Fracture model
A total of fifty-four fresh and unembalmed tibias were selected from fifty-four voluntarily donated adult male cadavers (Department of Anatomy, Air Force Medical University, Xi'an, China) between the ages of 18 and 50. The average length of the selected tibias was 340 mm (range from 310 mm to 375 mm). All selected tibias specimens were examined for bone mineral density, and osteoporosis was ruled out by means of dual energy X-ray absorptiometry (LUNAR IDAX, GE Inc., Boston, Massachusetts, USA). The tibias were then cleaned of any soft tissues for use in this study. T-scores were selected to represent the values of bone mineral density.
The novel external tibial fixation device prototype, namely, the plate-type external fixator, consists of a proximal tibial fixation lath with a transverse slat at the proximal end and a distal tibial fixation lath with a transverse slat at the distal end. The distal end of the proximal fixation slat is equipped with a slot, and the proximal end of the distal fixation slat can be inserted into the slot and can slide along lath to adjust the length of the fixator to accommodate various lengths of the human lower limb. In addition, the tibial fixation laths and the transverse slats are both equipped with locking screw holes, and all the screws used are fully threaded self-tapping locking screws (Fig. 1). With a lower profile than the traditional the unilateral external fixator, the novel external tibial fixator, designed to match perfectly with the crus, is expected to make it easier to adjust the plate close to the bone surface. For this study, we lengthened the plate-type external fixator by 30 mm, namely, twofold the hole spacing, to clarify whether the extended plate-type external fixator could also provide sufficient stiffness and strength.
The fifty-four tibias were randomly divided into three groups of eighteen specimens each for fixation by a classical plate-type external fixator (CPF), an extended plate-type external fixator (EPF) or the the unilateral external fixator (UEF). Subsequently, the eighteen specimens of each configuration group were randomly divided into three groups of six specimens each for axial compression, four-point bending, and torsion testing, respectively.
A standardized midshaft osteotomy by means of an oscillating saw was performed in all tibias to create a 20 mm fracture gap, measured with the aid of a Vernier caliper, to simulate a comminuted tibial shaft fracture and to ensure no contact between both ends of the fracture. Eighteen specimens were stabilized with a 13-hole stainless steel CPF (300 mm in length, 21 mm in width, 10 mm in thickness, Kangding Medical Alliance Co., Ltd., Shanghai, China), with three 5 mm diameter stainless steel locking screws placed proximally in the first, third and fifth locking holes and three 5 mm diameter stainless steel locking screws placed distally in the ninth, eleventh and thirteenth locking holes. Another eighteen specimens were stabilized with a 15-hole stainless steel extended plate-type external fixator (330 mm in length, 21 mm in width proximally and 16 mm in width distally, 10 mm in thickness proximally and 5 mm in thickness distally, Kangding Medical Alliance Co., Ltd., Shanghai, China), with three 5 mm diameter stainless steel locking screws placed proximally in the second, fourth and sixth locking holes and three 5 mm diameter stainless steel locking screws placed distally in the tenth, twelfth and fourteenth locking holes. Both plate-type external fixators have a hole spacing of 15 mm.
The final eighteen specimens were stabilized with a stainless steel UEF (Kangding Medical Alliance Co., Ltd., Shanghai, China) as the control group. Three stainless steel half-pins (5 mm in diameter) were fixed per fragment and linked with pin clamps to a stainless steel rod (300 mm in length, 11 mm in diameter). The positions of the half-pins corresponded to the proximal locking screws of the CPF in the first, third and fifth holes and to the distal locking screws in the ninth, eleventh and thirteenth holes.
The choice of three locking screws/half-pins per fracture fragment in our study adhered to the AO principles of external fixation that a minimum of three screws were needed to achieve stable fixation on either fragment of the fracture. The AO recommended having a screw near and a screw far from the fracture end in both fragments; however, for the sake of comparison, the most distant screws were inserted into the second and fourteenth locking holes in the extended plate-type external fixator group instead of into the first and fifteenth locking holes, so the same three locking screws/half-pins positions were used in both fragments of the fracture among the three fracture fixation configuration groups. We acknowledge that this represents a limitation of our study, as the adjustment of the locking screws may influence the fixation stiffness of the extended plate-type external fixator.
