Proximal third humeral shaft fractures fixed with long helical PHILOS plates in elderly patients: benefit of pre-contouring plates on a 3D-printed model—a retrospective study

We retrospectively identified patients, treated with long helical PHILOS plates between February 2012 and February 2015, from our trauma center’s database. Inclusion criteria were (I) aged 65 years or above, (II) presence of proximal third humeral shaft fractures (unilateral), (III) treated by the MIPO technique with a long helical PHILOS plate, and (IV) more than 12 months follow-up after surgery with complete follow-up data. Exclusion criteria included pathological fractures, open fractures, multiple fractures, or the presence of other diseases affecting the same upper limb.

Ultimately, 46 patients were included, and data regarding their medical histories, functional evaluations, and regular radiographic examinations were collected. All fractures were classified according to the AO/ASIF classification system [15]. Medical history information included gender, age, and surgical information (plate pre-contoured on a Synbone or 3D-printed model, the name of the surgeon, duration of surgery, and blood loss volume during surgery).

A bilateral humerus CT scan (1-mm layer interval) was taken from patients in the 3D-printed group, and CT data was extracted for constructing the 3D-printed model, using medical 3D processing software (Mimics 16.0, Materialise, Belgium). The scapula, clavicle, and relative soft tissue were separated. A mirror image of the intact side was used to simulate the other side, and the models were printed using a 3D-printing machine (Lite, RS6000, UnionTech, Shanghai, China) with ultraviolet rays (UV) curable resin material. Total preparation time of a 3D bone model was approximately 3 h as long as the CT data is available.

In the 3D-printed group, a long PHILOS plate was pre-contoured on an intact 3D-printed model, to ensure that it could be used on the fractured bone. In the Synbone group, the plate was pre-contoured on a standard Synbone model. The plate was twisted at approximately 60–90° (starting at the superior part of the humeral deltoid tuberosity), while the proximal part of the plate was located on the lateral side of the greater tubercle and the distal part on the anterior side of the distal humerus. Then, the contoured helical plate was sterilized on the day before the surgery. The 3D-printed models and the helical plate used in an elderly Asian female patient are shown in Fig. 1, and it is obvious that the true humerus is much shorter than the Synbone.

The basic surgical steps were similar for the two groups and were as described in our previous study [11]. After general anesthesia or brachial plexus block anesthesia, the patient was placed in the beach-chair position, with the upper limb in full supination, and an anterolateral acromial approach (ALA) was performed with a 5-cm skin incision proximally, while the anterior approach was performed with a 5-cm skin incision distally by splitting the brachialis longitudinally just along the lateral side of the biceps brachii. The site of the distal approach was decided according to the length of the pre-contoured plate. During the distal anterior approach, care was taken not to injure the musculocutaneous nerve, which lies between the biceps and the brachialis; the dissection through the brachialis was performed bluntly but gently, without routine nerve exposure. An extraperiosteal tunnel was made to connect with both approaches, from the lateral part of the greater tubercle proximally to the anterior part of the humerus distally. The key point was that, in order to protect the deltoid attachment point, the tunnel did not pass through the humeral deltoid tuberosity but across it anteriorly. Using the MIPO technique, the pre-contoured helical long PHILOS plate was then inserted from the proximal approach, passed through the tunnel distally, and fixed on the humerus with locking screws through the ALA approach proximally (five or more screws inserted). According to AO principles, at least two or three screw holes should be left open over the fracture to decrease stress concentration [15]. So, usually, three to four screws were inserted distally according to the different fracture types (Fig. 2). Finally, intraoperative fluoroscopy was used to ensure a correct fracture reduction and positioning of the plate. Some adjustments were required if the fluoroscopy results were unsatisfactory. The name of the surgeon, duration of surgery (from incision to skin closure), and blood loss volume were recorded.

After surgery, all patients were recommended the same rehabilitation plan. They were allowed to perform passive range of motion exercises immediately, and active exercises were allowed after 2 weeks. Patients were recommended to attend follow-up visits at 4 weeks, 12 weeks, 6 months, and 12 months after surgery. At each follow-up visit, a regular X-ray examination was taken, and at the 12-month follow-up, functional evaluation was added.

Categorical variables in each group are presented as count (percentage), and continuous variables are presented as mean ± standard deviation. All categorical variables were compared directly to each other using the chi-square test or Fisher’s exact tests. For continuous variables, Student’s two-tailed t tests were conducted. Results were considered significant when P < 0.05. The analysis was performed using the statistical program SPSS version 21 (IBM Corp., Armonk, NY, USA).