Biomechanical analysis of internal fixation techniques for proximal interphalangeal joint arthrodesis

Biomechanical analysis of internal fixation techniques for proximal interphalangeal joint arthrodesis Although numerous fixation techniques have been developed for performing small joint arthrodesis, no previous study of the biomechanical properties of these constructs has been published. The strength of specimens of arthrodesis of the proximal interphalangeal joint performed with cadaver material was determined for three-point anteroposterior bending, axial torsion, and lateral cantilever bending stress. Crossed Kirschner wires, intraosseous wiring, and figure of eight tension-band wiring were studied. Figure of eight tension-band wiring demonstrates superior strength in anteroposterior bending and torsion. For lateral bending stress, no significant difference exists among the techniques studied. (J HAND SURG llA:562-6, 1986.)

John C. Kovach, M.D., Fredrick W. Werner, M. Mech. Eng., Andrew K. Palmer, M.D., Seth Greenkey, M.D., and Dennis J. Murphy, B. S., Syracuse, N. Y.

In a review of 71 consecutive finger joint arthrodeses perfonned by one attending surgeon at the State University of New York, Upstate Medical Center, an overall nonunion rate of 7% was found. Fixation techniques used in this study included single longitudinal Kirschner wire, single oblique Kirschner wire, crossed Kirschner wires, external fixation, and intraosseous wiring as described by Lister. All fusion techniques used flat-end bone cuts. In examining nonunion rates, the intraosseous wiring technique was found to have the lowest rate at one in 32, or 3.1 %. This superior nonunion rate was hypothesized to be due to superior fixation stability. To study this concept, a biomechanical analysis of internal fixation techniques for small joint arthrodeses was undertaken. A profusion of internal and external fixation techniques have been described for arthrodesis of the proximal interphalangeal joint (PIP). 1-9 Although numerous biomechanical analyses exist for fixation techniques for small bone diaphyseal osteotomy and fracture fixation,5. 10-13 no previous study has been conducted for small joint arthrodesis fixation. Some conflict exists in From the Department of Orthopedic Surgery, State University of New York, Upstate Medical Center, Syracuse, N.Y. Received for publication June 6, 1985; accepted in revised form Sept. 18, 1985. Reprint requests: John C. Kovach, M.D., Bryden Canyon Center. 320 Warner Dr.. Lewiston. ID 83501.

562

THE JOURNAL OF HAND SURGERY

Dorsal TYPE I

TYPE

n

Fig. 1. Fixation techniques I and 2. Technique I consists of an oblique Kirschner wire and a tension axis intraosseous loop placed through holes labeled type I. Technique 2 consists of an oblique Kirschner wire and a neutral axis intraosseous loop placed through the holes labeled type II.

the data regarding stability of various fixation techniques in diaphyseal osteotomies. In addition, because of different bone composition and geometric characteristics at the epiphysis and metaphysis compared with the diaphysis, data from previous biomechanical studies may not apply to small joint arthrodesis fixation.

Material and methods An in vitro study was conducted to analyze the biomechanical properties of four fixation techniques. Embalmed cadaver hands were used with specimens selected to exclude those with significant osteopenia. Flatended resection of the articular surfaces of the PIP joint was used, since this was the method used in our clinical

Vol. llA, No.4 July 1986

Techniques for PIP arthrodesis

563

Fig. 2. Technique 3 (left) consists of one oblique and one longitudinal Kirschner wire. Technique 4 (right) consists of paired longitudinal Kirschner wires and a dorsal figure of eight wire loop.

Fig. 4. A specimen is positioned in the torsion stress jig for testing.

