 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1997 The Journal of Bone and Joint Surgery, Inc. A Comparison of Fixation Screws for the Scaphoid during Application of Cyclical Bending Loads* , KANSAS CITY, KANSASInvestigation performed at The University of Kansas Medical Center, Kansas City
Matched pairs of scaphoids from cadavera were stressed with ramped intensity cyclical bending loads after osteotomy and fixation of one scaphoid with a Herbert screw and fixation of the other with an AO 3.5-millimeter cannulated screw, a Herbert-Whipple screw, an Acutrak cannulated screw, or a Universal Compression screw. The AO screw, Acutrak screw, and Herbert-Whipple screw demonstrated superior resistance to cyclical bending loads compared with the Herbert screw. The Universal Compression screw did not provide better fixation than the Herbert screw because of fractures that occurred at the time of insertion. The AO screw and the Herbert screw were then tested in a separate setup in which a segment of volar cortex had been removed in addition to the simple osteotomy. The loss of volar cortex greatly diminished the quality of the fixation provided by both of the screws during application of ramped intensity cyclical bending loads. CLINICAL RELEVANCE: A fixation device in the scaphoid must be able to withstand the stresses that are placed on the scaphoid as a result of its position spanning the proximal and distal carpal rows. Also, because of the prolonged time required for healing of fractures or non-unions of the scaphoid, the device must be able to withstand many such cycles of stress. The present study demonstrates that commonly used screws for fixation of the scaphoid vary significantly (p < 0.005) in their ability to resist cyclical bending loads.
Fractures of the scaphoid are a challenge for hand surgeons. There is a risk of delayed healing or non-union not only because of the compromised blood supply but also because of the persistent forces applied to the scaphoid as a result of its critical position between the proximal and distal carpal rows. These substantial bending forces tend to cause dorsal angulation when the scaphoid is fractured1,3,7,21,22,25. Internal fixation of the scaphoid is difficult because of its small size, unique geometry, and predominance of articular surface area. In 1984, Herbert and Fisher reported their experience with a new bone screw designed for fixation of the scaphoid11. This screw had leading and trailing threads of different pitch to provide compression between the proximal and distal fragments when tightened. Because the screw is small and has no head, it has advantages for fixation of fractures and non-unions of the scaphoid compared with conventional screws. Several biomechanical studies comparing the fixation properties of commercially available devices for use in the scaphoid have been published recently5,13,15,17,19,20. Despite reports that it has inferior pull-out strength and generates less compressive force compared with standard screws, the Herbert screw has become popular for treatment of fractures and non-unions of the scaphoid. Additionally, a number of clinical studies of the device have demonstrated satisfactory results4,8,9,12,16,24. The purpose of our study was to subject devices for fixation of the scaphoid to a more physiological pattern of loading in order to assess the quality of fixation. Since the scaphoid is primarily subjected to bending forces and tends to subside into a palmar-flexed position with non-union, a cyclical bending test was chosen for this experiment. The Herbert screw was selected as the standard for comparison with other fixation devices because it has gained popularity for the treatment of fractures and non-unions of the scaphoid.
Thirty-five matched pairs of human cadaveric scaphoids were chosen from a pool of 100 matched pairs of scaphoids that had been collected and preserved in 10 per cent formalin solution. Each pair consisted of the right and left scaphoids from one individual. Scaphoids that were exceptionally small or large, had an unusual shape, or had evidence of previous injury were not used. The scaphoids were cleaned of all soft tissue and were symmetrically marked circumferentially at the narrowest aspect of the waist. A smooth transverse osteotomy was performed along the markings with a thin-bladed circular saw. The osteotomized pairs were then randomly assigned to testing groups. Six study groups were formed (Table I). Group 1 consisted of five pairs of scaphoids. The first pair was fixed with Herbert screws (Zimmer, Warsaw, Indiana) after the osteotomy; the second, with AO 3.5-millimeter cannulated, partially threaded screws (Synthes, Waldenburg, Switzerland); the third, with Herbert-Whipple screws (Zimmer); the fourth, with cannulated, conical Acutrak screws (Acumed, Beaverton, Oregon); and the fifth, with Universal Compression screws (Howmedica, Rutherford, New Jersey) (Fig. 1). This group was the control group, with the same fixation devices in the right and left scaphoids from the same individual tested with cyclical bending loads.
