Reasons for surgical revision after conservatively treated radial head fractures—retrospective study of 70 patients

In 35 cases (50.0%), retrospective grading according to the Mason classification could not be made because radiological images at the time of the trauma were absent and there was incomplete or missing data on the initial classification. In the remaining 35 patients, there were 20 type I, eight type II, five type III, and two type IV fractures.

Approximately half of our patients (55.7%) were immediately immobilized with a plaster after trauma. The duration of the immobilization was 3.4 weeks (range, 1–8 weeks) on average. Five patients (7.1%) were treated with orthopedic devices after trauma (e.g., hinged external fixator, etc.).

In all, 98.6% of patients complained of pain in the affected elbow; only one patient specified no pain.

Radial head fractures tend to be underestimated by clinicians, in particular because of the lack of evidence of displacement on plain radiographs and the good results achieved with conservative treatment. However, these injuries may often be accompanied by a high degree of ligamentous and soft tissue as well as intra-articular cartilaginous damage, which cannot be displayed on radiographs [4]. Therefore, a thorough assessment of radial head fractures is necessary in order to determine the true severity and extent of these lesions and to minimize the risk of delayed surgical therapy. The correlation between an osseous lesion of the radial head and ligamentous injuries is well known. Itamura et al. described medial collateral ligament ruptures in 54%, lateral collateral ligament ruptures in 80%, and bilateral ruptures in 50% of all cases classified as Mason II and III fractures, while Mason IV fractures were excluded [19].

Hausmann et al. found a partial lesion of the interosseous membrane in nine of 14 patients with Mason I fractures using MRI [16]. Kaas et al. supported these findings, detecting accompanying injuries in Mason I–III fractures including lateral collateral ligament ruptures (n = 28) and humeroradial cartilage damage (n = 8) in 35 of 46 cases [22‐24]. Whether concomitant injuries increase the necessity for surgical treatment remains controversial. Kaas et al. stated that most of the additional lesions in patients with radial head fractures are either not symptomatic or not of clinical importance. However, in their follow-up examination at least 12 months after trauma, flexion and extension deficits occurred in 45 and 43% of their patients, respectively. In addition, 13 of 40 patients described crepitus and one patient had locking due to a loose body. The authors claimed that no patient needed delayed surgery since symptoms were mild and without major restrictions [23]. In our case series, most of the patients were initially treated elsewhere and did not undergo MRI as they suffered from mainly Mason I and Mason II fractures (28/35 cases).

Conservative treatment for Mason I fractures is regarded the gold standard and yields good to excellent results with temporary immobilization for 5–7 days followed by early functional treatment [17, 43, 45]. However, Burkhart et al. reported on 16 patients with poor outcome after Mason I fractures due to fracture sequelae such as instability, loose bodies, and posttraumatic arthrosis. The trauma mechanism in radial head fractures is similar to that of elbow dislocations. An enhanced clinical and radiological evaluation (typically MRI) is required for type I Mason radial head fractures in order to detect soft tissue and ligamentous damage and to initiate adequate treatment [6]. MRI frequently reveals evidence of elbow dislocation in Mason type I fractures with severe soft tissue injuries. It remains unknown which of these soft tissue injuries might benefit from surgery. However, a study by Adolfsson et al. supports the assumption that the severity of soft tissue injuries correlates with complications. The authors reported on a cohort of patients who experience redislocation despite receiving proper conservative management for a simple elbow dislocation. During surgery they found complete avulsions of the medial and/or lateral collateral ligaments and muscle origins [1]. For this reason, the use of MRI to visualize the extent of soft tissue injuries is reasonable. This conclusion is supported by our study. In particular, subjective elbow instability, pre-arthrotic deformities, and restriction in elbow mobility frequently occur after Mason I and II fractures with a negative effect on the clinical outcome. Our approach always includes MRI in selected cases of insufficient clinical improvement during early follow-up even in non-dislocated Mason I fractures. Conservative treatment is aimed for, but concomitant injuries might call for surgery.

