If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Midfoot and subtalar arthrodesis surgeries are performed to correct foot deformities and relieve arthritic pain. These procedures often employ intramedullary (IM) devices. The aim of the present study was to evaluate the biomechanical performance of a sustained dynamic compression (SDC) IM device compared to mechanically static devices in withstanding the effects of simulated bone resorption. Mechanically static and SDC IM devices were implanted in simulated bone blocks (n = 5/device). Compressive loads were measured with a custom-made mechanism to simulate bone resorption. The construct bending stiffness was determined from a 4-point bend test. Resorption was simulated by cutting a 1 mm or 2 mm gap in the midpoint of each construct and repeating the loading (n = 6/device). Initial compressive loads after device insertion were greater in the SDC IM devices when compared to the static devices (p < .01). The SDC device was able to sustain compression from 2 mm to 5.5 mm of simulated resorption depending upon device length, while the static devices lost compression within 1 mm of simulated resorption regardless of implant length (p < .001). In the 4-point bend test, the SDC device maintained its bending stiffness during simulated resorption whereas the static device displayed a significant loss in bending stiffness after 1 mm of simulated resorption (p < .001). The SDC device exhibited a significantly higher bending stiffness than the static device (p < .001). The SDC IM device demonstrated superior biomechanical performance during simulated resorption compared to static devices (p < .001). In conclusion, the ability of SDC IM devices to maintain construct stability and sustain compression across the fusion site while adapting to bone resorption may lead to greater fusion rates and overall quicker times to fusion than static IM devices. Surgeons who perform midfoot and subtalar arthrodesis procedures should be aware of a device's ability to sustain compression, especially in cases where bone resorption and joint settling are prevalent postoperatively.
). Common indications for subtalar fusion include posttraumatic subtalar arthritis, primary osteoarthritis, talocalcaneal coalition, and inflammatory joint disease (
). In addition, medial column fusion is often performed for indications such as Charcot neuroarthropathy, posttraumatic degenerative joint disease, rheumatoid arthritis, and peritalar subluxation deformity (
). These procedures often utilize internal fixation devices such as screws, plates, staples, and intramedullary (IM) nails; external fixation; or a combination of external and internal fixation (
). Currently, internal fixation with IM devices, specifically screws, beams, and bolts, is often the chosen fixation method in subtalar and medial column arthrodesis. This method promotes bone healing and fusion as it has lower complication rates, increased patient tolerance, and is less invasive than external fixation alternatives (
Total ankle replacement conversion to tibiotalocalcaneal arthrodesis with bulk femoral head allograft and pseudoelastic intramedullary nail providing sustained joint compression.
). Most IM devices are mechanically static in nature which often leaves them unable to respond to changes in loading caused by postoperative bone resorption, joint settling, or bone relaxation, resulting in a loss of compression over time (
Benchtop comparison of a novel dynamic compression screw to a standard cortical screw: compression integrity and gap size over time during simulated resorption.
Total ankle replacement conversion to tibiotalocalcaneal arthrodesis with bulk femoral head allograft and pseudoelastic intramedullary nail providing sustained joint compression.
Benchtop comparison of a novel dynamic compression screw to a standard cortical screw: compression integrity and gap size over time during simulated resorption.
The key biomechanical principles of bone healing are stability, lack of micromotion, load-sharing, sustained compression, high contact area, and rigidity (
Benchtop comparison of a novel dynamic compression screw to a standard cortical screw: compression integrity and gap size over time during simulated resorption.
Intramedullary medial column support with the Midfoot Fusion Bolt (MFB) is not sufficient for osseous healing of arthrodesis in neuroosteoarthropathic feet.