Groups 2 through 6 each consisted of six matched pairs of scaphoids, with one scaphoid of each pair fixed with a Herbert screw. In group 2, the other scaphoid was fixed with an AO 3.5-millimeter cannulated screw. In group 3, the other scaphoid was also fixed with an AO 3.5-millimeter cannulated screw, but in each scaphoid of the pair the fixation was performed after not only simple osteotomy but also resection of a two-millimeter volar wedge from the proximal fragment with a small hand-held saw. The cut extended approximately to the mid-sagittal line of the waist. The volar wedge resection was performed to simulate volar comminution. The AO screw and the Herbert screw were selected for this group because they seem to represent the spectrum of fixation quality. In group 4, the contralateral scaphoid was fixed with a Herbert-Whipple screw; in group 5, with an Acutrak screw; and in group 6, with a Universal Compression screw. The left and right sides were alternated with respect to placement of the various fixation devices in groups 2 through 6. All screws were placed in accordance with the manufacturer's recommendation. The Herbert screws and the Herbert-Whipple screws were placed through the appropriate guide-jigs with precompression applied to the osteotomy site. When the AO screws and the Herbert screws were compared in groups 2 and 3 and the Acutrak screws and the Herbert screws were compared in group 5, the guide-wire was placed first and the Herbert jig was applied subsequently to the contralateral scaphoid at the corresponding entrance and exit sites of the guide-wire on the first scaphoid. This technique ensured nearly symmetrical placement of the fixation devices for individual pairs of scaphoids. The AO screws and the Universal Compression screws were countersunk into the scaphoid tuberosity and in general were approximately two millimeters longer than the corresponding Herbert screw. High-resolution radiographs in two planes were made of all specimens before testing to confirm symmetrical placement of the screw and to ensure that all leading screw threads were across the osteotomy site. On completion of testing, radiographs of the scaphoids were made to examine the mode of failure. Cyclical testing of the specimens was performed on a custom-designed and fabricated loading device attached to a servohydraulic materials-testing machine (MTS, Eden Prairie, Minnesota) (Fig. 2). The loading device consisted of lower and upper beams hinged together at one end. The free end of the lower beam was attached to the lower table (actuator) of the servohydraulic machine. The free end of the upper beam was connected through a slot with a bolt to the load cell in the fixed head of the servohydraulic machine. The proximal pole of each scaphoid was secured within a shallow square cup by sharpened screws from two opposite sides of the cup and by a pin across the other sides of the cup (Fig. 3). The cup filled with polymethylmethacrylate was fixed to the lower beam of the device 225 millimeters away from the hinged end. The distal pole was secured to a freely rolling sled that lay within the upper beam. A smooth pin was then driven across the distal fragment parallel to the plane of the osteotomy. This pin was accurately placed volar to the most volar aspect of the osteotomy to effect a desired bending moment to the volar cortex. Additionally, great care was taken to ensure that placement of the pin was symmetrical with that of the corresponding scaphoid by fixing both scaphoids simultaneously in separate loading devices. Cannulated screws secured on either side of the sled were then placed over the pin and were tightened to abut firmly against the cortex on each side of the distal fragment, just dorsal to the fixation screw. The specimens were kept moist with 10 per cent formalin solution throughout the testing sequence.
