Abstract
Background: Retrograde intramedullary nailing is an established procedure for tibiotalocalcaneal arthrodesis. The goal of this study was to evaluate the effects of angle-stable locking or compressed angle-stable locking on the initial stability of the nails and on the behavior of the constructs under cyclic loading conditions.
Methods: Tibiotalocalcaneal arthrodesis was performed in fifteen third-generation synthetic bones and twenty-four fresh-frozen cadaver legs with use of retrograde intramedullary nailing with three different locking modes: a Stryker nail with compressed angle-stable locking, a Stryker nail with angle-stable locking, and a statically locked Biomet nail. Analyses were performed of the initial stability of the specimens (range of motion) and the laxity of the constructs (neutral zone) in dorsiflexion/plantar flexion, varus/valgus, and external rotation/internal rotation. Cyclic testing up to 100,000 cycles was also performed. The range of motion and the neutral zone in dorsiflexion/plantar flexion at specific cycle increments were determined.
Results: In both bone models, the intramedullary nails with compressed angle-stable locking and those with angle-stable locking were significantly superior, in terms of a smaller range of motion and neutral zone, to the statically locked nails. The compressed angle-stable nails were superior to the angle-stable nails only in the synthetic bone model, in external/internal rotation. Cyclic testing showed the nails with angle-stable locking and those with compressed angle-stable locking to have greater stability in both models. In the synthetic bone model, compressed angle-stable locking was significantly better than angle-stable locking; in the cadaver bone model, there was no significant difference between these two locking modes. During cyclic testing, five statically locked nails in the cadaver bone model failed, whereas one nail with angle-stable locking and one with compressed angle-stable locking failed.
Conclusions: Regardless of the bone model, the nails with angle-stable or compressed angle-stable locking had better initial stability and better stability following cycling than did the nails with static locking.
Clinical Relevance: Angle-stable locking of retrograde nails used for tibiotalocalcaneal arthrodesis can enhance construct stability in the hindfoot and may reduce loosening, which may help to improve the clinical outcome.
Tibiotalocalcaneal arthrodesis may be performed with a variety of techniques. Apart from screws, the main devices that have been described are blade-plates, external fixators, and intramedullary nails1,2. As long ago as 1962, Küntscher3 described closed retrograde intramedullary tibiotalocalcaneal arthrodesis with use of an unlocked nail. Following the advent of nail locking, tibiotalocalcaneal arthrodesis with intramedullary implants came to be performed chiefly with retrograde femoral nails or retrograde ankle arthrodesis nails, with the locking screws arranged in the coronal plane. Most of the relevant biomechanical studies in the literature have dealt with procedures performed with use of these so-called first-generation nails2,4-6.
The so-called second-generation nails provide special locking options such as posteroanterior locking along the longitudinal axis of the calcaneus as well as external compression devices to encourage fixation. Mann et al.7 found a retrograde nail with a posterior-to-anterior interlocking screw passed through the calcaneus in a longitudinal fashion to provide significantly more rotational stiffness than the same nail construct with a conventional transverse calcaneal screw (p < 0.036). Berson et al.1 confirmed fusion-site compression by a nail with an external compression mechanism. Our group demonstrated the effect of compression on the initial stability of a tibiotalocalcaneal arthrodesis done with an intramedullary nail in a synthetic bone model8. However, that study did not answer the question of the applicability of the results to the human model and it did not provide evidence of the behavior of the fixation under cyclic loading conditions.
Since our earlier study8, an intramedullary nail for tibiotalocalcaneal arthrodesis that provides posterior-to-anterior interlocking in the calcaneus, includes internal and external devices for optional compression of the joint surfaces, and provides angle-stable locking in the hindfoot has been developed. To our knowledge, there have been no biomechanical studies of angle-stable intramedullary nails for tibiotalocalcaneal arthrodesis in the literature. This study was performed to establish the effects of angle-stable locking or compressed angle-stable locking on the initial stability of the nails and on the behavior of the constructs under cyclic loading conditions.
