Long-term results after non-operative and operative treatment of radial neck fractures in adults

Isolated fractures of the radial neck are a rare injury of the elbow in both adults and children. In children, they account for about 5–10% of all elbow injuries [1, 2]. The radial neck fractures in children are classified according to Judet as modified by Maitezeau in types 1–4, according to the initial grade of displacement [3]. In dislocated fractures (initial angulation > 30°, according to type 3 and above), operative treatment is recommended due to the increasing risk of avascular necrosis [1, 2, 4‐6].

In adults, fractures of the radial neck are even rarer still [7]. At present, very little data is available regarding the clinical and radiological outcome of these fractures in adults, especially when treated surgically [8, 9]. Radial neck fractures in adults can be classified according to Broberg-Morrey [10], which is a modification of the Mason classification [11]. A majority of isolated radial neck fractures, especially types 1 and 2 according to the Broberg-Morrey classification, can be treated by functional exercises alone, achieving excellent or good results in most patients [8, 12‐14]. In comminuted and dislocated fractures of the radial neck, operative treatment with open reduction and internal fixation with screws or plate osteosynthesis is common [15‐17]. In the available literature, results of radial neck fractures in adults are often evaluated and discussed together with radial head fractures [8, 9, 15‐18]. Therefore, specific recommendations for treatment decision, especially in isolated radial neck fractures, are currently not available.

In the present study, the functional and radiological outcome of non-operatively and operatively treated fractures of the radial neck in adults has been investigated in detail. The primary aim was to evaluate the clinical and radiological long-term outcome of these fractures. The secondary aim was to evaluate the risk of complications and revision surgery following the initial treatment decision.

Demographic data of the study population with variation of the treatment groups are shown in Table 1. Twenty patients (58.8%) were treated non-operatively; 14 patients (41.2%) received primary operative treatment. Four patients of the operative group (28.6%) were treated by crossed screw osteosynthesis, 10 patients (71.4%) by plate osteosynthesis. The mean follow-up was 5.7 (2.0 to 15.7) years (non-operative group 5.6 [2.0 to 15.7], operative group 5.7 [2.1 to 12.7] years). The distribution of the fractures according to the Broberg-Morrey classification showed that patients in the non-operative group had less severe injury types (Table 2). The initial RHSA was larger in the operative group with a significant difference in the anterior-posterior plane (p = 0.015 (Fig. 3, Table 3). At postoperative radiographs, the RHSA was significantly lower in the operative group (p = 0.043). Due to the significant difference of the RHSA in the initial radiographs, statistical tests for the comparison of the clinical results of the two groups were not performed. Regression analysis of the initial RHSA regarding the DASH score and MEPS outcome was performed, regarding a correlation between angulation and functional results. There was no strong correlation between RHSA and DASH (R = 0.253) or MEPS (R = 0.205).

At follow-up, the overall clinical outcome was good with a mean DASH score of 13.1 (0 to 71.6) points and a mean MEPS of 81.0 (30 to 95) points. The detailed analysis of the clinical scores showed good results in both groups (Table 4). The quality of life assessed by the SF-36 was comparable to the average values of the normative groups.

Five patients (25%) developed a total of 7 complications in the non-operative group, and 7 patients (50%) developed a total of 11 complications in the operative group (Table 5).

One patient (5%) received radial head resection in the non-operative group while 7 patients (50%) in the operative group underwent revision surgery (Table 6).

Two patients (10%) in the non-operative group did not reach ability to work, while all patients in the operative group were able to return to work. The mean time between the accident and return to work was 102.19 (31 to 319) days in the non-operative and 146.33 (28 to 459) days in the operative group.

Seventeen out of 20 patients (85%) in the non-operative group were doing sporting activities before the accident with a mean duration of 4.3 (1 to 13) hours per week. Seven out of 14 patients (50%) in the operative group were doing sporting activities with a mean duration of 8.4 (2 to 24) hours per week. At follow-up, 16 patients in the non-operative group (80%) were doing sporting activities with a mean duration of 4.1 (1 to 13) hours per week, while 6 patients in the operative group (40%) were doing sporting activities with a mean duration of 3.0 (0 to 8) hours per week.

