- Email: [email protected]
A biomechanical analysis of triangulation of anterior vertebral double-screw fixation
Clinical Biomechanics 18 (2003) S40–S45 www.elsevier.com/locate/clinbiomech
A biomechanical analysis of triangulation of anterior vertebral double-screw fixation Tsung-Jen Huang a
a,*
, Robert Wen-Wei Hsu a, Ching-Lung Tai
a,b
, Weng-Pin Chen
b
Biomechanical Laboratory, Department of Orthopaedic Surgery, Chang Gung Memorial Hospital at Chia-Yi, College of Medicine, Chang Gung University, No. 6, West Section, Chia Pu Road, Putz City, Chia-Yi 613, Taiwan, ROC b Department of Biomedical Engineering, Chung-Yuan Christian University, Chung-Li 320, Taiwan, ROC
Abstract Objective. This study tested the hypothesis that triangulation of two anterior vertebral screws without penetration of the cortex offers more resistance to pullout than two screws placed in parallel and penetrated. Design. The pullout strength for two parallel or two triangulated anterior vertebral screws fixation, with a uni-cortical or bicortical purchase, were tested and compared to the strength of a single-screw fixation with a bi-cortical purchase. Four porcine spines (six months old) were used for biomechanical test and bone mineral density was measured for each specimen before testing. Background. The potential hazards from penetration by anterior vertebral cortex screws including neurovascular and organs injuries are well documented. However, bi-cortical screw penetration is widely recognized as necessary for good anterior spinal stabilization. The authors are not aware of any biomechanical study on the anterior placement of triangulated vertebral screws without penetration and its effect on the fixation strength of anterior vertebral device remains unclear. Methods. In this study five modes of screw fixations in lateral vertebral bodies were performed: Group A, triangulated screws with one screw penetration; Group B, triangulated screws without penetration; Group C, parallel penetrating screws; Group D, parallel nonpenetrating screws; and Group E, a single-screw with bi-cortical purchase. Biomechanical analysis with a material testing system machine was performed to determine the pull out strength of each configuration. Results. The results showed that the pullout strength in the various double-screw fixation modes were statistically increased as compared to that of the single-screw with bi-cortical purchase mode. There existed statistical differences ðP < 0:05Þ between Groups A and B, Groups C and D and Groups D and E, respectively. However, no significant difference was found between Groups B and C ðP ¼ 0:144Þ. Conclusions. Based on the current data, triangulation of two anterior vertebral screws without penetration of the cortex (Group B) achieved pullout strengths similar to that of two-parallel double-cortical screws (Group C). The authors believe that this is an attractive alternative in anterior spinal instrumentation avoiding the potential risks of cortical penetration. However, in the event of pullout failure, the triangulation configuration will produce a more disastrous consequence. Relevance Triangulation of two anterior vertebral screws without penetration of the cortex achieve pullout strengths similar to that of twoparallel double-cortical screws. This is an attractive alternative in anterior spinal instrumentation that avoids the potential risks of cortical penetration. Ó 2003 Elsevier Science Ltd. All rights reserved. Keywords: Vertebral screw; Triangulation; Biomechanical test; Bi-cortical; Uni-cortical; Anterior spinal device
1. Introduction Anterior spinal instrumentation of the thoracic and lumbar spine has gained in popularity since Dwyer et al. developed the cable system in 1964 (Dwyer, 1970). However, implant loosening, screw breakage and cut-
*
Corresponding author. E-mail address: [email protected] (T.-J. Huang).
throughs have been noted clinically when using Dwyer or ZielkeÕs ventrale derotations spondylodese (VDS) instrumentations (Bauer et al., 1986; Breeze et al., 1988; Lieberman et al., 1998; Ogon et al., 1996; Yuan et al., 1988). To avoid these complications and pseudarthrosis, one may use an anterior vertebral screw to penetrate the second cortex, or even a double-screw instrumentation using two parallel and penetrated screws to provide a stronger fixation (Dunn, 1984; Kostuik et al., 1986; Kostuik, 1988). Bi-cortical screw purchase of vertebral
0268-0033/03/$ - see front matter Ó 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0268-0033(03)00083-4
T.-J. Huang et al. / Clinical Biomechanics 18 (2003) S40–S45
body in anterior spinal fixation is well recognized to achieve optimal stabilization of implants (Horton et al., 1996; Snyder et al., 1995). Kaneda et al. (1984) advocate that an anterior double-screw, double-rod system may offer optimal strength, removing the need for a postoperative brace. However, cortical penetration by vertebral screws is not without risks. A review of the literature reveals that penetrated screw tips can cause serious complications (Hsieh et al., 1999; Matsuzaki et al., 1993). There are potential risks to neurovascular structures, internal organs and the diaphragm itself. In the current study, five different anterior vertebral screw methods were studied namely triangulated screws with one screw penetration (Group A); triangulated screws without penetration (Group B), parallel penetrating screws (Group C); parallel nonpenetrating screws (Group D); and a singlescrew with bi-cortical purchase (Group E) (Fig. 1). This study tested the hypothesis that two triangulated anterior vertebral screws without penetration offers more resistance to pullout than two screws placed in parallel and penetrated.
2. Methods Thirty-one (31) fresh low thoracic and lumbar vertebrae taken from four porcine spines (six months old) were used to perform the biomechanical pullout test, including both single-screw and double-screw anterior vertebral placements in this current study. The specimens were stored at )20 °C until the day of testing. After thawing for 24 h, individual specimens were dissected free of soft tissue; dual-energy X-ray absorptiometry (DEXA), (QDR-2000, Hologic Inc., Boston, Mass, USA) was used to examine the variance in bone mineral density (BMD) for each individual specimen.