In this study, tibiotalocalcaneal arthrodesis was performed in human cadaver bones and in synthetic bones with use of two different retrograde nails and a total of three different locking modes. The three patterns (Fig. 1) were (1) a nail with compressed angle-stable locking (T2 Ankle Arthrodesis Nail [AAN]; Stryker Osteosynthesis, Schönkirchen/Kiel, Germany); (2) a nail with angle-stable locking (T2 Ankle Arthrodesis Nail [AAN]; Stryker Osteosynthesis); and (3) a nail with static locking (Biomet Ankle Arthrodesis Nail; Biomet, Berlin, Germany). The Biomet Arthrodesis Nail is an established implant and was used in the study to reflect the state of the art. The T2 AAN device had a diameter of 11 mm and a length of 200 mm; the Biomet Arthrodesis Nail had a diameter of 11 mm and a length of 180 mm. The two nails were made from the same material (Ti6-Al4-V). The Biomet nail was straight and had static locking. In contrast, the T2 AAN device had a 5° hindfoot valgus angulation at its distal end.
Bone Models
The fifteen synthetic bones used in the study were third-generation composite tibiae and hard-plastic tali and calcanei (Sawbones, Malmö, Sweden). The joint surfaces of the synthetic bones were resected to produce defined-plane parallel surfaces. Ankle fusion was performed in neutral flexion.
Twenty-four fresh-frozen below-the-knee specimens (including eight pairs) were provided by the Department of Anatomy I of Jena University. Following amputation, the bone mineral density of the cancellous bone in the calcaneus was determined with quantitative computed tomography (GE LightSpeed; GE Healthcare, Milwaukee, Wisconsin), and the specimens were stored at -25°. Dissection was performed and instrumentation was applied once the specimens had been thawed. The tibia was transected 25 cm proximal to the ankle joint, the forefoot was divided through the Chopart joint, and the fibula was removed. The joint surfaces were left in place to constitute an in situ arthrodesis.
The synthetic bone assemblies were divided into three groups of five specimens each for testing of the angle-stable, compressed angle-stable, or statically locked device. The human bone specimens were divided into three groups of eight specimens each and were treated in a like fashion.
Instrumentation
The T2 AAN device was tested with and without compression of the articular surfaces. As a result of the nail design (Fig. 1), locking was always angle-stable. For the compression of the articular surfaces, the compression screw and the external compression sleeve were tightened with a torque-limiting wrench to a torque of 2 Nm as recommended by the manufacturer.
All of the nails were implanted by an experienced surgeon (T.M.), following the manufacturer's instructions. All of the constructs were checked with radiographs.
Biomechanical Testing
The instrumented specimens were potted with use of RenCast FC 51 two-component polyurethane resin (Schurg, Bad Wildungen, Germany). Prior to potting, any implant parts protruding from the calcaneus, and the gap of the subtalar joint, were covered with an elastic rubber mass to prevent any interference with the potting. Testing was started as soon as the resin had cured.
Measurements were made on a testing machine (Instron 8874; Instron, Darmstadt, Germany) equipped with a 2-kN load cell (accuracy class, 0.1%). Bending moments in the tibiotalocalcaneal joints were created by application of a load to a lever (Fig. 2). The resulting displacement was measured by the displacement transducer of the testing system. At the start of our test series, the stiffness of the entire setup was determined with use of a steel tube. The stiffness of the steel tube was found to be approximately 0.001°/Nm, and the stiffness of the entire setup was found to be approximately 0.037°/Nm.
The sequence of initial stability tests was (1) dorsiflexion/plantar flexion, (2) varus/valgus, and (3) external rotation/internal rotation. Using a technique similar to the one described by Friedman et al.9, we loaded each specimen fifteen times with a bending moment of ±5 Nm, at a rate of 0.5 Hz. Load-displacement data were recorded. Owing to the design of the experimental setup, different lever-arm lengths were required for the different directions tested, resulting in maximum loads of 31.3 N for dorsiflexion/plantar flexion and varus/valgus (lever-arm length: 160 mm) and of 38.5 N for external rotation/internal rotation (lever-arm length: 130 mm).
Following these tests, fatigue testing was performed in dorsiflexion/plantar flexion with use of 100,000 cycles at a frequency of 1.5 Hz. On the basis of the studies by Chiodo et al.2, the synthetic bone models were loaded with a moment of 20 Nm. Preliminary tests at our center had shown the instrumented human-bone specimens to consistently and prematurely fail at this level of loading. The failure criteria were the same for the two bone models, which is why, in the cyclic testing, the human-bone specimens were loaded with a moment of 10 Nm. As a result of the lever lengths used, the maximum force was 62.5 N in the human bone model and 125 N in the synthetic bone model. If the construct failed or there was >10° of displacement, the test was discontinued.
