Abstracts

January 11-12, 2003

Kauai'i, Hawaii

Functional Recovery after Hand Transplantation-Results at Three-Year Follow-up. Warren C. Breidenbach (Hand and Microsurgery, Christine M. Kleinert Institute, Louisville, KY), Vijay S. Gorantla, Nestor R. Gonzalez, Diane J. Pidwell, and Darla K. Granger.

Nerve regeneration and functional recovery are the ultimate goals of composite tissue allografts such as hand transplants. Published and presented reports indicate that 13 hands have been transplanted to 10 recipients around the world. The first Louisville recipient currently has the longest surviving hand transplant in the world at 3 years after surgery.

These authors presented the results of functional testing of this latter patient at a 3-year post-transplantation follow-up. There was clinical evidence of motor return in the transplanted intrinsics, as confirmed by electrophysiologic testing: adductor pollicis (3/5); abductor pollicis (2/5); first dorsal interosseous (1/5); and lumbricals 3-4FDM (2/5). Wrist dorsiflexion, palmar flexion, radial and ulnar deviation were 4+/5. Grip strength at position II on the Jamar dynamometer was 12 lbs. Lateral pinch was 4.5 lbs and two-point pinch was 1 lb. Evaluation of integrated upper extremity and global hand function by the Carroll test revealed a score of 61/99.

The patient can perform such activities as tying shoelaces, dressing, turning book and newspaper pages, writing, throwing a ball, picking up checkers or poker chips, and can use both his hands in the activities of daily living and in his occupation. Physical therapy initially included electrical stimulation, anti-claw splinting, resistance exercises, and dynamic splinting. He now reports once every 2 weeks for stretching, and strengthening and resistance exercises. He no longer requires splint wear, and has shown rapid progression of Tinel's sign, with return of temperature, and pain and pressure sensation to the hand and fingers. He can localize touch to the thumb, ring, and small fingers; long and index still reversed (can localize on the long except for the tip). Perception when touched demonstrates different grades of tingling and pressure. Semmes-Weinstein monofilament scores at the tips of the fingers: 3.61 in the index, middle, and ring fingers; 3.22 in the thumb; and 3.84 in the little finger. Results of testing showed the most improvement in the thumb since last testing in September 2001 (3.84 to 3.22). There have been no episodes of ulceration. Static two-point discrimination is greater than 15 mm. There was no evidence of retardation of nerve regeneration, as measured by advancing Tinel's sign, during the rejection episodes.

The initial trial period of experimental human hand transplantation has proved successful, far beyond expectations. At the 3-year time point, function after hand transplantation has mirrored that achieved 3 or more years after forearm replantation, and is superior to that after prostheses.

Neuromuscular Reconstruction in Severe Brachial Plexus Lesions. Pietro Giovanoli (Department of Surgery, University of Vienna, Division of Plastic and Reconstructive Surgery, Vienna, Austria), Oskar Aszmann, and Manfred Frey.

In failed neurotization procedures or long-standing brachial plexus lesions, functioning free muscle transplants are necessary to restore function, especially below the elbow. In the past 5 years, the authors have used a free functioning muscle graft in nine cases, for neuromuscular reconstruction after irreparable brachial plexus injury or failed extraplexual neurotization.

Functional restoration is best accomplished by extraplexual neurotization. In limited cases, in which no donor nerve or nerve transfer is available for direct end-to-end suture, intraplexual neurotization by end-to-side neurorrhaphy was performed. Intraoperative electrical stimulation (SEP and evoked muscle response) was routinely used to help clarify the extent of injury. A functioning free gracilis muscle graft was innervated by intercostal nerve transfer in six cases, and by end-to-side neurorrhaphy in another three cases. In five cases, the goal of reconstruction was restoration of elbow flexion; in four cases, wrist and finger flexion. Ages of patients ranged from 10 to 46 years; the mean follow-up was 18.3 months.

The nine functioning muscle grafts demonstrated (with the exception of one case) good clinical results, with a minimum of M3 for elbow flexion and finger flexion 16 months postoperatively (range: M3 to M4). All patients developed reinnervation through the end-to-side coaptation site. Axonal regeneration could be documented electrophysiologically and later on, clinically. No downgrading of the donor nerve (or root) or muscle function occurred in all clinical cases.

To gain a prospect for success or improved results in brachial plexus reconstruction, neuromuscular reconstruction using functioning free muscle transplants or new additional techniques, such as end-to-side neurorrhaphy, have been introduced over the last few years. Over-dimensioning seems to be useful in functional muscle transplantation, and also in limited neural capacity situations. Synergic donor nerves should be preferred in end-to-side nerve repair.

Comparative Study of Results after Gracilis Muscle Transplant for Facial Animation in Children with the Innervated Motor Nerve: Cross-Facial Nerve Graft vs. Masseteric Nerve. Yong-Chan Bae (Department of Plastic and Reconstructive Surgery, Pusan National University, South Korea), Ronald M. Zuker, Ralph Manktelow, and Shawna Wade.

The microneurovascular transfer of a free muscle transplant is the procedure of choice for facial animation in patients with facial paralysis. One of the critical factors in this procedure is the selection of a motor nerve to innervate the transplanted muscle.

From 1989 to 1999, 165 free segmental gracilis muscle transfers were performed in 120 children for facial animation. The cross-facial nerve graft (CFNG) was used in 70 procedures; the masseteric nerve was used in 94 procedures (50 patients, masseteric group). To compare the operative procedures between the two groups, all charts were reviewed. The extent of oral commissure movement was determined by the measurements taken from the tragion to the commissure corner, both at rest and with full smile. In the masseteric group (n = 16), the extents were measured at the normal and operated sides; in the CFNG group (n = 20), these were measured at the left and right sides. No statistically significant difference was found between the two groups (p < 0.05) in the mean age at muscle transplantation, total operation time for muscle transplantation, and length and fraction section area of the used gracilis muscle. Although the operative variables were similar between the two groups, muscle excursion on the operated side in the CFNG group was less than that on the right (p = 0.0006) or left side (p = 0.0000) in the masseteric group, as well as that on the normal side (p = 0.0000). Concerning excursion, there was no significant difference between the left and right sides within the masseteric group (p < 0.05).

These results may be evidence that segmental gracilis muscle transplantation using the masseteric nerve for facial animation in children is a very reproducible procedure. The extent of oral commissure movement in the masseteric group was the same as that of the normal side in the CFNG group (p = 0.35, 0.61).

Aberrant Reinnervation in Obstetric Brachial Plexus Palsy. David Chwei-Chin Chuang (Plastic and Reconstructive Surgery, Chang Gung Hospital, Taipei, Taiwan).

Aberrant reinnervation (misdirection of regenerated axons) is more frequently seen in the shoulder and elbow, and less in the forearm and hand in obstetric brachial plexus palsy. The main reasons for aberrant reinnervation are dominant rupture of the upper trunk, short gap, and natural conduits by ruptured scalene muscle fibers.

One hundred babies with palsies were surveyed, based on intraoperative findings, to examine the relationship between C5, C6 spinal nerves and the quality of scar tissue that contributes the sources of aberrant reinnervation. There were eight types of lesion encountered: type 1 - C5 major + C6 minor with upper trunk lesion and acceptable scars; type 2A - C5 major and C6 minor with upper trunk lesion but worse scar on C6; type 2B - C5 major + C6 minor with upper trunk lesion with C5 rupture and C6 avulsion/rupture; type 2C - C5 major + C6 minor with upper trunk lesion with C5 rupture and C6 true avulsion; type 3A - C5 minor + C6 major with upper trunk lesion with worse scar on C5; type 3B - C5 minor + C6 major with upper trunk lesion with C5 avulsion/rupture; type 3C - C5 minor + C6 major with upper trunk lesion with C5 true avulsion; type 4 - C5, C6 minor with C5, C6 avulsion. Here, major means dominant, and minor means less in source contribution for reinnevation.

Results showed 47 cases (47%) of type 1; 26 cases (26%) of type 2A; three cases (3%) of type 2B; 16 cases (16%) of type 2C; one case (1%) of type 3A; no cases of type 3B; two cases (2%) of type 3C; five cases (5%) of type 4. In more than 90% (92%) of the pediatric palsy patients, shoulder function was recovered by aberrant reinnervation. Half came from C5 and C6 major contributions; the other half from C5 major and C6 minor contributions. In only 5% of patients was shoulder function of the paralytic type. Another new clinical finding was the “umbrella” sign, in addition to the trumpet sign. This umbrella sign is apparent during shoulder elevation: the elbow will flex spontaneously, similar to holding an umbrella. Severe umbrella sign, defined during shoulder abduction less than 90 degrees with elbow flexion, supposedly showed the main sources of C5 major and C6 major, or C5 minor and C6 minor growing to the anterior division of the upper trunk. This appears to be the first report evaluating aberrant reinnervation in obstetric brachial plexus palsy.

A Tissue-Engineered Conduit Seeded with Schwann Cells Promotes Peripheral Nerve Regeneratiion. Yasushi Nakao (Department of Orthopaedic Surgery, University School of Medicine, Tokyo) and Yoshito Ikada.

Autogenous donor-nerve grafts facilitate nerve regeneration; however, the source of the material is limited, and harvest of these grafts results in donor-site morbidity. Many types of artificial conduits have been studied as potential alternatives to short nerve grafts, but long nerve gaps have not been successfully reconstructed due a lack of growth potential.

Tissue engineering offers the potential of creating replacement structures from autologous cells and biodegradable polymer conduits. Since these contain living cells, the structures have the potential to grow and induce regeneration. The objectives of the reported study were to create Schwann cell-seeded biodegradable nerve guides, and to assess in vivo nerve regeneration through these guides.

