Antegrade nailing of proximal humeral fractures using a straight nail can damage the bony insertion of the supraspinatus tendon and may lead to varus failure of the construct. In order to establish the ideal anatomical landmarks for insertion of the nail and their clinical relevance we analysed CT scans of bilateral proximal humeri in 200 patients (mean age 45.1 years (sd 19.6; 18 to 97) without humeral fractures. The entry point of the nail was defined by the point of intersection of the anteroposterior and lateral vertical axes with the cortex of the humeral head. The critical point was defined as the intersection of the sagittal axis with the medial limit of the insertion of the supraspinatus tendon on the greater tuberosity. The region of interest, i.e. the biggest entry hole that would not encroach on the insertion of the supraspinatus tendon, was calculated setting a 3 mm minimal distance from the critical point. This identified that 38.5% of the humeral heads were categorised as ‘critical types’, due to morphology in which the predicted offset of the entry point would encroach on the insertion of the supraspinatus tendon that may damage the tendon and reduce the stability of fixation.

We therefore emphasise the need for ‘fastidious’ pre-operative planning to minimise this risk.

Cite this article: Bone Joint J 2014;96-B:249–53.

Untreated 3- and 4-part fractures of the proximal humerus have a poor functional outcome. Open operation increases the risk of avascular necrosis and percutaneous reduction and fixation may be preferable.

We report 27 patients, 9 with 3-part and 18 with 4-part fractures, treated by percutaneous reduction and screw fixation. Thirteen of the 4-part fractures were of the valgus type with no significant lateral displacement of the articular segment, and five showed significant shift. Instruments were introduced into the fracture through small incisions so that the fragments could be manoeuvred under the control of an image intensifier, taking advantage of ligamentotaxis as far as possible. A good reduction was achieved in most cases.

The average follow-up was 24 months (18 to 47). All the 3-part fractures showed good to very good functional results, with an average Constant score of 91% (84% to 100%), and no signs of avascular necrosis. Good radiological results were achieved in 4-part fractures when impacted in valgus except for one patient with partial avascular necrosis of the head. In those with lateral displacement of the head, revision to a prosthesis was required in one patient because of avascular necrosis and in another because of secondary redisplacement of the fracture. Avascular necrosis was seen in 11% of 4-part fractures. The average Constant score in patients with 4-part fractures who did not need further operation was 87% (75% to 100%).