Introduction: We investigated the use of suture repair for tuberosities in hemiarthroplasty for 4 part fractures, by examining cadaveric repairs after repetitive loading tests.

Sutures are the most popular repair tool currently. However tuberosity “disappearance” with this is common; and touted as the prime cause for failure. In some studies biochemical lysis has been suggested to explain this. The aim of this study was to analyse the mode of failure.

Materials and Methods: Tests were carried out in line with UK HTA regulations. Eight fresh frozen cadaver shoulders were stripped of all soft tissue except the rotator cuff. A 4 part fracture was then created by osteotomising the tuberosities. A standard hemiarthroplasty implant was cemented in at the native height and retroversion. A tuberosity repair with Ethibond 5 sutures, employed transverse cerclage sutures; and vertical figure of eight repairs, from the cuff to a shaft drill hole.

The repairs were subjected to cyclical tension on the cuff musculature, and simultaneous gleno-humeral motion; using a test process we have described separately. During the entire process the repair site was videoed

At the end of 8000 cycles the tuberosities were probed with a forceps, to record any movement in vertical, horizontal and axial planes. The humeral shaft was also rotated to check for dissociation between tuberosities and shaft.

Results: Uniform failure of the hold on the tuberosities occurred by 8000 cycles. Defining failure as movement of any tuberosity more than 3mm, failure rate was 100% (exact 95% confidence interval 65.2 – 100% due to sample size). Movements of at least 1cm were commonly observed. The sutures were loose but never snapped. Failure mechanisms involved digging of the sutures into the tendon, cutting of the sutures through bone, loss of cancellous supporting structure, and migration of the sutures. Failure occurred early in what we considered to the parallel of 1 to 2 weeks of rehabilitation.

Conclusion: Suture repair of tuberosities has a high chance of failure if the cuff can not be prevented from contracting. We believe failure is early and mechanical, rather than by lysis or biochemical means. More effective alternatives to suture techniques would be desirable.

Introduction: The upper edge of pectoralis major tendon insertion (PMTI) on the humerus is touted as a reference to restore humeral height, to be used in hemi-arthroplasty for proximal humeral fracture. The aim of the study was to verify the reliability of reference and ease of its use in routine practice.

Materials and Methods: Tests were carried out in line with UK HTA regulations.

21 fresh frozen shoulders were stripped of all tissues except the rotator cuff and the PMTI. In each specimen the tuberosities were osteotomised, and the humeral head dislocated. The top margin of the PMTI was sought and marked. A flat object was held on the top of the humeral head perpendicular to the shaft and the vertical height from upper edge of PMTI to this object was measured.

Results: The 21 measurements (12 right +9 left) had a mean height of 5.98cm (95% CI: 5.64 – 6.31cm). The range of heights was 5 to 7.3 cm for all shoulders and also for right and left sides separately. Eighteen of these 21 shoulders were a pair i.e. right and left of 9 cadavers. The average height for 12 right shoulders was 5.91cm and for the left side average height was 6.05cm. The heights for a pair matched within 0.5 cm only in four pairs (8 shoulders) and exactly in only two pairs (4 shoulders) of specimens. The maximum difference in heights for one pair of shoulders from the same cadaver was as large as 1.4 cm. In comparison to previous studies our measurements were significantly greater than 5.6 cm (P = 0.029). Furthermore, there were 4 measurements of at least 7cm (19.0%, 95%CI: 5.4 – 41.9%).

Conclusion: Proximal humeral fracture treatment is a clinical challenge. In these fractures, usually all the ‘bony’ landmarks are lost. In these circumstances an anatomical landmark is needed to navigate from, and the upper edge of PMTI serves as a reference in restoring the height.

We found the average values for height to be larger than previously reported. Additionally the notion that the contra-lateral side can be used to calculate the height as a percentage of the length, accepting an error of +/− 0.5cm; was not corroborated. In 3 cases of the 9 pairs available a difference of 14, 7 and 6mm was seen. Lastly in certain cases a flimsy layer of fascia curved superiorly at the top margin, and we were in doubt as to whether to take this fascial extension as the top margin; in one case this flimsy layer also had muscular tissue attached. Though use of the fixed value had on our testing, a wider range than reported, we feel that PMTI is a useful landmark in these operations.