A novel, biodegradable polymer conduit (2 mm in diameter, 0.5 mm in thickness, 25 mm in length) was composed of a polycaprolactone-polylactic acid copolymer reinforced with woven polyglycolic acid. First, Schwann cells were isolated from Wistar-rat sciatic nerve, and expanded in vitro by 8 weeks. Second, these cells were seeded into the polymer conduit via a low-pressure injection and centrifuge technique. Two weeks after seeding, Schwann cells culturing in a porous conduit wall were observed, using the electron microscope. Third, the Schwann cell-seeded conduits were used to bridge 20-mm defects in the rat sciatic nerve (n = 8). Polymer conduits lacking Schwann cells (n = 8) and silicone conduits (n = 8) were implanted into other sets of animals as controls. After 8 weeks, the sciatic nerve, including the conduit, was harvested and fixed by glutaraldehyde. At the mid-portion of the conduit, cross sections were cut and stained with toluidine blue, to evaluate the axonal contents.

Electron microscopy prior to grafting demonstrated numerous numbers of Schwann cells culturing in a porous conduit wall in a three-dimensional manner. Histologic assessment under light micrscopy at 8 weeks demonstrated excellent regeneration, with well-myelinated fibers in the bridging tissue through the Schwann cell-seeded polymer conduits. In the untreated polymer conduits, four animals showed bridging tissue, but nerve regeneration was poor. There was no regeneration through the silicone conduits. Under morphologic assessment, there was a statistically significant difference between the Schwann cell-seeded and untreated conduits in the total number of myelinated fibers (4125 ± 581 and 1220 ± 268, respectively, p < 0.01) and fiber density (fiber number /mm²).

F24554 biodegradable polymer materials and adherent living Schwann cells can facilitate nerve regeneration in reconstruction over longer nerve gaps.

Reinnervated Muscle Function is Related to Muscle-Fiber Denervation in the Rat. Melanie G. Urbanchek (Section of Plastic Surgery, University of Michigan, Ann Arbor, MI), Jack H. Van der Meulen, Mustafa Asim Aydin, Paul S. Cederna, Tomoaki Eguchi, and William M. Kuzon, Jr.

Reinnervated skeletal muscle generates lower maximal force (P_o). As a result of contraction-induced injury, skeletal muscle has an immediately reduced force capacity that is reportedly fully recovered by 60 days. It is not known whether force deficits following contraction-induced injury are attributable to muscle atrophy and muscle-fiber denervation after recovery from contraction-induced injury.

The specific hypothesis tested was that denervated/reinnervated muscle-force deficits, following contraction-induced injury, are attributable to muscle atrophy and muscle fiber denervation.

Adult rat peroneal nerve was either sham-operated (SHAM) or divided and repaired with all (NR-non-reduced) or a reduced number (DR-drastically reduced) of proximal axon stumps included in the repair. Following a recovery period of 4 months, extensor digitorum longus muscle P_o was measured in situ at optimal muscle length by supramaximal stimulation of the peroneal nerve. For some rats, this was the endpoint (no injury subgroup). For others, muscle testing continued with an EDL contraction-induced injury protocol (injury subgroup) consisting of 225 muscle-lengthening contractions at 20% strain. Po was re-measured at 60 days. Muscle cross sections were labeled to identify neural cell adhesion molecule, a marker for cell denervation, and fiber denervation was quantified.

For reinnervated muscle groups exposed to no injury, both the NR (3261 ± 636) and DR (2590 ± 975) groups demonstrated force deficits (mN), compared with the SHAM muscle group (3994 ± 346). When the no-injury subgroup force was normalized to the muscle cross-sectional area (CSA) (sP_o), the DR group continued to show an sPo deficit (SHAM 283 ± 30, NR 267 ± 43, DR 233 ± 60) (kN*m²). However, when the sP_o was corrected to eliminate the contribution on denervated fibers, there was no longer a significant force deficit (SHAM 293 ± 31, NR 292 ± 50, DR 280 ± 88) (kN*m²). At 60 days post-inury, the DR group denervated/reinnervated muscles had a lower P_o (2522 ± 540 mN), compared with SHAM (3608 ± 485 mN) and NR (2767 ± 531 mN) group muscles. Force normalized to muscle CSA (sPo) for the SHAM, NR, and DR groups was 260 ± 86, 221 ± 52, and 208 ± 35 kN*m², respectively. When the injured muscle sP_o was corrected to eliminate denervated fiber contribution, the corrected sPo (kN*m²) for each group was 210 ± 31 (SHAM), 231 ± 11 (NR), and 239 ± 26 (DR).

These preliminary data find force differences following contraction-induced injury attributable to muscle atrophy but not to muscle-fiber denervation.

Axon Regeneration and Functional Recovery Following Rat Cavernosal Nerve Repair Using a Collagen Nerve Guide. S.J. Archibald (Integra NeuroSciences, Plainsboro, NJ) and M.H. Spilker.

Impotence resulting from injury to the cavernosal nerve is one of the most devastating complications of radical prostatectomy, and the fear of impotence remains one of the main causes for the unpopularity of this treatment for prostate cancer. Although the nerve-sparing technique has reduced the overall impotence rate, following radical prostatectomy, there is still a need for approaches that may restore erectile function through nerve regeneration. Recent clinical studies suggest that nerve grafting may promote such regeneration. Bridging a nerve deficit with a tubular prosthesis offers an attractive alternative to nerve-graft repair, since it does not require harvest of a donor nerve.

The objective of the current experimental study was to test the efficacy of a collagen nerve guide conduit to support and maintain axon regeneration and functional recovery of the autonomic nervous system responsible for erectile function, following bilateral repair of 5-mm deficits in the rat cavernosal nerve.

The experimental model consisted of bilateral cavernosal nerve injury in young adult Sprague-Dawley rats. Three experimental groups were used: nerve crush, nerve transection and repair with a polyethylene tube, and transection and repair with a collagen nerve guide. Axon regeneration and function recovery were assessed at 28, 46, and 280 days post-injury, using three complimentary methods, including electrophysiologic investigation of the engorgement of the corpus cavernosum, scoring based on behavioral observation, and histologic examination of the regenerated nerves.

After 280 days, the intra-cavernosa pressure evoked by stimulation of the cavernosal nerves repaired using the collagen nerve guide (23.6 ± 4.9 mmHg, n = 6) returned to levels statistically equivalent to the nerve crush procedure (30.2 ± 6.4. n = 6), and were not significantly different from the normal baseline controls (40.7 ± 5.9, n = 8). The results for the polyethylene tube were significantly inferior (2.6 ± 1.2, n = 5). Scoring based on behavioral observation of sexual activity indicated that the nerve crush and collagen nerve guide repair groups were superior to the polyethylene repair group, but inferior to normal. Histologic examination verified that a population of myelinated and unmyelinated axons were present in all repair groups. The regenerated axon population included a great number of very small axons, which were not present in the normal controls.

The results of this study demonstrate that the collagen nerve guide supports physiologic recovery of erectile function following bilateral cavernosal nerve transection in the rat, and suggests that a collagen nerve guide has potential for use in clinical repair of cavernosal nerve following radical prostatectomy.

Differential Gene Expression in Human Peripheral Nerve and Neuromas. Henry Y.H. Siu (Division of Neurosurgery, University of Toronto, Toronto, Ontario, Canada), Todd G. Mainprize, Michael D. Taylor, Catherine A. Munro, and Rajiv Midha.

Neuromas are the result of frustrated regeneration of axons in the PNS after injury to the nerve. Painful neuromas and neuromas in continuity that are not capable of successful regeneration are traditionally treated via surgical resection. The advent of micro-array technology has allowed for large-scale simultaneous analysis of gene expression between two different tissue samples. The long-term goal of this project is to ascertain genes that are differentially expressed in human neuroma tissue. Such genes may include axonal guidance genes (e.g., LAR and Shh), and neuronal markers (e.g., S100 and Krox20). In addition to these genes, novel genes are expected to be discovered as well. How these novel genes are involved in axonal regeneration, neuroma formation, and pain generation, has the potential for leading to future studies of specific genes.

All tissue samples were collected at the time of surgery and immediately flash-frozen in liquid nitrogen. Samples were then stored at -80° C until the RNA was extracted. RNA from human peripheral nerve (n = 5) and neuroma tissue (n = 8) was extracted via the TRIzol method. The RNA was amplified and fluorescently labeled via RT-PCR. After overnight hybridization to a 19K micro-array chip, confocal laser microscopy and computer analysis yielded a list of differentially expressed genes.

Preliminary results showed that there are differentially expressed genes between the two tissue types. However, due to the vast amount of information obtained, definitive results are not yet available. Ongoing analysis will allow precise determination of the differentially expressed genes and gene families; this information was presented at the meeting.

Immunohistochemistry will provide protein confirmation of the differentially expressed genes. Additional molecular analysis will facilitate the development of possible genetic hypotheses aimed at explaining neuroma formation and maintenance.

Effect of ALS and CsA Protocol on Sciatic Nerve Allograft Regeneration in a Rat Model. Joseph Scharpf (Cleveland Clinic Foundation, Plastic Surgery, Cleveland, OH) and Maria Siemionow.

The objective of this study was the investigation of the potential use of nerve allograft, using a short-term combined protocol of anti-lymphocyte serum (ALS) and cyclosporine (CsA) to improve nerve regeneration.

A rat sciatic nerve model was used. Fifteen male Lewis and 15 Lewis-Brown-Norway (LBN) rats were randomly divided into three groups. In Group 1, control Lewis rats underwent the creation of a 1.5-cm deficit with conventional epineurial repair with allografts from LBN rats without treatment. In Group 2, Lewis rats had a 1.5-cm deficit with LBN allograft repair and treatment with CsA for 3 weeks. In Group 3, Lewis rats sustained 1.5-cm deficits with LBN allograft repair and treatment with ALS and CsA for 3 weeks.

