3D-Printed Double-Helical Biodegradable Iron Suture Anchor: A Rabbit Rotator Cuff Tear Model

Suture anchors are extensively used in rotator cuff tear surgery. With the advancement of three-dimensional printing technology, biodegradable metal has been developed for orthopedic applications. This study adopted three-dimensional-printed biodegradable Fe suture anchors with double-helical threads and commercialized non-vented screw-type Ti suture anchors with a tapered tip in the experimental and control groups, respectively. The in vitro study showed that the Fe and Ti suture anchors exhibited a similar ultimate failure load in 20-pound-per-cubic-foot polyurethane foam blocks and rabbit bone. In static immersion tests, the corrosion rate of Fe suture anchors was 0.049 ± 0.002 mm/year. The in vivo study was performed on New Zealand white rabbits and SAs were employed to reattach the ruptured supraspinatus tendon. The in vivo ultimate failure load of the Fe suture anchors was superior to that of the Ti suture anchors at 6 weeks. Micro-computed tomography showed that the bone volume fraction and bone surface density in the Fe suture anchors group 2 and 6 weeks after surgery were superior, and the histology confirmed that the increased bone volume around the anchor was attributable to mineralized osteocytes. The three-dimensional-printed Fe suture anchors outperformed the currently used Ti suture anchors.


Introduction
Rotator cuff repair is one of the most common surgeries in the upper extremities [1]. The current trend in rotator cuff repairs is the use of suture anchors (SAs), such as metallic SAs and nondegradable or biodegradable polymer SAs [2,3]. The two most commonly In vitro mechanical tests were conducted to evaluate the mechanical characteristics of the SAs. The tests were performed using a 20-pound-per-cubic-foot (pcf) polyurethane foam block (part# 1522-03; Sawbone, Pacific Research Laboratories, Vashon, WA, USA). A No. 2 ultra-high molecular weight polyethylene fiber suture (Ultrabraid, Smith & Nephew, London, UK) with equal limbs was threaded through the suture eyelet, looped, and fixed over a post on the adapter before mechanical testing ( Figure 2). The static ultimate pullout strength was determined at a displacement rate of 1 mm/s. The mechanical tests were performed using an Instron E3000 (ElectroPuls, Instron, MA, USA). In vitro mechanical tests were conducted to evaluate the mechanical characteristics of the SAs. The tests were performed using a 20-pound-per-cubic-foot (pcf) polyurethane foam block (part# 1522-03; Sawbone, Pacific Research Laboratories, Vashon, WA, USA). A No. 2 ultra-high molecular weight polyethylene fiber suture (Ultrabraid, Smith & Nephew, London, UK) with equal limbs was threaded through the suture eyelet, looped, and fixed over a post on the adapter before mechanical testing ( Figure 2). The static ultimate pullout strength was determined at a displacement rate of 1 mm/s. The mechanical tests were performed using an Instron E3000 (ElectroPuls, Instron, MA, USA).

Corrosion Rate of Pure Fe SAs Using Static Immersion Tests
Static immersion tests were performed to evaluate the weight loss of the Fe SA. Six samples were weighed to obtain their initial weights (g) and surface areas (mm 2 ). The experimental specimens were immersed in 10 mL of Hanks' solution for 3 months and the temperature was maintained at 37 °C using a heating mantle. The samples were removed from the solution after every 30 days of treatment, ultrasonically cleaned, air dried, and weighed. The mass changes were used to calculate the corrosion rate (mm/year) using Equation (1), which is based on the ASTM G31 standard [29]: where CR is the corrosion rate, K is a constant (8.76 × 10 4 ), Δm (g) is the weight loss of the sample, ρ (g/cm 3 ) is the density of the object, A (cm 2 ) is the initial surface area of the sample, and t (h) is the immersion time of the sample in the solution.

In Vivo Animal Study Design
All animal experiments were approved by the Ethics Committee of the Biomedical Technology and Device Research Laboratories of the Industrial Technology Research Institute in accordance with national animal welfare legislation (approval no.: ITRI-IACUC-2020-050), and the study protocol conformed to the National Institute of Health guidelines for the use of laboratory animals. A total of 24 New Zealand white rabbits (Animal Health Research Institute of the Council of Agriculture) with a mean body weight of 4.0 ± 0.4 kg at the age of 6 months were selected. Each rabbit shoulder joint was randomized into experimental and control groups by using a computer-generated randomization method. In the control group, Ti SAs were implanted in the greater tuberosity of the humerus. In the experimental group, Fe SAs were implanted using the same surgical procedure as that in The dimensions of the Fe SA were 11.0 × 3.5 mm. (B) An inserter handle was designed to hold the Fe SA, which was inserted into a predrilled hole on the polyurethane foam block. (C) After inserting the SA and removing the inserter handle, the suture exited the core of the iron SA. (D) The control group: the non-vented screw-type Ti SA had a tapered tip portion with sutures through the proximal eyelet. The suture was passed through the eyelet at the top of the SA. The Ti SA was pulled out of the polyurethane foam block during the mechanical test.

