MORPHOLOGICAL AND FUNCTIONAL EVALUATION OFLEG-MUSCLE REINNERVATION AFTER COUPLER COAPTATIONOF THE DIVIDED RAT SCIATIC NERVE

BARBARA S. LUTZ, M.D., Ph.D.,1* and DISA LIDMAN, M.D., Ph.D2

Mechanical couplers are successfully used for microvascular venous anastomoses. The advantages include a simple and fast technique anda high patency rate. Couplers offer a secluded coaptation site, and might also be of use in peripheral nerve repair. The present study wasdesigned to investigate coupler coaptation of the rat sciatic nerve, evaluating the number and locations of motor and sensory neuronsprojecting to the selected muscles as well as stimulation-induced muscle contraction force. Adult rats underwent either suture or couplerrepair after left sciatic nerve transection. In all rats, the experimental side was compared to the healthy right side. Evaluation after 20 weeksincluded retrograde labeling of motoneurons and dorsal root ganglion neurons projecting to the tibial anterior muscle and to the tibial posteriormuscle, histology, muscle contraction force (tibial anterior muscle and gastrocnemius muscle), and a pinch reflex test. The results show thatthe suture and the coupler groups did not differ significantly regarding the examined parameters, except for discrete signs of nerve com-pression at the coaptation site after coupler repair due to fibrous tissue ingrowth. However, this did not impair axonal regeneration. Impor-tantly, axonal outgrowth from the repair site to the surrounding tissue was not observed after coupler coaptation, but it was observed aftersuture repair. These results suggest that couplers may be of value for repair of nerves in adjacency to avoid axonal crisscrossing betweennerves during regeneration. ª 2005 Wiley-Liss, Inc. Microsurgery 25:235�240, 2005.

Mechanical coupler devices have long been successfullyused for microvascular venous anastomoses.1�3 Thistechnique is simple and fast, and has a high patencyrate.3 The use of similar couplers for experimental ratsciatic nerve coaptation showed promising results.4 Su-ture and coupler groups showed similar outcomes withrespect to axon counts and walking-track analyses aftera follow-up time of 14 weeks. Nerves repaired withcouplers showed smooth epineurial outlines, whereasnerves repaired with sutures showed signs of axonalsprouting from the coaptation site into the surroundingtissue and intrusion of scar tissue into the coaptationsite.4,5 This indicates that couplers provide a more se-cluded juncture site than sutures. However, couplersmay cause nerve compression, which bears a long-termrisk of axonal degeneration.4 The present study wasdesigned to investigate this possible risk further. In re-sponse to a previous study,4 investigations were refined,insofar as the number of relevant motoneurons anddorsal root ganglion neurons were counted. Musclecontraction force was assessed, and a pinch reflex test6,7

was used to evaluate regeneration of sensory axons.

MATERIALS AND METHODS

Animals

All experiments were approved by the local ethicscommittee for animal experiments. Twenty adult maleSprague-Dawley rats (300 g) were used. Each rat wasanesthetized with isofluran (Isotec 5 Low Flow Vapor-izer, Ohmeda, Hartfield, UK) during surgery. Addi-tionally, 5% noviform was locally applied to the eyes forprotection, and caprofen (s.c., 0.1 ml/kg) was systemi-cally administered for postoperative pain relief. Duringmuscle contraction force testings and perfusions, keta-min (80 mg/kg) and xylacin (12 mg/kg) were appliedintraperitoneally. The rats had water and standard chowad libitum, and were treated according to the standardsfor laboratory animals established by the SwedishCentral Committee of Laboratory Animal Resources.

Surgical Procedures

The animals were divided into two groups. In groupA (n = 10), transection of the left sciatic nerve atmidthigh was followed by coaptation with two epineural10/0 nylon stitches. In group B (n = 10), transection ofthe left sciatic nerve at midthigh was followed bycoaptation using 1.5-mm couplers (Medical CompaniesAlliance, Inc., Bessemer, AL), which matched thediameter of the sciatic nerves. The epineurium wascarefully set on the spikes of the couplers (rigid plasticrings), and the rings were closed. At completion of theprocedure, no nerve tissue was located between the twocouplers. For geometrical reasons, a fascicular repair ofthe tibial and peroneal fascicles was not possible withthe couplers in the rat.

