Assessment and management of cirrhosis in people older than 16 years: summary of NICE guidanceBMJ 2016; 354 doi: https://doi.org/10.1136/bmj.i2850 (Published 06 July 2016) Cite this as: BMJ 2016;354:i2850
All rapid responses
The scope, which defined what this guideline would (and would not) examine, was scrutinised comprehensively by stakeholders(a) at a meeting with the developers and via written submissions. This appraisal concluded the guideline would not consider the diagnosis, investigation and management of the diseases causing cirrhosis (NICE guidelines either published or in development, cover these areas(1-3)) and would be applicable to all causes of cirrhosis, including autoimmune liver diseases, as documented on page 45 of the guideline.
Since cirrhosis is not always clinically apparent, the guideline development group (GDG) desired a strategy to increase the recognition of cirrhosis. Therefore, we looked at the performance of numerous tests, both individually and in combination, to diagnose cirrhosis in people with risk factors for liver disease. A lifetime Markov decision model was constructed depicting the natural history of a patient diagnosed with compensated cirrhosis. Model inputs were populated primarily from published data sources and were subjected to extensive probabilistic and deterministic analysis to explore the combined parameter uncertainty. The GDG acknowledged some variation in the cost-effectiveness of the diagnostic tests across the different aetiologies: either transient elastography or ARFI could be used for people with NAFLD with advanced liver disease, and transient elastography or liver biopsy for patients with hepatitis C. Although no testing was an option for people with alcohol-related liver disease, the ICER for transient elastography compared to no testing was £22,438 per QALY gained and confidence limits around this base case showed testing could not be above £30,000 per QALY. Therefore, the GDG concluded that testing is the appropriate strategy for people with alcohol-related liver disease or those at high risk of cirrhosis from drinking. There is no universally agreed definition of high risk but the GDG adopted the criterion of people exceeding 50 units per week for men or 35 units for women.
The GDG noted the practicality of recommending a standard non-invasive test for cirrhosis for all aetiologies. For people with hepatitis B, CG165(2) recommends transient elastography and there is good clinical and economic evidence to support transient elastography as a cost-effective option in other causes of liver disease. The GDG did not find evidence to support the use of fibrosis blood tests across all aetiologies and in people with hepatitis C these tests were amongst the least cost-effective options. The early recognition of cirrhosis is cost effective since it triggers a more intensive clinical path that allows for more timely intervention and mitigates against the all too familiar occurrence of a person with previously unrecognised cirrhosis presenting to medical care with one of the major complications such as bleeding varices or advanced hepatocellular carcinoma. We did not consider the performance of the non-invasive tests as screening tools in the general population. A tool to better identify people at risk of cirrhosis in primary care is desirable and the GDG recommended this be an area of future research.
Doctors Rowe and Hirschfield question the recommendation that all people with cirrhosis should have endoscopy to look for varices, although this approach is advocated by the American College of Gastroenterology and the American Association for the Study of Liver Disease(4). The health-economic evidence identified that endoscopic surveillance every 3 years was cost-effective in people with cirrhosis due to NAFLD, ALD or hepatitis B at a threshold of £20,000 per QALY gained; whereas, 2-yearly surveillance was cost-effective in people with hepatitis C. This recommendation identifies people who might benefit from measures to decrease the risk of variceal haemorrhage. Bleeding is a significant event because each episode is associated with a 10–30% mortality risk. Consequently, the GDG examined the clinical and cost- effectiveness of non-selective beta-blockers compared to endoscopic variceal ligation (EVL) for the primary prevention of bleeding in patients with oesophageal varices due to cirrhosis. Data from a Cochrane systematic review and 2 additional randomised trials were included in the meta-analysis. The Cochrane review included 7 conference abstracts but their data were re-examined individually to extract the additional outcomes: survival and freedom from variceal bleeding as a time-to-event outcomes and hospital admissions. One study (Lo 2004) of nadolol in the Cochrane review was not included, as this drug is not used in the UK. The meta-analysis in people with medium or large varices found clinically significant benefits of EVL compared to non-selective beta-blockers for variceal bleeding (RR 0.44; 95% CI 0.27 to 0.71), upper GI bleeding from any cause (RR 0.71; 95% CI 0.54 to 0.92) and lethargy (OR 0.09; 95% CI 0.04 to 0.22). Using rates of all-cause mortality and variceal bleeding from the meta-analysis, with standard NHS costs for EVL, propranolol and treatment of variceal bleeding, the ICER for prophylactic EVL compared to beta-blockers was £14,641 per QALY gained. This does not include the additional financial benefit of preventing hospitalisation (161 fewer admissions per 1,000 people with EVL compared to beta-blockers) as the duration and therefore the cost of hospital stay in the studies was not known. The GDG concluded that EVL is clinically and cost-effective compared to beta-blockers for people with medium or large varices and we refute the suggestion that the evidence is weak. In people with cirrhosis and small oesophageal varices, the GDG did not find evidence to support prophylaxis based on the current evidence and made a research recommendation.