Corrosion Rate of Pure Fe SAs Using Static Immersion Tests
Static immersion tests were performed to evaluate the weight loss of the Fe SA. Six samples were weighed to obtain their initial weights (g) and surface areas (mm 2 ). The experimental specimens were immersed in 10 mL of Hanks' solution for 3 months and the temperature was maintained at 37 • C using a heating mantle. The samples were removed from the solution after every 30 days of treatment, ultrasonically cleaned, air dried, and weighed. The mass changes were used to calculate the corrosion rate (mm/year) using Equation (1), which is based on the ASTM G31 standard [29]: where CR is the corrosion rate, K is a constant (8.76 × 10 4 ), ∆m (g) is the weight loss of the sample, ρ (g/cm 3 ) is the density of the object, A (cm 2 ) is the initial surface area of the sample, and t (h) is the immersion time of the sample in the solution.

In Vivo Animal Study Design
All animal experiments were approved by the Ethics Committee of the Biomedical Technology and Device Research Laboratories of the Industrial Technology Research Institute in accordance with national animal welfare legislation (approval no.: ITRI-IACUC-2020-050), and the study protocol conformed to the National Institute of Health guidelines for the use of laboratory animals. A total of 24 New Zealand white rabbits (Animal Health Research Institute of the Council of Agriculture) with a mean body weight of 4.0 ± 0.4 kg at the age of 6 months were selected. Each rabbit shoulder joint was randomized into experimental and control groups by using a computer-generated randomization method. In the control group, Ti SAs were implanted in the greater tuberosity of the humerus. In the experimental group, Fe SAs were implanted using the same surgical procedure as that in the control group. Four rabbits were euthanized immediately after SA implantation and frozen for biomechanical testing. The other rabbits were further divided into two subcategories based on implantation periods of 2 weeks and 6 weeks after surgery (10 rabbits in each group). Micro-CT and biochemical tests were performed 2 and 6 weeks after the surgery. Subsequently, the specimens were frozen for use in further biomechanical tests.

Surgical Methods
All surgical procedures were performed under general anesthesia by administering an intramuscular injection of a Zoletil-Rompun mixture (Zoletil 15 mg/kg; Rompun 0.05 mL/kg; Zoletil, Virbac Taiwan, Taipei, Taiwan; Rompun, Bayer Taiwan, Taipei, Taiwan). To induce analgesia, the rabbits were given intramuscular ketoprofen (2 mg/kg, ASTAR, Hsinchu, Taiwan) 24 h preoperation, and for 7 consecutive days following the surgery. For the prophylaxis of infection, the rabbits were given intramuscular Gentamycin (5 mg/kg, Standard Chem. & Pharm, Tainan, Taiwan) 24 h preoperation and for 7 consecutive days following the surgery.
Surgical procedures were performed following the method reported by Louati et al. with some modifications [30]. The supraspinatus insertion was sharply detached from the greater tuberosity, simulating a complete tear. A scalpel blade was used to decorticate the SSP footprint. A hole was predrilled on the lateral and distal to the footprint in the cortical bone with a 1.5 mm drill bit. In the experimental group, Fe SAs with No. 2 Ultrabraid sutures were inserted into the predrilled hole. The tendon was repositioned to the footprint using a modified Mason-Allen stitch. In the control group, Ti SAs with No. 2 Ultrabraid sutures were used. The tendon was repositioned to the footprint with the same technique as the experimental group ( Figure 3). The deltoid was closed, followed by skin closure. The position of the SAs was confirmed postoperatively using X-ray imaging ( Figure 4). All animals were euthanized after the experiments were completed by administering an intravenous overdose of pentobarbital.
Materials 2022, 15, x FOR PEER REVIEW 5 the control group. Four rabbits were euthanized immediately after SA implantatio frozen for biomechanical testing. The other rabbits were further divided into two su egories based on implantation periods of 2 weeks and 6 weeks after surgery (10 rabb each group). Micro-CT and biochemical tests were performed 2 and 6 weeks after th gery. Subsequently, the specimens were frozen for use in further biomechanical test