1Department of Plastic Surgery, University Hospital Orebro, Orebro, andDivision of Cell Biology, Institute for Biology and Surgery, University ofLinkoping, Linkoping, Sweden2Department of Plastic Surgery, Hand Surgery, and Burns, University Hospitalof Linkoping, Linkoping, Sweden

Grant sponsor: Orebro County Council; Grant number: 458/01; Grant spon-sor: University Hospital of Linkoping; Grant sponsor: Swedish ScienceCouncil; Grant number: VR Project 3761.

*Correspondence to: Barbara S. Lutz, M.D., Ph.D., Department of PlasticSurgery, University Hospital Orebro, SE-70 185 Orebro, Sweden. E-mail:bslutz08@hotmail.com

Received 21 June 2005; Accepted 20 September 2004

Published online 4 February 2005 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/micr.20102

ª 2005 Wiley-Liss, Inc.

The right healthy extremity served as control in allanimals.

Retrograde Labeling of Neurons Projecting to the

Tibialis Anterior Muscle (TA) and to the Tibialis

Posterior Muscle (TP)

The purpose of this assessment was to compare thenumber of motoneurons and sensory neurons projectingto the tibialis anterior muscle (TA) and tibialis posteriormuscle (TP) after either method of coaptation, using fiverats in each group.

Five months postoperatively, the rats were reanes-thetized, and the TA and TP were dissected bilaterally.The fluorescent retrograde tracer Fast Blue (FB; Sigma,St. Louis, MO) was injected into the center of the TPusing a Hamilton syringe (3 ll of 3% dye). The fluo-rescent retrograde tracer Diamidino Yellow (DY; Sig-ma) was similarly injected into the TA (5 ll of 2% dye).To avoid dye leakage and contamination of adjacenttissues, the surroundings of the injection sites werecovered with Vaseline-impregnated filter paper, and themuscle fascia at the injection site was sutured with 9/0nylon during removal of the syringe needle. Since theperoneal nerve innervating the TA and the tibial nerveinnervating the TP are separated by the tibia, TA-TPcontamination was unlikely.

After a survival time of 12 days, all animals weredeeply anesthetized and perfused through the heart with250 ml 37�C Ringer solution containing heparin (1,000IE/kg), followed by 500 ml ice-cold 4% paraformalde-hyde in 0.1 M phosphate buffer. The L3�L6 spinal cordsegments and dorsal root ganglia were removed andplaced overnight in a solution of 30% sucrose and 4%phosphate-buffered paraformaldehyde. Transverse seri-al sections (30 lm) were cut from the spinal cord seg-ments with a cryostat and mounted on slides inantifading medium (Dako, Carpinteria, CA). Transversesections (20 lm) from the dorsal root ganglia (DRG)were cut and mounted similarly. In each animal, everysecond section was examined and subjected to countingin a Nikon Eclipse E600 fluorescence microscope with aNikon DXM 1200 digital camera (Tekno Optik AB,Huddinge, Sweden). Using a Nikon filter (DM 430, V-2A, BA 450), FB- and DY-labeled anterior horn andDRG neurons were counted at segmental levels L3�L6.Experimental sides were compared to control sides.8

Counting in spinal cord sections included all labeledventral horn units (both a- and c-motoneurons). Duringcounting, sections were systematically scanned, and eachlabeled neuron was recorded. Great care was taken toavoid counting individual neurons twice or to misscounting a labeled neuron. Three labeling patterns wereobserved: 1) neurons with blue cytoplasm = single-la-

beled with FB from the TP; 2) neurons with yellownucleus = single-labeled with DY from the TA; and 3)neurons with yellow/whitish nucleus and blue cytoplasm= double-labeled with DY and FB from TA and TP.Neurons with dubious labeling were not counted.

