Non-operative treatment of minimally displaced proximal humeral fractures provided good fracture healing and satisfactory functional results [3, 4]. However, severely displaced fractures have occasionally had poor clinical results due to nonunion or osteonecrosis [2, 10]. Two-part surgical neck fractures account for 28% of all proximal humeral fractures . Non-operative treatment of these fractures has often led to varus malunion [3, 5]. Several surgical techniques have been described in the literature, including percutaneous pinning, external fixation, intramedullary nailing, and plate fixation [5, 11, 12]. In elderly patients, poor fixation due to osteoporosis has been reported and surgical treatment did not lead to improvement . Recently, the locking intramedullary nail and locking plate were used to overcome osteoporotic fractures in the proximal humerus [14–16]. Some studies have compared the locking intramedullary nail and the locking plate for the treatment of 2-part proximal fractures [17, 18]. Zhu et al. reported that the Constant-Murley score (including pain and shoulder motion) was not significantly different between the two techniques but duration of surgery, blood loss, and complications were significantly greater in plate fixation . Gradle et al. compared the intramedullary locking nail and the locking plate in a prospective multicentre study . They found no significant differences between the two techniques in Constant-Murley score or in complications. However, secondary fracture displacement and malunion were seen exclusively in patients after plate fixation. In the present study, we used the locking intramedullary nail for the treatment of severely displaced 2-part surgical neck fractures in elderly patients. Radiographic fracture healing was obtained in all patients. About 90% of these patients gained satisfactory functional results. Complications included osteonecrosis in 1 patient. We found that treatment using the locking intramedullary nail for elderly patients gives good results with a low rate of complications.
Our secondary purpose was to assess the factors influencing outcomes in elderly patients. In our study, varus angulation of the humeral head was associated with poor JOA shoulder score and showed restriction of shoulder movement. It is reported that there is a good correlation between the degree of varus malunion and the outcome . Functional score was lower in the group with 25 to 34° varus angulation than in the group with 15 to 24° varus angulation . In a biomechanical study, 20° varus angulation caused greater forces to accomplish early shoulder abduction, and 45° varus angulation significantly decreased the abduction efficiency of the supraspinatus tendon . Our findings in this study support that varus angulation is associated with significantly poor shoulder function especially in ROM of shoulder joint. Three patients who had more than 30° of varus deformity (HSA < 110°) showed poor shoulder function.
We observed osteonecrosis in one patient after the surgical treatment. The patient had severe pain and severe restriction of shoulder movement. Postoperatively, re-displacement of the fracture site occurred, and bone support was lost on the medial side. The arterial blood supply of the proximal humerus consists of the anterior and posterior circumflex humeral arteries. We suppose that the distal fragment of the humerus disrupted the circumflex arterial blood supply causing the osteonecrosis. An analysis of other factors that affect outcome, the elderly group (75 years or Older) had significantly a lower activity of daily living. The elderly group had two patients with dementia and their clinical outcomes were poor. While follow-up radiographs of these patients showed fracture union and no deformity, both patients had severe restriction of shoulder movement. Their poor functional outcomes may be due to inadequate rehabilitation because they had been unable to follow the active range of motion exercise program.
There are limitations to this study. First, data were collected retrospectively. Selection bias may be a problem, as it is likely that the patients with symptoms were examined in the hospital, but the patients without symptoms were not followed up for a long time. Proximal humeral fracture in elderly patients is common, and a prospective study design would be appropriate to evaluate the factors influencing the outcomes. Secondly, the study population was small. Thirdly, rotation deformities of the shoulder were not assessed. In the future, another large-scale and 3-dimensional study should be performed to resolve these limitations.