Surgical Methods
All surgical procedures were performed under general anesthesia by administ an intramuscular injection of a Zoletil-Rompun mixture (Zoletil 15 mg/kg; Rompu mL/kg; Zoletil, Virbac Taiwan, Taipei, Taiwan; Rompun, Bayer Taiwan, Taipei, Tai To induce analgesia, the rabbits were given intramuscular ketoprofen (2 mg/kg, AS Hsinchu, Taiwan) 24 h preoperation, and for 7 consecutive days following the sur For the prophylaxis of infection, the rabbits were given intramuscular Gentamy mg/kg, Standard Chem. & Pharm, Tainan, Taiwan) 24 h preoperation and for 7 con tive days following the surgery.
Surgical procedures were performed following the method reported by Louati with some modifications [30]. The supraspinatus insertion was sharply detached fro greater tuberosity, simulating a complete tear. A scalpel blade was used to decortica SSP footprint. A hole was predrilled on the lateral and distal to the footprint in the co bone with a 1.5 mm drill bit. In the experimental group, Fe SAs with No. 2 Ultra sutures were inserted into the predrilled hole. The tendon was repositioned to the print using a modified Mason-Allen stitch. In the control group, Ti SAs with No. trabraid sutures were used. The tendon was repositioned to the footprint with the technique as the experimental group ( Figure 3). The deltoid was closed, followed by closure. The position of the SAs was confirmed postoperatively using X-ray imaging ure 4). All animals were euthanized after the experiments were completed by admin ing an intravenous overdose of pentobarbital.

Micro-CT Analysis
After the rabbits were euthanized, 10 specimens were retrieved from each group, and multi-scale nano-CT (Skyscan 2211, Bruker Micro-CT, Kontich, Belgium) was used for 30 μm voxel resolution. A voltage of 155 kVp, an 80 μA current, and a 6 W output in microfocus mode with a 360° scan was used for the Ti implants. A voltage of 180 kVp, a 100 μA current, and an 18 W output in high-power mode with a 360° scan was used for the Fe implants.
To reposition the reconstructed cross-section and select the region of interest (ROI), the 3.5 mm implant column was isolated. Three mm (100 slices) images were used for the analysis. CTAn software was used for automatic Ostu thresholding and bone growth analysis. A 200-1000 μm region around the implant was defined as the ROI for bone growth analysis ( Figure 5). The bone and metal structures could be separated according to the differences in X-ray absorption. CTAn software with a shrink-wrap algorithm was employed to identify the border of the metallic structure. The tissue volume (TV, mm 3 ), bone volume (BV, mm 3 ), and bone surface area (BS, mm 2 ) were measured for the 200-1000 μm ROI surrounding the metallic implant in the bone. To determine the BV percentage, the BV/TV ratio (BV/TS, %) was calculated. The BS area per total volume (BS/TV, 1/mm) in the 3D analysis was used for characterizing the bone to implant contact. Subsequently, a "sphere-fitting" measurement method was used to analyze the entire object volume (OV, mm 3 ), object surface (OS, mm 2 ), and object thickness (OT, mm) [31][32][33]. Bone formation and implant degradation curve analyses were performed. Avizo software (Thermo Fisher Scientific, MA, USA) and CTVox (Bruker Micro-CT, Kontich, Belgium) were used for the 3D visualization.

Micro-CT Analysis
After the rabbits were euthanized, 10 specimens were retrieved from each group, and multi-scale nano-CT (Skyscan 2211, Bruker Micro-CT, Kontich, Belgium) was used for 30 µm voxel resolution. A voltage of 155 kVp, an 80 µA current, and a 6 W output in micro-focus mode with a 360 • scan was used for the Ti implants. A voltage of 180 kVp, a 100 µA current, and an 18 W output in high-power mode with a 360 • scan was used for the Fe implants.
To reposition the reconstructed cross-section and select the region of interest (ROI), the 3.5 mm implant column was isolated. Three mm (100 slices) images were used for the analysis. CTAn software was used for automatic Ostu thresholding and bone growth analysis. A 200-1000 µm region around the implant was defined as the ROI for bone growth analysis ( Figure 5). The bone and metal structures could be separated according to the differences in X-ray absorption. CTAn software with a shrink-wrap algorithm was employed to identify the border of the metallic structure. The tissue volume (TV, mm 3 ), bone volume (BV, mm 3 ), and bone surface area (BS, mm 2 ) were measured for the 200-1000 µm ROI surrounding the metallic implant in the bone. To determine the BV percentage, the BV/TV ratio (BV/TS, %) was calculated. The BS area per total volume (BS/TV, 1/mm) in the 3D analysis was used for characterizing the bone to implant contact. Subsequently, a "sphere-fitting" measurement method was used to analyze the entire object volume (OV, mm 3 ), object surface (OS, mm 2 ), and object thickness (OT, mm) [31][32][33]. Bone formation and implant degradation curve analyses were performed. Avizo software (Thermo Fisher Scientific, MA, USA) and CTVox (Bruker Micro-CT, Kontich, Belgium) were used for the 3D visualization.

