Comments from the Editors
What's in a NAme?†
Potential conflict of interest: Nothing to report.
What's in a name? That which we call a rose by any other name would smell as sweet.
Romeo and Juliet, William Shakespeare, 1594
What Is Meant by “Nonalcoholic Steatohepatitis” (NASH)?
Despite sincere efforts on the parts of numerous published and unpublished individuals, myself included, the nomenclature for nonalcoholic fatty liver disease (NAFLD)/NASH has remained linked to the fact that it is a “non-association” with alcohol abuse rather than to the underlying condition(s) with which we recognize an association.1–4 Various suggestions that have been made, or could be made, include the earlier descriptors “steatofibrosis,” “fatty liver hepatitis,” and “pseudo-alcoholic fatty liver disease” and the subsequent descriptors “metabolic syndrome steatohepatitis” (MESH), “metabolic fatty liver disease,” “obesity-related fatty liver disease,” “drug-associated steatohepatitis (DASH),” and “insulin-resistance induced steatohepatitis” (IRISH). One could, of course, ponder many variations on this theme [e.g., “transplant-associated steatohepatitis” (TRASH)]; unfortunately, it seems that “NASH” is catchy and may be here to stay, much like the misnomer “primary biliary cirrhosis,” which is used for antimitochondrial antibody–positive nonsuppurative destructive cholangitis, whether it is cirrhotic or not.
Many pathologists in the field have tended to agree and tried to push for a change in the nomenclature for reasons that go beyond the fact that the determination of “non–alcohol consumption” must be ultimately made by the clinician, so our diagnoses, by definition, must be purely descriptive. Perhaps more important is the fact that although the histologic features of alcoholic liver disease (ALD) do indeed cover a spectrum ranging from steatosis to steatohepatitis to cirrhosis with or without steatohepatitis, which is similar to the spectrum reported for NAFLD, ALD also includes lesions that are not present in NASH. Fundamentally, NAFLD must include steatosis, but ALD does not; alcoholic hepatitis is not uncommonly without steatosis. NAFLD would be quite rarely considered in cases of pure microvesicular steatosis; however, ALD can occur as “foamy degeneration,” in which the steatosis is entirely microvesicular steatosis.5 Other findings of ALD that are not found in the obesity-related fatty liver disease spectrum include perivenular fibrosis,6 “veno-occlusive disease–like” lesions of terminal hepatic venules,7 sclerosing hyaline necrosis,8 pericholangitis with abundant polymorphs, canalicular cholestasis, large and well-formed Mallory-Denk bodies, and often, in cirrhosis, large regions of parenchymal extinction (scar) with micronodular cirrhosis (reviewed by investigators in the field8–10). Clearly, however, the histologic features of “milder” forms of alcoholic steatosis and steatohepatitis substantially overlap those of obesity/metabolic-related steatosis/steatohepatitis, and a pathologist cannot discern the clinical diagnosis.
What Is the Meaning of “Simple Steatosis”?
The concept that NAFLD is a spectrum ranging from simple steatosis, which is benign, to NASH, which is potentially progressive, to cirrhosis is so commonly stated in the introductory lines of studies on the subject that it is apparently accepted without question. As a person whose work is to observe, dissect, and evaluate patterns of injury and repair in order to put a story together, I will admit to questioning and concern in these areas. What truly is so simple about simple steatosis? Moreover, is it really a benign condition? In the early work that centered on establishing recognition and acceptance of NAFLD by the medical community, the histologic definitions of the entities within this process were neither well defined nor agreed upon. Descriptions ranged from fatty infiltration, steatosis (not otherwise defined or localized), and inflammation (likewise not defined or described) to steatosis, Mallory's hyaline (now called Mallory-Denk bodies), and pericellular fibrosis and many variants thereof; the result was wide variation in reports of the incidence of “fatty liver” versus “steatohepatitis”11, 12 and thus in comparisons of results between and among studies.13 Pleas were published for some kind of standardization of findings13 and for minimum hepatocellular injury to be recognized in order to identify/diagnose the potentially progressive form of this entity.2, 14 The application of reasonably well-defined criteria has largely gained acceptance, and at the least, most studies published now include detailed descriptions of how the diagnosis of “NASH” was derived.
