Finite Element Analysis of Traditional and New Fixation Techniques of the 3D-Printed Composite Interlocking Nail in Canine Femoral Shaft Fractures

Abstract

Since the removal of a metallic interlocking nail system leaves a blank cavity inside a healed bone, bioactive and biodegradation materials have been used instead to induce bone formation and eliminate complications of the material removal procedure. The previous study presented the possibility of an interlocking nail fabrication from polylactic acid (PLA), polycaprolactone (PCL), and hydroxyapatite (HA) using 3D printing, namely fused filament fabrication (FFF), for canine diaphyseal fractures. Therefore, a finite element analysis (FEA) was used to predict the maximum principal stress of this 3D-printed composite interlocking nail to stabilize a canine femoral fracture, and the biomechanical performance was evaluated for the treatment of canine femoral shaft fractures using both traditional and new fixation techniques. Three-dimensional FEA models were created, and the composite interlocking nail was tested for implant strength and stability. Three types of canine femoral shaft fracture (proximal shaft fracture, middle shaft fracture, and distal shaft fracture) fixed by traditional and new fixation techniques, consisting of two, four, and six locking screws, were analyzed with a multilevel factorial design technique. The maximum principal stresses of the composite interlocking nail were compared with each fixation technique. According to the multilevel factorial design, gap type, fracture gap, and fixation techniques are factors that affect the maximum principal stress of the composite interlocking nail for two and four locking screws. For six locking screws, all factors, including gap type, fracture gap, nail length, and fixation techniques, significantly affect the maximum principal stress. The use of a 3D-printed composite interlocking nail system with new fixation techniques demonstrated lower maximum principal stresses than the interlocking nail system that used a traditional fixation technique. The results of this study could help orthopedic veterinary surgeons to understand the biomechanical performances of traditional and new fixation techniques. Furthermore, surgeons may use the numerical results of this analysis to choose a fixation technique based on a patient’s condition.

1. Introduction

The metallic interlocking nail system has been widely used to treat canine femoral shaft fractures [1,2]. The biomaterials used in interlocking nails are titanium and stainless steel [3,4,5]. Nevertheless, the removal of a metallic interlocking nail is still necessary due to long-term implantation complications [6]. Biological support to living cells is not activated, and bones are not regenerated in the interlocking nail cavity according to the bioinert properties of these materials [7]. Thus, the metallic interlocking nail will leave a blank cavity inside the healed bone.

To eliminate the limitations of the metallic interlocking nail, alternative biodegradable materials such as polylactic acid (PLA) and polycaprolactone (PCL) were used to fabricate interlocking nail components for the model of a canine, porcine, and ovine femoral fracture [8,9,10]. During the early state and the complete healing state of bone fractures (4 and 12 weeks after implantation), there were no differences concerning the fracture healing, gait pattern, or tissue reaction between the PLA and stainless-steel interlocking nails in an animal experiment [9,11]. In a rabbit humerus osteotomy model, PCL intramedullary pins presented low inflammation compared to stainless-steel pins. All PCL-immobilized osteotomies have been healed with an extensive production of periosteal cells [12]. However, the PCL fixation was initially weaker than the stainless-steel fixation by mechanical testing. To enhance the properties of the fabricated materials, copolymers or mixed composites were presented [13]. PLA and PCL are additionally combined as a copolymer providing the necessary abilities for applications in the field of tissue engineering, for example, biocompatibility, biodegradability, non-toxicity, and environmentally friendly tissue [12,13]. Currently, PLA and PCL are commercially available for fused filament fabrication (FFF) 3D printing applications [14], which is the promised fabrication technique for medical applications [15,16]. The previous study presented the possibility of an interlocking nail fabrication from FFF 3D-printed PLA/PCL/hydroxyapatite (HA) composites [17]. Hydroxyapatite (HA), the main mineral in mammalian bone, was used to improve osteoconductivity and induce bone growth [18,19].

A common treatment for a canine femoral shaft fracture is an interlocking nail system with a traditional fixation technique [20]. Due to its improved fixation and reduced loss of soft tissue compared with traditional plating, this technique has biomechanical advantages [21].

However, some traditional fixation techniques have rigidly locked proximal as well as distal fragments by the locking screws and could not transfer the canine weight from the proximal fragment to the distal fragment. Furthermore, the distance between the screws, which fixes only the proximal or distal fragments without fixation in the middle, makes it easy for the stress to reach the middle of the interlocking nail [22,23]. Thus, the new fixation techniques have added middle locking screws and made it more stable to fasten the middle fragments, which allow for better support of the stresses in the middle, which normally occur in the interlocking nail without the middle holes [24]. Based on a literature review, there have been no research investigations using a finite element analysis (FEA) of the biomechanical characteristics of traditional locking nails in canine femoral shaft fractures. We hypothesized that the new technique of fixation would reduce the implant failure risk and stress of the interlocking nail. Moreover, new fixation techniques could produce different results.

In this study, traditional and new fixation techniques were used for three types of canine femoral shaft fractures. Additionally, to estimate the biomechanical performances, the interlocking nail stress and fixation stability were calculated. This paper aims towards the fixation technique modeled by SolidWorks and calculated by an FEA using the ANSYS software, gathered with a statistical tool, with an experimental design by the Minitab software, to determine the main effects and interactions among the variables. The variables treated were gap type, fracture gap, interlocking nail length, and fixation techniques. A 3D FEA provides detailed information regarding the maximum principal stress of the composite interlocking nails. This model is evaluated to provide the data to orthopedic veterinary surgeons and researchers using dogs as a model. This study will also be the basis for future biomechanics studies on the use of composite interlocking nails in the canine femur.

