- Email: [email protected]
Distal interphalangeal joint arthrodesis comparing tensionband wire and herbert screw: A biomechanical and dimensional analysis
Distal Interphalangeal Joint Arthrodesis Comparing Tensionband Wire and Herbert Screw: A Biomechanical and Dimensional Analysis Brad Wyrsch, MD, Cincinnati, OH, John Dawson, PhD, Saint Aufranc, BS, DouglasWeikert, MD, Michael Milek, MD, Nashville, TN Thirty cadaveric distal interphalangeal joints (15 male and 15 female joints) were prepared with either a Herbert screw or a tension-band wire technique to simulate an arthrodesis. To elucidate mechanical differences between these constructs, the strength of the specimens was determined for three-point anteroposterior and lateral bending and for axial torsion. The Herbert screw demonstrated significantly greater anteroposterior bending strength and greater torsional rigidity when compared to the tension-band wire technique. For dimensional analysis, the height and width of each distal phalanx was measured prior to fixation, 4 mm from the distal tip of the bone (the region that must accommodate the large-diameter threads of the Herbert screw). Results indicated that the mean height of the distal phalanx (3.55 mm) is smaller than the diameter of the screw (3.90 mm). Fracture or thread penetration at the tip of the distal phalanx during screw placement occurred in 25of the specimens overall and in all the female phalanges, often resulting in stretching or violation of the nail bed. Despite fracture or screw penetration, the Herbert screw appears to offer additional strength that may be clinically important for joint arthrodesis. (J Hand Surg 1996;21A:438-443.)
Distal interphalangeal (DIP) joint arthrodesis is indicated for relief of painful conditions such as osteoarthritis and post-traumatic arthritis and for stabilization of the joint following severe bony trauma and irreparable tendon injuries. Several techniques of fixation have been described for DIP arthrodeses, From the Department of Orthopaedics and Rehabilitation, Vanderbilt University Medical Center, Nashville, TN. Received for publication Jan. 30, 1995; accepted in revised form Aug. 8, 1995. Although the authol's have not received or will not receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article, benefits have been or will be received but are directed solely to a research fund, foundation, educational institution, or other nonprofit organization with which one or more of the authors are associated. Reprint requests: Brad Wyrsch, MD, Cincinnati Hand Center, 2800 Winslow Avenue, Cincinnati, OH 45206.
438
The Journal of Hand Surgery
including Kirschner wires (K-wires), bone pegs, interosseous or tension-band wires (TBW), and compression screws. 1-9 More recently, Herbert compression screws have been employed for interphalangeal joint arthrodeses, m Proximal interphalangeal (PIP) joint arthrodesis with the Herbert screw has achieved excellent clinical results, u In addition, equivalent biomechanical strength has been demonstrated for the Herbert screw and TBW techniques in PIP joints.ll However, despite wide clinical use of these techniques, there is a paucity of literature concerning the use of Herbert screws in the DIP joint. With limited experience at our institution, Herbert screw DIP arthrodesis has achieved a 100% union rate with no complications. This was a two-part study. In the first part, the height and width of distal phalangeal specimens
The Journal of Hand Surgery/Vol. 21A No. 3 May 1996 439
were measured to see how the distal phalanx of the adult finger accommodates the large threads of the Herbert screw. In the second part, biomechanical testing was performed, comparing simulated DIP fusion constructs of TBW and the Herbert screw.