S41
The mean BMD ranged from 1.101 (SD 0.084) g/cm2 to 1.379 (SD 0.097) g/cm2 . There is no significant difference in bone density among the groups ðP > 0:05Þ. Each specimen was then prepared for screw placement with Trifix vertebral screws––6.2 mm in diameter and 45 mm in length (Trifix, San Andrio, CA, USA)––as shown in Fig. 2A. A 4 mm drill was used to create a hole located at midway between the end plates as determined by radiological examination. This hole was located close to the posterior vertebral cortex. Single-screw with bi-cortical purchase and double-screw with four modes of fixation (Fig. 1A–E) were evenly distributed and allocated in sequence from the 11th thoracic vertebrae to the sixth lumbar vertebrae of the four porcine spines. In total, there were 31 vertebrae used for the biomechanical test. Twenty-four (24) vertebrae were assigned into four groups (six in each) to undergo different double-screw placements as follows: Group A, two triangulated screws (a 22° intersection angle) with one screw penetration; Group B, triangulated screws (a 22° intersection angle) without penetration; Group C, parallel penetrating screws; Group D, parallel nonpenetrating screws. The other seven vertebrae underwent single-screw penetration placement by insertion of the screw in such a way that two complete threads protruded from the distal cortex (Group E, Fig. 1E). The vertebrae allocation, fixation configuration, and specimen number for five different anterior vertebral screw modes was listed in Table 1. The implantation of the fixation screws was carried out under direct visualization radiologically. Radiological examinations were done at the same time to check for the implanted screwsÕ depth and position in all specimens. An angled jig device with a 22° angle was used to ensure the accuracy of the placement of the fixation screws. The depth of the second screw in Groups A and B
Fig. 1. Illustrations showing the four modes of anterior vertebral double-screw placements and a single-screw with bi-cortical purchase: (A) Group A, triangulated screws (a 22° intersection angle) with one screw penetration; (B) Group B, triangulated screws (a 22° intersection angle) without penetration; (C) Group C, parallel penetrating screws; (D) Group D, parallel nonpenetrating screws; and (E) Group E, a single-screw with bi-cortical purchase.
S42
T.-J. Huang et al. / Clinical Biomechanics 18 (2003) S40–S45
anterior vertebral plate. Each specimen was tested for failure in axial pullout using the material testing system (Bionix 858, MTS Corp., MN, USA) (Fig. 2B). The specimen was placed on a specially designed universal fixture with a self-aligned function, clamped on the upper side of the material testing systemÕs wedge grip (Fig. 2B). The heads of the screws were fixed in a 10 mm diameter cylindrical rod and clamped at the lower side of the material testing systemÕs wedge grip using the outer thread of the screw itself (Fig. 2B). After the specimens were mounted, the pullout force was applied at a constant loading rate of 20 N/s. The relation between force and displacement was recorded at an increment of 50 N by the material testing system Teststar-II software. To evaluate the effect of different modes of screw implantation on the stability of spinal fixation, the magnitude of force at failure for each individual specimen was selected for comparison by using two-way analysis of variance (A N O V A ) for statistical analysis. A significant difference was reported at P < 0:05. 3. Results
Fig. 2. (A) The implanted vertebral screw with a screw diameter of 6.2 mm and 45 mm in length––the outer thread of the screw at its upper end was clamped by a 10 mm diametric cylindrical rod; (B) Experimental pullout test for triangulated implantation. The pullout force was transferred to the specimen by a universal device with a selfaligning function above the upper grip. The intersection angle between the two implanted screws in the vertebral body was 22°.
was half the length of the initial one and was inserted without touching each other. In the current study, the authorsÕ choice of an intersection angle of 22° for the triangulated screw placement groups was based on the screw designs of the anterior locking plate system (ALPS, Trifix) (Fig. 2A). In this design, placement of the long locking screw with the short sag screw is at an angle of 22° and locked in the
The typical diagram of force and displacement versus time for the pullout test is shown as Fig. 3A. The corresponding diagram of force versus displacement is shown as Fig. 3B. The force–time curve characterizes the pullout force as it increase at a constant rate of 20 N/ s. The mean maximum pull-out forces for Groups A to E were 3283 (SD 553) N, 2717 (SD 250) N, 2450 (SD 327) N, 2067 (SD 211) N and 1700 (SD 490) N, respectively (Fig. 4). The current results revealed that there existed statistical differences ðP < 0:05Þ between Groups A and B, Groups C and D and Groups D and E, respectively. However, no significant difference was found between Groups B and C ðP ¼ 0:144Þ. 4. Discussion There are many factors that affect the fixation strength of an anterior vertebral device. These include BMD (Breeze et al., 1988; Snyder et al., 1995; Gilbert et al., 1993; Halvorson et al., 1994; Lim et al., 1995; Okuyama et al., 1993) and fixation device related issues,
Table 1 The vertebrae allocation, fixation configuration, and specimen number for five different anterior vertebral screw modes Vertebrae allocation
Fixation configuration
Group A
1 T11, L1 to T5 (one in each)
Group B
1 T12, L1 to L5 (one in each)
Group C Group D Group E
1 T11, L1 to L5 (one in each) 1 T12, L1 to L5 (one in each) 2 T11, 2 T12, 3 L6
Two triangulated screws (a 22° intersection angle) with one screw penetration Two triangulated screws (a 22° intersection angle) without penetration Two parallel penetrating screws Two parallel nonpenetrating screws One single-screw with bi-cortical purchase
Specimen number 6 6 6 6 7
T.-J. Huang et al. / Clinical Biomechanics 18 (2003) S40–S45
S43
Fig. 3. (A) The typical diagram of force (N) and displacement (mm) versus time (s) for pullout test. The force–time curve demonstrates how the pullout force increases at a constant rate of 20 N/s. (B) The typical diagram of force versus displacement for pullout test.
