- Email: [email protected]
Flexor tendon-tendon sheath interaction after tendon grafting: A biomechanical study in a human model in vitro
Flexor Tendon–Tendon Sheath Interaction After Tendon Grafting: A Biomechanical Study in a Human Model In Vitro Jun Nishida, MD, Morioka, Japan, Peter C. Amadio, MD, Paul C. Bettinger, MD, Kai-Nan An, PhD, Rochester, MN
A human cadaver tendon sheath model was used to study the differences in excursion resistance of tendons that might be considered as sources of clinical tendon grafts. The flexor digitorum profundus and superficialis tendons, the extensor indicis proprius tendon used in its normal proximal-distal orientation, the extensor indicis proprius tendon used in a reversed distal–proximal orientation, and the palmaris longus tendon were studied in 7 fingers. The intrasynovial tendons (the flexor digitorum profundus and superficialis tendons and the reversed extensor indicis proprius tendon) produced less excursion resistance (p , .05) than the extrasynovial tendons (the normally oriented extensor indicis proprius tendon and the palmaris longus tendon). In contrast to studies measuring resistance against a single pulley, resistance within a complete tendon sheath may be affected by contact with other structures, particularly in joint extension. (J Hand Surg 1999;24A:1097–1102. Copyright © 1999 by the American Society for Surgery of the Hand.) Key words: Gliding resistance, tendon graft.
The clinical results of flexor tendon grafting depend on restoration of grip strength and active range of motion of the finger.1–7 Postoperative adhesions between the grafted tendon and the surrounding tissues are frequently formed and limit the excursion of the tendon.2– 4,7 It recently has been suggested that tendon grafts of intrasynovial origin may perform better than those of extrasynovial origin8 –11 and that resistance to gliding under the A2 pulley is greater From the Department of Orthopedics, Mayo Clinic and Mayo Foundation, Rochester, MN. Received for publication August 27, 1997; accepted in revised form March 17, 1999. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. Reprint requests: Peter C. Amadio, MD, Mayo Clinic, 200 First St SW, Rochester, MN 55905. Copyright © 1999 by the American Society for Surgery of the Hand 0363-5023/99/24A05-0002$3.00/0
for the palmaris longus, an extrasynovial tendon, than for the flexor digitorum profundus, an intrasynovial tendon.12 A system has been developed that allows quantitative measurement of the excursion resistance between a tendon and its associated pulley.13,14 This technique has the advantage that the tendon–pulley interaction can be measured directly. Although friction between the tendon sheath and the tendon after repair or grafting may be an important factor in ultimate tendon excursion, the differences in excursion resistance between the entire tendon sheath and tendon, comparing different graft sources, have not been reported. The purpose of this study was to evaluate the difference in excursion resistance between intrasynovial and extrasynovial tendon graft sources against an intact sheath, including the A1–A4 pulleys and the phalanges and metacarpal, in an in vitro human model. The Journal of Hand Surgery 1097
1098 Nishida et al / Flexor Tendon–Tendon Sheath Interaction
Materials and Methods Seven fresh-frozen middle finger specimens were used. The specimens were obtained from 3 female and 3 male cadavers with ages ranging from 61 to 85 years (average, 74 years). Specimens were thawed immediately before testing. All specimens were kept moist in a saline bath throughout the testing procedure. The tendons used were the ipsilateral palmaris longus tendon, the ipsilateral extensor indicis proprius tendon, the ipsilateral flexor digitorum superficialis tendon of the ring finger, and the homodigital (that is, middle finger) flexor digitorum profundus tendon. Because the proximal part of the extensor indicis proprius passes beneath the extensor retinaculum and may therefore be considered intrasynovial, this tendon was tested twice, once for its distal extrasynovial portion and once for its proximal subretinacular portion. These 2 segments were tested simply by using both a distal–proximal and a proximal– distal orientation of the extensor indicis proprius. In the normal proximal– distal orientation, the subretinacular portion of the tendon remained proximal to the tendon sheath throughout testing. In the reversed distal–proximal orientation, the subretinacular portion of the tendon remained within the tendon sheath throughout the testing. Specimen preparation was identical for each hand. The palmaris longus and extensor indicis proprius tendons were harvested from their insertion to their musculotendinous junction. Any peritenon was left attached to the tendon. The flexor digitorum superficialis tendon of the ring finger was similarly harvested from its insertion to the musculotendinous junction. For the middle finger, the finger was disarticulated from the hand through the proximal portion of the metacarpal; care was taken to preserve the flexor tendons to the musculotendinous junction and the tendon sheath. A transverse incision through the synovial sheath was made just distal to the A4 pulley to mark the surface of the flexor digitorum profundus tendon with the finger in full extension. The flexor digitorum profundus tendon was then pulled proximally until full proximal interphalangeal and distal interphalangeal joint flexion was achieved. In this position, the tendon was again marked through the previous incision. The distance between these 2 marks represented the physiologic excursion range of the flexor digitorum profundus tendon for that finger. The flexor digitorum superficialis tendon was removed from the middle finger by pulling the tendon
