Frozen-section diagnosis in donor livers: error rate estimation of steatosis degree.

A high degree of macrovesicular steatosis is associated with a significant risk of graft dysfunction. Most centers, including ours, consider 60% the limit value for transplantability, while others have adopted 30% as a cutoff. Pretransplant frozen-section (PFS) evaluation is used for reliable quantification of steatosis. However, the accuracy of PFS analysis for the degree of steatosis has largely been debated due to its high grade of variability and subjectivity. The aim of our study was to evaluate the accuracy of PFS diagnosis compared with subsequent paraffin histology samples. We retrospectively analyzed PFS from 52 consecutive liver donors. All PFS were blindly reviewed by two pathologists. The results were graded according to two classifications: (A) lower or higher than 60% steatosis, or (B) mild (0%-30%), moderate (30%-60%), or severe (>60%) steatosis. The rate of error for A (two-grade classification score) was 1.9% with the erroneous discharge of a viable organ. The error rate for B (three-grade classification score) was 7.7% with four discrepancies. In three cases, the discrepancy was related to the mild to moderate cutoff value, leading to a clinical error rate of 5.8% (discharging organs with statosis >30%, when we used more strict criteria). Our study validated PFS analysis as a reliable technique when the maximum value for organ transplantation was 60% steatosis. There was a higher error rate when a cutoff value of 30% was used. This finding suggests the usefulness of another technique to support a more precise steatosis evaluation. Copyright 2010 Elsevier Inc. All rights reserved.