Rationale and impact
These sections briefly explain why the committee made the recommendations and how they might affect practice.
People are usually offered an upper gastrointestinal endoscopy to detect oesophageal varices. If varices are found, they can be treated with either non-selective beta-blockers or endoscopic variceal band ligation. The 2023 update of this guideline recommends carvedilol or propranolol as potential options for preventing decompensation in people who have clinically significant portal hypertension (see the recommendations on primary prevention of decompensation). Therefore, the committee agreed that endoscopy is unnecessary for this group of people because these medicines prevent variceal bleeding. They amended the 2016 recommendations to clarify this.
Based on their knowledge and experience, the committee agreed that endoscopy presents an opportunity to carry out one‑off endoscopic variceal band ligation on medium or large varices at the same time. This is because the equipment and preparation are the same for both procedures. They noted that it may not always be possible to offer the endoscopic variceal band ligation during endoscopy because some healthcare professionals will not be trained, or have experience, in carrying out the procedure.
The additional resource impact of carrying out endoscopic variceal band ligation at the same time as endoscopy, where possible, is expected to be small.
Carvedilol and propranolol are commonly used in practice to prevent variceal bleeding. However, the committee noted that healthcare professionals often prefer to prescribe carvedilol because it has anti-alpha adrenergic vasodilatory effects in addition to beta-blockade and is also better tolerated. Both medicines need to be used with caution because they have a much greater effect on the heart rate and blood pressure of people with liver disease than in people who do not have liver disease. The committee further noted that the BNF states carvedilol should be used with caution in people with moderate hepatic impairment and avoided in those with severe hepatic impairment. Based their experience and expertise, the committee agreed that 'moderate' would apply to many people with cirrhosis. They were also aware of a growing body of research suggesting it should not be given to some people with large-volume or refractory ascites, so used this as an example of 'severe' hepatic impairment.
The committee noted that the BNF recommends a lower starting dose for propranolol for preventing variceal bleeding in people with cirrhosis, which the committee agreed to include in the recommendation. The same approach should also be taken with carvedilol; however, the committee were concerned that healthcare professionals may not be aware of this, because this is an off-label use of the medicine. Therefore, they recommended that, when used, carvedilol should also be started at a much lower dose than in people without liver disease and included an example of a suitable dosage in the recommendation to help guide healthcare professionals.
The committee noted that the use of non-selective beta-blockers to prevent decompensation is an emerging practice, but that they are already widely used for preventing variceal bleeding. Health economic analysis shows that the cost of prescribing these medicines is outweighed by savings in terms of preventing decompensation and reducing the number of endoscopic variceal band ligation sessions required.
The committee acknowledged that portal hypertension is not measured directly as a matter of routine in the UK, but agreed that clinically significant portal hypertension can be diagnosed on the basis of clinical features (for example, ascites or small varices) as well as through non-invasive methods, including tests to measure liver stiffness, or serum biomarkers.
Emerging evidence on the effectiveness of carvedilol and propranolol in preventing primary decompensation in people with clinically significant portal hypertension is unclear but looks promising. Therefore, the committee agreed that either medicine could be a potential treatment option. In the committee's view, carvedilol may be more effective in reducing pressure in the portal vein because it has anti-alpha adrenergic vasodilatory effects in addition to beta-blockade.
The committee made a recommendation for research on the effectiveness of non-selective beta-blockers in people with clinically significant portal hypertension that has been diagnosed through non-invasive investigations. They also noted that a large, National Institute for Health and Care Research-funded trial, underway in the UK, is looking at the use of carvedilol for this indication. They agreed this evidence could help inform future updates of this guideline.
The committee noted that prescribing non-selective beta-blockers for this indication is a new practice, and that they are not widely used in this way across the NHS. However, health economic analysis shows that the additional cost of prescribing these medicines is outweighed by the savings in terms of preventing decompensation.
Non-selective beta-blockers or endoscopic variceal band ligation can be used to prevent oesophageal variceal bleeding. The committee agreed that most people would prefer to take medication instead of undergoing an invasive procedure, especially as this would prevent further hospital visits. However, some will need endoscopic variceal band ligation instead, if they have had adverse reactions to non-selective beta-blockers, find it difficult to take tablets every day, or if the medicines do not work as expected. Therefore, the committee agreed it was important to discuss all treatment options, and talk with the person about their preferences and personal circumstances, to identify the right treatment for them.
The evidence did not distinguish between the clinical effectiveness of non-selective beta-blockers and endoscopic variceal band ligation. However, non-selective beta-blockers were found to be more cost effective when people were able to take these medicines regularly.
The studies did not look at the use of endoscopic variceal band ligation in conjunction with a non-selective beta-blocker to prevent variceal bleeding, which the committee agreed could have an additional benefit. They therefore made a recommendation for research to look at the effectiveness of combining both treatments.
In the committee's experience, both non-selective beta-blockers and endoscopic variceal band ligation are in common use for preventing oesophageal variceal bleeding. These recommendations will increase the use of non-selective beta-blockers to prevent variceal bleeding and should reduce both the use of endoscopic variceal band ligation and the costs associated with that procedure.
The committee looked at evidence that compared different types of antibiotics for preventing spontaneous bacterial peritonitis (SBP) in people with cirrhosis and ascites. They noted that 1 of the antibiotics in the evidence (norfloxacin) was not available in the UK, and that the fluoroquinolone class of antibiotics (which includes ciprofloxacin) was the subject of a 2019 MHRA drug safety update that includes restrictions and precautions for their use.
The evidence did not demonstrate any overall benefit of antibiotic prophylaxis when compared with no prophylactic treatment. Therefore, the committee agreed that antibiotics should not be routinely offered to prevent SBP in people with cirrhosis and ascites, unless they were at high risk of developing the infection, or its impact could be severe. Most of the evidence focused on people who had low ascitic protein count, so this was included as an example of a measure that could be used to identify risk factors. The committee also included other measures for assessing the severity of liver disease.
The committee agreed that, because the evidence showed no overall benefit of using 1 antibiotic over another, prescribers should follow local microbiological advice. This would mean local pathogen distribution would be taken into account and would lead to better antimicrobial stewardship. The committee also noted the importance of good antimicrobial stewardship to reduce the potential emergence of drug-resistant bacteria.
The committee agreed that the evidence was not methodologically robust, the sample sizes were small, and none of the studies looked at quality-of-life data. Therefore, they made a recommendation for research on antibiotic prophylaxis to prevent SBP to encourage studies that are more rigorous and look at clinically important outcomes such as mortality, health-related quality of life, and serious adverse events.
In the committee's experience, there is wide variation in practice in the prescribing of antibiotics for preventing SBP. The 2023 recommendations suggest using antibiotics for people at high risk, but do not specify which antibiotics. This means that antibiotic choice can be based on local microbiological advice and funding agreements, which may reduce the costs of prescribing antibiotics where necessary. It also makes clear that antibiotics are not necessary for everyone with cirrhosis and ascites.