We describe the results after open reduction and internal fixation of 22 consecutive displaced fractures of the glenoid with a mean follow-up of ten years. A posterior approach was used in 16 patients and an anterior in six, the approach being chosen according to the Ideberg classification of the fractures. The fixation failed in two patients, one of whom required a further operation. There were two cases of deep infection.

At follow-up the median Constant score was 94% (mean 79%, range 17 to 100). The score was less than 50% in four patients, including the two who became infected. A further two had an associated complete palsy of the brachial plexus.

Intramedullary nailing of metaphyseal fractures may be associated with deformity as a result of instability after fixation. Our aim was to evaluate the clinical use of Poller screws (blocking screws) as a supplement to stability after fixation with statically locked intramedullary nails of small diameter.

We studied, prospectively, 21 tibial fractures, 10 in the proximal third and 11 in the distal third in 20 patients after the insertion of Poller screws over a mean period of 18.5 months (12 to 29).

All fractures had united. Healing was evident radiologically at a mean of 5.4 ± 2.1 months (3 to 12) with a mean varus-valgus alignment of −1.0° (−5 to 3) and mean antecurvatum-recurvatum alignment of 1.6° (−6 to 11). The mean loss of reduction between placement of the initial Poller screw and follow-up was 0.5° in the frontal plane and 0.4° in the sagittal plane. There were no complications related to the Poller screw.

The clinical outcome, according to the Karström-Olerud score, was not influenced by previous or concomitant injuries in 18 patients and was judged as excellent in three (17%), good in seven (39%), satisfactory in six (33%), fair in one (6%), and poor in one (6%).

The accuracy of templates used for the preoperative planning of the fixation of intramedullary fractures depends on radiological magnification. To study the accuracy of these templates, we randomly selected 100 femoral and 100 tibial radiographs taken after stabilisation by an intramedullary nail using a standard technique. We then compared the known nail length with the corresponding measurements on the radiographs.

The mean magnification factor for the femur was 9% and for the tibia 7%; these differ considerably from the range of magnification of the manufacturers’ templates (femur, 15% to 17%; tibia 10% to 15%). We conclude that templates are unreliable for the selection of implant length and that this should be done by intraoperative measurements.