There is no clear consensus regarding the indications for surgical treatment of middle third clavicle fractures. An initial shortening of 2 cm or more of the clavicle was associated with poor clinical outcomes and higher rate of non-union. The number needed to treat (NNT) clavicle fractures in order to prevent non-union ranges in the recent literature from 4.5 to 9.2. A direct relationship between shortening of the clavicle and a poor clinical outcome has not yet been demonstrated.

Prospective cohort study performed in a Level one trauma centre including 148 clavicle fractures treated conservatively. Eighty-five patients met the inclusion criteria (healed fracture in the middle third, no other upper limb lesions) and 63 were enrolled. A single assessment was realised at a minimum one year follow-up by an independent examiner and consisted in Constant and DASH scores, range of motion, strength in abduction (Isobex) and a specific radiographic evaluation using a calibrated AP radiographs of both clavicles. Two groups were constituted and analysed according to a radiologic shortening > 2 cm (patients and assessor blinded). Sub-analyses were performed to find any relevant clinical threshold.

The rate of shortening > 2cm in this cohort is 16.1% (10 patients). No clinical differences between the two groups for Constant scores (shortened > 2 cm = 96.0 ± 6.0 vs 95.2 ± 6.6, p=0,73) and DASH scores (8.4 ± 11.9 vs 5.4 ± 8.1, p=0,32). A slight loss in flexion was observed with a shortening > 2cm (175 deg ± 8.5 vs 179.3 ± 3.4, p=0,007). No clinical threshold (in absolute or relative length) was associated with lower functional scores. No relationship between clinical results and patient characteristics. Interestingly, cosmesis was not an issue for patients.

This study could not demonstrate any clinical impact of the shortening of the clavicle in patients treated conservatively for a fracture in the middle third. Functional scores are excellent and the slight difference in flexion is not clinically significant. We were not able to found patients unsatisfied with their treatment. The poor functional outcomes described in previous studies are mainly related to non-unions. Just after the trauma, protraction of the scapula and single AP views centered on the clavicle can overestimate the real shortening. An initial shortening of the clavicle > 2 cm is not a surgical indication for fractures in the middle third; patient selection for surgery should focus on risk factors for non-unions.