Materials and Methods Measurement of the Distal Phalanx
The index, middle, and ring fingers of five male and five female embalmed cadaveric hands were stripped of soft tissue, leaving only the palmar plates. The terminal slip of the extensor tendon was excised, as were the collateral ligaments. The transverse dimensions of distal phalanges were measured 4 mm proximal to the distal tips of the bones, using a caliper. This area of the distal phalanx is the region that must accommodate the large-diameter (lagging) threads of the Herbert screw. Measurements were taken in two mutually perpendicular planes: the dorsal-palmar dimension (the height) and the mediolateral dimension (the width). Fixation Techniques The DIP joint cartilage was excised in all specimens, using an osteotome or rongeur, to create a cup-and-cone configuration of cancellous bone. In 15 of the fingers, the DIP joints were fixed using Herbert screws (24-28 mm in length) in the following manner: a small K-wire was used to produce a guide hole, proceeding from proximal to distal through the center of the medullary canal of the distal phalanx; the 2-ram-diameter Herbert drill was passed manually through the guide hole from distal to proximal through both phalanges, while care was taken to center the drill in both bones; the large Herbert drill was then used to enlarge the screw hole in the distal phalanx; and finally, a Herbert screw was placed retrograde until the large-diameter (lagging) threads of the screw completely engaged in the distal phalanx and compression was achieved. The authors who prepared these specimens (B. W. and S. A.) believed that excellent placement and fixation of the screw was achieved in all specimens. In the remaining 15 fingers, the DIP joints were fixed using K-wires and a dorsal TBW in the following manner: two medial-to-lateral transverse holes were created in both the middle and distal phalanges (the holes were dorsal to the midline of the bones) using a 0.35-mm K-wire. A 26-gauge wire was passed through these holes and crossed dorsally, then twisted counterclockwise, to create a figure-of-eight
dorsal tension band with the DIP joint in flexion (0~176 next, two 0.35-ram K-wires were placed in a crossed fashion palmar to the TBWs. The authors who prepared these specimens (B. W. and S. A.) believed that excellent fixation was achieved in all specimens. Hardware placement was then evaluated by direct inspection and x-ray films. For the Herbert screw, those instances when the screw threads penetrated bone cortices were noted. No specimens were discarded. Mechanical Testing
All specimens were mechanically tested; tests included anteroposterior (AP) bending, lateral bending, and axial torsion (five specimens for each test of both constructs). Repeated use of a specimen was not possible because each test was carried out to failure of the joint. Failure was defined as a displaced fracture of the specimen or cutout of a wire or screw through cortical bone. Each DIP fusion construct was potted into a specially fabricated test fixture. This consisted of two stainless-steel square blocks with drill holes for screw fixation to mount onto an MTS (Material Testing Systems, 3M Corps., Minneapolis, MN). A specimen could be positioned in this fixture in one of three orientations, corresponding to the type of test to be conducted. The proximal aspect of the middle phalanx and the distal tip of the distal phalanx were secured with four small locking screws at the base of each wall of the recessed area of the block. In addition, the recessed area of each block was filled with resin (Technovit 4000, Kulzer Inc., Germany), which was allowed to solidify. For the AP bending tests and the lateral bending tests, the distances from the proximal pot to the DIP joint (a) and from the distal pot to the DIP joint (b) and the total specimen length (L) were measured. These distances (moments) are needed to compute the bending stiffness (EI) of the DIP joint construct, using the following equation: EI = Pa3b3/3dL3 At the time of failure of the specimen, the load (P) and displacement (d) were used to calculate the bending stiffness of the specimen. The bending tests were conducted under displacement control at 5 mm/min, to a maximum displacement of 10 ram. For the AP bending test, specimens were placed horizontally, with the dorsal aspect of the joint facing up. Thus, the test apparatus created a bending moment to
440
Wyrsch et al./Tension-band Wire vs Herbert Screw
the dorsal surface of the DIP joint. For the lateral bending test, specimens were placed horizontally with the radial aspect of the DIP joint facing up. For this test, the apparatus created a bending moment to the lateral (radial) surface of the DIP joint. All specimens failed prior to the maximum displacement. The applied axial rotation and the resulting torque were measured throughout a torsion test. The torsion tests were conducted under angle control at 10~ to a maximum angle of 30 ~ The maximum torque and its associated angle o f twist were used to calculate the torsional rigidity of the specimen. The torsional rigidity at the DIP joint was expressed as the maximum torque divided by the associated angle of twist.
Inspection of the fingers prior to mechanical testing indicated that fracture or thread penetration through the dorsal or palmar cortex occurred in 10 of 15 male fingers and 15 of 15 female fingers (Figs. 2, 3).
Results
Measurement of the Distal Phalanx For the 30 distal phalanges, the dorsal-palmar dimension (the height) was 3.55 mm + 0.55 mm (mean + 1 SD). The mediolateral dimension (the width) was 7.23 mm + 1.17 mm. For the 15 male fingers, the mean height was 3.73 mm + 0.49 mm and the mean width was 8.17 mm + 0.64 mm; for the 15 female fingers, the mean height was 3.42 m m + 0.57 m m and the mean width was 6.47 mm + 0.91 mm. The dimensions of the male and female fingers were not significantly different from each other, using a Student's t-test (p < .05 being significant). In addition, there was no significant difference between dimensions of index, middle, and ring fingers. It is important to note that the average height of the specimens (3.55 mm + 0.55 mm) is less than the diameter (3.90 mm) of the lagging threads of the Herbert screw (Fig. 1).
Figure 2. An x-ray film of a cadaveric finger specimen with a Herbert screw across the prepared distal interphalangeal joint. Note a small but visible fracture of the dorsal cortex overlying the lagging threads.
DISTAL PHALANX |
/
T 355m~mav he, ht
~t
\O.mm
/difference 3.90 mmdiameter
HERBERT SCREW
Figure 1. Schematic of average distal phalanx dorsalpalmar dimension compared to diameter of a standard Herbert compression screw. Note that the lagging thread diameter is greater than the average dorsal and palmar height of a distal phalanx.
