Search

Elevations in the liver enzyme alanine transaminase (ALT), traditionally regarded as a highly specific indicator of liver injury, are now recognised as a novel marker of impaired metabolic health. Elevated ALT has a positive, linear association with the metabolic syndrome and its individual components including truncal adiposity, type 2 diabetes and dyslipidaemia. Prospective data indicates that elevated ALT predicts incident metabolic syndrome and type 2 diabetes in a dose dependent relationship, and may predict incident cardiovascular mortality, although the latter relationship appears non-linear. Similarly, elevations in the liver enzyme gamma-glutamyltransferase (GGT), may also represent impaired metabolic health, but used alone is less specific than ALT. Despite these robust data, the prevalence of elevated ALT and associated mortality outcomes for people with elevated liver enzymes in the Australian population is largely unstudied. ALT is also the enzyme most strongly correlated with hepatic fat accumulation, termed non-alcoholic fatty liver disease (NAFLD), itself frequently referred to as “the hepatic manifestation of the metabolic syndrome”. NAFLD is an important public health problem in developed countries, where it affects up to 30% of people, and 2-4% may progress to liver cirrhosis, hepatocellular carcinoma or liver transplantation. Longitudinal studies indicate that the most important risk factor for disease progression is advanced fibrosis identified at liver biopsy, however it is not feasible to perform biopsy at a widespread level. Therefore, non-invasive scoring systems have been developed and have good accuracy in high risk populations, but their diagnostic accuracy in a population with a low prevalence of advanced fibrosis has not been studied. Furthermore, to effectively treat people with advanced fibrosis and reduce disease progression, interventions that are supported by high Abstract 9 quality evidence on the benefits and harms should be instituted. However, many advocated lifestyle interventions in NAFLD are based on observational data alone, which is subject to selection bias and confounders, and evidence based management is further compromised by the conspicuous lack of information provided on the harms of interventions. In this body of work, a range of epidemiological techniques were used to explore these evidence gaps in the identification, treatment and counselling of people at high risk of metabolic liver disease. In Chapter 2, a cross sectional study established the prevalence of elevated ALT in the Australian population. In Chapter 3, these data were extended by a longitudinal, prospective cohort study to determine if there was an excess risk of cardiovascular and all-cause mortality for individuals with elevations in ALT and GGT. In Chapter 4, a diagnostic test study assessed the accuracy of non-invasive fibrosis scoring systems for detection of people with advanced fibrosis in the general population. To then support evidence informed decision making in NAFLD, Chapter 5 explores the evidence base for standard lifestyle recommendations in NAFLD, and in Chapter 6, a methodological study assessing the balance of reporting of benefits and harms in systematic reviews was undertaken.Prevalence of elevated alanine transaminase (ALT) in Australia and its relationship to metabolic risk factors: a cross sectional study of 9,447 people. Elevated alanine transaminase (ALT) is a strong and reproducible marker of metabolic syndrome, and may help to identify individuals at risk of future metabolic disease, but there is little data from the Australian population. We aimed to determine the prevalence of elevated ALT and its association with metabolic risk factors in the general population. We utilised data from a large, population based cross sectional study (N=9,447) from a nationwide survey. We used standard thresholds [defined as ≥ 40 IU/L (males) and ≥ 30 IU/L (females) and lower thresholds (ALT as ≥ 30 IU/L for males and ≥ 19 IU/L for females), and evaluated for independent predictors of elevated ALT in logistic regression models. Analyses were weighted using population weights. Using standard thresholds, elevated ALT levels were found in 11.2% of the Australian population overall. People with elevated ALT were younger (43 vs 46 years) with more truncal adiposity (100 vs 91 cm), higher levels of pro-atherogenic lipids and glucose and exercised less (120 vs 160 minutes per week, p51 and ALT >40 IU/L, and GGT >33 and ALT >31 IU/L for females), and we included people with elevations in both enzymes to increase detection of those most at risk of metabolic liver disease. Adjusted Cox proportional hazards models assessed the association of elevated enzymes and mortality with long duration follow up. Over a median follow up of 10 years (20,145 person years), 701 people died, including 203 (34%) from cardiovascular disease. Cox regression models adjusted for sex, age, smoking and alcohol intake indicated that people with elevated enzymes had an increased risk of all-cause mortality that was modified by age (test for interaction p=0.01). Age-stratified analyses demonstrated no increased risk at younger ages (age ≤ 59 years, H.R. 0.46 95% confidence interval 0.06-3.49), but increased risk with age; age 60-69, H.R. 1.05 (0.53-2.07), age 70-79 years, H.R. 1.54 (0.81-2.93), and age ≥ 80 years, H.R. 3.53 (1.55-8.04). Similarly, the risk of cardiovascular mortality with elevated enzymes was also modified by, and increased with age (test for interaction p=0.02); age 70-79, H.R. 3.15 (1.37-7.23), age ≥ 80, H.R. 6.86 (2.44-19.30). These data suggest that in community dwelling elderly persons, an elevation in ALT and GGT are associated with an excess risk of all-cause and cardiovascular mortality which increases with age. Identification of people with NAFLD and advanced fibrosis from within the general population. Despite the high prevalence of NAFLD within developed countries, only a small proportion will progress to liver cirrhosis and failure, particularly those with advanced fibrosis. Fibrosis is most accurately detected using liver biopsy or transient elastography, however these tests are not widely available. Therefore, non-invasive fibrosis scoring systems which use a combination of available variables such as ALT, BMI and platelet count are increasingly used. The diagnostic accuracy of non-invasive scores was established in tertiary liver clinics where the prevalence of advanced fibrosis is high, at up to 30%, and whether they are equally accurate in a low prevalence population has not been studied. We used data from a population based, cross-sectional study, where participants underwent clinical assessment, fasting blood tests and transient elastography as the reference standard for advanced fibrosis. We assessed test performance characteristics of three