The Clinical Relevance Of Non-Invasive Measures Of Liver Health And Non-Alcoholic Fatty Liver Disease (NAFLD) In Diabetes Mellitus

Abstract

Introduction Non-alcoholic fatty liver disease (NAFLD) is common in adults with diabetes mellitus. In addition, it tends to occur as non-alcoholic steatohepatitis (NASH) and/or with liver fibrosis more frequently than in the general population and this has important implications for morbidity and mortality. While simple steatosis due to NAFLD can be determined by non-invasive means with reasonable ease and reliability, establishing the presence and severity of NASH and liver fibrosis degree requires more sophisticated approaches, which are in need of further development. Furthermore, our understanding and interpretation of currently available clinical measures of NAFLD and its severity need improvement. Aims This thesis aimed to explore the frequency of NAFLD overall, and severe fibrosis due to NAFLD specifically, in cohorts with diabetes mellitus using non-invasive techniques and to define the clinical and biochemical associations with these. It also sought to determine broader clinical associations with liver enzymes, the most commonly used clinical markers of NAFLD, with the objective of improving our understanding of other factors that might influence these measures. The work also planned to investigate the use of both established and novel circulating biomarkers for grading liver disease severity in NAFLD. The integration of these biomarkers into liver fibrosis screening algorithms was explored. Methods Data sourced from 4 complimentary and relevant study populations were used in this thesis. One cohort was developed by a process of deliberate clinical recruitment, while data from the other three cohorts were obtained by either sub-analysis of completed trial data, post-hoc analysis of archived samples, or an audit of a pre-existing clinical database. Given the significant differences between cohorts, no direct comparisons were made between them. Results Steatosis was present in 84% of those with type 2 diabetes from a tertiary diabetes centre who had causes of liver disease other than NAFLD excluded. Non-invasive measure of liver fibrosis by transient elastography (TE) was consistent with severe liver fibrosis in ~ 1 in 5 of this group. Across cohorts, markers of fibrosis in NAFLD, but not steatosis per se, were associated with a more severe diabetes phenotype overall and with peripheral neuropathy specifically as a novel finding in this thesis. Collectively across cohorts, ALT was associated with steatosis but not fibrosis. It was higher in the presence of adverse metabolic features, although it was also noted to be related to other factors. These factors included important negative associations with age after 50 years, higher exercise capacity, full-time employment and higher level of education completed. Consistent with these findings, ALT had a prospective, inverse relationship with cardiovascular events and cardiovascular mortality in a cohort of 9795 individuals with type 2 diabetes in a community-based setting. Across cohorts, GGT was associated with non-invasive markers of NAFLD and liver disease severity in NAFLD, as well as with the presence of adverse metabolic features. It was higher in those assessed as having decreased level of physical and emotional wellbeing by the SF-36 and, consistent with this, higher GGT was independently associated with all-cause and cardiovascular mortality in the same 9795 with type 2 diabetes. Additionally, fenofibrate use was associated with decreases in GGT and ALT with time, when compared to placebo, in this group. Changes in weight, HbA1c and triglycerides were also associated with significant changes in these liver enzymes. Liver stiffness measurement by TE appeared to be a reasonable screening tool for liver fibrosis in those with type 2 diabetes from a tertiary diabetes centre who had liver disease due to causes other than NAFLD excluded, although repeat measure may improve assessment. Algorithms with serial assessment by NAFLD fibrosis score, followed by a biomarker for fibrosis, such as circulating Fibroblast Activation Protein (cFAP), and then TE are likely to be clinically useful for triaging those who may require specialist assessment and liver biopsy. Related to FAP, circulating dipeptidyl peptidase 4 (DPP4) activity was significantly associated with cytokeratin M30 Apoptope, a validated marker of hepatocyte apoptosis. Conclusions NAFLD is a real and present complication of diabetes. Moreover, it is likely to increasingly contribute to liver and perhaps systemic morbidity and mortality due to escalating rates of obesity, an aging population and also better treatments for other diabetes-associated illnesses. It is timely to work towards the development of adequate biomarkers for NAFLD severity and become familiar with their use in people with diabetes in order to stratify appropriate investigations and to initiate care in those at risk for unfavourable health outcomes. This stratification will also help to facilitate the development of effective treatments for NAFLD. Such treatments may not only improve liver pathology but could influence the outcomes in diabetes overall.