The end-to-side peripheral nerve repair

Journal of Neuroscience Methods 136 (2004) 45–53

The end-to-side peripheral nerve repair Functional and morphometric study using the peroneal nerve of rats J.M.R. De Sá1 , N. Mazzer2 , C.H. Barbieri∗,3 , A.A. Barreira4 Department of Biomechanics, Medicine and Rehabilitation of the Locomotor Apparatus (Laboratory of Microsurgery), Ribeirão Preto School of Medicine, São Paulo University, Ribeirão Preto, SP 14049-900, Brazil Received 29 August 2003; received in revised form 22 December 2003; accepted 23 December 2003

Abstract Morphologic and functional recovery following an end-to-side repair was studied comparatively with conventional end-to-end repair in a model of peroneal nerve lesion in rats. Twenty-eight rats were used and divided into four groups according to the reparative procedure following nerve division: (1) nerve stumps buried into neighboring muscles (n = 8); (2) conventional end-to-end repair (n = 7); (3) end-to-side repair onto the tibial nerve (n = 8); (4) sham operation (n = 5). The sciatic functional index (SFI) was evaluated at weekly intervals for 8 weeks, the peroneal nerve being resected on the 56th day for histologic and morphometric studies. The SFI progressively improved in Groups 2 (−16.9) and 3 (−22.7), although it did not reach normal values (around −8). The average nerve fiber density increased to normal values in both Groups 2 and 3, although accompanied by a marked decrease of average minimal and maximal nerve fiber diameter, myelin sheath area and G quotient. The differences between Groups 2 and 3 or Groups 2 and 4 were not significant. We conclude that, although resulting in significant morphologic and functional recovery, end-to-side repair is not as efficient as the conventional end-to-end nerve repair. However, end-to-side repair has a potential for application in selected cases in humans. © 2004 Elsevier B.V. All rights reserved. Keywords: Peroneal nerve; End-to-side repair; End-to-end repair; Sciatic functional index; Morphometry; Rat

1. Introduction Despite the marked progress experienced in the field of peripheral nerve reconstruction over the last 30 years, correcting a wide nerve gap is still a problem for most surgeons. Actually, direct reconstruction with a conventional free or vascularized nerve graft is not always possible due to the relatively limited number of donor nerves, or it does not work well due to the distance between nerve stumps. The use of end-to-side nerve repair to treat long lasting facial nerve palsy was proposed as early as at the beginning of the last century (Ballance et al., 1903; Sherren, 1906) but functional results were very poor and this kind of nerve repair was no longer investigated nor recommended

∗

Corresponding author. Tel.: +55-16-6337559; fax: +55-16-6330336. E-mail address: [email protected] (C.H. Barbieri). 1 Orthopedic and hand surgeon, doctorate level postgraduate student. 2 Associate professor (adviser). 3 Full professor (co-adviser). 4 Associate professor (collaborator), Department of Neurology and Psychiatry, Laboratory of Neurosciences. 0165-0270/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.jneumeth.2003.12.018

by other authors until recently. Despite the initially negative experience with end-to-side nerve repair, experimental evidence that it enhances axonal growth and nerve regeneration inside the recipient stump of motor or sensory nerves has been recently produced (Belzberg and Campbell, 1998; Campbell and Belzberg, 1997; Lohman et al., 1997; Noah et al., 1997a,b; Terzis et al., 1997; Viterbo et al., 1992, 1994a,b; Zhao et al., 1997). The fact that sensory and motor collateral nerve sprouting can be induced from an intact peripheral nerve by end-to-side repair has definitely been demonstrated (Lundborg et al., 1994), thus, confirming that this technique has a potential for clinical use and deserves additional investigation. However, all previous investigations were focused on the morphologic aspect of nerve regeneration induced by the end-to-side repair, so that evidence that morphologic regeneration actually results in functional recovery, particularly in mixed nerves, is still lacking. Functional evaluation of nerve regeneration in animals is obviously difficult, but De Medinaceli et al. (1982, 1984) developed the sciatic functional index (SFI) method to evaluate hind paw function following a sciatic nerve injury and repair in rats, later

46

J.M.R. De S´a et al. / Journal of Neuroscience Methods 136 (2004) 45–53

modified by Bain et al. (1989) and by Carlton and Goldberg (1986). According to the modification by Bain and colleagues, the method requires the measurement of three parameters on the rat’s hind paw print, which are introduced into a specific mathematics formula. The SFI actually is a negative indicator of the degree of nerve dysfunction and varies from 0 to −100, with 0 corresponding to normal function and −100 indicating complete dysfunction. The SFI was demonstrated to be directly correlated with nerve fiber density and to be a reliable method to evaluate sciatic nerve functional deficiency in the rat (Bain et al., 1989; Maeda et al., 1999; Oliveira et al., 2001). The purpose of the present investigation was to study nerve regeneration following an end-to-side repair of the peroneal nerve onto the sciatic nerve trunk and to correlate morphometry and SFI data.