Analysis
As proposed by Wilke et al.10, we determined the range of motion and the neutral zone in each construct. The range of motion was defined as the angulation under maximum load; the neutral zone was defined as the range over which the instrumented specimen moved essentially free of applied loading and reflected the laxity of the entire construct. For the determination of the range of motion and neutral zone, each specimen was preconditioned with five cycles, following which it was subjected to ten measuring cycles, which were analyzed and averaged. The results of the cyclic fatigue testing were analyzed in a similar fashion at 1000, 25,000, 50,000, 75,000, and 100,000 cycles.
Statistical Analysis
The results were analyzed for normal distribution with the Kolmogorov-Smirnov test and the Shapiro-Wilk test. Significance was measured with the Mann-Whitney U test and the Wilcoxon test. The hardware failures during cyclic testing were plotted as a Kaplan-Meier graph, and significance was measured with the log-rank test. Significance was set at p < 0.05. Spearman correlation analysis was used to investigate the dependence of construct stability on the bone mineral density of the cadaveric specimens.
In the synthetic bone model, the nails with angle-stable locking and those with compressed angle-stable locking showed significantly greater initial stability, in all loading directions, than did the statically locked nails (Figs. 3-A, 3-B, and 3-C). The construct with compressed angle-stable locking was significantly (p = 0.008) stiffer under external rotation/internal rotation loading than was the construct with angle-stable locking (Fig. 3-C). In the other loading directions, there was no significant difference between the constructs with angle-stable locking and those with compressed angle-stable locking. The same ranking was seen when the synthetic bone specimens were analyzed with regard to their neutral zones.
The mean bone mineral density of the human specimens was 0.515 ± 0.14 g/cm3 (range, 0.266 to 0.750 g/cm3). There were no significant differences between the groups. In the human bone model, too, the initial range of motion and the initial neutral zone were significantly smaller, in all loading directions, in the constructs with compressed angle-stable or angle-stable locking than they were in the statically locked constructs (Figs. 3-A, 3-B, and 3-C). There were no significant differences, in any of the loading directions, between the angle-stable and the compressed angle-stable constructs. The power calculation for these results, based on eight samples, showed low confidence levels (7% for dorsiflexion/plantar flexion, 8% for varus/valgus, and 11% for external rotation/internal rotation).
In the cyclic tests of the synthetic bone specimens, the constructs with compressed angle-stable locking were more stable, at all of the cycling steps that were analyzed (1000, 25,000, 50,000, 75,000, and 100,000 cycles), than the angle-stable or statically locked constructs (Fig. 4). None of the constructs failed during the 100,000 loading cycles.
Cyclic testing of the human bone specimens produced similar ranking. The angle-stable and compressed angle-stable constructs had a smaller range of motion than did the statically locked constructs. This difference was significant at 1000 cycles (p < 0.001) and at 25,000 cycles (p = 0.016). In the subsequent course of fatigue testing, five of the eight statically locked nails failed (Fig. 5); thus, the sample size available for the remainder of the cycles was too small to allow significance to be calculated.
There was one failure of a compressed angle-stable device and one failure of an angle-stable device with cyclic loading. The failure pattern (breakage of the posterior-to-anterior calcaneal screw and fracture of the calcaneus) was the same in all specimens that failed. The statically locked constructs had significantly poorer survival than did the compressed angle-stable constructs (p < 0.02) and the angle-stable constructs (p < 0.01) (Fig. 5). There was no significant difference, in the human bone model, between the nails with angle-stable locking and those with compressed angle-stable locking.
Spearman correlation analysis did not show any linear relationship between the bone mineral density of the specimens and the initial biomechanical properties (the initial range of motion and the initial neutral zone) of the different constructs. Cyclic loading of the human bone specimens above 50,000 cycles showed a linear relationship between bone mineral density and the range of motion (rho = 0.82 for 75,000 cycles and rho = 0.93 for 100,000 cycles) and between bone mineral density and the neutral zone (rho = 0.80 for 75,000 cycles and rho = 0.79 for 100,000 cycles) in the constructs with angle-stable locking and those with compressed angle-stable locking. Too few (three) of the statically locked nails survived until this cycling step for significance to be calculated.