proximally, cutting its insertion and short vinculum through a window in the C2 pulley, and cutting the long vinculum proximal to the A1 pulley. The incision in the C2 pulley was repaired with a 6 – 0 nylon suture. The tendon sheath, including A1–A4 pulleys, C1–C3 pulleys, the parietal membrane in this region, and the visceral membrane of the flexor digitorum profundus tendon, were all preserved. The finger was disarticulated at the distal interphalangeal joint, detaching the flexor digitorum profundus from the distal phalanx. A 1.5-mm K-wire was inserted through the proximal phalanx parallel to the longitudinal axis of the bone. This wire was used to stabilize the specimen in the mounting jig. The measurement system consisted of one custombuilt mechanical actuator with a linear potentiometer, 2 custom-made tensile load transducers accurate to 1 g,13 and a pulley13,14 (Fig. 1). The palmar side of the bone sheath preparation faced upward and the proximal end faced toward the actuator. Each tendon to be tested was prepared similarly. After the tendon was aligned in the proper proximal– distal orientation, it was passed through the tendon sheath. Load transducers were connected to the proximal and distal ends of the tendon with Dacron cord. The proximal load transducer (F2) was connected to the actuator. The distal transducer (F1) was connected to a
Figure 1. Experimental setup for the measurement of excursion resistance between the tendon for graft material and the adjacent tissues. Tensions of F1 and F2 are measured by the tensile load transducers. Excursion is measured by a linear potentiometer. a and b are the angles between the proximal and distal tendon ends, respectively, and the reference axis of the proximal phalanx. (Modified and reprinted with permission.12)
The Journal of Hand Surgery / Vol. 24A No. 5 September 1999 1099
250-g weight. The actuator was positioned at the preselected angle a, which was defined as the angle in degrees formed between the horizontal plane of the proximal phalanx and the proximal cable extension. This motion occurred primarily at the metacarpophalangeal joint. The mechanical pulley between the load and the distal load transducer was positioned at the preselected angle b, which was defined as the angle formed between the horizontal plane of the proximal phalanx and the distal cable extension. This motion occurred primarily at the proximal interphalangeal joint. The sum of angles a and b was considered the angle of the arc of contact. The tendon to be tested was pulled proximally at a rate of 2.0 mm/s by the actuator and was opposed by the weight. The movement of the tendon toward the actuator was regarded as flexion. F1, F2, and the corresponding excursion were recorded by computer at a sampling rate of 10 Hz. Excursion was limited to the distance between the 2 tendon markers on the profundus tendon of that finger. Angles a and b were varied to 5 different positions by raising or lowering the actuator and weight. Three trials were performed for each of the 5 positions (a, b): 15, 5; 20, 10; 30, 10; 30, 20; and 30, 30. These positions, and the testing rate, were selected because they were comparable to those used in previous studies.12,14,15 Each graft was tested once, except for the extensor indicis proprius tendon, which was tested once in each of 2 orientations. The order of testing in each finger was palmaris longus, reversed extensor indicis proprius, normally oriented extensor indicis proprius, flexor digitorum superficialis, and flexor digitorum profundus. Plots of F1 and F2 measurements versus excursion were examined for each trial and evaluated by their shape. Because the trials were generally identical and the first run was considered to be preconditioning, the average of the last 2 runs was selected for analysis for each angle. The mean force differences of F2 and F1 for the whole excursion were obtained and regarded as the resistance at the interface between the tendon and the pulley for the given arc of contact. The tendons were classified as follows: intrasynovial sources (flexor digitorum superficialis, flexor digitorum profundus, and reversed extensor indicis proprius) and extrasynovial sources (palmaris longus and normally oriented extensor indicis proprius). Significant differences in excursion resistance between tendons and at different angles were assessed with a 2-factor repeated-measures ANOVA. Tukey’s honestly significant difference test was then used for
a post hoc comparison of individual means in the mean effects.16 The level of significance was set at a 5 .05. To compare intrinsic versus extrinsic tendons, a 1-factor repeated-measures ANOVA with an orthogonal polynomial breakdown for means across angles of measurement was used.16 A significance level of a 5 .01 was used to correct for the error rate of a second ANOVA test.
Results The results are shown in Figure 2. ANOVA showed no difference between results when characterized by finger or by angle. There was a significant difference when the results were characterized by tendon (p , .01). Specifically, the normally oriented extensor indicis proprius was significantly (p , .05) different from the flexor digitorum profundus, the flexor digitorum superficialis, and the reversed extensor indicis proprius tendons. The intrasynovial tendons as a group were significantly different from the extrasynovial tendons (p , .01). Univariate and multivariate repeated-measures analysis within subjects showed that the intrasynovial and extrasynovial groups were affected differently by joint angle; specifically, as contact angle increased, the resistance of intrasynovial tendons tended to decrease, whereas for extrasynovial tendons the resistance tended to increase (p , .001). The excursion resistance of the extrasynovial tendons was significantly higher than that of the intrasynovial tendons (p , .01). Among extrasynovial tendons, the excursion resistance was higher in the normally oriented extensor indicis proprius tendon than in the palmaris longus tendon, and these differences were statistically significant at all angles (p , .05). There were no differences in resistance between the 3 intrasynovial tendons. The mean length of the subretinacular portion of the extensor indicis proprius was 25 mm compared with a mean A1–A4 length of 71 mm (range, 61–79 mm).
