Forty-two patients with nonalcoholic steatohepatitis were followed for a median of 4.5 yr (range = 1.5 to 21.5 yr). Except for two patients with lipodystrophy, all were obese; 35 of 42 were women, 26 of 32 were hyperlipidemic and 15 were hyperglycemic. Upper abdominal pain was the most common reason for presentation. Initial liver biopsy specimens showed the presence of macrovesicular fatty infiltration, lobular (acinar) inflammation, apoptosis, Mallory bodies (in four cases) and fibrosis (in 18 cases). Cirrhosis was present at initial diagnosis in one subject and in another two subjects liver biopsy showed marked fibrosis with disturbed architecture. Serial liver biopsy specimens revealed minimal or no apparent progression of the disorder in most of the patients, in keeping with their benign clinical course. However, one patient showed progression from fibrosis to cirrhosis during the 5-yr observation period, and in the patients with extensive fibrosis the liver disease evolved from one of active inflammation to one of inactive cirrhosis without fat or inflammation. The patient with cirrhosis later died of hepatocellular carcinoma. The severity or type of hepatic change did not correlate with the degree of obesity, hyperlipidemia or hyperglycemia. However, in individual patients, poorly controlled diabetes and rapid weight loss preceded the onset of steatohepatitis. We conclude that nonalcoholic steatohepatitis is a cause of hepatic inflammation histologically resembling that of alcohol-induced liver disease but usually slowly progressive and of low-grade severity. However, the disorder may ultimately result in cirrhosis. Nonalcoholic steatohepatitis should be distinguished from alcoholic steatohepatitis and recognized as a further cause of "cryptogenic cirrhosis."

The role of the measurement of hepatic iron in the diagnosis of genetic hemochromatosis was studied, with particular reference to the differentiation of early hemochromatosis from alcoholic siderosis and the critical hepatic iron concentration associated with fibrosis in hemochromatosis. Hepatic iron was measured in 30 homozygous relatives of 17 hemochromatosis probands, 8 heterozygous relatives, 51 patients with alcoholic liver disease and 40 control subjects. Hepatic iron concentrations were greatly increased in the majority of homozygous hemochromatosis subjects, and there was little overlap with the other groups. In the absence of alcoholism, fibrosis or cirrhosis in hemochromatosis was present only with hepatic iron concentrations above a threshold of approximately 400 mumoles per gm (22.3 mg per gm) dry weight. In some heterozygous hemochromatosis subjects and in some alcoholic patients, hepatic iron concentrations were in the range seen in young homozygous subjects. However, an age-related rise in hepatic iron was seen only in hemochromatosis homozygotes, and calculation of an hepatic iron index (hepatic iron/age) resulted in a clear distinction between homozygotes and the other three groups. It is concluded: that chemical measurement of hepatic iron concentration, when corrected for the age of the subject, reliably distinguishes early hemochromatosis from alcoholic siderosis, and, that there appears to be a threshold level of hepatic iron above which there is a high risk of fibrosis.

No abstract

We determined the prevalence of iron overload due to homozygous haemochromatosis in an asymptomatic Australian (predominantly Caucasian) population by surveying 1968 employees of two large corporations. Subjects were screened by measurement of transferrin saturation and serum ferritin concentration and, in all subjects with elevation of both indices, percutaneous liver biopsy was performed to establish whether significant iron overload was present. The prevalence of iron overload due to haemochromatosis in this population was 0.36%. The prevalence rate was not significantly different between males and females, suggesting that this autosomal recessive disease is expressed equally in females given an adequate dietary iron supply. The positive predictive value of a transferrin saturation consistently greater than 45% together with an elevated serum ferritin concentration was 64%. It is concluded that the prevalence of significant iron overload due to homozygous haemochromatosis warranting treatment is approximately 1:300 and that transferrin saturation should be included in existing adult health screening programmes.