However, when and how was “simple steatosis” defined, and why is it seemingly so universally agreed that it is simple and benign? Is this related to a common lack of alanine aminotransferase (ALT) elevation? If so, how many cases of progressive fatty liver disease are we ignoring by referring to it as “simple steatosis”? Browning et al. found that 79% of subjects with a >5.6% hepatic triglyceride content by proton magnetic resonance spectroscopy had normal ALT.15 Biopsy-based studies have documented the full range of histologic lesions, including cirrhosis, with normal ALT in adults16, 17 and children.18 Would it not be more accurate, therefore, to state that at the current time, it is recognized that the majority of subjects with obesity-related fatty liver likely will not progress to serious, end-stage liver disease, instead of stating unequivocally that the condition is benign? Are we justified in calling this condition benign when studies have shown that hepatic steatosis itself can be a significant cofactor in the progression of other forms of liver disease?19–22 Have we put aside the data of the original Matteoni study,23 which showed that 4% of patients with type 1 fatty liver (histologically determined “steatosis only”) progressed to cirrhosis over time? There is a significant extant and growing body of literature indicating that the presence of liver steatosis, in itself, is a marker of other serious metabolic disease processes, such as cardiovascular disease, stroke, hypertension, and diabetes.24–30 Studies have shown that an accumulation of lipids in hepatocytes may directly lead to organelle dysfunction (specifically in the mitochondria and endoplasmic reticulum) and may itself participate in the development of hepatic insulin resistance.31–32
It is recognized that some ethnicities/races are less likely to develop steatosis despite similar rates of obesity measured by the body mass index and features of metabolic syndrome.33–34 Recent work continues to implicate visceral adiposity rather than subcutaneous adiposity as a more significant contributor to the features of metabolic syndrome than the body mass index; this may help explain the lower prevalence of obesity-related fatty liver in African Americans, as they have relatively less visceral adipose tissue than other ethnic/racial groups.34 Alternatively, the increased prevalence of fatty liver in Asians with the otherwise “thin” body habitus has been correlated with the relatively increased visceral compartment of adiposity,35 and recommendations for the use of anthropometric standards specific to Asians in the determination of overweight and obesity are encouraged.36 Do we honestly know if there are differences in the significance of steatosis based on the underlying genetic differences among ethnicities?
From intriguing studies demonstrating that triglyceride formation is actually protective of hepatocytes,37 one can infer that the intracellular fatty acid metabolism related to insulin resistance is a significant component of the lipotoxicity of fatty liver disease if it is not adequately sequestered and stored as triglycerides. Altered compositions of intracellular lipids of several classes were reported in a lipidomic analysis of fatty livers and steatohepatitis compared to normal livers.38 Among other results, the study found no differences in hepatic free fatty acids between normal livers, fatty livers, and steatohepatitis, but it did find an accumulation of free cholesterol and decreased arachidonic acid in the fatty livers. The former is recognized to be cytotoxic, and the latter has potentially protective anti-inflammatory effects. This intriguing study suggests that steatosis itself is not so simple.
The histologic definition of “simple steatosis” is equally questioned. I grant that the majority of studies with pathology are based on livers that are biopsied or resected, and the majority of liver biopsies are performed because of some evidence of clinical dysfunction. Having acknowledged this, I have found that several published studies of human fatty liver disease include “simple steatosis” as a category, and this is what is being questioned. Is “simple steatosis” more than just the generally accepted 5.6%?15 Is it still “simple” when present in a cirrhotic liver? Does it, as it almost seems by definition, exclude inflammation, Kupffer cells and occasional eosinophils around a single steatotic hepatocyte (lipogranulomas), or collections of pigmented Kupffer cells (microgranulomas) and other forms of lobular inflammation? In a recent correlation study of 933 centrally reviewed liver biopsies with the full spectra of histologic findings of obesity-related fatty liver disease from the National Institute of Diabetes and Digestive and Kidney Diseases–sponsored Nonalcoholic Steatohepatitis Clinical Research Network, all but 4 of 728 (0.55%) adult biopsies and all of the 205 (100%) pediatric biopsies had some degree of lobular inflammation.39 In the same series, portal inflammation was present in 52% of the adult biopsies and 90% of the pediatric biopsies. With the exclusion of the 58% of adult biopsies and 30% of pediatric biopsies with “definite steatohepatitis,” this finding seriously questions the consideration that “simple steatosis” in human fatty liver (i.e., without inflammation) exists.