In summary, the guideline’s recommendations were based on the best available research evidence relating to both clinical and cost effectiveness. When there was no evidence to endorse a treatment or the evidence indicated that the treatment was not clinically or cost effective, the GDG made a clear statement to this effect and made recommendations to guide future research. Detailed summaries of the GDG’s discussions are in the “linking evidence to recommendations” section of each chapter. A draft guideline containing the recommendations and methodology was sent out for consultation and, following peer review by multiple stakeholders(a), the revised final guideline was published. Consequently, we believe the cirrhosis guideline’s recommendations are based on good quality evidence and we reject the criticisms made by Doctors Rowe and Hirschfield.
(a) Stakeholders included:
Royal College of Pathologists, Royal College of Physicians, Royal College of General Practitioners, Royal College of Nursing, Royal College of Radiologists, Welsh Association of Gastroenterology and Endoscopy, British Liver Trust, British Medical Association, British Association for the Study of the Liver, British Society of Gastroenterology, NHS England, The British Society for Gastrointestinal and Abdominal Radiology.
1 National Institute for Health and Care Excellence. Alcohol-use disorders: diagnosis and management of physical complications. NICE clinical guideline CG100. London. National Clinical Guideline Centre, 2010. Available from: https://www.nice.org.uk/guidance/cg100
2 National Institute for Health and Care Excellence. Hepatitis B (chronic) : diagnosis and management . NICE clinical guideline CG165. London. National Clinical Guideline Centre, 2013.
Available from: https://www.nice.org.uk/Guidance/CG165
3 National Institute for Health and Care Excellence. Non-alcoholic fatty liver disease (NAFLD): assessment and management. NICE clinical guideline NG49. London. National Clinical Guideline Centre, 2016. Available from: https://www.nice.org.uk/guidance/ng49
4 Garcia-Tsao G, Sanyal AJ, Grace ND, Carey WD. Prevention and management of gastroesophageal varices and variceal hemorrhage in cirrhosis. Am J Gastroenterol 2007;102:2086-2102
Competing interests: No competing interests
We read with interest the summary of the NICE guidance on the assessment and management of cirrhosis. The rising prevalence of cirrhosis has been recognised as of great concern.1 Efforts to increase the identification of patients with significant liver disease are seen as vital so that those individuals can receive treatment of the underlying cause of liver disease as well as for any complications of cirrhosis. The causes of liver disease are diverse and disappointingly this in itself does not appear to be adequately addressed in the NICE guidance produced.2, 3 Patients with common and rare liver disease merit equal care, and the absence of mention of autoimmune liver disease, for example, which has a disproportionate burden on patients, is telling of a guideline that fails to be representative and thus immediately in need of rebuttal.
Strategies to increase the diagnosis of cirrhosis forms a major part of the guidance produced. Although numerous tests have been developed the guidance recommends the use of transient elastography or acoustic radiation force impulse imaging. These technologies are not routinely available in primary care and the cost and service implications of using these rather than validated combinations of biochemical parameters in that setting do not seem clearly thought through. Indeed, the infrastructure and capital costs are not considered in the cost-effectiveness models and the associated resource impact summary significantly underestimates the extent of testing required. The published estimate of hazardous drinking in the UK adult population is 7%, that is in excess of 3 million individuals who would be potentially offered testing for cirrhosis according to this guidance.4 Accepting only a minority of these will be identified, testing just 10% of those drinking alcohol in a hazardous fashion would place a significant burden on already overstretched primary care infrastructure with the potential to overwhelm secondary care liver services without clear evidence that this would improve outcomes in this patient group.
Furthermore there are other more defined areas within the guideline that raise cause for significant concern. It is striking that no high quality evidence exists for any of the recommendations and several recommendations are made based exclusively on the experience and opinion of the guideline development group (GDG). It is recognised that the quality of evidence supporting recommendations in international practice guidelines for patients with liver disease is most often of low quality and this is a major issue for a disease that is the third most common cause of death in the UK.5 Nevertheless it is important to clearly demarcate the gaps in evidence in order to drive appropriate studies for the future, which can contribute important, but currently missing, evidence for truly informed practice.
