Eur Spine J DOI 10.1007/s00586-015-3904-3
ORIGINAL ARTICLE
Effect of augmentation techniques on the failure of pedicle screws under cranio-caudal cyclic loading Richard Bostelmann1 • Alexander Keiler2 • Hans Jakob Steiger1 Armin Scholz3 • Jan Frederick Cornelius1 • Werner Schmoelz2
•
Received: 7 October 2014 / Revised: 22 March 2015 / Accepted: 22 March 2015 Ó Springer-Verlag Berlin Heidelberg 2015
Abstract Purpose Augmentation of pedicle screws is recommended in selected indications (for instance: osteoporosis). Generally, there are two techniques for pedicle screw augmentation: inserting the screw in the non cured cement and in situ-augmentation with cannulated fenestrated screws, which can be applied percutaneously. Most of the published studies used an axial pull out test for evaluation of the pedicle screw anchorage. However, the loading and the failure mode of pullout tests do not simulate the craniocaudal in vivo loading and failure mechanism of pedicle screws. The purpose of the present study was to assess the fixation effects of different augmentation techniques (including percutaneous cement application) and to investigate pedicle screw loosening under physiological cyclic cranio-caudal loading. Methods Each of the two test groups consisted of 15 vertebral bodies (L1–L5, three of each level per group). Mean age was 84.3 years (SD 7.8) for group 1 and 77.0 years (SD 7.00) for group 2. Mean bone mineral density was 53.3 mg/cm3 (SD 14.1) for group 1 and 53.2 mg/cm3 (SD 4.3) for group 2. 1.5 ml high viscosity PMMA bone cement was used for all augmentation techniques. For test group 1, pedicles on the right side of the vertebrae were instrumented with solid pedicle screws in standard fashion without augmentation and & Richard Bostelmann [email protected] 1
Department of Neurosurgery, University Hospital of Duesseldorf, Moorenstrasse 5, 40225 Du¨sseldorf, Germany
2
Department of Trauma Surgery, Medical University Innsbruck, Innsbruck, Austria
3
Department of Trauma Surgery, University Hospital of Duesseldorf, Du¨sseldorf, Germany
served as control group. Left pedicles were instrumented with cannulated screws (Viper cannulated, DePuy Spine) and augmented. For test group 2 pedicles on the left side of the vertebrae were instrumented with cannulated fenestrated screws and in situ augmented. On the right side solid pedicle screws were augmented with cement first technique. Each screw was subjected to a cranio-caudal cyclic load starting at 20–50 N with increasing upper load magnitude of 0.1 N per cycle (1 Hz) for a maximum of 5000 cycles or until total failure. Stress X-rays were taken after cyclic loading to evaluate screw loosening. Results Test group 1 showed a significant higher number of load cycles until failure for augmented screws compared to the control (4030 cycles, SD 827.8 vs. 1893.3 cycles, SD 1032.1; p \ 0.001). Stress X-rays revealed significant less screw toggling for the augmented screws (5.2°, SD 5.4 vs. 16.1°, SD 5.9; p \ 0.001). Test group 2 showed 3653.3 (SD 934) and 3723.3 (SD 560.6) load cycles until failure for in situ and cement first augmentation. Stress X-rays revealed a screw toggling of 5.1 (SD 1.9) and 6.6 (SD 4.6) degrees for in situ and cement first augmentation techniques (p [ 0.05). Conclusion Augmentation of pedicle screws in general significantly increased the number of load cycles and failure load comparing to the nonaugmented control group. For the augmentation technique (cement first, in situ augmented, percutaneously application) no effect could be exhibited on the failure of the pedicle screws. By the cranio-caudal cyclic loading failure of the pedicle screws occurred by screw cut through the superior endplate and the characteristic ‘‘windshield-wiper effect’’, typically observed in clinical practice, could be reproduced. Keywords Biomechanical investigation Load measurement Cyclic loading Spine Spinal fusion
123
Eur Spine J
Cement augmentation Osteoporosis Cannulated screws Revision surgery Cadaver study Osteoporotic fractures
Introduction It is projected, that the continuous increase of the adults ‘‘65-year-old and older’’ in the industrial countries over the last years will proceed. Consequently, also the number of osteoporosis, a typical disease of this age, will rise and with it the costs for health care systems (all time high in 2010:1.9 billion € in Germany) [1]. Every year more than 400,000 vertebral body fractures are anew diagnosed in Europe [2]. Their number will double probably till 2050 due to the demographic change. The prevalence of this fracture form rises at age and amounts to 5–20 % in the group of the 50-year-old and [50 % in the group of about 80-year-old [3, 4]. If surgical therapy for spinal instrumentation is necessary, minimal access methods seem to be favorable. It is still not clear, if percutanously inserted screws, with or without additional cement augmentation reach the same stability as in open procedures. Generally, there are two techniques for pedicle screw augmentation: inserting the screw into the prefilled, noncured cement and in situ-augmentation using cannulated fenestrated screws [5–8]. With both techniques cement can be applied percutaneously. Most of the published studies used an axial pull out test for evaluation of the pedicle screw anchorage. However, the loading and the failure mode of pullout tests does not simulate the cranio-caudal in vivo loading and failure mechanism of pedicle screws [9–13]. The purpose of the study was to investigate pedicle screw loosening under physiological cyclic cranio-caudal loading. First the effect of screw augmentation technique was compared to a non-augmented control-group and in a second part the pedicle screw anchoring of two different cementing techniques was investigated.
Materials and methods Specimens Osteoporotic spines with a bone mineral density (BMD) \80 mg/cm3 [14] were used for biomechanical testing. Three lumbar spines of the levels L1–L5 (n = 15) were selected for each of the two test groups (group 1: mean BMD 53.3 mg/cm3, SD (standard deviation) 14.1; mean age 84.3 a, SD 7.8; group 2: mean BMD 53.2 mg/cm3, SD 4.3; mean age 77.00 a SD 7.00). Trabecular bone mineral
123
density (BMD) of the specimens was measured using a preoperative quantitative computer tomograph scan (qCT, Lightspeed VCT16, GE Healthcare, Waukesha, Wisconsin, USA) including a calibration phantom (EFP, European Forearm Phantom). The double shrink-wrapped specimens were frozen at minus 20 °C. For preparation they were thawed overnight at 6 °C and all soft tissue such as muscle and fat was dissected. The different tested pedicle screw-systems were: 1.
2. 3.
SI polyaxial screw (Expedium Spine System—DePuy Spine, Raynham, MA, USA. SI stands for single innie. It is a single stage closure cap to fix polyaxiality of the screw and movement of the rod.) Cannulated polyaxial screw (Viper System, cannulated—DePuy Spine, Raynham, MA, USA). Cannulated and fenestrated polyaxial screw (Viper System CorticalFix, double threading in the area of pedicle anchorage, cannulated and with 6 holes at the distal end of the screw—DePuy Spine, Raynham, MA, USA).
