Fractures of the anteromedial facet (AO/OTA 21-B1.1, O'Driscoll Type 2, subtype 3) are associated with varus posteromedial rotational instability of the ulnohumeral joint and early post-traumatic arthritis. The purpose of this study was to examine the stability of plate (locking and non-locking) vs screw constructs in the fixation of anteromedial coronoid facet fractures in a sawbone model.

An anteromedial coronoid facet fracture (AO/OTA 21-B1.1) was simulated in 24 synthetic ulna bones. They were then assigned into 3 fracture fixation groups: non-locking plate fixation, locking plate fixation, and dual cortical screw fixation. An AO 2.0 mm screw and plate system was used for the plate fixation groups and 2.0 mm cortical screws were used for the screw-only group. Following fixation, each construct was potted in bismuth alloy and secured to a servohydraulic load frame. Each construct was cycled in tension and then in compression at 0.5Hz. For both cycling modalities, an incremental loading pattern was used starting at 40 N and increased by 20 N every 200 cycles up to 200N. Fracture fragment displacement was recorded with an optical tracking system. Following cyclic loading each construct was loaded to failure (displacement >2 mm) at 10mm/min.

Tension cycling – All constructs in the plated groups (locking and non-locking constructs) survived the cyclic tension loading protocol (to 200N) with maximum fragment displacement of 12.60um and 14.50um respectively. There was no statistical difference between the plated constructs at any load level. No screw-only fixed construct survived the tension protocol with mean force at failure of 110N (range 60–180N).

Compression Testing – All constructs in the plated groups (locking and non-locking constructs) survived the cyclic compression loading protocol (to 200N), while all but one of the screw-only fixation constructs survived. Fracture fragment displacement was significantly greater in the screw-only repair group across all loading levels when compared to the plated constructs. There was no statistically significant difference in fragment motion between the locking and non-locking groups.

Failure Testing – The maximum load at failure in the screw-only group (281.9 N) was significantly lower than locking and non-locking constructs (587.0 N and 515.5N respectively, p <0.05). There was no difference between the locking and non-locking group in mean load to failure or mean stiffness. Screw construct stiffness (337.2 N/mm) was lower than the locking and non-locking constructs (682.9 N/mm and 479.1 N/mm respectively) however this did not reach statistical significance (p=0.051).

Fixation of anteromedial coronoid fractures is best achieved with a plating technique. Locking plates did not offer any advantage over conventional plates. Isolated screw fixation might not provide adequate stability for these fractures which could result in loss of reduction leading to post-traumatic arthrosis or instabilility.

Nearly a quarter of screws cause damage during insertion by stripping the bone, reducing pullout strength by over 80%. Studies assessing surgically achieved tightness have predominately shown that variations between individual surgeons can lead to underpowered investigations. Further to the variables that have been previously explored, several basic aspects related to tightening screws have not been evaluated with regards to how they affect screw insertion. This study aims to identify the achieved tightness for several variables, firstly to better understand factors related to achieving optimal intraoperative screw purchase and secondly to establish improved methodologies for future studies. Two torque screwdrivers were used consecutively by two orthopaedic surgeons to insert 60 cortical, non-locking, stainless-steel screws of 3.5 mm diameter through a 3.5 mm plate, into custom-made 4 mm thick 20 PCF sheets of Sawbone, mounted on a custom-made jig. Screws were inserted to optimal tightness subjectively chosen by each surgeon. The jig was attached to a bench for vertical screw insertion, before a further 60 screws were inserted using the first torque screwdriver with the jig mounted vertically, enabling horizontal screw insertion. Following the decision to use the first screwdriver to insert the remaining screws in the vertical position for the other variables, the following test parameters were assessed with 60 screws inserted per surgeon: without gloves, double surgical gloves, single surgical gloves, non-sterile nitrile gloves and, with and then without augmented feedback (using digitally displayed real-time achieved torque). For all tests, except when augmented feedback was used, the surgeon was blinded to the insertion torque. Once the stopping torque was reached, screws were tightened until the stripping torque was found, this being used to calculate tightness (stopping/stripping torque ratio). Screws were recorded to have stripped the material if the stopping torque was greater than the stripping torque. Following tests of normality, Mann-Whitney-U comparisons were performed between and combining both surgeons for each variable, with Bonferroni corrections for multiple comparisons. There was no significant (p=0.29) difference in the achieved tightness between different torque screw drivers nor different jig positions (p=0.53). The use of any gloves led to significant (p < 0 .001) increases in achieved tightness compared to not using gloves for one surgeon but made no difference for the other (p=0.38–0.74). Using augmented feedback was found to virtually eliminate stripping. For one surgeon average tightness increased significantly (p < 0 .001) when torque values were displayed from 55 to 75%, whilst for the other, this was associated with significantly decreases (p < 0 .001), 72 to 57%, both surgeons returned to their pre-augmentation tightness when it was removed. Individual techniques make a considerable difference to the impact from some variables involved when inserting screws. However, the orientation of screws insertion and the type of screwdriver did not affect achieved screw tightness. Using visual feedback reduces rates of stripping and investigating ways to incorporate this into clinical use are recommended. Further work is underway into the effect of other variables such as bone density and cortical thickness