2. Materials and Methods

For the canine femoral models, 120 transverse computed tomography scans of the femur were obtained from the cadaver of a male dog weighing 22 kg. The dog was healthy and was not associated with any abnormality of the femur. This investigation was approved by the Institutional Animal Care and Use Committee, Faculty of Veterinary Medicine, Chiang Mai University, Chiang Mai, Thailand (Ref. No. R7/2558). The entire left femur was scanned using the 3D scanner EinScan-SE from SHINING 3D (EinScan SE, Shining 3D, Hangzhou, China) and the EinScan-S series v2 software. This canine femur model consists only of cortical bone. A canine femur with a transverse fracture was developed with the use of SolidWorks 2018 (SolidWorks Corporation, Concord, MA, USA), and the fracture gap between the proximal and distal bone fragments was assumed to be 1–3 mm [23,25]. In the current research, three types of femoral shaft fractures were considered, including a proximal shaft fracture, a middle shaft fracture, and a distal shaft fracture, as shown in Figure 1.

Each type of canine femoral shaft fracture was fixed using a composite interlocking nail system, including traditional fixation techniques: proximal and distal shaft fracture fixed with two proximal locking screws (2P), two distal locking screws (2D), one proximal locking screw with one distal locking screw (P–D), and two proximal locking screws with two distal locking screws (2P–2D); and new fixation techniques: proximal and distal shaft fracture fixed with two middle locking screws (2M), one proximal locking screw with one middle locking screw (P–M), one middle locking screw with one distal locking screw (M–D), two proximal locking screws with two middle locking screws (2P–2M), two middle locking screws with two distal locking screws (2M–2D), one proximal locking screw with two middle locking screw and one distal locking screws (P–2M–D), and two proximal locking screws with two middle locking screw and two distal locking screws (2P–2M–2D). The composite interlocking nail was implanted from the proximal fragment to the distal fragment. In the current study, three types of canine femoral shaft fractures fixed by 11 types of fixation techniques, as seen in Figure 2 (proximal and distal), and 4 types of fixation techniques, as seen in Figure 3 (middle), were analyzed.

The material properties were as follows: The 3D models of the canine femoral fracture fixed with the composite interlocking nail system were transferred to the commercial software ANSYS R14.5 Workbench (ANSYS, Inc., Canonsburg, PA, USA). The canine femoral model length was about 180 mm, and the canine femur made from the isotropic material for the cortical bone with a Young’s modulus of 15 GPa, a Poisson’s ratio of 0.3, and a density of 2 g/cm³ [26,27]. The 3D-printed composite interlocking nail and locking screws used in this study were made from PLA and PCL in a 70:30 mixing ratio with hydroxyapatite 15% w/w (PLA/PCL/15HA). The test specimens were 3D-printed from the extruded composite filaments using an FFF printer. The mechanical properties (

Table 1. Properties of materials used to compose the models of the composite interlocking nail system.

Material Properties	Cortical Bone	PLA/PCL/15HA
Compressive strength (MPa)	193	82.72
Tensile strength (MPa)	150	52.02
Young’s modulus (GPa)	15	1.06
Poisson’s ratio	0.3	0.34
Density (g/cm3)	2.0	2.0

) were obtained from the mechanical tests of Pitjamit et al. [17], with a Poisson’s ratio of 0.34 and a density of 2 g/cm³ [28]. The composite interlocking nail had a diameter of 6 mm and a length of 120–160 mm, with a 2.7 mm diameter [29], and the new hole position was located on two proximal holes, two middle holes, and two distal holes (Figure 4). The locking screws were modeled without threads, with a diameter of 2.7 mm.

The fractured canine femoral model with a composite interlocking nail system was free meshed using tetrahedral elements [30]. Interlocking nails and screws were included in the canine femoral model using SolidWorks 2018 to generate output files in a Parasolid file format (.x_t), which were used as the input to the FEA software. By considering that the screws were connected to the bone and the nails, contact interactions were established between the different materials. Table 1 provides the mechanical properties in all the materials: cortical bone and PLA/PCL/15HA, which were modeled as being homogenous, isotropic, and linearly elastic.

The loading and boundary conditions were as follows: Two force configurations were loaded onto the model. Loading was employed to investigate the effect of the hip joint reaction force (numbers 1 and 2) and the forces applied by all muscles inserted on the femur (numbers 3–18). The theoretical calculation with the optimization techniques from Shahar et al. [26] produces magnitudes and force directions (see Figure 5 and

Table 2. The x, y, and z components of the joint reaction forces (numbers 1 and 2) and the muscle forces (numbers 3–18) acting on the canine left femur during a three-legged stance [25].

Number	Force	Fx (N)	Fy (N)	Fz (N)
1	Hip-joint reaction force	−10.10	52.42	−232.74
2	Patellofemoral reaction force	−9.18	81.46	20.03
3	Medial gluteal	7.25	−36.92	15.29
4	Piriformis	4.99	−5.65	3.37
5	Superficial gluteal	1.01	−9.02	1.48
6	Articularis coxae	−0.05	−0.13	0.29
7	Gemelli	1.52	2.05	2.46
8	External obturator	2.11	2.85	3.41
9	Internal obturator	0.38	0.15	0.61
10	Lateral and intermediate vastus	4.49	−1.55	−38.30
11	Iliopsoas	−0.17	−1.25	0.88
12	Quadratus femoris	1.12	3.44	3.64
13	Medial vastus	0.38	−3.46	−40.52
14	Adductor magnus and brevis	5.91	2.55	9.17
15	Pectineus	0.11	−0.05	0.44
16	Cranial tensor fascia latae	−0.17	−2.43	8.44
17	Semimembranosus	0.09	7.10	19.64
18	Long digital extensor	3.53	6.78	−10.40

). The forces were applied to nodes chosen based on data from the cadaver dissections and anatomical textbooks.

Different fixation techniques could affect the implant failure risk in terms of the maximum tensile stress and compression stress of the 3D composite materials (PLA/PCL/15HA). For the evaluation of their mechanical strength, the maximum principal stresses of the composite interlocking nail were selected. Additionally, the maximum principal stresses of the composite interlocking nail system have also been compared with either the traditional fixation technique or the new fixation techniques.