2. Material and methods The experiment was carried out on 28 male Wistar rats weighing on average 300–350 g. Before and after the operation the animals were kept in individual cages and fed specific rat chow and water ad libitum. All were given a single weight related subcutaneous dose of antibiotic (Penicillinprocaine, 400 000 IU) the day before the operation. The peroneal nerve was used in the study and the animals were divided into four groups according to the operative procedure: Group 1: segmentary resection (n = 8); Group 2: end-to-end repair (n = 7); Group 3: end-to-side repair (n = 8); Group 4: sham operation (n = 5). 2.1. Operative procedure Immediately before the operation, the rats were taught to walk on the walking track by repetitive trials and normal preoperative hind paw prints were then obtained, identified, and stored. The animals were then anaesthetized with a single intraperitoneal injection of sodium pentobarbiturate (Nembutal Abbott® , 60 mg/kg body weight) and placed prone on the table, with both fore and hind paws fixed with rubber bands. The lateral aspect of the right thigh, hip, and flank was then routinely prepared, including trimming off the hair and antisepsis (20% iodine alcohol solution), and the operative field was protected with sterile towels. The sciatic nerve was exposed through a posterolateral longitudinal straight incision going down from the greater trochanter to the lateral condyle of the femur, followed by blunt dissection between the gluteus maximus and quadriceps muscles. The entire length of the nerve was made visible and its three main distal branches, the common peroneal, tibial, and sural nerves, were carefully identified in the popliteal fossa (Fig. 1). The common peroneal nerve was carefully dissected under magnification with a surgical microscope (16×), trans-

versely divided 7 mm distally to its emergence from the sciatic trunk and repaired according to grouping. In Group 1 (segmentary resection), both proximal and distal nerve stumps were buried into the neighboring muscles and fixed with a single U-shaped 8/0 nylon (Ethicon® ) stitch to prevent the occurrence of any regeneration (Fig. 1B). In Group 2 (end-to-end repair), the nerve was repaired with four isolated simple epineural 11/0 nylon sutures (Ethicon® ) placed at 90◦ intervals (Fig. 1C). In Group 3 (end-to-side repair), a 1 mm-wide segment of the epineurium was resected on the lateral aspect of the tibial nerve, the distal stump of the common peroneal nerve being sutured into the defect so created with three isolated epineural 11/0 nylon sutures placed at 120◦ intervals and the proximal stump being buried into the neighboring muscles (Fig. 1D). In Group 4, the sciatic nerve and branches were simply exposed (sham operation). In all groups the wound was closed in layers, isolated simple 8/0 silk sutures being used for the muscles and isolated simple 5/0 nylon sutures for the skin. 2.2. Footprint recording and analysis Hind foot prints were obtained on paper strips impregnated with a 1% bromophenol blue acetone solution and left do dry, according to the method proposed by De Medinaceli’s group and later modified (Lowdon et al., 1988). The bromophenol blue impregnated paper becomes yellow after drying but turns immediately and permanently blue when in contact with water or any water based solution. Instead of plain water the animals’ hind paws were immersed into ordinary domestic detergent, which prevents excessive dispersion of the print. The animals were made to walk on the walking track where they left about three imprints of each paw. The paper strips containing the imprints were left to dry and copied with a high-resolution scanner. The digitized imprints were then stored and analyzed in a computer with the assistance of a graphic software package specially developed for this purpose, which allows handling of the stored imprints and automatically calculates the SFI (Oliveira et al., 2001). The parameters measured in the footprints were print length (PL, or the maximal distance between the tip of the longest toe and the heel), toe spread (TS, or the distance between the first and fifth toe), and intermediate toe spread (IT, or the distance between the second and fourth toe), both in the normal (N) and experimental (E) paw prints. The landmarks of each parameter were simply clicked with the mouse on the monitor screen according to a previously established sequence and the SFI was automatically calculated, including a graphic plot of the regeneration curve as a function of time. The formula used to calculate the SFI was that proposed by Bain and colleagues, as follows: EPL − NPL ETS − NTS + 109.5 NPL NTS EIT − NIT +13.3 − 8.8 NIT