Fig. 4. The pullout forces of the single-screw with bi-cortical purchase and double-screw with four modes of fixation: Group A, two triangulated (a 22° intersection angle) and one screw penetration; Group B, two triangulated (a 22° intersection angle) and no screw penetration; Group C, two parallel and two-screw penetration; and Group D, two parallel without any screw penetration.
such as the necessity of double-cortex penetration (Lieberman et al., 1998; Horton et al., 1996), single- or double-screws fixation (Dunn, 1984; Kostuik et al., 1986; Kaneda et al., 1984), special screw construction with nuts (Lieberman et al., 1998) or staples (Snyder et al., 1995), and specific triangulation modes for double-screws (Ogon et al., 1996). Pullout strength for vertebral screws is highly correlated to BMD (Gilbert et al., 1993; Lim et al., 1995; Okuyama et al., 1993). However, Krag et al. (1986) noted that the pitch and tooth patterns of the screws were not significant factors in fixation strength. A bi-cortical penetration is mandatory for single-screw fixations is widely accepted (Horton et al., 1996). Clinically, screw pullout is a common problem in anterior fixation, regardless of the implant system, particularly during a lateral flexion moment at the proximal and distal ends of a multi-level construct (Spiegel et al.,
S44
T.-J. Huang et al. / Clinical Biomechanics 18 (2003) S40–S45
2000). Breeze et al. (1988) noted that bi-cortical screws were significantly stronger in resisting pullout than unicortical screws, and the engaged second cortex increases 25–44% in pullout strength depending upon BMD. Lieberman et al. (1998) noted that a large-diameter cancellous thread pattern, such as Zielke and Kaneda screws, could improve the pullout strength from a vertebral body, and that a bi-cortical purchase improved pullout strength as well. To lessen the pseudarthrosis rate in anterior spinal fixation, Kostuik et al. (1986) suggested the placement of two parallel Zielke screws. The double-screw design was also seen in Dunn and Kaneda devices (Dunn, 1984; Kaneda et al., 1984). Kaneda et al. (1984) found that the double penetrated screws and double-rod system can offer sufficient device strength, allowing patients to be mobile postoperatively without the need for braces. Thus far, triangulation of two anterior vertebral screws is not yet widely accepted (Ogon et al., 1996). Several studies have investigated the pullout strength or resistance against bending of the posterior pedicle screw instrumentation. Krag et al. (1986) postulated that angulated placement of pedicle screws increases the pullout strength of posterior spinal devices. Their theory was later verified by a biomechanical test performed by Ruland et al. (1991). Ruland et al. (1991) declared that the fixation strength of triangulation for two pedicle screws is determined not only by the amount of bone within the pedicle screw threads, but also by the area of bone within the trapezoid formed by the screws. This concept was then applied to anterior vertebral fixation devices. In this current study, the typical outcome of specimens after pullout tests for two-parallel screws with penetration (Group C) and triangulated screws without penetration (Group C) are shown in Fig. 5A and B, respectively. In triangulated vertebral screw placements, the area within the trapezoid (Fig. 5C) is considered to contribute to the increase of pullout strength, which might cause a severe damage after testing (Fig. 5B). This phenomenon was also found in pullout testing for posterior pediculate screws (Spiegel et al., 2000). Ogon et al. (1996) noted that the instrumentation of Cotrel– Dubousset–Hopf screws on condition of triangulated fixation with double-screw penetration would lead to a 73% increase in pullout strength as compared to that of ZielkeÕs VDS single-screw with bi-cortical instrumentation. This leads the authors to suggest that in anterior spinal surgery, fixation of the vertebral device interface can be improved considerably by applying two triangulated screws. However, Ogon et al. did not analyze the effects of triangulated screws without penetration in their study. Gilbert et al. (1993) noted that while pedicle screws failed by pulling out of bone, the anterior vertebral screws failed by cut-throughs and pullout of the vertebral body due to the complex forces they resist. Screw pullouts are widely accepted as a mode of clinical
Fig. 5. Photographs showing specimens after the biomechanical pullout test. (A) Two parallel screws with penetration (in Group C). (B) Two triangulated screws without penetration (in Group B). A severe vertebral damage occurs in the triangulated mode was noted after testing. (C) The area of bone within the trapezoid formed by the two vertebral screws might contribute to the increase of pullout strength.
failure. Screw failures may result from a variety of factors, including repeated cyclic loading of the screws in multiple planes or the biological response of bone to the screws overtime (Spiegel et al., 2000). Breeze et al. (1988) noted that screw failure was evaluated using axial pullout test because this method is simple and extremely reproducible. Spiegel et al. (2000) noted that the force acting on the screws may cause implant loosening owing to the progressive toggling following microfractures in the surrounding trabecular bone at the end vertebral levels, which correlates with clinical observation. In anterior vertebral instrumentation, bi-cortical screw purchase shows better fixation strength and is always recommended clinically (Lieberman et al., 1998; Snyder et al., 1995). However, bi-cortical screw penetration technique is not without any risks. Previous reports in literature noted that there had been injuries to great vessels or the diaphragm causing hemothorax, or great vessel tear (Hsieh et al., 1999; Matsuzaki et al., 1993). Biomechanically, fixation strength of single-screw with bi-cortical fixation is better than uni-cortical fixation. The newly designed anterior implant with double-screw fixation at each vertebra can decrease pseudarthrosis and the possibility of implant failure (Kostuik et al., 1986; Kaneda et al., 1984). However, the risks of visceral or vessel injuries might increase accordingly. Technically, two parallel vertebral screws with bi-cortical fixation are recommended (Dunn, 1984; Kostuik et al., 1986; Horton et al., 1996; Kaneda et al., 1984). In the current biomechanical study, the triangulation mode of two vertebral screws was located subcortically (Fig. 1B). In Group B, although both triangulated screws did not penetrate the cortex, they could achieve pullout strength not less than two parallel, double-screws penetration
T.-J. Huang et al. / Clinical Biomechanics 18 (2003) S40–S45
mode (Group C) (2717 (SD 250) N versus 2450 (SD 327) N, P ¼ 0:144). The authors believe that this is an attractive alternative in anterior spinal instrumentation avoiding the potential risks of cortical penetration. The normal spine is able to sustain multi-directional physiological forces resulting from flexion, extension, rotation, shearing, and lateral bending movements. Clinically, the most common failure mode of the anterior vertebral implants is lateral migration of screws and plates. To the authorsÕ understanding, the lateral bending moment of the normal spine may induce pullout force acting on the fixation screws. In the current study, the failure mode of the implants that resulted from pullout force was quite similar to that of anterior vertebral implants that failed due to lateral migration. It would be logical to assume that implants with high pullout force would be more stable clinically. It is interesting to note that the pullout strength for Group C was only 1.44 times that of Group E. Theoretically, it should be about twice as strong. The results revealed that the average pullout strength of Groups C and E were 2450 (SD 327) N and 1700 (SD 490) N, respectively. The data showed a substantial amount of variation in both groups. An important factor that may lead to the difference could be the entrance point of the screw placement. In Group E, the screw was placed at the diametric line of vertebral body, which resulted in the maximum contact area between screw and bone. While in Group C, the entrances of the parallel screws were placed one anteriorly and the other posteriorly to the diametric line of vertebral body. The total contact area of the two parallel screws and bone would be less than twice that of Group E. Although data from the current study have shown the Group B––trianguration configuration achieved similar pullout strength to Group C––parallel configuration, in the event of pullout failure, the trianguration configuration would produce a more disastrous consequence. Theoretically, establishing a mathematical model will be a useful tool to predict the pullout strength of any configuration of screw insertions. The experimental results of the current study could be used to validate this model. The model could then predict any configuration of screw insertion. However, this work is not done in the current study, an alternative finite element analysis is now being carried out to simulate all the configurations of screw instrumentations described in the current study.