Biomechanical Analysis
Two rabbits were euthanized immediately postimplantation, and ten rabbits (20 shoulders) were euthanized 2 and 6 weeks postoperation. Ten shoulders were retrieved and used for biomechanical analyses to investigate the failure load and the site of failure [30], which could explain the in vivo failure load and failure site of the repairs of SST using SAs. The SST and proximal humerus were isolated to ensure that only the SST attached to the humerus head contributed to the mechanical evaluation. The biomechanical testing was performed at room temperature (25 °C). After removing the redundant tissue, the humerus was embedded into a custom-made metallic clamp at the bottom of the testing machine (Instron E3343; ElectroPuls, Instron, MA, USA). The tendons were positioned along their anatomic direction of pull, at an angle of 45° to the longitudinal axis of the humeral shaft. This was followed by tensile loading to failure at 1 mm/s, where a 50% drop in tensile strength was defined as the breaking point. The load and displacement data were collected, and the mode of failure was noted. The load at failure was determined using Bluehill LE v3.71.4609 (Instron, Norwood, MA, USA). The failure modes were defined as failure at the tendon-suture junction, failure at the suture-anchor junction, and SA pullout.

Histological Analysis
Ten specimens were retrieved from each group for histological analysis. All the harvested samples were fixed in 10% formalin for 14 days and sequentially dehydrated with increasing concentrations of ethanol (70, 95, and 100%) for at least 1 d and infiltrated for 5 d by using polymethylmethacrylate [34,35]. After embedding, the samples were cut vertically, perpendicular to the long axis of the SA, at the level of the bone-implant interfaces. The sections were cut to approximately 150 μm in thickness by using a low-speed saw (IsoMet, Buehler, Lake Bluff, IL, USA) and ground to 60 μm by using a grinding and polishing machine [36]. The ground sections were stained with Sanderson's rapid bone stain (Dorn & Hart Microedge Inc., Loxley, AL, USA) and then counterstained with acid

Biomechanical Analysis
Two rabbits were euthanized immediately postimplantation, and ten rabbits (20 shoulders) were euthanized 2 and 6 weeks postoperation. Ten shoulders were retrieved and used for biomechanical analyses to investigate the failure load and the site of failure [30], which could explain the in vivo failure load and failure site of the repairs of SST using SAs. The SST and proximal humerus were isolated to ensure that only the SST attached to the humerus head contributed to the mechanical evaluation. The biomechanical testing was performed at room temperature (25 • C). After removing the redundant tissue, the humerus was embedded into a custom-made metallic clamp at the bottom of the testing machine (Instron E3343; ElectroPuls, Instron, MA, USA). The tendons were positioned along their anatomic direction of pull, at an angle of 45 • to the longitudinal axis of the humeral shaft. This was followed by tensile loading to failure at 1 mm/s, where a 50% drop in tensile strength was defined as the breaking point. The load and displacement data were collected, and the mode of failure was noted. The load at failure was determined using Bluehill LE v3.71.4609 (Instron, Norwood, MA, USA). The failure modes were defined as failure at the tendon-suture junction, failure at the suture-anchor junction, and SA pullout.

Histological Analysis
Ten specimens were retrieved from each group for histological analysis. All the harvested samples were fixed in 10% formalin for 14 days and sequentially dehydrated with increasing concentrations of ethanol (70, 95, and 100%) for at least 1 d and infiltrated for 5 d by using polymethylmethacrylate [34,35]. After embedding, the samples were cut vertically, perpendicular to the long axis of the SA, at the level of the bone-implant interfaces. The sections were cut to approximately 150 µm in thickness by using a lowspeed saw (IsoMet, Buehler, Lake Bluff, IL, USA) and ground to 60 µm by using a grinding and polishing machine [36]. The ground sections were stained with Sanderson's rapid bone stain (Dorn & Hart Microedge Inc., Loxley, AL, USA) and then counterstained with acid fuchsin. All bone-implant interfaces were examined using a light microscope (Nikon Eclipse Ti-series, Melville, NY, USA).

Biochemical Analysis
Blood samples were obtained before surgery and 2 weeks and 6 weeks after surgery. The blood serum was processed in an ISO 15189:2012 [37] accreditation of medical laboratories using an automated spectrophotometer (ADVIA Chemistry XPT System, Siemens Healthineers, Germany) for analysis of the following parameters: blood urea nitrogen (BUN), creatinine (Cr), alanine transaminase (ALT), and albumin (Alb).

Statistical Analysis
All experimental data are presented as the mean ± standard deviation. The Wilcoxon rank-sum test was used for the nonparametric analysis. A p-value of <0.05 was considered statistically significant. Statistical analysis was performed using SPSS statistics (version 26, Chicago, IL, USA). fuchsin. All bone-implant interfaces were examined using a light microscope (N Eclipse Ti-series, Melville, NY, USA).

Biochemical Analysis
Blood samples were obtained before surgery and 2 weeks and 6 weeks after surg The blood serum was processed in an ISO 15189:2012 [37] accreditation of medical la atories using an automated spectrophotometer (ADVIA Chemistry XPT System, Siem Healthineers, Germany) for analysis of the following parameters: blood urea nitro (BUN), creatinine (Cr), alanine transaminase (ALT), and albumin (Alb).

Statistical Analysis
All experimental data are presented as the mean ± standard deviation. The Wilco rank-sum test was used for the nonparametric analysis. A p-value of <0.05 was consid statistically significant. Statistical analysis was performed using SPSS statistics (ver 26, Chicago, IL, USA).