The offset distance was restricted to 15 mm between the bone surface and the external plates/rods to allow the swelling of soft tissue without disturbance of the configuration and to provide sufficient space for postoperative care. We chose an offset of 15 mm instead of 20 mm or 30 mm for the purpose of increasing the fixation stability of the configuration to prevent excessive interfragmentary movements (4, 19, 20). The inner locking screws/half-pins were inserted at a distance of 20 mm from the fracture end. The locking screws/half-pins used were long enough to ensure adequate purchase of the bilateral cortex.
Mechanical testing
The proximal and distal ends of all the fracture fixation configurations were potted in polymethylmethacrylate for mechanical testing (Fig. 2) (6). Subsequently, the bone-implant constructs were mounted in the testing machine with a customized clamp. The classical plate-type external fixation constructs, the extended plate-type external fixation constructs and the unilateral external fixation constructs were tested to determine the fixation stiffness under three loading conditions (axial compression, four-point bending and torsion) (Fig. 3) (6, 21). The relative displacements at the fracture site were recorded on a computer to calculate the stiffness of the configuration. Subsequently, the three constructs underwent dynamic loading until failure under each loading mode to determine the construct strength and the failure modes. Construct strength was defined as the peak load at the moment of construct failure during progressive dynamic loading to failure under each loading mode. Configuration failure was defined either by catastrophic fracture or by nonrecoverable deformation in the region of fracture, whichever occurred first (5, 22-24).
Axial compression test
Both ends of the constructs were mounted with the use of a customized axial compression clamp in the Zwick/Roell-Z005 electronic materials testing machine (Ulm, Germany) (Fig. 3a). The applied loading was gradually increased from 0 N to a maximum load of 700 N, corresponding to the weight of a 70 kg person during a one-legged stance (25), at a rate of 0.1 mm/s for six cycles. The interfragmentary displacements at the fracture site were determined by means of laser displacement sensors (LK-G10, KEYENCE Inc., JAPAN). Axial compression stiffness was determined by dividing the axial load values by the vertical displacement values and was expressed in N/mm.
After the static test, sinusoidal loading with a constant load amplitude was applied for each construct. Every 100 loading cycles, the load amplitude was increased stepwise by 100 N until configuration failure occurred, while preloading was applied to 50 N to ensure that construct failure occurred within a reasonable number of loading cycles (<10,000) by increasing the load stepwise to failure (5, 21).
Four-point bending test
The constructs were placed in turn by means of a customized bending clamp on a Zwick/Roell-Z005 electronic materials testing machine (Ulm, Germany) (Fig. 3b). The bending moment was calculated by multiplying the bending force by the bending length. The distance between the lower supports was set to 200 mm, while the upper supports were separated by 100 mm. The bending length, defined as the distance between the upper and lower supports on either side of the fracture, was set to 50 mm. The bending force applied was constantly increased up to 400 N, corresponding to a bending moment of 20 Nm, at a rate of 1 mm/min. The bending stiffness was calculated by dividing the bending moment by the bending angle and was expressed in Nm/deg (26, 27). Afterwards, sinusoidal loading with a constant amplitude was applied for each configuration. The load amplitude was increased gradually every 100 loading cycles by 1 Nm until configuration failure occurred, while the preload was applied to 1 Nm (5).
Torsion test
The torsional testing was performed by using a CTS-500 microcomputer controlled torsion test machine (Hualong Testing Instrument Co., Ltd., Shanghai, China) equipped with a custom-made torsional clamp, with the proximal and distal ends of the constructs being rigidly clamped by means of the clamp (Fig. 3c). The implemented torque was constantly increased from 0 Nm to 10 Nm at a rate of 0.1 deg/s for six cycles. Torsional stiffness was obtained by dividing the torque value by the relative rotation value and was expressed in Nm/deg. Subsequently, sinusoidal loading with a constant amplitude was applied for each configuration. The load amplitude was increased every 100 loading cycles in steps of 1 Nm until construct failure occurred, while the preload was adjusted to 1 Nm (5).
Statistical analysis
The collected data were statistically analyzed with SPSS 23.0 software (SPSS, Chicago, Illinois, USA). First, the results were tested for normality and homogeneity of variance. When a normal distribution and homogeneity of variance were found, the data were analyzed by means of one-way analysis of variance to determine the significance of differences in the means among the three groups. The LSD test was used for post hoc testing, if necessary. A p < 0.05 was considered statistically significant.