Muscle Contraction Force

In five rats of each group, the contractile force ofthe gastrocnemius muscle (GM) and tibialis anteriormuscle (TA) were evaluated bilaterally after a survivaltime of 5 months. The GM is supplied by the tibialnerve, and the TA is innervated by the peroneal nerve.In this experiment, the GM was used instead of the TP,because the latter was too small to allow reliablephysiological experiments. In preparation for thisexperiment, both muscles and the sciatic nerves weredissected free without touching the coaptation sites.The knee and ankle joints were immobilized with K-wires. A 3/0 silk thread was sutured to the respectivetendons and connected to a force transducer (PFI50N,Mecmesin, UK). The electric signal was led through anamplifier (our own construction) and then to a re-corder (Memory Hi Corder, Hiaki 8830, Japan). Astimulating silver electrode was applied to the sciaticnerve 1 cm proximal to the coaptation site (stimulationfrequency, 50 Hz; duration, 0.2 ms). Resting muscletension was adjusted so that maximal active tensionwas obtained.9 For each preparation, the response seenat the optimal muscle length was used as tetanic musclecontraction force. In each animal, the response ob-served on the experimental side was compared to theresponse obtained on the control side.

Pinch Reflex Test and Histology

After completion of muscle contraction forcerecordings, a pinch reflex test was used to assess thepresence of sensory axons.6,7 Starting distally, consecu-tive segments of the sciatic nerve were pinched with apair of forceps. When regenerated sensory axons werepinched, a reflex abdominal muscle contraction wasobserved.6 Subsequently the coaptation site was dis-sected and examined. Specimens were collected andprepared for Toluidine blue and hematoxylin-and-eosin(H&E) histology.

Statistics

Statistics were used to compare the two experimentalgroups after nerve regeneration, and to compareexperimental and control sides within each group.Approximate Student’s t-test for equal and unequalvariances was employed for statistics. P < 0.05 wasconsidered significant.

236 Lutz and Lidman

RESULTS

General Observations

In both groups, several rats showed signs of autot-omy laterally in the left foot. This was more pronouncedin group A (5 of 10 rats) than in group B (2 of 10 rats).Because of the high self-mutilation rate in group A,walking-track analysis was omitted.

When the coaptation site was dissected 5 monthsafter nerve suture, the anesthetized group A rats neededan extra injection of ketamin, indicating sensory nerveoutgrowth into the surrounding tissue. In contrast, thegroup B rats subjected to coupler repair did not show anypain reaction during dissection of the coaptation site.

Gross Examination and Histology

All coaptation sites were macroscopically intact, andthe coupler site did not show any signs of infection orinflammation. The couplers were surrounded by looseconnective tissue, without adhesions to surroundingtissue or capsule formation (Fig. 1). The couplers per sedid not strangle the nerve, and were easily stripped off.However, growth of connective tissue between thenerves and the couplers resulted in discrete nerve com-pression (Fig. 2). Axonal sprouting into the surroundingtissue was not observed in the coupler group B rats bylight microscopy, but this was observed in rats of suturegroup A.

Retrograde Labeling of Neurons projecting to TA

and TP

Anterior horn. On the control side, DY-positive neu-rons labeled from the TA were located laterally in theanterior horn from caudal L3 to, and including, therostral L5 segments. The mean number of DY-labeledneurons was 97 ± 28 (SD; range, 54�140). FB-positiveneurons labeled from the TP were observed from thelower L3 to, and including, the lower L5 segments, beinglocalized more medially in the anterior horn than theneurons labeled from the TA. The mean number of FB-labeled motoneurons per animal was 42 ± 13 (SD;range, 24�58). For both motoneuron pools, a peakoccurrence was found in the segment L4, which is con-sistent with previous data.10,11 On the experimental sidesin both groups, FB- and DY-labeled neurons wereintermingled medially as well as laterally in the anteriorhorn. Most neurons occurred at the L3�L5 levels, but afew neurons were found at the L6 level. Sporadic dou-ble-labeled neurons were found. There were noremarkable differences between groups A and B. In bothgroups, the number of labeled neurons was subnormalon the experimental sides. On average, 83 ± 19 (groupA) and 78 ± 21 (group B) DY-labeled neurons, and 37