In Vivo Biomechanical Analysis
The biomechanical analysis results revealed no difference between the ultimate failure load of the Fe SA and that of the Ti SA (60.60 ± 15.28 N, p = 1.000) at 0 weeks (60.22 ± 28.73 N). At 2 weeks, there was no significant difference between the ultimate failure load of the Fe SA and Ti SA (69.94 ± 16.18 N and 53.36 ± 15.74 N, p = 0.093, respectively). At 6 weeks, there was a significant difference between the ultimate failure load of the Fe SA and Ti SA (116.64 ± 33.80 N and 52.14 ± 28.20 N, p = 0.043, respectively) ( Figure 8).

Failure Sites
At week 0, all repaired tendons failed at the tendon-suture junction in both the Fe and Ti SA groups. At week 2, there were five out of six (83.3%) failures at the tendonsuture junction, one out of six (16.7%) failures at the suture-anchor junction, and no SA was pulled out in the Fe SA group. In the control group, there were three out of six (50%)

In Vivo Biomechanical Analysis
The biomechanical analysis results revealed no difference between the ultimate failure load of the Fe SA and that of the Ti SA (60.60 ± 15.28 N, p = 1.000) at 0 weeks (60.22 ± 28.73 N). At 2 weeks, there was no significant difference between the ultimate failure load of the Fe SA and Ti SA (69.94 ± 16.18 N and 53.36 ± 15.74 N, p = 0.093, respectively). At 6 weeks, there was a significant difference between the ultimate failure load of the Fe SA and Ti SA (116.64 ± 33.80 N and 52.14 ± 28.20 N, p = 0.043, respectively) ( Figure 8).

In Vivo Biomechanical Analysis
The biomechanical analysis results revealed no difference between the ultimate failure load of the Fe SA and that of the Ti SA (60.60 ± 15.28 N, p = 1.000) at 0 weeks (60.22 ± 28.73 N). At 2 weeks, there was no significant difference between the ultimate failure load of the Fe SA and Ti SA (69.94 ± 16.18 N and 53.36 ± 15.74 N, p = 0.093, respectively). At 6 weeks, there was a significant difference between the ultimate failure load of the Fe SA and Ti SA (116.64 ± 33.80 N and 52.14 ± 28.20 N, p = 0.043, respectively) ( Figure 8).

Failure Sites
At week 0, all repaired tendons failed at the tendon-suture junction in both the Fe and Ti SA groups. At week 2, there were five out of six (83.3%) failures at the tendonsuture junction, one out of six (16.7%) failures at the suture-anchor junction, and no SA was pulled out in the Fe SA group. In the control group, there were three out of six (50%) Figure 8. In vivo biomechanical ultimate pullout strength assessment for different SAs at 0, 2, and 6 weeks after surgery. Mean ± SEM. * p < 0.05.

Failure Sites
At week 0, all repaired tendons failed at the tendon-suture junction in both the Fe and Ti SA groups. At week 2, there were five out of six (83.3%) failures at the tendon-suture junction, one out of six (16.7%) failures at the suture-anchor junction, and no SA was pulled out in the Fe SA group. In the control group, there were three out of six (50%) failures at the tendon-suture junction, two out of six (33.3%) failures at the suture-anchor junction, and one SA (16.7%) was pulled out from the bone. At week 6, there were four out of six (66.7%) failures at the tendon-suture junction ( Figure 9A), two out of six (33.3%) failures at the suture-anchor junction, and no SA was pulled out in the Fe SA group ( Figure 9B). In the control group, there were five out of six (83.3%) failures at the tendon-suture junction, and one SA (16.7%) was pulled out of the bone ( Figure 9C).
failures at the tendon-suture junction, two out of six (33.3%) failures at the suture-anchor junction, and one SA (16.7%) was pulled out from the bone. At week 6, there were four out of six (66.7%) failures at the tendon-suture junction ( Figure 9A), two out of six (33.3%) failures at the suture-anchor junction, and no SA was pulled out in the Fe SA group (Figure 9B). In the control group, there were five out of six (83.3%) failures at the tendonsuture junction, and one SA (16.7%) was pulled out of the bone ( Figure 9C).

Micro-CT Analysis
Micro-CT was performed to evaluate the bone formation between the implant and bone tissue. The Fe SA exhibited a higher postoperative BV/TV at 2 weeks (35.84 ± 3.80 vs. 27.18 ± 4.46, p = 0.003) and 6 weeks (33.47 ± 3.78 vs. 27.46 ± 2.14, p = 0.001) as compared to the Ti SA ( Figure 10A). The Fe SA exhibited a higher postoperative BS/TV at 2 weeks (5.66 ± 0.76 vs. 4.47 ± 0.53, p = 0.005) and 6 weeks (5.58 ± 0.89 vs. 4.36 ± 0.56, p = 0.005) as compared to the Ti SA ( Figure 10B). There was no difference in the BV/TV intra-group analysis between the 2-and 6-week samples in the Ti SA group (p = 0.453) and in the Fe group SA (p = 0.294). There was no difference in the BS/TV intra-group analysis between the 2-and 6-week samples in the Ti SA group (p = 0.860) and in the Fe group SA (p = 0.793). Figures  11 and 12 show examples of the Ti SA and Fe SA at 2 and 6 weeks, respectively, after the SA implantation.