The used screws of all systems had a size of 6 9 45 mm. For augmentation 1.5 ml of high viscosity PMMA cement (Confidence, DePuy Spine, Raynham, MA, USA) was used for all screws. Instrumentation technique Pedicle screws were inserted with standard instruments. Laminar cortex overlying the pedicle screw entry point was broached with a starter awl. A thread for the placement of the pedicle screw was cut. Care was taken to place the pedicle screws on both sides without crossing the midline of the vertebral body. Moreover for application of the cement, a needle with two different diameters was used to simulate the percutaneous application technique. For test group 1 (Fig. 1a) right pedicles of each vertebral body were instrumented with solid SI screws in standard fashion without augmentation and served as control (RNC, right no cement). Left pedicles were instrumented with cannulated screws. After full implantation the pedicle screws were turned back for three turns to open up space for the augmentation. An augmentation-needle was inserted through the cannula of the screws and 1.5 ml of high-viscous PMMA cement was injected. Subsequently screws were fully inserted (LCM, left cemented; cannulated cemented). For test group 2 (Fig. 1b) the right pedicles of each vertebral body were threaded and augmented with 1.5 ml high-viscous PMMA cement via a cannula. Then the solid SI screws were inserted into the dough cement (RSC, right screw after cement; cement-first technique). After
Eur Spine J
Fig. 2 Test setup in the material testing machine showing the specimen mounted on an x–y bearing table and the motion analysis system attached to the head of the screw. The red arrows are showing the load application and the pivot axis to allow a tilting motion of the screw in the vertebral body
Fig. 1 Axial X-rays a group 1, on the left side cementing technique through the turned back cannulated screws LCM was applied, on the right side no cement was injected and served as control (RNC). b Group 2, left pedicle screws were fenestrated and in situ augmented (LFC), on the right side the screw was inserted after the cement first technique (RSC) was applied
placement of screws in the left pedicles 1.5 ml of high viscous PMMA cement were injected in situ through the cannulated fenestrated screws (LFC, left fenestrated cement; in situ augmentation). After curing time the lumbar spines were separated in single vertebral bodies and each single vertebral body was embedded in PMMA (Technovit 3040, Heraeus Kulzer GmbH, Wehrheim, Germany) to form an endplate mold for fixation purpose in the testing-machine and stored frozen until testing. The implantation and augmentation of the screws was radiologically documented. (Fig. 1a, b). Cyclic loading Instrumented specimens were fixed on an x–y table allowing translational movements in two axes in a servohydraulic biaxial material testing machine (858 Mini Bionix II, MTS, Eden Prairie, MN, USA) (Fig. 2). A loading rod was connected to the pedicle screw of one side and a 3D motion analysis system (Zebris, Winbiomechanics, Isny,
Germany) was mounted to the pedicle screw and the base plate of the x–y table (Fig. 2). Pedicle screws were loaded with a force of 20–50 N with increasing of the upper limit of 0.1 N per cycle and a frequency of 1 Hz for 5000 cycles (20–550 N) or until total loosening of the screw (10 mm axial displacement of the actuator in caudal direction). This criterion for termination of the loading does not correspond to 10 mm screw head displacement in clinical practice, as the setup also allows a rotation around an axis above the pedicle with a lever arm of 15 mm relative to the screw head. A constantly increasing load magnitude is a wellsuited load protocol to provoke implant loosening and osteosynthesis failure in varying bone quality and osteosynthesis stiffness [15–18]. During all testing the displacement and force at the actuator was recorded. The 3D motion analysis system recorded every 50 cycles for two cycles the angle of the screw head relatively to the base plate to which the vertebral body was rigidly fixed. For post testing data evaluation of screw loosening the failure criteria was set to an absolute tilt of the pedicle screw of eight degrees to the starting point or to a relative increase in pedicle screw tilt of more than 1° within 50 load cycles. Stress X-rays After cyclic loading each vertebral body was cut in the midline and each half was loaded in caudal and cranial
123
Eur Spine J
direction in a custom made holding device for stress X-rays with 20 N. Lateral X-rays were taken for documenting the screw-position in both extreme-points. These X-rays of the maximum deflection in cranial and caudal direction were digitally superimposed and the angle between the two loaded screw positions (Fig. 3) was measured using ICO view software package (version 3.0.0, x86 edition, ITH icoserve technology for healthcare GmbH, Innsbruck, Austria). Pullout tests After cyclic loading and stress X-rays pullout-tests were performed in a servohydraulic biaxial material testing machine (858 Mini Bionix II, MTS, Eden Prairie, MN, USA). Each half of a vertebral body was clamped in a bench vice mounted on a ball-and-socket joint, allowing 3-dimensional spatial adjustment. Pullout tests were performed under displacement control at a constant speed of 10 mm/min until complete removal of the screw or fracturing of the pedicle. During all testing axial displacement and force at the actuator were recorded at 100 Hz. Maximum load during pullout tests and load–displacement graphs for each test were analysed. Statistical analysis Statistical analysis was carried out using SPSS software package (version 21.0, SPSS, Chicago, Illinois, USA). Jarque–Bera-tests were performed for evaluating the distribution pattern. As all data were normal distributed, paired t tests were applied for comparing screws inserted on the left side vs. screws inserted on the right side. Oneway analysis of variances (ANOVA) with Bonferroni post
Fig. 3 Digitally superimposed X-ray of a right half of a vertebral body of the control group (RNC) loaded with 20 N in caudal and cranial direction. Axes of the screw in maximum cranial and caudal deflection are marked in orange
123
hoc testing was used for comparison of the different augmentation techniques. Level of significance was set to a value of a = 0.05.
Results In test group 1, two of 15 vertebral bodies had to be excluded from data evaluation as the screw tips were too close together to allow an uninfluenced loading of a single screw, leaving 13 vertebral bodies for statistical evaluation. Some of the screw heads tilted in the polyaxial joint during the testing prior to loosening and were not included in the analysis (details see below). Cyclic loading In group 1 all (13) of the screws in the nonaugmented control group RNC failed within the 5000 load cycles. Twelve due to a relative tilt of more than 1° within 50 load cycles and one due to an absolute tilt of more than 8°. For the cannulated cementing technique LCM 12 (of 13) screws failed during the cyclic loading period with four screws reaching eight degrees of angulation and eight screws tilting more than 1° within 50 cycles. In one screw of the LCM augmentation the polyaxial screw head tilted prior to screw loosening, the test was stopped and the screw was not included in the analysis. The mean failure cycle was 1893.3 (SD 1032) for the control screws and 4030 (SD 827.8) for the augmented screws (p \ 0.001). In group 2 for in situ augmentation (LFC) 14 (of 15) screws failed within the 5000 load cycles, eight due to relative tilt and six due to an absolute tilt. Analysing the cement-first technique (RSC) 13 (of 15) screws failed during the cyclic loading period—all of them because of tilting more than one degree within 50 cycles. In one screw of the LFC augmentation and two screws of the RSC augmentation the polyaxial screw head tilted prior to screw loosening, tests were stopped and the screws were not included in the analysis. The mean failure cycle for in situ augmented screws (LFC) was 3653.3 (SD 934.0) and 3723.3 (SD 560.6) for cement-first technique (RSC) (p = 0.824). Comparison of the mean failure cycles of all cementing techniques to the control showed a significant difference (p \ 0.001), while there was no significant difference in between the three augmentation techniques (p [ 0.999) (Fig. 4). Stress X-rays Stress X-rays of group 1 revealed significant less screw toggling for cemented screws (LCM) compared to the non-
Eur Spine J
Fig. 4 The control screwing technique exhibited significant lower load cycles to failure compared to all other groups. The boxplots are showing the median, the 25th and the 75th quartile of the load cycles to failure for the four groups. Statistical significant differences to the control group are marked with an asterisk
Fig. 5 The control screwing technique exhibited a significant larger toggeling after cyclic loading compared to all other groups. The boxplots are showing the median, the 25th and the 75th quartile of the screw toggeling after cyclic loading for the four groups. Statistical significant differences to the control group are marked with an asterisk
augmented control RNC (mean angle 5.2°, SD 5.4° vs. 16.1°, SD 6.0; p \ 0.001). Group 2 showed a screw toggling of 5.1° (SD 1.9) and 6.6° (SD 4.6) for in situ (LFC) and cement-first augmentation (RSC), respectively. None of the differences between the cementing techniques of group 2 were statistically significant (p = 0.282). Comparing the means of screw toggling of all of the cementing techniques (LCM, LFC, RSC) to the non-augmented control (RNC) revealed a statistical significant difference (p \ 0.001), while there was no significant difference in between the three augmentation techniques (p [ 0.999) (Fig. 5). Pullout tests Comparing the means of the maximum pullout forces after cyclic loading for screws of group 1 showed 776.0 N (SD 380.3) for cannulated augmented screws (LCM) and 168.6 N (SD 107.9) for the nonaugmented control (RNC). Statistical analysis showed a significant higher maximum pullout force for cannulated cemented screws (p \ 0.001) (Fig. 6). In group 2 mean values of 629.9 N (SD 159.1) for the in situ augmentation (LFC) and 497.5 N (SD 255.1) for the cement-first technique (RSC) were measured. None of the differences between the cementing techniques of group 2 were statistically significant (p [ 0.999) (Fig. 6). Means of all of the cementing techniques LCM, LFC, RSC revealed no significant difference in between the three
Fig. 6 The control screwing technique exhibited a significant lower pullout force after cyclic loading compared to all other groups. The boxplots are showing the median, the 25th and the 75th quartile of the ultimate pull out force after cyclic loading for the four groups. Statistical significant differences to the control group are marked with an asterisk
augmentation techniques (p [ 0.999), but documented a statistical significant difference compared to the non-augmented control RNC (p \ 0.001) (Fig. 6). Testing the maximal pullout force of the inserted screws resulted in pedicle fractures in 19 of 43 (44.2 %) of the augmented screws.