The design of the experiment using a multilevel factorial design technique, Minitab 19 (Minitab, Ltd., Coventry, UK), typically determines statistically significant factors in several types of application.

According to this study, the 3 levels of experimental design were used to evaluate the effect of the maximum principal stress, as

Table 3. Two locking screws: multilevel factorial design of relevant factors and their levels. (Number of levels: 3, 3, 3, and 6.).

Factors	Levels						Unit
Gap Type (A)	Proximal		Middle		Distal		Type
Fracture gap (B)	1		2		3		mm
Nail length (C)	120		140		160		mm
Fixation techniques (D)	2P	2D	2M	P–D	P–M	M–D	type

,

Table 4. Four locking screws: the multilevel factorial design of relevant factors and their levels. (Number of levels: 3, 3, 3, and 4.).

Factors	Levels				Unit
Gap Type (A)	Proximal	Middle		Distal	Type
Fracture gap (B)	1	2		3	mm
Nail length (C)	120	140		160	mm
Fixation techniques (D)	2P–2D	2P–2M	2M–2D	P–2M–D	type

and

Table 5. Six locking screws: the multilevel factorial design of relevant factors and their levels. (Number of levels: 2, 3, and 3.).

Factors	Levels			Unit
Gap Type (A)	Proximal		Distal	Type
Fracture gap (B)	1	2	3	mm
Nail length (C)	120	140	160	mm
Fixation techniques *	2P–2M–2D			Type

* The case for using six-screw fixation techniques can be fixed only once.

demonstrate, and the experimental boundary of the factors including gap type (A), fracture gap (B), nail length (C), and fixation techniques (D), with two locking screws, four locking screws, and six locking screws. The combinations of experimental conditions for these were 162, 108, and 18, respectively, and were created to find the appropriate boundary condition of the canine femoral nailing techniques with the optimal maximum principal stress.

3. Results

3.1. Numerical Modeling

The static structural mechanical analysis module was conducted by utilizing the tetrahedron element type [23,25,30], with 1 mm in size for the elements’ interlocking nail and screws, and 2 mm in size for the canine femoral bone, as shown in Figure 6. The total number of elements and nodes are shown in

Table 6. All elements and nodes of each interlocking nail system.

	Elements	Nodes
Two locking screws
120 mm interlocking nail	140,762	228,472
140 mm interlocking nail	146,210	237,583
160 mm interlocking nail	150,939	245,140
Four locking screws
120 mm interlocking nail	143,179	232,836
140 mm interlocking nail	148,984	242,533
160 mm interlocking nail	152,973	248,812
Six locking screws
120 mm interlocking nail	144,881	235,920
140 mm interlocking nail	151,282	246,595
160 mm interlocking nail	154,759	252,052

. Tetrahedral elements have been used with respect to the femur because these elements allow us to model complex bone, nail, and screw geometries without significant calculation costs.

3.2. Analysis of Variance

3.2.1. Two Locking Screws

The maximum principal stress of two locking screws results from running the FEA model at various testing conditions, as shown in

Table 7. FEA model of the two locking screws’ results for the testing conditions.