In clinical practice, both intramedullary nails and blade-plates are used for tibiotalocalcaneal arthrodesis in patients with particularly demanding cases (such as those undergoing revision surgery and those with poor bone stock). One major difference between the two devices is the extent of the soft-tissue approach required for their insertion. Patients for whom tibiotalocalcaneal arthrodesis is being considered frequently have compromised soft tissues in the region concerned and usually cannot tolerate large incisions. Unlike a blade-plate, a nail can be introduced percutaneously, through small additional incisions. Given the less invasive access required for nailing, the questions of whether tibiotalocalcaneal arthrodesis with a retrograde intramedullary nail will be sufficiently stable and of which locking modes (angle-stable, compressed angle-stable, or static) should be used assume considerable clinical importance. To our knowledge, only Mann et al.7 have studied these questions with regard to second-generation arthrodesis nails; all of the other biomechanical studies have involved first-generation nails, which permit static locking in the coronal plane only.
In our study, the angle-stable locked constructs were significantly superior to the statically locked constructs in terms of initial stability (a smaller range of motion) and of laxity (a smaller neutral zone), independent of the bone model and direction tested. Regardless of the bone model used, angle-stable and compressed angle-stable locking conferred better initial stability and better fatigue behavior to the constructs than did static locking. The nails used were comparable in diameter, length, locking-screw diameter, and material. Therefore, the biomechanical differences that we observed may be attributed to the different locking modes that were used.
The use of synthetic bones for biomechanical testing is controversial. In principle, fresh human bones should be used. However, the qualitative and dimensional variability of human cadaveric bones poses a problem. The biomechanical properties of the shafts of third-generation composite femora and tibiae have been demonstrated to be comparable with those of normal human specimens11,12.
The mean bone density of our cadaver specimens was comparable with the data obtained in other biomechanical studies2,6. It ranged from normal to reduced and was thus typical of the patient population considered for tibiotalocalcaneal arthrodesis. Cyclic testing of specimens with this bone stock quality showed better results for angle-stable and compressed angle-stable locking than for static locking. Also, none of the talar locking screws used for the angle-stable devices loosened during the cyclic tests. There was one case of screw loosening with the compressed angle-stable pattern with cyclic loading. This failure mode is understandable since, with compressed angle-stable locking, the talar locking screw is placed in the dynamic position in the longitudinal hole and is subsequently pushed proximally with the compression screw. This means that, in contrast to the situation with the angle-stable pattern, the talar screw is not blocked in an angle-stable fashion. Whereas compression conferred significant benefit in the synthetic bone model, there was no significant benefit in the human bone model. The lesser effectiveness of compression is likely attributable to hardware loosening in the cancellous bone and to impaction of the osseous fusion surfaces into each other.
We are aware of only one biomechanical study of intramedullary nails with angle-stable locking in the lower extremity, which was performed by Kaspar et al.13. In that animal model, a tibial nail that had been modified to provide angle-stable locking was found to be superior to a statically locked nail, in that it reduced interfragmentary movement and accelerated bone-healing. Our findings are consistent with the observations in that study.
In our analysis of the different cycling steps, the human cadaveric bones showed two phenomena between 1000 and 100,000 cycles: (1) an absolute increase in the range of motion and the neutral zone of the constructs with angle-stable locking and those with compressed angle-stable locking, and (2) an increase in the percentage share of the neutral zone in the range of motion. These findings reflect the loosening of the angle-stable locking screws in the bone, especially in the calcaneus, in the course of cyclic testing. It follows that future work aimed at improving the devices for tibiotalocalcaneal arthrodesis will need to focus on the bone-implant interfaces and on loosening at these sites as well as on the overall rigidity of the angle-stable device. The statically locked constructs exhibited range-of-motion and neutral-zone patterns, from the start of cycling, that were similar to those seen with the angle-stable constructs at the 100,000-cycle level. Thus, the statically locked nails had greater flexibility from the outset, which accounts for the device failures seen in the course of fatigue testing.
Our test setup was essentially comparable with the systems used in other biomechanical studies9,14,15. In order to reduce variability, we followed the policy adopted by Alfahd et al.6 and Chiodo et al.2 of dispensing with the manual preparation of the joint surfaces. Cyclic loading in our synthetic bone model was based on the pattern described by Chiodo et al. The stiffness found by Chiodo et al. was not entirely comparable with our results, being higher than our values by a factor of two to five. This may, to some extent, have been due to the test setup used by Chiodo et al. They did not provide details of the potting technique, and they used an eccentric force transmission into the load cell. In addition, loading was in dorsiflexion only, and Chiodo et al. did not consider the increase in the play (the neutral zone) of the construct over the course of cyclic loading when they determined stiffness.
