3D printed locking osteosynthesis screw threads have comparable strength to machined or hand‐tapped screw threads

Additive manufacturing, aka three dimensional (3D) printing, is increasingly being used for personalized orthopedic implants. Additively manufactured components normally undergo further processing, in particular 3D printed locking osteosynthesis plates require post‐printing screw thread creation. The aim of this study was to compare 3D printed threads with machined and hand‐tapped threads for a locking plate application. Pushout tests were performed on 115 additively manufactured specimens with tapered screw holes; additive manufacture was performed at 0°, 20°, 45°, or 90° build orientations. The screw holes were either machined, hand‐tapped or 3D printed. The 3D printed screw holes were left as printed, or run through with a tap lubricated with water or with thread cutting oil. Printed threads run through using oil, with a build orientation of 90°, had comparable pushout force (median: 6377 N 95% confidence interval [CI]: 5616‐7739 N) to machined (median: 6757 N; 95% CI: 6682‐7303 N) and hand‐tapped (median: 7805 N; 95% CI: 7154‐7850 N) threads. As printed threads and those run through using water had significantly lower pushout forces. This study shows for the first time that 3D printed screw threads for a locking osteosynthesis plate application have comparable strength to traditionally produced screw threads.

for use is not correct. Additively produced components still require postprocessing procedures. In the case of locking osteosynthesis plates the screw threads need to be machined after the component has been printed. The postprocessing operations add to the cost of the component as well as potentially being a source of deviation from the original planned 3D geometry.
As the 3D printing technology improves greater fidelity between the printed component and the source geometry can be achieved. Recently it has become possible to 3D print locking threads. This potentially will allow the production of high fidelity complex parts at lower cost; however it is not known how the performance of printed threads is influenced by 3D build parameters. Orientation is known to influence the mechanical properties of Ti-6Al-4V 7 the most popular material currently used for additively manufactured implants. Screw threads are a challenging feature to include in additively manufactured parts due to the small size of thread detail in absolute terms as well as in relation to the rest of the part. Additive manufacturing machine parameters are generally optimized to a specific scale and this means they may not be optimal for both the global part and the small thread features.
The aim of this study was to evaluate how 3D printed threads compare to those made by five axis machining and by hand tapping in terms of pushout strength.

| Test specimens
Test specimens were designed as rectangular blocks containing either a non-threaded pilot hole or a threaded screw hole suitable for 5 mm diameter locking screws, with a head major diameter tapering from 6.7 mm to 5.9 mm, appropriate for a personalized high tibial osteotomy plate. 8 The blocks were 16 × 16 mm with depths of 3.58, 4.08, 4.58, and 5.08 mm corresponding to 3, 4, 5, and 6 threads, respectively ( Figure 1). The locking thread had a 14°taper, 0.5 mm pitch, 1.0 mm lead, and 0.3 mm thread height. The test specimens used were additively manufactured using selective laser sintering (Renishaw AM250; Renishaw plc, Wotton-under-Edge, UK) with titanium alloy (Ti-6Al-4V) grade 23.The powder particle size was between 15 and 45 μm, the laser power was 200 W and the layer thickness was 40 µm.The following laser parameters were used: 80 µm point distance and 50 µs exposure time for borders; 60 µm point distance and 50 µs exposure time for fill hatching. Four borders were used with a hatch offset of 160 µm. Specimens were additively manufactured in different orientations (0°, 20°, 45°, 90°) to examine the influence of this parameter on thread connection strength. The 0°, 20°, and 90°orientations are illustrated in Figure 2. Prior to testing, the supporting scaffold structures ( Figure 2) were removed so that all test specimens could be placed flat for testing. The custom blocks were tested in combination with 5 mm orthopedic locking screws made from titanium alloy ( Figure 3).