Condition	A	B	C	D	Principal Stress	Condition	A	B	C	D	Principal Stress
1	Proximal	1	120	2P	129.75	38	Proximal	3	120	2D	158.2
2	Proximal	1	120	2D	128.35	39	Proximal	3	120	2M	59.3
3	Proximal	1	120	2M	50.24	40	Proximal	3	120	P–D	153.86
4	Proximal	1	120	P–D	127.43	41	Proximal	3	120	P–M	57.89
5	Proximal	1	120	P–M	47.89	42	Proximal	3	120	M–D	60.23
6	Proximal	1	120	M–D	50.24	43	Proximal	3	140	2P	164.12
7	Proximal	1	140	2P	209.63	44	Proximal	3	140	2D	163.07
8	Proximal	1	140	2D	145.02	45	Proximal	3	140	2M	58.17
9	Proximal	1	140	2M	50.07	46	Proximal	3	140	P–D	141.89
10	Proximal	1	140	P–D	112.6	47	Proximal	3	140	P–M	53.67
11	Proximal	1	140	P–M	46.13	48	Proximal	3	140	M–D	58.2
12	Proximal	1	140	M–D	50.11	49	Proximal	3	160	2P	171.5
13	Proximal	1	160	2P	198.89	50	Proximal	3	160	2D	168.02
14	Proximal	1	160	2D	120.17	51	Proximal	3	160	2M	49.68
15	Proximal	1	160	2M	44.44	52	Proximal	3	160	P–D	211
16	Proximal	1	160	P–D	84.52	53	Proximal	3	160	P–M	92.14
17	Proximal	1	160	P–M	39.28	54	Proximal	3	160	M–D	49.92
18	Proximal	1	160	M–D	44.43	55	Middle	1	120	2P	36.7
19	Proximal	2	120	2P	110.87	56	Middle	1	120	2D	36.73
20	Proximal	2	120	2D	129.73	57	Middle	1	120	2M	*
21	Proximal	2	120	2M	50.37	58	Middle	1	120	P–D	36.72
22	Proximal	2	120	P–D	114.36	59	Middle	1	120	P–M	*
23	Proximal	2	120	P–M	48.3	60	Middle	1	120	M–D	*
24	Proximal	2	120	M–D	50.37	61	Middle	1	140	2P	39.13
25	Proximal	2	140	2P	137.89	62	Middle	1	140	2D	39.96
26	Proximal	2	140	2D	130.24	63	Middle	1	140	2M	*
27	Proximal	2	140	2M	46.34	64	Middle	1	140	P–D	39.69
28	Proximal	2	140	P–D	105.97	65	Middle	1	140	P–M	*
29	Proximal	2	140	P–M	43.28	66	Middle	1	140	M–D	*
30	Proximal	2	140	M–D	46.35	67	Middle	1	160	2P	27.24
31	Proximal	2	160	2P	152.42	68	Middle	1	160	2D	28.2
32	Proximal	2	160	2D	136.94	69	Middle	1	160	2M	*
33	Proximal	2	160	2M	46.94	70	Middle	1	160	P–D	27.93
34	Proximal	2	160	P–D	118	71	Middle	1	160	P–M	*
35	Proximal	2	160	P–M	46.01	72	Middle	1	160	M–D	*
36	Proximal	2	160	M–D	46.95	73	Middle	2	120	2P	36.34
37	Proximal	3	120	2P	138.4	74	Middle	2	120	2D	36.37
Condition	A	B	C	D	Principal Stress	Condition	A	B	C	D	Principal Stress
81	Middle	2	140	2M	*	118	Distal	1	140	P-D	132.12
82	Middle	2	140	P-D	48.63	119	Distal	1	140	P-M	132.19
83	Middle	2	140	P-M	*	120	Distal	1	140	M-D	104.47
84	Middle	2	140	M-D	*	121	Distal	1	160	2P	126.23
85	Middle	2	160	2P	43.41	122	Distal	1	160	2D	131.05
86	Middle	2	160	2D	45.37	123	Distal	1	160	2M	162.22
87	Middle	2	160	2M	*	124	Distal	1	160	P-D	86.37
88	Middle	2	160	P-D	44.99	125	Distal	1	160	P-M	142.48
89	Middle	2	160	P-M	*	126	Distal	1	160	M-D	98.69
90	Middle	2	160	M-D	*	127	Distal	2	120	2P	118.59
91	Middle	3	120	2P	44.84	128	Distal	2	120	2D	131.05
92	Middle	3	120	2D	45.44	129	Distal	2	120	2M	81.53
93	Middle	3	120	2M	*	130	Distal	2	120	P-D	120.33
94	Middle	3	120	P-D	45.13	131	Distal	2	120	P-M	89.82
95	Middle	3	120	P-M	*	132	Distal	2	120	M-D	86.22
96	Middle	3	120	M-D	*	133	Distal	2	140	2P	186.02
97	Middle	3	140	2P	48.57	134	Distal	2	140	2D	159.14
98	Middle	3	140	2D	49.02	135	Distal	2	140	2M	190.35
99	Middle	3	140	2M	*	136	Distal	2	140	P-D	158.46
100	Middle	3	140	P-D	50.85	137	Distal	2	140	P-M	190.02
101	Middle	3	140	P-M	*	138	Distal	2	140	M-D	144.26
102	Middle	3	140	M-D	*	139	Distal	2	160	2P	134.75
103	Middle	3	160	2P	39.62	140	Distal	2	160	2D	127.98
104	Middle	3	160	2D	45.51	141	Distal	2	160	2M	176.06
105	Middle	3	160	2M	*	142	Distal	2	160	P-D	92.66
106	Middle	3	160	P-D	42.86	143	Distal	2	160	P-M	154.41
107	Middle	3	160	P-M	*	144	Distal	2	160	M-D	98.84
108	Middle	3	160	M-D	*	145	Distal	3	120	2P	287.54
109	Distal	1	120	2P	154.55	146	Distal	3	120	2D	182.01
110	Distal	1	120	2D	131.01	147	Distal	3	120	2M	297.06
111	Distal	1	120	2M	134.98	148	Distal	3	120	P-D	202.58
112	Distal	1	120	P-D	120.44	149	Distal	3	120	P-M	314.64
113	Distal	1	120	P-M	146.98	150	Distal	3	120	M-D	99.25
114	Distal	1	120	M-D	91.41	151	Distal	3	140	2P	229.15
115	Distal	1	140	2P	122.8	152	Distal	3	140	2D	137.82
116	Distal	1	140	2D	104.52	153	Distal	3	140	2M	228.98
117	Distal	1	140	2M	129.35	154	Distal	3	140	P-D	125.09
						155	Distal	3	140	P-M	218.22
						156	Distal	3	140	M-D	104.35
						157	Distal	3	160	2P	283.41
						158	Distal	3	160	2D	164.47
						159	Distal	3	160	2M	249.58
						160	Distal	3	160	P-D	128.83
						161	Distal	3	160	P-M	315.93
						162	Distal	3	160	M-D	112.42

* According to orthopedic veterinary surgeons, this condition could not be treated by locking screws in that position because the locking screws could not be locked near the fracture site of the middle shaft fracture.

. All main factors were identified in importance by the analysis of variance (ANOVA) shown in

Table 8. Analysis of variance (ANOVA) for principal stress (two locking screws).

Source	DF	Adj. SS	Adj. MS	F-Value	P-Value
Model	23	505,293	21,969.3	26.18	0.000
Linear	9	230,383	25,598.1	30.51	0.000
A	2	151,544	75,772.0	90.30	0.000
B	2	34,878	17,439.0	20.78	0.000
D	5	23,384	4676.7	5.57	0.000
2-Way Interactions	14	127,635	9116.8	10.86	0.000
A*B	4	24,258	6064.4	7.23	0.000
A*D	10	100,187	10,018.7	11.94	0.000
Error	115	96,498	839.1
Total	138	601,791

R-Square = 83.96%, R-Square (adj.) = 80.76%.

. Furthermore, R-squared and adjusted R-squared statistics were analyzed for checking the fit of the dataset. Moreover, if the p-value of each main effect and its interaction was less than 0.05, it was considered statistically significant. According to this analysis, gap type (A), fracture gap (B), fixation techniques (D), the interaction between the variables A and D, and the interaction between the variables A and B were factors that affected the maximum principal stress of the composite interlocking nail.

The R² value shows that the model explains 83.96% of the variance in the maximum principal stress, which indicates that the model fits the data well.

