An evaluation of the reliability of muscle fiber cross-sectional area and fiber number measurements in rat skeletal muscle

Background: The reliability of estimating muscle fiber cross-sectional area (measure of muscle fiber size) and fiber number from only a subset of fibers in rat hindlimb muscle cross-sections has not been systematically evaluated. This study examined the variability in mean estimates of fiber cross-sectional area as a function of the number of fibers measured, and tested whether counting a subset of fibers in a cross-section could predict total fiber number in middle-aged rats. Results: Soleus and extensor digitorum longus (EDL) muscle cross-sections from 23-month-old, male Fisher 344 x Brown Norway rats were stained for myofibrillar ATPase activity to identify muscle fiber type (either type I [slow-twitch] or II [fast-twitch]) and laminin to facilitate fiber cross-sectional measurements. We outlined the circumference of 1000 to 1600 single muscle fibers for measurement of fiber cross-sectional area within muscle sections. Mean type I fiber cross-sectional area was based on soleus muscle sections which were predominantly composed of type I muscle fibers. Mean type II fiber cross-sectional area was based on EDL muscle sections which were predominantly composed of type II muscle fibers. A bootstrapping resampling technique demonstrated that variability in sampling distribution of mean type I and II fiber cross-sectional areas decreased and gradually stabilized as the number of fibers measured increased with large declines in variability occurring at numbers below 150 fibers. Coefficients of variation for bootstrapped mean type I fiber cross-sectional areas were lower than for type II. In the same muscle sections, total fiber number was compared to fiber numbers within 1, 2, 3, and 4 fixed field areas (10x magnification; 1000 x 1500 pixels in size/field) on the cross-section. Fiber numbers from 3 to 4 fields (approximating 15 to 20% of the cross-section) provided a reasonably predictive value of total fiber number (r=0.57-0.59, P=0.003). Conclusions: These data describe a pattern of improved precision in estimating mean fiber cross-sectional area as sample size of fibers measured increases to at least 150 in this rat model. Counting 15-20% of the fibers in cross-sections provides a reasonably reliable estimate of the total fiber number.