The mechanism of loading resulted in bending of the scaphoid due to a force applied at the distal pole with the proximal fragment fixed. This was accomplished when a load was applied at the free end of the upper beam and a bending force was induced through the pin in the distal pole by the upper beam acting as a simple rigid lever. The bending force (Q') acting on thq scaphoid was related to the compressive load (P') because of the compression load (P) applied at the free end of the upper beam by the relation P' = 1.56 x P' (Fig. 2). The line of action of the bending force from the proximal pole was offset by e' = 55 per cent of the width of the scaphoid (Fig. 2) and hence the bending moment on the scaphoid reduced to M = Q'e' The bending rigidity of the experimental apparatus was not tested before each specimen was tested, but since the same apparatus, as well as the same-size smooth pin, was used in all cases errors due to bending of the apparatus itself were minimized. Initial deflections from motion between the smooth pin and the cannulated screw were subtracted from the total displacements when the cycles required for a given displacement were calculated. A hemisinusoidal cyclical loading at two cycles per second was applied, starting with a compression load (P) of five newtons. A tension load of five newtons was applied between each cycle of compression loading. This unloaded the device and brought the apparatus back to near the beginning deflection point, allowing full displacement to occur with each compression cycle. The compression load was increased 0.01 newton per cycle in a ramp fashion to a maximum compression load of 105 newtons at 10,000 cycles2,6,23. The corresponding bending force (Q') on the scaphoid was therefore increased 0.0158 newton per cycle to a maximum force of 166 newtons. For an average value of e' = 7.81 millimeters, the bending moment was increased 0.123 newton-millimeter per cycle to a maximum bending moment of 1300 newton-millimeters. An extensometer was attached to the proximal fragment fixation cup and the distal fragment sled, and displacement data were recorded simultaneously with a strip-chart recorder. The numbers of cycles needed to produce 0.5 millimeter, 1.0 millimeter, and 1.5 millimeters of displacement were recorded. In addition, the specimens were observed during the cyclical loading and were inspected after testing for clues to patterns of failure. The significance of differences between experimental groups was analyzed with paired sample t testing. The level of significance was p < 0.05.
In group 1, the results for the right and left scaphoids, which had the same fixation device, were similar. The Hotelling T square test for paired observation was performed to test the difference between the right and the left side for the three types of displacements simultaneously. It showed no significance, with the numbers available. Also, the paired t test was performed for the difference between the right and the left side separately for each type of displacement. Again, there was no significance with the numbers available (Table II). These data reflected that the fixation and mounting techniques could be performed in a reproducible and symmetrical manner.
The scaphoids fixed with the Herbert screw demonstrated a different pattern of failure than those fixed with the AO screw. The Herbert screw consistently failed first at the proximal (leading) threads, which are smaller, while the distal (trailing) threads remained firmly attached to the distal pole. As the proximal threads lost fixation, the narrow shaft of the screw was observed to piston easily in the cancellous bone at the site of the osteotomy. This resulted in a rapidly progressive pattern of failure (Fig. 4). In contrast, the AO screw remained fixed in the proximal fragment throughout testing, moved little during most of the progressive cyclical loading, and failed as the distal fragment rocked back and forth on the threadless trailing aspect of the screw. The larger shaft of the AO screw appeared to resist pistoning at the site of the osteotomy, accounting in part for the more gradual pattern of failure (Fig. 4). The pattern of failure of the Herbert-Whipple screw was very similar to that of the Herbert screw, but there was a greater number of ramped intensity loading cycles at each displacement. The Acutrak screw failed gradually and not catastrophically (Fig. 5). The Universal Compression screw provided strong fixation in four of the six pairs of scaphoids; in the other two pairs, there was rapid failure when a fracture occurred at the distal pole of the scaphoid as the screw was tightened and the head of the screw caused comminution of the distal pole (Fig. 6). Thus, there were two different patterns of failure for the Universal Compression screw (Fig. 6).
The AO screw showed superior resistance to ramped intensity cyclical bending loads compared with the Herbert screw. The number of loading cycles required to produce displacements of 0.5, 1.0, and 1.5 millimeters with the AO screw was more than twice that required to produce the same displacements with the Herbert screw (p < 0.005) (Table III). Since the cyclical intensity increased with the increasing number of cycles, these findings indicate meaningful differences in clinical performance. The addition of volar comminution (group 3) dramatically reduced the number of loading cycles required for displacements of 0.5, 1.0, and 1.5 millimeters during testing of both the Herbert screw and the AO screw (Table III). However, the relative differences in the quality of fixation between the scaphoids tested with and without volar comminution appeared to be similar for the two devices.