± 14 (group A) and 35 ± 12 (group B) FB-labeledneurons per animal were counted in groups A and B. Inorder to compensate for individual variations in theabsolute number of labeled neurons in different animals,the results of counting are presented as percentages ofthe number of labeled cells on the contralateral controlside (Table 1).

Dorsal root ganglia. On the control side, most labeledcells were located in the L4 ganglion, with a few addi-tional cells in the L3 and L5 ganglia. The mean numberof DY-labeled cells per animal was 313 ± 67 (SD;range, 168�389). The corresponding number of FB-la-beled cells was 320 ± 80 (SD; range, 198�431). On thenerve-injured side, labeled cells were found in theL3�L5 dorsal root ganglias (DRGs) without any pref-

Figure 1. Coupler surrounded by loose connective tissue, 5 months

postoperatively. [Color figure can be viewed in the online issue,

which is available at www.interscience.wiley.com.]

Figure 2. Discrete nerve compression induced by growth of con-

nective tissue between coupler and nerve (H&E stain, ·40). Arrowindicates slight nerve compression. [Color figure can be viewed in

the online issue, which is available at www.interscience.wiley.com.]

Nerve Coaptation With Couplers 237

erence. Double-labeled cells were sporadically observedin both groups. In both groups, the number of labeledDRG neurons was significantly subnormal on theexperimental side as compared to the control side. Therewere no significant differences between groups A and B(Table 1), with an average of 284 ± 86 (group A) and268 ± 63 (group B) DY-labeled cells and 302 ± 87(group A) and 294 ± 76 (group B) FB-labeled cells.. -Muscle Contraction Force

In both groups, muscle contraction force (MCF)values obtained on the experimental side were signifi-cantly lower than the corresponding control values(Table 2). In this respect, groups did not differ statisti-cally.

Pinch Reflex Test

All animals showed a positive response to the pinchreflex test, indicating the presence of sensory axonsdistal to the coaptation site.

DISCUSSION

Aberrant reinnervation is a common feature afternerve transection and repair,12�18 and was also observedin the present study. Prevention of crisscrossing ofregenerating axons between two nerves in adjacency canimprove the functional outcome after nerverepair.13�15,17 The achievement of a sealed coaptationsite offered by the couplers suggested that couplersmight be useful in preventing this type of aberrantreinnervation. In the present study, the couplers werenot used as a ‘‘barrier’’ between two adjacent nervefascicles due to they were too large to use in rats. In-stead, possible side effects of coupler coaptation wereevaluated in the rat sciatic nerve.

In agreement with previous investigators, who usedsimilar couplers,4 we did not find any significant differ-ences between the sutured group and the coupler groupafter a follow-up time of 20 weeks. Supplemental toprevious investigations,4 which included axon countsand walking-track analysis only, we did not observesignificant A/B-differences with respect to nerve cellcounts. The muscle contraction force testing of tibialnerve- and peroneal nerve-related muscles revealed

comparable outcomes in the sutured and couplergroups. The pinch reflex test, though a rough investi-gation, showed the presence of regenerated sensory ax-ons in all animals. Hence, although the growth offibrous tissue between couplers and nerves caused adiscrete nerve compression, this did not have any majornegative impact.