123
Eur Spine J
Discussion To investigate if a biomechanically sufficient augmentation is possible with a percutaneous technique, we carried out the present study with cyclic cranio-caudal loading of the screws. First it was found that augmentation of pedicle screws significantly increased the number of load cycles and failure load compared to the non-augmented control group. A second result is that augmentation in percutaneous technique (that is using cannulated screws with percutaneous equipment) showed no significant difference to the conventional technique. As final result we found that after cranio-caudal cyclic loading screw cut out through the superior endplate. This ‘‘windshield-wiper effect’’ is typically observed in clinical practise, underlining that this test set up reproduces clinical findings. Axial pull out tests for evaluation of the pedicle screw anchorage are easy to conduct and give a practical parameter, but it must also be considered, that they do not reflect the clinical setting [9, 19, 20] as they apply an unphysiological loading to the pedicle screw. Loosening due to fatigue loading and screw breakage are much more common reasons for screw failure. In accordance to literature we therefore simulated cranio-caudal in vivo loading and failure mechanism of pedicle screws as a more clinically applicable testing mode [15–18]. A constantly increasing load magnitude was shown to be a well suited load protocol to provoke implant loosening and osteosynthesis failure in varying bone quality and osteosynthesis stiffness for trabecular anchoring of implants and screws such as in the femoral head or the humeral head. To achieve a good transferability into clinical practice we decided to test most used screw diameters (6 mm) and length (45 mm). The same screw size was used for all vertebral bodies to avoid differences in fixation strength. On the other hand also the kind of cement being used for augmentation matters. We used a high viscosity PMMA cement. Biomechanical studies reported a higher number of load cycles and a higher bending moment to failure for PMMA augmented screws compared to screws augmented with for instance calcium triglyceride bone cement. Also high-viscosity cement may enhance the fixation effect of the screws compared to low-viscosity cement. It is reported that in lumbar spine a significant relative increase in pullout strength could be reached by using a volume of 1.0–3.0 cc. With 1.5 ml we aimed for a compromise between using enough volume to improve mechanical strength while avoiding the risks of cement leakage associated with higher cement volumes. Moreover with the injected volume of 1.5 ml, we were only slightly above the optimum filling of 75 % (about 1.03 ml) of the volume suggested for bone cement augmentation of the trajectory [21].
123
Augmentation of pedicle screws in osteoporotic bone increases the number of load cycles until failure, significantly (LCM 212.8 %, LFC 192.9 %, RSC 196.6 %). Moreover the tested augmentation techniques showed no significant difference, no matter if they are done percutaneously or not. This finding underlines, to the author’s knowledge for the first time described in biomechanical reports, that an augmentation in a percutaneous technique is feasible and as effective as with usual techniques. Although there is first report on clinical experience with short term follow up in managing elderly thoracolumbar fractures using percutaneous cemented screws [22]. The advantage of a percutaneous procedure would be for instance a less invasive approach, shorter operation time and the benefit of in situ augmented screws, moreover the autochthonous back muscles is not devascularized. However, in situ augmentation of the screws requires a higher pressure for cement application as the cannulas have a smaller diameter [23]. Some authors advocate that the technique of cement injection does matter and reported higher pullout and fatigue strength for in situ screw augmentation techniques than for prefilled technique [20, 24]. From a physical point of view, the cement is distributed according to the falling pressure gradient. So it will concentrate around the distal third of the fenestrated screw, serving the first holes first. Injection into a tapped hole can lead to a filling along the thread, for instance if cement with low viscosity is used [24]. This difference in the distribution may promote a lower force to failure. We used high viscosity cement in all cases and observed only a distribution around the distal third of the thread, regardless of the application. This may be due to the high viscosity of the cement. In addition a trend towards higher failure loads for high-viscosity cement compared to low-viscosity cement has been described. Compared to the non-augmented control group, we also found a substantial increase in the mean pull-out strength after cyclic testing (LCM 460 %, LFC 373.6 %, RSC 295.1 %). This goes well together with similar observations in literature: 119 % [8], 162 % [8], 181 % [25], 187 % [6], 206 % [25], 213 % [25], 250 % [26]. However, this might be due differences in preceding cyclic loading, in bone mineral density, used screw types, filling volume and augmentation material and technique. In 44.2 % (19 of 43) of the augmented screws the failure mode was pedicle fracture with the cement still attached to the screw and no axial screw pullout as for non-augmented pedicle screws. This failure mode of augmented pedicle screws in pullout tests was already reported by other authors, e.g. Sarzier et al. [25] reports on 9 of 21 (42.9 %) and Bullmann et al. [23] on 10 of 23 (43.5 %) pedicle fractures of augmented screws in their trials.
Eur Spine J
In the present investigation with cranio-caudal cyclic loading screw cut out through the superior endplate and the ‘‘windshield-wiper effect’’, which is usually observed in clinical practice, could be reproduced. This underlines the assumption that our loading model is closer to reality. On the other hand it is difficult to determine the physiological load magnitude for the cranio-caudal cyclic testing. In measurements with a telemetrized internal fixator of 10 patients Rohlmann et al. [27, 28] reported axial craniocaudal peak force magnitudes in everyday activities between 200 and 250 N. With the constantly increasing load magnitude during cyclic loading the mean failure load of the non-cemented control group (RNC) was 239 N, while the mean failure load for all augmented screws was 453, 415 and 422 N for the LCM, LFC and RSC group, respectively. However, it should be considered that in the present study only osteoporotic vertebrae (mean age of 84.3 years) were used while the active patients in the study by Rohlmann et al. had a mean age of 50.6 years. Combining the in vivo measurements with the results of the present study, it can be assumed, that in patients with reduced bone quality pedicle screw loosening might occur during everyday activity, while augmentation raises the failure load above the forces occurring during everyday activities.
Limitations of the study First, the pull out test was done after cyclic loading with loosened pedicle screws. Therefore, the absolute values should be regarded with this limitation. Nevertheless this limitation occurred to all screws in a similar way, thus the proportional trend could be transferred to clinical practice. A second limitation in this study is that it was conducted on cadavers, so we have no in vivo situation. Cadaver models have of course their inherent limitations. They cannot replicate biological factors such as screw ingrowth and bone remodeling around screws. Osseous resistance, muscular loading and in vivo loading of structures can only be approximated.