In the literature, conservative treatment is still regarded the method of choice for Mason II fractures. In a long-term follow-up study, Akesson et al. reported good to excellent results in about 82% of cases after conservative management of Mason II fractures. The rate of degenerative changes reached 82% for the injured and 21% for the uninjured elbow [2]. Surgical treatment led to similar clinical results (82% good to excellent) 22 years after open reduction and internal fixation (ORIF) in a study by Lindenhovius and colleagues. However, arthrosis was only detected in one of 16 cases in their study [28]. To date, there is no prospective study comparing ORIF with conservative treatment in Mason II fractures. Yoon et al. compared nonsurgical treatment with ORIF in partial articular radial head fractures, but the study was compromised by several biases and the conclusion is therefore limited [44]. The RAMBO trial was initiated in 2014 and aimed to address the question of whether Mason II fractures should be treated conservatively or surgically [5]. Unfortunately, no results from the trial have been published yet. A systematic review by Zwingmann et al. favored ORIF with screws in Mason type II fractures over osteosynthesis with pins or K‑wires and over conservative treatment because of better outcomes with the former approach. However, the selection of conservative studies included in the review, dated from 1981 to 1992, represents a possible bias [46].

Owing to the lower rate of degenerative changes, especially in young patients, we recommend surgery for Mason II fractures. However, the operative approach (screws, plate osteosynthesis, pins etc.) has to be assessed individually.

By contrast, in Mason type III and IV fractures, a surgical approach is the gold standard. The method to be followed, however, is still under discussion since ORIF, as the preferred therapy, and implantation of a radial head prosthesis or radial head replacement represent suitable treatment options [3, 20, 25, 26, 37, 40, 43]. All these strategies exhibit advantages and disadvantages and are mainly dependent on the expertise of the surgeon.

Our study involved only patients undergoing surgery for fracture sequelae such as instability, restricted range of motion (stiffness), or painful weight-bearing. Elbow stiffness described in the literature, which is significantly correlated with the duration of immobilization, is a serious complication after radial head fractures [8, 29, 30, 33‐35, 41, 42]. This was also found in the present study. In most cases, a relevant restriction of elbow movement was one of the main reasons for delayed surgery after initial conservative treatment of non- or only slightly displaced radial head fractures. Regarding the soft tissue damage, some patients with Mason I radial head fractures likely had an injury mechanism similar to an elbow dislocation. Furthermore, the radial head is an important stabilizer against valgus stress in combination with the medial collateral ligament, which was confirmed by several biomechanical studies [12‐14, 25, 36].

Soft tissue damage, especially in Mason I fractures, can lead to the classification being adapted after surgical intervention, thereby resulting in a Mason IV fracture.

In contrast to reports in the literature, in which cartilage lesions are described as either asymptomatic or clinically not relevant, our patients suffered remarkably from cartilage damage, which is mostly detected as crepitus during clinical examination. In our series, all complications and all pathologies represented surgical indications due to restricting symptoms. The difficulty of treating radiocapitellar arthritis especially in young patients is a frequent topic of discussion in the literature. Therefore, prevention of radiocapitellar arthritis seems logical. ORIF of Mason II fractures is a simple procedure with a very high rate of good results and low complication rates.

On the basis of our patient population, we cannot recommend primary surgical treatment for Mason I or Mason II fractures since we did not compare outcomes. Instead, appropriate and extensive diagnostics are necessary to detect concomitant injuries, which might influence the decision on whether the patient will benefit from surgery or not. Furthermore, continuous clinical examinations are highly recommended so as to change conservative treatment when required. We tend to recommend operative therapy for Mason II fractures in cases of concomitant lesions, since degenerative lesions mostly prevailed in our patient group with Mason II fractures.

Thus, measurements that possibly promote complications, such as lengthy immobilization or patients with suspicious injuries who refuse MRI, should be avoided. Although not applicable to all patients and all clinics, we recommend acquiring radiographs in two planes and performing MRI independent of the Masson classification of the fracture. The necessity for surgical treatment has to be evaluated individually in every patient with consideration of age, occupation, sports, and handedness [4, 7, 16, 19, 22, 23, 32].

Especially the treatment of intraoperatively detected instabilities by means of stabilization surgery (ligament repair or ligament reconstruction) should be assessed individually. In our study, not every arthroscopically detected instability needed surgical stabilization. Even without stabilization, adequate clinical results could be achieved, similar to the findings of Kaas and coworkers [22‐24].

The lack of a classification system that combines bony and ligamentous lesions as well as the lack of prospective studies comparing surgical with conservative treatment does not allow for a general therapy algorithm. It is important to pay attention to so-called red flags like restriction of movement, unchanging high pain level, or simply unambiguous elbow instabilities in stress testing a few days after trauma.