Acknowledgement The authors would like to thank the National Science Council of the Republic of China for the financially supporting this research under contract no. NSC 892320-B-182-018.
S45
References Bauer, R., Mostegel, A., Eichenauer, M., 1986. An analysis of the results of Dwyer and Zielke instrumentations in the treatment of scoliosis. Arch. Orthop. Trauma Surg. 105, 302–309. Breeze, S.W., Doherty, B.J., Noble, P.S., LeBlanc, A., Heggeness, M.H., 1988. A biomechanical study of anterior thoracolumbar screw fixation. Spine 23, 1829–1831. Dunn, H.K., 1984. Anterior stabilization of thoracolumbar injuries. Clin. Orthop. 189, 116–124. Dwyer, A.F., 1970. Anterior instrumentation in scoliosis. J. Bone Joint Surg. 52B, 782–787. Gilbert, S.G., Johns, P.C., Chow, D.C., Black, R.C., 1993. Relationship of vertebral bone screw axial pullout strength to quantitative tomographic trabecular bone mineral content. J. Spinal Disord. 6, 513–521. Halvorson, T.L., Kelley, L.A., Thomas, K.A., Whitecloud III, T.S., Cook, S.D., 1994. Effects of bone mineral density on pedicle screw fixation. Spine 19, 2415–2420. Horton, W.C., Blackstock, S.F., Norman, J.T., 1996. Strength of fixation of the anterior vertebra body screw. Spine 21, 439–444. Hsieh, P.H., Chen, W.J., Chen, L.H., Niu, C.C., 1999. An unusual complication of anterior spinal instrumentation: hemothorax contralateral to the side of the incision. A case report. J. Bone Joint Surg. 81A, 998–1001. Kaneda, K., Abumi, K., Fujiya, M., 1984. Burst fractures with neurological deficits of the thoracolumbar spine: results of the anterior decompression and stabilization with anterior instrumentation. Spine 9, 788–795. Kostuik, J.P., 1988. Anterior fixation of burst fractures of the thoracic and lumbar spine with or without neurological involvement. Spine 13, 286–293. Kostuik, J.P., Errico, T.J., Gleason, T.F., 1986. Techniques of internal fixation for degenerative conditions of the lumbar spine. Clin. Orthop. 203, 219–231. Krag, M.H., Beynnon, B.D., Pope, M.H., 1986. An internal fixator for posterior application to short segments of the thoracic, lumbar, or lumbosacral spine: design and testing. Clin. Orthop. 203, 75–98. Lieberman, I.H., Khazim, R., Woodside, T., 1998. Anterior vertebral body screw pullout testing. A comparison of Zielke, Kaneda, universal spine system, and universal spine system with pulloutresistant nut. Spine 23, 908–910. Lim, T.H., An, H.S., Evanich, C., 1995. Strength of anterior vertebral screw fixation in relation to bone mineral density. J. Spinal Disord. 8, 121–125. Matsuzaki, H., Tokuhashi, Y., Wakabayashi, K., Kitamura, S., 1993. Penetration of a screw into the thoracic aorta in anterior spinal instrumentation. A case report. Spine 18, 2327–2331. Ogon, M., Haid, C., Krismer, M., Sterzinger, W., Bauer, R., 1996. Comparison between single-screw and triangulated, double-screw fixation in anterior spine surgery. A biomechanical test. Spine 21, 2728–2734. Okuyama, K., Sato, K., Abe, E., 1993. Stability of transpedicle screwing for the osteoporotic spine: An in vitro study of mechanical stability. Spine 18, 2240–2245. Ruland, C.M., McAfee, P.C., Wareden, K.E., Cunningham, B.W., 1991. Triangulation of pedicular instrumentation: a biomechanical analysis. Spine 16, 270–276. Snyder, B.D., Zaltz, I., Hall, J.E., Emans, J.B., 1995. Predicting the integrity of vertebral bone screw fixation in anterior spinal instrumentation. Spine 20, 1568–1574. Spiegel, D.A., Cunningham, B.W., Oda, I., 2000. Anterior vertebral screw strain with and without solid interspace support. Spine 25, 2755–2761. Yuan, H.A., Mann, K.A., Found, E.M., 1988. Early clinical experience with the Syracuse I-plate: an anterior spinal fixation device. Spine 23, 278–285.