Discussion The concept of friction measurement and its application to the tendon–pulley unit has been the focus of previous reports.12–15 We extend the investigation to include other components of the tendon sheath, beyond the basic single tendon–single pulley interaction. A tendon sliding through a curved pulley is analogous to a belt wrapped around a fixed mechanical pulley.13,14 If the tensions in the belt are F1 and F2
1100 Nishida et al / Flexor Tendon–Tendon Sheath Interaction
Figure 2. Comparison between intrasynovial and extrasynovial tendons: flexor digitorum superficialis tendon (FDS), palmaris longus tendon (PL), normally oriented extensor indicis proprius tendon (EIP), reversed extensor indicis proprius tendon (EIP-R), and flexor digitorum profundus (FDP). Error bars represent standard error of the mean.
on each end of the tendon, respectively, and if the impending motion of the tendon is from F1 to F2, then F2 is greater than F1 because of the friction (f), and f 5 F2 2 F1. In previous studies,12–14 bone was carefully excised so that a pure tendon–A2 interaction could be evaluated. In this study, in contrast, the bones and joints were not disturbed, so F2 and f also potentially include the friction between the tendon and other surrounding tissues, including bone. Our hypothesis was that intrasynovial tendon would have less resistance to gliding under an intact tendon sheath than would extrasynovial tendon. This is, in fact, what we observed, and we believe that our data thus far support the proposed hypothesis. We also observed some differences in excursion resistance when comparing A1–A4 with previously reported work on resistance versus A2 alone.12 These differences are discussed below. The final goal of tendon grafting is to achieve maximal excursion of the tendon with as little resistance as possible. The kinematics of tendon gliding are complex. The surface of the flexor digitorum profundus tendon contacts its guiding pulleys as well as the flexor digitorum superficialis tendon and the phalangeal bases.17 Each type of interaction (tendon– pulley, tendon–tendon, tendon– bone) may have its own characteristic friction relationship. Most investigators have discussed tendon graft sources from the viewpoint of only utility or convenience.3,18 –20 An important body of experimental
evidence suggests, however, that the source of the tendon graft may be an important factor in determining the clinical result of the grafting procedure. Intrasynovial tendons heal with less cellular necrosis and with less extensive adhesions than extrasynovial tendons.9 –11 Intrasynovial tendon grafts survive better than extrasynovial tendon grafts.9,11 In addition, intrasynovial grafts integrate with less scar formation between the tendon surface and the surrounding tissues.9,11 Intrasynovial tendon grafts show significant differences from extrasynovial tendon grafts in both surface morphology and vascularity.10,11 The synthesis of proteoglycan matrix proteins and DNA is different between these 2 types of tendon surfaces.8 Surface lubrication mechanisms are also different between these 2 types of tendons,15 and excursion resistance of extrasynovial tendons exceeds that of intrasynovial tendons in a tendon–A2 pulley model.12 We believe that our data add to this body of experimental evidence, supporting the advantages of intrasynovial tendons as sources of tendon grafts. Furthermore, the measured excursion resistance between the flexor digitorum profundus and the A1–A4 pulleys at the highest angle (a, b 5 30°, 30°, or 60°) was not greater than that recorded in a previous study that measured resistance between the flexor digitorum profundus and the A2 pulley alone.12 Because friction is not dependent on contact area, this finding is consistent with the hypothesis that we are truly measuring tendon–pulley friction;
The Journal of Hand Surgery / Vol. 24A No. 5 September 1999 1101
that is, even though the tendon is contacting a much longer pulley in A1–A4 than A2, no increase in resistance was observed. In contrast with previous studies12,21 that evaluated friction between the tendon and A2 pulley, in this study of tendon contact with a more complete sheath we found a higher resistance for the tendons at lower angles of a and b but not at higher angles. We believe that this can be explained by tendon contact with other components of the tendon sheath complex, such as bone.17 The bony element was specifically excluded in previous studies by excising the bony condyles.12–14 Direct observation by Lin et al.17 showed that loss of contact of tendon with underlying bone occurred at approximately 30° of flexion at the proximal interphalangeal joint in normal cadaver fingers. At a lower angle, the excursion resistance may therefore include friction not only between the tendon and pulley but also between the tendon and bone. Although we did not directly measure the friction between the tendon and bone, the fact that excursion resistance was greater at lower values of a and b, when it would be reasonable to assume that tendon made contact with bone, suggests that the coefficient of friction may well be greater for the tendon– bone interaction than it is for the tendon– pulley interaction. Supporting this hypothesis, at higher values of a and b, where tendon– bone contact no longer occurs, this difference disappears. This effect, again in comparison with previous studies of resistance against A2 alone, is greater for the flexor digitorum superficialis and flexor digitorum profundus, that is, the steadily increasing resistance at higher angles observed for all tendons against the A2 pulley alone12,21 and for the extrasynovial tendons versus the entire sheath in this study is not observed for the flexor digitorum profundus and superficialis versus the entire sheath. We believe this is due to the effect of tendon cross-section, which is greater for the flexor digitorum profundus and flexor digitorum superficialis tendons than it is for the extensor indicis proprius or palmaris longus tendons. This greater cross-section does not, of course, increase friction per se because frictional force is not affected by the surface area of contact. Increased cross-section, however, does allow these larger tendons to contact other surfaces, such as bone, which may have a different coefficient of friction than tendon sheath. We believe that this contact with other structures within the sheath system explains the difference we observed, and we plan to investigate this with future studies of friction of tendon versus the various components of
an intact sheath system, for example, tendon/tendon and tendon/bone. There was not a statistically significant difference between the excursion resistance of the palmaris longus tendon and the intrasynovial tendons. We believe that this relates more to sample size than to anything else because in previous studies12,21 we have been able to demonstrate differences in excursion resistance between the palmaris longus and the intrasynovial tendons. In this study there was certainly a trend in that direction, even though it did not meet the test of statistical significance. We did not think it worthwhile to test more specimens simply to get the palmaris longus to a level of statistical significance because we have demonstrated that previously. We do not believe that the order of testing, which was the same in each specimen in this study, can explain the differences we observed. We have demonstrated previously that the order of testing does not affect the results in our system, even when harsh enzymatic treatments are used15 or when the tendon is subjected to as many as 100 cycles of testing (unpublished data). This study does have several limitations. We did not duplicate normal finger kinematics, and the gliding of the tendons was not physiologic because the proximal phalanx was fixed and the distal interphalangeal joint was disarticulated. A method of measurement of excursion resistance that permits normal joint movement should be developed. The resistance to gliding of the tendons in this study was measured in a sheath without a flexor digitorum superficialis tendon. The effect of inclusion of the flexor digitorum superficialis tendon on gliding resistance should be evaluated in the future. Finally, the results of this in vitro study need to be confirmed in vivo. We acknowledge that intrasynovial donor tendons are easier to find in an experimental model than in a clinical situation. We believe, however, that occasions do arise in which an otherwise unsalvageable digit may serve as a tendon donor source, and we are optimistic that in the future it may be possible to consider other sources, such as allografts or even heterografts, for tendon reconstruction. Furthermore, in the future it may be possible to genetically alter the gliding characteristics of extrasynovial tendons.22 We believe that the strength of this study lies in our ability to directly measure the complex interaction of a tendon sheath that extends both proximal and distal to the previously studied A2, which also affords an opportunity to study tendon interactions in
1102 Nishida et al / Flexor Tendon–Tendon Sheath Interaction
a more complex environment. This in vitro model supports the hypothesis that intrasynovial tendons have advantages over extrasynovial tendons for tendon grafting. In addition to the flexor digitorum superficialis and flexor digitorum profundus tendons, the reversed extensor indicis proprius tendon also has excursion resistance properties consistent with an intrasynovial source. An in vivo animal model to test the properties of the grafted tendon with healing over time will be necessary in the future because in vitro models cannot address the complex effects of wound healing on tendon gliding. The authors thank David A. Gabriel, PhD, for assistance with statistical methods and analysis.