Nerve regeneration was evaluated at 6 and 12 weeks by somatosensory evoked potentials (SSEP) and standardized pin-prick and toe-spread tests. Nerve samples were harvested at 12 weeks and were stained with toluidine blue to assess the total number of myelinated axons, axon area, and myelin-sheath thickness. Each animal served as its own control, to derive a ratio of operated vs. normal outcome for SSEP and muscle weights.

No significant difference was found in somatosensory evoked potentials between the SSEP ratios of operated vs. non-operated legs among the three groups, using Dunn's multiple comparison procedure with a Bonferroni correction. There was a significant (p < 0.05) median decrease in muscle denervation atrophy of non-treatment animals, compared to CsA-treated animals, as evidenced by a greater gastrocnemius weight ratio. Functional outcome based on toe-spread and pin-prick tests displayed a trend toward better functional outcome for the ALS/CsA treatment group, and preliminary histologic analysis revealed a greater axon number and myelin-sheath thickness for the ALS/CsA treatment group.

In this preliminary study, peripheral-nerve repair using allografts and a combined treatment protocol of ALS/CsA resulted in better functional recovery and morphometric outcome, without a significant difference in electrophysiologic status, compared to the non-treatment allograft group. This technique may provide an expanded source of nerve tissue to circumvent the morbidity of harvesting peripheral-nerve autografts from multiple sites for those afflicted with extensive peripheral-nerve injuries.

Effect of VEGF Gene Therapy to Improve Nerve Regeneration in Diabetic Rats. Maria Siemionow (Cleveland Clinic Foundation, Plastic Surgery, Cleveland, OH), Maciej S. Zielinski, Romed Meirer, and Dileep Nair.

The Vascular Endothelial Growth Factor 165 (VEGF-165) demonstrates angiogenic and neurogenic activity. An experimental study was conducted to investigate the potential of VEGF to stimulate regeneration in the sciatic nerves of diabetic rats, following crush injury.

Thirty-six sciatic nerves were evaluated in Zucker diabetic fatty male rats (ZDF/Gmi-fa/fa). The potential of VEGF for improvement of nerve regeneration was studied, using local injection of Ad-VEGF-165 into the area of the crushed sciatic nerve. Crush injury was performed using the authors' standardized model (60 sec compression with 17.4 N force). Animals were divided into three groups (n = 6). Group 1 (negative control) received no treatment. In Group 2 (GFP control), potential for transfection with adenovirus was evaluated by sub-epineurial injection of 0.3 ml solution containing 108 pfu of Ad-HGP. Animals in Group 3 (VEGF treatment) received injections containing 0.3 ml of 108 pfu of Ad-VEGF. Functional evaluation at 3, 6, and 12 weeks included: pin-prick test, toe-spread test, and somatosensory evoked potential (SSEP) analysis. Histomorphometric analysis of axon number, area, and myelin thickness was performed at 6 and 12 weeks. Histoimmunostaining with monoclonal antibodies against VEGF and Neuropilin (Np-1) was performed 7 days after transfection.

SSEP analysis showed a significant (p < 0.05) prolongation of the N2 latencies in the crush control and crush+GFP, compared to the crush+VEGF groups: at 3 weeks (33.63 ms and 32.32 ms vs. 27.59 ms); at 6 weeks (31.23 ms and 31.64 ms vs. 24.79 ms); at 12 weeks (29.70 ms and 32.01 ms vs. 26.14 ms), respectively. Histomorphometric evaluation showed improved nerve architecture, with statistically better (p < 0.05) results in the crush+VEGF group, compared to the crush and crush+GFP groups: at 6 weeks-axon area (54.11 um² vs. 41.02 um², 31.95% improvement), myelin thickness (2.75 um vs. 1.87 um and 1.99 um; 47.00% and 37.75% improvement); at 12 weeks-axon area (129.55 um² vs. 99.54 um² and 100.93 um²; 30.16% and 28.36% improvement), myelin thickness (3.75 um vs. 3.06 um and 3.10 um; 22.47% and 21.48% improvement, respectively. Finally, histoimmunostaining revealed higher levels of Np-1 and VEGF expression in the vasa nervorum and Schwann cells in the VEGF-treated group.

Local Ad-VEGF administration improved nerve regeneration following crush injury in the sciatic nerves of diabetic rats. It was confirmed by SSEP and histomorphometry analysis. The angiogenic and neurogenic mechanisms of VEGF action were proven by histoimmunostaining, showing expression of Neuropilin and VEGF not only in the endothelial cells of the vasa nervorum, but also in the Schwann cells.

Histologic Study of Peripheral-Nerve Regeneration, Macrophage Recruitment, and Schwann-Cell Migration in the Autogenous Venous Nerve Conduit (Work in Progress). Charles Y. Tseng (Columbia University College of Physicians and Surgeons, New York, NY), Guoli Hu, and David T.W. Chiu.

In this study, the spatial-temporal progression of peripheral-nerve regeneration, macrophage recruitment, and Schwann-cell migration across a 10-mm gap bridged by a venous conduit, was examined using immunocytochemical techniques.

In male Lewis rats, a 10-mm segment of sciatic nerve was resected. The femoral vein was resected and interposed between the proximal and distal sciatic nerve stumps. Interrupted 10-0 nylon sutures were used to approximate the full thickness of the epineurium with the endothelium of the vein graft. AVNC samples were harvested every other day, beginning POD 1 and ending POD 31. Silver staining was used to identify nerve axons. Macrophage cells were labeled with anti-ED1 monoclonal antibody immunoperoxidase staining. Schwann cells were identified with anti-S100 polyclonal antibody immunoperoxidase staining.

In the post vein graft period up to POD 5, the conduit distends secondary to endoneurial vessel bleeding and red blood cell clot formation. As the clot is resorbed, a pauci-cellular matrix is left behind. By POD 9, macrophage cells are observed in the lumen of the vein conduit and throughout the length of the distal nerve stump. Schwann cells proliferate at both nerve stumps and begin to migrate into the matrix. By POD 11, regenerating axons grow from the proximal nerve stump, and the density of macrophage cells in the proximal stump is markedly decreased. At the distal stump, macrophage density is markedly increased, and extensive Wallerian degeneration is observed. Schwann cells completely bridge the gap by 13 days. By POD 14, regenerating axons have reached the middle of the venous conduit. Macrophage cells are not observed at the proximal stump or within segments of the conduit that have filled with regenerating axons. By POD 30, axons regenerate into the remnant of the distal nerve stump. Macrophage cells are localized only to areas of the distal nerve stump still undergoing Wallerian degeneration.

Within one day, endoneurial vessel bleeding fills the venous conduit, thereby preventing collapse of the vessel lumen and also gives rise to a fibrin and red blood cell clot. Macrophage cells are localized to the tips of the proximal and distal nerve stumps. Schwann cells proliferate at the proximal and distal stumps, and migrate into the matrix. Axonal regeneration from the proximal stump consistently lags behind Schwann cell migration into the conduit. Macrophage density at the more proximal portions of the conduit begins to decrease as Schwann cells and axons advance into the conduit.

These observations suggest that macrophage cells may induce Schwann-cell activation, proliferation, and migration. In turn, the activated Schwann cells may modify the matrix, making it suitable for axonal regeneration or, more likely, they provide the necessary cellular framework for axonal growth by offering their own cell surfaces as substrata, or some combination thereof.

Immunomodulation with Anti-CD40L Monoclonal Antibody (MR1) Improves Muscle Function Following Peripheral-Nerve Allografting. David L. Brown (Section of Plastic Surgery, University of Michigan, Ann Arbor, MI), Keith Bishop, Sherri Y. Chan, William M. Kuzon, Jr., and Paul S. Cederna.

Anti-CD40L monoclonal antibody (MR1) treatment blocks the CD40-CD40L costimulatory rejection pathway, and has been shown to increase survival of murine cardiac allotransplants. Previous work in the authors' laboratory has demonstrated the effectiveness of MR1 in dramatically reducing the Th1 and Th2 responses to peripheral nerve allografts. This study investigated the effect of a 3-day course of MR1 inductive therapy on functional muscle recovery following peripheral-nerve allografting.

A two-by-two design was used to investigate the effects of MR1 therapy vs. vehicle on murine peripheral nerve isografts (C57BL/6 into C57BL/6) and allografts (BALB/c into C57BL/6), with six animals per group. Grafts of 0.8 cm were harvested from the left peroneal nerve, extending from just distal to its division from the sciatic to just proximal to its division into the superficial and deep branches. Grafts were transplanted into similar 0.8-cm defects in the recipients, utilizing 2 to 3-stitch coaptations with 11-0 nylon. On days 0, 1, and 2, each recipient received an IP injection of either MR1 (1.0 mg in 0.2 ml saline) or vehicle alone. All mice convalesced for 60 days following nerve grafting. The extensor digitorum longus (EDL) was then harvested, maximum tetanic isometric force was measured in vitro, the muscle was weighed, and the muscles and peroneal nerves were preserved for histomorphometric analysis.

Sixty days following peroneal nerve grafting, mean P_o for the EDL muscles in the vehicle-treated allograft group was 16% lower than the mean for animals in the MR1-treated allograft group. Two-way analysis of variance showed a significant effect for the variable of graft type (isograft vs. allograft), but not for treatment type (with or without MR1).

Short-term MR1 inductive therapy appears to hold promise for reducing or eliminating peripheral-nerve allograft rejection through blockade of the CD40-CD40L costimulatory pathway. In this experiment, a difference in mean values of 19% between Po for MR1-treated and untreated allografts was observed, but was not statistically significant. This lack of significance can possibly be explained as a result of small sample sizes, or may be related to the model and the time-point chosen. Efforts are currently underway to increase animal numbers in each experimental group.

Peripheral-Nerve Allografting Combined with Anti- CD40L Monoclonal Antibody Administration Reduces Cellular and Humoral Responses. Anil Mungara (Plastic Surgery, University of Michigan, Ann Arbor, MI), J.M. Rovak, D.K. Bishop, S.Y. Chan, and P.S. Cederna

The CD40-CD40L costimulatory pathway plays a crucial role in allograft rejection. Blockade of this pathway by anti-CD40L monoclonal antibody has been shown to prevent rejection of cardiac allografts. The purpose of this study was to determine the effectiveness of anti-CD40L mAb treatment as a method to induce tolerance to peripheral-nerve allografts.

Sciatic nerve grafts were transplanted from BALB/c mice into a subcutaneous pocket on the back of C57BL/6 mice. One gram of anti-CD40L mAb was then administered intraperitoneally on the day of transplantation, and on postoperative days 1 and 2. The control group received no anti-CD40L mAb. The mice were then allowed to convalesce for either 14 or 28 days. Following recovery, recipient mice were sacrificed and their spleens and brachial lymph nodes were harvested. Splenocytes and brachial node cells were stimulated with donor alloantigens, and subsequent production of interferon (INF)-γ and interleukins (IL)-5, -4, and 2 were quantified using the ELISPOT technique.

At 12 and 28 days post-transplant, splenocytes from anti-CD40L mAb-treated allograft recipients produced significantly less INF-γ, IL-4, and IL-2, compared to untreated control animals. Anti-CD40L mAb treatment did not affect IL-5 production at either time-point. In the brachial nodes 14 days post-transplant, cells producing INF-γ were considerably less in the treatment group, compared with the untreated group. Similarly, 28 days post-transplant, cells producing INF-γg, IL-4, and IL-2 were significantly less in the treatment group.

The tolerance induced by treatment administration on days 0, 1, and 2 persists for at least 24 days after the cessation of the anti-CD40L mAb treatment. These data imply that short-term anti-CD40L treatment at the time of peripheral-nerve allografting may increase long-term graft survival through the induction of tolerance.

Prevention of Neuroma Formation by Selectively Inhibiting Nerve Regeneration with Intraneural Injection of Ox7-Saporin. Zijie Zhang (Department of Neurosurgery, Johns Hopkins University, Baltimore, MD), Allan J. Belzberg, Richard A. Meyer, and John W. Griffin.

These authors investigated whether axonal transport of a kind of neuron-specific toxin (OX7-saporin) would lead to selective loss of neurons in the DRG and ventral horn of the spinal cord.

Thirty adult Sprague-Dawley rats, weighing 200 to 250 g, were randomly divided into three groups. In Group A (GA, n = 10), the left common peroneal nerve and posterior tibial nerve were severed to allow neuroma formation. In Group B (GB, n = 10), the nerves were cut as in GA, and then 2 ml of OX7-saporin was injected into the posterior tibial nerve. In Group C (GC, n = 10), 2 ml of Dulbecco's PBS was injected into the posterior tibial nerve. After 3 and 6 weeks, the rats were sacrificed by transcardial perfusion. The common peroneal, tibial, and dorsal and ventral roots of the sciatic nerve were harvested and routinely processed for light and electronmicroscope evaluations.

There was typical neuroma formation by gross observation in GA; the same findings were evident in GC. However, there was a tapered tip instead of a neuromatous bulb in GB. Microscopically, the neuroma was composed of entangled masses of nerve fibers with a large amount of regenerating axons. In marked contrast, there were extensive degenerating changes in the nerve fibers in the OX7-saporin-treated rats. The degeneration was in a well-defined portion of the sciatic nerve corresponding to the fascicles that form the posterior tibial nerve. Degeneration was also found in the dorsal and ventral roots in GB.

This study demonstrated that intraneural injection of OX7-saporin could effectively prevent neuroma formation by wiping out and/or inhibiting the regeneration of injured nerve fibers in this rat model. In the future, this type of neuron-specific neurotoxin could be seen as a potential molecular neurosurgical approach for prevention of neuroma formation. The study convincingly demonstrated that intraneural injection of OX7-saporin could effectively prevent neuroma formation.

Effects of Nerve Growth Factors on Sciatic Nerve Regeneration in Vein Grafts. James Chao (Plastic and Reconstructive Surgery, University of Texas Southwestern Medical Center, Dallas, TX), Mariam Awada, Tarek Abulezz, and Spencer Brown.

Peripheral-nerve injuries with a nerve gap are commonly encountered clinical problems, which result in long-term functional deficits. Investigations using artifical conduits and growth factors in nerve regeneration have been numerous. On the other hand, the influence of the extracellular matrix (ECM) has not been defined. This study aimed to identify the role of the ECM in nerve regeneration through the assessment and comparison of structural, functional, and electrophysiologic characteristics.

A rat sciatic nerve injury model was used. In Group 1, the nerve gap was bridged with a 1 cm long vein graft injected with nerve growth factor. In Group 2, a vein graft injected with a biodegradabale extracellular matrix was used. In Group 3, the vein graft was filled with a mixture of nerve growth factor and an extracellular matrix scaffold. The final two groups used reversed nerve graft or empty vein graft as a conduit and served as controls. The rats were analyzed at weeks 2, 4, 6, 8, and 12 for differences in functional recovery using digital walking- track analysis. After completion of functional analysis, the motor nerve conduction velocity across the repaired nerve was measured and compared to the contralateral normal sciatic nerve, to identify differences in the structural quality and quantity of the different groups of regenerated nerve segments.

Results of the assessment and comparison of the structural, functional, and electrophysiologic characteristics suggest that the addition of an extracellular matrix scaffold to nerve growth factor has an impact on nerve regeneration. Preliminary study results indicate that the ECM may accelerate nerve regeneration in a bimodal fashion. Future experiments for examining the mechanism of this cellular interaction are necessary to identify the most ideal scaffold and growth factor combination.

Susceptibility to and Recovery from Contraction-Induced Injury of the Extensor Digitorum Longus Muscle in Rat after Division and Immediate Repair of the Peroneal Nerve. Jack H. Van der Meulen (Department of Surgery, University of Michigan, Ann Arbor, MI), Mustafa Asim Aydin, Melanie G. Urbanchek, Paul S. Cederna, and William M. Kuzon, Jr.

The authors hypothesized that deficits in the maximal force generation (Po) of muscles, observed after repair of injured nerves, are caused by lengthening contractions. The specific hypothesis tested was that, compared with control muscle, other muscles that have been denervated and reinnervated are more susceptible to a single bout of lengthening contractions and recover less well from contraction-induced injury.

In adult rats, the peroneal nerve was divided and repaired with either all (NR = non-reduced) or a reduced number (DR = drastically reduced) of motor axons from the proximal nerve stump included in the repair. Sham-operated animals served as controls. After a recovery period of 4 months, the Po of the EDL muscles was measured in situ at optimal muscle length by supramaximal stimulation of the peroneal nerve. The muscles were then subjected to 225 lengthening contractions with 20% strain. Po was re-measured at either 3 days, when the Po deficit is maximal, or at 2 months, when the recovery is complete.

Data were presented as means ± SEM; *different from sham (p < 0.05). Compared with sham muscles, reinnervated muscles (NR and DR groups) generated a lower Po. There were no significant differences observed in the Po deficit at 3 days between sham, NR, and DR muscles. Two months following the contraction-induced injury protocol, sham, NR, and DR muscles recovered to pre-injury levels.

Based on these data, the authors rejected their hypothesis and conclude that, on the whole muscle level, reinnervated muscles are not more prone to a single bout of contraction-induced injury, and recover equally well from contraction-induced injury, compared with sham muscles.

Denervation of the Sinus Tarsi. Steve Barrett (Institute for Peripheral Nerve Surgery, Tucson, AZ) and A. Lee Dellon.

The sinus tarsi represents a collection of ligaments on the dorsolateral aspect of the ankle that can be torn during ankle inversion sprains or fractures. The innervation of the sinus tarsi has been demonstrated by Rab, Ebmer, and Dellon (2001) to be from terminal branches of the deep peroneal nerve. Just as injury to the wrist joint can damage the terminal branch of the posterior interosseous nerve, it was hypothesized that dorsolateral ankle pain after ankle injury, the sinus tarsi syndrome, might be due to traumatic neuromata of these terminal nerve branches. The authors' first clinical experiences with the results of denervation of the sinus tarsi were presented.

In 10 patients, the etiology of the pain was inversion sprain in 8, ankle fracture in 1, and MBA-flatfoot implant in 1 patient. The mean post-injury time was 13.5 months. Each patient had failed conservative treatment, including NSAIDs, orthotics, steroid injections; and each had no residual bone deformity requiring treatment. Each patient had relief of pain with a block of the deep peroneal nerve proximal to the ankle joint. Denervation was done 4 cm proximal to the ankle joint, preserving the sensory component to the first dorsal webspace and a portion of the innervation of the extensor brevis muscle. The technique was illustrated.

At a minimum of 6 months of postoperative follow-up, seven of the patients (70%) were completely free of pain, able to wear normal shoes, and had returned to work. In two of the patients, there was a small degree of residual pain, but they were able to resume normal activities and wear normal footwear (20%). In one patient, there was some pain relief, but this patient was not able to resume normal activities.

Outcome of Surgical Decompression of Nerves in Diabetics on Incidence of Ulceration and Amputation. A. Lee Dellon (Plastic Surgery, Johns Hopkins University, Baltimore, MD), Oskar Aszmann, and Patricia Tassler.

Historically, diabetic neuropathy is progressive and irreversible. A predictable outcome of this neuropathy is ulceration in 1 of 6 diabetics, occurring at a rate of 2.3% per year of diabetic neuropathy. One of six patients with ulceration will have an amputation. In 1998, there were 88,000 amputations in diabetics. Since 1992, four independent published studies have documented that sensation can be restored to the feet of diabetics by decompression of the posterior tibial nerve. Those studies did not use ulceration and/or amputation as the primary outcome parameters.

The present study evaluated 49 diabetics who had a unilateral decompression of the four medial ankle tunnels (posterior tibial nerve and its branches). The contralateral foot was not decompressed. The glycemic control was therefore the same for both feet. Outcomes were assessed by both a questionnaire and telephone call follow-up. Data were analyzed using SPSS 9.0 for Windows software. An unfavorable outcome was defined as an amputation or the development of an ulcer. Fisher's exact test was used to test the null hypothesis that the proportions with an unfavorable outcome in the population were the same: that is, that operated extremities were equally likely to develop unfavorable outcomes as non-operated extremities.

At a mean of 4.5 years after surgical decompression, there were no ulcerations and no amputations to the foot that had surgery, while there were 2 amputations and 7 ulcerations (9 different patients) in the contralateral (control) foot. This difference is statistically significant at the p < 0.002 level.

The results of this study suggest that decompression of the four medial ankle tunnels in the diabetic patient with symptomatic neuropathy can change the natural history of diabetic neuropathy in terms of ulceration and amputation.

Anatomy of the Ulnar Nerve Branches in the Forearm. Reid Abrams (Orthopedic Surgery, Hand and Microvascular Surgery Service, University of California, San Diego, CA), John Eggleston, and Clark Robinson.

Nerve identification and safe dissection are facilitated by knowledge of nerve anatomy. New neuroanatomic insights provide understanding of anomalous innervation, and the discovery of redundant motor innervation may identify potential motor axon donors for nerve transfers. Knowing the distances along a nerve from a reference point to the end organ allows estimation of recovery time after nerve repair.

The ulnar nerve was dissected under X3.5 magnification in 25 fresh cadaveric arms from the mid-brachium to the wrist. Using the medial humeral epicondyle (MHE) as a reference (0.0 mm point), distances were measured along the nerve to branch points (with “-” connoting “proximal” to the MHE) and along the branches to their targets. Branch patterns were recorded. Means and ranges of branch intervals and lengths were calculated.

One to three elbow joint branches were found in 48% of arms with an average branch point of -16.8 mm. The most proximal motor branch point was to the FCU at -2.6 mm. Average branch points to the FCU humeral and ulnar heads and the common belly were 19.1 mm, 22.3 mm, and 41.5 mm, respectively. Mean branch lengths to the humeral, ulnar, and common bellies averaged 18.6 mm, 23.2 mm, and 53.2 mm, respectively. The humeral head had branches in 88% of arms with 0 to 2 branches per specimen. One hundred percent of the ulnar heads had 1 to 3 branches, and the common belly had one branch in 2/25 specimens. The FCU was the most richly innervated muscle, having a mean of 2.4 branches (range: 2 to 4). The FXS received ulnar efferents in 3/25 specimens, with a mean branch point of 36.5 mm and length of 32.2 mm. The FDP was ulnar innervated in 96% of arms, with the most proximal and mean branch points being 11.7 mm and 42.0 mm, respectively, and mean branch length being 24.1 mm. A Martin-Gruber connection from the ulnar to the median nerve was found in 6/25 specimens, with a mean branch point of 72.4 mm. Twenty-one of 25 specimens had branches to the ulnar artery. The superficial ulnar nerve sensory branch was present in all specimens, branching at an average of 186.4 mm.

The most proximal branch was not always articular. Average muscle innervation order from proximal was FCU, FDS (when present), and FDP. Occasionally, the FDP was not motored by, and sometimes the FDS was partially innervated by, the ulnar nerve. The FDS and FCU may have motor axons that can be spared for nerve transfer.

Reverse “Baby-Sitter” Procedure: A New Technique to Optimize a Motor Nerve Prior to Functional Microvascular Muscle Transplantation. Darrell Brooks (Replantation/Transplantation Service, Buncke Clinic, San Francisco, CA).

Early heterotrophic nerve repair has been utilized as a “baby sitter,” to protect facial musculature from denervation atrophy. This technique optimizes eventual spontaneous mimetic function created by cross-facial nerve regeneration. Experimental studies have shown that early motor nerve repair to a recipient muscle can have a protective influence on the motor nerve by promoting axon regeneration and neuron survival. The author described early neuromuscular transplantation used to optimize the motor nerve in preparation for FMMT, a reverse of the standard baby-sitter technique.

A 37-year-old man sustained a crush-avulsion of his forearm, resulting in an 18 × 22-cm soft-tissue defect involving the dorsal and palmar compartments, as well as the radial nerve proximal to its sensory branch. The patient underwent latissimus muscle transplantation to obtain primary wound cover. The thoracodorsal nerve measuring 10 cm, was repaired to the mixed motor-sensory radial nerve, which was debrided 2 cm proximal to healthy-appearing fascicles outside the zone of injury. Between the 6th and 8th weeks after muscle transplantation, the latissimus began to contract. By the 13th week, the patient could contract the entire muscle.

Eight months after latissimus transplantation, the contralateral gracilis was harvested to restore finger and wrist extension. During elevation of the latissimus muscle, an avascular septo-alveolar plane, similar to that between two healthy muscle bellies, was developed with gentle finger dissection. After the muscle was inset and revascularized, the thoracodorsal nerve was transected and repaired to the obturator nerve insertion into the gracilis muscle. At 7 weeks after FMMT, there was palpable contraction. At 8 weeks, the patient had regained active wrist and finger extension. At 7 months, he had regained 92% of wrist extension, compared to the uninjured hand. Strength was 4+ based on the Medical Research Council scale.

The simple technique of innervating the muscle used for coverage of a large upper extremity wound helped optimize the motor nerve and wound bed. Latissimus excursion signaled the appropriateness of the target motor nerve, and eventually created a recipient bed comparable to the plane between two healthy muscles. The thoracodorsal nerve provided enough graft length to both allow initial proximal repair outside the zone of injury, and eventual distal repair as close as possible to the functional muscle, decreasing reinnervation time. The theoretic benefits of early nerve repair and restoration of neurotrophic influences on nerve regeneration, neuron survival, and ultimately the function of the FMMT were discussed.

Time Course of Recovery of Skeletal Muscle Following Botulinum Toxin Injection in Juvenile Rats. Jianjun Ma (Orthopaedic Research Laboratory, Department of Orthopaedic Surgery, Wake Forest University School of Medicine, Winston-Salem, NC), Beth P. Smith, Thomas L. Smith, Francis O. Walker, and Andrew L. Koman.

Botulinum toxin A (Btx-A) blocks neurotransmission at the neuromuscular junction by preventing acetylcholine release from the presynaptic motor nerve terminal. Btx-A has been used to manage spasticity associated with cerebral palsy in the last several decades. However, the functional recovery of the injected muscles following toxin injections remains unclear. An improved understanding of the time course of recovery of the targeted muscle would provide information necessary to standardize the techniques of Btx-A injection in pediatric cerebral palsy patients.

Btx-A was injected into the left gastrocnemius of 29 one-month-old Sprague-Dawley rats at a dosage of 6 units/kg. The same volume of saline was injected into the right gastrocnemius to serve as a control. Muscle mass, motor evoked action potentials (latency, amplitude, and area under the curve), and muscle force generation were evaluated at 2, 7, and 14 days and 1, 2, and 3 months after injection (n = 4-5 at each time point).

Muscle mass was reduced by 31.5% at the contralateral side within 2 weeks, and at 3 months was still reduced 27.1%. Latency was delayed to 144.9% at 3 days and recovered to 112.3% at 3 months. Amplitude and area under the curve decreased to approximately 30% at 3 days, then gradually recovered to 80% at 3 months. Single twitch was decreased to 12.3% at 1 week but recovered to 86.3% at 3 months; tetanus decreased to 24.4% at 3 days, then gradually recovered to 91.5% at 3 months.

The motor evoked action potentials and muscle force generation recovered faster during the first month than during the second and third months. Btx-A at a dosage of 6 units/kg could induce a substantial muscle atrophy but reversible muscle weakness within 3 months following toxin injection. The recovery of motor action potentials and muscle force generation indicates that most neuromuscular junctions have been re-established functionally.

Protection of Chronically Denervated Peripheral Nerve Stumps. Catherine A. Munro (Division of Neurosurgery, Sunnybrook and Women's College HSC, University of Toronto, Ontario, Canada), Rajiv Midha, Steven Y.T. Chan, Akkapong Nitising, and Tessa Gordon.

Delayed repair of peripheral nerve injuries often results in poor motor functional recovery. This may be due to the deterioration or loss of endoneurial pathways in the distal nerve stump before motor axons can regenerate into it. In this study, the authors developed a rat hindlimb model to determine if regeneration of either motor or sensory nerve down a distal nerve stump would allow the successful regeneration of a cut motor nerve 2 months later, compared to the regeneration of a cut motor nerve in a 2-month chronically denervated distal nerve pathway.

Using the rat femoral nerve, the authors protected the distal endoneurial pathways of the saphenous nerve with either cross suture of the quadriceps motor nerve (Group A) or resuture of the saphenous nerve (Group B), to compare later motor regeneration into the “protected” saphenous nerve pathway to chronic denervation and “unprotected” saphenous nerve (Group C). A total of 45 rats, 15 per group, were operated on. After this protection (or lack of protection) for 2 months, the motor branch of the femoral nerve was cut and sutured to the distal saphenous, to allow motor regeneration into protected and unprotected saphenous nerve stumps. The quantitative assessment of axonal regeneration was carried out after 6 weeks, 30 mm distal to the suture site, by applying a 4% solution of Fluorogold (FG). Rats were perfusion fixed 6 days later, the T11 to L2 spinal cord collected, and the number of FG-labelled femoral motoneurons in all the spinal cord sections counted.

These experiments will determine if using an expendable sensory nerve to prevent distal endoneurial pathway deterioration will allow for successful motor regeneration down long pathways.

Management of the Paralyzed Eyelid. Hallene A. Maragh (Cosmetic and Plastic Surgery Specialists, Richmond, VA).