A biomechanical analysis of triangulation of anterior vertebral double-screw fixation Tsung-Jen Huang a
a,*
, Robert Wen-Wei Hsu a, Ching-Lung Tai
a,b
, Weng-Pin Chen
b
Biomechanical Laboratory, Department of Orthopaedic Surgery, Chang Gung Memorial Hospital at Chia-Yi, College of Medicine, Chang Gung University, No. 6, West Section, Chia Pu Road, Putz City, Chia-Yi 613, Taiwan, ROC b Department of Biomedical Engineering, Chung-Yuan Christian University, Chung-Li 320, Taiwan, ROC
Abstract Objective. This study tested the hypothesis that triangulation of two anterior vertebral screws without penetration of the cortex offers more resistance to pullout than two screws placed in parallel and penetrated. Design. The pullout strength for two parallel or two triangulated anterior vertebral screws fixation, with a uni-cortical or bicortical purchase, were tested and compared to the strength of a single-screw fixation with a bi-cortical purchase. Four porcine spines (six months old) were used for biomechanical test and bone mineral density was measured for each specimen before testing. Background. The potential hazards from penetration by anterior vertebral cortex screws including neurovascular and organs injuries are well documented. However, bi-cortical screw penetration is widely recognized as necessary for good anterior spinal stabilization. The authors are not aware of any biomechanical study on the anterior placement of triangulated vertebral screws without penetration and its effect on the fixation strength of anterior vertebral device remains unclear. Methods. In this study five modes of screw fixations in lateral vertebral bodies were performed: Group A, triangulated screws with one screw penetration; Group B, triangulated screws without penetration; Group C, parallel penetrating screws; Group D, parallel nonpenetrating screws; and Group E, a single-screw with bi-cortical purchase. Biomechanical analysis with a material testing system machine was performed to determine the pull out strength of each configuration. Results. The results showed that the pullout strength in the various double-screw fixation modes were statistically increased as compared to that of the single-screw with bi-cortical purchase mode. There existed statistical differences ðP < 0:05Þ between Groups A and B, Groups C and D and Groups D and E, respectively. However, no significant difference was found between Groups B and C ðP ¼ 0:144Þ. Conclusions. Based on the current data, triangulation of two anterior vertebral screws without penetration of the cortex (Group B) achieved pullout strengths similar to that of two-parallel double-cortical screws (Group C). The authors believe that this is an attractive alternative in anterior spinal instrumentation avoiding the potential risks of cortical penetration. However, in the event of pullout failure, the triangulation configuration will produce a more disastrous consequence. Relevance Triangulation of two anterior vertebral screws without penetration of the cortex achieve pullout strengths similar to that of twoparallel double-cortical screws. This is an attractive alternative in anterior spinal instrumentation that avoids the potential risks of cortical penetration. Ó 2003 Elsevier Science Ltd. All rights reserved. Keywords: Vertebral screw; Triangulation; Biomechanical test; Bi-cortical; Uni-cortical; Anterior spinal device
1. Introduction Anterior spinal instrumentation of the thoracic and lumbar spine has gained in popularity since Dwyer et al. developed the cable system in 1964 (Dwyer, 1970). However, implant loosening, screw breakage and cut-
*
Corresponding author. E-mail address: [email protected] (T.-J. Huang).
throughs have been noted clinically when using Dwyer or ZielkeÕs ventrale derotations spondylodese (VDS) instrumentations (Bauer et al., 1986; Breeze et al., 1988; Lieberman et al., 1998; Ogon et al., 1996; Yuan et al., 1988). To avoid these complications and pseudarthrosis, one may use an anterior vertebral screw to penetrate the second cortex, or even a double-screw instrumentation using two parallel and penetrated screws to provide a stronger fixation (Dunn, 1984; Kostuik et al., 1986; Kostuik, 1988). Bi-cortical screw purchase of vertebral
0268-0033/03/$ - see front matter Ó 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0268-0033(03)00083-4
T.-J. Huang et al. / Clinical Biomechanics 18 (2003) S40–S45
body in anterior spinal fixation is well recognized to achieve optimal stabilization of implants (Horton et al., 1996; Snyder et al., 1995). Kaneda et al. (1984) advocate that an anterior double-screw, double-rod system may offer optimal strength, removing the need for a postoperative brace. However, cortical penetration by vertebral screws is not without risks. A review of the literature reveals that penetrated screw tips can cause serious complications (Hsieh et al., 1999; Matsuzaki et al., 1993). There are potential risks to neurovascular structures, internal organs and the diaphragm itself. In the current study, five different anterior vertebral screw methods were studied namely triangulated screws with one screw penetration (Group A); triangulated screws without penetration (Group B), parallel penetrating screws (Group C); parallel nonpenetrating screws (Group D); and a singlescrew with bi-cortical purchase (Group E) (Fig. 1). This study tested the hypothesis that two triangulated anterior vertebral screws without penetration offers more resistance to pullout than two screws placed in parallel and penetrated.
2. Methods Thirty-one (31) fresh low thoracic and lumbar vertebrae taken from four porcine spines (six months old) were used to perform the biomechanical pullout test, including both single-screw and double-screw anterior vertebral placements in this current study. The specimens were stored at )20 °C until the day of testing. After thawing for 24 h, individual specimens were dissected free of soft tissue; dual-energy X-ray absorptiometry (DEXA), (QDR-2000, Hologic Inc., Boston, Mass, USA) was used to examine the variance in bone mineral density (BMD) for each individual specimen.
