This study investigated the effect of the articulation position on joint load transfer in total elbow arthroplasty. To quantify loading, an adjustable humeral component, instrumented with a load cell, was developed to measure ulnohumeral loads in-vitro. Computer guidance was implemented to accurately place the linked articulation into eight cadaveric elbows. Axial compression and bending about the flexion axis produced the greatest loads during simulated active elbow flexion. An anteriorly malpositioned flexion-extension axis resulted in increased joint loads during flexion. Translational positional errors were more influential than rotational position on articular loading.

To quantify the relationship between total elbow arthroplasty position and elbow joint loading.

Eight cadaveric upper extremities were tested using a motion-controlled testing device, which simulated muscle activity. Computer guidance was employed to accurately position a linked implant consisting of a custom-designed adjustable humeral component and commercial ulnar component. The testing apparatus was instrumented with a six-degree-of-freedom load cell to measure axial and bending loads. Seven implant positions were tested including anterior-posterior translation (−5.0, −2.5, 0.0, 2.5, 5.0 mm) and internal-external rotation (−5, 0, 5°) during supinated and pronated flexion.

The resultant joint force decreased for all prosthetic hinge positions as elbow flexion increased (p< 0.001). Anterior hinge positions produced greater ulnohumeral loads (p< 0.001) and moments (p< 0.001) than posterior hinge positions during simulated elbow flexion. The greatest bending moment occurred about the flexion axis which reached maximum magnitudes during mid-flexion. Implant hinge malrotation did not have a significant effect on axial (p=0.07) or bending (p=0.6) forces experienced at the joint. The distance between the flexion (hinge) axis and the muscular line-of-action of flexors is reduced with anterior hinge placement, likely increasing the force necessary to produce flexion.

An anteriorly malpositioned flexion-extension axis resulted in an increase in joint loading during flexion and should be avoided during elbow arthroplasty.

This is the first reported study to measure the effect of elbow prosthesis positioning on joint loading. The knowledge gained about joint loads should improve future prosthetic designs and treatment options.

Funding: Canadian Institute for Health Research.

Please contact author for graphs and/or diagrams.

Passive and active elbow flexion was performed in eight cadaveric arms to determine the effect of Type 1 coronoid fractures and suture repair on kinematics. Testing was performed in ligamentously intact and MCL deficient elbows; with radial head arthroplasty (RHA); with an intact coronoid, following a Type 1 fracture, and with suture repair of the coronoid. There was an alteration in elbow kinematics and stability following Type 1 coronoid fractures that was not corrected with coronoid repair. Suture fixation of the coronoid is probably unnecessary if the lateral ligaments are repaired and the radial head is repaired or replaced.

To determine the effect of fixation of Type 1 coronoid fractures on elbow stability and kinematics in ligamentously intact and medial collateral ligament (MCL) deficient elbows with radial head arthroplasty (RHA).

Type 1 coronoid fractures cause changes in elbow kinematics and stability that are not corrected with suture repair.

Suture fixation of Type 1 coronoid fractures is probably unnecessary if the lateral ligaments are repaired and the radial head is repaired or replaced.

With intact ligaments, there was an increase in valgus angulation following a Type 1 coronoid fracture (p< 0.05) that was not corrected with fixation. With MCL deficiency, there was no change in valgus angulation for all coronoid states. For both ligament states, there was an increase maximum varus-valgus laxity after a Type 1 coronoid fracture with forearm pronation (p=0.03) that was not corrected with fixation (p=0.4). Kinematic data was collected from eight cadaveric arms during passive and simulated active elbow motion. The protocol was performed in stable and MCL deficient elbows with RHA. Testing occurred with the coronoid intact, following Type 1 coronoid fracture, and with suture repair of the fracture. Valgus angulation and maximum varus-valgus laxity were measured.

With intact ligaments, Type 1 coronoid fractures cause an alteration in elbow kinematics and laxity that is not corrected with suture fixation. With MCL disruption, Type 1 coronoid fractures have no effect on elbow kinematics and a small effect on laxity that is not corrected with coronoid repair.

Funding: Research and Institutional Support received from Wright Medical Technologies.

Please contact author for graphs and/or diagrams.