Conclusion Augmentation of pedicle screws significantly increased the number of load cycles and failure load compared to the non-augmented control group. The augmentation technique in situ with perforated screws or cement first even percutaneously, had no effect on the loosening of the pedicle screws. By the cranio-caudal cyclic loading screw cut out through the superior endplate and the typical ‘‘windshield-wiper effect’’, typically observed in clinical practice, could be reproduced. The augmentation technique, even
percutaneously, had no effect on the failure of the pedicle screws. Conflict of interest The study was supported by DePuy Spine Germany (receiving the implants, cement, and laboratory cost, no wages have been paid), but none of the authors has any potential conflict of interest.
References 1. Office GFS (2002–2010) Total cost of illness in millions of Euro. Classification: years, gender, ICD10, provider (Primary source: Federal Statistical Office, Total Health Expenditure, Statistics). In., Berlin 2. Masala S, Lunardi P, Fiori R, Liccardo G, Massari F, Ursone A, Simonetti G (2004) Vertebroplasty and kyphoplasty in the treatment of malignant vertebral fractures. J Chemother 16(Suppl 5):30–33 3. Melton LJ 3rd, Thamer M, Ray NF, Chan JK, Chesnut CH 3rd, Einhorn TA, Johnston CC, Raisz LG, Silverman SL, Siris ES (1997) Fractures attributable to osteoporosis: report from the National Osteoporosis Foundation. J Bone Mineral Res 12:16–23. doi:10.1359/jbmr.1997.12.1.16 4. Papaioannou A, Watts NB, Kendler DL, Yuen CK, Adachi JD, Ferko N (2002) Diagnosis and management of vertebral fractures in elderly adults. Am J Med 113:220–228 5. Becker S, Chavanne A, Spitaler R, Kropik K, Aigner N, Ogon M, Redl H (2008) Assessment of different screw augmentation techniques and screw designs in osteoporotic spines. Eur Spine Jy 17:1462–1469. doi:10.1007/s00586-008-0769-8 6. Burval DJ, McLain RF, Milks R, Inceoglu S (2007) Primary pedicle screw augmentation in osteoporotic lumbar vertebrae: biomechanical analysis of pedicle fixation strength. Spine 32:1077–1083. doi:10.1097/01.brs.0000261566.38422.40 7. Chen LH, Tai CL, Lai PL, Lee DM, Tsai TT, Fu TS, Niu CC, Chen WJ (2009) Pullout strength for cannulated pedicle screws with bone cement augmentation in severely osteoporotic bone: influences of radial hole and pilot hole tapping. Clin Biomech (Bristol, Avon) 24:613–618. doi:10.1016/j.clinbiomech.2009.05. 002 8. Frankel BM, D’Agostino S, Wang C (2007) A biomechanical cadaveric analysis of polymethylmethacrylate-augmented pedicle screw fixation. J Neurosurg Spine 7:47–53. doi:10.3171/SPI-07/ 07/047 9. Zindrick MR, Wiltse LL, Widell EH, Thomas JC, Holland WR, Field BT, Spencer CW (1986) A biomechanical study of intrapeduncular screw fixation in the lumbosacral spine. Clin Orthop Relat Res (203):99–112 10. Rohlmann A, Bergmann G, Graichen F (1997) Loads on an internal spinal fixation device during walking. J Biomech 30:41–47 11. Rohlmann A, Bergmann G, Graichen F, Weber U (1997) Comparison of loads on internal spinal fixation devices measured in vitro and in vivo. Med Eng Phys 19:539–546 12. Rohlmann A, Graichen F, Bergmann G (2000) Influence of load carrying on loads in internal spinal fixators. J Biomech 33:1099–1104 13. Wu JC, Chen CS, Yip SW, Hsu ML (2012) Stress distribution and micromotion analyses of immediately loaded implants of varying lengths in the mandible and fibular bone grafts: a three-dimensional finite element analysis. Int J Oral Maxillofac Implants 27:e77–e84 14. Felsenberg D, Gowin W (1999) Bone densitometry by dual energy methods. Der Radiologe 39:186–193
123
Eur Spine J 15. Windolf M, Maza ER, Gueorguiev B, Braunstein V, Schwieger K (2010) Treatment of distal humeral fractures using conventional implants. Biomechanical evaluation of a new implant configuration. BMC Musculoskelet Disord 11:172. doi:10.1186/ 1471-2474-11-172 16. Unger S, Erhart S, Kralinger F, Blauth M, Schmoelz W (2012) The effect of in situ augmentation on implant anchorage in proximal humeral head fractures. Injury 43:1759–1763. doi:10. 1016/j.injury.2012.07.003 17. Kiner DW, Wybo CD, Sterba W, Yeni YN, Bartol SW, Vaidya R (2008) Biomechanical analysis of different techniques in revision spinal instrumentation: larger diameter screws versus cement augmentation. Spine 33:2618–2622. doi:10.1097/BRS.0b013 e3181882cac 18. Goel VK, Winterbottom JM, Weinstein JN (1994) A method for the fatigue testing of pedicle screw fixation devices. J Biomech 27:1383–1388 19. Law M, Tencer AF, Anderson PA (1993) Caudo-cephalad loading of pedicle screws: mechanisms of loosening and methods of augmentation. Spine 18:2438–2443 20. Kueny RA, Kolb JP, Lehmann W, Puschel K, Morlock MM, Huber G (2014) Influence of the screw augmentation technique and a diameter increase on pedicle screw fixation in the osteoporotic spine: pullout versus fatigue testing. Eur Spine J. doi:10. 1007/s00586-014-3476-7 21. Fan H, Zhang R, Shen C, Dong F, Li Y, Song P, Gong C, Wang Y (2000) The biomechanical properties of pedicle screw fixation combined with trajectory bone cement augmentation in
123
22.
23.
24.
25.
26.
27.
28.