Figure 3. The same specimen as in Figure 2 after the stretched nail bed was incised with a scalpel and elevated. The dorsal fracture with visible screw threads is easily demonstrated.
The Journal of Hand Surgery / Vol. 21A No. 3 May 1996
Mechanical Testing The results of mechanical testing are summarized in Table 1 and Figure 4. Again, Student's t-test for significance was used. In AP bending, the increased bending strength of the Herbert screw constructs compared to the TBW constructs was statistically significant. However, the difference in the bending stiffnesses of the Herbert screw constructs and the TBW constructs was not statistically significant. For the Herbert screw specimens, bending of the screw and breakout from the cortex of the distal phalanx accompanied failure of the constructs; for the tension band wire specimens, failure was accompanied by fracture at a K-wire-bone interface, usually in the distal phalanx. In lateral bending, the bending strengths of the constructs were not statistically different. A statistically significant difference did exist, however, between the two constructs, as the Herbert screw demonstrated less bending stiffness. For the Herbert screw specimens, failure of the constructs occurred with bending of the screw and breakout from the medial cortex of the distal phalanx; for the TBW constructs, failure occurred at a K-wire-bone interface (usually in the distal phalanx). In axial torsion, the Herbert screw constructs were significantly more rigid than those of the TBW. For the Herbert screw constructs, no fractures occurred and no screws were bent or broken during testing becausc only rotation of the distal phalanx with respect to the middle phalanx occurred. In the TBW constructs, fracture (usually of the distal phalanx) at a K-wire-bone interface usually accompanied failure.
Discussion A tenet in arthrodesis is that minimal nonunion rates are associated with fixation techniques that ensure stable, strong, motion-limiting constructs. It is our anecdotal experience that nonunion rates for
441
DIP joint arthrodeses are greater than those for PIP joint arthrodeses; in addition, pain secondary to hardware loosening or hardware prominence often leads to another surgical procedure for removal. Clinical experiences cited in the literature likewise suggest that nonunion rates for small-joint arthrodeses in the hand may be great, ranging from 0% to 30%. 1,2,4,6 Bucko-Gramcko2 reported on a series of 363 arthrodeses over a 9-year period. Nonunion occurred in 29 (8%) of patients treated with screw arthrodesis, compared to a 30% nonunion rate in patients who underwent K-wire arthrodesis. Union rates were 7 weeks and 10 weeks, respectively, in the two groups. This suggests that the compression arthrodesis technique is superior to the K-wire technique in both achievement of union and time to union. This report included both DIP and PIP joint arthrodesis. In the DIP joint specifically, one study comparing K-wires and compression screws reported 6 nonunions in 30 consecutive patients. Nonunion rates were identical between the two groups, but time to union and return to work was faster in the compression screw group. 12 Because of the large variation in union rates among different techniques for small-joint arthrodesis, it remains unclear in the literature if variation in union rates is attributable to fixation stability. Certainly, other important factors for small-joint arthrodesis, such as concomitant external immobilization, cigarette smoking, peripheral vascular disease, bone quality, and patient compliance, may also play a role and are difficult to analyze. Therefore, since a poor clinical outcome may be multifactorial, fixation stability and surgical technique may be the only variables over which the surgeon has control. For this reason, it may be relevant to quantify the biomechanical natures of joint arthrodesis constructs. The two techniques that were compared are ones that are presently used at our institution. This study employed a single load to failure in AP bending, lateral bending, and axial rotation. These configura-
Table 1. Summary of Mechanical Testing Anteroposterior Bending Maximum Load KW-TBW Herbert screw
Lateral Bending
Torsion
Bending Stiffness (Nmm e)
Maximum Load
(N)
(N)
Bending Stiffness (Nmm 2)
Torsional Rigidity (Nmm 2)
135.8 + 45.3 195.0 + 58.9
16,720 + l 2,490 11,830 +_9,3602
245.7 + 36.2 12.6 + 57.5
14,180 + 14,420 10,160 + 2,500
18,740 + 7,120 64,150 _+24,310
KW-TBW, Kirschner-wire tension-bandwire.
