Steatosis in South African women: How much and why?

BACKGROUND: Globally, steatosis is the commonest type of liver pathology and is closely associated with obesity and the metabolic syndrome. Obesity is common in urban African females but no data is available on hepatic fat content in this population group when compared to other ethnic groups. The aim of this study was therefore to compare hepatic fat content in woman from different ethnic groups in South Africa and to characterise the principle determinants of liver fat.

MATERIALS AND METHODS: A convenience sample of 106 (48 Indian, 29 African and 29 Caucasian) female volunteers aged 20-60 years and having no history of cardiometabolic disorders were recruited. Hepatic fat was determined from CT scans using the liver-spleen attenuation ratio (LAR), which decreases with increasing levels of hepatic fat. Anthropometric and cardiometabolic parameters were measured with insulin resistance determined using the HOMA index and dysglycaemia defined as fasting glucose ≥5.60 mmol/L.

RESULTS: The African subjects had significantly lower hepatic fat content (LAR as median [interquartile range]: 1.35 [1.28, 1.41]) than the Indian (1.22 [1.10, 1.35]; p<0.005) and Caucasian (1.27 [1.16, 1.33]; p<0.05) females even though they had significantly higher BMIs than both groups (p<0.0005 and p<0.05, respectively). Linear regression showed that: subcutaneous abdominal fat was a significant (unstandardised β = 0.007; p = 0.03) negative, whilst insulin resistance (β = -0.97; p = 0.01) and dysglycaemia (β = -3.58; p = 0.01) were significant positive determinants of liver fat; higher hepatic fat levels in subjects with the metabolic syndrome were explained by insulin resistance and dysglycaemia.

DISCUSSION: African ethnicity is associated with low liver fat content. Subcutaneous abdominal fat protects against steatosis, possibly by acting as a triglyceride reservoir. Insulin resistance and dysglycaemia lead to greater hepatic fat deposition and explain higher liver fat levels in subjects with the metabolic syndrome. These observations must be further investigated in longitudinal surveys.