Micro-CT Analysis
Micro-CT was performed to evaluate the bone formation between the implant and bone tissue. The Fe SA exhibited a higher postoperative BV/TV at 2 weeks (35.84 ± 3.80 vs. 27.18 ± 4.46, p = 0.003) and 6 weeks (33.47 ± 3.78 vs. 27.46 ± 2.14, p = 0.001) as compared to the Ti SA ( Figure 10A). The Fe SA exhibited a higher postoperative BS/TV at 2 weeks (5.66 ± 0.76 vs. 4.47 ± 0.53, p = 0.005) and 6 weeks (5.58 ± 0.89 vs. 4.36 ± 0.56, p = 0.005) as compared to the Ti SA ( Figure 10B). There was no difference in the BV/TV intra-group analysis between the 2-and 6-week samples in the Ti SA group (p = 0.453) and in the Fe group SA (p = 0.294). There was no difference in the BS/TV intra-group analysis between the 2-and 6-week samples in the Ti SA group (p = 0.860) and in the Fe group SA (p = 0.793). Figures 11 and 12 show examples of the Ti SA and Fe SA at 2 and 6 weeks, respectively, after the SA implantation.          The Fe SA degradation analysis showed an OV increase at 6 weeks as compared to that at 2 weeks (27.42 ± 0.81 vs. 26.71 ± 0.41, p = 0.021); however, the OS and OT at 2 and 6 weeks exhibited no significant difference (157.13 ± 1.20 vs. 156.89 ± 2.66, p = 0.173 and 0.64 ± 0.02 vs. 0.63 ± 0.01, p = 0.240, respectively), as shown in Figure 13. Figure 14 shows the reconstructed micro-CT images of the Fe SA at 2 and 6 weeks. The Fe SA degradation analysis showed an OV increase at 6 weeks as compared to that at 2 weeks (27.42 ± 0.81 vs. 26.71 ± 0.41, p = 0.021); however, the OS and OT at 2 and 6 weeks exhibited no significant difference (157.13 ± 1.20 vs. 156.89 ± 2.66, p = 0.173 and 0.64 ± 0.02 vs. 0.63 ± 0.01, p = 0.240, respectively), as shown in Figure 13. Figure 14 shows the reconstructed micro-CT images of the Fe SA at 2 and 6 weeks.

Histological Analyses
The mineralized bone formation was observed in the Ti SA and Fe SA groups at 2 and 6 weeks postoperation ( Figure 15). In all the histological analyses, mineralized osteocytes were observed in the region that closely contacted the SA. At 2 weeks, new bone formation was observed around the Fe SA, which penetrated the suture fiber. More degradation products were observed surrounding the FE SA at 6 weeks as compared to that of the Fe SA at 2 weeks. The Fe SA degradation analysis showed an OV increase at 6 weeks as compared to that at 2 weeks (27.42 ± 0.81 vs. 26.71 ± 0.41, p = 0.021); however, the OS and OT at 2 and 6 weeks exhibited no significant difference (157.13 ± 1.20 vs. 156.89 ± 2.66, p = 0.173 and 0.64 ± 0.02 vs. 0.63 ± 0.01, p = 0.240, respectively), as shown in Figure 13. Figure 14 shows the reconstructed micro-CT images of the Fe SA at 2 and 6 weeks.

Histological Analyses
The mineralized bone formation was observed in the Ti SA and Fe SA groups at 2 and 6 weeks postoperation ( Figure 15). In all the histological analyses, mineralized osteocytes were observed in the region that closely contacted the SA. At 2 weeks, new bone formation was observed around the Fe SA, which penetrated the suture fiber. More degradation products were observed surrounding the FE SA at 6 weeks as compared to that of the Fe SA at 2 weeks.

Histological Analyses
The mineralized bone formation was observed in the Ti SA and Fe SA groups at 2 and 6 weeks postoperation ( Figure 15). In all the histological analyses, mineralized osteocytes were observed in the region that closely contacted the SA. At 2 weeks, new bone formation was observed around the Fe SA, which penetrated the suture fiber. More degradation products were observed surrounding the FE SA at 6 weeks as compared to that of the Fe SA at 2 weeks.

Biochemical Analysis
Blood samples were collected from all the rabbits preoperation and at 2 and 6 weeks postoperation for biochemical analysis. Table 1 and Figure 16 show that the serum ALT level was increased gradually at 2 and 6 weeks postoperation. However, the serum Cr, Alb, and BUN levels were comparable at both time points.