442
Wyrsch et al. /Tension-band Wire vs Herbert Screw Property Value Normalized I~y K W - T B W 5
T
Strength
Stiffness
AP Bending
Strength
Stiffness
Lateral Bending
-
Rigidity Axial Torsion
Figure 4. Loading modes and structural properties of Kirschner wire tension-band wires versus Herbert compression screws. The data have been normalized to the tension-band wire so that the Herbert screw data can be viewed as a percentage comparison. The increased strength and torsional rigidity for the Herbert screw construct in anteroposterior bending and axial torsion testing, respectively, are significant.
tions were chosen to represent loads associated with vigorous activity, heavy labor, trauma, and passive manipulation. Activities of daily living, such as gentle tip or key pinch, are likely to generate forces that do not lead to catastrophic failures like those induced in this study. Nevertheless, by quantifying the strengths of the constructs, the results of this study describe not only the mechanical stiffnesses of the constructs but also their functional limits. Another difference between the loads of this study and physiologic loads is that, in testing, a single load cycle was used, whereas most activities of daily living are repetitive. The consequences of repetitive loading on DIP joint arthrodeses have not been reported in the literature and are outside the scope of this study. Our study compares biomechanical characteristics of two DIP fusion constructs. Previous studies in the literature have evaluated biomechanical characteristics of PIP fusion constructs. Kovach et al. 13 compared crossed K-wires, interosseous wiring, and figure-of-eight TBW, analyzing three-point AP bending, axial torsion, and lateral cantilever bending. No screw constructs were tested. Ayres ll et al. compared biomechanical properties of PIP joints fixed with either a TBW or a Herbert screw. Only AP bending testing was performed, yielding similar biomechanical properties between the two constructs. The functional stiffness requirements of constructs for joint fusion are not known. Likewise, the consequences of the decreased lateral bending stiffness and of the increased torsional rigidity of the Herbert compression screw construct with respect to the Kwire-TBW construct are not known. However, it is important to recognize that the Herbert compression screw construct possessed adequate stiffness to pre-
vent internal or external rotation of the distal phalanx with respect to the middle phalanx. Overall, the biomechanical testing performed demonstrates that the Herbert compression screw appears to offer additional strength that may be clinically important for joint arthrodesis. Of serious interest in this study is the incidence of screw thread penetration or fracture of the palmar or dorsal cortex of the distal phalanx with screw placement. The measured dimensions of the phalanges demonstrate that the screw thread diameter and the distal phalanx diameter are frequently similar. Indeed, fracture or screw penetration occurred at the tip of the distal phalanx in 25 of the specimens overall and in all 15 of the female fingers. The clinical significance of these fractures or of screw thread penetration is not known. One half of the fractures occurred in the distal cortex, resulting in stretching or violation of the nail bed. Despite the almost inevitable occurrence of thread penetration or prominence with their use, Herbert compression screws produce DIP joint fixation that is adequately strong. Because screw thread penetration and, perhaps, some degree of fracture of the phalanx can occur, palmar cortex penetration of the Herbert compression screw threads may be preferable to dorsal cortex penetration. If the dorsal cortex is penetrated, the nail bed may be injured, leading to the potential complications of postoperative pain, subungual hematoma, and nail deformity. Patients undergoing DIP arthrodesis with the Herbert compression screw should be informed of these potential surgical complications. Cortex penetration may be avoided by using the Herbert mini screw (Zimmer Inc., Warsaw, IN),
The Journal of Hand Surgery / Vol. 21A No. 3 May 1996
which has a smaller trailing thread diameter of 3.2 mm and thus is more easily accommodated by the distal phalanx. Our experience with this device, however, is very limited, and biomechanical analysis has yet to be performed. In addition, the available lengths for this particular screw may not be adequate for DIP arthrodesis, as most of our arthrodeses require screw lengths of 2 mm or more to achieve adequate purchase and compression. Certainly, a controlled clinical trial comparing these fixation techniques is needed to compare rates of union, failure, and other potential complications. With the advent of new compression screws and other fixation techniques, stronger biomechanical constructs likely exist. Ideally, further studies may be able to predict a construct that provides the optimum degree of stability and compression needed for promoting bony union. We would like to thank Holly Quick for assistance in preparing the manuscript.
References 1. Allende BT, Engelman JC. Tension band arthrodesis in finger joints. J Hand Surg 1980;5:269-271.
443
2. Bucko-Gramcko D. Compression arthrodesis of joints in the hand. In: Tubiana R, ed. The hand. Vol 2. Philadelphia: WB Saunders, 1985:703-706. 3. Burton RI. Small-joint arthrodesis in the hand. J Hand Surg 1986; 11A:678-681. 4. Carroll RE, Hill NA. Small joint arthrosis in hand reconstruction. J Bone Joint Surg 1969;51A: 1219-1221. 5. Herbert TJ. Management of the fractured scaphoid bone using a new surgical technique. J Bone Joint Surg 1982; 64B:633. 6. Lister G. Intraosseous wiring of the digital skeleton. J Hand Surg 1978;3:427-434. 7. Moberg E. Arthrodesis of finger joints. Surg Clin North Am 1960;40:465-470. 8. Potenza AD. A technique for arthrodesis of finger joints. J Bone Joint Surg 1973;55A:1534--1536. 9. Stern PJ, Drury WJ. Tension band arlhrodesis of small joints. Contemp Ortbop 1984;8:59-62. 10. Herbert TJ, Faithful DK. Small joint fusions of the hand using the Herbert bone screw. J Hand Surg 1984;9B: 167-168. 11. Ayres JR, Goldstrohm GL, Miller GJ, Dell PC. Proximal interphalangeal joint arthrodesis with the Herbert screw. J Hand Surg 1988;13A:60~603. 12. Engel J, Tsur H, Farin I. A comparison between K-wire and compression screw fixation after arthrodesis of the DIP joint. Plast Reconstr Surg 1977;60:611-614. 13. Kovach JC, Werner FW, Palmer AK, Greenkey S, Murphy DJ. Biomechanical analysis of internal fixation techniques for PIP joint arthrodesis. J Hand Surg 1986;1 IA:562-566.
