August 2001
D R Gaya
Department of Gastroenterology
Gartnavel General Hospital
Great Western Road
Glasgow
Corresponding author: D R Gaya:
e mail: : danielgaya@aol.com
Since Paul Ehrlich’s first percutaneous liver biopsy in 1883 in Germany, the technique has been modified greatly. It is now a well established diagnostic procedure and is accepted as the most specific test to assess the nature and severity of many liver diseases.
Liver biopsy can now be done via the percutaneous, transjugular, laparoscopic or open routes. The practice throughout the UK varies greatly between centres and thus the British Society of Gastroenterology (BSG) produced guidelines in 19991 in an attempt to standardise current practice in the light of the best available evidence. The guidelines themselves are easy to understand and implement and are a fair representation of the current available evidence. In addition they will hopefully reduce the large variation in current practice throughout the UK and stimulate the audit cycle in this field. However, one must be under no illusions that the guidelines themselves are based on largely weak evidence and many of the recommendations are based on arbitrary cut?off values. This is by no means a criticism of the researching physicians, more that there is a dearth of good evidence in this field; as morbidity and mortality rates related to liver biopsy are so small, large numbers of biopsies have to be included in studies in order to see statistically significant results. As a general medical SHO, I was asked to see a man one hour post ultrasound guided liver biopsy. He was in pain and clinically shocked; I immediately commenced intravenous fluid resuscitation. I must site ignorance on my part for being falsely reassured that this gentleman was unlikely to have had a significant intra?abdominal bleed as his biopsy was under imaging guidance. He actually dropped his haemoglobin five grams and almost bled to death due to the procedure. So does ultrasound guidance decrease the rate of bleeding after liver biopsy?
Definitely not is the answer. Indeed in the case of focal liver lesions (esp. neoplastic deposits) ultrasound guidance is likely to result in a higher pathological yield but at the cost of a higher rate of haemorrhage as the tissue in question is almost invariably more vascular than the surrounding hepatic parenchyma.
McGill’s series from the Mayo Clinic in 19902 retrospectively reviewed 9500 liver biopsies in an attempt to identify risk factors for haemorrhage post biopsy over a 21 year period. The only significant positive correlation with haemorrhagic risk was presence of malignancy in the liver. That is to say, platelet count, coagulation profile and imaging guidance did not appear to influence the risk.
An audit by Gilmore et al 3 which formed much of the foundation of the BSG guidelines,1 confirmed the notion that haemorrhage post liver biopsy is largely unpredictable. For example 90% of bleeds occurred with an INR of <1.3 and hence having a normal INR is no reassurance that one will not bleed post biopsy (yet having a high INR is a contraindication to having the procedure itself).
In 1991, at the time of the audit by Gilmore,3 one third of all liver biopsies were done under ultrasound guidance. That figure is likely to be nearer half by now. In the discussion section of this paper, it is said that “. . . it may become increasingly difficult to justify blind needle insertion when ultrasonography is so widely available”. However, as Vautier4 pointed out in 1994, there is no evidence to suggest that ultrasound guidance increases the safety of the procedure. Also, it only increases the diagnostic yield if a focal lesion needs to be biopsied. The latter is reflected in the BSG1 guidelines which do not stipulate that ultrasound guidance is necessary or even recommended at the time of biopsy. Only that “. . . some form of imaging of the liver within the preceding four weeks” be done, “in order to detect anatomical variations and any focal anomalies needing ultrasound guidance”
Intuitively, the risk of pneumothorax and perforation of the gall bladder will be decreased by ultrasound guidance, although there is no evidence available from randomised controlled trials to confirm this. If I were to have a liver biopsy, I would want it under ultrasound guidance if only to reduce the risk of the latter two unfortunate complications from arising. I would however do this in the full knowledge that my haemorrhagic risk would be essentially unchanged, if not increased.
In summary, the risk of haemorrhage post liver biopsy is related not to the presence or absence of imaging guidance, but to the actual process of penetrating vascular hepatic parenchyma with a needle. This fact should be understood by doctors and patients alike.