One of the rather difficult issues is to decide when a person has Barrett’s oesophagus and when they do not. It should be straightforward but it isn’t.
The first thing is that instead of the usual pale pink lining, the person must have a visible abnormality of ‘salmon pink’ or red lining in the wall of the lower oesophagus. Note that this ‘salmon pink’ lining looks exactly the same as normal stomach lining.
In the USA, the second requirement is the presence of a particular type of ‘columnar metaplasia’ which is called ‘intestinal metaplasia’. This is diagnosed when biopsy samples are removed from the area of suspected Barret’s oesophagus and viewed down the microscope. Intestinal metaplasia is diagnosed when cells which belongs in the intestine are found. These are called goblet cells. These cells can be seen easily when biopsy tissue samples are viewed down the microscope. There is, however, another type of columnar cell which is found normally in the stomach.
So, in the USA, the diagnosis requires both a visible abnormality and characteristic findings when biopsy samples are viewed down the microscope. This combination makes it unlikely that the sample has been taken from the normal stomach or a hiatus hernia because most people will not have intestinal metaplasia there.
In the UK, we also need two abnormalities. The patient must have the visible ‘salmon pink’ lining but as long as there are ‘columnar cells’ (which are also present in the normal stomach), the person can be said to have Barrett’s oesophagus.
What happens if the British doctor sees what he thinks is salmon pink lining in the bottom of the oesophagus and takes a sample? Well – he is very likely to diagnose Barrett’s oesophagus. The problem occurs if he has actually taken the sample from a hiatus hernia which he has mistaken for being the bottom of the oesophagus. This is rather easy to do.
I see patients all the time who have been given a diagnosis of Barrett’s oesophagus who actually turn out not to have it at all.
I will explain more on the subtleties of this issue in another post.