Distal interphalangeal (DIP) joint arthrodesis is indicated for relief of painful conditions such as osteoarthritis and post-traumatic arthritis and for stabilization of the joint following severe bony trauma and irreparable tendon injuries. Several techniques of fixation have been described for DIP arthrodeses, From the Department of Orthopaedics and Rehabilitation, Vanderbilt University Medical Center, Nashville, TN. Received for publication Jan. 30, 1995; accepted in revised form Aug. 8, 1995. Although the authol's have not received or will not receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article, benefits have been or will be received but are directed solely to a research fund, foundation, educational institution, or other nonprofit organization with which one or more of the authors are associated. Reprint requests: Brad Wyrsch, MD, Cincinnati Hand Center, 2800 Winslow Avenue, Cincinnati, OH 45206.
438
The Journal of Hand Surgery
including Kirschner wires (K-wires), bone pegs, interosseous or tension-band wires (TBW), and compression screws. 1-9 More recently, Herbert compression screws have been employed for interphalangeal joint arthrodeses, m Proximal interphalangeal (PIP) joint arthrodesis with the Herbert screw has achieved excellent clinical results, u In addition, equivalent biomechanical strength has been demonstrated for the Herbert screw and TBW techniques in PIP joints.ll However, despite wide clinical use of these techniques, there is a paucity of literature concerning the use of Herbert screws in the DIP joint. With limited experience at our institution, Herbert screw DIP arthrodesis has achieved a 100% union rate with no complications. This was a two-part study. In the first part, the height and width of distal phalangeal specimens
The Journal of Hand Surgery/Vol. 21A No. 3 May 1996 439
were measured to see how the distal phalanx of the adult finger accommodates the large threads of the Herbert screw. In the second part, biomechanical testing was performed, comparing simulated DIP fusion constructs of TBW and the Herbert screw.
Materials and Methods Measurement of the Distal Phalanx
The index, middle, and ring fingers of five male and five female embalmed cadaveric hands were stripped of soft tissue, leaving only the palmar plates. The terminal slip of the extensor tendon was excised, as were the collateral ligaments. The transverse dimensions of distal phalanges were measured 4 mm proximal to the distal tips of the bones, using a caliper. This area of the distal phalanx is the region that must accommodate the large-diameter (lagging) threads of the Herbert screw. Measurements were taken in two mutually perpendicular planes: the dorsal-palmar dimension (the height) and the mediolateral dimension (the width). Fixation Techniques The DIP joint cartilage was excised in all specimens, using an osteotome or rongeur, to create a cup-and-cone configuration of cancellous bone. In 15 of the fingers, the DIP joints were fixed using Herbert screws (24-28 mm in length) in the following manner: a small K-wire was used to produce a guide hole, proceeding from proximal to distal through the center of the medullary canal of the distal phalanx; the 2-ram-diameter Herbert drill was passed manually through the guide hole from distal to proximal through both phalanges, while care was taken to center the drill in both bones; the large Herbert drill was then used to enlarge the screw hole in the distal phalanx; and finally, a Herbert screw was placed retrograde until the large-diameter (lagging) threads of the screw completely engaged in the distal phalanx and compression was achieved. The authors who prepared these specimens (B. W. and S. A.) believed that excellent placement and fixation of the screw was achieved in all specimens. In the remaining 15 fingers, the DIP joints were fixed using K-wires and a dorsal TBW in the following manner: two medial-to-lateral transverse holes were created in both the middle and distal phalanges (the holes were dorsal to the midline of the bones) using a 0.35-mm K-wire. A 26-gauge wire was passed through these holes and crossed dorsally, then twisted counterclockwise, to create a figure-of-eight
dorsal tension band with the DIP joint in flexion (0~176 next, two 0.35-ram K-wires were placed in a crossed fashion palmar to the TBWs. The authors who prepared these specimens (B. W. and S. A.) believed that excellent fixation was achieved in all specimens. Hardware placement was then evaluated by direct inspection and x-ray films. For the Herbert screw, those instances when the screw threads penetrated bone cortices were noted. No specimens were discarded. Mechanical Testing