The number of ramped intensity loading cycles required for a given displacement during testing of the Herbert-Whipple screw was 1.67 (p < 0.03), 1.87 (p < 0.0009), and 2.13 (p < 0.0001) times higher than that required with the Herbert screw (Table III). This difference was significant. Approximately 1.6 times more ramped intensity loading cycles were required to cause 0.5 (p < 0.12), 1.0 (p < 0.05), and 1.5 (p < 0.05) millimeters of displacement with the Acutrak screw, compared with that required to cause the displacements with the Herbert screw. The number of ramped intensity loading cycles required for each chosen displacement during testing of the Universal Compression screw was not significantly different from that required during testing of the Herbert screw. However, the fixation of the scaphoids that had not fractured at the distal pole during tightening of the Universal Compression screw was superior to that of the scaphoids fixed with the Herbert screw. The average loads on the scaphoids needed to produce 0.5, 1.0, and 1.5 millimeters of displacement paralleled the cycle data (Table III). The loads needed to produce each displacement during testing of the AO screw, the Acutrak screw, and the Herbert-Whipple screw were greater than that needed during testing of the Herbert screw. The fractures of the distal pole of the scaphoid that occurred during insertion of the Universal Compression screw decreased the average loads required to produce the given displacements with that screw; thus, the loads were not significantly different from those required to produce displacement with the Herbert screw.
Because of the great variability among scaphoids from different cadavera, all of the screws in this study were compared with the Herbert screw in a matched-pair analysis. Therefore, the Acutrak screw was not directly compared with the AO screw, for example. We believe, however, that the relative quality of the fixation provided by the devices can be inferred with use of the Herbert screw as the standard. We used the AO 3.5-millimeter cannulated screw and the Herbert screw when we tested the effect of volar comminution as these two screws seem to represent the spectrum of quality of fixation in our experiment and in previous studies17,19,20. In our study, the loss of volar bone, which was simulated by the wedge resection, substantially diminished the performance of both the Herbert screw and the AO screw. The AO screw and the Herbert-Whipple screw resisted cyclical bending loads significantly better than the Herbert screw did. The Acutrak screw was significantly better than the Herbert screw only with regard to the number of cycles required to produce 1.0 and 1.5 millimeters of displacement. The Universal Compression screw provided better fixation than the Herbert screw in four of the six pairs. However, it was not significantly different from the Herbert screw with regard to the number of cycles required to produce any of the displacements. This was because the distal pole of the scaphoid fractured during insertion of the screw in the remaining two specimens. The Universal Compression screw in these two specimens failed with a minimum number of ramped intensity loading cycles. Previous biomechanical studies have concerned compression produced by as well as the pull-out strength of fixation devices for the scaphoid. Shaw19 measured compressive forces generated by a Herbert screw and an ASIF 4.0-millimeter cancellous-bone screw with use of a custom-designed load washer and found that the ASIF screw produced significantly more compression than the Herbert screw (p < 0.0001). On the basis of that experiment, Shaw proposed that a small cannulated screw is ideal for fixation of the scaphoid. In a follow-up study, Shaw20 tested small ASIF cannulated screws (3.5-millimeter thread diameter) and found that they provided 2.5 times more compression capability than Herbert screws did. Rankin et al. performed a similar study, comparing the compressive forces generated by a Herbert screw with that generated by a variety of other standard screws. Again, the 3.5-millimeter cannulated screw generated significantly greater compression forces than the Herbert screw did (p < 0.001). Kaulesar Sukul et al. found no large differences among Herbert screws, 2.7 and 3.5-millimeter cortical-bone screws, and 4.0-millimeter cancellous-bone screws with regard to bending strength, tensile strength, and torsion stability. The test was not cyclical in nature, and the experimental models were made from ashwood. Carter et al. found little difference between the resistance to bending failure of a Herbert screw and that of a 3.5-millimeter cannulated screw used to repair a transverse osteotomy at the waists of cadaveric scaphoids. Repetitive cyclical bending was selected as the mode of testing for the present study because non-unions of the scaphoid often collapse into a flexed position. Increasing the load with increasing cycles is an engineering technique used to