Fig. 3. A specimen is positioned in the three-point AP-bending stress jig for testing.

series. Joints were fused in 30° of palmar flexion after all soft tissue was removed except the palmar plate. The fixation techniques evaluated in this initial study were those techniques commonly used at our medical center. The four techniques evaluated were: (1) one oblique 0.045-inch Kirschner wire with a dorsal coronal (tension axis) plane, 26-gauge stainless steel wire loop, (2) a variation of the first technique in which the wire loop was located in the midcoronal plane (neutral axis) (Fig. 1), (3) one oblique and one longitudinal neutral axis 0.045-inch Kirschner wire with the intersection of the Kirschner wires occurring in the middle phalanx as seen on an anteroposterior (AP) view, and (4) two longitudinal 0.045-inch Kirschner wires protruding dorsally from the proximal phalanx with a dorsal figure of

Fig. 5. A specimen is positioned in the lateral cantileverbending stress jig for testing.

eight, 26-gauge wire loop (Fig. 2). This is commonly referred to as tension-band fixation. I. 14 In all techniques with stainless steel wire loops, the ends were twisted and not tied. All procedures were performed by one person. Specimens were kept moist and refrigerated until testing.

564

The Journal of HAND SURGERY

Kovach et al.

Table I. Mechanical testing results for three-point bending of PIP fusions (eight fingers in each technique) Maximum bending moment (N - m)

Technique

0.749 0.352 0.696 1.22

2 3 4

± ± ± ±

0.393 0.144 0.183 0.43

Table ll. Mechanical testing results for torsional tesing of PIP fusions (six fingers in each technique) Maximum torque

Torsional energy

(N - m)

(N - m)

Technique

1 2 3 4

0.330 0.469 0.226 0.536

± ± ± ±

0.133 0.102 0.051 0.188

0.131 0.177 0.092 0.217

± ± ± ±

0.060 0.046 0.023 0.074

Mechanical testing The instrumented PIP joints were tested to failure in three-point AP bending, torsion, and lateral cantilever bending. Six to eight fingers were used for each mechanical test of a given fixation technique. All tests were conducted on an Instron Testing Machine (Canton, Mass.) that displayed force as a function of displacement or torque as a function of angular rotation on a chart recorder. Specially designed bone-holding jigs were used for this study (Figs. 3 to 5). In the three-point bending test, the proximal and middle phalanges were supported with the center of the joint displaced downward 10.2 mm at a rate of 5.1 mml min, thus increasing finger flexion. Torsional testing of each finger was designed to twist the middle phalanx internally relative to the proximal phalanx as would occur in thumb-finger key pinch. Each specimen was rotated in the absence of axial load to 40° at a rate of 19.2°/min. Lateral cantilever bending of the fingers was conducted by applying a force to the middle phalanx via a V-shaped bar that was located 13 mm from the joint line. The middle phalanx was thus displaced ulnarly relative to the proximal phalanx until the bar came down 10.2 mm. The range of deformation selected for each type of loading was selected on the basis that in a clinical situation, a single loading episode of the degree of deflection achieved would clearly be interpreted as clinical failure of the fixation. Each specimen was examined at the conclusion of

Bending rigidity (N - mlm)

242 109 194 340

± ± ± ±

121 57 85 78

Energy (N - mm) - mm

3390 1550 3390 6080

± ± ± ±

1550 770 1030 1720

the tests for gross failure of bones, bending of the pins, stretching of the wires, or migration of the pins or wire through the bones. The following data were extracted or calculated from chart recordings for the three-point AP bending tests: maximum bending moment, three-point bending rigidity, and bending energy until the location of maximum bending moment. Bending rigidity is the steepest tangential slope of the bending moment-displacement curve. It is analogous to the stiffness on a force-displacement curve. For torsional loading, maximum torque and torsional energy up to the location of the maximum torque were determined. For lateral bending, maximum bending moment, cantilever bending rigidity, and bending energy up to the location of maximum bending moment were determined. For each specimen, the cross-sectional area of bone apposition at the osteosynthesis site was measured for normalization of the biomechanical data. All data were statistically compared with an analysis of variance technique. Results The averaged results and the standard deviations for three-point AP bending testing are shown in Table I. Technique 4 had the statistically largest maximum bending moment, bending rigidity, and energy (p < 0.01), while technique 2 had the smallest moment, rigidity, and energy (p < 0.01). Techniques 1 and 3 had mechanical properties similar to each other. In torsion testing, technique 4 had the largest energy of the four techniques (p < 0.01). Technique 4 also had a larger maximum torque than did techniques 1 and 3, but had torque similar to technique 2 (p < 0.01). Technique 3 had the smallest maximum torque and energy of the four techniques (Table II). No statistical differences were found between techniques in the maximum bending moment, bending rigidity, or bending energy for the lateral cantilever bending tests (Table III).

