Pitfalls of using amylase to confirm pancreatitis

A 30 year old man presented intoxicated to A&E. He didn’t speak much English, and was triaged as ‘abdominal pain and vomiting’.  With the help of a multilingual colleague, we managed to work out that he had actually had three episodes of hematemesis that morning, followed by three episodes of maelena and was an alcoholic. There was abdominal pain, but it had been there three months and not changed recently.

His pulse was 110, BP 113/80 with a cap refill of less than 2 seconds. He was walking and talking.

His abdomen was soft and non tender.

He had a non zero Blatchford score because of his pulse >100. According to NICE, only those scoring zero should be considered for early discharge.

The patient was accepted by the medics. One hour later, I got a call back from them.

“That patient you referred…well, his amylase is 311. He’s probably got a pancreatitis.  He really shouldn’t be under the medics.”

A moderately raised amylase is a common problem, especially in patients vomiting and/or with abdominal pain.

Where does serum amylase come from?


It’s also important to realise how amylase is eliminated. Although the exact metabolism of amylase has not been established, patients with one kidney have a roughly three fold higher baseline amylase level compared to patients with two kidneys, which suggests renal clearance is important. There is probably also a role for hepatic clearance.

Amylase in pancreatitis


In pancreatitis, the serum level of the pancreatic isoenzyme usually rises. This rise diminishes with time. This means that using any arbitrary value e.g. four times the upper limit of normal as ‘diagnostic’ of pancreatitis without a timeframe is as meaningless as saying a random troponin at any time above a certain value is diagnostic of MI (see page iii3 of http://www.bsg.org.uk/pdf_word_docs/pancreatic.pdf)

In patients with chronic pancreatitis, or recurrent acute pancreatitis, there may be a reduced amylase reserve, leading to a less spectacular amylase rise if any.

Some non pancreatitis causes of a raised amylase

In perforation, there is increased absorption of amylase from the peritoneal cavity. A perforated posterior peptic ulcer could cause severe pain radiating to the back +/- vomiting and return an amylase value that is moderately raised.


In parotid hypertrophy, the baseline amylase level may be raised. This occurs in alcoholism, which constitutes one of the populations at risk of pancreatitis.


In any cause of mucosal inflammation such as inflammatory bowel disease, there will be a disturbed blood gut barrier. This will lead to an increase absorption of amylase.


In a GI bleed, this barrier will also be disturbed. In a patient with a history of a GI bleed with a non tender abdomen, I think that a moderately raised amylase should be attributed to the upper GI bleed unless there were further evidence to consider a diagnosis of acute pancreatitis.

For a full list, see emedicine.

None of this matters

We should be using lipase for acute pancreatitis anyway.  According to the UK Pancreatitis guidelines:

“Although amylase is widely available and provides acceptable accuracy of diagnosis, where lipase is available it is preferred for the diagnosis of acute pancreatitis (recommendation grade A).”