SFI = −38.3

J.M.R. De S´a et al. / Journal of Neuroscience Methods 136 (2004) 45–53

47

Fig. 1. Operative procedure details. (A) The sciatic nerve and main branches, peroneal (1), tibial (2), and sural (3) nerves. (B) Segmentary resection of the peroneal nerve (Group 1), with the proximal and distal stumps (arrow heads) buried into the neighboring muscle. (C) Conventional end-to-end repair (Group 2) of the peroneal nerve showing the site of repair (arrow). (D) End-to-side repair (Group 3) of the peroneal nerve onto the tibial nerve showing the proximal nerve stump (arrow head) and the site of repair (arrow).

Hind paw prints were obtained and analyzed at weekly intervals from the first to the eighth postoperative week and comparisons were made between the postoperative prints from the normal and operated paw and between preoperative and postoperative data. 2.3. Histologic and morphometric studies The animals were killed on the 56th postoperative day with an intraperitoneal injection of sodium pentobarbiturate (Nembutal Abbott® , 120 mg/kg body weight) and the sciatic nerve was entirely removed through a posterolateral approach to the thigh in Groups 2 through 4. In Groups 2 and 4, the peroneal nerve was separated from the sciatic trunk, a 10 mm-long segment being obtained, with the site of end-to-end repair in the middle portion in the specimens of Group 2. In Group 3, the recipient tibial nerve was divided immediately above and below the site of the end-to-side repair and the peroneal nerve was divided 10 mm below, so that it was only the peroneal nerve that was prepared for the histologic and morphometric studies. In Group 1, a 4 mm-long segment of the distal stump of the peroneal nerve was harvested. The nerve segments were placed on a piece of filter paper (Whatman® ), identified and fixed in Karnowisky solution (2.5% glutaraldehyde, 4% paraphormaldehyde and 0.1 M sodium cacodylate buffered aqueous solution) for 12 h at room temperature, after which they were thoroughly washed with an isotonic sodium cacodylate aqueous solution and

post-fixed with 2% osmium tetroxide for 12 h at room temperature. The nerve fragments were then dehydrated in an ethyl alcohol aqueous solution of growing concentrations, from 25 to 100% at 25% increases for 1 h each, and embedded in epoxy resin (Poly Bed-812® , Polysciences Inc.) at 60 ◦ C for 72 h. Serial 5 ␮m-thick sections were cut with an ultra-microtome (MT 6000-XL, RMC Inc.) from the nerve segments, beginning from the site of repair downwards for 5 mm. A total of 1000 sections were cut from each nerve segment, 200 (1:5) of which were examined and 20 (1:10) were counted. The sections were stained with 1% toluidine blue and examined with a light microscope (Zeiss Axiophoto) equipped with a video camera linked to a microcomputer loaded with the KS 400 Measure Interactive (version 2.0) software. The first step of morphometric analysis consisted of capturing the entire fascicle image, followed by measuring the fascicle area by contouring its internal epineural edge with the mouse (magnification: objective 2.5×, optovar 1.6×, camera 0.5×). The next step consisted of capturing sequential inner areas of the fascicle (magnification: objective 100×, optovar 1.6×, camera 0.5×), with the examiner selecting those with the most adequate appearance for analysis. The selected histologic images were then converted into binary (black and white) images and cleaned of any blood vessels, degenerated nerve fibers and artifacts, following which the individual myelinated nerve fibers were manually counted and the corresponding nerve fiber density (fiber/mm2 ) was calculated. Other parameters selected for

48

J.M.R. De S´a et al. / Journal of Neuroscience Methods 136 (2004) 45–53

analysis were the minimal and maximal fiber diameter, the myelin sheath area and the G quotient (quotient between axon diameter and fiber diameter), which were automatically calculated by the computer software. An average figure was obtained for each parameter in each group, in addition to histograms of the nerve fiber diameter frequency distribution. The Kruskal–Wallis non-parametric multiple comparison test was used for statistical analysis of both SFI and nerve fiber density figures according to grouping at the 6% level of significance (P ≤ 0.06). The Mann–Whitney non-parametric test was preferred for statistical analysis of nerve fiber density according to nerve fiber diameter intervals at the 6% level of significance (P ≤ 0.06).

