New nerve repair techniques should only be introduced into general clinical practice if they can be conclusively proved efficient in improving the results obtained from previous techniques. To reach this goal, evaluation methods that provide an objective measure of recovery are required. Animal models also provide objective measures of functional recovery in a manner not presently obtainable in clinical studies. Morphological and electrophysiological measures reflect the inherent variability in the rate of nerve regeneration, myelination, and functional recovery; therefore, a combination of electrophysiological and morphometric measures may yield the best indication of recovery, especially over multiple time points. We demonstrated that recovery of MCV and mean fiber diameter were well correlated with time after sciatic nerve transection. Although mean myelin thickness, axonal diameter, and g-ratio decreased after transection, they were not well correlated with time or MCV recovery.
Conventional MCV measurements tend to reflect primarily upon the faster conducting fibers and provide little information about the conduction properties of the entire population of regenerating fibers (Rosen and Jewett 1980
; Dorfman 1984
). The present study showed that MCV progressively increased through 50–200 days after transection, although it did not return to normal by 200 days. These observations reflect the recovery process of the regenerated fibers. Conduction velocity increases in appropriate proportion to fiber diameter (Rushton 1951
; Moore et al. 1978
); therefore, the increase in MCV should reflect an increase in the relative number of fibers with large diameters. Indeed, the histograms plotted in our study revealed a substantial increase in the number of fibers with large diameters during recovery. While peak posttransection MCV was within 80% of that measured in intact nerves, mean fiber diameter remained substantially below that of the intact nerves. Moreover, the histograms for fiber diameter in the transection group revealed a unimodal distribution at all time points up to 200 days, while the fiber diameter distribution for the control group was bimodal, with a significantly higher proportion of fibers with large diameters. Dissociation between MCV recovery and mean fiber diameter recovery, which was calculated from the whole fibers, is therefore expected. This may simply imply that many nonfunctional regenerating fibers could not be eliminated morphologically, or that there were no significant differences in MCV between the various groups.
Many of the fibers with small diameters may in fact be nonconducting and degenerating. As the nerve fibers regenerate distally and reach the appropriate target organ, fiber diameter increases and the myelin sheath grows (Weiss et al. 1945
; Schröder 1972
; Myles and Glasby 1991
). If sprouting axons do not make an appropriate connection with the target organ, they are denied vital growth factors and degenerate. It has been demonstrated that in rat sciatic nerves, there is an initial increase in the number of fibers distal to the site of transection, followed by a gradual decrease (Mackinnon et al. 1991
). The initial increase can last for approximately six months before axonal number slowly decreases back to pretransection levels over the following two years. It may be difficult to distinguish smaller, successfully regenerated fibers from atrophic, dying fibers, especially during the early phase of regeneration. Therefore, if studies on the morphological evaluation of rat sciatic nerves are completed within six months, their results may be considered inappropriate.
There was a marked dissociation between axon diameter and myelin thickness during regeneration (Cragg and Thomas 1964
; Schröder 1972
). Regenerated fibers have thinner myelin sheaths than those of normal fibers, although axonal diameters may approach normal values. In the present study, mean fiber diameters increased with time, and they increased to 46% of the normal value at 200 days after nerve repair; however, mean myelin thickness decreased at 150 days. There is an optimal myelin thickness relative to fiber diameter (as measured by the g
-ratio) to maximize conduction velocity (Rushton 1951
). The scatter plots of g
-ratio against axon diameter and their regression curves showed that larger fibers had higher g
-ratios, whereas smaller fibers had excessively low g
-ratios. The mean axon diameter increased between 50 and 150 days; however, it decreased at 200 days. In contrast, the number of fibers with low g
-ratios increased at 200 days. The highest number of small-caliber axons with much thicker myelin sheaths (low g
-ratio fibers) were observed at 100 and 200 days after nerve transection. These fibers with low g
-ratios may be those that failed to reach their target organ, with ensuing collapse of the myelin sheath around a shrinking axon (Beuche and Friede 1985
). Therefore, neither mean axon diameter nor myelin thickness provided an accurate morphological index of recovery because of the prevalence of thin, nonfunctional fibers with relatively thick sheaths in the regenerating nerves.
Historically, internodal length has been regarded as an important determinant of MCV (Waxman 1980
). Internodal length is also roughly proportional to fiber diameter in normal fibers (Hiscoe 1947
; Vizoso 1950
). On the other hand, regenerating fibers have shorter internodes relative to normal fibers of the same diameter, and the regression line for the relationship between internodal length and fiber diameter is represented by a flatter slope (Vizoso and Young 1948
; Cragg and Thomas 1964
; Friede and Beuche 1985
; Guttuso et al. 1988
). These observations are consistent with our data. The internodal length in the regenerated fibers remained at around 300 μm, although fiber diameter increased with time. This indicates that internodal length does not increase as significantly as does diameter in regenerating fibers, and the decrease in the internodal length of regenerated fibers is not considered to alter MCV significantly. Hence, the slope of the regression lines for intermodal length between 50 and 200 days may not be considered as a sensitive morphological index of recovery in regenerated fibers.
The relationship between internodal length and MCV exhibited a peak conduction velocity over a broad quotient IL/FD range (between 100 and 200) (Brill et al. 1977
). This quotient is thought to maximize the MCV. In the present study, the regression curves of IL/FD against fiber diameter showed a similar trend at all four posttransection time points. Thus, it appears reasonable to assume that this function may be an appropriate relation to maximize the MCV in regenerating fibers. Therefore, the relationship between fiber diameter and internodal length is not a sensitive recovery index.
Thus, we concluded that MCV and mean fiber diameter were the most reliable indices of functional recovery during sciatic nerve regeneration. Furthermore, the regression relation between fiber diameter and internodal length was not a sensitive index of recovery.