The residual plot of the maximum principal stress is shown in Figure 7, representing the normal probability plot of the residuals. The residuals of the response randomly scatter around zero in the residual plots. Regarding the residuals versus the fits, the observation points (blue dots) scattered randomly in the normal probability plot show that the process parameters in this study were normal and independently distributed. The histogram of the residuals is approximately symmetric. Regarding the residuals versus the order, the residuals exhibit no clear pattern.

The magnitude and significance of an effect were determined by the Pareto chart shown in Figure 8. The Pareto chart shows the absolute values from the largest to the smallest effect of the standardized effects. The standardized effects are t-statistics that test the null hypothesis that the effect is zero [31]. In this Pareto chart, the bars that represent factors A, AD, B, AB, and D cross the reference line, which is at 1.981. These factors are statistically significant at the 0.05 level with the current model terms. In the selected range, the infection order of each factor is A, AD > B > AB > D. This means that the variable A and the interaction between variables A and D are potentially the most and second-most important, respectively. The second in terms of importance is the variable B. The interaction between variables A and C and the variable D are considered the third- and fourth-most important effects, regarding the maximum principal stress by the statistical model analysis.

3.2.2. Four Locking Screws

The maximum principal stress of four locking screws results from running the FEA model at various testing conditions, as shown in

Table 9. FEA model of the four locking screws’ results for the testing conditions.

Condition	A	B	C	D	Principal Stress	Condition	A	B	C	D	Principal Stress	Condition	A	B	C	D	Principal Stress
1	Proximal	1	120	2P-2D	121.95	38	Middle	1	120	2P-2M	*	75	Distal	1	120	2M-2D	82.77
2	Proximal	1	120	2P-2M	47.31	39	Middle	1	120	2M-2D	*	76	Distal	1	120	P-2M-D	80.56
3	Proximal	1	120	2M-2D	50.22	40	Middle	1	120	P-2M-D	*	77	Distal	1	140	2P-2D	107.99
4	Proximal	1	120	P-2M-D	47.93	41	Middle	1	140	2P-2D	39.14	78	Distal	1	140	2P-2M	115
5	Proximal	1	140	2P-2D	195.93	42	Middle	1	140	2P-2M	*	79	Distal	1	140	2M-2D	98.92
6	Proximal	1	140	2P-2M	62.29	43	Middle	1	140	2M-2D	*	80	Distal	1	140	P-2M-D	95.63
7	Proximal	1	140	2M-2D	50.07	44	Middle	1	140	P-2M-D	*	81	Distal	1	160	2P-2D	77.88
8	Proximal	1	140	P-2M-D	46.17	45	Middle	1	160	2P-2D	27.25	82	Distal	1	160	2P-2M	122.95
9	Proximal	1	160	2P-2D	77.32	46	Middle	1	160	2P-2M	*	83	Distal	1	160	2M-2D	97.98
10	Proximal	1	160	2P-2M	36.57	47	Middle	1	160	2M-2D	*	84	Distal	1	160	P-2M-D	84.28
11	Proximal	1	160	2M-2D	44.44	48	Middle	1	160	P-2M-D	*	85	Distal	2	120	2P-2D	119.7
12	Proximal	1	160	P-2M-D	39.37	49	Middle	2	120	2P-2D	36.34	86	Distal	2	120	2P-2M	79.73
13	Proximal	2	120	2P-2D	112.28	50	Middle	2	120	2P-2M	*	87	Distal	2	120	2M-2D	78.06
14	Proximal	2	120	2P-2M	47.95	51	Middle	2	120	2M-2D	*	88	Distal	2	120	P-2M-D	76.05
15	Proximal	2	120	2M-2D	50.36	52	Middle	2	120	P-2M-D	*	89	Distal	2	140	2P-2D	154.48
16	Proximal	2	120	P-2M-D	48.34	53	Middle	2	140	2P-2D	47.84	90	Distal	2	140	2P-2M	185.57
17	Proximal	2	140	2P-2D	129.17	54	Middle	2	140	2P-2M	*	91	Distal	2	140	2M-2D	125.85
18	Proximal	2	140	2P-2M	40.48	55	Middle	2	140	2M-2D	*	92	Distal	2	140	P-2M-D	143.57
19	Proximal	2	140	2M-2D	46.34	56	Middle	2	140	P-2M-D	*	93	Distal	2	160	2P-2D	78.23
20	Proximal	2	140	P-2M-D	43.34	57	Middle	2	160	2P-2D	43.44	94	Distal	2	160	2P-2M	126.69
21	Proximal	2	160	2P-2D	85.46	58	Middle	2	160	2P-2M	*	95	Distal	2	160	2M-2D	98.1
22	Proximal	2	160	2P-2M	40.52	59	Middle	2	160	2M-2D	*	96	Distal	2	160	P-2M-D	87.92
23	Proximal	2	160	2M-2D	46.94	60	Middle	2	160	P-2M-D	*	97	Distal	3	120	2P-2D	275.77
24	Proximal	2	160	P-2M-D	45.05	61	Middle	3	120	2P-2D	44.85	98	Distal	3	120	2P-2M	265.51
25	Proximal	3	120	2P-2D	131.4	62	Middle	3	120	2P-2M	*	99	Distal	3	120	2M-2D	90.19
26	Proximal	3	120	2P-2M	59.22	63	Middle	3	120	2M-2D	*	100	Distal	3	120	P-2M-D	95.91
27	Proximal	3	120	2M-2D	59.29	64	Middle	3	120	P-2M-D	*	101	Distal	3	140	2P-2D	106.69
28	Proximal	3	120	P-2M-D	57.89	65	Middle	3	140	2P-2D	42.26	102	Distal	3	140	2P-2M	207.8
29	Proximal	3	140	2P-2D	113.3	66	Middle	3	140	2P-2M	*	103	Distal	3	140	2M-2D	98.49
30	Proximal	3	140	2P-2M	49.82	67	Middle	3	140	2M-2D	*	104	Distal	3	140	P-2M-D	95.64
31	Proximal	3	140	2M-2D	58.16	68	Middle	3	140	P-2M-D	*	105	Distal	3	160	2P-2D	93.04
32	Proximal	3	140	P-2M-D	53.70	69	Middle	3	160	2P-2D	39.64	106	Distal	3	160	2P-2M	272.78
33	Proximal	3	160	2P-2D	156.01	70	Middle	3	160	2P-2M	*	107	Distal	3	160	2M-2D	111.55
34	Proximal	3	160	2P-2M	50.26	71	Middle	3	160	2M-2D	*	108	Distal	3	160	P-2M-D	102.97
35	Proximal	3	160	2M-2D	49.68	72	Middle	3	160	P-2M-D	*
36	Proximal	3	160	P-2M-D	60.24	73	Distal	1	120	2P-2D	119.79
37	Middle	1	120	2P-2D	36.71	74	Distal	1	120	2P-2M	130.08