All specimens were mechanically tested; tests included anteroposterior (AP) bending, lateral bending, and axial torsion (five specimens for each test of both constructs). Repeated use of a specimen was not possible because each test was carried out to failure of the joint. Failure was defined as a displaced fracture of the specimen or cutout of a wire or screw through cortical bone. Each DIP fusion construct was potted into a specially fabricated test fixture. This consisted of two stainless-steel square blocks with drill holes for screw fixation to mount onto an MTS (Material Testing Systems, 3M Corps., Minneapolis, MN). A specimen could be positioned in this fixture in one of three orientations, corresponding to the type of test to be conducted. The proximal aspect of the middle phalanx and the distal tip of the distal phalanx were secured with four small locking screws at the base of each wall of the recessed area of the block. In addition, the recessed area of each block was filled with resin (Technovit 4000, Kulzer Inc., Germany), which was allowed to solidify. For the AP bending tests and the lateral bending tests, the distances from the proximal pot to the DIP joint (a) and from the distal pot to the DIP joint (b) and the total specimen length (L) were measured. These distances (moments) are needed to compute the bending stiffness (EI) of the DIP joint construct, using the following equation: EI = Pa3b3/3dL3 At the time of failure of the specimen, the load (P) and displacement (d) were used to calculate the bending stiffness of the specimen. The bending tests were conducted under displacement control at 5 mm/min, to a maximum displacement of 10 ram. For the AP bending test, specimens were placed horizontally, with the dorsal aspect of the joint facing up. Thus, the test apparatus created a bending moment to
440
Wyrsch et al./Tension-band Wire vs Herbert Screw
the dorsal surface of the DIP joint. For the lateral bending test, specimens were placed horizontally with the radial aspect of the DIP joint facing up. For this test, the apparatus created a bending moment to the lateral (radial) surface of the DIP joint. All specimens failed prior to the maximum displacement. The applied axial rotation and the resulting torque were measured throughout a torsion test. The torsion tests were conducted under angle control at 10~ to a maximum angle of 30 ~ The maximum torque and its associated angle o f twist were used to calculate the torsional rigidity of the specimen. The torsional rigidity at the DIP joint was expressed as the maximum torque divided by the associated angle of twist.
Inspection of the fingers prior to mechanical testing indicated that fracture or thread penetration through the dorsal or palmar cortex occurred in 10 of 15 male fingers and 15 of 15 female fingers (Figs. 2, 3).
Results
Measurement of the Distal Phalanx For the 30 distal phalanges, the dorsal-palmar dimension (the height) was 3.55 mm + 0.55 mm (mean + 1 SD). The mediolateral dimension (the width) was 7.23 mm + 1.17 mm. For the 15 male fingers, the mean height was 3.73 mm + 0.49 mm and the mean width was 8.17 mm + 0.64 mm; for the 15 female fingers, the mean height was 3.42 m m + 0.57 m m and the mean width was 6.47 mm + 0.91 mm. The dimensions of the male and female fingers were not significantly different from each other, using a Student's t-test (p < .05 being significant). In addition, there was no significant difference between dimensions of index, middle, and ring fingers. It is important to note that the average height of the specimens (3.55 mm + 0.55 mm) is less than the diameter (3.90 mm) of the lagging threads of the Herbert screw (Fig. 1).
Figure 2. An x-ray film of a cadaveric finger specimen with a Herbert screw across the prepared distal interphalangeal joint. Note a small but visible fracture of the dorsal cortex overlying the lagging threads.
DISTAL PHALANX |
/
T 355m~mav he, ht
~t
\O.mm
/difference 3.90 mmdiameter
HERBERT SCREW
Figure 1. Schematic of average distal phalanx dorsalpalmar dimension compared to diameter of a standard Herbert compression screw. Note that the lagging thread diameter is greater than the average dorsal and palmar height of a distal phalanx.
Figure 3. The same specimen as in Figure 2 after the stretched nail bed was incised with a scalpel and elevated. The dorsal fracture with visible screw threads is easily demonstrated.