S41
The mean BMD ranged from 1.101 (SD 0.084) g/cm2 to 1.379 (SD 0.097) g/cm2 . There is no significant difference in bone density among the groups ðP > 0:05Þ. Each specimen was then prepared for screw placement with Trifix vertebral screws––6.2 mm in diameter and 45 mm in length (Trifix, San Andrio, CA, USA)––as shown in Fig. 2A. A 4 mm drill was used to create a hole located at midway between the end plates as determined by radiological examination. This hole was located close to the posterior vertebral cortex. Single-screw with bi-cortical purchase and double-screw with four modes of fixation (Fig. 1A–E) were evenly distributed and allocated in sequence from the 11th thoracic vertebrae to the sixth lumbar vertebrae of the four porcine spines. In total, there were 31 vertebrae used for the biomechanical test. Twenty-four (24) vertebrae were assigned into four groups (six in each) to undergo different double-screw placements as follows: Group A, two triangulated screws (a 22° intersection angle) with one screw penetration; Group B, triangulated screws (a 22° intersection angle) without penetration; Group C, parallel penetrating screws; Group D, parallel nonpenetrating screws. The other seven vertebrae underwent single-screw penetration placement by insertion of the screw in such a way that two complete threads protruded from the distal cortex (Group E, Fig. 1E). The vertebrae allocation, fixation configuration, and specimen number for five different anterior vertebral screw modes was listed in Table 1. The implantation of the fixation screws was carried out under direct visualization radiologically. Radiological examinations were done at the same time to check for the implanted screwsÕ depth and position in all specimens. An angled jig device with a 22° angle was used to ensure the accuracy of the placement of the fixation screws. The depth of the second screw in Groups A and B
Fig. 1. Illustrations showing the four modes of anterior vertebral double-screw placements and a single-screw with bi-cortical purchase: (A) Group A, triangulated screws (a 22° intersection angle) with one screw penetration; (B) Group B, triangulated screws (a 22° intersection angle) without penetration; (C) Group C, parallel penetrating screws; (D) Group D, parallel nonpenetrating screws; and (E) Group E, a single-screw with bi-cortical purchase.
S42
T.-J. Huang et al. / Clinical Biomechanics 18 (2003) S40–S45
anterior vertebral plate. Each specimen was tested for failure in axial pullout using the material testing system (Bionix 858, MTS Corp., MN, USA) (Fig. 2B). The specimen was placed on a specially designed universal fixture with a self-aligned function, clamped on the upper side of the material testing systemÕs wedge grip (Fig. 2B). The heads of the screws were fixed in a 10 mm diameter cylindrical rod and clamped at the lower side of the material testing systemÕs wedge grip using the outer thread of the screw itself (Fig. 2B). After the specimens were mounted, the pullout force was applied at a constant loading rate of 20 N/s. The relation between force and displacement was recorded at an increment of 50 N by the material testing system Teststar-II software. To evaluate the effect of different modes of screw implantation on the stability of spinal fixation, the magnitude of force at failure for each individual specimen was selected for comparison by using two-way analysis of variance (A N O V A ) for statistical analysis. A significant difference was reported at P < 0:05. 3. Results
Fig. 2. (A) The implanted vertebral screw with a screw diameter of 6.2 mm and 45 mm in length––the outer thread of the screw at its upper end was clamped by a 10 mm diametric cylindrical rod; (B) Experimental pullout test for triangulated implantation. The pullout force was transferred to the specimen by a universal device with a selfaligning function above the upper grip. The intersection angle between the two implanted screws in the vertebral body was 22°.
was half the length of the initial one and was inserted without touching each other. In the current study, the authorsÕ choice of an intersection angle of 22° for the triangulated screw placement groups was based on the screw designs of the anterior locking plate system (ALPS, Trifix) (Fig. 2A). In this design, placement of the long locking screw with the short sag screw is at an angle of 22° and locked in the
The typical diagram of force and displacement versus time for the pullout test is shown as Fig. 3A. The corresponding diagram of force versus displacement is shown as Fig. 3B. The force–time curve characterizes the pullout force as it increase at a constant rate of 20 N/ s. The mean maximum pull-out forces for Groups A to E were 3283 (SD 553) N, 2717 (SD 250) N, 2450 (SD 327) N, 2067 (SD 211) N and 1700 (SD 490) N, respectively (Fig. 4). The current results revealed that there existed statistical differences ðP < 0:05Þ between Groups A and B, Groups C and D and Groups D and E, respectively. However, no significant difference was found between Groups B and C ðP ¼ 0:144Þ. 4. Discussion There are many factors that affect the fixation strength of an anterior vertebral device. These include BMD (Breeze et al., 1988; Snyder et al., 1995; Gilbert et al., 1993; Halvorson et al., 1994; Lim et al., 1995; Okuyama et al., 1993) and fixation device related issues,
Table 1 The vertebrae allocation, fixation configuration, and specimen number for five different anterior vertebral screw modes Vertebrae allocation
Fixation configuration
Group A
1 T11, L1 to T5 (one in each)
Group B
1 T12, L1 to L5 (one in each)
Group C Group D Group E
1 T11, L1 to L5 (one in each) 1 T12, L1 to L5 (one in each) 2 T11, 2 T12, 3 L6
Two triangulated screws (a 22° intersection angle) with one screw penetration Two triangulated screws (a 22° intersection angle) without penetration Two parallel penetrating screws Two parallel nonpenetrating screws One single-screw with bi-cortical purchase
Specimen number 6 6 6 6 7
T.-J. Huang et al. / Clinical Biomechanics 18 (2003) S40–S45
S43
Fig. 3. (A) The typical diagram of force (N) and displacement (mm) versus time (s) for pullout test. The force–time curve demonstrates how the pullout force increases at a constant rate of 20 N/s. (B) The typical diagram of force versus displacement for pullout test.