References 1. Amadio PC, Wood MB, Cooney WP III, Bogard SD. Staged flexor tendon reconstruction in the fingers and hand. J Hand Surg 1988;13A:559 –562. 2. Boyes JH, Stark HH. Flexor-tendon grafts in the fingers and thumb. A study of factors influencing results in 1000 cases. J Bone Joint Surg 1971;53A:1332–1342. 3. Hunter JM, Salisbury RE. Flexor-tendon reconstruction in severely damaged hands. A two-stage procedure using a silicone-Dacron reinforced gliding prosthesis prior to tendon grafting. J Bone Joint Surg 1971;53A:829 – 858. 4. McClinton MA, Curtis RM, Wilgis EF. One hundred tendon grafts for isolated flexor digitorum profundus injuries. J Hand Surg 1982;7A:224 –229. 5. Silfverskio¨ld KL, May EJ, To¨rnvall AH. Tendon excursions after flexor tendon repair in zone II: results with a new controlled-motion program. J Hand Surg 1993;18A:403–410. 6. Small JO, Brennen MD, Colville J. Early active mobilisation following flexor tendon repair in zone 2. J Hand Surg 1989;14B:383–391. 7. Strickland JW. Flexor tendon injuries. Part 4. Staged flexor tendon reconstruction and restoration of the flexor pulley. Orthop Rev 1987;16:78 –90. 8. Abrahamsson SO, Gelberman RH, Lohmander SL. Variations in cellular proliferation and matrix synthesis in intrasynovial and extrasynovial tendons: an in vitro study in dogs. J Hand Surg 1994;19A:259 –265.
9. Gelberman RH, Seiler JG III, Rosenberg AE, Heyman P, Amiel D. Intercalary flexor tendon grafts. A morphological study of intrasynovial and extrasynovial donor tendons. Scand J Plast Reconstr Surg Hand Surg 1992;26: 257–264. 10. Gelberman RH, Chu CR, Williams CS, Seiler JG III, Amiel D. Angiogenesis in healing autogenous flexor-tendon grafts. J Bone Joint Surg 1992;74A:1207–1216. 11. Seiler JG III, Gelberman RH, Williams CS, et al. Autogenous flexor-tendon grafts. A biomechanical and morphological study in dogs. J Bone Joint Surg 1993;75A:1004 – 1014. 12. Uchiyama S, Amadio PC, Coert JH, Berglund LJ, An K-N. Gliding resistance of extrasynovial and intrasynovial tendons through the A2 pulley. J Bone Joint Surg 1997;79A: 219 –224. 13. An K-N, Berglund L, Uchiyama S, Coert JH. Measurement of friction between pulley and flexor tendon. Biomed Sci Instrum 1993;29:1–7. 14. Uchiyama S, Coert JH, Berglund L, Amadio PC, An K-N. Method for the measurement of friction between tendon and pulley. J Orthop Res 1995;13:83– 89. 15. Uchiyama S, Amadio PC, Ishikawa J-I, An K-N. Boundary lubrication between the tendon and the pulley in the finger. J Bone Joint Surg 1997;79A:213–218. 16. Kirk RE. Experimental design: procedures for the behavioral sciences. Belmont, CA: Brooks/Cole Publishing Company, 1968. 17. Lin GT, Amadio PC, An K-N, Cooney WP. Functional anatomy of the human digital flexor pulley system. J Hand Surg 1989;14A:949 –956. 18. Harvey FJ, Chu G, Harvey PM. Surgical availability of the plantaris tendon. J Hand Surg 1983;8A:243–247. 19. Wehbe´ MA. Tendon graft donor sites. J Hand Surg 1992; 17A:1130 –1132. 20. White WL. Tendon grafts: a consideration of their source, procurement and suitability. Surg Clin North Am 1960;40: 403– 413. 21. Nishida J, Amadio PC, Bettinger PC, An K-N. Excursion properties of tendon graft sources: interaction between tendon and A2 pulley. J Hand Surg 1998;23A:274 –278. 22. Lou J, Manske PR, Aoki M, Joyce ME. Adenovirus-mediated gene transfer into tendon and tendon sheath. J Orthop Res 1996;14:513–517.
