Nonalcoholic Fatty Liver Disease Progression Rates to Cirrhosis and Progression of Cirrhosis to Decompensation and Mortality

A Real World Analysis of Medicare Data

Rohit Loomba; Robert Wong; Jeremy Fraysse; Sanatan Shreay; Suying Li; Stephen Harrison; Stuart C. Gordon


Aliment Pharmacol Ther. 2020;51(11):1149-1159. 

In This Article

Abstract and Introduction


Background: Risk factors and timing associated with disease progression and mortality in nonalcoholic fatty liver disease (NAFLD) are poorly understood.

Aims: To evaluate the impact of disease severity, demographics and comorbidities on risk of mortality and time to progression in a large, real-world cohort of diagnosed NAFLD patients.

Methods: Claims data from a 20% Medicare representative sample between 2007 and 2015 were analysed retrospectively. Adults were categorised into disease severity groups: NAFLD/nonalcoholic steatohepatitis (NASH) alone, compensated cirrhosis, decompensated cirrhosis, liver transplant or hepatocellular carcinoma. Cumulative incidence of mortality and disease progression were calculated for each group and multivariate analyses performed adjusting for demographics, comorbidities and disease severity.

Results: A total of 10 826 456, patients were assessed and the prevalence of NAFLD was 5.7% (N = 621 253). Among patients with NAFLD, 71.1% had NAFLD/NASH alone and 28.9% had NAFLD cirrhosis. Overall, 85.5% of patients had hypertension, 84.1% dyslipidemia, 68.7% had cardiovascular disease and 55.5% diabetes. The cumulative risk of progression of NAFLD to cirrhosis, and compensated cirrhosis to decompensated cirrhosis was 39% and 45%, respectively, over 8 years of follow-up. The independent predictors of progression included cardiovascular disease, renal impairment, dyslipidemia and diabetes. The cumulative risk of mortality for NAFLD, NAFLD cirrhosis, decompensated cirrhosis and hepatocellular carcinoma was 12.6%, 31.1%, 51.4% and 76.2%, respectively.

Conclusions: The present report (a) demonstrates that NAFLD is grossly underdiagnosed in real-world clinical settings and (b) provides new evidence on the progression rates of NAFLD and risk factors of mortality across the spectrum of severity of NAFLD and cirrhosis.


The prevalence of nonalcoholic steatohepatitis (NASH) in the US is estimated at 3%-5% of adults.[1] This represents an estimated 17.3 million individuals in 2016 with the prevalence of NASH projected to increase to 27 million cases by 2030.[1] Approximately 20% of adult patients with nonalcoholic fatty liver disease (NAFLD) are thought to have NASH,[1] and NASH is a leading cause of chronic liver disease in the United States.[2] NASH is associated with increased risk of progression to compensated cirrhosis, decompensated cirrhosis, need for liver transplant and hepatocellular carcinoma.[3,4]

Historically, advanced liver fibrosis was considered the most important predictor of clinical outcomes in NASH,[3,5] with advanced fibrosis (METAVIR fibrosis stage F3-F4) associated with an increased risk for overall- and liver-related mortality.[6] More recently, an increased risk in overall- and liver-related mortality has been reported for those with early-stage disease as well (METAVIR fibrosis stage F1-F2).[7] However, to date limited data are available regarding the risk of overall mortality for patients with late stages of the disease, including for patients with progression to decompensated cirrhosis or hepatocellular carcinoma.

A systematic review and meta-analysis based on five cohort studies of 1495 patients with 17 452 patient-years of follow-up quantified relative all-cause and liver-related mortality risk by histologically confirmed stage of fibrosis. There was an exponential increase in the risk of all-cause and liver-related mortality with each increase in the stage of fibrosis.[8] The carefully designed and executed histologically based studies in this systematic included review include inherent limitations that necessitate additional evaluations to elucidate the natural history of NASH as well as factors associated with clinical outcomes in NASH patients. These limitations include findings based on small cohort studies with entry requiring referral by an experienced hepatologist and limited to only patients with liver biopsies. As real-world rates of biopsy in NASH patients have been found to be low (less than 2%),[9] these limitations likely introduced selection bias in previous natural history studies towards patients with worse prognosis. Furthermore, while biopsy is considered the reference standard to assess liver disease severity, it is associated with well-recognised sampling variability and low uptake in real-world settings.[10]

Given these limitations, data gaps exist in our current understanding of the natural history of patients with NASH in the real-world setting. The objectives of this retrospective cohort study were to further our understanding of the natural history of NASH by evaluating the impact of liver disease severity, demographics and comorbid health conditions on risk of all-cause mortality and risk of disease progression in a generalisable sample of Medicare patients with a diagnosis of NAFLD/NASH. This study fills a major gap in the natural history of NAFLD and NAFLD cirrhosis regarding the real-world evidence of progression and mortality in this patient population. The Medicare 20% sample evaluated, with beneficiaries randomly selected for entry, is generalisable to the over 50 million Medicare beneficiaries and is an ideal claims-based data source for this evaluation.[11] Once an individual is enrolled in Medicare, their withdrawal is extremely rare and individuals are typically followed until death.[11]