osteoporotic vertebrae. J Spin Disord Tech. doi:10.1097/BSD. 1090b1013e3182a14870 Pesenti S, Blondel B, Peltier E, Adetchessi T, Dufour H, Fuentes S (2014) Percutaneous cement-augmented screws fixation in the fractures of the aging spine: is it the solution? BioMed Res Int 2014:610675. doi:10.1155/2014/610675 Bullmann V, Schmoelz W, Richter M, Grathwohl C, Schulte TL (2010) Revision of cannulated and perforated cement-augmented pedicle screws: a biomechanical study in human cadavers. Spine 35:E932–E939. doi:10.1097/BRS.0b013e3181c6ec60 Choma TJ, Pfeiffer FM, Swope RW, Hirner JP (2012) Pedicle screw design and cement augmentation in osteoporotic vertebrae: effects of fenestrations and cement viscosity on fixation and extraction. Spine 37:E1628–E1632. doi:10.1097/BRS.0b013e3182740e56 Sarzier JS, Evans AJ, Cahill DW (2002) Increased pedicle screw pullout strength with vertebroplasty augmentation in osteoporotic spines. J Neurosurg 96:309–312 Cook SD, Salkeld SL, Stanley T, Faciane A, Miller SD (2004) Biomechanical study of pedicle screw fixation in severely osteoporotic bone. Spine J 4:402–408. doi:10.1016/j.spinee.2003. 11.010 Rohlmann A, Graichen F, Weber U, Bergmann G (2000) 2000 Volvo Award winner in biomechanical studies: monitoring in vivo implant loads with a telemeterized internal spinal fixation device. Spine 25:2981–2986 Rohlmann A, Bergmann G, Graichen F (1999) Loads on internal spinal fixators measured in different body positions. Eur Spine J 8:354–359
ORIGINAL ARTICLE
Effect of augmentation techniques on the failure of pedicle screws under cranio-caudal cyclic loading Richard Bostelmann1 • Alexander Keiler2 • Hans Jakob Steiger1 Armin Scholz3 • Jan Frederick Cornelius1 • Werner Schmoelz2
•
Received: 7 October 2014 / Revised: 22 March 2015 / Accepted: 22 March 2015 Ó Springer-Verlag Berlin Heidelberg 2015
Abstract Purpose Augmentation of pedicle screws is recommended in selected indications (for instance: osteoporosis). Generally, there are two techniques for pedicle screw augmentation: inserting the screw in the non cured cement and in situ-augmentation with cannulated fenestrated screws, which can be applied percutaneously. Most of the published studies used an axial pull out test for evaluation of the pedicle screw anchorage. However, the loading and the failure mode of pullout tests do not simulate the craniocaudal in vivo loading and failure mechanism of pedicle screws. The purpose of the present study was to assess the fixation effects of different augmentation techniques (including percutaneous cement application) and to investigate pedicle screw loosening under physiological cyclic cranio-caudal loading. Methods Each of the two test groups consisted of 15 vertebral bodies (L1–L5, three of each level per group). Mean age was 84.3 years (SD 7.8) for group 1 and 77.0 years (SD 7.00) for group 2. Mean bone mineral density was 53.3 mg/cm3 (SD 14.1) for group 1 and 53.2 mg/cm3 (SD 4.3) for group 2. 1.5 ml high viscosity PMMA bone cement was used for all augmentation techniques. For test group 1, pedicles on the right side of the vertebrae were instrumented with solid pedicle screws in standard fashion without augmentation and & Richard Bostelmann [email protected] 1
Department of Neurosurgery, University Hospital of Duesseldorf, Moorenstrasse 5, 40225 Du¨sseldorf, Germany
2
Department of Trauma Surgery, Medical University Innsbruck, Innsbruck, Austria
3
Department of Trauma Surgery, University Hospital of Duesseldorf, Du¨sseldorf, Germany
served as control group. Left pedicles were instrumented with cannulated screws (Viper cannulated, DePuy Spine) and augmented. For test group 2 pedicles on the left side of the vertebrae were instrumented with cannulated fenestrated screws and in situ augmented. On the right side solid pedicle screws were augmented with cement first technique. Each screw was subjected to a cranio-caudal cyclic load starting at 20–50 N with increasing upper load magnitude of 0.1 N per cycle (1 Hz) for a maximum of 5000 cycles or until total failure. Stress X-rays were taken after cyclic loading to evaluate screw loosening. Results Test group 1 showed a significant higher number of load cycles until failure for augmented screws compared to the control (4030 cycles, SD 827.8 vs. 1893.3 cycles, SD 1032.1; p \ 0.001). Stress X-rays revealed significant less screw toggling for the augmented screws (5.2°, SD 5.4 vs. 16.1°, SD 5.9; p \ 0.001). Test group 2 showed 3653.3 (SD 934) and 3723.3 (SD 560.6) load cycles until failure for in situ and cement first augmentation. Stress X-rays revealed a screw toggling of 5.1 (SD 1.9) and 6.6 (SD 4.6) degrees for in situ and cement first augmentation techniques (p [ 0.05). Conclusion Augmentation of pedicle screws in general significantly increased the number of load cycles and failure load comparing to the nonaugmented control group. For the augmentation technique (cement first, in situ augmented, percutaneously application) no effect could be exhibited on the failure of the pedicle screws. By the cranio-caudal cyclic loading failure of the pedicle screws occurred by screw cut through the superior endplate and the characteristic ‘‘windshield-wiper effect’’, typically observed in clinical practice, could be reproduced. Keywords Biomechanical investigation Load measurement Cyclic loading Spine Spinal fusion
123
Eur Spine J
Cement augmentation Osteoporosis Cannulated screws Revision surgery Cadaver study Osteoporotic fractures
Introduction It is projected, that the continuous increase of the adults ‘‘65-year-old and older’’ in the industrial countries over the last years will proceed. Consequently, also the number of osteoporosis, a typical disease of this age, will rise and with it the costs for health care systems (all time high in 2010:1.9 billion € in Germany) [1]. Every year more than 400,000 vertebral body fractures are anew diagnosed in Europe [2]. Their number will double probably till 2050 due to the demographic change. The prevalence of this fracture form rises at age and amounts to 5–20 % in the group of the 50-year-old and [50 % in the group of about 80-year-old [3, 4]. If surgical therapy for spinal instrumentation is necessary, minimal access methods seem to be favorable. It is still not clear, if percutanously inserted screws, with or without additional cement augmentation reach the same stability as in open procedures. Generally, there are two techniques for pedicle screw augmentation: inserting the screw into the prefilled, noncured cement and in situ-augmentation using cannulated fenestrated screws [5–8]. With both techniques cement can be applied percutaneously. Most of the published studies used an axial pull out test for evaluation of the pedicle screw anchorage. However, the loading and the failure mode of pullout tests does not simulate the cranio-caudal in vivo loading and failure mechanism of pedicle screws [9–13]. The purpose of the study was to investigate pedicle screw loosening under physiological cyclic cranio-caudal loading. First the effect of screw augmentation technique was compared to a non-augmented control-group and in a second part the pedicle screw anchoring of two different cementing techniques was investigated.
Materials and methods Specimens Osteoporotic spines with a bone mineral density (BMD) \80 mg/cm3 [14] were used for biomechanical testing. Three lumbar spines of the levels L1–L5 (n = 15) were selected for each of the two test groups (group 1: mean BMD 53.3 mg/cm3, SD (standard deviation) 14.1; mean age 84.3 a, SD 7.8; group 2: mean BMD 53.2 mg/cm3, SD 4.3; mean age 77.00 a SD 7.00). Trabecular bone mineral
123
density (BMD) of the specimens was measured using a preoperative quantitative computer tomograph scan (qCT, Lightspeed VCT16, GE Healthcare, Waukesha, Wisconsin, USA) including a calibration phantom (EFP, European Forearm Phantom). The double shrink-wrapped specimens were frozen at minus 20 °C. For preparation they were thawed overnight at 6 °C and all soft tissue such as muscle and fat was dissected. The different tested pedicle screw-systems were: 1.
2. 3.
SI polyaxial screw (Expedium Spine System—DePuy Spine, Raynham, MA, USA. SI stands for single innie. It is a single stage closure cap to fix polyaxiality of the screw and movement of the rod.) Cannulated polyaxial screw (Viper System, cannulated—DePuy Spine, Raynham, MA, USA). Cannulated and fenestrated polyaxial screw (Viper System CorticalFix, double threading in the area of pedicle anchorage, cannulated and with 6 holes at the distal end of the screw—DePuy Spine, Raynham, MA, USA).
