The natural history of non-alcoholic fatty liver disease

Nonalcoholic fatty liver disease (NAFLD) is a potentially serious condition present in 20-30% ofthe general population. Insulin resistance is an important pathogenic factor for NAFLD. Subsequently, as the prevalence of conditions associated with insulin resistance such as diabetes and obesity increase in the community, the prevalence of NAFLD is also likely to be increasing. Thus NAFLD is becoming a frequently encountered clinical scenario for physicians and general practitioners. Knowledge of the natural history of NAFLD is vital to guide clinical management decisions regarding investigation, monitoring and treatment as well as for patient counselling. At a population level, knowledge of the natural history guides resource allocation and public health policy on prevention and treatment.

It is recognized that some patients with NAFLD progress to cirrhosis which may be complicated by hepatocellular carcinoma, liver failure and death. However, previous studies examining the natural history of NAFLD have had significant methodological problems including selection biases, limited numbers and short follow-up periods. Therefore the natural history is not clearly defined, with the magnitude and rate of disease morbidity and mortality unclear and prognostic factors related to these outcomes unknown.

This thesis examined the natural history of NAFLD in a number of cohorts generated from the Rochester Epidemiology Project and the diagnostic index of the Mayo Clinic. Firstly, the clinical evolution to cirrhosis, liver related death and overall death was examined in a community based cohort. Secondly, the histological evolution of subclinical NAFLD was detailed in a cohort of patients with serial liver biopsies. The third and fourth experimental chapters examined the clinical outcomes of subjects with NAFLD and diseases associated with insulin resistance, namely hypopituitarism/hypothalamic disease and type 2 diabetes mellitus.

To evaluate the clinical evolution of NAFLD, 420 community based subjects were followed for a median duration of 7.1 years (range 0.1-23.5). Death occurred in 12.6% ofthe cohort and 1.7% died fro'm liver related causes. NAFLD was associated with an increased risk of death (standardized mortality ratio 1.34 (95% CI 1.003-1.76], p=0.03) in comparison to the general population. Liver disease accounted for 13% of all deaths among NAFLD patients in contrast to less than one percent of all deaths in the general population. Impaired fasting glycemia and high body mass index (BMI) were risk factors for overall and liverrelated death. The occurrence of cirrhosis and its complications was relatively low, at 5% and 3.1% respectively.

The histological evolution of NAFLD was examined in a cohort of 103 subjects who had undergone serial liver biopsies a mean of 3.2 years (range 0.7-21) apart. Progressive fibrosis was observed in 37% of subjects with nine percent becoming cirrhotic. Fibrosis progressed slowly at an average rate 0.02 + 0.66 stages/year or at 0.09 + 0.67 stages/year when cirrhotics were excluded. Diabetes mellitus and BMI were significantly associated with fibrosis rate on multivariate analysis. In contrast to fibrosis, the histological features of steatosis and necro-inflammation and serum aminotransaminase levels improved over time.

Following the identification of diabetes mellitus and obesity as significant adverse prognostic factors, the natural history of NAFLD was examined among other conditions associated with insulin resistance. Firstly, 21 subjects with hypopituitarism/hypothalamic disease were identified and followed. From time of diagnosis, patients were observed to develop obesity, glucose intolerance and dyslipidemia and were subsequently diagnosed with NAFLD. The course of the liver disease in these patients was severe. Cirrhosis developed in 29% of the cohort and liver related death or transplantation occurred in 14% after a median duration of follow-up of only 6.0 years (range 1.0 to 10 years).

To further examine the natural history of NAFLD among conditions associated with insulin resistance, a cohort of 337 community-based subjects with type 2 diabetes mellitus were followed for a median of 9.9 years (range 0.1-25.0). During follow-up, a diagnosis of NAFLD was found in 116 patients and was an independent risk factor for death (hazard ratio 1.63, 95% confidence interval 1.04-2.56). Liver related deaths were more frequent among patients with NAFLD occurring in 19% compared to 0% among those without NAFLD.

In summary, NAFLD is a slowly progressive disease that leads to cirrhosis and liver related death in a minority of individuals. These sequelae contribute to the increased risk of overall death present among NAFLD patients in the community and among NAFLD patients with diabetes mellitus. Therefore, because NAFLD is common in the general community, the potential disease burden is considerable. Clinicians should be aware that patients with clinical phenotypes of insulin resistance, namely obesity, diabetes mellitus and hypopituitarism/hypothalamic disease are at increased risk of progressive liver disease and death. These patients should be appropriately counselled, considered for liver biopsy and targeted for interventions aimed at reducing liver related morbidity and mortality.