). Comorbidities such as smoking, poor local vascularity, diabetes, obesity, and compliance with weightbearing make patients much more likely to experience nonunion (
). One retrospective study found that smokers were 3.8 times more likely to have nonunion occur than nonsmokers, and diabetic patients were 18.7 times more likely to have a nonunion (
). In patients with poor healing potential or where traditional fixation methods may have failed, there is a need for an IM device that provides sustained compression (
). A sustained dynamic compression (SDC) IM device has been designed to aid in achieving successful fusion with its unique, internal compressive nitinol element that sustains compression by adapting to changes in loading caused by joint settling or bone resorption (
). Shape recovery is possible through either of these mechanisms. The pseudoelastic effect occurs via an applied deformation on the stable austenite phase to the detwinned martensite phase, which is accompanied by a large reversible strain at a near constant stress. Upon release of this applied load, the material reverts to the stable austenite phase. Most current nitinol staples utilize this shape recovery mechanism. The shape-memory effect occurs when the martensite phase is stable at ambient temperature. The martensite phase allows for large strains as it deforms from the twinned martensite to detwinned martensite state. This deformed state is held until heating above the austenite phase transformation occurs, upon which the material recovers its shape back to the parent stable austenite phase. Early nitinol staples required deformation and storage at cold temperatures, and heating activated these staples to recover the original shape (
). These studies have established that SDC devices provide more stability than static IM devices due to the continuous compression and load sharing between the SDC device and bone (
). A retrospective study by Steele, et.al. found that SDC devices lead to increased fusion rates and significantly quicker times to fusion than static devices, despite the SDC patient population having more risk factors for nonunion (
Simultaneous intraoperative measurement of cadaver ankle and subtalar joint compression during arthrodesis with intramedullary nail, screws, and tibiotalocalcaneal plate.
). However, there is a lack of studies analyzing the biomechanical performance of SDC IM devices for subtalar and medial column arthrodesis and biomechanical performance of these devices in response to bone resorption and joint settling. The primary aim of this study was to examine the effect of simulated resorption on the biomechanical performance of static and SDC IM devices, specifically in sustaining compressive loads and maintaining bending stiffness during simulated resorption. SDC and static IM devices underwent axial compression and 4-point bend testing to evaluate the mechanical performance in synthetic bone blocks with simulated bone resorption.
Materials and Methods
Intramedullary Device Design
Two different types of IM devices were tested using the methods described in this paper. Static headless compression devices (Wright Medical Group N.V., SALVATION™ Beam) of 7 mm diameter and lengths of 95 mm and 140 mm were compared to SDC devices (MedShape, Inc. - acquired by DJO®, DynaNail Mini®) of 8 mm diameter and lengths of 70 mm, 80 mm, 90 mm, 120 mm, 130 mm, and 140 mm. For the axial compression test, the 95 mm and 140 mm static devices were tested along with all lengths of the SDC devices; for the 4-point bend test method, the 140 mm static device and the 140 mm SDC device were tested. The static devices consist of a cannulated, unthreaded working length, with threads of differing pitch at either end. The static device tested featured 34 mm of cancellous threading and 11 mm of cortical threading, as seen in Fig. 1.
Fig. 1Wright Salvation Beam; intramedullary device with cortical and cancellous threads.
The SDC device featured an unthreaded outer body, with a transverse screw hole on the proximal end and a slotted transverse screw hole on the distal end of the outer body to allow for bone fixation, as seen in Fig. 2. A nitinol element stretched to 2.2 mm, 2.8 mm, 3.4 mm, 4.5 mm, 5.1 mm, or 5.6 mm was used to maintain compression for the 70 mm, 80 mm, 90 mm, 120 mm, 130 mm, and 140 mm implants, respectively. This was achieved by fixating the implant with a transverse screw through the slotted sliding element affixed to the end of the nitinol element. In the case of postoperative bone resorption or joint settling, the stretched compressive element will automatically recover, and the construct thus sustains compression across the fusion site while it does so. An environmental chamber (Electro-Tech Systems Model 5506-00) was set to maintain temperatures at 37°C to replicate internal body temperature. Both IM devices were placed in the environmental chamber for 15 minutes prior to implantation to allow the devices to fully warm up to 37°C.
Fig. 2DynaNail Mini; intramedullary nail incorporating pseudoelastic nitinol element.
To simulate bone, rigid polyurethane foam blocks (Sawbones®, Pacific Research Laboratories) were used during testing to minimize bone quality variability. Foam with density of 20 pounds per cubic foot (PCF) was used to simulate bone of normal bone density. Multiple studies have shown that polyurethane foam blocks are good substitutes for in vitro testing because the blocks have similar compressive Young's modulus and yield strength values to human cancellous bone (
). Additionally, the usage of Sawbone blocks with uniform foam reduces variability and allows for a more precise comparison of mechanical characteristics between IM designs than if using cadaveric specimens (
). The Sawbone blocks were cut down to size according to the test method and implant length. For the axial compression test, the Sawbone blocks were cut to a height of 40 mm, with a cross-section of 45 mm by 45 mm. For the 4-point bending test, the Sawbone blocks were cut with a cross-section of 40 mm by 40 mm, with a length of 70 mm. Each of the prepared Sawbone blocks had pilot holes drilled through the center of the face of the cross-section following the manufacturer's recommended surgical technique.