measure the strength of materials2,6,23. It is particularly useful when there is substantial variability among specimens, as was the case in the present study. For example, the amount of bending force required to cause failure of fixation with the Herbert screw within a reasonable amount of time would require so much time to cause failure of a contralateral scaphoid fixed with an AO 3.5-millimeter cannulated screw the experiment would be impractical to perform. Because differences in the number of cycles required to produce a certain displacement are also associated with a greater force to reach that number of cycles, significant differences between matched pairs, as in this testing situation, should be associated with clinically meaningful differences. Although it would have been ideal to perform this test on fresh specimens, it was not possible for our institution to obtain a sufficient number of matched, unembalmed, fresh scaphoids. However, we do not believe that testing the fixation devices in preserved cadaveric specimens was a problem because comparisons were made in matched pairs that had identical alterations in the biomechanical properties due to the preservation technique. We also decided that it was critical to use actual scaphoids, rather than models made of wood or Styrofoam (polystyrene) because the geometry of the scaphoid is difficult to simulate and the geometry itself leads to the typical dorsal angulation pattern of failure. Numerous authors have studied the alterations in the mechanical properties of bone caused by embalming, and most have found little alteration other than that of energy to failure if the specimens are kept moist10,14,18. The different screws, by their design, tolerated different amounts of tightening torque to produce the desired compression. Additionally, the pairs of scaphoids differed in their ability to resist insertion torque before stripping. Therefore, all screws were inserted to so-called fingertip tightness, as is done clinically. Our results corresponded to those of the biomechanical studies by Rankin et al. and by Shaw19,20, all of which showed superior fixation by the AO 3.5-millimeter cannulated screw compared with the Herbert screw. This was expected because compression results in interdigitation between the osseous fragments and prevents displacement with cyclical bending loads. The differences between these devices were highly significant. The Herbert-Whipple screw seemed to provide more satisfactory fixation than the standard Herbert screw did, primarily because of its size. However, because of its size, insertion may be more difficult. The Acutrak screw provided somewhat superior fixation compared with the Herbert screw. Because it is totally threaded, there is a greater surface area for fixation between the bone and the screw. Its conical shape also may be an advantage with regard to avoiding pistoning within the scaphoid. The Universal Compression screw seems to have a potential for providing fixation with considerable compression. In two of the six pairs in the present study, a fracture occurred at the distal pole of the scaphoid. This resulted in a rapid loss of fixation in the ramped intensity cyclical bending test. It is possible that unembalmed bone or specimens from younger individuals would not have fractured. The potential for fracture, however, seems to be considerable with this device. The Herbert screw, the Herbert-Whipple screw, and the Acutrak screw share the advantage of a headless design. Additionally, the Acutrak screw and the Herbert-Whipple screw are cannulated for ease of placement. Volar comminution puts fixation devices in the scaphoid to their most severe test. Both the Herbert screw and the AO screw showed significantly decreased fixation capability with a small amount of volar bone removed from the scaphoid. It seems appropriate, in these settings, to consider use of the AO 3.5-millimeter cannulated screw or another screw that provides superior fixation, despite the inconvenience of the protruding head or the large diameter. Similarly, when an operation is to involve opening-wedge bone-grafting, consideration might be given to the AO 3.5-millimeter cannulated screw, Herbert-Whipple screw, or Acutrak screw rather than to the Herbert screw. The selection of a device for fixation of a fracture or non-union of the scaphoid involves numerous factors, such as ease of insertion, prominence of the hardware, and quality of the fixation. The results of the present study, which dealt with cyclical bending loads, should be considered when such a selection is made.
*Although none of the authors have received or will 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 non-profit organization with which one or more of the authors are associated. No funds were received in support of this study.
  CiteULike  Connotea  Del.icio.us  Technorati What's this?
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
| home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |