Interobserver variability in analysis of asbestos fibres and asbestos bodies in human lung tissue.

Abstract

Two different methods of quantifying asbestos fibre burden were assessed and the counts obtained were compared with semi-quantitative asbestos body counts in corresponding tissue sections. Comparison of the two methods found significantly different asbestos fibre counts between specimens. Each technique showed wide limits of agreement for reproducibility and interobserver variability as assessed by Bland-Altman plots, such that a repeated count could not necessarily be expected to lie within the same exposure category. Asbestos body counts in tissue sections were reproducible with good correlation between observers. Asbestos body and asbestos fibre counts showed correlation in some samples but not others. Counting of asbestos bodies is a valuable screening technique as the finding of asbestos bodies is accepted as a marker of significant asbestos exposure. When no asbestos bodies are identified asbestos fibres estimations may be useful in proving asbestos exposure. Different techniques are not interchangeable and each laboratory should establish a background range from unexposed individuals.