Vol. IIA, No.4 July 1986

Techniques for PIP arthrodesis

565

Table III. Mechanical testing results for lateral bending of PIP fusions (six fingers in each technique) Technique

2 3 4

Maximum bending moment (N - m)

0.506 0.715 0.668 0.706

± ± ± ±

0.133 0.254 0.143 0.234

In examining gross mechanisms of failure for techniques 1 and 2, the intraosseous wiring techniques, the holes for the wire loops elongated in both the AP and lateral bending modes. This was particularly true for the radial holes on the lateral cantilever-bending specimens . With AP bending and torsional stresses, the wire loops were also noted to stretch slightly. When subjected to AP bending or torsional stresses, technique 3, one longitudinal and one oblique Kirschner wire, failed by rotation about the oblique pin with the pins pulling out longitudinally. In AP bending the longitudinal pin also bent permanently. In lateral cantilever bending, the pins both pulled out and showed some permanent bending. Technique 4, figure of eight tension-band technique, failed in AP and lateral cantilever bending by minimal permanent bending of the pins, with minimal stretching of the wire loops occurring with AP bending stress only. In torsion minimal permanent bending of the Kirschner wires occurred. With technique 4, substantially less permanent displacement of the bone ends existed after release of the deforming forces compared with the other three techniques. Stretching of wire loops was always accompanied by some unwinding of the twisted ends . In general, bone failure, as evidence of the wire loops cutting through bone, was manifest more commonly than stretching or unwinding of the wire loops. Kirschner wires tended to fail by bending and pulling out of the bone rather than failure of the bone about the pin. Discussion We selected single-cycle loading of the bone/instrumentation construct to a predetermined stopping point, since this mode of stress application can be easily correlated to clinical failure. In some cases the location of maximum moment or torque did not occur before the stopping point; in these cases maximum moment or torque was measured at the stopping point. Slow rates of load application were used so that mechanisms of failure could be visually observed during loading. Loads of the magnitude applied in this study are greater than those encountered in most daily activities. How-

Bending rigidity (N -mlm)

258 344 237 375

± ± ± ±

196 326 123 202

Bending energy (N - mm) - mm

2840 2620 3750 3700

± ± ± ±

1280 1130 1320 2090

ever, angular moments and torques of such large magnitude are encountered in situations such as vigorous activities, trauma, and passive manipulation. Internal fixation should therefore be capable of withstanding these high energy loads. Six or eight specimens were subject to each stress for each of the four fixation techniques. Normally we instrumented the PIP joints of the index, long, ring, and small fingers unless the small finger was deleted . The cross-sectional area of the osteosynthesis surfaces was measured, and the data were normalized to account for the surface area variations seen. This normalization did not significantly change the relative magnitude or ranking of one technique to another, and therefore these results are not reported since the prenormalization data are sufficient. Many more fusion fixation techniques remain to be tested. The techniques tested were selected on the basis of current use in our medical center. In addition to flatended bone resection fusion techniques, other joint resection configurations, in particular ball and cup and chevron, should be tested since surface geometry may significantly influence fixation stability. The biomechanical testing data from this study indicate that of the four fixation techniques evaluated, technique 4 was superior except in lateral bending in which all techniques were statistically equivalent. In examining specimens for gross mechanisms of failure, technique 4 showed significantly less permanent deformation after the release of the applied stress than did the other three techniques. In examining the mode of failure of wire loops, failure occurred chiefly, in order of significance, by cutting through the bone, stretching of the wire, and unwinding of the twisted wire ends. The first two problems could be addressed by the use of additional wire loops and larger diameter wire, respectively. However, these remedies carry the obvious disadvantage of increasing the burden of foreign-body mass or additional dissection and operating time. Unwinding of the wire ends could be avoided by tying rather than twisting the ends. II The stainless steel pins were observed to fail by pull-

566

Kovach et ai.