Fig. 4. The pullout forces of the single-screw with bi-cortical purchase and double-screw with four modes of fixation: Group A, two triangulated (a 22° intersection angle) and one screw penetration; Group B, two triangulated (a 22° intersection angle) and no screw penetration; Group C, two parallel and two-screw penetration; and Group D, two parallel without any screw penetration.
such as the necessity of double-cortex penetration (Lieberman et al., 1998; Horton et al., 1996), single- or double-screws fixation (Dunn, 1984; Kostuik et al., 1986; Kaneda et al., 1984), special screw construction with nuts (Lieberman et al., 1998) or staples (Snyder et al., 1995), and specific triangulation modes for double-screws (Ogon et al., 1996). Pullout strength for vertebral screws is highly correlated to BMD (Gilbert et al., 1993; Lim et al., 1995; Okuyama et al., 1993). However, Krag et al. (1986) noted that the pitch and tooth patterns of the screws were not significant factors in fixation strength. A bi-cortical penetration is mandatory for single-screw fixations is widely accepted (Horton et al., 1996). Clinically, screw pullout is a common problem in anterior fixation, regardless of the implant system, particularly during a lateral flexion moment at the proximal and distal ends of a multi-level construct (Spiegel et al.,
S44
T.-J. Huang et al. / Clinical Biomechanics 18 (2003) S40–S45
2000). Breeze et al. (1988) noted that bi-cortical screws were significantly stronger in resisting pullout than unicortical screws, and the engaged second cortex increases 25–44% in pullout strength depending upon BMD. Lieberman et al. (1998) noted that a large-diameter cancellous thread pattern, such as Zielke and Kaneda screws, could improve the pullout strength from a vertebral body, and that a bi-cortical purchase improved pullout strength as well. To lessen the pseudarthrosis rate in anterior spinal fixation, Kostuik et al. (1986) suggested the placement of two parallel Zielke screws. The double-screw design was also seen in Dunn and Kaneda devices (Dunn, 1984; Kaneda et al., 1984). Kaneda et al. (1984) found that the double penetrated screws and double-rod system can offer sufficient device strength, allowing patients to be mobile postoperatively without the need for braces. Thus far, triangulation of two anterior vertebral screws is not yet widely accepted (Ogon et al., 1996). Several studies have investigated the pullout strength or resistance against bending of the posterior pedicle screw instrumentation. Krag et al. (1986) postulated that angulated placement of pedicle screws increases the pullout strength of posterior spinal devices. Their theory was later verified by a biomechanical test performed by Ruland et al. (1991). Ruland et al. (1991) declared that the fixation strength of triangulation for two pedicle screws is determined not only by the amount of bone within the pedicle screw threads, but also by the area of bone within the trapezoid formed by the screws. This concept was then applied to anterior vertebral fixation devices. In this current study, the typical outcome of specimens after pullout tests for two-parallel screws with penetration (Group C) and triangulated screws without penetration (Group C) are shown in Fig. 5A and B, respectively. In triangulated vertebral screw placements, the area within the trapezoid (Fig. 5C) is considered to contribute to the increase of pullout strength, which might cause a severe damage after testing (Fig. 5B). This phenomenon was also found in pullout testing for posterior pediculate screws (Spiegel et al., 2000). Ogon et al. (1996) noted that the instrumentation of Cotrel– Dubousset–Hopf screws on condition of triangulated fixation with double-screw penetration would lead to a 73% increase in pullout strength as compared to that of ZielkeÕs VDS single-screw with bi-cortical instrumentation. This leads the authors to suggest that in anterior spinal surgery, fixation of the vertebral device interface can be improved considerably by applying two triangulated screws. However, Ogon et al. did not analyze the effects of triangulated screws without penetration in their study. Gilbert et al. (1993) noted that while pedicle screws failed by pulling out of bone, the anterior vertebral screws failed by cut-throughs and pullout of the vertebral body due to the complex forces they resist. Screw pullouts are widely accepted as a mode of clinical
Fig. 5. Photographs showing specimens after the biomechanical pullout test. (A) Two parallel screws with penetration (in Group C). (B) Two triangulated screws without penetration (in Group B). A severe vertebral damage occurs in the triangulated mode was noted after testing. (C) The area of bone within the trapezoid formed by the two vertebral screws might contribute to the increase of pullout strength.
failure. Screw failures may result from a variety of factors, including repeated cyclic loading of the screws in multiple planes or the biological response of bone to the screws overtime (Spiegel et al., 2000). Breeze et al. (1988) noted that screw failure was evaluated using axial pullout test because this method is simple and extremely reproducible. Spiegel et al. (2000) noted that the force acting on the screws may cause implant loosening owing to the progressive toggling following microfractures in the surrounding trabecular bone at the end vertebral levels, which correlates with clinical observation. In anterior vertebral instrumentation, bi-cortical screw purchase shows better fixation strength and is always recommended clinically (Lieberman et al., 1998; Snyder et al., 1995). However, bi-cortical screw penetration technique is not without any risks. Previous reports in literature noted that there had been injuries to great vessels or the diaphragm causing hemothorax, or great vessel tear (Hsieh et al., 1999; Matsuzaki et al., 1993). Biomechanically, fixation strength of single-screw with bi-cortical fixation is better than uni-cortical fixation. The newly designed anterior implant with double-screw fixation at each vertebra can decrease pseudarthrosis and the possibility of implant failure (Kostuik et al., 1986; Kaneda et al., 1984). However, the risks of visceral or vessel injuries might increase accordingly. Technically, two parallel vertebral screws with bi-cortical fixation are recommended (Dunn, 1984; Kostuik et al., 1986; Horton et al., 1996; Kaneda et al., 1984). In the current biomechanical study, the triangulation mode of two vertebral screws was located subcortically (Fig. 1B). In Group B, although both triangulated screws did not penetrate the cortex, they could achieve pullout strength not less than two parallel, double-screws penetration
T.-J. Huang et al. / Clinical Biomechanics 18 (2003) S40–S45
mode (Group C) (2717 (SD 250) N versus 2450 (SD 327) N, P ¼ 0:144). The authors believe that this is an attractive alternative in anterior spinal instrumentation avoiding the potential risks of cortical penetration. The normal spine is able to sustain multi-directional physiological forces resulting from flexion, extension, rotation, shearing, and lateral bending movements. Clinically, the most common failure mode of the anterior vertebral implants is lateral migration of screws and plates. To the authorsÕ understanding, the lateral bending moment of the normal spine may induce pullout force acting on the fixation screws. In the current study, the failure mode of the implants that resulted from pullout force was quite similar to that of anterior vertebral implants that failed due to lateral migration. It would be logical to assume that implants with high pullout force would be more stable clinically. It is interesting to note that the pullout strength for Group C was only 1.44 times that of Group E. Theoretically, it should be about twice as strong. The results revealed that the average pullout strength of Groups C and E were 2450 (SD 327) N and 1700 (SD 490) N, respectively. The data showed a substantial amount of variation in both groups. An important factor that may lead to the difference could be the entrance point of the screw placement. In Group E, the screw was placed at the diametric line of vertebral body, which resulted in the maximum contact area between screw and bone. While in Group C, the entrances of the parallel screws were placed one anteriorly and the other posteriorly to the diametric line of vertebral body. The total contact area of the two parallel screws and bone would be less than twice that of Group E. Although data from the current study have shown the Group B––trianguration configuration achieved similar pullout strength to Group C––parallel configuration, in the event of pullout failure, the trianguration configuration would produce a more disastrous consequence. Theoretically, establishing a mathematical model will be a useful tool to predict the pullout strength of any configuration of screw insertions. The experimental results of the current study could be used to validate this model. The model could then predict any configuration of screw insertion. However, this work is not done in the current study, an alternative finite element analysis is now being carried out to simulate all the configurations of screw instrumentations described in the current study.