The severity of the complications of the paralyzed eyelid dictates the need for timely surgical intervention. This report focused on the management of 41 patients over a period of 6 years.

The charts of 41 consecutive patients who underwent eyelid procedures were reviewed from July 1995 to June 2001. Patients were grouped into the following categories: (1) those who had split hypoglossal and cross-facial procedures alone (9 cases); (2) those who had split hypoglossal, gold weight, and lower lid support (10 cases); (3) those who had traumatic facial nerve repair (15 cases); (4) those who had lid-loading procedures and lower lid support (7 cases). The cross-facial nerve grafts were repaired in a single stage in 3 patients, and in two stages in 6 patients. In Group 2, the lower lid support was provided with a sling or hard palate graft. In Group 3, all facial nerves were repaired with nerve grafts. Three patients had facial nerve grafted to the upper facial nerve branches and split hypoglossal grafted to the lower face. In the other 12 patients, only the facial nerve was grafted.

Of the 42 patients, there were 21 females and 20 males, with a mean age of 41 years and an age range of 2 to 68 years. The mean duration of the facial palsy was 21 months, with a range of 6 months to 5.5 years. The mean follow-up time was 4.2 years (range: 5 months to 6 years). In Group 1, six patients (66%) had excellent improvement, with no requirements for tears or ointment. No secondary procedures have been required. In Group 2, five patients (50%) had excellent results, with no requirement for tears or ointment. One patient (10%) had had repositioning of a gold weight. In Group 3, nine patients (75%) had excellent results, requiring no tears or ointment. In Group 4, four patients (57%) had excellent results, with no requirement for tears or ointment. The others have had implant adjustment. Two of these were metallic springs and the other was a gold weight. In summary, 58% of the patients had excellent results. Only 23% of 17 patients with metallic implants required revision.

When the denervation time is less than 24 months, neurotization of the sphincter must be provided, as this improves the function, aesthetics of the eyelids, and reduces the need for medical treatment.

Cavernous Nerve Reconstruction with Sural Nerve Graft during Radical Prostatectomy or Cystectomy in Patients with Advanced Local Stage Prostate Cancer. Vishal Sarwahi (Orthopaedics, Columbia University Medical Center, New York, NY), R.J. Strauch, M.P. Rosenwasser, Ridwan Shabsigh, and Mitchell Benson.

Advanced local-stage prostate and bladder cancers frequently require wide local resection and sacrifice of cavernous nerves for negative surgical margins, leading to erectile dysfunction. This study documented preliminary experience with unilateral cavernous nerve reconstruction using sural nerve graft, its safety and efficacy in recovery of erectile function.

Sural nerve grafting was performed unilaterally after primary resection. Postoperative treatment with sidenafil and alprostadil was given until recovery of spontaneous function; recovery was documented by the International Index of Erectile Function (IIEF).

Twelve unilateral nerve grafts were performed, 10 during radical prostatectomy and 2 during radical cystoprostatectomy. Patient mean age was 57.5 years, and the mean follow-up was 16.1 months. Surgical margins were positive in 5 while 2 had positive lymph nodes. At the most recent follow-up, 4 patients were fully potent, while 1 had improving partial erections. The only complication was one superficial donor-site wound infection.

Animal studies using genitofemoral nerve grafts have documented recovery of erectile function in denervated rats. In humans, this is a technically challenging procedure, due to the complexity of the cavernous nerves, short distal cavernous stumps, difficulty using a microscope in the depth of the plexus, a bleeding resection bed, and an uneven operating field. Four of 12 patients in the study had return of spontaneous erectile function with a score >20, while one had partial recovery. The average time to return of function was 16 months, with faster recovery seen in younger patients. Patients receiving external beam radiation or hormonal therapy had no recovery, as irradiation compromises the revascularization of the nerve graft. Kim et al. reported 3 of 12 patients with rigid erections, while 9 had varying degree of recovery after sural nerve graft. They indicated that full benefits might not be appreciated until 24 to 36 months.

Sural nerve grafts are extensively used in peripheral nerve surgery and may be a better donor tissue than the genitofemoral nerve. The study indicates that unilateral sural nerve grafts after radical resections are technically feasible and are associated with minimum morbidity. Younger patients and patients with negative surgical margins (4/7) can have restored potency. Extracapsular spread is a vexing problem, since it carries a high risk of recurrence and also requires adjuvant hormonal or radiation therapy, which further compromises recovery. The technique may provide patients with wide local disease the opportunity of a spontaneous erectile function, and may allow surgeons to perform more aggressive tumor resection.

Restoration of Motor Function of the Deep Peroneal Nerve by Direct Nerve Transfer of Branches from the Tibial Nerve. Kale D. Bodily (Mayo Medical School, Rochester, MN), Robert J. Spinner, and Allen T. Bishop.

Stretch injuries to the common peroneal nerve, frequent at the time of varus knee injury, result in significant morbidity due to loss of ankle dorsiflexion. Reconstruction is often unsuccessful because of the length of the nerve lesion, sometimes extending from its proximal origin of bifurcation high in the popliteal fossa to the point of its terminal bifurcation several centimeters distal to the fibular head. Such lengthy injuries are poor candidates for sural nerve grafts. Furthermore, tendon transfer often does not eliminate the need for bracing and may result in late hind-foot valgus and arthritis. When these options fail, chronic use of an ankle-foot orthosis is necessary. The objective of this reported study was to determine the anatomic feasibility of an alternative option, consisting of nerve transfer of motor branches from the tibial nerve to the deep peroneal nerve.

This technique bypasses the zone of injury, obviates the need for an intercalary nerve graft, and decreases regeneration time by reconstructing the nerve as close as possible to the end organ. All of these advantages should improve the results of primary nerve reconstruction.

In 10 cadaveric limbs, the branching pattern, length, and diameter of motor branches of the tibial nerve in the proximal leg were characterized. Nerve transfer of each of these motor branches was then simulated to the proximal deep peroneal nerve.

A consistent, reproducible pattern of innervation was seen with minor variability. Branches to the flexor hallucis longus and flexor digitorum longus muscles were determined to be adequate for direct nerve transfer in all specimens, without interpositional graft, based on their diameter and length. Other branches, including those to the tibialis posterior, popliteus, gastrocnemius, and soleus muscles were not consistently adequate for repair of injuries extending to the bifurcation of the common peroneal nerve. For neuromas that do not extend as far distally, branches to the soleus and lateral head of the gastrocnemius may be adequate, if the intramuscular portions of the nerves are dissected.

This study confirmed the anatomic feasibility of direct nerve transfer, using nerves to toe-flexor muscles as a treatment option to restore ankle dorsiflexion in cases of common fibular nerve injury.

Static and Dynamic Forces to Control the Shoulder Joint. Hanno Millesi (Ludwig Boltzmann Institute for Experimental Plastic Surgery, Vienna, Austria), Roland Reihsner, and Dagmar Millesi.

At the shoulder joint, the humerus is kept in place by a resting force of the following muscles: deltoid supraspinatus, long head of the biceps, long head of the triceps. If the deltoid and supraspinatus in particular are paralyzed, the head of the humerus luxates partially until the soft connective tissue becomes elongated enough to prevent further luxation. Different techniques have been designed to transfer non-paralyzed muscles, e.g., part of the trapezius or the levator scapulae, to improve active motion in a paralytic shoulder joint after a brachial plexus lesions. Even after successful surgery, one can observe that a large percentage of the power of the transferred muscle is used for static correction. If the head of the humerus is lifted into the correct position, only a fraction is left for the desired active motion.

The authors have analyzed this problem from a mechanical point of view, and concluded that a better result can be achieved if the static and dynamic forces are divided. The static force to maintain the head of the humerus in place is provided by a fascia lata sling across the acromion and the head of the humerus. The transferred muscle is than protected against elongation, and its force is fully available for motion. This principle was demonstrated by a series of cases.

Anatomic Evaluation of the Triceps Brachii Motor Branches as Recipients in Nerve Transfers. Robert J. Spinner (Departments of Neurologic Surgery and Orthopedics, Mayo Clinic, Rochester, MN), Edward C. Stauber, and Allen T. Bishop.

Triceps function as an elbow extensor and stabilizer is an increasingly recognized goal of brachial plexus reconstruction. Intercostal nerve transfer to the triceps brachii is employed to this end, especially in double free muscle transfer. However, the particular triceps branch to which intercostals are coapted has not been characterized, and the anatomy of the triceps branches relevant to nerve transfers has not been described.

The anatomy of nerves innervating the triceps brachii was defined in 10 cadaveric limbs to evaluate these motor branches as recipients in intercostal nerve transfers. Length, diameter, and branching patterns of triceps motor branches, as well as their location relative to relevant anatomic landmarks, were recorded. Intercostal nerve transfers (T3 and T4) to triceps branches were simulated.

Only two triceps branches consistently originated in the axilla, allowing relatively simple exposure via the anterior or axillary approach-long head nerve and ulnar collateral nerve. The long head nerve was the most proximal branch of the radial nerve and provided the sole innervation to the long head in all but one specimen. On average, the long head nerve had a diameter of 1.2 mm and traveled 76 mm before penetrating the long head. The ulnar collateral nerve originated from the radial nerve 12.2 mm distal to the origin of the long head nerve, had a diameter of 0.9 mm, and traveled 100 mm before penetrating the medial head. Direct intercostal nerve transfer to the long head nerve and ulnar collateral nerve was anatomically feasible in all specimens.

Anatomically, the long head nerve is the preferred recipient for intercostal nerve transfer for several reasons. First, this branch is most accessible at surgery, either through a deltopectoral or axillary incision. Second, as the shortest of triceps branches, nerve transfer to the long head nerve would minimize muscle denervation time. Third, the long head nerve is the sole source of innervation to its head, while the ulnar collateral nerve supplies only part of the medial head (along with the nerve to the anconeous/medial head). Fourth, the cross-sectional area of the long head nerve is greater than the ulnar collateral nerve. Fifth, recovery of long-head action as an arm adductor would potentially stabilize the shoulder.