* According to orthopedic veterinary surgeons, this condition could not be treated by locking screws in that position because the locking screws could not be locked near the fracture site of the middle shaft fracture.

. All main factors were identified in importance by an ANOVA, as shown in

Table 10. Analysis of variance for principal stress (four locking screws).

Source	DF	Adj. SS	Adj. MS	F-Value	P-Value
Model	17	200,355	11,785.6	14.50	0.000
Linear	7	110,188	15,741.2	19.37	0.000
A	2	83,008	41,504.1	51.07	0.000
B	2	7771	3885.4	4.78	0.011
D	3	19,172	6390.7	7.86	0.000
2-Way Interactions	10	48,725	4872.5	6.00	0.000
A*B	4	8580	2145.1	2.64	0.040
A*D	6	40,557	6759.4	8.32	0.000
Error	82	66,637	812.7
Total	99	266,993

R-Square = 75.04%, R-Square (adj.) = 69.87%.

. Furthermore, R-squared and adjusted R-squared statistics were analyzed for checking the fit of the dataset. Moreover, if the p-value of each main effect and its interaction was less than 0.05, it was considered statistically significant. According to this analysis, gap type (A), fracture gap (B), fixation techniques (D), the interaction between the variables A and D, and the interaction between the variables A and B were factors that affected the maximum principal stress of the composite interlocking nail in the same way as two screws.

The R² value shows that the model explains 75.04% of the variance in the maximum principal stress, which indicates that the model fits the data well.

The residual plot of the maximum principal stress is shown in Figure 9, representing the normal probability plot of the residuals. The residuals of the response randomly scatter around zero in the residual plots. Regarding the residuals versus the fits, the observation points (blue dots) scattered randomly in the normal probability plot show that the process parameters in this study were normal and independently distributed. The histogram of the residuals is approximately symmetric. Regarding the residuals versus the order, the residuals exhibit no clear pattern.

The magnitude and significance of an effect were determined by the Pareto chart shown in Figure 10. The Pareto chart shows the absolute values from the largest to the smallest effect of the standardized effects. The standardized effects are t-statistics that test the null hypothesis that the effect is zero [31]. In this Pareto chart, the bars that represent factors A, AD, D, B, and AB cross the reference line, which is at 1.989. These factors are statistically significant at the 0.05 level with the current model terms. In the selected range, the infection order of each factor is A > AD > D > B > AB. This means that the variable A and the interaction between variables A and D are potentially the most and second-most important, respectively. The third-most important is the variable D. The variable B and the interaction between variables A and B are considered the fourth- and the fifth-most important effects regarding the maximum principal stress by the statistical model analysis.

3.2.3. Six Locking Screws

The maximum principal stress of six locking screws results from running the FEA model at various testing conditions as shown in

Table 11. FEA model of the six locking screws’ results for the testing conditions.

Condition	A	B	C	Principal Stress
1	Proximal	1	120	47.23
2	Distal	1	120	80.50
3	Proximal	3	120	59.18
4	Distal	3	120	148.48
5	Proximal	1	160	36.58
6	Distal	1	160	75.83
7	Proximal	3	160	50.24
8	Distal	3	160	91.14
9	Proximal	1	120	47.23
10	Distal	1	120	80.50
11	Proximal	3	120	59.18
12	Distal	3	120	148.48
13	Proximal	1	160	36.58
14	Distal	1	160	75.83
15	Proximal	3	160	50.24
16	Distal	3	160	91.14
17	Proximal	2	140	40.50
18	Distal	2	140	123.70
19	Proximal	2	140	40.50
20	Distal	2	140	123.70

. All main factors were identified in importance by an ANOVA, as shown in

Table 12. Analysis of variance for principal stress (six locking screws).

Source	DF	Adj. SS	Adj. MS	F-Value	P-Value
Model	8	23,878.3	2984.8	38.81	0.000
Linear	3	20,979.5	6993.2	90.92	0.000
A	1	16,350.0	16350.0	212.58	0.000
B	1	2964.8	2964.8	38.55	0.000
C	1	1664.6	1664.6	21.64	0.001
2-Way Interactions	3	1930.8	643.6	8.37	0.004
A*B	1	831.7	831.7	10.81	0.007
A*C	1	449.9	449.9	5.85	0.034
B*C	1	649.2	649.2	8.44	0.014
3-Way Interactions	1	739.3	739.3	9.61	0.010
A*B*C	1	739.3	739.3	9.61	0.010
Error	11	846.0	76.9
Total	19	24724.3

R-Square = 96.58%, R-Square (adj.) = 94.06%.

. Furthermore, R-squared and adjusted R-squared statistics were analyzed for checking the fit of the dataset. Moreover, if the p-value of each main effect and its interaction was less than 0.05, it was considered statistically significant. According to this analysis, gap type (A), fracture gap (B), nail length (C), the interaction between variables A and B, the interaction between variables A and C, the interaction between variables B and C, and the three-way interaction A*B*C are factors that affect the maximum principal stress of the composite interlocking nail.