3. Results Both anesthetic and operative procedures were well tolerated by all animals, which also did not present any sign of infection at any time. Completely normal gait, with the right hind paw toes fully spread, was resumed around the second postoperative day in Group 4 (sham operation), while a flexion contracture of the paw (“drop foot”) and adduction of the toes remained unchanged throughout the entire experiment in Group 1 (segmentary resection). In Groups 2 (end-to-end repair) and 4 (end-to-side repair), the initial flexion contracture of the paw and adduction of the toes gradually recovered, together with the capacity to stand and walk on the paw, although completely normal gait was never regained, even during the eighth postoperative week. 3.1. Sciatic functional index A total of 252 footprints, each containing six prints, were studied. The imprints were essentially normal in Group 4 from the first to the eighth postoperative evaluation, with virtually no difference between operated on and untouched paw. In Groups 1 through 3 they were shorter and narrower than normal, and remained so only in Group 1, in which the SFI became worse from the first to the eighth evaluation. In Groups 2 and 3, the imprints gradually became longer and wider, thus, resuming a nearly normal appearance. The average preoperative SFI obtained for intact animals was never 0, but oscillated between −0.2 in Group 2 and −10 in Group 3, with no significant difference between

Fig. 2. Graphic of the SFI behavior according to grouping. The SFI remained nearly unchanged throughout the experiment in Group 4. In Groups 2 and 3, the SFI progressively improved, evolving from a moderate to a slight dysfunction, a bit worse in Group 3. In Group 1 the SFI became progressively worse with time, probably because of a flexion contracture due to unoposed flexor muscles.

groups (P = 0.23). Between the first and the eighth postoperative evaluation, the average SFI decreased from −34.57 to −53.78 in Group 1. In Groups 2 and 3, the SFI increased from −33.76 to −16.89 and from −35.67 to −22.69, respectively, thus, characterizing a remarkable functional improvement. In Group 4, the SFI remained virtually unchanged throughout the experiment although oscillating from −18.62 to −6.19 on average (Fig. 2). Statistical analysis showed that preoperative SFI values were not significantly different between groups. During the first postoperative week, there was a significant difference between Groups 1–3 and Group 4 (P = 0.02). During the fourth week, a significant difference was detected between Groups 2–4 and Group 1 (P = 0.005), but not between Groups 2 and 3 and Group 4 (P > 0.06). On the eighth week, differences remained significant between Groups 2–4 and Group 1 (P = 0.05) and became significant between Groups 3 and 4 (P < 0.001), but not between Groups 2 and 3 or between Groups 2 and 4 (Table 1). 3.2. Histologic and morphometric studies Nerve fiber counting was not done in Group 1, considering the advanced morphologic derangement observed in the distal stump of the peroneal nerve in this group. A total of about 400 peroneal nerve sections were counted in Groups

Table 1 SFI figures (mean and standard deviation) according to grouping and postoperative period Group

n

Postoperative week First

1 2 3 4

8 7 8 5

−12.68 −17.61 −27.91 −11.58

Fourth to to to to

−54.09 −51.02 −46.08 −16.46

(−34.57 (−33.76 (−35.67 (−13.62

G4 × G1–3, P = 0.02

± ± ± ±

17.98) 12.48) 5.9) 2.21)

−27.51 −11.71 −18.68 −14.01

Eighth to to to to

−56.12 (41.96 ± 11.49) −36.44 (−20.59 ± 11.47) −32.96 (−25.57 ± 4.29) −37.4 (−18.62 ± 14.57)

G1 × G2–4, P = 0.005

−34.16 −13.08 −16.89 −2.72

to to to to

−78.75 (−53.78 ± 16.64) −35.72 (−16.89 ± 10.66) −25.29 (−22.69 ± 3.63) −10.6 (−6.19 ± 8.07)

