NONALCOHOLIC FATTY LIVER-DISEASE - ASSESSMENT OF VARIABILITY IN PATHOLOGICAL INTERPRETATIONS

The exact cause, prevalence, and rate of progression of nonalcoholic f atty liver disease (NAFLD) are unclear because of a lack of agreement on the pathologic features associated with the different types of NAFL D, their clinical syndromes, and because of a lack of accuracy in the interpretation of these pathologic features. Studies of NAFLD would be aided by a consistent and standardized approach to the interpretation of pathologic features. The aim of our study was to assess interobser ver and intraobserver variation in the histologic abnormalities associ ated with NAFLD. We identified histologic features of NAFLD as reporte d in the literature, and we identified patients with the diagnosis of NAFLD through the databases of two large institutions. Histologic para meters were evaluated for each liver biopsy specimen by four hepatopat hologists and twice by two of the four pathologists (blindly). Interob server and intraobserver concordance among the pathologists was measur ed by kappa statistics. Nineteen histologic parameters compartmentaliz ed into steatosis, inflammation, liver cell injury, and fibrosis were evaluated on 53 liver biopsy specimens. Significant, substantial, or m oderate concordance was present in only six items: the extent of steat osis, sinusoidal location of fibrosis, perivenular fibrosis, grade of fibrosis, ballooning degeneration, and the presence of vacuolated nucl ei. Substantial or moderate concordance also was seen for interobserve r readings for location of steatosis and periportal injury. Parameters of inflammation were not scored as reliably as parameters of fibrosis and cell injury. We conclude that only some histologic features previ ously reported in NAFLD (especially those with substantial and moderat e concordance for both interobserver and intraobserver interpretation) are interpreted uniformly by experienced pathologists. These histolog ic features might prove useful for the development of a standardized a nd reliable pathologic scoring system that includes the full histologi c spectrum of NAFLD and its various clinical outcomes.