The R² value shows that the model explains 96.58% of the variance in the maximum principal stress, which indicates that the model fits the data extremely well.

The residual plot of the maximum principal stress is shown in Figure 11, representing the normal probability plot of the residuals. The residuals of the response randomly scatter around zero in the residual plots. Regarding the residuals versus the fits, the observation points (blue dots) scattered randomly in the normal probability plot show that the process parameters in this study were normal and independently distributed. The histogram of the residuals is approximately symmetric. Regarding the residuals versus the order, the residuals exhibit no clear pattern.

The magnitude and significance of an effect were determined by the Pareto chart shown in Figure 12. The Pareto chart shows the absolute values from the largest to the smallest effect of the standardized effects. The standardized effects are t-statistics that test the null hypothesis that the effect is zero [31]. In this Pareto chart, the bars that represent factors A, B, C, AB, ABC, BC, and AC cross the reference line, which is at 2.20. These factors are statistically significant at the 0.05 level with the current model terms. In the selected range, the infection order of each factor is A > B > C > AB > ABC > BC > AC.

3.3. Maximum Principal Stresses of the Composite Interlocking Nailing System

All values for the maximum principal stresses in the biocomposite interlocking nails are presented in Table 7, Table 9, and Table 11. In the proximal and middle shaft fracture, the interlocking nail’s maximum principal stress values were mostly below the compressive and tensile strength for cortical bone and PLA/PCL/15HA. In the distal shaft fracture, the interlocking nail’s stress values were mostly below the compressive and tensile strength for cortical bone but higher than the mechanical properties of PLA/PCL/15HA. In the proximal shaft fracture, the maximum principal stress contours occur at the middle screw at the top (Figure 13 top row). In the distal shaft fracture, the maximum principal stress contours occur at the distal screw holes that are fixed near the fracture gap at the top. However, in the middle shaft fracture, the maximum principal stress contours occur in the middle of the interlocking nail (Figure 13 bottom row), whether using interlocking nail lengths of 120, 140, or 160 mm in different fracture gaps. The maximum principal stress contours occur in similar areas. The new fixation technique, both in the two and four locking screws, can reduce the maximum principal stress that occurs at the proximal, middle, and distal locking hole of the composite interlocking nail. Figure 14 shows an example of the composite interlocking nails with a length of 120 mm at a 1 mm proximal shaft fracture using the P–D and P–M fixation techniques and at a 1 mm distal shaft fracture using the traditional P–D fixation technique and the new M–D fixation technique. The maximum principal stress occurs at the edge of the locking hole in all positions (proximal, middle, and distal locking screws) for the traditional fixation technique. The new fixation technique, however, reduced the maximum principal stress, and stress occurred only at the edges of the locking holes in the proximal–middle positions (proximal shaft fracture) and distal–middle positions (distal shaft fracture).

4. Discussion

In the current research, the maximum principal stresses on the canine femoral shaft fracture stabilized with the 3D-printed composite interlocking nail were estimated effects (gap type, fracture gap, nail length, and fixation techniques) based on a computer-based finite element analysis. Failure in implants is one surgical complication of interlocking nails in femoral shaft fracture treatment, and, according to previous research, interlocking nail failure occurs frequently at the middle shaft [32,33,34]. In comparison with the traditional fixation technique, the use of a composite interlocking nail system with new fixation techniques revealed lower maximal principal stresses. The new fixation techniques added middle locking screws, fixing middle fragments to be more stable, allowing better support for the stresses in the middle that usually occur in the interlocking nail without the middle holes, and permitting contact between the proximal and distal fragments. Increasing the locking hole distance can prevent the failure of the interlocking nail [35,36]. Therefore, the fractured bone can share the canine weight of the composite interlocking nailing system.

Nevertheless, some traditional fixation techniques (P–D and 2P–2D) were used to stabilize the proximal fragments and the distal fragments by locking screws rigidly, thus the proximal fragment cannot transfer the canine’s weight to the distal fragment [25]. Moreover, the distance between the screw positions that fix only the proximal or distal fragments (2P and 2D) without fixation in the middle of the canine bone makes the stress to the middle of the interlocking nail easy. That is why the new fixation techniques decreased the maximum principal stresses, which were not reduced by the traditional fixation technique. The stabilization of a fractured canine femur is related to several factors, consisting of the length and diameter of the composite interlocking nail system [37], fracture type [38], fracture gap, and the condition of the soft tissue of the canine femur [39]. This research established a relationship between the different fixation techniques and fixation stability.

The composite interlocking nail system with the new fixation technique had higher stability, as the proximal, middle, and distal locking screws (P–2M–D and 2P–2M–2D) remained in place. However, the proximal and distal fragments were not interconnected. In patients with bone formation problems, the lack of contact could lead to a fracture nonunion [40]. The new fixation techniques (2M, P–M, 2P–2M, 2M–2D, and M–D) have decreased the gap between the fractures and allowed proximal–distal contact between the fragments. The promotion of bone callus formation was reported [41]. The new fixation techniques remove the distal or proximal locking screws, but the wrong side would produce different clinical results. This rule states that the proximal locking screws can be removed (M–D and 2M–2D) for a distal shaft fracture and that the distal locking screws can be removed (P–M and 2P–2M) for a proximal shaft fracture [42,43,44].

