Participants

The Ethics Committee of our hospital approved this study, and informed consent was obtained from all participants. This was a single-center, retrospective study of patients treated between April 2008 and July 2018. The diagnosis and classification of distal humeral fractures were verified using radiography and three-dimensional computed tomography (Fig. 1). Sixty-five consecutive patients who underwent the SFDB or OO approach for type AO 13C3 distal humeral fractures were included. Between April 2008 and May 2014, the surgeon in our unit performed OO for all patients with type AO 13C3 distal humeral fractures. In 2014, we found that by dislocating the lateral condyle in patients with simple lateral humeral condylar fractures, it was possible to obtain sufficient visualization of the joint to achieve anatomical reduction of the articular surface fracture. Therefore, we started attempting this approach for type AO 13C3 distal humerus fractures.

Fig. 1

The radiographs of the distal humeral fracture. a–d X-Rays and reconstructed 3D CT images of AO type13C3 distal humeral fracture

The inclusion criteria were age > 18 years, 2018 AO 13C3 intercondylar fracture of the distal humerus using AO/ASIF classification, and achievement of anatomical restoration and internal fixation of type AO 13C3 distal humeral fracture through the triceps-sparing participial approach. The exclusion criteria were severe injury (open fracture), nonunion, or delayed fracture of the distal humerus; pathological fracture of the distal humerus; multiple fractures in the same extremity or a pre-existing joint pathology; and less than 12 months follow-up. One patient was lost to follow-up because she died of cerebral hemorrhage 6 months postoperatively. Thus, 64 patients were finally included. In total, 33 patients underwent surgical flip dislocation using a bicolumnar approach without OO; the remaining patients underwent OO.

Surgical techniqueStep 1. Exposure to the fracture

One senior orthopedic surgeon performed all procedures. The patient was placed supine, and the shoulder on the affected side was supported by the injured arm resting on the chest. This position allows the elbow to be placed intraoperatively on the receiving end of a C-arm by abducting the affected limb. Fluoroscopic guidance was used for the surgical procedure. The limb was prepared and draped, and a tourniquet was applied to the proximal upper extremity. A posterior midline incision was made around the top of the elbow in the distal portion; if necessary, the incision was extended proximally (Fig. 2).

Fig. 2

The photograph shows skin incision and dissection. a A posterior midline incision of the elbow. b–c The bilateral borders of the triceps muscle were identified and bluntly split from the intermuscular septum to form medial and lateral windows. The ulnar nerve was identified and protected

Full-thickness fasciocutaneous flaps were elevated over the deep fascia. The ulnar nerve was identified and protected. The bilateral borders of the triceps muscle were determined and bluntly split from the intermuscular septum to form medial and lateral windows (Fig. 2). The two windows were connected through the space between the posterior muscles and distal humerus by blunt dissection; muscle stripping was limited to the distal half of the humerus to protect the radial nerve. The triceps muscle was elevated, and the fat pad was excised from the olecranon fossa, which allowed space for straightforward visualization of the posterior articular surface of the distal humerus. However, the anterior articular surface remained obscure, making anatomical restoration of the anterior articular surface challenging.

Step 2. Exposure and anatomical restoration of the articular surface of the distal humerus

To achieve anatomical restoration of comminuted intercondylar fractures, including anterior articular surface fractures, we used a novel method based on the triceps-sparing participial approach, in which the total articular surface was exposed intraoperatively (Fig. 3).

Fig. 3

a The images of AO type13C3 distal humeru fracture. b–c The lateral window and the medial window of the fracture site. d–g First, the medial and lateral condyles fracture fragments were turned 180° through the two-window space, using the collateral ligament as the axis, respectively. Then, the comminuted condyles fracture fragments were anatomically fixed under direct visualization, using a 1.0 mm K-wire for a small fragment or a 2.7 mm screw for a bigger fragment. Finally, the medial and lateral condyles of the humerus were turned back toward the bottom of the triceps muscle and matched with olecranon, and fixed to the humerus shaft

First, the medial and lateral condyles of the humerus were rotated 180° through the two-window space using the collateral ligament as the axis. Following this process, the posterior insertion of the medial collateral ligament was partially cut, if necessary, to obtain a better view of the anterior articular surface of the medial condyle. The medial ulnar collateral ligament (MUCL) complex of the elbow comprises the anterior bundle (AB), posterior bundle (PB), and transverse ligament. The AB is the strongest component of the ligamentous complex and the primary restraint to valgus stress. The PB of the MUCL is a fan-shaped area of capsular thickening that extends from the medial epicondyle to the semilunar notch of the ulna. Following anterior transposition of the ulnar nerve, posterior dislocation of the medial humeral condyle can be easily achieved by extending a longitudinal resection of Osborn's ligament distally. However, in some cases where the soft tissue around the medial condyle is under more tension and it is difficult to dislocate the condyle posteriorly, the PB of the ulnar collateral ligament is resected partially in the insertion of the posterior bundle of MUCL to achieve posterior dislocation of the medial condyle of the distal humerus. The Osborne ligament and posterior bundle are repaired once the fracture has been reduced and fixed. During the partial resection of the medial collateral ligament, we protected the main static stabilizing structure of the AB of the MUCL and the dynamic stabilizing structure of the ulnar carpal flexor due to the anti-rotation stress. Therefore, the stability of the elbow joint was not impaired. Fractures of the humeral trochlea were simultaneously removed and cleaned. This method exposes the anterior articular surface, with complete exposure of the articular surfaces of the humeroulnar and humeroradial joints (Fig. 4).