Of note the guideline contains several controversial and highly debatable recommendations. For example, monitoring of patients with cirrhosis for the development of oesophageal varices is central in the ongoing management of these individuals. The recommendation for the use of endoscopic variceal banding ligation as primary prevention of variceal haemorrhage is in contrast to all other guidelines in the field, including recent specialty based guidelines from the British Society of Gastroenterology.6 The alternative treatment is with non-selective beta-blockers and there are a number of studies comparing these treatments. In addition to the NICE evidence review there has also been a Cochrane review of this topic.7 Both reviews find that there is no difference in overall mortality between banding ligation and beta blocker treatment. A reduction in variceal bleeding however is reported with banding ligation when all studies are included but in the Cochrane review this effect is lost when studies either published in abstract form or with high risk of attrition bias are excluded questioning this advantage of banding ligation. The recommendation for banding ligation in the NICE guideline is supported by an economic model that includes a point estimate of decreased mortality as well as the reduction in variceal bleeding. It is arguable whether this economic evaluation is valid given that the decrease in mortality is assessed as very low quality evidence due to risk of bias and is not statistically significant. In addition, the question of the harms of treatment are partially addressed in the guideline where only fatigue with beta blockers is reported. The adverse events associated with banding ligation are not reported and these include clinically important bleeding, retrosternal pain, and there are also reports of mortality with banding ligation.7 As a consequence of each of these factors the recommendation for banding ligation as primary prevention is questionable and an alternative view would be to involve the patient in decision making regarding prevention of variceal haemorrhage with balanced information regarding both banding ligation and treatment with beta blockers.
The purpose of the NICE clinical guidelines is to provide evidence-based recommendations to support effective, good value care. The recommendations made in this NICE guideline and summarised in the BMJ indicate a degree of certainty that is not supported by the available evidence. Whilst the incorporation of measures of cost-effectiveness is potentially valuable their use where evidence is weak and when models incorporate erroneous parameters increases the risk that incorrect and potentially misleading recommendations are made. It is critical that gaps in the evidence base for patients with cirrhosis are addressed in high quality prospective randomised studies such that outcomes for this growing patient group might be improved. Implementation of this current guidance risks changes to services that are neither more effective nor higher value to patients.
1. Williams R, Aspinall R, Bellis M, et al. Addressing liver disease in the UK: a blueprint for attaining excellence in health care and reducing premature mortality from lifestyle issues of excess consumption of alcohol, obesity, and viral hepatitis. Lancet 2014;384:1953-97.
2. NICE. Clinical Guideline 50: Cirrhosis in over 16s: Assessment and management. 2016.
3. Harrison P, Hogan BJ, Floros L, et al. Assessment and management of cirrhosis in people older than 16 years: summary of NICE guidance. BMJ 2016;354:i2850.
4. Robinson S, Bulger C. General Lifestyle Survey 2008: Smoking and Drinking Among Adults, 2008. Newport: Office for National Statistics, 2010.
5. Rowe IA, Parker R, Armstrong MJ, et al. Assessment of the quality of evidence underlying international guidelines in liver disease. Am J Gastroenterol 2012;107:1276-82.
6. Tripathi D, Stanley AJ, Hayes PC, et al. U.K. guidelines on the management of variceal haemorrhage in cirrhotic patients. Gut 2015;64:1680-704.
7. Gluud LL, Krag A. Banding ligation versus beta-blockers for primary prevention in oesophageal varices in adults. Cochrane Database Syst Rev 2012:CD004544.
Competing interests: Ian A Rowe has no competing interests to declare. Gideon Hirschfield has received honorariums as a speaker at education events for Falk Pharmaceuticals and for attendance at an advisory board from Intercept Pharmaceuticals in the area of immune mediated liver disease. He is also an investigator on a number of clinical trials in immune mediated liver disease.
Distorted view of Ultrasound in cirrhosis.
I was not sure whether I should be sad or upset at the summary dismissal of the role of ultrasound by the learned authors stating that ‘ultrasound examinations are not sensitive’ with no reservations, not even in the case of well established cirrhosis when diagnosis from the echo pattern, surface and internal nodularity, evaluation of liver compliance and evidence of portal hypertension can be made with considerable ease. In an article that is meant to provide guidance for hospital based practice and general practice they declare ‘non-invasive transient elastography or acoustic radiation force impulse imaging is the first line investigation’. In this article the only place for ultrasound is in the six monthly surveillance for hepatocellular carcinoma.