Growth of fibrous tissue between a biodegradabletube and a repaired nerve was assumed to be due to anoversized internal diameter of the tube.19 In the presentexperiment, the inner diameter of the couplers matchedthe diameter of the sciatic nerve, but the growth ofconnective tissue was not completely prevented. On theother hand, the couplers were rigid and thereby pro-tected the repair site against radial forces20,21 fromsurrounding tissues. When a silicone cuff was used as asheath around a nerve coaptation site, such forces wereimposed by surrounding fibrous tissue formation20,21

and induced nerve compression. This impaired neuronalsurvival,16 nerve morphology,22�24 nerve conductionvelocity,24 and walking-track performance.25 A foreign-body reaction with outgrowth of myofibroblasts wasassumed to be the main reason for the formation offibrous tissue.22,26 A perineurial foreign-body reactionalso seems to be associated with other materials. This isimportant, because all materials which are used as a‘‘sheath’’ around a nerve coaptation site can theoreti-cally prevent aberrant centrifugal escape of regeneratingaxons from the coaptation site, as observed in the cou-pler group.

Table 1. Relative Numbers of Anterior Horn Neurons and Dorsal Root Ganglion Neurons After Retrograde Tracing From Anterior TibialMuscle and Posterior Tibial Muscle�

Anterior Horn Dorsal Root Ganglia

Group n Ant. tib. muscle DY Post. tib. muscle FB Ant. tib. muscle DY Post. tib. muscle FBA 5 89.8 ± 16 87.4 ± 13* 91.6 ± 3* 93.4 ± 5*B 5 81.0 ± 6* 86.2 ± 5* 88.4 ± 3* 92.8 ± 3*

�Numbers are expressed as percentages of numbers on contralateral healthy side. (± standard deviation; DY; Diamidino Yellow; FB; Fast Blue).*Indicates significant difference compared to control side (P < 0.05). There are no significant differences between groups A and B.

Table 2. Averaged Muscle Contraction Force of Left Gastro-cnemius Muscle and Left Tibialis Anterior Muscle, Expressed asPercentage of MCF Seen in Corresponding Muscles on Con-

tralateral Control Side*

Group N MCF (%) P-value to control side

GMA 5 77.2 ± 11.3 0.005B 5 79.3 ± 15.3 0.01TAA 5 54.2 ± 22.7 0.005B 5 52.3 ± 9.3 0.0005

*There were no statistical differences between groups A and B (± standarddeviation, P < 0.05 was considered statistically significant).

238 Lutz and Lidman

For example, when absorbable ‘‘sheathing’’ materi-als such as polyglycolic acid (PGA)25 or poly-b-hy-droxybutyrate (PHB)27 were employed for end-to-endsuture sites, no significant advantages were found com-pared to conventional epineural suture repairs without asheath. A fibrous capsule was observed at 6 months,when tubes made of PHB were wrapped around dividedcat nerves.28 The use of collagen tubes to cover nervesuture sites was recently advocated,29,30 but the collagentube method did not give better results than standardsutures at 330 or 1229 months. On the other hand,Madorsky et al.30 reported that nerve repair with col-lagen tubes gave inferior results in terms of motor andsensory neuron counts compared to conventional epi-neural suture repair. Although contractile myofibro-blasts were less evident around collagen tubes thanaround silicone tubes,20 a foreign-body reaction was stillobvious.20,31,32 As pointed out above, nerve compres-sion due to radial forces does occur, if the employed cuffmaterial is soft. However, the couplers used in thepresent experiment were stiff and prevented nervecompression by radial forces from the surrounding tis-sue. The observed limited growth of fibrous tissue be-tween coupler and nerve did not have any obviousnegative effects on nerve regeneration within a follow-uptime of 20 weeks. However, we cannot rule out that thecouplers may cause notable nerve compression and/orlocal irritation after longer survival times, as observed inhuman nerves repaired with silicone tubes.33,34 Thecouplers offer the advantage of preventing axonalsprouting from the coaptation site into surroundingtissue. This advantage was evident in our own and aprevious experiment,4 regardless of whether two (groupA) or five4 stitches were used for the conventional sciaticnerve coaptation.

CONCLUSIONS

Coupler coaptation of nerve fascicles in adjacencyhas the potential to avoid crisscrossing of regeneratingaxons between fascicles without impairing nerve regen-eration. Though a removal, which was easily done in theexperiment, might be necessary in the long term, thisadvantage may be of clinical value for the repair ofnerves in adjacency such as the fascicles of the sciaticnerve, of the facial nerve, or of the brachial plexus,aiming at improved motor recovery.