Implantation of Devices for Axial Compression Test
For the axial compression test, the static device was inserted through the prepared synthetic bone blocks with the donut load cell and custom-made resorption stage in between the Sawbone blocks to form the test construct. The donut load cell (Transducer Techniques® Load Cell, Model THD-2K-W) was a load cell with a hollow center to allow for IM devices to span between the Sawbone blocks. The custom-made resorption setup is a set of two screw-driven parallel aluminum plates. Resorption was simulated by turning the drive screws and lowering the plates, thus creating slack in the construct. The axial compression test set-up is depicted in Fig. 3. The Sawbone-device construct was placed back in the environmental chamber for at least 15 minutes until it reached 37°C. The SDC device was inserted into the prepared Sawbone blocks with the donut load cell and resorption stage in between the Sawbone blocks. The SDC device was affixed with two 4 mm by 40 mm transverse screws, with one each through the transverse screw hole and the slotted screw hole. Before complete fixation through the slotted screw hole, manual compression was applied with the manual compression wheel. The Sawbone-device construct was placed back in the environmental chamber, and after 15 minutes at 37°C, the nitinol element was released, and the Sawbone-device construct was removed from the targeting frame.
During the axial compression testing, all device constructs were tested inside the environmental chamber for the entirety of the testing process. Once the implants had been fully inserted, the initial distance between the two plates of the resorption stage was measured with digital calipers and the initial compressive load was recorded. The test method used in this study has previously been established to measure compressive forces of TTC fixation devices with simulated bone resorption (
). Resorption was simulated by lowering the upper plate of the custom stage by turning the 2 screws on the resorption stage by quarter turns. After each quarter turn, the distance between the two plates of the stage and the compressive loads on the donut load cell were measured. The distance between the two plates and the compressive loads measured after each quarter turn until no further compressive loads were experienced by the load cell.
Implantation of IM Devices for 4-Point Bend Test
The implantation process for the IM devices for the 4-point bend test is the same as the implantation process for the axial compression test; however, the resorption stage and load cell are not placed between the Sawbone blocks.
4-Point Bend Testing
The devices were tested with an Instron Materials Testing System (Instron® 5567) using custom 4-point bending fixtures set-up per ASTM F1264, as depicted in Fig. 4. The SDC device was tested in an Instron thermal chamber (Instron SFL Heatwave Thermal Chamber) set to maintain temperatures at 37°C to replicate internal body temperature. Once the IM devices were prepared and implanted into the Sawbone blocks, they were centered onto the custom fixtures and preloaded with 5N of force. The devices were quasi-statically loaded by cyclically loading from 6.8 N to 68.8 N of force at a displacement rate of 1mm/min for 10 cycles in order to mimic worst-case scenarios of slow-walking loads experienced in the midfoot. The load values were calculated from the average force along the medial column of a 95th percentile male, representing the worst-case loading scenario (
). After each set of 10 cycles, reciprocating saw blades were used to create a 1mm gap at the interface between the two blocks of Sawbone around the device to simulate bone resorption. Custom cutting guides were used to ensure precision in sawing gaps to simulate bone resorption. The quasi-static testing was repeated for up to 2 mm of simulated resorption for each device. The linear region of the load-displacement data from the 10th cycle was analyzed to determine bending stiffness (n = 6/group).
Fig. 44-point bend test set-up, as per ASTM F1264.
Data is reported as mean ± st.dev. A one-way ANOVA test with Tukey's post-hoc test was performed to statistically compare results between groups (α = 0.05; Igor Pro, v.6.37, Wave Metrics, 2014).
