Experimental analysis of error sources in fibre type counts of biopsies in horses.

The contribution to total variance of different error sources in fibre type counts of equine gluteus medius muscle biopsies was determined to quantify and possibly improve the resolution of the method. Fibre types were defined on the basis of myosin heavy chain immunostaining. Errors were determined at levels: (1) positioning the insertion channel, (2) positioning the needle tip (3) biopsy heterogeneity (4) observer interpretation. Errors at levels 1 and 2 were considerable. Confidence intervals for individual observations were +/- 10-15%. In longitudinal studies a group size of 4 animals is necessary to resolve fibre composition changes of 10%. Comparison with multiple counts on post mortem specimens showed that local muscle fibre heterogeneity is responsible to a considerable extent for the error variance. Variance is effectively reduced by processing 3-4 shavings from the same insertion channel.