A human cadaver tendon sheath model was used to study the differences in excursion resistance of tendons that might be considered as sources of clinical tendon grafts. The flexor digitorum profundus and superficialis tendons, the extensor indicis proprius tendon used in its normal proximal-distal orientation, the extensor indicis proprius tendon used in a reversed distal–proximal orientation, and the palmaris longus tendon were studied in 7 fingers. The intrasynovial tendons (the flexor digitorum profundus and superficialis tendons and the reversed extensor indicis proprius tendon) produced less excursion resistance (p , .05) than the extrasynovial tendons (the normally oriented extensor indicis proprius tendon and the palmaris longus tendon). In contrast to studies measuring resistance against a single pulley, resistance within a complete tendon sheath may be affected by contact with other structures, particularly in joint extension. (J Hand Surg 1999;24A:1097–1102. Copyright © 1999 by the American Society for Surgery of the Hand.) Key words: Gliding resistance, tendon graft.
The clinical results of flexor tendon grafting depend on restoration of grip strength and active range of motion of the finger.1–7 Postoperative adhesions between the grafted tendon and the surrounding tissues are frequently formed and limit the excursion of the tendon.2– 4,7 It recently has been suggested that tendon grafts of intrasynovial origin may perform better than those of extrasynovial origin8 –11 and that resistance to gliding under the A2 pulley is greater From the Department of Orthopedics, Mayo Clinic and Mayo Foundation, Rochester, MN. Received for publication August 27, 1997; accepted in revised form March 17, 1999. No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. Reprint requests: Peter C. Amadio, MD, Mayo Clinic, 200 First St SW, Rochester, MN 55905. Copyright © 1999 by the American Society for Surgery of the Hand 0363-5023/99/24A05-0002$3.00/0
for the palmaris longus, an extrasynovial tendon, than for the flexor digitorum profundus, an intrasynovial tendon.12 A system has been developed that allows quantitative measurement of the excursion resistance between a tendon and its associated pulley.13,14 This technique has the advantage that the tendon–pulley interaction can be measured directly. Although friction between the tendon sheath and the tendon after repair or grafting may be an important factor in ultimate tendon excursion, the differences in excursion resistance between the entire tendon sheath and tendon, comparing different graft sources, have not been reported. The purpose of this study was to evaluate the difference in excursion resistance between intrasynovial and extrasynovial tendon graft sources against an intact sheath, including the A1–A4 pulleys and the phalanges and metacarpal, in an in vitro human model. The Journal of Hand Surgery 1097
1098 Nishida et al / Flexor Tendon–Tendon Sheath Interaction
Materials and Methods Seven fresh-frozen middle finger specimens were used. The specimens were obtained from 3 female and 3 male cadavers with ages ranging from 61 to 85 years (average, 74 years). Specimens were thawed immediately before testing. All specimens were kept moist in a saline bath throughout the testing procedure. The tendons used were the ipsilateral palmaris longus tendon, the ipsilateral extensor indicis proprius tendon, the ipsilateral flexor digitorum superficialis tendon of the ring finger, and the homodigital (that is, middle finger) flexor digitorum profundus tendon. Because the proximal part of the extensor indicis proprius passes beneath the extensor retinaculum and may therefore be considered intrasynovial, this tendon was tested twice, once for its distal extrasynovial portion and once for its proximal subretinacular portion. These 2 segments were tested simply by using both a distal–proximal and a proximal– distal orientation of the extensor indicis proprius. In the normal proximal– distal orientation, the subretinacular portion of the tendon remained proximal to the tendon sheath throughout testing. In the reversed distal–proximal orientation, the subretinacular portion of the tendon remained within the tendon sheath throughout the testing. Specimen preparation was identical for each hand. The palmaris longus and extensor indicis proprius tendons were harvested from their insertion to their musculotendinous junction. Any peritenon was left attached to the tendon. The flexor digitorum superficialis tendon of the ring finger was similarly harvested from its insertion to the musculotendinous junction. For the middle finger, the finger was disarticulated from the hand through the proximal portion of the metacarpal; care was taken to preserve the flexor tendons to the musculotendinous junction and the tendon sheath. A transverse incision through the synovial sheath was made just distal to the A4 pulley to mark the surface of the flexor digitorum profundus tendon with the finger in full extension. The flexor digitorum profundus tendon was then pulled proximally until full proximal interphalangeal and distal interphalangeal joint flexion was achieved. In this position, the tendon was again marked through the previous incision. The distance between these 2 marks represented the physiologic excursion range of the flexor digitorum profundus tendon for that finger. The flexor digitorum superficialis tendon was removed from the middle finger by pulling the tendon
proximally, cutting its insertion and short vinculum through a window in the C2 pulley, and cutting the long vinculum proximal to the A1 pulley. The incision in the C2 pulley was repaired with a 6 – 0 nylon suture. The tendon sheath, including A1–A4 pulleys, C1–C3 pulleys, the parietal membrane in this region, and the visceral membrane of the flexor digitorum profundus tendon, were all preserved. The finger was disarticulated at the distal interphalangeal joint, detaching the flexor digitorum profundus from the distal phalanx. A 1.5-mm K-wire was inserted through the proximal phalanx parallel to the longitudinal axis of the bone. This wire was used to stabilize the specimen in the mounting jig. The measurement system consisted of one custombuilt mechanical actuator with a linear potentiometer, 2 custom-made tensile load transducers accurate to 1 g,13 and a pulley13,14 (Fig. 1). The palmar side of the bone sheath preparation faced upward and the proximal end faced toward the actuator. Each tendon to be tested was prepared similarly. After the tendon was aligned in the proper proximal– distal orientation, it was passed through the tendon sheath. Load transducers were connected to the proximal and distal ends of the tendon with Dacron cord. The proximal load transducer (F2) was connected to the actuator. The distal transducer (F1) was connected to a
