Nonalcoholic fatty liver disease (NAFLD) describes fatty infiltration of the liver in individuals with features of the metabolic syndrome in the absence of excessive alcohol consumption or of secondary causes such as medications and hereditary disorders. Histologically, the spectrum of NAFLD ranges from a non-progressive phenotype called nonalcoholic fatty liver (NAFL) to progressive nonalcoholic steatohepatitis (NASH) [1]. In NAFL, only macrosteatosis is present histologically, whereas in NASH, macrosteatosis is present with a variable mix of hepatocyte ballooning and inflammation with or without fibrosis. The distinction of different forms of NAFLD is important due to its significant impact on disease prognosis and management [1]. The natural history of NAFLD depends on the histologic subtype: Patients with NAFL rarely progress to cirrhosis, whereas a sizable proportion of patients with NASH will progress over time to cirrhosis [1]. One-third of patients with NASH progress to the next stage of fibrosis over a 3- to 5-year span [1, 2] with 9–20 % developing cirrhosis over 10 years [2]. Approximately half of NASH patients who develop cirrhosis decompensate; of these, one-third need to be transplanted in order to survive. The most rapidly increasing indication for liver transplantation (LT) in the USA is NASH [3]. It is therefore not surprising that liver fibrosis has emerged as the strongest predictor of long-term outcomes in patients with NAFLD [4], in whom liver biopsy is often performed as a prognosticator.

Since NAFLD patients commonly manifest components of the metabolic syndrome such as obesity, type 2 diabetes mellitus, and hypertriglyceridemia, NAFLD is currently considered a hepatic manifestation of this syndrome. In patients with type 2 diabetes (T2DM), the prevalence of NAFLD is up to 70 % and is often present in association with cardiovascular disease [5]. With increased prevalence of T2DM, NAFLD has been linked to a higher burden of liver disease and associated hepatocellular cancer (HCC) [6]. A population-based study reported increased incidence of T2DM with hospitalizations for chronic liver disease and HCC when compared to non-diabetics despite adjusting for age and socioeconomic status [6]. Diabetes is a key predictor of NASH and advanced fibrosis in NAFLD. In the prospective studies conducted by NASH Clinical Research Network, NASH and advanced fibrosis were present in 69 and 41 %, respectively, in diabetic patients with NAFLD [7]. With such a high prevalence of NAFLD with advanced phenotypes in T2DM, it is logical to ask whether individuals with T2DM should be screened for NAFLD and whether such screening would be cost-effective. Currently, practice guidelines do not support routine screening for NAFLD/NASH in patients with T2DM [8].

In this issue of Digestive Diseases and Sciences, Corey et al. [9] examine the cost-effectiveness of screening for NASH in middle-aged individuals with T2DM. The screening strategy included: a) one-time screening ultrasound, b) liver biopsy in those who have fatty liver, and c) medical therapy in those who have NASH. Although screening for NASH predicted decreased the progression to cirrhosis and liver-related deaths, it was not cost-effective due to the disutility (lack of efficacy) and the negative impact on quality-adjusted life years (QALYs) of pioglitazone. When the model excluded the disutility associated with treatment, the screening became cost-effective, at an incremental cost-effectiveness ratio of $42,134 per QALY.

Why is screening necessary when there is no approved treatment or when there are no treatments to date that improve patient outcomes? Although there is a temptation to use pioglitazone and vitamin E based on the PIVENS (pioglitazone vs. vitamin E for treatment of non-diabetic NASH) trial [10] that reported significant efficacy for these agents, the PIVENS trial did not include patients with T2DM. Moreover, weight gain during treatment, recurrence of NASH after discontinuation of treatment, and prior concerns about its association with bladder cancer have reduced the popularity of pioglitazone for the treatment of NASH. The treatment strategy proposed by Corey et al. is therefore not only hypothetical but also impractical. The authors additionally excluded vitamin E in their modeling due to a lack of evidence for the efficacy of vitamin E in diabetic patients with NASH, a meta-analysis showing no histologic benefit with vitamin E and possible increased all-cause mortality with high doses [11, 12]. Yet, a recent pooled analysis from the PIVENS and FLINT (Farnesoid X Receptor Ligand Obeticholic Acid in Nonalcoholic Steatohepatitis Treatment) clinical trials reported that vitamin E treatment was associated with histologic improvement in diabetic (OR 4.4, 95 % CI 1.1, 18.0, p = 0.04) and in non-diabetic patients (OR 3.1, 95 % CI 1.7, 5.8, p < 0.001) [13]. The similar significant improvement in NASH histology in both diabetic and non-diabetic patients was not associated with any safety concerns [13]. Furthermore, the predictive model proposed by Corey et al. included only medical therapy for NASH without any consideration of bariatric surgery in at least a subset of T2DM patients with advanced fibrosis. Several prospective studies have indicated that bariatric surgery in patients with NASH was associated with resolution of NASH and T2DM [14].

In the NASH screening strategy, all patients receive a one-time screening ultrasound, with all individuals with fatty infiltration undergoing a liver biopsy [6]. Nevertheless, adherence to guidelines is never 100 % in actual practice. Moreover, many clinicians use blood-based noninvasive tools such as NAFLD fibrosis score and fibrosis 4 index to identify NAFLD patients at high risk of NASH in order to limit unnecessary liver biopsies [8]. One limitation with this approach is that these tests are geared toward identification of patients with advanced fibrosis (NASH CRN fibrosis stage of ≥3) [15]. A simple, noninvasive test that can be performed at bedside that is reliable for identification of NAFLD patients with early but clinically significant liver fibrosis (NASH CRN Fibrosis stage ≥2) is therefore very desirable. One such option is to use transient elastography by Fibroscan® or magnetic resonance imaging that simultaneously estimates hepatic steatosis and liver stiffness (a marker for liver fibrosis) [16]. Availability of such a test may help healthcare providers intervene early with prognostication by liver biopsy or therapy with liver-directed treatments, or perhaps even with enrollment into a therapeutic clinical trial. A combination of targeted screening with accurate biomarkers, selective use of liver biopsy, and early intervention with effective therapy could lead to cost-saving interventions.

Quantifying the value proposition of screening for NASH in T2DM is premature at this point due to lack of accurate noninvasive diagnostic tests, lack of approved treatment options, and a rapidly changing diagnostic and therapeutic landscape (Table 1). Still, the Corey study is an important exercise that mirrors current practice regarding the diagnosis and treatment of NASH. The negative conclusion of the Corey study underscores the many unmet needs that remain in the care of individuals with NAFLD and NASH.

Table 1 Factors that could influence the cost-effectiveness analysis of screening for nonalcoholic steatohepatitis in patients with type 2 diabetes mellitus
Full size table