This study confirms the anatomic feasibility of direct intercostal nerve transfer to the long head nerve and demonstrates anatomic factors that make it the superior triceps branch recipient in nerve transfers.

Prognostic Ability of a Positive Tinel Sign in Determining Ability to Restore Sensation to Feet in Diabetic and Non-Diabetic Neuropathy. Cathy H. Lee (Institute for Peripheral Nerve Surgery, Johns Hopkins University, Baltimore, MD) and A. Lee Dellon.

It has been demonstrated in four independent reports from 1992 to 2000, that decompression of peripheral nerves in diabetics can restore sensation and relieve pain in patients with symptomatic neuropathy. Two of these studies required patients to have a positive Tinel sign over the tibial nerve in the tarsal tunnel as an indication for surgery, while two of the studies did not. There have been no reports of being able to restore sensation to the feet of patients with neuropathy of unknown etiology.

The purpose of the present study was to evaluate prospectively, beginning in 1997, the results of surgery to decompress the tibial nerve and its branches, based on the presence or absence of a positive Tinel sign at the tarsal tunnel in patients with symptomatic neuropathy due to diabetes of unknown etiology (26). Outcome was determined at a minimum of 1-year follow-up, ending in January 2001. Outcome was dichotomized into either good/excellent or failure/poor, permitting analysis for sensitivity and specificity.

For patients with diabetic neuropathy, the presence of a positive Tinel sign prior to surgery gave a sensitivity of 88%, a specificity of 50%, and a positive predictive value of 88% in identifying patients who would have a good-to-excellent outcome. For patients with neuropathy of unknown etiology, the presence of a positive Tinel sign prior to surgery gave a sensitivity of 95%, a specificity of 56%, and a positive predictive value of 93% in identifying patients who would have a good-to-excellent outcome.

Changes in Spinal Cord Architecture after Brachial Plexus Lesion in the Neonate. Oskar C. Aszmann (Division of Plastic and Reconstructive Surgery, Department of Surgery, University of Vienna, Austria), Klaus Korak, and Manfred Frey.

Obstetric brachial plexus injuries present a dramatic insult not only to the peripheral nervous system, but also to its central components. Thus far, only the peripheral pathology has been given adequate attention, both regarding diagnosis/prognosis and treatment. In the current work, the authors have investigated the hypothesis that obstetric brachial plexus injuries will lead to permanent changes in spinal-cord architecture, and that they not only represent an important part of the pathology but also of the mechanisms of recovery.

A total of 40 Sprague-Dawley rats were used, and divided into three groups. Control groups A and B consisted of 10 adult animals each; the experimental group C consisted of 20 neonate animals. In control group A, the size of the normal musculocutaneous (MC) motoneuron pool was determined through retrograde tracing using fast blue and fluorogold. In control group B, the relative contribution of spinal root C7 to the normal adult MC motoneuron pool was determined. Further, total force generation of the biceps muscle, as well as C7 individual contribution of force generation, were measured. In the experimental group C, the upper trunk of the brachial plexus was crushed at the time of birth. After 8 weeks, force measurement of the biceps with individual C7 and MC stimulation were performed, and quantification of the corresponding motoneuron pool was done.

In the normal animals, the motoneuron pool of the MC nerve extends from segments C4 to C7. However, the contribution of C7 is minimal (28 ± 24) and not present in all animals. In the experimental animals in which the upper trunk was crushed at birth, a major portion (132 ± 56) of the motoneuron pool of the MC nerve was found in the C7 segment, that also contributed significantly to biceps function 12 weeks after the injury.

Injury to the brachial plexus at birth leads to irreversible changes in spinal-cord architecture. The experiments described indicate that these changes are not only an important part of the primary pathology of obstetric brachial plexus injuries, but also of the subsequent recovery process. This more complete understanding of the pathology lays a better groundwork for understanding the natural course of recovery that may also lead to new therapeutic concepts.

Evaluation of Sensitivity and Specificity of Diagnostic Tests in Carpal Tunnel Syndrome Using Latent Class Analysis. Binu P. Thomas (Christine M. Kleinert Institute of Hand and Microsurgery, Louisville, KY), Scott LaJoie, and Steven McCabe.

There are many reports in the literature assessing the specificity (Sp) and sensitivity (Se) of diagnostic tests for carpal tunnel syndrome (CTS), such as Tinel's sign, the Phalen test, and electrodiagnostic (EDX) tests. Most of these assessments are done with the premise that the EDX test is the gold standard in the diagnosis of CTS. This assumption is not entirely correct. When a gold standard is assumed where actually none exists, the Sp and Se of the tests may be erroneously low. These authors aimed to calculate the Se and Sp of diagnostic tests in CTS when a gold standard is not available.

The Latent Class Analysis (LCA) is a mathematical model that allows calculation of Se and Sp in the absence of a gold standard. The model is available as software that can be downloaded from the internet. The LCA was used to calculate the Se and Sp of Tinel's sign, the Phalen test, and the EDX test in the diagnosis of CTS. The prevalence rate in a sample population can also be calculated with the LCA.

Charts of patients presenting with symptoms of upper extremity nerve compression were scrutinized. One hundred such patients were selected who had concurrent recordings of the Tinel sign, Phalen test, and value of EDX test expressed as a binary (i.e., positive or negative) result. In the 100 patients, 185 hands with these data were identified. The data were analyzed using the LCA software.

The Se and Sp of the Tinel sign was 0.97 and 0.92, that of the Phalen test was 0.9 and 0.89, and of the EDX tests was 0.92 and 0.86, respectively. The prevalence rate of CTS in the sample population was 0.62. The estimates obtained using the LCA were higher than in previous reports. The authors believe that this reflects the true state of these tests, as this method proportions the error between the three tests based on the best estimate from the observed data.

Evaluation of Accessory to Suprascapular Nerve Transfer in Reconstructive Brachial Plexus Surgery. Martijn J.A. Malessy (Leiden University Medical Center, Neurosurgery, Leiden, The Netherlands), Godard de Ruiter, and Ralph T.W.M. Thomeer.

Restoration of abduction and exorotation is one of the goals in treating patients with severe brachial plexus (BP) injuries. The accessory to suprascapular nerve (XI-SS) transfer is used for this purpose when spinal roots are avulsed and proximal stumps are not available. In this reported study, the authors analyzed the results obtained in a consecutive series of 22 patients with a BP traction/crush lesion treated with XI-SS transfer.

The nerve transfer was part of an extended BP reconstruction with restoration of elbow flexion as the primary goal. Patients with neurotization of the axillary nerve were excluded. The surgical intention was to perform direct XI-SS coaptation. The XI nerve was therefore transected as distally as necessary or technically possible. All patients were followed for at least 2 years postoperatively. The mean age at surgery was 24 years. The mean interval between trauma and operation was 127 days ± 62 (SD). Nine patients were shown to have avulsion of C5-T1, and 13 had rupture or avulsion of C5, and avulsion of C6 (C7). The force exerted by the supraspinatus and infraspinatus muscles was assessed according to the MRC scale. The glenohumeral abduction and exorotation angles and total (scapulothoracic and glenohumeral) range of movement were noted. Repeated postoperative needle EMG studies of the supraspinatus and infraspinatus muscles were performed in all patients.

EMG signs of reinnervation in the supraspinatus muscle were found in 19 patients (86%) and in the infraspinatus muscle in 16 patients (72%). In 15 patients (68%), direct coaptation of XI to SS nerve was achieved. In this group, 5 patients (33%) attained supraspinatus muscle force of MRC grade ≥ 3. In 4 patients (27%), glenohumeral abduction was 45 degrees and total abduction range was 60 to 90 degrees. Three patients (20%) achieved infraspinatus MRC ≥ 3. Glenohumeral exorotation was 45 degrees in 2 patients and 70 degrees in 1. Only 2 patients (13%) had both supra- and infraspinatus MRC ≥ 3. In 7 patients (32%), a graft (mean length: 4.8 cm ± 1.4 SD) was required to connect the XI to the SS nerve. In this group, none of the muscles achieved MRC ≥ 3. Six of these 7 patients had avulsions of C5-T1.

Although EMG signs of reinnervation were present in the vast majority of patients, restoration of useful shoulder function was limited, especially when grafts were needed to bridge the gap between the XI and SS nerves.

External Rotation Following Suprascapular Nerve Reconstruction in Obstetric Brachial Plexus Palsy: Accessory Nerve Transfer vs. C5 Grafting. Jeffrey R. Marcus (Hospital for Sick Children, Division of Plastic Surgery, Toronto, Ontario, Canada), Christine G. Curtus, and Howard Clarke.

Following neuroma resection in obstetric brachial plexus palsy, distal targets are generally reconstructed by interposition nerve grafts from the proximal nerve stumps or by nerve transfer. Transfer of the spinal accessory nerve (CNXI) to the suprascapular nerve (SSN) is one of the most common neurotization procedures, but the comparative efficacy of this technique vs. intraplexal grafting has never been objectively established. The null hypothesis of this study was that there is no difference in functional outcome-external rotation at the shoulder-following reconstruction of the SSN, using either CNXI nerve transfer or grafting from the 5th cervical nerve stump.

The study is an outcomes analysis of prospectively acquired data from consecutive patients. All patients had the diagnosis of obstetric brachial plexus palsy and underwent exploration of the brachial plexus with neuroma resection. Other than SSN, all distal targets were reconstructed using multi-strand intraplexual grafting. SSN was reconstructed using either CNXI nerve transfer or grafting from the C5 stump. All patients had a minimum of 3-years follow-up with a standardized schedule for regular postoperative reexamination. The primary outcome measure was external rotation at the shoulder, as determined during each serial examination of the Active Movement Grading Scale.