G1 × G2–4, P = 0.05; G3 × G4, P < 0.001

J.M.R. De S´a et al. / Journal of Neuroscience Methods 136 (2004) 45–53

2, 3, and 4 (20 sections per animal, 20 animals). The nerve’s internal morphology during the eighth postoperative week was entirely normal in Group 4, with the typical myelinated fibers of several diameters distributed into a single fascicle. In Groups 2 and 3, the histological appearance was that of an ongoing regeneration process, with predominance of small diameter myelinated fibers with a proportionally thin myelin sheath, organized in small diameter fascicles; nerve fibers with axonal atrophy were occasionally seen. In Group 1, the appearance was that of a typical Wallerian degeneration, with advanced axonal loss, complete absence of the myelin sheath and occasional endoneurial large vacuoles and macrophages phagocytizing degenerated myelin (Fig. 3). In view of the advanced degree of axonal degeneration, morphometric evaluation was not carried out in Group 1. In Group 2, the average myelinated nerve fiber density was 13 397 fiber/mm2 (range: 9203–25 335 fiber/mm2 ) and the average minimal fiber diameter was 1.67 ␮m, with a left shift in the distribution according to fiber diameter, thus, showing a trend to a uni-modal and asymmetric regeneration pattern. The average myelin sheath area was 35.59 ␮m2 , the average maximal fiber diameter was 10.37 ␮m and the average G quotient was 0.55. The G quotient distribution according to fiber density showed a slight left shift, indicating a trend to axonal degeneration.

49

In Group 3, the average myelinated fiber density was 10 924 fiber/mm2 (range: 5923–19 705 fiber/mm2 ) and the average minimal fiber diameter was 1.51 ␮m, also with a left shift in the distribution according to fiber diameter and the same trend to a uni-modal and asymmetric regeneration pattern. The average myelin sheath area was 33.92 ␮m2 , the average maximal fiber diameter was 7.42 ␮m and the average G quotient was 0.53, with a slight left shift in the distribution according to fiber density and a trend to axonal degeneration similar to that observed in Group 2. In Group 4, the average myelinated fiber density was 9966 fibers/mm2 (range: 8953–11 568 fibers/mm2 ) and the average minimal fiber diameter was 1.68 ␮m, with a normal distribution according to fiber diameter. The average myelin sheath area was 58.98 ␮m2 , the average maximal fiber diameter was 13.12 ␮m and the average G quotient was 0.58, also with a normal distribution according to fiber density (Figs. 4 and 5). Statistical analysis of the average myelinated nerve fiber densities showed significant difference between Groups 2 and 3 and Group 4 (P = 0.01) and between Group 2 and Groups 3 and 4 (P = 0.01), but not between Group 3 and Group 4 (P > 0.06). For the average maximal and minimal nerve fiber diameter and average myelin sheath area, the differences between Groups 2 and 3 and Group 4 were

Fig. 3. Severe wallearian degeneration in the distal stump of the peroneal nerve in Group 1 (A). Many small diameter myelinated and unmyelinated fibers, and fibers undergoing wallerian degeneration in Groups 2 (B) and 3 (C); although nerve fiber density is similar in both groups, fiber diameter is obviously smaller in Group 3. Normal appearance of the peroneal nerve fibers, with small and large diameter myelinated fibers, in Group 4 (D) (1% toluidine blue; scale bar 10 ␮m (A–C) and 15 ␮m (D)).

50

J.M.R. De S´a et al. / Journal of Neuroscience Methods 136 (2004) 45–53

Fig. 4. Graphic of the nerve fiber density distribution according to fiber diameter in Groups 2 through 4. A high concentration of small diameter fibers and a peak at 2 ␮m are observed in Groups 2 and 3, indicating a uni-modal and asymmetric regeneration pattern. Density of fibers above 5 ␮m of diameter is virtually irrelevant in Group 3. Group 4 shows the usual equanimous distribution.

Fig. 5. Graphic of the G quotient distribution according to nerve fiber density in Groups 3 through 4. Groups 2 and 3 show a slight trend to a left shift with a small step at 0.35 and a peak at 0.6, while Group 4 shows the usual normal distribution with a peak at 0.6. Although the slight left shift indicates a trend to axonal degeneration, the higher peak at 0.6 observed in Group 2 compared to Group 3 indicates a more advanced myelin regeneration process.

significant (P = 0.002, P = 0.001, and P = 0.001, respectively). For the G quotient, there was a significant difference between Group 3 and Group 4 (P = 0.03), but not between Group 2 and Groups 3 and 4 (P > 0.06) (Table 2).