Acknowledgement The authors would like to thank the National Science Council of the Republic of China for the financially supporting this research under contract no. NSC 892320-B-182-018.
S45
References Bauer, R., Mostegel, A., Eichenauer, M., 1986. An analysis of the results of Dwyer and Zielke instrumentations in the treatment of scoliosis. Arch. Orthop. Trauma Surg. 105, 302–309. Breeze, S.W., Doherty, B.J., Noble, P.S., LeBlanc, A., Heggeness, M.H., 1988. A biomechanical study of anterior thoracolumbar screw fixation. Spine 23, 1829–1831. Dunn, H.K., 1984. Anterior stabilization of thoracolumbar injuries. Clin. Orthop. 189, 116–124. Dwyer, A.F., 1970. Anterior instrumentation in scoliosis. J. Bone Joint Surg. 52B, 782–787. Gilbert, S.G., Johns, P.C., Chow, D.C., Black, R.C., 1993. Relationship of vertebral bone screw axial pullout strength to quantitative tomographic trabecular bone mineral content. J. Spinal Disord. 6, 513–521. Halvorson, T.L., Kelley, L.A., Thomas, K.A., Whitecloud III, T.S., Cook, S.D., 1994. Effects of bone mineral density on pedicle screw fixation. Spine 19, 2415–2420. Horton, W.C., Blackstock, S.F., Norman, J.T., 1996. Strength of fixation of the anterior vertebra body screw. Spine 21, 439–444. Hsieh, P.H., Chen, W.J., Chen, L.H., Niu, C.C., 1999. An unusual complication of anterior spinal instrumentation: hemothorax contralateral to the side of the incision. A case report. J. Bone Joint Surg. 81A, 998–1001. Kaneda, K., Abumi, K., Fujiya, M., 1984. Burst fractures with neurological deficits of the thoracolumbar spine: results of the anterior decompression and stabilization with anterior instrumentation. Spine 9, 788–795. Kostuik, J.P., 1988. Anterior fixation of burst fractures of the thoracic and lumbar spine with or without neurological involvement. Spine 13, 286–293. Kostuik, J.P., Errico, T.J., Gleason, T.F., 1986. Techniques of internal fixation for degenerative conditions of the lumbar spine. Clin. Orthop. 203, 219–231. Krag, M.H., Beynnon, B.D., Pope, M.H., 1986. An internal fixator for posterior application to short segments of the thoracic, lumbar, or lumbosacral spine: design and testing. Clin. Orthop. 203, 75–98. Lieberman, I.H., Khazim, R., Woodside, T., 1998. Anterior vertebral body screw pullout testing. A comparison of Zielke, Kaneda, universal spine system, and universal spine system with pulloutresistant nut. Spine 23, 908–910. Lim, T.H., An, H.S., Evanich, C., 1995. Strength of anterior vertebral screw fixation in relation to bone mineral density. J. Spinal Disord. 8, 121–125. Matsuzaki, H., Tokuhashi, Y., Wakabayashi, K., Kitamura, S., 1993. Penetration of a screw into the thoracic aorta in anterior spinal instrumentation. A case report. Spine 18, 2327–2331. Ogon, M., Haid, C., Krismer, M., Sterzinger, W., Bauer, R., 1996. Comparison between single-screw and triangulated, double-screw fixation in anterior spine surgery. A biomechanical test. Spine 21, 2728–2734. Okuyama, K., Sato, K., Abe, E., 1993. Stability of transpedicle screwing for the osteoporotic spine: An in vitro study of mechanical stability. Spine 18, 2240–2245. Ruland, C.M., McAfee, P.C., Wareden, K.E., Cunningham, B.W., 1991. Triangulation of pedicular instrumentation: a biomechanical analysis. Spine 16, 270–276. Snyder, B.D., Zaltz, I., Hall, J.E., Emans, J.B., 1995. Predicting the integrity of vertebral bone screw fixation in anterior spinal instrumentation. Spine 20, 1568–1574. Spiegel, D.A., Cunningham, B.W., Oda, I., 2000. Anterior vertebral screw strain with and without solid interspace support. Spine 25, 2755–2761. Yuan, H.A., Mann, K.A., Found, E.M., 1988. Early clinical experience with the Syracuse I-plate: an anterior spinal fixation device. Spine 23, 278–285.