Fig. 4

Intraoperative photograph showing the reduction to restore the articular surface without olecranon osteotomy. a–d exposure of fracture  fragments. e–h reduction and fixation of fracture fragments

Additionally, the comminuted fracture fragments were anatomically reduced to either side of the condyle and fixed through direct visualization using a 1.0-mm K-wire for small fragments or a 2.7-mm screw for large fragments. A complicated type AO 13C3 intercondylar fracture was then simplified to a type C2 simple intercondylar fracture. Finally, the medial and lateral condyles of the humerus were turned back to fix the distal humeral fragment.

For some complex and challenging cases, the reduction of medial column fragments was performed first, and the medial column was then temporarily fixed on the olecranon ulna with a 1.5-mm Kirschner wire, using the humeroulnar joint as a reference. The humeroulnar joint, radial head, and absent lateral column formed a ring through which the reduction and fixation procedures could be assisted. Following anatomical restoration, the medial and lateral condyles were temporarily fixed with 2.0-mm Kirschner wires under fluoroscopic guidance. Using this technique, a complex C3 fracture was transformed into a C2 fracture.

Step 3. Fracture reduction and internal fixation

The medial and lateral columns of the distal humerus were reconstructed using the orthogonal plating technique, with one plate on the dorsolateral side of the humerus and the other on the medial column (Fig. 4). Direct visualization and fluorescence were used to confirm the reduction, alignment, and length of the screws. The stability of fracture fixation and elbow range of motion (ROM) was assessed intraoperatively. A surgical drain tube was placed beneath the triceps muscle and wound during surgery. The fascia profunda, subcutaneous fat, and skin were sutured layer by layer. Anterior transposition of the ulnar nerve was performed in all patients.

Postoperative protocol

To achieve early functional restoration, passive and active movement of the elbow joint was performed as soon as possible. Per the patient's general condition, dynamic exercises of the shoulder, wrist, and fingers were started on the day of surgery, and functional movements of the elbow joint were started on the day after surgery. Indomethacin was administered orally (25 mg, three times per day) for 6 weeks from the day after surgery. After removing the drainage tube, fully active and assisted elbow motions were initiated. Implant removal could be performed 1–2 years after the remodeling. Routine metal removal was not recommended because although younger patients may require implant removal, older patients may tolerate it.

Assessments

All patients underwent routine evaluations, including clinical and radiographic assessments and ROM evaluations, at 1, 3, 6, and 12 months postoperatively and once yearly thereafter (Figs. 5 and 6).

Fig. 5

The postoperative radiographs were shown during the follow-up. a–d Postoperative X-Rays and postoperative reconstructed 3D CT images of AO type13C3 distal humerus fracture

Fig. 6

No significant limitation in flexion, extension, pronation, and supination of the left elbow at 1-year follow-up. a Flexion, b extension, c pronation d supination function of the injured limb. e–f the showing of the wound incision

The fracture healing time and incidence of complications, including delayed union, nonunion, malunion, and heterotopic ossification, were recorded. The range of daily elbow motion was evaluated using Morrey's criteria [15], which includes 30° extension to 130° flexion, > 50° pronation, and > 50° supination at the final follow-up. Elbow pain, ROM, stability, and function were comprehensively assessed using the Mayo Elbow Performance Index (MEPI). The maximum MEPI score totals 100 points, with a 90–100 score indicating an excellent result; 75–89, a good result; 60–74, a fair result; and < 60, a poor result [16]. Triceps muscle strength was assessed subjectively, with comparison to the uninjured arm, and muscle strength was graded as usual, good, or fair [17, 18]. Disabilities of the Arm, Shoulder, and Hand (DASH) questionnaire scores were obtained during follow-up [19].

Statistical analysis

Propensity score matching was used to minimize bias due to possible confounding factors. The SFDB and OO groups were propensity score-matched in a 1:1 ratio, based on age, sex, body mass index, and American Society of Anesthesiology (ASA) score, using the propensity scores generated by logistic regression. After matching, all clinical and functional outcomes were compared between the groups using the paired t test. Before starting this trial, we performed a power analysis to estimate the sample size. The sample size in this study was sufficient to detect a significant difference between the cohorts at an alpha value of 0.05 and a power of 80%. Statistical analysis was performed using IBM SPSS Statistics (version 22, New York, NY, USA). Statistical significance was set at P < 0.05.