For the two and four locking screws, according to the multilevel factorial design, gap type, fracture gap, and fixation techniques are the main factors that affect the maximum principal stress of the composite interlocking nail. Thus, this section discusses the relationship between these main factors because the nail length does not significantly impact the maximum principal stress. The fixation techniques were divided into two groups: traditional fixation techniques and new fixation techniques. For proximal shaft fractures of 1 mm, 2 mm, and 3 mm, the difference in the maximum principal stress of the new fixation techniques was lower than that of the traditional fixation technique group, and the maximum principal stresses in the interlocking nail were lower than the tensile stresses of the 3D-printed composite specimen (PLA/PCL/15HA) in a proximal shaft fracture of 1–2 mm (Figure 15 and Figure 16 top row). However, in the middle hole (Figure 13 top row, Figure 14), these stresses reached values that were very close to the tensile stresses of the 3D-printed composite specimen (52.05 MPa). For distal shaft fractures of 1 mm, 2 mm, and 3 mm, the maximum principal stresses in the interlocking nail were greater than the tensile stresses of the 3D-printed composite specimen in all models (Figure 15 and Figure 16, middle row). Such findings confirm the in vivo study in which all the interlocking nails failed at one of the locking holes nearer to the fractured line at the interface. Previous research also reported that interlocking nails frequently failed at the locking hole close to the fracture line [35,45]. The first locking hole at the distal site is the hole that is thought to be most at risk, especially if it is near the fracture site [35]. Figure 13 (middle row) shows that the maximum principal stresses occur in the first locking hole in all models. For middle shaft fractures of 1 mm, 2 mm, and 3 mm, the maximum principal stresses in the interlocking nail were lower than the tensile stresses of the 3D-printed composite specimen in all models (Figure 15 and Figure 16 bottom row), as well as the stress that occurs in the area between the middle locking holes near the fracture site (Figure 13, bottom row).

For the six locking screws in the proximal shaft fracture, according to the multilevel factorial design, gap type, fracture gap, nail length, and fixation techniques are the main factors that affect the maximum principal stress of the composite interlocking nail. Therefore, this paragraph discusses the relationship between the main factors. The maximum principal stresses in the interlocking nail were lower than the tensile stresses of the 3D-printed composite specimen (52.05 MPa) in conditions 1, 3, 4, 5, 6, 8, and 9. In contrast to the proximal shaft fracture, the maximum principal stresses in the interlocking nail were greater than the tensile stresses of the 3D-printed composite specimen (52.05 MPa) for the distal shaft fracture (Figure 17). In distal shaft fractures, the distal locking hole has a greater bending load than the middle locking hole because it is nearer to the fracture site and causes more interlocking nail failure [46,47]. A key factor in reducing the stress of these interlocking nails was the nail length necessary to be pushed down to the subchondral zone to increase the distance between the fracture locking hole [35]. At the distal shaft fracture, the interlocking nail with a length of 160 mm had less stress than the interlocking nails with lengths of 120 and 140 mm. Thus, smaller diameter nails with insufficient length should be avoided (where nail length does not reach the subchondral zone). Modifications to the designs that increase the distal thickness of the interlocking nail could decrease nail failure.

The new fixation techniques have decreased the maximum principal stresses of the composite interlocking nail in canine shaft fractures. Generally, each fixation technique has advantages and disadvantages. Due to modeling assumptions, there are some limitations. One of the limitations is the assumption that bone is a homogeneous isotropic material. This assumption is justified in this case as we modeled compact bone only. It cannot be assumed that the apparent density of the compact bone varies too much from point to point or that Young’s modulus has a single value. Moreover, the compact bone in long-bone diaphysis is transversely isotropic, but the assumption of bone isotropy was a reasonable approach to analyzing implant behavior. The effects of the screw threads were ignored for all the locking screws, and cylindrical rods were created for the locking screws. The main stresses of the composite locking nail method may also have been underestimated.

The mechanical properties and degradation ability of polymers are related to the material composition, molecular weight, density, glass transition point, and melting temperature [48,49,50]. These properties are the boundary for the material’s applications, and it could be improved to be used beyond its estimations by several techniques such as chemical treatments, thermal treatments, and blending with other materials. In the previous study, PCL blended with PLA could decrease the molecular weight reduction during the degradation period (up to 20 weeks), which prolonged the mechanical conditions and prevented degradation compared with a neat PLA [51]. According to the literature reviews, canine bone fracture healing times are approximately 16 weeks [52,53]. Therefore, the interlocking nails must provide proper mechanical properties during this healing period. In our previous study, the HA was added to accelerate bone formation by increased osteoblast proliferation [17]. Therefore, the healing period should be decreased as well. Moreover, the biodegradation test conditions were set to mimic the living body conditions, which performed under the material’s boundary conditions [17]. On day 30, the degradation was approximately 0.25% and did not present any significant difference compared with day 7 or day 15. The prediction equation of PLA/PCL/15HA’s biodegradation was Y = 0.0115X +0.1898 (where Y is the degradation percentage, and X is the degradation days) [17]. According to this equation, the degradation percentage of 16 weeks of implantation (bone completely healed) is approximately 1.48%. The slight degradation of the composite materials and the accelerated bone formation by HA might eliminate the mechanical problem of the PLA/PCL/15HA interlocking nail during the healing period. However, further investigations are still required.

Due to the biodegradability, biocompatibility, low cost, and reduced stress protection or stress protection atrophy that can occur using rigid metal interlocking nails [12,13], a 3D-printed composite material was tested in the present study [17,54,55]. The advantages of 3D-printed composite materials for bone fracture treatments were reported and include avoiding a later implant removal operation [56,57].

5. Conclusions

The biomechanical performance of the 3D-printed composite interlocking nail system with either the traditional or new fixation techniques was investigated using an FEA, and the effect of the maximum principal stress in the experimental boundary of factors including gap type (A), fracture gap (B), nail length (C), and fixation techniques (D) for two locking screws, four locking screws, and six locking screws were evaluated. From the FEA results, the 3D-printed composite interlocking nail system with the new fixation techniques presented a high possibility to be used in canine femoral fracture treatments at the proximal and middle positions, while the distal shaft fracture must be further improved and tested to reduce stress. The results of the present study can provide data for the veterinary orthopedic surgeon to understand the traditional and new fixation techniques. Moreover, veterinary orthopedic surgeons could decide on one of the fixation techniques based on the numerical results of this research and their patients’ specific situations.