The used screws of all systems had a size of 6 9 45 mm. For augmentation 1.5 ml of high viscosity PMMA cement (Confidence, DePuy Spine, Raynham, MA, USA) was used for all screws. Instrumentation technique Pedicle screws were inserted with standard instruments. Laminar cortex overlying the pedicle screw entry point was broached with a starter awl. A thread for the placement of the pedicle screw was cut. Care was taken to place the pedicle screws on both sides without crossing the midline of the vertebral body. Moreover for application of the cement, a needle with two different diameters was used to simulate the percutaneous application technique. For test group 1 (Fig. 1a) right pedicles of each vertebral body were instrumented with solid SI screws in standard fashion without augmentation and served as control (RNC, right no cement). Left pedicles were instrumented with cannulated screws. After full implantation the pedicle screws were turned back for three turns to open up space for the augmentation. An augmentation-needle was inserted through the cannula of the screws and 1.5 ml of high-viscous PMMA cement was injected. Subsequently screws were fully inserted (LCM, left cemented; cannulated cemented). For test group 2 (Fig. 1b) the right pedicles of each vertebral body were threaded and augmented with 1.5 ml high-viscous PMMA cement via a cannula. Then the solid SI screws were inserted into the dough cement (RSC, right screw after cement; cement-first technique). After
Eur Spine J
Fig. 2 Test setup in the material testing machine showing the specimen mounted on an x–y bearing table and the motion analysis system attached to the head of the screw. The red arrows are showing the load application and the pivot axis to allow a tilting motion of the screw in the vertebral body
Fig. 1 Axial X-rays a group 1, on the left side cementing technique through the turned back cannulated screws LCM was applied, on the right side no cement was injected and served as control (RNC). b Group 2, left pedicle screws were fenestrated and in situ augmented (LFC), on the right side the screw was inserted after the cement first technique (RSC) was applied
placement of screws in the left pedicles 1.5 ml of high viscous PMMA cement were injected in situ through the cannulated fenestrated screws (LFC, left fenestrated cement; in situ augmentation). After curing time the lumbar spines were separated in single vertebral bodies and each single vertebral body was embedded in PMMA (Technovit 3040, Heraeus Kulzer GmbH, Wehrheim, Germany) to form an endplate mold for fixation purpose in the testing-machine and stored frozen until testing. The implantation and augmentation of the screws was radiologically documented. (Fig. 1a, b). Cyclic loading Instrumented specimens were fixed on an x–y table allowing translational movements in two axes in a servohydraulic biaxial material testing machine (858 Mini Bionix II, MTS, Eden Prairie, MN, USA) (Fig. 2). A loading rod was connected to the pedicle screw of one side and a 3D motion analysis system (Zebris, Winbiomechanics, Isny,
Germany) was mounted to the pedicle screw and the base plate of the x–y table (Fig. 2). Pedicle screws were loaded with a force of 20–50 N with increasing of the upper limit of 0.1 N per cycle and a frequency of 1 Hz for 5000 cycles (20–550 N) or until total loosening of the screw (10 mm axial displacement of the actuator in caudal direction). This criterion for termination of the loading does not correspond to 10 mm screw head displacement in clinical practice, as the setup also allows a rotation around an axis above the pedicle with a lever arm of 15 mm relative to the screw head. A constantly increasing load magnitude is a wellsuited load protocol to provoke implant loosening and osteosynthesis failure in varying bone quality and osteosynthesis stiffness [15–18]. During all testing the displacement and force at the actuator was recorded. The 3D motion analysis system recorded every 50 cycles for two cycles the angle of the screw head relatively to the base plate to which the vertebral body was rigidly fixed. For post testing data evaluation of screw loosening the failure criteria was set to an absolute tilt of the pedicle screw of eight degrees to the starting point or to a relative increase in pedicle screw tilt of more than 1° within 50 load cycles. Stress X-rays After cyclic loading each vertebral body was cut in the midline and each half was loaded in caudal and cranial
123
Eur Spine J
direction in a custom made holding device for stress X-rays with 20 N. Lateral X-rays were taken for documenting the screw-position in both extreme-points. These X-rays of the maximum deflection in cranial and caudal direction were digitally superimposed and the angle between the two loaded screw positions (Fig. 3) was measured using ICO view software package (version 3.0.0, x86 edition, ITH icoserve technology for healthcare GmbH, Innsbruck, Austria). Pullout tests After cyclic loading and stress X-rays pullout-tests were performed in a servohydraulic biaxial material testing machine (858 Mini Bionix II, MTS, Eden Prairie, MN, USA). Each half of a vertebral body was clamped in a bench vice mounted on a ball-and-socket joint, allowing 3-dimensional spatial adjustment. Pullout tests were performed under displacement control at a constant speed of 10 mm/min until complete removal of the screw or fracturing of the pedicle. During all testing axial displacement and force at the actuator were recorded at 100 Hz. Maximum load during pullout tests and load–displacement graphs for each test were analysed. Statistical analysis Statistical analysis was carried out using SPSS software package (version 21.0, SPSS, Chicago, Illinois, USA). Jarque–Bera-tests were performed for evaluating the distribution pattern. As all data were normal distributed, paired t tests were applied for comparing screws inserted on the left side vs. screws inserted on the right side. Oneway analysis of variances (ANOVA) with Bonferroni post
Fig. 3 Digitally superimposed X-ray of a right half of a vertebral body of the control group (RNC) loaded with 20 N in caudal and cranial direction. Axes of the screw in maximum cranial and caudal deflection are marked in orange
123
hoc testing was used for comparison of the different augmentation techniques. Level of significance was set to a value of a = 0.05.
Results In test group 1, two of 15 vertebral bodies had to be excluded from data evaluation as the screw tips were too close together to allow an uninfluenced loading of a single screw, leaving 13 vertebral bodies for statistical evaluation. Some of the screw heads tilted in the polyaxial joint during the testing prior to loosening and were not included in the analysis (details see below). Cyclic loading In group 1 all (13) of the screws in the nonaugmented control group RNC failed within the 5000 load cycles. Twelve due to a relative tilt of more than 1° within 50 load cycles and one due to an absolute tilt of more than 8°. For the cannulated cementing technique LCM 12 (of 13) screws failed during the cyclic loading period with four screws reaching eight degrees of angulation and eight screws tilting more than 1° within 50 cycles. In one screw of the LCM augmentation the polyaxial screw head tilted prior to screw loosening, the test was stopped and the screw was not included in the analysis. The mean failure cycle was 1893.3 (SD 1032) for the control screws and 4030 (SD 827.8) for the augmented screws (p \ 0.001). In group 2 for in situ augmentation (LFC) 14 (of 15) screws failed within the 5000 load cycles, eight due to relative tilt and six due to an absolute tilt. Analysing the cement-first technique (RSC) 13 (of 15) screws failed during the cyclic loading period—all of them because of tilting more than one degree within 50 cycles. In one screw of the LFC augmentation and two screws of the RSC augmentation the polyaxial screw head tilted prior to screw loosening, tests were stopped and the screws were not included in the analysis. The mean failure cycle for in situ augmented screws (LFC) was 3653.3 (SD 934.0) and 3723.3 (SD 560.6) for cement-first technique (RSC) (p = 0.824). Comparison of the mean failure cycles of all cementing techniques to the control showed a significant difference (p \ 0.001), while there was no significant difference in between the three augmentation techniques (p [ 0.999) (Fig. 4). Stress X-rays Stress X-rays of group 1 revealed significant less screw toggling for cemented screws (LCM) compared to the non-
Eur Spine J
Fig. 4 The control screwing technique exhibited significant lower load cycles to failure compared to all other groups. The boxplots are showing the median, the 25th and the 75th quartile of the load cycles to failure for the four groups. Statistical significant differences to the control group are marked with an asterisk
Fig. 5 The control screwing technique exhibited a significant larger toggeling after cyclic loading compared to all other groups. The boxplots are showing the median, the 25th and the 75th quartile of the screw toggeling after cyclic loading for the four groups. Statistical significant differences to the control group are marked with an asterisk
augmented control RNC (mean angle 5.2°, SD 5.4° vs. 16.1°, SD 6.0; p \ 0.001). Group 2 showed a screw toggling of 5.1° (SD 1.9) and 6.6° (SD 4.6) for in situ (LFC) and cement-first augmentation (RSC), respectively. None of the differences between the cementing techniques of group 2 were statistically significant (p = 0.282). Comparing the means of screw toggling of all of the cementing techniques (LCM, LFC, RSC) to the non-augmented control (RNC) revealed a statistical significant difference (p \ 0.001), while there was no significant difference in between the three augmentation techniques (p [ 0.999) (Fig. 5). Pullout tests Comparing the means of the maximum pullout forces after cyclic loading for screws of group 1 showed 776.0 N (SD 380.3) for cannulated augmented screws (LCM) and 168.6 N (SD 107.9) for the nonaugmented control (RNC). Statistical analysis showed a significant higher maximum pullout force for cannulated cemented screws (p \ 0.001) (Fig. 6). In group 2 mean values of 629.9 N (SD 159.1) for the in situ augmentation (LFC) and 497.5 N (SD 255.1) for the cement-first technique (RSC) were measured. None of the differences between the cementing techniques of group 2 were statistically significant (p [ 0.999) (Fig. 6). Means of all of the cementing techniques LCM, LFC, RSC revealed no significant difference in between the three
Fig. 6 The control screwing technique exhibited a significant lower pullout force after cyclic loading compared to all other groups. The boxplots are showing the median, the 25th and the 75th quartile of the ultimate pull out force after cyclic loading for the four groups. Statistical significant differences to the control group are marked with an asterisk
augmentation techniques (p [ 0.999), but documented a statistical significant difference compared to the non-augmented control RNC (p \ 0.001) (Fig. 6). Testing the maximal pullout force of the inserted screws resulted in pedicle fractures in 19 of 43 (44.2 %) of the augmented screws.