ing out of the bone and to a lesser degree by bending. These problems could be addressed by altering the surface of the pin to increase the coefficient of friction, bending the pins at the bone surface,1O or using larger diameter pins.13 Altering the surface of the pin could be accompanied by problems in the ultimate removal of the fixation material, and increasing the diameter of the pins has obvious disadvantages in performing a small joint arthrodesis in which the quantity of available bone stock is limited. Moberg? stated that the goal in arthrodesis is to produce a solid arthrodesis in the desired position in the shortest duration. We would add that it is desirable to avoid the use of additional external splinting to protect the internal fixation. Because secure internal fixation facilitates the achievement of these goals, dorsal figure of eight tension-band wiring with two longitudinal Kirschner wires is superior to the other three techniques evaluated in this study. Because of the limited number of techniques evaluated, this study should be viewed as preliminary, especially with respect to clinical relevance. Clinical factors such as degree of soft tissue dissection, subcutaneous implant irritation, and difficulty of implant removal are not addressed in this study. The biomechanical data collected in this study serve as a baseline on which to expand with future studies. A controlled clinical study comparing these fixation techniques is needed to establish rates of healing and incidence of failure. Because failure of fixation is associated with the development of pseudarthroses, stable fixation is desirable. However, the optimum degree of rigidity for promoting osseous union is unknown. Varying the rates of load application to correspond more closely with the clinical state would yield additional useful information. Cyclic loading within the elastic range of deformation is more commonly encountered than are the catastrophic loads used in this study. Failure of internal fixation in small bone and joint fractures and arthrodesis applications as a response to cyclic loading has not been studied to date.

Conclusion Figure of eight tension-band internal fixation for arthrodesis of the PIP joint is superior biomechanically

The Journal of HAND SURGERY

when subject to three-point AP bending and torsional stress loading compared with the other three fixation techniques evaluated in this study. In cantilever lateral bending, all four fixation techniques evaluated were similar biomechanically. In examining specimens for mechanisms of gross failure, the figure of eight tensionband technique demonstrated significantly less permanent deformation than did the other three techniques evaluated. REFERENCES I. Allende BT, Engelem JC: Tension-band arthrodesis in the finger joints. J HAND SURG 5:269-71, 1980 2. Carroll RE, Hill NA: Small joint arthrosis in hand reconstruction. J Bone Joint Surg [Am] 51:1219-21, 1969 3. Harrison SH, Nicolle FV: A new intramedullary bone peg for digital arthrodesis. Br J Plast Surg 27:240-1, 1974 4. Hogh J, Jensen PQ: Compression-arthrodeses of finger joints using Kirschner wires and cerclage. Hand 14: 14952, 1982 5. Fyfe IS, Mason S: The mechanical stability of internal fixation of fractured phalanges. Hand 11:50-4, 1979 6. Lister G: Intraosseous wiring of the digital skeleton. J HAND SURG 3:427-34, 1978 7. Moberg E: Arthrodesis of finger joints. Surg Clin North Am 40:465-70, 1960 8. Watson HK, Schaffer SR: Concave-convex arthrodeses in joints of the hand. Plast Reconstr Surg 46:368-71, 1970 9. Wexler MR, Rousso M, Weinberg H: Arthrodeses of finger joints by dynamic external compression. Plastic Reconstr Surg 59:882-5, 1977 10. Massengill JB, Alexander H, Parsen JR, Schecler MJ: Mechanical analyses of Kirschner wire fixation in a phalangeal model. J HAND SURG 4:351-6, 1979 II. Massengill JB, Alexander H, Lagrana N, Mylod A: A phalangeal fracture model-quantitative analysis of rigidity and failure. J HAND SURG 7:264-70, 1982 12. Rayhack 1M, Belsole RI, Skelton WH: A strain recording model: analyses of transverse osteotomy fixation in small bones. J HAND SURG 9:383-7, 1984 13. Yanik RK, Weber RC, Matloub HS, Sanger JR, Gingrass RP: The comparative strengths of internal fixation techniques. J HAND SURG 9:216-21, 1984 14. Stem PI, Drury WI: Tension band arthrodesis of small joints. Contemp Orthop 8:59-62, 1984