''Pure'' alcoholic fatty liver has been widely assumed to be ''benign'
' with very low risk of progression to cirrhosis. Studies thus far hav
e included either patients with coexisting recognised precursor lesion
s of cirrhosis or have been restricted to short-term histological foll
ow-up. We have followed 88 patients, first seen between 1978 and 1985,
with a histological diagnosis of pure alcoholic fatty liver and no ev
idence of fibrosis or alcoholic hepatitis, for a median of 10.5 years,
to determine any factors predictive of disease progression. Of the 88
, at follow-up nine had developed cirrhosis and a further seven fibros
is. Eight of nine patients with cirrhosis had continuing alcohol consu
mption of more than 40 units per week at follow-up; in the other patie
nts, consumption was unknown. Independent histological predictors of p
rogression on index biopsy were: presence of mixed macro/microvesicula
r fat, and presence of giant mitochondria. We can no longer regard alc
oholic fatty liver as benign. in the presence of continuing high alcoh
ol consumption the above histological features identified those at hig
h risk (47-61%) of disease progression. Therefore, patients with these
features should be counselled intensively regarding their alcohol con
sumption.