The strength of the intact and four reconstruction techniques (figure-eight, docking, single strand utilizing interference screws, and a single strand) of the medial collateral ligament (MCL) of the elbow were compared. Twenty cadaveric specimens were tested with a cyclic valgus loading protocol. The peak loads to failure of the MCL reconstructions were inferior compared to the intact ligament (p< 0.05). The docking and single strand reconstruction utilizing an endobutton for ulnar fixation were equivalent and had greater initial strength than the interference screws or figure-eight technique. It is suggested that improved interference screws are required for this repair.

The purpose of this study was to compare the initial strength of the intact medial collateral ligament (MCL) of the elbow and four reconstruction techniques.

The docking and endobutton reconstructions showed equivalent peak load to failure.

Improved interference screws are required before they are employed clinically.

The average peak load to failure or 5mm of joint gapping was 142.5±39.4N for the intact, 53.0±9.5N for the docking, 52.5±10.4N for the endobutton, 41.0±16.0N for the interference screw, and 33.3±7.1N for the figure-eight reconstructions. The peak load to failure was higher for the intact specimens compared to any of the reconstructions (p< 0.001). The docking reconstruction showed higher peak loads than the figure-eight or interference screw reconstruction, and the endobutton reconstruction showed higher peak loads than the figure-eight reconstruction (p< 0.004). There was no difference in peak loads between the docking and endobutton reconstructions (p> 0.05).

Twenty (ten pairs) unpreserved cadaveric upper extremities were mounted in a custom jig with the elbow at 90°, and a valgus force was applied 12cm from the elbow joint. The specimens were loaded starting at 20N with the load increased in increments of 10N (200 cycles at each load), until either complete ligament failure or a 5mm increase in the distance between the attachment sites of the MCL. The results support that a single strand or multistrand ligament reconstruction can be equivalent with respect to maximal peak loads and cyclic loading. There are concerns with regard to the use of interference screw fixation in the clinical situation.

Purpose: Metal radial head arthroplasty is a proven technique for the treatment of complex radial head fractures. Little previous research is available on the utility and longevity of metal radial head arthroplasty for elbow reconstruction. The purpose of this study was to evaluate the functional outcome of patients with metal radial head arthroplasty (RHA) for elbow reconstruction.

Methods: This was a retrospective review from one institution with three senior orthopaedic surgeons. 23 consecutive patients with 23 RHA were included in the study, 4 patients were lost to follow up. RHA was performed for conditions of rheumatoid arthritis, post traumatic radial head nonunion, post traumatic radial head malunion, elbow instability following previous radial head excision and failed silicone radial head implants. Patients were excluded from the study if they were treated for an acute injury with RHA (under 2 months from injury). There were 8 males and 15 females with a minimum of 2 year follow up. Analysis included chart review, personal interview, physical examination, radiographs and strength testing. Region specific questionnaires were used including: DASH, ASES, MEPI, PRWE, WOS and SF-36. Mean follow up was 9 yrs.

Results: Mayo elbow performance scores were excellent in 11 patients, good in 4 patients, fair in 4 patients and poor in 4 patients. Subjective patient satisfaction was averaged at 8.4 out of 10. Patients demonstrated significantly less isometric strength and grip strength in the affected versed unaffected elbow. There was no statistically significant difference seen in ulnar variance or ulnohumeral joint space between the affected and unaffected limbs.

Conclusions: In conclusion, RHA for elbow reconstruction is a safe procedure that provides patients with long term functional ROM and pain relief.

The purpose of this study was to determine if arthroscopic release is safe and effective in the management of elbow contracture. Twenty patients (mean age of forty-two), undergoing arthroscopic contracture release were retrospectively reviewed at a minimum follow-up of one year (mean twenty-five months). Most patients had combined extrinsic & intrinsic causes for contractures. Mean flexion improved from 122 to 137°. Mean extension improved from 38° to 18°. The mean arc improvement was 35° (p < 0.001). None of the patients had instability and there were no major neuro-vascular complications. All patients had decreased pain and improved elbow function.

To determine if arthroscopic release is safe and effective in the management of elbow contracture.

Twenty patients (mean age of forty-two), undergoing arthroscopic contracture release were retrospectively reviewed at a minimum follow-up of one year (mean twenty-five months). Most patients had combined extrinsic & intrinsic causes for contractures. Motion and strength were measured with standard goniometry and the LIDO isokinetic system by independent evaluators.