Results
Axial Compression Testing
Compressive force curves as a function of simulated resorption distance for the SDC and static devices were generated, as shown in Fig. 5A and B. During the axial compression test, the static devices had average initial compression of 186.4 ± 6.7 N and 299.9 ± 47.5 N for the 95 mm (Fig. 6A) and 140 mm (Fig. 6B) implant lengths, respectively, and lost all compression within 1 mm of simulated bone resorption regardless of the implant length. Regardless of device length, the SDC device provided significantly greater initial compressive loads than the static devices (p < .01) (Fig. 6A, and B). The SDC devices had average initial compressions of 309.1 N ± 23.0 N, 292.1 N ± 6.5 N, 311.7 N ± 21.1 N, 380.6 N ± 20.9 N, 403.9 N ± 35.9 N, and 389.9 N ± 13.6 N for the 70 mm, 80 mm, 90 mm, 120 mm, 130 mm, and 140 mm lengths, respectively (Fig. 6A, and B).
Fig. 5(A) SDC and static devices’ (subtalar fusion lengths: 70 mm-95 mm) compressive force as a function of simulated resorption distance. (B) SDC and static devices’ (medial column fusion lengths: 120 mm-140 mm) compressive force as a function of simulated resorption distance.
Fig. 6(A) SDC and static devices (subtalar fusion lengths: 70 mm-95 mm) initial compressive force at 0 mm of simulated bone resorption. (B) SDC and static devices (medial column fusion lengths: 120 mm-140 mm) initial compressive force at 0mm of simulated bone resorption. *, **, † signify p < .01.
For the 95 mm static device, the compression dropped to 43.2 N after 0.3 mm of simulated bone resorption and lost all compression within 0.65 mm of simulated bone resorption (Fig. 7A). For the 140 mm static device, the compression dropped to 44.0 N after 0.5 mm of simulated resorption and lost all compression within 0.78 mm of simulated bone resorption (Fig. 7B). For the 70 mm length SDC device, the compression dropped to 224 N after 0.5 mm of simulated bone resorption and continued to sustain compression until 2.1 mm of simulated resorption had been reached (Fig. 7A). Similar levels of sustained compression were seen for the other SDC devices, with compression achieved through 2.98 mm, 3.70 mm, 4.51 mm, 5.01 mm, and 5.48 mm of simulated bone resorption for the 80 mm, 90 mm, 120 mm, 130 mm, and 140 mm implant lengths, respectively (Fig. 7A, and B). The forces generated by the SDC device were sustained for greater amounts of simulated bone resorption than the static devices (p < .001; Fig. 7A, and B). Additionally, as the length of the SDC devices increased, the amounts of simulated bone resorption significantly increased (p < .05) (Fig. 7A, and B).
Fig. 7(A) SDC and static devices (subtalar fusion lengths: 70 mm-95 mm) amounts of simulated bone resorption through which compression was sustained. (B) SDC and static devices (medial column fusion lengths: 120 mm-140 mm) amounts of simulated bone resorption through which compression was sustained. *, **, #, ##, †, ‡ signify p < .001; *** signify p < .05.
During the 4-point bending test, the static device had an average bending stiffness of 608 ± 40 N/mm at 0 mm of simulated bone resorption. At 1 mm of simulated bone resorption, the average bending stiffness of the static device was 228 ± 50 N/mm, and at 2 mm of resorption the average bending stiffness was 202 ± 44 N/mm (Fig. 8).
Fig. 8SDC and static devices’ bending stiffness vs simulated resorption distance; *,**,#,† signify p < .001.
The SDC device had an average bending stiffness of 578 ± 58 N/mm at 0 mm of simulated bone resorption. At 1 mm of simulated bone resorption, the average bending stiffness of the SDC device was 503 ± 20 N/mm, and at 2 mm of resorption, the average bending stiffness was 548 ± 52 N/mm (Fig. 8).
There was no statistically significant difference between the static and the SDC IM devices at 0 mm of resorption. There was a statistically significant decrease in the bending stiffness of the static device during simulated resorption of 1 mm and 2 mm compared to 0 mm (p < .001). There was no statistically significant difference in bending stiffness for the SDC device during simulated resorption of 1 mm and 2 mm compared to 0 mm (p = .17 for 0 mm vs 1 mm; p = .21 for 0 mm vs 1 mm; p = .99 for 1 mm vs 2 mm for SDC). At both 1 mm and 2 mm of simulated resorption, the SDC IM device demonstrated statistically significantly higher bending stiffness when compared to the static device at the same resorption level (p < .001).
