyponatremia has developed over a period of days or weeks the compensatory mechanisms discussed in this case may minimise the effects on CNS function.

It is rarely necessary to treat hyponatremia directly and the decision to do so must be made by a senior doctor. Most cases of hyponatremia will correct with treatment of the underlying cause.

Hypertonic saline

In a situation where hyponatremia is responsible for severe CNS symptoms (low GCS, seizures etc), the hand of a senior doctor may be forced. Direct treatment of the hyponatremia with hypertonic saline may be required. A change to practice in Europe is likely to follow the publication of a set of guidelines by a group of European medical organisations from relevant specialities. Perhaps the most useful suggestion in this paper is that when faced with absolutely no choice but to correct a low serum sodium because of severe symptoms related to the hyponatremia, we can use an IV infusion of 150 ml 3% hypertonic saline administered over 20 minutes. While repeating the infusion, the serum sodium is rechecked at the 20 minute time point, aiming for a rise in concentration of 5 mmol/l. The rise in serum sodium should be limited to 10 mmol/l in the first 24 hours and an additional 8 mmol/l every 24 hours thereafter. This is irrespective of the patients weight or the chronicity of the hyponatremic state. Using a simple standard infusion avoids the need for mathematical equations previously used to calculate the volume of hypertonic infusate. These equations put people off as there is a risk of miscalculation, a need to estimate the patients weight and a lack of general agreement as to which equation is best. I am not aware of any transatlantic guidelines and you should learn what is considered appropriate in your part of the world.