123
Eur Spine J
Discussion To investigate if a biomechanically sufficient augmentation is possible with a percutaneous technique, we carried out the present study with cyclic cranio-caudal loading of the screws. First it was found that augmentation of pedicle screws significantly increased the number of load cycles and failure load compared to the non-augmented control group. A second result is that augmentation in percutaneous technique (that is using cannulated screws with percutaneous equipment) showed no significant difference to the conventional technique. As final result we found that after cranio-caudal cyclic loading screw cut out through the superior endplate. This ‘‘windshield-wiper effect’’ is typically observed in clinical practise, underlining that this test set up reproduces clinical findings. Axial pull out tests for evaluation of the pedicle screw anchorage are easy to conduct and give a practical parameter, but it must also be considered, that they do not reflect the clinical setting [9, 19, 20] as they apply an unphysiological loading to the pedicle screw. Loosening due to fatigue loading and screw breakage are much more common reasons for screw failure. In accordance to literature we therefore simulated cranio-caudal in vivo loading and failure mechanism of pedicle screws as a more clinically applicable testing mode [15–18]. A constantly increasing load magnitude was shown to be a well suited load protocol to provoke implant loosening and osteosynthesis failure in varying bone quality and osteosynthesis stiffness for trabecular anchoring of implants and screws such as in the femoral head or the humeral head. To achieve a good transferability into clinical practice we decided to test most used screw diameters (6 mm) and length (45 mm). The same screw size was used for all vertebral bodies to avoid differences in fixation strength. On the other hand also the kind of cement being used for augmentation matters. We used a high viscosity PMMA cement. Biomechanical studies reported a higher number of load cycles and a higher bending moment to failure for PMMA augmented screws compared to screws augmented with for instance calcium triglyceride bone cement. Also high-viscosity cement may enhance the fixation effect of the screws compared to low-viscosity cement. It is reported that in lumbar spine a significant relative increase in pullout strength could be reached by using a volume of 1.0–3.0 cc. With 1.5 ml we aimed for a compromise between using enough volume to improve mechanical strength while avoiding the risks of cement leakage associated with higher cement volumes. Moreover with the injected volume of 1.5 ml, we were only slightly above the optimum filling of 75 % (about 1.03 ml) of the volume suggested for bone cement augmentation of the trajectory [21].
123
Augmentation of pedicle screws in osteoporotic bone increases the number of load cycles until failure, significantly (LCM 212.8 %, LFC 192.9 %, RSC 196.6 %). Moreover the tested augmentation techniques showed no significant difference, no matter if they are done percutaneously or not. This finding underlines, to the author’s knowledge for the first time described in biomechanical reports, that an augmentation in a percutaneous technique is feasible and as effective as with usual techniques. Although there is first report on clinical experience with short term follow up in managing elderly thoracolumbar fractures using percutaneous cemented screws [22]. The advantage of a percutaneous procedure would be for instance a less invasive approach, shorter operation time and the benefit of in situ augmented screws, moreover the autochthonous back muscles is not devascularized. However, in situ augmentation of the screws requires a higher pressure for cement application as the cannulas have a smaller diameter [23]. Some authors advocate that the technique of cement injection does matter and reported higher pullout and fatigue strength for in situ screw augmentation techniques than for prefilled technique [20, 24]. From a physical point of view, the cement is distributed according to the falling pressure gradient. So it will concentrate around the distal third of the fenestrated screw, serving the first holes first. Injection into a tapped hole can lead to a filling along the thread, for instance if cement with low viscosity is used [24]. This difference in the distribution may promote a lower force to failure. We used high viscosity cement in all cases and observed only a distribution around the distal third of the thread, regardless of the application. This may be due to the high viscosity of the cement. In addition a trend towards higher failure loads for high-viscosity cement compared to low-viscosity cement has been described. Compared to the non-augmented control group, we also found a substantial increase in the mean pull-out strength after cyclic testing (LCM 460 %, LFC 373.6 %, RSC 295.1 %). This goes well together with similar observations in literature: 119 % [8], 162 % [8], 181 % [25], 187 % [6], 206 % [25], 213 % [25], 250 % [26]. However, this might be due differences in preceding cyclic loading, in bone mineral density, used screw types, filling volume and augmentation material and technique. In 44.2 % (19 of 43) of the augmented screws the failure mode was pedicle fracture with the cement still attached to the screw and no axial screw pullout as for non-augmented pedicle screws. This failure mode of augmented pedicle screws in pullout tests was already reported by other authors, e.g. Sarzier et al. [25] reports on 9 of 21 (42.9 %) and Bullmann et al. [23] on 10 of 23 (43.5 %) pedicle fractures of augmented screws in their trials.
Eur Spine J
In the present investigation with cranio-caudal cyclic loading screw cut out through the superior endplate and the ‘‘windshield-wiper effect’’, which is usually observed in clinical practice, could be reproduced. This underlines the assumption that our loading model is closer to reality. On the other hand it is difficult to determine the physiological load magnitude for the cranio-caudal cyclic testing. In measurements with a telemetrized internal fixator of 10 patients Rohlmann et al. [27, 28] reported axial craniocaudal peak force magnitudes in everyday activities between 200 and 250 N. With the constantly increasing load magnitude during cyclic loading the mean failure load of the non-cemented control group (RNC) was 239 N, while the mean failure load for all augmented screws was 453, 415 and 422 N for the LCM, LFC and RSC group, respectively. However, it should be considered that in the present study only osteoporotic vertebrae (mean age of 84.3 years) were used while the active patients in the study by Rohlmann et al. had a mean age of 50.6 years. Combining the in vivo measurements with the results of the present study, it can be assumed, that in patients with reduced bone quality pedicle screw loosening might occur during everyday activity, while augmentation raises the failure load above the forces occurring during everyday activities.
Limitations of the study First, the pull out test was done after cyclic loading with loosened pedicle screws. Therefore, the absolute values should be regarded with this limitation. Nevertheless this limitation occurred to all screws in a similar way, thus the proportional trend could be transferred to clinical practice. A second limitation in this study is that it was conducted on cadavers, so we have no in vivo situation. Cadaver models have of course their inherent limitations. They cannot replicate biological factors such as screw ingrowth and bone remodeling around screws. Osseous resistance, muscular loading and in vivo loading of structures can only be approximated.