Discussion
Fusion rates in midfoot arthrodesis range between 54.4% and 100%, based on fusion technique, fusion hardware, method of fusion assessment, and any underlying comorbidities that the patient may possess (
Total ankle replacement conversion to tibiotalocalcaneal arthrodesis with bulk femoral head allograft and pseudoelastic intramedullary nail providing sustained joint compression.
Total ankle replacement conversion to tibiotalocalcaneal arthrodesis with bulk femoral head allograft and pseudoelastic intramedullary nail providing sustained joint compression.
Benchtop comparison of a novel dynamic compression screw to a standard cortical screw: compression integrity and gap size over time during simulated resorption.
). Furthermore, bone resorption reduces construct compressive stability and changes load-sharing properties. Tibiotalocalcaneal (TTC) fusion case studies using a SDC IM nail with either a bulk femoral head allograft or with Charcot patients showed nitinol element recovery suggesting between 3.1 mm and 5.9 mm of bone resorption and joint settling (
Total ankle replacement conversion to tibiotalocalcaneal arthrodesis with bulk femoral head allograft and pseudoelastic intramedullary nail providing sustained joint compression.
). In addition, studies on bone viscoelasticity and stress relaxation found that 43.35% and 65.77% of compressive load across tested bone cylinders was lost over the first three minutes and 30 minutes, respectively, postloading (
). The same viscoelasticity study was done with a lag screw and found that 20% of compression was lost over the first 34 seconds postimplantation. After retightening the screws, there was still a further average loss in compression of 25.5% after 30 minutes (
). Thus, multiple factors contribute to the loss of device compression, and these SDC devices can adapt to clinically relevant amounts of bone resorption and joint settling.
Previous biomechanical studies have assessed the ability of SDC devices in sustaining compression during resorption and joint settling. Anderson et al. evaluated a SDC IM nail in terms of compression and load sharing (
). It was concluded that the SDC nail helps to prevent stress shielding, therefore allowing a greater amount of load during gait to be taken by the bone than by the nail. In a different study, Yakacki et al. measured compression forces of TTC fixation devices with simulated bone resorption in 2 compressive IM devices, one SDC and one static device. These were tested in synthetic and cadaveric bone using a custom mechanism to simulate resorption (
). It was found that the static IM nails lost 90% or more of initial compression within 1mm of simulated resorption, and the SDC device demonstrated sustained compressive loads for 7 mm of simulated resorption (
). In this study, the primary aim was to examine the effect of simulated resorption on the biomechanical performance of static and SDC IM devices, specifically in sustaining compressive loads and maintaining bending stiffness during simulated resorption. Overall, our results showed that the SDC device maintains bending stiffness during simulated bone resorption compared to the static device because of the addition of the pseudoelastic nitinol element (p < .001). In addition to providing significantly larger initial compressive forces (p < .01), the forces generated by the SDC device were sustained for significantly greater amounts of simulated bone resorption than the static devices (p < .001). The increased initial compressive loads before simulated resorption for the SDC device are due to the SDC device's delivery frame and its ability to provide intraoperative compression via the manual compression wheel. When choosing a SDC device for a given case, maximizing SDC device length should be considered. As the SDC device length increases, the SDC device will adapt to more bone resorption due to having a longer nitinol element (p < .05).
Regarding the bending stiffness of the IM devices, the static IM device experienced significant decreases in bending stiffness as bone resorption was simulated as shown in Fig. 8 (p < .001). However, the SDC IM device did not see significant differences in bending stiffness when bone resorption was simulated. This can be explained by the presence or lack thereof, respectively, of a gap in the Sawbone-device construct. In the static case, the simulated resorption left a 1 mm or 2 mm gap in the Sawbone-device construct. As the device was mechanically static, the device was unable to close any gap that formed in the construct after implantation. This gap formation led to the decrease in bending stiffness observed as bone resorption was simulated. In contrast, the SDC device's unique internal pseudoelastic nitinol element allowed for sustained compression postimplantation. This allowed the SDC device to immediately close any gaps in the Sawbone-device construct when resorption was simulated, unlike the static device. As a result, no significant changes in bending stiffness were observed in the SDC device as bone resorption was simulated.
