• fatty liver
• cardiovascular disease
• diabetes mellitus
• nonalcoholic steatohepatitis

Many,1 but not all,2 previous studies investigating the association between non-alcoholic fatty liver disease (NAFLD) and cardiovascular disease (CVD) have shown that NAFLD is a risk factor for CVD morbidity and mortality. The debate over the last 5 years as to whether NAFLD is an independent CVD risk factor has focused on the suggestion that the association between NAFLD and increased CVD risk occurs because of residual confounding by metabolic syndrome-associated cardiovascular risk factors.

With the burgeoning 21st problem of obesity, NAFLD has become a common disease that is often present but remains often undiagnosed in the adult population. Thus, in large registry studies investigating associations between NAFLD and outcomes such as CVD,2 it is not possible to prove that subjects in the control (reference) group do not have NAFLD. When undiagnosed NAFLD occurs in subjects in the reference group, this causes misclassification bias, and misclassification bias always attenuates the strength of any association between the exposure variable (ie, NAFLD) and the outcome (CVD), towards the null. Moreover, among the few published NAFLD histology cohorts, to date, that have investigated the association between NAFLD and risk of incident CVD, all have been limited by small sample sizes (previous largest study, n=603) with few recorded outcomes and imprecise estimates of risk across NAFLD histological …