ACKNOWLEDGMENTS

We thank Prof. Claes Hildebrand, University ofLinkoping, Linkoping, Sweden, for his valuable com-ments and suggestions on the manuscript. Results ofgroup A were also used in a different work.

REFERENCES

1. Berggren A, Ostrup LT, Ragnarsson R. Clinical experience withthe Unilink/3M precise microvascular anastomotic device. Scand JPlast Reconstr Hand Surg 1993;27:35�39.

2. Ostrup LT, Berggren A. The Unilink instrument system for fastand safe microvascular anastomosis. Ann Plast Surg 1986;17:521�525.

3. Shaw WW. What is the best technique for venous anastomosis? Issuturing of vessels obsolete? J Reconstr Microsurg 1997;13:257�262.

4. Prevel CD, Eppley BL, McCarty M, Brock C. Mechanical anas-tomosis of nerves: a histological and functional comparison toconventional suturing. Ann Plast Surg 1994;33:600�605.

5. Marshall DM, Grosser M, Stephanides MC, Keeley RD, RosenJM. Sutureless nerve repair at the fascicular level using a nervecoupler. J Rehabil Res Dev 1989;26:63�76.

6. Kanje M, Lundborg G, Edstrom A. A new method for studies ofthe effects of locally applied drugs on peripheral nerve regenerationin vivo. Brain Res 1988;439:116�121.

7. Lubinska L, Olekiewicz M. The rate of regeneration of amphibianperipheral nerves at different temperatures. Acta Biol Exp (War-saw) 1959;15:125�145.

8. Aldskogius H, Molander C, Persson J, Thomander L. Specific andnonspecific regeneration of motor axons after sciatic nerve injuryand repair in the rat. J Neurol Sci 1987;80:249�257.

9. Lutz BS, Ma SF, Chuang DC, Wei FC. Effects of systemicallyapplied IGF-1 on motor nerve recovery after peripheral nervetransection and repair in the rat—a functional study. Hand Surg1999;4:131�136.

10. Nikolopoulos-Stournaras S, Iles JF. Motor neuron columns in thelumbar spinal cord of the rat. J Comp Neurol 1983;217:75�85.

11. Swett JE, Wikholm RP, Blanks RH, Swett AL, Conley LC.Motoneurons of the sciatic nerve. Exp Neurol 1986;93:227�252.

12. Bodine-Fowler SC, Meyer RS, Moskovitz A, Abrams R, BotteMJ. Inaccurate projection of rat soleus motoneurons: a compari-son of nerve repair techniques. Muscle Nerve 1997;20:29�37.

13. Kline DG, Kim D, Midha R, Harsh C, Tiel R. Management andresults of sciatic nerve injuries: a 24-year experience. J Neurosurg1998;89:13�23.

14. Lutz BS, Ma SF, Chuang DC, Chan KH, Wei FC. Interposition ofa pedicle fat flap significantly improves specificity of reinnervationand motor recovery after repair of transected nerves in adjacency.Plast Reconstr Surg 2001;107:116�123.

15. Lutz BS, Ma SF, Chuang DC, Lidman D, Wei FC. Specificity ofreinnervation and motor recovery after interposition of an arti-ficial barrier between transected and repaired nerves in adja-cency�an experimental study in the rat. Acta Neurochir (Wien)2001;143:393�399.

16. Madison RD, Archibald SJ, Brushart TM. Reinnervation accuracyof the rat femoral nerve by motor and sensory axons. J Neurosci1996;16:5698�5703.

17. Meyer RS, Abrams RA, Botte MJ, Davey JP, Bodine-Fowler SC.Functional recovery following neurorrhaphy of the rat sciaticnerve by epineural repair compared with tubulization. J OrthopRes 1997;15:664�669.