4. Discussion Repairing a peripheral nerve affected by a wide tissue loss may be a problem for the surgeon involved in its treatment. For cases in which well-established conventional or vascularized nerve grafting procedures cannot be done or have not worked satisfactorily there is virtually nothing else to be done and both surgeon and patient may have to accept a permanent undesirable sequel. The end-to-side nerve repair, introduced long before nerve grafting procedures were known as they are nowadays and studied in this investigation, may be an adequate solution for these cases, with the

advantage that it does not inflict any additional damage or sacrifice any donor nerve. Previous investigators have already proved that morphologic regeneration occurs inside the distal stump of the damaged nerve following an end-to-side nerve repair (Belzberg and Campbell, 1998; Campbell and Belzberg, 1997; Lohman et al., 1997; Lundborg et al., 1994; Noah et al., 1997a,b; Terzis et al., 1997; Viterbo et al., 1992, 1994a,b; Zhao et al., 1997), but evidence that axonal growth results in acceptable function was still lacking. Actually, all previous studies focused mostly on the morphologic aspect of nerve regeneration so that functional recovery could not occur, since no specific sensory or motor nerve fiber alignment is possible with this procedure and axonal growth apparently occurs at random, “perhaps induced by factors emanating from the attached nerve segment, and subsequently making functional peripheral connections” (Lundborg et al., 1994). The peroneal nerve pro tibial nerve end-to-side repair was preferred instead of the contrary because the tibial nerve is about twice as thick as the peroneal nerve, causing the end-to-side suture of the tibial nerve distal stump onto the lateral aspect of the peroneal nerve to be a technical problem. Furthermore, the tibial nerve contains about twice as many fibers, thus, being a better donor than a recipient nerve, and for this reason it was not used as a recipient nerve in the present model. Such a discrepancy in nerve fiber contents between the donor and recipient nerve might induce an inadequate regeneration rate and cause misinterpretation of the results. Whether an epineural window is open (Viterbo et al., 1994a,b) or not (Viterbo et al., 1992) at the site of coaptation of the distal end of the recipient nerve does not seem to influence the final outcome, since morphologic regeneration was observed in both cases, which “validate(s) that the regenerating fibers after terminolateral neurorraphy (TLN) have the ability to penetrate the endoneurium, and perineurium” (Zhao et al., 1997). However, a larger amount of axons was found in the recipient nerve when an epineural window was opened with the resection of both epineurium and perineurium or a one-third transverse partial neurectomy was made in the donor nerve (Noah et al., 1997a,b). It was our option to open a 1 mm-wide window in the epineurium, considering that this is the most frequently used procedure and that the procedure with resection of a transverse segment of the donor nerve is not a real end-to-side repair. The proximal stump of the peroneal nerve was buried into the neighboring muscles in order to minimize the risk that regenerating fibers could bridge the gap and reach the 10 mm-distant distal stump sutured directly onto the lateral aspect of the tibial nerve. According to our results, a quite satisfactory morphologic regeneration occurred 8 weeks after an end-to-side nerve repair and was accompanied by progressive functional recovery, although not yet complete at that time. The morphometric parameters used were more favorable in the nerves with an end-to-end repair (Group 2), with a

Group

n

Density (fiber/mm2 )

Maximum diameter (␮m)

Minimum diameter (␮m)

Myelinated sheath area (␮m2 )

G quotient

2 3 4

7 8 5

9203–25 335 (13 397 ± 5480) 5923–19 705 (10 924 ± 4453) 8953–11 568 (9966 ± 1014)

7.92–15.79 (10.37 ± 2.8) 6.81–9.51 (7.42 ± 0.99) 12.10–14.01 (13.12 ± 0.81)

1.47–2.12 (1.67 ± 0.26) 1.31–1.81 (1.51 ± 0.21) 1.60–1.99 (1.68 ± 0.19)

14.96–85.81 (35.59 ± 23.86) 23.11–52.97 (33.92 ± 10.39) 45.01–75.53 (58.98 ± 12.46)

0.54–0.57 (0.55 ± 0.01) 0.51–0.55 (0.53 ± 0.02) 0.55–0.59 (0.58 ± 0.02)

G2–3 × G4, P = 0.01; G2 × G3–4, P = 0.01

G2–3 × G4, P = 0.002

G2–3 × G4, P = 0.001

G2–3 × G4, P = 0.001

G3 × G4, P = 0.03

J.M.R. De S´a et al. / Journal of Neuroscience Methods 136 (2004) 45–53

Table 2 Range, average, and standard deviation of the morphometric parameters evaluated, according to grouping

