In 1986, Shaffner and Thaler coined the definition of nonalcoholic fatty liver disease (NAFLD) although what we now call NAFLD had previously been reported under different names [1]. Essentially NAFLD describes a clinico-pathological syndrome of variable types of fatty liver changes [from steatosis to nonalcoholic steatohepatitis (NASH) with/without fibrosis, cirrhosis and hepatocellular carcinoma (HCC)] that may occur either primarily or in a complex relationship with various other conditions (Table 1) [2, 3].

In 2020, a panel of authors proposed that NAFLD should be renamed metabolic-associated fatty liver disease (MAFLD) [4]. However, NAFLD and MAFLD are not entirely equivalent and the modification is more than a single-letter change in an acronym [5]. Additionally, the first and somewhat neglected hint to disease variability is the finding that a reduction of steatosis to the point of complete fat loss may accompany the development of advanced liver fibrosis in NASH [6]. This apparent paradox suggests that even the name “fatty liver” is not invariably appropriate in all NAFLD/MAFLD forms and stages. As such, it seems that we are failing to personalize diagnosis from the very beginning, namely from disease definition.

The importance of NAFLD may be envisaged under both a clinical and public health perspective. Clinically, NAFLD poses the challenge of a common condition whose diagnosis, by definition, is based on histological grounds, though no licensed treatment is available [7]. As a public health problem, large strata of global population have one in four individuals suffering from NAFLD, which places a formidable burden on health systems [8].

Within this complex and worrisome scenario, a milestone article recently published by Arrese and colleagues [9], addresses heterogeneity as a fundamental aspect of NAFLD which dictates, through an extraordinary variability of pathogenic pathways, the development of heterogeneous clinical phenotypes and disease outcomes [9]. The present commentary briefly discusses the bases of NAFLD heterogeneity first and next critically illustrates the work by Arrese et al. [9].

Epidemiological models have clearly indicated that the principal physiologic modifiers of NAFLD are age, sex and ethnicity, along with belonging to “non-alcoholic fatty liver disease families” and carrying risk alleles for selected genetic polymorphisms [10]. Among the metabolic modifiers, metabolic syndrome and its components should be considered as the major determinants of NAFLD risk [10].

Why do some, but not other, individuals with simple hepatic steatosis progress to developing NASH? And, importantly, why do only a fraction of those with NASH progress to cirrhosis? Although we have no definite answer to these questions, from a pathobiological perspective, it is fully conceivable that the heterogeneity of disease course reflects variability in exposure and response to metabolic stress, susceptibility to hepatocyte lipotoxicity, differences in repair-response efficacy and “metabolic inflammation” [11, 12]. The net balance among these various pathogenic factors will eventually promote the progression of fibrosis and risk of HCC in NASH [13]. Therefore, the identification of simple biomarkers identifying and scoring each of the above pathomechanisms is a major challenge in translational research.

Studies evaluating biopsy-proven disease suggest that progressive fibrosis develops in up to about 50% of NASH cases, with advanced fibrosis/cirrhosis eventually ensuing in about 10% of the overall NAFLD population [14]. Among NAFLD patients, the only histological independent predictor of liver-related morbidity/mortality, and mortality owing to all-causes, is liver fibrosis, not NASH [14, 15]. However, individual patients exhibit a striking variability in disease progression owing to alcohol consumption, obesity, type 2 diabetes/insulin resistance and genetic factors all playing a key role as disease modifiers [16].

The multiplicity of pathogenic pathways involved in disease progression dictate a substantial heterogeneity of clinical phenotypes of NAFLD and also affect and hinder the diagnostic process [14, 17]. Schematically, a “hepatic NAFLD” form exists as opposed to an “extra-hepatic NAFLD”, the latter accounting for the leading cause of mortality, namely cardiovascular and cancer-related and the former following the course of all chronic liver diseases owing to other etiologies [18]. The confluence, in variable proportions, of the two principal NAFLD forms may ultimately contribute to shaping a potentially uncountable variety of clinical phenotypes and disease outcomes.

Today, little has been proposed to translate the above clinical and pathobiological premises into a consistent and feasible categorization scheme. However, the so called “LDE” classification system (Figure 1) [2, 19] or similar systematic approaches, may eventually lead to selecting more homogenous subgroups of patients with comparable natural course and predictable response to therapy in the setting of advancing the frontiers of personalized NAFLD medicine [20].