Figure 1. Experimental setup for the measurement of excursion resistance between the tendon for graft material and the adjacent tissues. Tensions of F1 and F2 are measured by the tensile load transducers. Excursion is measured by a linear potentiometer. a and b are the angles between the proximal and distal tendon ends, respectively, and the reference axis of the proximal phalanx. (Modified and reprinted with permission.12)
The Journal of Hand Surgery / Vol. 24A No. 5 September 1999 1099
250-g weight. The actuator was positioned at the preselected angle a, which was defined as the angle in degrees formed between the horizontal plane of the proximal phalanx and the proximal cable extension. This motion occurred primarily at the metacarpophalangeal joint. The mechanical pulley between the load and the distal load transducer was positioned at the preselected angle b, which was defined as the angle formed between the horizontal plane of the proximal phalanx and the distal cable extension. This motion occurred primarily at the proximal interphalangeal joint. The sum of angles a and b was considered the angle of the arc of contact. The tendon to be tested was pulled proximally at a rate of 2.0 mm/s by the actuator and was opposed by the weight. The movement of the tendon toward the actuator was regarded as flexion. F1, F2, and the corresponding excursion were recorded by computer at a sampling rate of 10 Hz. Excursion was limited to the distance between the 2 tendon markers on the profundus tendon of that finger. Angles a and b were varied to 5 different positions by raising or lowering the actuator and weight. Three trials were performed for each of the 5 positions (a, b): 15, 5; 20, 10; 30, 10; 30, 20; and 30, 30. These positions, and the testing rate, were selected because they were comparable to those used in previous studies.12,14,15 Each graft was tested once, except for the extensor indicis proprius tendon, which was tested once in each of 2 orientations. The order of testing in each finger was palmaris longus, reversed extensor indicis proprius, normally oriented extensor indicis proprius, flexor digitorum superficialis, and flexor digitorum profundus. Plots of F1 and F2 measurements versus excursion were examined for each trial and evaluated by their shape. Because the trials were generally identical and the first run was considered to be preconditioning, the average of the last 2 runs was selected for analysis for each angle. The mean force differences of F2 and F1 for the whole excursion were obtained and regarded as the resistance at the interface between the tendon and the pulley for the given arc of contact. The tendons were classified as follows: intrasynovial sources (flexor digitorum superficialis, flexor digitorum profundus, and reversed extensor indicis proprius) and extrasynovial sources (palmaris longus and normally oriented extensor indicis proprius). Significant differences in excursion resistance between tendons and at different angles were assessed with a 2-factor repeated-measures ANOVA. Tukey’s honestly significant difference test was then used for
a post hoc comparison of individual means in the mean effects.16 The level of significance was set at a 5 .05. To compare intrinsic versus extrinsic tendons, a 1-factor repeated-measures ANOVA with an orthogonal polynomial breakdown for means across angles of measurement was used.16 A significance level of a 5 .01 was used to correct for the error rate of a second ANOVA test.
Results The results are shown in Figure 2. ANOVA showed no difference between results when characterized by finger or by angle. There was a significant difference when the results were characterized by tendon (p , .01). Specifically, the normally oriented extensor indicis proprius was significantly (p , .05) different from the flexor digitorum profundus, the flexor digitorum superficialis, and the reversed extensor indicis proprius tendons. The intrasynovial tendons as a group were significantly different from the extrasynovial tendons (p , .01). Univariate and multivariate repeated-measures analysis within subjects showed that the intrasynovial and extrasynovial groups were affected differently by joint angle; specifically, as contact angle increased, the resistance of intrasynovial tendons tended to decrease, whereas for extrasynovial tendons the resistance tended to increase (p , .001). The excursion resistance of the extrasynovial tendons was significantly higher than that of the intrasynovial tendons (p , .01). Among extrasynovial tendons, the excursion resistance was higher in the normally oriented extensor indicis proprius tendon than in the palmaris longus tendon, and these differences were statistically significant at all angles (p , .05). There were no differences in resistance between the 3 intrasynovial tendons. The mean length of the subretinacular portion of the extensor indicis proprius was 25 mm compared with a mean A1–A4 length of 71 mm (range, 61–79 mm).
Discussion The concept of friction measurement and its application to the tendon–pulley unit has been the focus of previous reports.12–15 We extend the investigation to include other components of the tendon sheath, beyond the basic single tendon–single pulley interaction. A tendon sliding through a curved pulley is analogous to a belt wrapped around a fixed mechanical pulley.13,14 If the tensions in the belt are F1 and F2
1100 Nishida et al / Flexor Tendon–Tendon Sheath Interaction
Figure 2. Comparison between intrasynovial and extrasynovial tendons: flexor digitorum superficialis tendon (FDS), palmaris longus tendon (PL), normally oriented extensor indicis proprius tendon (EIP), reversed extensor indicis proprius tendon (EIP-R), and flexor digitorum profundus (FDP). Error bars represent standard error of the mean.