51

52

J.M.R. De S´a et al. / Journal of Neuroscience Methods 136 (2004) 45–53

nerve fiber density about 60% higher (13 379 fiber/mm2 ) than in nerves with an end-to-side repair (Group 3, with 10 924 fiber/mm2 ) and about 80% higher than normal (Group 4, with 9966 fiber/mm2 ). Also, the average minimal nerve fiber diameter was nearly normal in Group 2 (1.67 ␮m) compared to Group 4 (1.68 ␮m), while in Group 3 it was much lower (1.51 ␮m). The minimal nerve fiber diameter is probably the most reliable parameter of nerve regeneration, together with the myelin sheath area, which was 35.39, 33.92, and 58.98 ␮m2 , respectively, in Groups 2, 3, and 4. It is our opinion that the obvious morphologic regeneration observed in Group 3 can only be the result of nerve sprouting from the donor towards the recipient nerve, probably in response to neurotrophins such as nerve growth factors (NGF), insulin-like growth factors (IGF-I, IGF-II), and fibroblast growth factor (FGF), “released from the distal stump of an injured nerve, incorporated by the axons of the donor nerve and transported back to the cell body in the anterior horn by retrograde axoplasm transport” (Noah et al., 1997a,b). However, we would not rule out a local action of the neurotrophins, very much in the same way as bone morphogenetic protein (BMP), released by necrotic bone undergoing resorption, stimulates bone growth and regeneration. From a theoretical standpoint nerve sprouting causes a single proximal nerve fiber to become responsible for two or more distal nerve fibers (the original one of the donor nerve and at least a new one in the recipient nerve), thus, increasing the size and probably decreasing the efficiency of motor and sensitive units, with a possible negative influence on function. Furthermore, under normal conditions the muscles innervated by the peroneal nerve are opposed by those innervated by the tibial nerve and perfect function is possible due to the contraction–relaxation mechanism of antagonistic muscles. With end-to-side repair, only the tibial nerve will innervate antagonistic muscles, so that the contraction–relaxation mechanism will be lost, also with a negative influence on function. However, function was surprisingly good in our Group 3, a fact that is still beyond our understanding at the present stage of knowledge. Low G quotient values (around 0.4) indicate axonal degeneration and high values (around 0.7) indicate either myelin degeneration or regeneration. According to our results, the G quotient presented an entirely normal distribution with a single peak at 0.6 in Group 4. In Groups 2 and 3, the G quotient showed a slight trend to a bi-modal distribution, with a high peak at 0.6 and a step around 0.35 and 0.4 (Fig. 5). Such events were interpreted as myelin regeneration, which was obviously less evident in Group 3 and this theoretically may also have a negative effect on function. Despite the apparently worse morphologic regeneration in Group 3, functional evaluation showed that the SFI improved almost similarly with time in both Groups 2 and 3, so that there was no significant difference in SFI figures between these groups at the final evaluation 8 weeks after the operation (−16.9 and −22.7, respectively), although in both cases they were much below normal (−6.2). Such a

discrepancy between SFI and morphometry data is difficult to explain but may indicate either that the SFI method is still defective and needs some improvement or that it is not entirely suitable for evaluation of the peroneal nerve. Actually, the SFI method was developed and is more suitable for evaluation of a complete lesion of the sciatic nerve which produces a long and narrow footprint, particularly due to the lack of both plantar flexion of the ankle and spread of the toes, while a dysfunction of the peroneal nerve produces a short and narrow footprint due to the lack of dorsal flexion of the ankle and extension of the toes. The method developed by Bain et al. (1989) also includes a peroneal functional index (PFI) and a tibial functional index (TFI) based on the same parameters measured for the SFI but with different correction factors. Both PFI and TFI were automatically calculated together with the SFI by the software used in this investigation but it was our option to refer only to the SFI figures, since there was no significant difference between the two indices and the SFI is a more widely known parameter. Neither clinical inspection nor SFI evaluation showed any evidence that the end-to-side repair might have impaired function of the tibial nerve as described earlier (Lundborg et al., 1994). Since the tibial component of the sciatic nerve was normal, partial weight bearing was resumed as early as during the first week in a nearly normal fashion. The animal barely touched the ground on the first days but improved over the next weeks, with the imprint acquiring a nearly normal appearance during the eighth week. This fact may well indicate that a reasonable functional recovery is possible with a still incomplete morphologic regeneration of the peroneal nerve. Although many questions regarding the end-to-side nerve repair still remain unanswered, we conclude that the technique resulted in a reliable degree of morphologic regeneration and functional recovery for the peroneal nerve. It certainly has a potential for application in humans, with the advantage that it does not require a nerve graft to be performed for most of the situations that we can imagine and, as far as we could detect, does not down-grade the donor nerve function. Acknowledgements The authors acknowledge Professor Amilton Antunes Barreira, from the Department of Neurology and Psychiatry for his kind permission to use the Laboratory of Neurosciences facilities and equipments, and Mr. Antonio Renato Meirelles e Silva and Ms. Maria Cristina Lopes Schiavoni for their kind technical assistance. References Bain JR, Mackinnon SE, Hunter DA. Functional evaluation of complete sciatic, peroneal, and posterior tibial nerve lesions in the rat. Plast Reconstr Surg 1989;83(1):129–36.