Do we know what type or types of steatosis are being seen when “simple steatosis” is recorded? Are there differences in disease progression when the steatosis is large-droplet only or a large-droplet/small-droplet mix? Surely, this is a reflection of different metabolic activities, just as “true” microvesicular steatosis is. Pathologists can aid in the process of clarifying many of these questions by diagnosing exactly what is seen, as suggested by the Matteoni type 1 and type 2 classifications for fatty liver disease23 (steatosis and steatosis with inflammation), and by carefully documenting the types of steatosis, acinar location, and types and locations of inflammation. However, unless the community as a whole can adapt more appropriate and accurate terminology than “Simple Steatosis,” this lesion related to MEtabolic dysfunction will remain an area of apparent contradictions and thus a bit of a MESS.
What Is New in Obesity-Related Fatty Liver Disease Pathology?
As much as we have learned about fatty liver pathophysiology in the past decades, it is worth noting that histopathologic lesions are continuing to be studied and refined. The most apparent result in the evolution is the variance of lesions and patterns of lesions in fatty liver disease related to nondrinking, obese children (primarily boys).40 Schwimmer et al.40 codified the different patterns for pediatric fatty liver disease into three types: type 1 (similar to the adult pattern), type 2 (more common in children and boys), and overlap.40 Other groups have since agreed that the pediatric liver disease of obesity and/or diabetes is more often not similar to that of adults,41 but they have also shown that the type 2 pattern is not as common as stated in the initial report from California. Nobili et al.42 studied an Italian cohort of children, and Carter-Kent et al.43 evaluated children from multiple sites in North America. What all these studies have had in common is this: regardless of ethnicity or race, pediatric livers do not commonly manifest the effects of increased hepatocellular lipids and inflammation with zone 3 steatosis, ballooning, Mallory-Denk bodies, and perisinusoidal fibrosis seen in adults but rather show either panacinar steatosis or zone 1 predominance of steatosis. Rarely discussed but personally observed is the fact that steatosis in children is also more likely to be purely large-droplet macrovesicular steatosis in affected hepatocytes rather than the common mixed large-droplet and small-droplet steatosis noted in adults. Likewise, portal inflammation and portal fibrosis are much more likely to occur prior to significant zone 3 fibrosis and architectural remodeling in children than in adults. (However, zone 3 perisinusoidal fibrosis can occur in pediatric biopsies.) Why do these differences occur? Why is ballooning (as seen in adults) so uncommon in children? Why are Mallory-Denk bodies so rare? Has the disease not been present for a proper length of time for cytoskeletal filaments to aggregate? That seems unlikely given that cirrhosis is known to occur in children. What additional “toxic” event is missing? Is the zone 1 predominance a manifestation of an “immature” gut releasing increased endotoxin to zone 1 hepatocytes? Is the different histology and predominance in boys a result of a lack of female or adolescent/adult hormones? Is this an age-related difference in cellular metabolic functions in hepatocytes of zones 1 and 3 related to insulin “resistance”? Are there differences in the lipid droplet–associated proteins of children and adults, and if so, are they secondary or primary? Is the immune response of children (and boys) differently activated/manifested? Moreover, one of the very fundamental questions of pediatric fatty liver disease is what are the histologic criteria that fulfill a diagnosis of “steatohepatitis.” As noted in a recent publication,44 at this point in time, the closest that dedicated pathologists can come is to identify the pediatric findings as the “borderline, zone 1 pattern.” Whether a formal interobserver variability study would prove this to be true or not has not been tested, but a personal observation is that a pediatric liver biopsy with obesity-related fatty liver disease is reasonably straightforward to identify as such. Interestingly, although many rodent models with high-fat diet–induced “NAFLD” show little evidence of histologic steatosis visible as macrovesicular lipid droplets by routine light microscopy, a dietary model that includes transfats and high fructose corn syrup has large lipid droplets initially deposited in periportal hepatocytes, with eventual zone 3 expansion by hepatocellular small droplet–only steatosis45; with time, periportal fibrosis develops.46
There are several other active areas of investigation in our field, but for the sake of space, I will mention only one more: our considerations of mechanisms of hepatic fibrogenesis are being expanded by studies of the interactions of hepatic steatosis, hepatocyte senescence, hepatic progenitor cell proliferation, and periportal matrix deposition. First showing this in human livers with hepatitis C,47 Clouston et al.48 have expanded their work to obesity-related fatty liver disease. The possible role of epithelial-mesenchymal transition, noted in chronic liver injury in animals and humans,49, 50 awaits investigation for fatty liver disease. It is a great time to be a pathologist in liver disease!