Conclusion Augmentation of pedicle screws significantly increased the number of load cycles and failure load compared to the non-augmented control group. The augmentation technique in situ with perforated screws or cement first even percutaneously, had no effect on the loosening of the pedicle screws. By the cranio-caudal cyclic loading screw cut out through the superior endplate and the typical ‘‘windshield-wiper effect’’, typically observed in clinical practice, could be reproduced. The augmentation technique, even
percutaneously, had no effect on the failure of the pedicle screws. Conflict of interest The study was supported by DePuy Spine Germany (receiving the implants, cement, and laboratory cost, no wages have been paid), but none of the authors has any potential conflict of interest.
References 1. Office GFS (2002–2010) Total cost of illness in millions of Euro. Classification: years, gender, ICD10, provider (Primary source: Federal Statistical Office, Total Health Expenditure, Statistics). In., Berlin 2. Masala S, Lunardi P, Fiori R, Liccardo G, Massari F, Ursone A, Simonetti G (2004) Vertebroplasty and kyphoplasty in the treatment of malignant vertebral fractures. J Chemother 16(Suppl 5):30–33 3. Melton LJ 3rd, Thamer M, Ray NF, Chan JK, Chesnut CH 3rd, Einhorn TA, Johnston CC, Raisz LG, Silverman SL, Siris ES (1997) Fractures attributable to osteoporosis: report from the National Osteoporosis Foundation. J Bone Mineral Res 12:16–23. doi:10.1359/jbmr.1997.12.1.16 4. Papaioannou A, Watts NB, Kendler DL, Yuen CK, Adachi JD, Ferko N (2002) Diagnosis and management of vertebral fractures in elderly adults. Am J Med 113:220–228 5. Becker S, Chavanne A, Spitaler R, Kropik K, Aigner N, Ogon M, Redl H (2008) Assessment of different screw augmentation techniques and screw designs in osteoporotic spines. Eur Spine Jy 17:1462–1469. doi:10.1007/s00586-008-0769-8 6. Burval DJ, McLain RF, Milks R, Inceoglu S (2007) Primary pedicle screw augmentation in osteoporotic lumbar vertebrae: biomechanical analysis of pedicle fixation strength. Spine 32:1077–1083. doi:10.1097/01.brs.0000261566.38422.40 7. Chen LH, Tai CL, Lai PL, Lee DM, Tsai TT, Fu TS, Niu CC, Chen WJ (2009) Pullout strength for cannulated pedicle screws with bone cement augmentation in severely osteoporotic bone: influences of radial hole and pilot hole tapping. Clin Biomech (Bristol, Avon) 24:613–618. doi:10.1016/j.clinbiomech.2009.05. 002 8. Frankel BM, D’Agostino S, Wang C (2007) A biomechanical cadaveric analysis of polymethylmethacrylate-augmented pedicle screw fixation. J Neurosurg Spine 7:47–53. doi:10.3171/SPI-07/ 07/047 9. Zindrick MR, Wiltse LL, Widell EH, Thomas JC, Holland WR, Field BT, Spencer CW (1986) A biomechanical study of intrapeduncular screw fixation in the lumbosacral spine. Clin Orthop Relat Res (203):99–112 10. Rohlmann A, Bergmann G, Graichen F (1997) Loads on an internal spinal fixation device during walking. J Biomech 30:41–47 11. Rohlmann A, Bergmann G, Graichen F, Weber U (1997) Comparison of loads on internal spinal fixation devices measured in vitro and in vivo. Med Eng Phys 19:539–546 12. Rohlmann A, Graichen F, Bergmann G (2000) Influence of load carrying on loads in internal spinal fixators. J Biomech 33:1099–1104 13. Wu JC, Chen CS, Yip SW, Hsu ML (2012) Stress distribution and micromotion analyses of immediately loaded implants of varying lengths in the mandible and fibular bone grafts: a three-dimensional finite element analysis. Int J Oral Maxillofac Implants 27:e77–e84 14. Felsenberg D, Gowin W (1999) Bone densitometry by dual energy methods. Der Radiologe 39:186–193
123
Eur Spine J 15. Windolf M, Maza ER, Gueorguiev B, Braunstein V, Schwieger K (2010) Treatment of distal humeral fractures using conventional implants. Biomechanical evaluation of a new implant configuration. BMC Musculoskelet Disord 11:172. doi:10.1186/ 1471-2474-11-172 16. Unger S, Erhart S, Kralinger F, Blauth M, Schmoelz W (2012) The effect of in situ augmentation on implant anchorage in proximal humeral head fractures. Injury 43:1759–1763. doi:10. 1016/j.injury.2012.07.003 17. Kiner DW, Wybo CD, Sterba W, Yeni YN, Bartol SW, Vaidya R (2008) Biomechanical analysis of different techniques in revision spinal instrumentation: larger diameter screws versus cement augmentation. Spine 33:2618–2622. doi:10.1097/BRS.0b013 e3181882cac 18. Goel VK, Winterbottom JM, Weinstein JN (1994) A method for the fatigue testing of pedicle screw fixation devices. J Biomech 27:1383–1388 19. Law M, Tencer AF, Anderson PA (1993) Caudo-cephalad loading of pedicle screws: mechanisms of loosening and methods of augmentation. Spine 18:2438–2443 20. Kueny RA, Kolb JP, Lehmann W, Puschel K, Morlock MM, Huber G (2014) Influence of the screw augmentation technique and a diameter increase on pedicle screw fixation in the osteoporotic spine: pullout versus fatigue testing. Eur Spine J. doi:10. 1007/s00586-014-3476-7 21. Fan H, Zhang R, Shen C, Dong F, Li Y, Song P, Gong C, Wang Y (2000) The biomechanical properties of pedicle screw fixation combined with trajectory bone cement augmentation in
123
22.
23.
24.
25.
26.
27.
28.
osteoporotic vertebrae. J Spin Disord Tech. doi:10.1097/BSD. 1090b1013e3182a14870 Pesenti S, Blondel B, Peltier E, Adetchessi T, Dufour H, Fuentes S (2014) Percutaneous cement-augmented screws fixation in the fractures of the aging spine: is it the solution? BioMed Res Int 2014:610675. doi:10.1155/2014/610675 Bullmann V, Schmoelz W, Richter M, Grathwohl C, Schulte TL (2010) Revision of cannulated and perforated cement-augmented pedicle screws: a biomechanical study in human cadavers. Spine 35:E932–E939. doi:10.1097/BRS.0b013e3181c6ec60 Choma TJ, Pfeiffer FM, Swope RW, Hirner JP (2012) Pedicle screw design and cement augmentation in osteoporotic vertebrae: effects of fenestrations and cement viscosity on fixation and extraction. Spine 37:E1628–E1632. doi:10.1097/BRS.0b013e3182740e56 Sarzier JS, Evans AJ, Cahill DW (2002) Increased pedicle screw pullout strength with vertebroplasty augmentation in osteoporotic spines. J Neurosurg 96:309–312 Cook SD, Salkeld SL, Stanley T, Faciane A, Miller SD (2004) Biomechanical study of pedicle screw fixation in severely osteoporotic bone. Spine J 4:402–408. doi:10.1016/j.spinee.2003. 11.010 Rohlmann A, Graichen F, Weber U, Bergmann G (2000) 2000 Volvo Award winner in biomechanical studies: monitoring in vivo implant loads with a telemeterized internal spinal fixation device. Spine 25:2981–2986 Rohlmann A, Bergmann G, Graichen F (1999) Loads on internal spinal fixators measured in different body positions. Eur Spine J 8:354–359
