Abstract and Introduction
Abstract
Background Consensus guidelines aid in the diagnosis and management of autoimmune hepatitis, but they are frequently based on low-quality clinical evidence, conflicting experiences and divergent opinions. Recommendations may be weak, discrepant or non-existent at critical decision points.
Aims To identify the decision points where guidelines are weak or non-existent and review the evidence essential in the decision process.
Methods Full-text articles published in English using the keyword 'autoimmune hepatitis' were identified by PubMed from 1972 to 2013. Personal experience and investigations in autoimmune hepatitis also identified important contributions.
Results Seventy per cent of the guidelines developed by the American Association for the Study of Liver Diseases and 48% of those proposed by the British Society of Gastroenterology are based on low-quality evidence, conflicting experiences or divergent opinions. The key uncertainties in diagnosis relate to the timing of liver biopsy, recognising acute severe (fulminant) disease, interpreting coincidental nonclassical histological changes, accommodating atypical or deficient features in non-White patients, differentiating drug-induced from classical disease and identifying overlap syndromes. The key uncertainties in management relate to pre-treatment testing for thiopurine methyltransferase activity, treating asymptomatic mild disease, determining treatment end points, managing suboptimal responses, incorporating nonstandard medications as front-line and salvage agents, using azathioprine in pregnancy and instituting surveillance for hepatocellular carcinoma.
Conclusions Consensus guidelines are fraught with uncertainties in the diagnosis and management of autoimmune hepatitis. Each decision point must counterbalance the current available evidence and tailor the application of this evidence to the individual patient.
Introduction
Autoimmune hepatitis has codified diagnostic criteria and management strategies that have been promulgated as guidelines by the International Autoimmune Hepatitis Group (IAIHG), American Association for the Study of Liver Diseases (AASLD), and the British Society of Gastroenterology (BSG). These guidelines have been developed from experiences mainly in White North American and European patients. They define a classical phenotype that has been characterised by certain serological markers, especially antinuclear antibodies (ANA), smooth muscle antibodies (SMA) and antibodies to liver kidney microsome type 1 (anti-LKM1), hypergammaglobulinemia, increased serum immunoglobulin G (IgG) level, interface hepatitis on histological examination and responsiveness to corticosteroid therapy.
The guidelines have been useful in standardising the diagnosis and initial treatment of autoimmune hepatitis, but they have not encompassed all phenotypes, defined actions at every decision point, accommodated promising new treatments or applied uniformly to non-White populations. Furthermore, the current guidelines have been based on varying levels of clinical evidence, and most have been derived from small single-centre experiences and expert opinions rather than rigorously designed, randomised clinical trials. Only 11 randomised clinical trials have been performed in autoimmune hepatitis over the past 40 years, and most decisions in the diagnosis and management of this disease depend on the strength and consistency of clinical observations made in diverse referral centres.
Thirty-one of the 44 recommendations (70%) promulgated by the AASLD in 2010 are based on conflicting evidence or divergent opinion (Class II evidence), and the source of evidence used in developing 39 recommendations (89%) is from the consensus opinion of experts or case studies (Level C evidence) (Table 1). Similarly, 15 of the 31 recommendations (48%) proposed by the BSG in 2011 are based on low-quality clinical evidence (Class C evidence), and 12 recommendations (39%) developed from this evidence are considered weak (Level 2 evidence) (Table 1). Only the recommendation for corticosteroid therapy is based on high-quality randomised clinical trials (Class IA evidence) and strongly supported (Level 1 recommendation) by the IAIHG, AASLD and BSG.
Guidelines based on low-quality clinical evidence and divergent clinical opinions are not confident clinical directives, and they are likely to change as additional clinical experiences and hypothesis-driven investigations emerge. Highly individualised clinical judgments are required at decision points that are outside confident guidelines, and these judgments can be difficult and controversial. The goals of this review were to identify the clinical circumstances in which the evidence-based guidelines are weak, discrepant or non-existent and to provide the bases for decision-making in these difficult clinical situations.