Regardless of the treatment modality, it was possible to a show an overall good to very good functional long-term outcome of radial neck fractures regarding the clinical parameters DASH and MEPS. Range of motion and pain level also showed good results in both the operative and non-operative group at final follow-up.

A relevant number of the patients in the operative group suffered from complications and needed revision surgery. This can be assumed to be a consequence of common risks of surgery but also of this specific operative technique. As known from displaced radial head fractures that are operatively treated by plate osteosynthesis, it is crucial to position the plate in the safe zone (meaning the non-articulating portion of the radial head) to avoid plate impingement [20, 23, 24].

None of the 4 patients that were treated by crossed screw osteosynthesis needed revision surgery, while 6 of the 10 patients who received plate osteosynthesis needed revision surgery. This might be due to a higher complexity of these fractures with a multi-fragmentary situation. These results are in agreement of Li et al. who compared crossed screw osteosynthesis with plate osteosynthesis in radial neck fractures and found better results for the screw osteosynthesis [25].

The available literature often analyses the outcome of radial neck fractures together with fractures of the radial head. Duckworth et al. published a larger series of 237 patients who were suffering from a radial head or radial neck fracture and analysed the clinical outcome according to the MEPS after a mean follow-up of 6 months [8]. They recorded an excellent mean MEPS of 92 points, which is better than the results of the current study. They also reported of only two complications, which might be due to the rather short follow-up time. The revision surgery in our patients was performed after an average of 9.6 months. Most patients in this study were treated non-operatively.

Kang et al. published a follow-up of six patients that suffered from non-union after surgically treated displaced radial head and neck fractures [18]. All of them went well under non-operative treatment with a mean MEPS of 96.7 points at follow-up of 7.6 years. This is in contrast with the results of the present study, as two out of three patients with pseudoarthrosis needed revision surgery due to painful limitation of motion. Both patients were treated by excision of the radial head. The mean MEPS of these 3 patients was 78.3 points.

Herbertsson et al. published a long-term follow-up study (mean follow-up 19 years) including radial head and neck fractures Mason types 2 and 3 [12]. A minority of 2 of 100 patients received open reduction and osteosynthesis, while 19 patients were treated with radial head excision alone. The functional results were good with a tendency of impaired flexion of the injured side. In 2004, Herbertsson et al. published the results following radial head excision after radial head and neck fractures [26]. The functional outcome was overall fair to good with worse results of the initial Mason type 4 fractures. It should be noted, however, that a separate analysis of the radial neck fractures was not performed in these studies.

In summary, overall good results can be expected after non-operative and operative treatment of radial neck fractures with higher rates of revision surgery in the operative group. Nonetheless, the operatively treated patients with more severe injury types according to the Broberg-Morrey classification reached the same good clinical outcome parameters as the non-operatively treated patients at final follow-up. We therefore conclude that open reduction and osteosynthesis seem favourable for patients with comminuted and dislocated fractures.

There are several limitations of this study. First of all, the study is limited by its retrospective nature and its small sample size. Due to the rare incidence of this specific fracture, the observational groups are rather small. Secondly, the degree of angulation of the radial neck that led to the initial decision of operative or conservative treatment was significantly different in the two groups. This made a statistical testing between both groups impossible. The decision whether to treat non-operatively or operatively was made on an individual basis rather than by randomisation. Third, the radiological assessment of the initial angulation depends on the exact radiological technique. Due to the acute painful situation, some of the initial radiographs are not exactly anterior-posterior or lateral views. Furthermore, an interobserver and intraobserver analysis of the measurement method of the determination of the RHSA was not performed.

Radial neck fractures are a rare entity in adults. Depending on the initial degree of angulation, a non-operative treatment can achieve good results in clinical scores. If open reduction and osteosynthesis is needed, crossed screw osteosynthesis seems to have a lower risk of surgery-associated complications. However, from the data of this study, it is not possible to determine a RHSA that necessitates open reduction. Further prospective studies might show a certain degree of RSHA that necessitates open reduction and osteosynthesis.