Mean flexion improved from 122 ± 16° to 137 ± 12°. Mean extension improved from 38 ± 18° to 18 ± 14°. The mean arc improvement was 35 ± 21° (p < 0.001). Arthroscopic release did not affect forearm rotation or strength. One patient developed a permanent medial antebrachial cutaneous neuroma. One patient required a repeat surgery to remove a loose body. There were no instability and no major neurovascular complications. All patients had improved elbow function with a mean ASES score of thirty-one out of thirty-six. Most patients were satisfied with their surgery, had minimal pain, considered themselves in good physical health on the SF-36, and had minimal impairment on the DASH.

Arthroscopic release is safe and effective in experienced hands.

Results are comparable to traditional open techniques. The theoretical advantages of arthroscopy include improved joint visualization, decreased morbidity and earlier rehabilitation.

Disadvantages include the potential for serious neurovascular complications, and the inability to deal with ulnar nerve pathology or heterotopic ossification. Indications for conversion to open release include excessive swelling, and failure to maintain adequate view.

This study was conducted to determine the effect of passive and active muscle loading on humeral head translation during glenohumeral abduction. A shoulder simulator produced unconstrained active glenohumeral abduction using several sets of loading ratios. Significantly greater translations occurred in passive motion as compared to active motion between 30 and 70 degrees of elevation in three dimensions and in the anterosuperior plane. No difference was found between the active motions. Also, translations of the humeral head decreased with active simulation of abduction emphasizing the importance of the rotator cuff muscles in creating and maintaining the ball-and-socket kinematics of the shoulder.

This in-vitro study was conducted to determine the effect of passive and active loading on humeral head translation during glenohumeral abduction.

Five cadaveric shoulders were tested using a shoulder simulator designed to produce unconstrained abduction of the humerus. Forces were applied to simulate loading of the supraspinatus, subscapularis, infraspinatus/teres minor, anterior, middle, and posterior deltoid muscles using four different sets of loading ratios. These were based on:

equal loads to all cables (Constant-Constant);

In three dimensions, significantly greater translations occurred in passive motion as compared to active motion between 30 and 70 degrees of elevation (p< 0.001). No difference was found between the active motions. Similar results were observed in the two-dimensional resultant translations in the anterosuperior plane of the scapula, with more translation occurring during passive motion (3.6 ± 1.1mm) than active (2.1 ± 1.0mm) (p=0.002), and no significant differences between the active loading methods (Figure 1). The majority of translation tended to occur in the superior-inferior direction for all loading ratios employed.

It was clearly shown that the translations of the humeral head decreased with active simulation of abduction. These findings are in agreement with other in-vivo and in-vitro investigations.

This emphasizes the importance of the rotator cuff muscles in creating and maintaining the ball-and-socket kinematics of the shoulder.

The influence of the supinator and pronator quadratus (PQ) muscles on distal radioulnar joint stability were evaluated using a joint simulator capable of producing forearm rotation, before and after ulnar head excision. Multiple pronation trials were conducted with incremental loading of the PQ relative to the pronator teres; supination trials were similarly conducted with the supinator and biceps. Incremental supinator muscle loading did not alter forearm kinematics. Increased PQ loading did not affect intact kinematics, but did alter joint motion following ulnar head excision. PQ activation will likely aggravate forearm instability following ulnar head excision; suggesting rehabilitation should incorporate immobilization in supination.

The purpose of this study was to study the effect of pronator quadratus (PQ) and supinator loads on forearm kinematics in both an intact distal radioulnar joint (DRUJ) and following ulnar head excision.

The PQ muscle appears to aggravate instability of the DRUJ following ulnar head excision, while incremental loading of the supinator muscle had no effect.

Patients with DRUJ instability and/or who have undergone surgical removal of the ulnar head should be rehabilitated in supination to limit the influence of the PQ muscle.

Eight cadaveric upper extremities were tested in a custom joint simulator employing motion and load-controlled tendon loading to produce forearm rotation. Pronation was achieved via loading of the pronator teres and PQ muscles. Repeated trials were conducted in which the ratio of the PQ load was increased incrementally relative to the pronator teres load. Supination trials were similarly conducted using the biceps and supinator muscles. Testing was conducted in the intact forearm and following ulnar head excision. An electromagnetic tracking device was used to record motion of the radius and ulna. Kinematic data was analyzed with a planar analysis that measured dorsal palmar displacements and diastasis of the DRUJ.