The authors are not entirely to blame for their belief. Despite the wide spread use of ultrasound by GPs and in hospital practise there seem to be some radiologists who spread the heresy that there are certain fields that are taboo and ultrasound is of little or no value. These include chronic liver disease and bowel ultrasound. Regrettably there are teaching institutions where this is imparted to the trainee registrars. It is my conviction that radiologists are only as good as the demands placed on them by the clinicians. Failure of the radiologists to provide the service results in the clinicians having recourse to CT scans or other investigations for the evaluation of cirrhosis. Radiologists have very little training in pathology and the appreciation of images in relation to pathological changes and histology is not their forte. Perhaps the Royal College of Radiologists could help. It is rather strange that a radiologist is taught more physics to appreciate the physics behind image production than the pathology that gives rise to the image.
A rapid response is not a forum to justify the value of ultrasound in cirrhosis. This article however takes me back to the eighties when doubts were expressed and ultrasound evaluation of chronic liver disease was in its infancy. In a joint paper submitted to the BMJ by the departments of medicine and radiology, St. Georges Hospital, in 48 patients with histologically established steatosis ultrasound identified steatosis in 45 patients. Ultrasound was less useful in the detection of fibrosis. In 35 patients ranging from increased fibrosis to established cirrhosis fibrosis was detected with 57% sensitivity and 88% specificity (1). The response of the then editor Richard Smith was that the better results were ‘because your radiologist is so good’. This belief in observer dependence is one of the factors that has plagued ultrasound and is still widely held. Following the review of the images from which the diagnosis was made the editor proceeded to publish the paper.
Another editorial commissioned by Clinical Radiology on the assessment of parenchymal liver disease with ultrasound (2), prompted the following comment from one of the referees. ‘Within the international community the author is generally recognised as being in the minority of one in his opinion for the various features seen in chronic liver disease’. Many of us were beginning to believe that the recognition of the value of ultrasound had moved on. I doubt if this article could have turned the clock back. If nothing else there would be very few clinicians who would express any doubt about the ability of ultrasound to detect fatty change and fibrosis with established cirrhosis. I have seen the progress of ultrasound over the last four decades and I hope that articles such as these could not detract from the value of ultrasound.
It is heartening to see the other response to this article also expressing the value of ultrasound and it is my hope that clinicians and radiologists who have greater belief in the value of ultrasound would also express their opinion.
1. Ultrasound scanning in the detection of hepatic fibrosis and steatosis. S.H. Saverymuttu, A.E.A.Joseph and J.D.Maxwell. BMJ (1986) 292, 13-15
2. Ultrasound in the assessment of diffuse parenchymal liver disease A.E.A.Joseph Clinical Radiology (1991) 44 219-221
Competing interests: No competing interests
Liver cirrhosis is becoming one of the major killers and is emerging as a serious public health issue. The article brings to attention the means of assessment and effective management of cirrhosis(1). Recent statistics published by the chief medical officer for England state liver diseases are on the rise in England whereas they are decreasing among other European countries (2). High alcohol consumption, obesity and chronic viral hepatitis B and C are leading contributors to liver cirrhosis worldwide(3).
Fibroscan is still widely used for the early diagnosis of cirrhosis, although other modern non-invasive tools such as acoustic radiation force impulse (ARFI), real-time elastography (RT-E), magnetic resonance elastography (MRE), and shear wave elastography (SWE) are being used as well.(4) Absence of contraindications, easy manageability, time and cost effectiveness and higher safety make fibroscan more acceptable than previously predominantly used liver biopsy. However, readings are unreliable for people presenting conditions such as obesity, ascites, certain metabolic syndromes (type 2 diabetes, increased waist circumference), old age, active hepatitis, biliary obstruction and mass lesions within the liver. Elastography does not replace either conventional ultrasound or liver biopsy. While ultrasound examination provides a panoramic view of the structural integrity of the liver, and features of portal hypertension, liver biopsy is essential for the staging of cirrhosis. Liver biopsy is associated with bleeding, but is rare, recording to only around 0.35%(5).
Apart from patients with refractory ascites, transjugular intrahepatic portosystemic shunt (TIPS) should be also offered to patients suffering from esophageal bleeding, hepatorenal syndrome and portal vein thrombosis(6) as these conditions are also common complications of cirrhosis.
Cirrhotic patients with hepatocellular carcinoma, portopulmonary hypertension should perhaps be considered as exceptional groups (7) as their calculated MELD score doesn’t represent the real severity of the condition and can change unexpectedly over a short period time.