Biochemical Analysis
Blood samples were collected from all the rabbits preoperation and at 2 and 6 weeks postoperation for biochemical analysis. Table 1 and Figure 16 show that the serum ALT level was increased gradually at 2 and 6 weeks postoperation. However, the serum Cr, Alb, and BUN levels were comparable at both time points.

Discussion
The results showed that the in vitro biomechanical data of the 3D-printed Fe SA was similar to that of the Ti SA ( Figure 6). The corrosion rate of the Fe SA was 0.049 ± 0.002 mm/year (Figure 7). The in vivo data in the rabbit model demonstrated that the Fe SA exhibited a higher pullout strength and showed more mineralized bone formation 2 and 6 weeks postoperation as compared to the Ti SA ( Figure 8).
Corrosion behavior is one of the most important factors for biodegradable materials. The optimal value of corrosion rate for plates and screws is approximately 0.5 mm/year [14]. If the corrosion rate is too high, the mechanical strength will deteriorate before healing, leading to implant failure. If the corrosion rate is too low, delays in the healing of the bone tissue might cause adverse effects. The in vitro corrosion rates vary with either the static immersion test or polarization test using different physiological solutions. The Fe SA employed in this study was produced using SLM technology and Fe powder with an Fe purity of > 99.5%. The corrosion rate calculated using the 3-month static immersion test was 0.049 ± 0.002 mm/year, which was similar to scaffolds produced using extrusionbased 3D printing with an Fe power purity of 99.88% (0.05 mm/year) [13], but was slower than that of pure Fe with a refined structure produced using electroforming (0.40 mm/year) [38], soft ingot Fe (>99.8% purity) in the form of 2 mm thick sheets produced using cross-rolling (0.11-0.14 mm/year) [39], and composites produced using SLM with C nanotubes/Fe (99% purity) (0.085 mm/year) [40]. Overall, pure Fe's corrosion rate ranges from 0.049 to 0.4 mm/year. In comparison to biodegradable Fe, the in vitro corrosion rate of pure Mg ranges from 0.33 to 0.99 mm/year [17,41,42], and that of pure Zn ranges from 0.014 to 0.75 mm/year [43][44][45]. The in vitro corrosion rate of pure Fe is slower than that of pure Mg and Zn.
The architecture of small mammal rotator cuffs shows greater similarity to that of humans than that of large mammals [27]. Some rabbit rotator cuff tear studies have been proposed to demonstrate the outcomes of repairing these using SAs [30,46]. Chaler et al. [46] demonstrated that repairing SST ruptures using SAs yielded higher loads to failure immediately after the repair and after the first postoperative week as compared to the transosseous repair. Louati et al. [30] demonstrated the outcomes of the channeling and