Greater diastasis and dorsal translation of the radius relative to the ulna were noted under increased PQ loading following ulnar head excision (p< 0.05). Increased supinator load had no effect on kinematics before or after ulnar head excision. This effect is likely due to the location of the two muscles. The effect of PQ muscle loading was only noted in neutral to full pronation. These results suggest that rehabilitation of the forearm following ulnar head excision should be conducted with the forearm in supination to minimize joint instability.

Funding: Natural Sciences and Engineering Research Council of Canada, The Arthritis Society (Canada)

Single-strand medial collateral elbow ligament (MCL) reconstruction strength was evaluated using double docking (DD) and interference screw (IS) methods with either palmaris longus (PL) or Graft Jacket_ (GJ) as the reconstruction material. Thirteen upper-extremities were mounted in 90° valgus orientations, and subjected to increasing cyclic valgus loading until failure. DD reconstructions outperformed IS reconstructions (P< 0.05), while PL and GJ performed comparably (P> 0.05). The initial Graft Jacket strength makes it a potential alternative to palmaris longus tendons; Laboratory evaluation of graft strength during healing is required. For its simplicity and strength, the DD technique should be considered, clinically.

Single-strand medial collateral elbow ligament (MCL) reconstruction strength was evaluated using double docking (DD) and interference screw (IS) methods with either palmaris longus (PL) or Graft Jacket_ (GJ) as the reconstruction material.

Thirteen, fresh-frozen upper-extremities (66 ±5 years) were cleaned of all soft tissues except the medial and lateral collateral ligaments, flexed to 90° and mounted in a rigid, valgus testing system. DD or IS reconstructions were performed using either PL or GJ. A cyclic (0.5Hz) load was applied 12cm distal to the medial epicondyle. After 500 cycles, the load was increased by 10N until catastrophic failure or a length increase of 10mm.

The mean maximum load for the DD with GJ was 65 ±12N; for the IS with GJ: 45 ±5N; for the DD with PL: 59 ±11N; and for the IS with PL: 56 ±14N. The mean maximum number of cycles endured by the DD with GJ was 1292 ±562; for the IS with GJ: 356 ±292; for the DD with PL: 1104 ±479; and for the IS with PL: 924 ±690. For both the maximum load and number of cycles, the DD outperformed the IS (P< 0.05) and the GJ and PL performed comparably (P> 0.05).

Single-strand reconstructions using the double dock method outperform the interference screw technique. For its simplicity and strength, the DD technique should be considered, clinically. The initial Graft Jacket strength makes it a potential alternative to palmaris longus tendons; laboratory evaluation of graft strength during healing is required.

Funding: This study was partially funded by Wright Medical Technology (Arlington, TN) and the Canadian Institute for Health Research.

Please contact author for graphs and/or diagrams.

The reliability and accuracy of plain radiographs, MRI and CT Arthrography to detect the presence of loose bodies was evaluated in twenty-six patients with mechanical elbow symptoms. The location of loose bodies found by the imaging studies was compared to arthroscopic findings. Overall sensitivity for the detection of loose bodies was 88 – 100% and specificity was 20 – 70%. Plain radiographs had a similar sensitivity and specificity of 84% and 71% respectively. MRI and CT Arthrography were similar to plain radiography, suggesting that routine use of these modalities is not indicated.

The purpose of this study was to determine the clinical utility of MRI and CT Arthrography (CTA) to reliably and accurately predict the presence of loose bodies in the elbow.

Twenty-six patients with mechanical elbow symptoms underwent plain radiography, MRI and CTA, followed by standard elbow arthroscopy. Three musculoskeletal radiologists reviewed the ‘blinded’ plain radiographs with both the MRI and CTA at separate sittings. The location and number of loose bodies on the MRI and CTA were recorded. The preoperative plain radiographs, MRI and CTA were compared to the arthroscopic findings.

Agreement between radiologists was higher for the number of loose bodies identified in the posterior compartment (ICC=0.72 for both MRI and CTA) than in the anterior compartment (ICC=0.41 and 0.52 for MRI and CTA respectively). The correlation between the number of lose bodies observed on MRI and CTA compared to those found arthroscopically was also higher in the posterior compartment (r=0.54–0.85) than in the anterior compartment (r=0.01–0.45). Both MRI and CTA had excellent sensitivity (92–100%) but moderate to low specificity (15–77%) in identifying posteriorly located loose bodies. Neither MRI nor CTA were consistently sensitive (46–91%) or specific (13–73%) in predicting anterior loose bodies. Overall sensitivity for the detection of loose bodies in either compartment was 88–100% and specificity was 20–70%. The preoperative radiographs had a similar sensitivity and specificity of 84% and 71% respectively.