The Journal of Hand Surgery / Vol. 21A No. 3 May 1996
Mechanical Testing The results of mechanical testing are summarized in Table 1 and Figure 4. Again, Student's t-test for significance was used. In AP bending, the increased bending strength of the Herbert screw constructs compared to the TBW constructs was statistically significant. However, the difference in the bending stiffnesses of the Herbert screw constructs and the TBW constructs was not statistically significant. For the Herbert screw specimens, bending of the screw and breakout from the cortex of the distal phalanx accompanied failure of the constructs; for the tension band wire specimens, failure was accompanied by fracture at a K-wire-bone interface, usually in the distal phalanx. In lateral bending, the bending strengths of the constructs were not statistically different. A statistically significant difference did exist, however, between the two constructs, as the Herbert screw demonstrated less bending stiffness. For the Herbert screw specimens, failure of the constructs occurred with bending of the screw and breakout from the medial cortex of the distal phalanx; for the TBW constructs, failure occurred at a K-wire-bone interface (usually in the distal phalanx). In axial torsion, the Herbert screw constructs were significantly more rigid than those of the TBW. For the Herbert screw constructs, no fractures occurred and no screws were bent or broken during testing becausc only rotation of the distal phalanx with respect to the middle phalanx occurred. In the TBW constructs, fracture (usually of the distal phalanx) at a K-wire-bone interface usually accompanied failure.
Discussion A tenet in arthrodesis is that minimal nonunion rates are associated with fixation techniques that ensure stable, strong, motion-limiting constructs. It is our anecdotal experience that nonunion rates for
441
DIP joint arthrodeses are greater than those for PIP joint arthrodeses; in addition, pain secondary to hardware loosening or hardware prominence often leads to another surgical procedure for removal. Clinical experiences cited in the literature likewise suggest that nonunion rates for small-joint arthrodeses in the hand may be great, ranging from 0% to 30%. 1,2,4,6 Bucko-Gramcko2 reported on a series of 363 arthrodeses over a 9-year period. Nonunion occurred in 29 (8%) of patients treated with screw arthrodesis, compared to a 30% nonunion rate in patients who underwent K-wire arthrodesis. Union rates were 7 weeks and 10 weeks, respectively, in the two groups. This suggests that the compression arthrodesis technique is superior to the K-wire technique in both achievement of union and time to union. This report included both DIP and PIP joint arthrodesis. In the DIP joint specifically, one study comparing K-wires and compression screws reported 6 nonunions in 30 consecutive patients. Nonunion rates were identical between the two groups, but time to union and return to work was faster in the compression screw group. 12 Because of the large variation in union rates among different techniques for small-joint arthrodesis, it remains unclear in the literature if variation in union rates is attributable to fixation stability. Certainly, other important factors for small-joint arthrodesis, such as concomitant external immobilization, cigarette smoking, peripheral vascular disease, bone quality, and patient compliance, may also play a role and are difficult to analyze. Therefore, since a poor clinical outcome may be multifactorial, fixation stability and surgical technique may be the only variables over which the surgeon has control. For this reason, it may be relevant to quantify the biomechanical natures of joint arthrodesis constructs. The two techniques that were compared are ones that are presently used at our institution. This study employed a single load to failure in AP bending, lateral bending, and axial rotation. These configura-
Table 1. Summary of Mechanical Testing Anteroposterior Bending Maximum Load KW-TBW Herbert screw
Lateral Bending
Torsion
Bending Stiffness (Nmm e)
Maximum Load
(N)
(N)
Bending Stiffness (Nmm 2)
Torsional Rigidity (Nmm 2)
135.8 + 45.3 195.0 + 58.9
16,720 + l 2,490 11,830 +_9,3602
245.7 + 36.2 12.6 + 57.5
14,180 + 14,420 10,160 + 2,500
18,740 + 7,120 64,150 _+24,310
KW-TBW, Kirschner-wire tension-bandwire.
442
Wyrsch et al. /Tension-band Wire vs Herbert Screw Property Value Normalized I~y K W - T B W 5
T
Strength
Stiffness
AP Bending
Strength
Stiffness
Lateral Bending
-
Rigidity Axial Torsion
Figure 4. Loading modes and structural properties of Kirschner wire tension-band wires versus Herbert compression screws. The data have been normalized to the tension-band wire so that the Herbert screw data can be viewed as a percentage comparison. The increased strength and torsional rigidity for the Herbert screw construct in anteroposterior bending and axial torsion testing, respectively, are significant.