1. P Harrison, BJ Hogan, E Davies “Assessment and management of cirrhosis in people older than 16 years:” summary of NICE guidance .BMJ 2016; 354 doi: http://dx.doi.org/10.1136/bmj.i2850 (Published 06 July 2016) Cite this as: BMJ 2016;354:i2850
2. Department of Health. Chief medical officer publishes her first annual report. 2012. www.dh.gov.uk/health/2012/11/cmo/
3. Rohit Loomba, Hwai-I Yang, Jun Su et al “Synergism Between Obesity and Alcohol in Increasing the Risk of Hepatocellular Carcinoma: A Prospective Cohort Study”. Am. J. Epidemiol. (2013) 177(4):333-342
4. Australian Family Physician. The right upper quadrant, July 2013 “Fibroscan and transient elastography”. Volume 42, No7, July 2013 Pages 468-471
5. Dotan Y, Carlebach M, Zuckerman E, Maruf M, Schiff E “Delayed bleeding after percutaneous liver biopsy.” EJCRIM 2016;3:doi:10.12890/2016_000359
6. Jason T Sasamendi, FranciscoJ Gortes, Michelle Shnayder et al “Transsplenic portal vein reconstruction–transjugular intrahepatic portosystemic shunt in a patient with portal and splenic vein thrombosis”. Radiology case report, ELSEVIR. Available online 21 June 2016
7. Saleh Elwir, John Lake “Current Status of Liver Allocation in the United States” Gastroenterology & Hepatology March 2016, Volume 12, Issue 3
Competing interests: No competing interests
Could changes in Transient Elastography (TE) measurements in cirrhosis be attributed only to progression of fibrosis?
The article provides the outline of non invasive methods of assessing liver fibrosis. The imaging methods estimate liver stiffness ‘as a surrogate’ of fibrosis. The article does not pay much attention to the acoustic pattern often found during a routine ultrasound scan of the liver such as surface nodularity, observable especially in the presence of ascites, the liver pattern resulting from the presence of regenerating nodules and the morphological changes to the blood vessels.
Based on my experience I wish to describe a very simple method of assessment of liver ‘compliance’ as a reflection of liver stiffness during a routine ultrasound scan of the liver. Compliance of the liver varies depending on the pathological process. Fibrosis reduces liver compliance and could be visually estimated to complement the other features produced by fibrosis.
During a left subchondral scan performed the heart can be readily seen. With appropriate positioning of the probe it is possible to interpose the left lobe of the liver between the probe and the heart.
Cardiac pulsations apply a compression force on the liver surface which is transmitted into the liver causing alternate compression and relaxation of the liver tissue. This effect is maximal near the liver surface fading with distance into the liver. The response of the liver is a reflection of liver compliance could be observed with relative ease. The drawback, however, is the inability to assess this quantitatively. It is, however, of significant value in complementing the other pattern of changes that are observed in cirrhosis.
Assessment of liver stiffness or fibrosis with TE has been presented as a surrogate measure of fibrosis. There is hardly any literature on the effect of steatosis on measurements obtained with TE. It has long been recognised that steatosis can be recognised with relative ease on a standard ultrasound scan. The drawback again is that cannot be reliably quantified. Before the introduction of multi frequency probes, with narrow range frequencies with a predominant frequency such as 3Mz or 5Mz frequencies attenuation of the ultrasound beam was well recognised. Attenuation increased with frequency. Controlled attenuation parameter has more recently been introduced as a measure of steatosis.
Diminishing degree of compliance between fatty liver, normal liver and fibrotic liver can be observed at post mortem. There do not appear to be studies relating TE measurements with compliance. It is my educated guess that propagation of sheer waves would be related to compliance. In patients with cirrhosis particularly those with alcoholic liver disease there is co-existing steatosis and fibrosis. Disappearance of steatosis is, however, observed with advancing fibrosis. Fatty change may also be noted in patients with Hepatitis C. It is therefore reasonable to try and understand the influence of fatty change on TE before attributing changes entirely to fibrosis.
It is interesting that this would not be the first time that the influence of steatosis and fibrosis on a physical parameter has not been recognised. Physicists unaware of the histological findings in cirrhosis attributed attenuation of the beam to fibrosis on the assumption that histological change in cirrhosis was entirely that of fibrosis. It was also assumed that there was no attenuation due to steatosis. In a study done at St. Georges Hospital, published in the BMJ, we showed that attenuation was a predominant feature of steatosis and not fibrosis. We also showed that the attenuation of the beam observed in many cirrhotic livers was due to the co-existing steatosis.
I therefore believe the high reliability claimed for the detection of fibrosis and progression of fibrosis should be interpreted in the light of the possible influence of steatosis which may often be present in cirrhotic livers. The relationship between compliance and shear wave propagation should be studied particularly in view of the increased compliance associated with steatosis.
Competing interests: No competing interests