18. Rende M, Granato A, Lo Monaco M, Zelano G, Toesca A.Accuracy of reinnervation by peripheral nerve axons regeneratingacross a 10-mm gap within a permeable chamber. Exp Neurol1991;111:332�339.

19. Meek MF, Den Dunnen WFA, Bartels HL, et al. Peripheral nerveregeneration and functional nerve recovery after reconstructionwith a thin-walled biodegradable poly (DL-lactide-�-caprolactone)nerve guide. Cells Mater 1997;7:53�61.

20. Chamberlain LJ, Jannas IV, Hsu HP, Spector M. Connective tis-sue response to tubular implants for peripher nerve regeneration:the role of myofibroblasts. J Comp Neurol 2000;417:415�430.

21. Ducker TB, Hayes GJ. Experimental improvements in the use ofsilastic cuff for peripheral nerve repair. J Neurosurg 1968;28:582�587.

Nerve Coaptation With Couplers 239

22. Chamberlain LJ, Yannas IV, Arrizabalaga A, Hsu HP, NorregaardTV, Spector M. Early peripheral nerve healing in collagen and sil-icone tube implants: myofibro-blasts and the cellular response. Bi-omaterials 1998;19:1393�1403.

23. Heijke GCM, Klopper PJ, Baljet B, van Doorn IB. Silicone rubbertubulization in peripheral sensory nerve reconstruction: an exper-imental study in rabbits. Microsurgery 2001;21:306�316.

24. Smahel J, Meyer VE, Morgenthaler W. Silicone cuffs for periph-eral nerve repair: experimental findings. J Reconstr Microsurg1993;9:293�297.

25. Lolley RD, Bose WJ, Bastian F, Bassam B, Meyer FN, AndersonLD. Vein, silastic, and polyglycolic acid fine mesh: a comparativestudy in peripheral nerve repair. Ann Plast Surg 1995;35:266�271.

26. Rudolph R, Abraham J, Vecchione T, Guber, Woodward M.Myofibroblasts and free silicon around breast implants. PlastReconstr Surg 1978;62:185�196.

27. Hazari A, Johansson-Ruden G, Junemo-Bostrom K, Ljungberg C,Terenghi G, Green C, Wiberg M. A new resorbable wrap-aroundimplant as an alternative nerve repair technique. J Hand Surg [Br]1999;24:291�295.

28. Gogolewski S, Jovanovic M, Perren SM, Dillon JG, Hughes MK.Tissue response and in vivo-degradation of selected polyhydroxy-acids: polylactides (PLA), poly(3-hydroxybutyrate) (PHB), and

poly(3-hydroxybutyrate-co-3-hydroxyvalerate) (PHB-VA). J Bio-med Mater Res 1993;27:1135�1148.

29. Heijke GCM, Klopper PJ, Van Doorn IBM, Baljet B. Processedporcine collagen tubulization versus conventional suturing inperipheral nerve reconstruction: an experimental study in rabbits.Microsurgery 2001;21:84�95.

30. Madorsky SJ, Swett JE, Crumley RL. Motor versus sensory neu-ron regeneration through collagen tubes. Plast Reconstr Surg1998;102:430�438.

31. Chen HH, Liu HM. The use of collagen polymer tube and fibrinclot in peripheral nerve repair. Proc Natl Sci Counc Repub China[B] 1994;18:58�63.

32. Navarro X, Rodriguez FJ, Labrador RO, Buti M, Ceballos D,Gomez N, Cuadras J, Perego G. Peripheral nerve regenerationthrough bioresorbable and durable nerve guides. J Periph NervSyst 1996;1:53�64.

33. Braga-Silva J. The use of silicone tubing in the late repair of themedian and ulnar nerves in the forearm. J Hand Surg [Br]1999;24:703�706.

34. Dahlin LB, Anagnostaki L, Lundborg G. Tissue response tosilicone tubes used to repair human median and ulnar nerves.Scand J Plast Reconstr Surg Hand Surg 2001;35:29�34.

240 Lutz and Lidman