A critical analysis of three quantitative methods of assessment of hepatic steatosis in liver biopsies.

The issue of adequately quantitatively evaluating hepatic steatosis is still unresolved. Therefore, we compared three methods of quantitative assessment. Two groups of mice (n = 10 each) were fed standard chow (10% fat, SC group) or a high-fat diet (60% fat, HF group) for 16 weeks, and hepatic triglyceride (HT) and liver tissue were then studied. Paraplast-embedded tissues stained by hematoxylin and eosin (H-E) were compared to frozen sections stained by Oil Red-O (ORO). In addition, the volume density of steatosis (Vv[steatosis, liver]) was measured by point counting (P-C, sections H-E or ORO) or by image analysis (I-A, sections ORO). HT was significantly higher in the HF group (104% greater, P = 0.0004) than in the SC group. With P-C and H-E, Vv[steatosis, liver] was 4.80 ± 0.90% in the SC group and 33.50 ± 3.17% in the HF group (600% greater, P < 0.0001). With P-C and ORO, Vv[steatosis, liver] was 4.86 ± 0.89% in the SC group and 25.21 ± 1.27% in the HF group (420% greater, P < 0.0001). With I-A and ORO, Vv[steatosis, liver] was 4.17 ± 0.85% in the SC group and 23.35 ± 1.58% in the HF group (460% greater, P < 0.0001). Correlations between Vv[steatosis, liver] and HT were strong and significant in all methods. In conclusion, all methods were appropriate and reproducible. In P-C and H-E, there is a slight overestimation of steatosis in the HF animals in comparison to frozen sections and ORO; in frozen sections, differences between P-C and I-A are insignificant.