MRI and CTA were similar to plain radiography in the prediction of elbow loose bodies.

Introduction and Aims: Computer-assisted bone and soft tissue balancing in total knee arthroplasty (TKA) may aid in achieving perfect knee alignment leading to better function and prosthesis survival. The ‘Measured Resection’ technique was compared to a ‘Computer Assisted Gap Equalisation’ (CAGE) technique of knee balancing in TKA.

Method: TKAs were performed on eight pairs of cadaver knees. One side of each pair was randomly selected to the control group in which measured resection was used for balancing. The experimental technique (CAGE) using a computer-assisted ligament-tensioning device to equalise gap symmetry and load was used on the contralateral side. Post-operatively, a simulator applied forces to the quadriceps and hamstring tendons and a tibial load transducer measured compartmental force at five flexion angles (0, 30, 45, 60, 90 degrees). Outcome assessment consisted of measuring gap loads and symmetry under ligament tension pre-component insertion and compartmental force post-component insertion.

Results: Although there was no significant difference between the two groups in the symmetry of the extension (p = 0.27) and flexion (p=0.07) gaps pre-component insertion, there was a trend towards improved gap symmetry in the CAGE group. As well, pre-component insertion there was a significant (p< 0.05) equalisation of flexion and extension gap loads in the CAGE group. However, post-component insertion there was no significant difference (p> 0.05 using 2-way repeated measures ANOVA) in medial to lateral compartment load balance between the two groups. As well, the measured loads with the knee in full extension (zero degrees of flexion) were significantly higher (p< 0.001) in both groups compared to other flexion angles.

Conclusion: CAGE improves knee balance pre-component insertion by improving surgical accuracy with computer-assistance. However, component design, posterior capsular tension and tibial rotation preclude sustaining the improved balance post-component insertion leaving final knee balance unchanged. Further work is needed to translate the improved surgical accuracy into improved balance following component insertion.

Introduction and Aims: Suture anchors allow consistent reattachment of tendons and ligaments to bone. Many options are available. The purpose of this study was to compare the initial strength of two rotator cuff repair techniques. The hypothesis was that rotator cuff repair strength with anchors would be inferior to transosseous sutures.

Method: Eight paired shoulders with a standardised supraspinatus defect were randomised to bioabsorbable nonsuture-based anchor or transosseous suture repair. Each specimen was then subjected to a stepwise cyclic loading protocol, utilising a custom-designed loading apparatus. Repair site migration was measured using an optical measurement system, consisting of a digital camera and custom software. Mode of failure, number of cycles and load to failure were measured for 50% (5 mm) and 100% (10 mm) loss of repair. These results were correlated with bone mineral density, age and gender. Statistical analysis utilised paired t-tests and Pearson correlations.

Results: The anchors failed at the anchor-tendon interface, whereas the sutures failed through the sutures. Mean values for 50 percent loss of repair were 206 ± 88 cycles and 44 ± 15 N for the sutures, and 1193 ± 252 cycles and 156 ± 20 N for the anchors (p< 0.05). The corresponding values for 100 percent loss of repair were 2458 ± 379 cycles and 294 ± 27 N for the sutures, and 2292 ± 333 cycles and 263 ± 28 N for the anchors (p> 0.05). These results may be due to the relative less deformability of the anchors. This may be relevant clinically, as in the early post-operative period, while tendon healing to bone is occurring, anchors may offer improved strength, allowing improved initial healing. Strength was unaffected by bone quality. This may be attributed to each repair failing primarily through the repair construct or at the anchor-tendon interface, and not through bone. Strengths of this study include the use of paired specimens, the stepwise cyclic loading protocol, as well as increased accuracy of our measurement system. Limitations include the use of an in vitro model, as well as a simulated, standardised rotator cuff tear.

Conclusion: Repair strength with anchors was superior to sutures. Strength was unaffected by bone quality. Anchors facilitate an arthroscopic procedure, decrease operative time, and may allow a faster post-operative recovery. This study has described a new high-resolution method of measuring tendon repair failure and may be useful in future studies.