on each end of the tendon, respectively, and if the impending motion of the tendon is from F1 to F2, then F2 is greater than F1 because of the friction (f), and f 5 F2 2 F1. In previous studies,12–14 bone was carefully excised so that a pure tendon–A2 interaction could be evaluated. In this study, in contrast, the bones and joints were not disturbed, so F2 and f also potentially include the friction between the tendon and other surrounding tissues, including bone. Our hypothesis was that intrasynovial tendon would have less resistance to gliding under an intact tendon sheath than would extrasynovial tendon. This is, in fact, what we observed, and we believe that our data thus far support the proposed hypothesis. We also observed some differences in excursion resistance when comparing A1–A4 with previously reported work on resistance versus A2 alone.12 These differences are discussed below. The final goal of tendon grafting is to achieve maximal excursion of the tendon with as little resistance as possible. The kinematics of tendon gliding are complex. The surface of the flexor digitorum profundus tendon contacts its guiding pulleys as well as the flexor digitorum superficialis tendon and the phalangeal bases.17 Each type of interaction (tendon– pulley, tendon–tendon, tendon– bone) may have its own characteristic friction relationship. Most investigators have discussed tendon graft sources from the viewpoint of only utility or convenience.3,18 –20 An important body of experimental
evidence suggests, however, that the source of the tendon graft may be an important factor in determining the clinical result of the grafting procedure. Intrasynovial tendons heal with less cellular necrosis and with less extensive adhesions than extrasynovial tendons.9 –11 Intrasynovial tendon grafts survive better than extrasynovial tendon grafts.9,11 In addition, intrasynovial grafts integrate with less scar formation between the tendon surface and the surrounding tissues.9,11 Intrasynovial tendon grafts show significant differences from extrasynovial tendon grafts in both surface morphology and vascularity.10,11 The synthesis of proteoglycan matrix proteins and DNA is different between these 2 types of tendon surfaces.8 Surface lubrication mechanisms are also different between these 2 types of tendons,15 and excursion resistance of extrasynovial tendons exceeds that of intrasynovial tendons in a tendon–A2 pulley model.12 We believe that our data add to this body of experimental evidence, supporting the advantages of intrasynovial tendons as sources of tendon grafts. Furthermore, the measured excursion resistance between the flexor digitorum profundus and the A1–A4 pulleys at the highest angle (a, b 5 30°, 30°, or 60°) was not greater than that recorded in a previous study that measured resistance between the flexor digitorum profundus and the A2 pulley alone.12 Because friction is not dependent on contact area, this finding is consistent with the hypothesis that we are truly measuring tendon–pulley friction;
The Journal of Hand Surgery / Vol. 24A No. 5 September 1999 1101
that is, even though the tendon is contacting a much longer pulley in A1–A4 than A2, no increase in resistance was observed. In contrast with previous studies12,21 that evaluated friction between the tendon and A2 pulley, in this study of tendon contact with a more complete sheath we found a higher resistance for the tendons at lower angles of a and b but not at higher angles. We believe that this can be explained by tendon contact with other components of the tendon sheath complex, such as bone.17 The bony element was specifically excluded in previous studies by excising the bony condyles.12–14 Direct observation by Lin et al.17 showed that loss of contact of tendon with underlying bone occurred at approximately 30° of flexion at the proximal interphalangeal joint in normal cadaver fingers. At a lower angle, the excursion resistance may therefore include friction not only between the tendon and pulley but also between the tendon and bone. Although we did not directly measure the friction between the tendon and bone, the fact that excursion resistance was greater at lower values of a and b, when it would be reasonable to assume that tendon made contact with bone, suggests that the coefficient of friction may well be greater for the tendon– bone interaction than it is for the tendon– pulley interaction. Supporting this hypothesis, at higher values of a and b, where tendon– bone contact no longer occurs, this difference disappears. This effect, again in comparison with previous studies of resistance against A2 alone, is greater for the flexor digitorum superficialis and flexor digitorum profundus, that is, the steadily increasing resistance at higher angles observed for all tendons against the A2 pulley alone12,21 and for the extrasynovial tendons versus the entire sheath in this study is not observed for the flexor digitorum profundus and superficialis versus the entire sheath. We believe this is due to the effect of tendon cross-section, which is greater for the flexor digitorum profundus and flexor digitorum superficialis tendons than it is for the extensor indicis proprius or palmaris longus tendons. This greater cross-section does not, of course, increase friction per se because frictional force is not affected by the surface area of contact. Increased cross-section, however, does allow these larger tendons to contact other surfaces, such as bone, which may have a different coefficient of friction than tendon sheath. We believe that this contact with other structures within the sheath system explains the difference we observed, and we plan to investigate this with future studies of friction of tendon versus the various components of
an intact sheath system, for example, tendon/tendon and tendon/bone. There was not a statistically significant difference between the excursion resistance of the palmaris longus tendon and the intrasynovial tendons. We believe that this relates more to sample size than to anything else because in previous studies12,21 we have been able to demonstrate differences in excursion resistance between the palmaris longus and the intrasynovial tendons. In this study there was certainly a trend in that direction, even though it did not meet the test of statistical significance. We did not think it worthwhile to test more specimens simply to get the palmaris longus to a level of statistical significance because we have demonstrated that previously. We do not believe that the order of testing, which was the same in each specimen in this study, can explain the differences we observed. We have demonstrated previously that the order of testing does not affect the results in our system, even when harsh enzymatic treatments are used15 or when the tendon is subjected to as many as 100 cycles of testing (unpublished data). This study does have several limitations. We did not duplicate normal finger kinematics, and the gliding of the tendons was not physiologic because the proximal phalanx was fixed and the distal interphalangeal joint was disarticulated. A method of measurement of excursion resistance that permits normal joint movement should be developed. The resistance to gliding of the tendons in this study was measured in a sheath without a flexor digitorum superficialis tendon. The effect of inclusion of the flexor digitorum superficialis tendon on gliding resistance should be evaluated in the future. Finally, the results of this in vitro study need to be confirmed in vivo. We acknowledge that intrasynovial donor tendons are easier to find in an experimental model than in a clinical situation. We believe, however, that occasions do arise in which an otherwise unsalvageable digit may serve as a tendon donor source, and we are optimistic that in the future it may be possible to consider other sources, such as allografts or even heterografts, for tendon reconstruction. Furthermore, in the future it may be possible to genetically alter the gliding characteristics of extrasynovial tendons.22 We believe that the strength of this study lies in our ability to directly measure the complex interaction of a tendon sheath that extends both proximal and distal to the previously studied A2, which also affords an opportunity to study tendon interactions in
1102 Nishida et al / Flexor Tendon–Tendon Sheath Interaction
a more complex environment. This in vitro model supports the hypothesis that intrasynovial tendons have advantages over extrasynovial tendons for tendon grafting. In addition to the flexor digitorum superficialis and flexor digitorum profundus tendons, the reversed extensor indicis proprius tendon also has excursion resistance properties consistent with an intrasynovial source. An in vivo animal model to test the properties of the grafted tendon with healing over time will be necessary in the future because in vitro models cannot address the complex effects of wound healing on tendon gliding. The authors thank David A. Gabriel, PhD, for assistance with statistical methods and analysis.