J.M.R. De S´a et al. / Journal of Neuroscience Methods 136 (2004) 45–53 Ballance CA, Balance HA, Stewart P. Operative treatment of chronic facial palsy of peripheral origin. Br Med J 1903;2:1009–13. Belzberg AJ, Campbell JN. Evidence for end-to-side sensory nerve regeneration in a human. Case report. J Neurosurg 1998;89:1055–7. Campbell JN, Belzberg AJ. Nerve fibers in the proximal end of the distal segment of a previously severed nerve in man: evidence for side-to-side nerve sprouting. J Reconstr Microsurg 1997;13(2):137 [ASPN Abstracts]. Carlton JM, Goldberg NH. Quantitating integrated muscle function following reinnervation. Surg Forum 1986;37:611–2. DeMedinaceli L, Freed WJ, Wyatt RJ. An index of the functional condition of rat sciatic nerve based on measurements made from walking tracks. Exp Neurol 1982;77:634–43. De Medinaceli L, Derenzo E, Wyatt RJ. Rat sciatic functional index data management system with digitized input. Comput Biomed Res 1984;17:185–92. Lohman R, Bullock F, McNaughton T, Siemionow M. End-to-end vs end-to-side neurorrhaphy. J Reconstr Microsurg 1997;13(2):135 [ASPN Abstracts]. Lowdon IMR, Seaber AV, Urbaniak JR. An improved method of recording rat tracks for measurement of the sciatic functional index of De Medinaceli. J Neurosci Methods 1988;24:279–81. Lundborg G, Zhao Q, Kanje M, Danielsen N, Kerns JM. Can sensory and motor collateral sprouting be induced from intact peripheral nerve by end-to-side anastomosis? J Hand Surg [Br] 1994;19(3):277–82. Maeda T, Hori S, Sasaki S, Maruo S. Effects of tension at the site of coaptation on recovery of sciatic nerve function after neurorrhaphy: evaluation by walking-track measurement, electrophysiology, histomorphometry and electron probe X-ray microanalysis. Microsurgery 1999;19(4):200–7.

53

Noah EM, Williams A, Fortes W, Terzis JK. A new animal model to investigate axonal sprouting after end-to-side neurorraphy. J Reconstr Microsurg 1997a;13(5):317–25. Noah EM, Williams A, Jorgenson C, Skoulis TG, Terzis JK. End-to-side neurorrhaphy: a histologic and morphometric study of axonal sprouting into an end-to-side nerve graft. J Reconstr Microsurg 1997b;13(9):99– 106. Oliveira EF, Mazzer N, Barbieri CH, Selli M. Correlation between functonal index and morphometry to evaluate recovery of the rat sciatic nerve following crush injury: experimental study. J Reconstr Microsurg 2001;17(1):69–75. Sherren J. Some points in the surgery of the peripheral nerves. Edinb Med J 1906;20:297–332. Terzis JK, Fortes WM, Liuzzi F, Noah M. End-to-side neurorrhaphy: evaluation of nerve growth factor elaboration using cRNA probes (in situ hybridization). J Reconstr Microsurg 1997;13(2):136 [ASPN Abstracts]. Viterbo F, Trindade JC, Hoshino K, Mizzen Neto A. Latero-terminal neurorrhaphy without removal of the epineural sheath. Experimental study in rats. Rev Paul Med 1992;110(6):267–75. Viterbo F, Trindade JC, Hoshino K, Mizzen A. Two end-to-side neurorraphies and nerve graft with removal of the epineural sheath: experimental study in rats. Br J Plast Surg 1994a;47:75–80. Viterbo F, Trindade JC, Hoshino K, Mizzen Neto A. End-to-side neurorrhaphy with removal of the epineural sheath: an experimental study in rats. Plast Reconstr Surg 1994b;94(7):1038–47. Zhao J, Chen Z, Chen T. Nerve regeneration after termolateral neurorrhaphy: experimental study in rats. J Reconstr Microsurg 1997;13(1):31– 7.