scores, the NAFLD fibrosis score, FIB-4 and Aspartate Transaminase to Platelet Ratio Index (APRI), using standard thresholds. To calculate diagnostic test characteristics, we constructed a 2 by 2 table with the presence or absence of advanced fibrosis according to the transient elastography reading against the presence or absence of advanced fibrosis according to the scoring systems. Area under the receiver operating curve was calculated to assess overall diagnostic accuracy. The response rate was 36%. After exclusions, 922 underwent transient elastography, of which 749 had a valid reading and fifteen had advanced fibrosis (2%). The specificity of non-invasive scores for advanced fibrosis approximated 100% (95% CI 99-100) with a negative predictive value of 98% (95% CI 97-99) for all systems. The sensitivities were low, at 7% (95% CI 0-32) to 13% (95% CI 2-40). Positive predictive values ranged from 50% (95% CI 7-93) to 67% (9-99). Negative likelihood ratios approached one (0.87, 95% CI 0.71-1.06 to 0.93, 95% CI 0.82-1.07), and positive likelihood ratios were uninformative due to the small number of people with positive scores. These data suggest that when non-invasive fibrosis scores are used in the general population, their specificity and negative predictive values are very high and can reliably exclude those without advanced fibrosis without need for further testing, with consequent cost savings and avoidance of unnecessary procedures. However positive results are most likely to be false positives, and should prompt further testing. Lifestyle recommendations for people with NAFLD- which have the highest level of evidence? Non-alcoholic fatty liver disease (NAFLD) is considered the hepatic manifestation of the metabolic syndrome as it is tightly linked with the global obesity epidemic. Indeed, some propose that NAFLD may even precede development of the metabolic syndrome, although confirmatory studies are lacking. Pharmacological therapies for NAFLD including Vitamin E or pioglitazone may improve surrogate outcomes such as fibrosis, although there are no data on long term clinical outcomes, and concerns remain regarding long term safety. Thus lifestyle modification with diet and exercise counselling remains standard of care, but commonly recommended interventions such as the Mediterranean diet, omega 3 polyunsaturated fatty acids supplements or increased intake of monounsaturated fats are Abstract 14 largely derived from observational data with low validity. We conducted a review to evaluate which diet and exercise recommendations were underpinned by a high level of evidence, in order to prioritise these recommendations for clinical use. We found that no trials have been conducted with clinical endpoints. Trials of the Mediterranean diet and omega 3 polyunsaturated fatty acids show improvement in surrogate outcomes only, such as liver fat. Improvement in surrogate outcomes is also seen for aerobic, anaerobic and combined modality exercise, and a reduction in sedentary time also appears important. Overall, however, there remains a substantial lack of high quality evidence to support lifestyle recommendations. To provide a sound evidence base for NAFLD management, longer duration trials of standardised lifestyle regimes, and inclusion of robust clinical endpoints are essential. Quality of reporting of systematic reviews within gastroenterology- is there balanced reporting of benefits and harms? Systematic reviews are an integral component of evidence-based health care, however reviews tend to focus on purported benefits of therapies without due consideration to potential harms. Information on both outcomes is needed to adequately inform patients. We assessed the reporting of harms in systematic reviews relevant to clinical gastroenterology. We identified all systematic reviews of randomised trials of gastroenterology interventions published from 2008 to 2012 in highly cited gastroenterology and general medical journals. We adapted the Consolidated Standards of Reporting Trials (CONSORT) guidelines for harms and assessed qualitative and quantitative parameters of harms reporting. Regression analyses determined predictors of more Abstract 15 comprehensive harms reporting. In total, 78 systematic reviews were identified, with 72 published in gastroenterology journals and six in general medical journals. Overall, one in three systematic reviews (26/78, 33%) did not refer to harms of the intervention anywhere in the article. Less than half of the studies included adverse events as an outcome measure, and data on absolute rates of adverse events were only provided in 28%. Most (65%) did not include any figures or tables on adverse event; however, all included these on efficacy outcomes with an average of three tables and figures per study. Regression analyses indicated that the use of reporting guidelines was significantly associated with better harms reporting (P = 0.04). These findings suggest that the reporting of harms in systematic reviews pertinent to gastroenterology is largely inadequate and highly asymmetrical compared with the reporting of benefits. We suggest that systematic review authors routinely assess both efficacy and harms outcomes of an intervention, and that reporting guidelines specifically targeting harms reporting be developed.

Objectives: To describe radiologist’s attitudes and perspectives on evidence-based medicine (EBM) and their practice. Design: Face-to-face semistructured interviews, thematic analysis. Setting: 24 institutions across six Australian states and New Zealand. Transcripts were imported into HyperRESEARCH software and thematically analysed. Participants: 25 radiologists. Results: Six themes were identified: legitimising decisions (validated justification, prioritising patient preferences, reinforcing protocols), optimising outcomes (ensuring patient safety, maximising efficiency), availability of access (requiring immediacy, inadequacy of evidence, time constraints, proximity of peer networks, grasping information dispersion), overriding pragmatism ( perceptibly applicability, preserving the art of medicine, technical demands), limited confidence (conceptual obscurity, reputation-based trust, demands constant practice, suspicion and cynicism), and competing powers (hierarchical conflict, prevailing commercial interests). Conclusions: Radiologists believe EBM can support clinical decision-making for optimal patient outcomes and service efficiency but feel limited in their capacities to assimilate and apply EBM in practice. Improving access to evidence, providing ongoing education and training supplemented with practical tools for appraising evidence; and developing evidence-based guidelines and protocols may enhance feasibility and promote the confidence and skills among radiologists in applying EBM in radiology practice for better patient care.