References 1. Amadio PC, Wood MB, Cooney WP III, Bogard SD. Staged flexor tendon reconstruction in the fingers and hand. J Hand Surg 1988;13A:559 –562. 2. Boyes JH, Stark HH. Flexor-tendon grafts in the fingers and thumb. A study of factors influencing results in 1000 cases. J Bone Joint Surg 1971;53A:1332–1342. 3. Hunter JM, Salisbury RE. Flexor-tendon reconstruction in severely damaged hands. A two-stage procedure using a silicone-Dacron reinforced gliding prosthesis prior to tendon grafting. J Bone Joint Surg 1971;53A:829 – 858. 4. McClinton MA, Curtis RM, Wilgis EF. One hundred tendon grafts for isolated flexor digitorum profundus injuries. J Hand Surg 1982;7A:224 –229. 5. Silfverskio¨ld KL, May EJ, To¨rnvall AH. Tendon excursions after flexor tendon repair in zone II: results with a new controlled-motion program. J Hand Surg 1993;18A:403–410. 6. Small JO, Brennen MD, Colville J. Early active mobilisation following flexor tendon repair in zone 2. J Hand Surg 1989;14B:383–391. 7. Strickland JW. Flexor tendon injuries. Part 4. Staged flexor tendon reconstruction and restoration of the flexor pulley. Orthop Rev 1987;16:78 –90. 8. Abrahamsson SO, Gelberman RH, Lohmander SL. Variations in cellular proliferation and matrix synthesis in intrasynovial and extrasynovial tendons: an in vitro study in dogs. J Hand Surg 1994;19A:259 –265.
9. Gelberman RH, Seiler JG III, Rosenberg AE, Heyman P, Amiel D. Intercalary flexor tendon grafts. A morphological study of intrasynovial and extrasynovial donor tendons. Scand J Plast Reconstr Surg Hand Surg 1992;26: 257–264. 10. Gelberman RH, Chu CR, Williams CS, Seiler JG III, Amiel D. Angiogenesis in healing autogenous flexor-tendon grafts. J Bone Joint Surg 1992;74A:1207–1216. 11. Seiler JG III, Gelberman RH, Williams CS, et al. Autogenous flexor-tendon grafts. A biomechanical and morphological study in dogs. J Bone Joint Surg 1993;75A:1004 – 1014. 12. Uchiyama S, Amadio PC, Coert JH, Berglund LJ, An K-N. Gliding resistance of extrasynovial and intrasynovial tendons through the A2 pulley. J Bone Joint Surg 1997;79A: 219 –224. 13. An K-N, Berglund L, Uchiyama S, Coert JH. Measurement of friction between pulley and flexor tendon. Biomed Sci Instrum 1993;29:1–7. 14. Uchiyama S, Coert JH, Berglund L, Amadio PC, An K-N. Method for the measurement of friction between tendon and pulley. J Orthop Res 1995;13:83– 89. 15. Uchiyama S, Amadio PC, Ishikawa J-I, An K-N. Boundary lubrication between the tendon and the pulley in the finger. J Bone Joint Surg 1997;79A:213–218. 16. Kirk RE. Experimental design: procedures for the behavioral sciences. Belmont, CA: Brooks/Cole Publishing Company, 1968. 17. Lin GT, Amadio PC, An K-N, Cooney WP. Functional anatomy of the human digital flexor pulley system. J Hand Surg 1989;14A:949 –956. 18. Harvey FJ, Chu G, Harvey PM. Surgical availability of the plantaris tendon. J Hand Surg 1983;8A:243–247. 19. Wehbe´ MA. Tendon graft donor sites. J Hand Surg 1992; 17A:1130 –1132. 20. White WL. Tendon grafts: a consideration of their source, procurement and suitability. Surg Clin North Am 1960;40: 403– 413. 21. Nishida J, Amadio PC, Bettinger PC, An K-N. Excursion properties of tendon graft sources: interaction between tendon and A2 pulley. J Hand Surg 1998;23A:274 –278. 22. Lou J, Manske PR, Aoki M, Joyce ME. Adenovirus-mediated gene transfer into tendon and tendon sheath. J Orthop Res 1996;14:513–517.