The results obtained by Alfahd et al.6, in a study of tibiotalocalcaneal arthrodesis with a statically locked intramedullary nail, were comparable with those found in the analyses of the statically locked constructs in our study. Alfahd et al. observed failure of two (14%) of their fourteen specimens, at a low level of loading. In contrast, Chiodo et al.2 reported a failure rate of only two (10%) of twenty specimens, with an exacting load regimen of 22.2 Nm and 250,000 cycles. In our study, the overall failure rate, at 10 Nm and 100,000 cycles, was seven (29%) of twenty-four constructs, with failure of five of the eight locked nails, one of the eight angle-stable locking constructs, and one of the eight compressed angle-stable locking constructs.
The biomechanical results obtained in this study cannot be translated directly into the clinical context. Greater initial and cyclic stability is not necessarily synonymous with better clinical outcomes. However, our results provide evidence of the superiority of the stability of the compressed angle-stable and angle-stable locking modes over the static locking mode in tibiotalocalcaneal arthrodesis constructs. Angle-stable locking may confer some clinical benefit for the typical candidates for this form of arthrodesis, who tend to have reduced bone density. 
Berson L, McGarvey WC, Clanton TO. Evaluation of compression in intramedullary hindfoot arthrodesis. Foot Ankle Int.2002;23:992-5.23992
2002
[PubMed]
Chiodo CP, Acevedo JI, Sammarco VJ, Parks BG, Boucher HR, Myerson MS, Schon LC. Intramedullary rod fixation compared with blade-plate-and-screw fixation for tibiotalocalcaneal arthrodesis: a biomechanical investigation. J Bone Joint Surg Am.2003;85:2425-8.852425
2003
Küntscher G. Praxis der Marknagelung. Stuttgart: Schattauer; 1962. p 349.
1962
Berend ME, Glisson RR, Nunley JA. A biomechanical comparison of intramedullary nail and crossed lag screw fixation for tibiotalocalcaneal arthrodesis. Foot Ankle Int.1997;18:639-43.18639
1997
Fleming S, Moore TJ, Hutton WC. Biomechanical analysis of hindfoot fixation using an intramedullary rod. J South Orthop Assoc.1998;7:19-26.719
1998
Alfahd U, Roth SE, Stephen D, Whyne CM. Biomechanical comparison of intramedullary nail and blade plate fixation for tibiotalocalcaneal arthrodesis. J Orthop Trauma.2005;19:703-8.19703
2005
[CrossRef]
Mann MR, Parks BG, Pak SS, Miller SD. Tibiotalocalcaneal arthrodesis: a biomechanical analysis of the rotational stability of the Biomet Ankle Arthrodesis Nail. Foot Ankle Int.2001;22:731-3.22731
2001
Mückley T, Eichhorn S, Hoffmeier K, von Oldenburg G, Speitling A, Hofmann GO, Bühren V. Biomechanical evaluation of primary stiffness of tibiotalocalcaneal fusion with intramedullary nails. Foot Ankle Int.2007;28:224-31.28224
2007
[CrossRef]
Friedman RL, Glisson RR, Nunley JA 2nd. A biomechanical comparative analysis of two techniques for tibiotalar arthrodesis. Foot Ankle Int.1994;15:301-5.15301
1994
Wilke HJ, Wenger K, Claes L. Testing criteria for spinal implants: recommendations for the standardization of in vitro stability testing of spinal implants. Eur Spine J.1998;7:148-54.7148
1998
[CrossRef]
Cristofolini L, Viceconti M. Mechanical validation of whole bone composite tibia models. J Biomech.2000;33:279-88.33279
2000
[CrossRef]
Cristofolini L, Viceconti M, Cappello A, Toni A. Mechanical validation of whole bone composite femur models. J Biomech.1996;29:525-35.29525
1996
[CrossRef]
Kaspar K, Schell H, Seebeck P, Thompson MS, Schutz M, Haas NP, Duda GN. Angle stable locking reduces interfragmentary movements and promotes healing after unreamed nailing. Study of a displaced osteotomy model in sheep tibiae. J Bone Joint Surg Am.2005;87:2028-37.872028
2005
[CrossRef]
Nasson S, Shuff C, Palmer D, Owen J, Wayne J, Carr J, Adelaar R, May D. Biomechanical comparison of ankle arthrodesis techniques: crossed screws vs. blade plate. Foot Ankle Int.2001;22:575-80.22575
2001
Noonan T, Pinzur M, Paxinos O, Havey R, Patwardhin A. Tibiotalocalcaneal arthrodesis with a retrograde intramedullary nail: a biomechanical analysis of the effect of nail length. Foot Ankle Int.2005;26:304-8.26304
2005