Discussion
The results showed that the in vitro biomechanical data of the 3D-printed Fe SA was similar to that of the Ti SA ( Figure 6). The corrosion rate of the Fe SA was 0.049 ± 0.002 mm/year (Figure 7). The in vivo data in the rabbit model demonstrated that the Fe SA exhibited a higher pullout strength and showed more mineralized bone formation 2 and 6 weeks postoperation as compared to the Ti SA ( Figure 8).
Corrosion behavior is one of the most important factors for biodegradable materials. The optimal value of corrosion rate for plates and screws is approximately 0.5 mm/year [14]. If the corrosion rate is too high, the mechanical strength will deteriorate before healing, leading to implant failure. If the corrosion rate is too low, delays in the healing of the bone tissue might cause adverse effects. The in vitro corrosion rates vary with either the static immersion test or polarization test using different physiological solutions. The Fe SA employed in this study was produced using SLM technology and Fe powder with an Fe purity of > 99.5%. The corrosion rate calculated using the 3-month static immersion test was 0.049 ± 0.002 mm/year, which was similar to scaffolds produced using extrusion-based 3D printing with an Fe power purity of 99.88% (0.05 mm/year) [13], but was slower than that of pure Fe with a refined structure produced using electroforming (0.40 mm/year) [38], soft ingot Fe (>99.8% purity) in the form of 2 mm thick sheets produced using cross-rolling (0.11-0.14 mm/year) [39], and composites produced using SLM with C nanotubes/Fe (99% purity) (0.085 mm/year) [40]. Overall, pure Fe's corrosion rate ranges from 0.049 to 0.4 mm/year. In comparison to biodegradable Fe, the in vitro corrosion rate of pure Mg ranges from 0.33 to 0.99 mm/year [17,41,42], and that of pure Zn ranges from 0.014 to 0.75 mm/year [43][44][45]. The in vitro corrosion rate of pure Fe is slower than that of pure Mg and Zn.
The architecture of small mammal rotator cuffs shows greater similarity to that of humans than that of large mammals [27]. Some rabbit rotator cuff tear studies have been proposed to demonstrate the outcomes of repairing these using SAs [30,46]. Chaler et al. [46] demonstrated that repairing SST ruptures using SAs yielded higher loads to failure immediately after the repair and after the first postoperative week as compared to the transosseous repair. Louati et al. [30] demonstrated the outcomes of the channeling and nochanneling techniques in the repair of SST ruptures using SAs. Both animal studies showed no difference in the load to failure after 4 weeks of SA implantation [30,46]. Consequently, the authors hypothesized that 6 weeks after SA implantation is a long enough period to examine the ultimate load to failure. The in vivo biomechanical analysis mimicked the supraspinatus muscle pullout direction. The biomechanical results showed the most common failure site was at the tendon-suture junction, implying that the tendon was cut by the sutures (Figure 9). Although the load to failure was similar between the two groups, there was no SA pullout failure in the Fe SA group but there was one SA pullout failure in the Ti SA group.
Micro-CT was used to quantify the BV fraction (BV/TV) and BS density (BS/TV) of the ROI surrounding the SA [47]. A higher BV/TV is better for bone growth and a higher BS/TV indicates an increased bone growth closer to the implant surface region. The results of this study demonstrate that the Fe SA yielded increased bone growth as compared to the Ti SA, especially in the region near the implant surface. This result corresponded to the histology results, where the BV, identified using micro-CT, was mineralized bone. Due to the open-construct design of the Fe SA, new mineralized bone growth into the core of the anchor occurred as early as week 2 ( Figure 14). The micrograph demonstrated new bone grew into the Fe SA and embedded the sutures, which may provide stability, even if degradation occurred in the Fe SA ( Figure 15). In our previous study, the cell sensitivity assays showed that the Fe SA exhibited no cytotoxicity and lamellipodial extrusions from the cells, and the attachment on the implant surface could be identified using a scanning electron microscope [25]. Additionally, no significant Fe deposition in the visceral organs was found [25]. Overall, Fe SA is a biocompatible implant and shows better new mineralized bone growth around the implant.
In this study, a mild elevation in serum ALT was noted. The biocompatibility of the Fe SA was confirmed in a previous study of Fe SA implantation in rabbit tibia for three months [25], in which the histopathology of Prussian blue staining of the liver showed that no Fe was detected in the Fe SA group. Additionally, the percentage of Fe stores in the spleen in the Fe SA group showed no significant difference from those in the polymer SA group. In this study, some degradation products surrounding the Fe SA were found. A longer-term follow-up study is needed to confirm any occurrence of local inflammation reactions after further degradation of the Fe SA.
This study showed a similar short-term ultimate pullout strength between Ti and Fe SAs. Because of the structural difference, the stress concentration during the ultimate pullout test was located around the proximal site in the Ti SAs and around the distal site in the Fe SAs. A biomechanical study conducted in a polyurethane foam block and porcine bone showed a similar ultimate pullout strength and cyclic loading test results between open-construct coil-type PEEK SAs and Ti non-vented screw-type SAs [48]. The authors cannot attribute this to the superior biomaterial properties of Ti or Fe or the structural design of either the non-vented screw-type or open-construct coil-type SAs. Future studies comparing the same open-construct coil-type design with different biomaterials may determine whether nondegradable Ti or biodegradable Fe is superior for use in SAs.
3D printing technology allows for the rapid fabrication of porous implants and composite scaffolds which can be tailored to different orthopedic applications [25,49,50]. The Fe SA used in this study was produced by laser sintering using high purity biodegradable spherical Fe powder. 3D printing selective laser sintering technology allowed the accurate construction of the two-helical structure and the incorporation of porous structures in the thread, which could facilitate bone ingrowth to stabilize the anchor. Overall, the 3D printed open-construct biodegradable Fe SA showed good initial mechanical pullout strength, mineralized bone growth, and optimal gradual degradation.
This study had some limitations that should be addressed. First, the rabbit SST tear study analyzed the osteointegration between the bone and SA using histological evidence and micro-CT results. The ultimate pullout strength of the SST was evaluated. However, whether the occurrence of local inflammation reactions during Fe degradation may influence the SST healing requires further study. Second, this study showed that biodegradable Fe SA promoted better bone growth than commercialized Ti SAs. However, the two SAs in the experimental and control groups had different material properties and structural designs. Further studies using the same SA design with different materials are needed to clarify which material (Fe, Ti, or PEEK) is superior.

Conclusions
The in vitro test confirmed that the static biodegradable property of Fe SAs was 0.049 ± 0.002 mm/year and the pullout strength was similar to that of the Ti SAs. The micro-CT and histology confirmed that the Fe SAs exhibited better mineralized bone growth around the Fe SA as compared to the Ti SA in the rabbit rotator cuff tear model. Overall, the open-construct and double-helical Fe SAs produced using 3D printing technology could outperform the currently used Ti SAs in terms of their biomechanical properties and osteointegration capacity. Future research should focus on the biocompatibility of the long-term implantation of Fe SAs and the in vivo degradation rate of the biodegradable Fe SAs.