| Thread production or processing
Five different thread production/processing methods were employed. The test specimens produced with a pilot hole had threads produced either by machining on a five-axis mill (5Axis group) using a v-point cutting tool (Scorpion Tooling UK Ltd, Dursley, UK) or manually (Hand Tap group) using a tapered drill (Scorpion Tooling UK Ltd) and a custom tapered tap (Mercury Tool and Gauge Ltd, Coventry, UK). The test specimens produced with a 3D printed thread were either left as printed (RawPrinted group), or were run through with the custom tap using either water (Printed Tapped Water group) or cutting oil (Rocol Cutting Fluid, Rocol, Leeds, UK) (Printed Tapped Oil group) as a lubricant. Note for these two last groups the tapping operation effectively cleaned/repaired the 3D printed thread rather than creating a new thread. Table 1 lists the combinations of thread type, build orientation and thread count tested as well as the number of specimens per combination. The original intention was to test five specimens for set of considered parameter combinations, however some combinations displayed higher variability during testing and the numbers of specimens were increased for these to the maximum number of specimens available, which varied between 6 and 15.

| Testing
Prior to commencing the tests, the locking screws were tightened into each test specimen to a measured torque of 14 Nm with an

| Analysis
The maximum pushout force and corresponding extension were extracted from the Instron tests data files using a custom Matlab script T A B L E 1 Summary of thread groups tested, with details of build orientation, number of threads, and number of specimens  The CT scan images demonstrated that the fidelity of the threads was found to be lowest in the RawPrinted group with the 5Axis and PrintedTappedOil group show much more clearly defined threads ( Figure 8).

| DISCUSSION
Additive manufacture has been adopted into a number of orthopedic applications 1,2 ; one of the most disruptive applications is the relatively low cost, compared with traditional manufacturing methods, production of personalized implants. 9 However, 3D printed parts still require considerable finishing processes. In the manufacture of locking osteosynthesis plates this usually includes cutting of screw threads, either by hand tapping or further machining. 10 Additional post-printing manufacturing steps increase production costs and introduce risk of deviation from the original planned component geometry, particularly for anatomic parts lacking readily identifiable datums. In some cases, particularly in veterinary applications, the threads may be too small to be additively manufactured, 10  90°. For the particular thread geometry we used, the highest pushout was achieved for the 90°build ( Figure 5); again supporting the proposal that highest build fidelity occurs when the thread axis is orientated with the build direction. The fact that build orientation had no effect on pushout force for the 5Axis group ( Figure 6) suggests further that it is geometric matching of the corresponding screw faces which is important for pushout strength. This is supported by the CT scans demonstrating that the groups which produced larger pushout forces also had higher geometric fidelity around the threads (Figure 8).
The cleaning/repair of the geometry by running a tap through the printed thread gave rise to increases in pushout force. It was interesting to note that use of thread cutting oil gave a significantly higher pushout force compared with using water. Thread cutting oils are formulated to cool the tool and aid with chip removal; we propose the use of cutting oil will increase the geometric fidelity of the thread. The PrintedTappedOil group managed to achieve comparable pushout strengths to the 5Axis and HandTap groups.
This study used a quasi-static push-out test as a measure of threadinterlock strength; however,the long-term strength of the thread interlock will also be influenced by the cyclic behavior of the printed material. This aspect should therefore be investigated in the future.This study has used pushout force as an indicator of screw F I G U R E 8 CT scans through specimens showing the details of the thread geometries, (A) as printed, (B) 5 axis, and (C) printed and tapped with oil. CT, computed tomography thread efficacy. In reality, the thread junction will be subject to a more complex loading environment. However, this approach allowed us to compare the different thread groups in a straightforward manner. To our knowledge this is the first study to assess the efficacy of 3D printed threads compared with traditionally created threads for custom 3D printed osteosynthesis plates.
In conclusion, 3D printed threads created with a build orientation of 90°and cleaned/repaired by running a tap lubricated with thread cutting oil achieve comparable pushout strength to traditional hand or machine created threads.

ACKNOWLEDGMENT
The lead author was funded by Arthritis Research UK (grant number: 21495). We thank Renishaw plc for providing the test specimens.