) evaluated the function of Charcot-specific IM implants designed for midfoot reconstruction. In that study, multiple static devices were tested to examine biomechanical properties, specifically testing 4-point bending, cantilever bending, and thread pullout resistance. Their study utilized a similar test set-up to the 4-point bend testing conducted in our study; however, their devices were tested without implantation in any bone analog or cadaveric constructs. Testing device performance in bone analog blocks or cadaveric specimens is critical to better replicate the in vivo response of the device constructs in response to loading. In our study, the loads were distributed across both the IM device and bone analog. In a traditional 4-point bend set-up without implantation, the forces applied throughout the testing are applied onto the device only. This means the test is unable to assess how the applied loads are distributed across the device and bone and analyze how loads are handled by the bone-device construct overall. In our study, by implanting the devices into simulated bone prior to 4-point bend testing, we can assess how changes in the loads would be handled by the device in conditions that better reflect in vivo loading. Additionally, bone resorption reduces the compressive stability of device constructs, thus changing the load-sharing properties of device constructs (
). Therefore, we suggest that to analyze the effects of simulated resorption on biomechanical performance, the devices must be implanted into bone analogs or cadaveric bone.
The main limitation of this study was that it was an in vitro biomechanical study, thus future in vivo clinical studies will be needed to confirm our results and conclusion. Another limitation of this study was that synthetic bone blocks were used to simulate human bone, as opposed to cadaveric bone or an in vivo study. However, our method was based upon work by Yakacki et al. (
), which performed a similar test of compressive load during simulated bone resorption using both synthetic Sawbone blocks and cadaveric specimens. They noted that the greatest drawback to using synthetic bone blocks is the inability to match the exact compressive response of real bone. However, they affirmed that both Sawbone models and cadaveric bone constructs can undergo resorption testing and maintain compressive loads across the construct sites. They concluded that because synthetic Sawbone is stiffer and less viscous than cadaveric specimens, testing with synthetic bone blocks represents “best-case” scenarios for testing of IM devices. However, since this test study used synthetic bone blocks of regular size and shapes, they may not accurately depict how the differing shapes and surfaces in live human bone may affect the absolute magnitude of compression applied upon the fusion site by the IM devices. Furthermore, the simulated resorption test was meant to mimic long-term bone resorption that happens over the timeframe of several months, thus it was conducted at 37°C to show the behavior of the SDC device at body temperature. Initial intraoperative compression in a patient may vary in magnitude due to bone quality, operating temperature, and bone's viscoelastic response.
In conclusion, this study suggests that the SDC device can provide sustained compression in response to bone resorption and/or joint settling. The ability of SDC IM devices to maintain construct stability and sustain compression across the fusion site while adapting to bone resorption may lead to greater fusion rates and overall quicker times to fusion than static IM devices. Future studies on SDC IM technology in midfoot arthrodesis should assess biomechanical response to joint settling, bone viscoelasticity, and bone resorption in a cadaver model or in vivo.
References
Tuijthof GJ
Beimers L
Kerkhoffs GM
Dankelman J
Dijk CN.
Overview of subtalar arthrodesis techniques: options, pitfalls and solutions.
Total ankle replacement conversion to tibiotalocalcaneal arthrodesis with bulk femoral head allograft and pseudoelastic intramedullary nail providing sustained joint compression.
Benchtop comparison of a novel dynamic compression screw to a standard cortical screw: compression integrity and gap size over time during simulated resorption.
Intramedullary medial column support with the Midfoot Fusion Bolt (MFB) is not sufficient for osseous healing of arthrodesis in neuroosteoarthropathic feet.
Simultaneous intraoperative measurement of cadaver ankle and subtalar joint compression during arthrodesis with intramedullary nail, screws, and tibiotalocalcaneal plate.
Financial Disclosure: Safranski and Dupont are paid employees of DJO.
Conflicts of Interest: Safranski reports stock ownership and other compensation from MedShape-acquired by DJO during the conduct of this study and outside the submitted work. Dupont reports stock ownership and other compensation from MedShape-acquired by DJO during the conduct of this study and outside the submitted work. Wong and Beals have nothing to disclose.
00084-9&id=gr1.jpg){kind=link}
00084-9&id=gr2.jpg){kind=link}
00084-9&id=gr3.jpg){kind=link}
00084-9&id=gr4.jpg){kind=link}
00084-9&id=gr5.jpg){kind=link}
00084-9&id=gr6.jpg){kind=link}
00084-9&id=gr7.jpg){kind=link}
00084-9&id=gr8.jpg){kind=link}