Eighty-eight patients met inclusion criteria; 33 underwent CNXI transfer and 55 underwent C5 grafting. For either reconstruction technique, there was a statistically significant increase in graded external rotation at 3 years. However, comparing the two techniques, the treatment results were statistically equivalent.

Regardless of the type of reconstruction, all patients experienced significant measurable improvement in external rotation. There was no difference in the magnitude of this improvement for infants undergoing accessory nerve transfer or grafting from the C5 stump. Therefore, the null hypothesis of this study was upheld. The results of the study indicate that both procedures are suitable options for reconstruction of the suprascapular nerve.

Functional MRI Imaging after Brachial Plexus Repair by Means of Cross-C7 Nerve Transfer. J.Y. Beaulieu (Hospital Bichat, Paris, France), Christophe Oberlin, J. Blustajn, F. Teboul, S. De Schonen, P. Baud, and J.B. Thiebaud.

These authors used functional magnetic resonance imaging (fMRI) to evaluate cortical activation in the sensory-motor areas of the contra- and ipsilateral cerebral hemisphere during motor tasks and imagined movement with a biceps neurotised by a cross-C7 graft.

Seven patients with total brachial plexus avulsion underwent a transfer from a part of the contralateral C7 root to the musculocutaneous nerve or the lateral cord in order to restore elbow flexion. They were studied with echo-planar MRI (1.5 T) while performing flexion tasks with both elbows and imagined flexion sequences on both sides. During the tasks, the areas of activated pixels were studied and compared on the ipsi- and contralateral cortex.

There was a bilateral coactivation between the two sensory-motor cortexes (primary motor area) when the patients were using the neurotized elbow. During the imagined tasks, the activated area of the contralateral cortex was significantly larger than the ipsilateral cortex activated area; the contralateral area appeared similar to the contralateral area activated during motor tasks.

After cross-C7 nerve graft, the contralateral cortex is already involved in the elaboration of recovered movements in association with other motor areas.

Pores in Artificial Nerve Grafts Improve Electrophysiologic Properties of Regenerating Peripheral Nerves. Carmen L.A.M. Vleggeert-Lankamp (Leiden University Medical Center, Neuroregulation Group, Dept. of Neurosurgery, Leiden, The Netherlands), Ana P. Pego, Egbert A.J.F. Lakke, Enrico Marani, Andre A. Poot, Dirk W. Grijpma, Jan Feijen, and Ralph T.W.M. Thomeer.

Previously, these authors introduced a quantitative electrophysiologic method for comparing nerve grafts in vitro. This method was used to evaluate the effects of the presence of pores in biodegradable artificial nerve grafts on regeneration. Pores in an artificial nerve graft will allow easier access of nutrients to the regenerating nerve within the graft; consequently, the authors expected improved regeneration through the porous grafts, compared to non-porous grafts.

Artificial nerve grafts of porous (1-10 um pore size; n = 6) and non-porous (n = 6) poly (ε-caprolactone) (CL), and of a porous (1-10 um pore size; n = 6) copolymer of trimethylenecarbonate and ε-caprolactone (10:90 mol%) (TMC/CL) were used to bridge a 6-mm gap in the rat sciatic nerve. Autografted nerves served as controls. Twelve weeks after surgery, increasing stimulus voltages were applied to the (grafted) sciatic nerve in vitro, and the resulting compound action currents were measured. The maximum charge displaced (Q_max), the mean conduction velocity (MCV), and the mean voltage threshold (V50) were determined.

Q_max measured in the non-porous nerve grafts was too low to be measured in 5 of 6 regenerated nerves, and was therefore significantly lower, compared to the autografts and porous grafts. There were no significant differences in Qmax, MCV, and V₅₀ between the two porous grafts. However, in porous grafts, Qmax was significantly smaller, compared to the autografted nerves, while MCV and V₅₀ were not significantly different. Of the two porous grafts, the TMC/CL graft provided the best surgical handling capability.

Qmax is mainly determined by the number and diameter of the (regenerated) axons. Thus, the results indicate that more axons or larger axons or more and larger axons had regenerated through the porous grafts, compared to the non-porous grafts, providing evidence for the beneficial effect of pores in artificial nerve grafts on regeneration. Moreover, although Q_max in the porous grafted nerves was still significantly smaller, compared to the autografted nerves, MCV and V50 were mainly determined by the diameter of the (regenerated) axons, indicating that the diameter of the regenerated axons in the porous grafts is comparable to that of the regenerated axons in the autografts. Since, of the two porous grafts tested, the TMC/CL graft provided the best surgical handling capabilities, this will be the preferred graft for further experiments.

Regeneration of the Nerve Fibers after End-to-Side Coaptation of the Musculocutaneous Nerve to the Median or Ulnar Nerve in Rats. Minoru Shibata (Niigata University Hospital, Division of Plastic Surgery, Niigata, Japan), Way Chen, Kuihui Han, Masaji Shirokura, and Akiyoshi Kakita.

Nerve regeneration through end-to-side coaptation has been confirmed, and the fibers are thought to often sprout from the internodal space of the recipient nerve. These authors used recombinant adenovirus vector containing LacZ gene for the tracing of the peripheral nerve fibers, and showed that the regenerating nerve fibers can be traced along their length. In a previous study using peroneal nerve coapted to the tibial nerve in end-to-side fashion, the regenerating fibers sprouted from very proximal to the original neuron of the peroneal nerve, and there was no discernible internodal sprouting from the recipient nerve.

The experimental model involved 84 rats divided into two experimental groups. In experiment 1, root levels of the normal musculocutaneous, median, and ulnar nerves were detected using 28 rats, and 4 weeks after infection by adenovirus of these three nerves. In experiment 2, the severed proximal musculocutaneous nerve was connected to either median or ulnar nerve through a perineurial window in an ETS manner at the first surgical procedure. A second operation for adenovirus infection was carried out 6 weeks after the first operation. Four weeks after the second procedure, the animals were sacrificed for 8-galactoisadase chemistry.

Positive tracing ratios: 12 rats demonstrated positive staining of the infected nerve (12/27; 44% positive ratio). In experiment 2, the median nerve model showed a 50% positive ratio (15/30) and the ulnar nerve model, 66.7% (18/27). Distribution of traced nerve fibers: experiment 1-traced nerve fibers in the brachial plexus with infection to the musculocutaneous nerve distributed to C5 to C7 root level. Those in the median nerve control ranged in C6 to Th 1 roots. Ulnar nerve control demonstrated traced nerve fibers in the C8 and Th1 roots. In experiment 2, in the median nerve recipient group, most of the regenerating fibers were coming down from far proximally. The ulnar nerve recipient group demonstrated internodal sprouting at near to the end-to-side coaptation site, if there was no demonstrable pathway from the upper brachial plexus roots to the lower roots. A mixture of regenerating fibers from the proximal upper brachial plexus roots and internodal space at near to the coaptation site was demonstrated in the group with a demonstrable pathway from upper to lower roots.

This study suggests the novel fact that neurotropism to attain more appropriate reinnervation, such as sprouting from the original neurons of the injured nerve, may work in an end-to-side nerve repair.

Accuracy of Diagnostic Imaging in Assessing the Integrity of the Flexor Retinaculum Following Failed Carpal Tunnel Release: Pilot Study. Zvi Margaliot (University of Toronto, Department of Surgery, Division of Plastic Surgery, Toronto, Ontario, Canada), Dimitri J. Anastakis, David Salonen, Angela A. Luong, and Edna Becker.

Carpal tunnel syndrome is a common and debilitating condition of the upper extremity. Failure of surgical treatment and persistence of symptoms are reported in 10 to 20% of patients, with 12% of patients requiring reoperation. Incomplete release of the flexor retinaculum is a frequent finding at reoperation. The management of recurrent carpal tunnel syndrome remains a clinical dilemma due, in part, to the lack of a gold standard for assessing the integrity of the flexor retinaculum after previous release. The purpose of this study was to assess the accuracy of non-invasive diagnostic imaging in determining the integrity of the flexor retinaculum after surgical release of the carpal tunnel.

In part 1 of the study, 15 cadaver hands underwent three sequential operations: palmar skin incision, partial division of the flexor retinaculum, and complete division of the flexor retinaculum. T1 and T2 weighted MR images were obtained for each specimen after each procedure, and the protocol images assessed by two blinded radiologists. Sensitivity, specificity, intra-rater, and inter-rater reliability were calculated for the diagnosis of complete or incomplete release using Spearman rank correlation and kappa statistic. Part 2 of the study consisted of a retrospective clinical review of patients with recurrent carpal tunnel syndrome who presented to a single hand surgeon's practice over a 3-year period. Only patients who underwent ultrasound or MR imaging of the wrist, followed by surgical exploration, were included. The radiologic diagnoses were compared to intraoperative findings to estimate sensitivity and specificity.

Correct identification of retinacular integrity in the cadaver was obtained in 92% of specimens, when based on combined T1 and T2 weighted images (sensitivity, 94%; specificity, 93%). Inter-rater reliability, measured by kappa statistic, was very good (κ = 0.71), while intra-rater reliability was excellent (1.0 and 0.8 for radiologist 1 and 2, respectively). Diagnoses based on T1 and T2 weighted images were significantly more accurate and reliable than those based on T1 images alone. Ten patients were included in the clinical pilot cohort. The accuracy of non-invasive imaging, compared with intraoperative findings, was 70% overall (sensitivity = 86%; specificity = 33%).

MRI is sensitive, specific, and reliable in the diagnosis of partially released flexor retinaculum in a cadaver model. Clinically, although the sample size was small, this pilot study suggests a potentially high diagnostic value of non-invasive imaging in the clinical diagnosis of incomplete release of the flexor retinaculum in cases of recurrent carpal tunnel syndrome.