tions were chosen to represent loads associated with vigorous activity, heavy labor, trauma, and passive manipulation. Activities of daily living, such as gentle tip or key pinch, are likely to generate forces that do not lead to catastrophic failures like those induced in this study. Nevertheless, by quantifying the strengths of the constructs, the results of this study describe not only the mechanical stiffnesses of the constructs but also their functional limits. Another difference between the loads of this study and physiologic loads is that, in testing, a single load cycle was used, whereas most activities of daily living are repetitive. The consequences of repetitive loading on DIP joint arthrodeses have not been reported in the literature and are outside the scope of this study. Our study compares biomechanical characteristics of two DIP fusion constructs. Previous studies in the literature have evaluated biomechanical characteristics of PIP fusion constructs. Kovach et al. 13 compared crossed K-wires, interosseous wiring, and figure-of-eight TBW, analyzing three-point AP bending, axial torsion, and lateral cantilever bending. No screw constructs were tested. Ayres ll et al. compared biomechanical properties of PIP joints fixed with either a TBW or a Herbert screw. Only AP bending testing was performed, yielding similar biomechanical properties between the two constructs. The functional stiffness requirements of constructs for joint fusion are not known. Likewise, the consequences of the decreased lateral bending stiffness and of the increased torsional rigidity of the Herbert compression screw construct with respect to the Kwire-TBW construct are not known. However, it is important to recognize that the Herbert compression screw construct possessed adequate stiffness to pre-
vent internal or external rotation of the distal phalanx with respect to the middle phalanx. Overall, the biomechanical testing performed demonstrates that the Herbert compression screw appears to offer additional strength that may be clinically important for joint arthrodesis. Of serious interest in this study is the incidence of screw thread penetration or fracture of the palmar or dorsal cortex of the distal phalanx with screw placement. The measured dimensions of the phalanges demonstrate that the screw thread diameter and the distal phalanx diameter are frequently similar. Indeed, fracture or screw penetration occurred at the tip of the distal phalanx in 25 of the specimens overall and in all 15 of the female fingers. The clinical significance of these fractures or of screw thread penetration is not known. One half of the fractures occurred in the distal cortex, resulting in stretching or violation of the nail bed. Despite the almost inevitable occurrence of thread penetration or prominence with their use, Herbert compression screws produce DIP joint fixation that is adequately strong. Because screw thread penetration and, perhaps, some degree of fracture of the phalanx can occur, palmar cortex penetration of the Herbert compression screw threads may be preferable to dorsal cortex penetration. If the dorsal cortex is penetrated, the nail bed may be injured, leading to the potential complications of postoperative pain, subungual hematoma, and nail deformity. Patients undergoing DIP arthrodesis with the Herbert compression screw should be informed of these potential surgical complications. Cortex penetration may be avoided by using the Herbert mini screw (Zimmer Inc., Warsaw, IN),
The Journal of Hand Surgery / Vol. 21A No. 3 May 1996
which has a smaller trailing thread diameter of 3.2 mm and thus is more easily accommodated by the distal phalanx. Our experience with this device, however, is very limited, and biomechanical analysis has yet to be performed. In addition, the available lengths for this particular screw may not be adequate for DIP arthrodesis, as most of our arthrodeses require screw lengths of 2 mm or more to achieve adequate purchase and compression. Certainly, a controlled clinical trial comparing these fixation techniques is needed to compare rates of union, failure, and other potential complications. With the advent of new compression screws and other fixation techniques, stronger biomechanical constructs likely exist. Ideally, further studies may be able to predict a construct that provides the optimum degree of stability and compression needed for promoting bony union. We would like to thank Holly Quick for assistance in preparing the manuscript.
References 1. Allende BT, Engelman JC. Tension band arthrodesis in finger joints. J Hand Surg 1980;5:269-271.
443
2. Bucko-Gramcko D. Compression arthrodesis of joints in the hand. In: Tubiana R, ed. The hand. Vol 2. Philadelphia: WB Saunders, 1985:703-706. 3. Burton RI. Small-joint arthrodesis in the hand. J Hand Surg 1986; 11A:678-681. 4. Carroll RE, Hill NA. Small joint arthrosis in hand reconstruction. J Bone Joint Surg 1969;51A: 1219-1221. 5. Herbert TJ. Management of the fractured scaphoid bone using a new surgical technique. J Bone Joint Surg 1982; 64B:633. 6. Lister G. Intraosseous wiring of the digital skeleton. J Hand Surg 1978;3:427-434. 7. Moberg E. Arthrodesis of finger joints. Surg Clin North Am 1960;40:465-470. 8. Potenza AD. A technique for arthrodesis of finger joints. J Bone Joint Surg 1973;55A:1534--1536. 9. Stern PJ, Drury WJ. Tension band arlhrodesis of small joints. Contemp Ortbop 1984;8:59-62. 10. Herbert TJ, Faithful DK. Small joint fusions of the hand using the Herbert bone screw. J Hand Surg 1984;9B: 167-168. 11. Ayres JR, Goldstrohm GL, Miller GJ, Dell PC. Proximal interphalangeal joint arthrodesis with the Herbert screw. J Hand Surg 1988;13A:60~603. 12. Engel J, Tsur H, Farin I. A comparison between K-wire and compression screw fixation after arthrodesis of the DIP joint. Plast Reconstr Surg 1977;60:611-614. 13. Kovach JC, Werner FW, Palmer AK, Greenkey S, Murphy DJ. Biomechanical analysis of internal fixation techniques for PIP joint arthrodesis. J Hand Surg 1986;1 IA:562-566.