Epidemiology of chronic liver disease in older people with type 2 diabetes mellitus: the Edinburgh Type 2 Diabetes Study

Abstract

Increasingly chronic liver disease is being acknowledged as a complication of type 2 diabetes, in particular non-alcoholic fatty liver and non-alcoholic fatty liver disease. Rates of non-alcoholic fatty liver are higher in people with type 2 diabetes than in the general population, with prevalence rates believed to be between 40-70%. Given the aging Scottish population and the obesity driven diabetes epidemic, the problem of chronic liver disease is likely to increase. Despite this there has been little investigation into the natural history of nonalcoholic fatty liver disease and the risks of clinically significant chronic liver disease in community based cohorts because diagnosis has been heavily reliant on liver biopsy. The use of liver biopsy is limited in both research and clinical practice due to its associated high mortality (1/1000) and morbidity and also due to practical limitations (sampling variability, semi-quantitative scoring systems). As a result the use of non-invasive markers of liver injury (non-specific liver injury, steatosis, steatohepatitis, liver fibrosis and surrogates of advanced portal hypertension) are rising, in the diagnosis of chronic liver disease, however, their utility in both community cohorts and patients with type 2 diabetes has not been widely studied. The aims of the studies presented in the thesis, using the Edinburgh Type 2 Diabetes Study, were: (i) to describe the distributions of a range of non-invasive markers of steatohepatitis and liver fibrosis in older people with type 2 diabetes, their relationship with metabolic and liver disease risk factors, and to compare the agreement of different non-invasive markers of hepatic fibrosis; (ii) to determine the frequency (prevalence and incidence) of and risk factors for clinically significant chronic liver disease in people with type 2 diabetes; and (iii) to determine the importance of chronic liver disease as a risk factor (or risk marker) for cardiovascular mortality or morbidity in type 2 diabetes. Prior to undertaking this work I undertook a detailed systematic review of the literature relating to the use of non-invasive markers of hepatic fibrosis to inform the choice of markers used in the study. Examination of a wide range of potential markers of steatohepatitis and liver fibrosis found varied relationships with diabetes history. Most commonly, elevated markers of steatohepatitis and liver fibrosis were associated with older age and higher body fat measures. However, most of these relationships between liver markers and body fat measures lost statistical significance when limiting the population to only those with hepatic steatosis and/or non-alcoholic fatty liver disease. There were marked differences in the associations between different liver fibrosis markers and potential diabetes and metabolic risk factors, suggesting that these markers are not actually measuring the same underlying “fibrosis” condition. There was poor correlation between the five markers of liver fibrosis studied. Using the top vigintile (5%) of each marker resulted in excellent agreement on the absence of advanced liver disease but poor agreement on the presence of advanced liver disease. The prevalence of clinically significant CLD (defined as cirrhosis, HCC or gastrooesophageal varices) was 2.2% - 0.9% diagnosed prior to enrolment with an additional 1.4% identified by study investigations. Over nearly 6 years of follow-up, only 1.4% of the cohort developed incident clinically significant CLD. Higher levels of systemic inflammation, steatohepatitis and hepatic fibrosis markers were associated with both unknown prevalent and incident clinically significant chronic liver disease. Less than half of participants developing incident significant disease were identified as high risk by the study investigations. Abnormal liver enzymes were statistically significantly associated with incident cases, however the presence of hepatic steatosis was not. There were 372/1033 (36.0%) patients with prevalent CVD and 319 (30.9%) with prevalent CAD at baseline. After mean follow-up of 4.4 years there were 44/663 incident CVD events, including 27 CAD events. There were 30/82 CVD related deaths. However, risk of dying from or developing CVD was no higher in subjects with steatosis than in those without. There was also no statistically significant relationship between CVD and steatohepatitis or liver fibrosis. The only statistically significant relationship between CVD and any liver markers was with GGT (prevalent CVD, OR 1.28, p=0.007; incident CAD, OR 2.35, p=0.042), suggesting that in our study population, CLD may have little effect on the development of, or mortality from, CVD. In conclusion, the potential for using non-invasive biomarkers to diagnose clinically significant chronic liver disease in type 2 diabetes remains limited, however chronic liver disease is a significant problem in older people with type 2 diabetes and is frequently undiagnosed.