Minimally invasive plate osteosynthesis with a locking compression plate is superior to open reduction and internal fixation in the management of the proximal humerus fractures

This was a retrospective case control study including two groups. Between September 2007 and April 2012, 184 patients with displaced proximal humerus fractures were treated with LCP (Locking Compression Plate) according to Neer criteria in Union hospital, Tongji Medical College, Huazhong University of Science and Technologe. Of these, 118were treated with ORIF via a traditional deltopectoral approach between September 2007 and June 2010, while 66 were treated with MIPO via anterolateral deltoid splitting between June 2010 and April 2012. Patients with pathological fractures, head split fractures, open fractures, fractures with primary neurovascular damage and cases lost to follow-up were excluded from the study. After these exclusions, 158 patients remained in the study. Of these 158 patients, 86 (43 pairs) were selected for a retrospective matched-paired analysis according to age (±3 years), gender, and fracture type with a minimum follow-up of 12 months (range, 12–17months). Thus, inclusion bias could be excluded.

The median age of the MIPO group was 63 years compared to 61 years in the ORIF group. The distribution of age and sex by group was 27 (62%) women and 16 (38%) men in the MIPO group, and 31 (72%) women and 12 (28%) men in the ORIF group. According to the AO/ASIF classification system, the most frequent type of fracture was type B (n = 46; 53.5%), with 19(22.1%) type A and 21 (24.4%) type C fractures. There were 10 type A, 24 type B and 9 type C fractures in the MIPO group, and there were 9 type A, 22 type B and 12 type C fractures in the ORIF group (Table 1).

The LCP plate (titanium; thickness: 4.2 mm; width: 12 mm; length: 105-231 mm; Double Engine Medical Material Company, China) was anatomically pre-contoured with threeto ten holes on the plate shaft and nine holes for head screws. The proximal suture holes were applied to secure the tuberosity fragments and the plate.

In both the MIPO and ORIF groups, the patients were positioned in the beach-chair or supine position to allow two plane intraoperative C-arm image intensifier views[8, 11, 12, 16, 18, 19, 30‐32]. All of the procedures were performed under general anesthesia with administration of a broad-spectrum antibiotic prophylaxis[16, 18, 22, 30]. In the ORIF group, ORIF was performed using a standard deltopectoral approach with a LCP. The 12-14 cm incision started at the tip of the coracoid process and ran laterally in the direction of the insertion of the deltoid muscle (Figure 1F). Reduction was enabled with a K-wire under fluoroscopy according to the landmarks of the long head of biceps, the greater and lesser tubercles, and the intertubercular groove[26, 30]. After the fracture was anatomically reduced, a LCP was placed 1 cm posterior to the intertubercular groove[11, 30] and 1 cm distal to the tip of the greater tubercle[15, 27].

In the MIPO group, the anterolateral deltoid splitting approach was utilized, and the tip of the acromion was palpated and used as a landmark[14, 19]. A line perpendicular to the palpable shaft of the humerus was drawn 5 cm distal to the tip of the acromion. A second parallel line was drawn 2 cm further distal to the first line (Figure 1B). The area between these two lines contained the axillary nerve and was considered the unsafe zone. A longitudinal incision was made on the lateral side of the humerus starting from the lateral acromial border and ending distally 5 cm was to access the proximal humerus, the greater tuberosity and the humeral head (Figure 1E). To insert distal screws, a distal incision was made 7 cm distal to the acromion and was approximately 2.5-3.5 cm long (Figure 1E). In most cases, good reduction was achieved by applying axial traction on the humerus and pulling the rotator cuff[6, 8, 22, 23, 33]. In some cases, indirect reduction techniques, such as ligamentotaxis, were used. Plate reduction was performed in cases of a valgus displaced fracture configuration[22, 28]. The LCP was placed proximally below the apex of the greater tuberosity to maintain reduction. If non-absorbable sutures were used, they were secured to the suture wires holes in the LCP[6, 11, 16, 19, 21‐23, 26, 28, 33]. The plate was anchored proximally with multiple angled stable screws into the humeral head fragment. After removing the aiming arm, the non-absorbable sutures were tightened to the LCP[21‐23].

Patients in both groups had individual patient-related postoperative management. In the majority of cases, the patients’ arm was placed in a sling for a maximum of two weeks. Passive and active range of motion exercises were started after surgery, depending on pain and activity level[6, 11].

This study has been performed in compliance with the Helsinki Declaration and has been granted an exemption from Hospital’s Ethics Committee of Union Hospital of Tongji Medical College. The patients were informed and have written informed consent; all patients were over 18 years old. The classic cases in this article have been undertaken with the patient's consent.

We used SPSS 18 statistical software for Windows for all analyses. For normally distributed data (patient age, time between fracture and fixation, operative time, blood loss, hospital stay, and follow-up time), an independent sample t-test was used. For data that was not normally distributed data (AO classification and the Constant shoulder score), the Mann–Whitney rank sum test was used. For categorical data (gender and fracture pattern), a chi-square test was used. A P-value < 0.05 was considered statistically significant.

We performed clinical and radiographic assessments 3, 6 and 12 months after surgery[6, 11]. At each follow-up, the Constant score was used to assess shoulder function. Standardized X-rays were obtained in anteroposterior and transscapular views to evaluate fracture healing, avascular necrosis, placement of the plate, and quality of reduction. Complications were evaluated based on follow-up radiographs and a retrospective chart review of the patients’ medical records to determine the incidence of humeral head necrosis, delayed union, implant failure or a neurological deficit.

To perform the retrospective matched-paired analysis, we selected patients according to age (±3 years), gender and fracture types. The patient demographics of 86 patients are listed in Table 1. The analysis revealed no significant differences in group demographics, including AO/ASIF classification and mean age (p > 0.05). However, there was a higher female ratio in the ORIF group compared to the MIPO group (p > 0.05). The average duration between trauma and surgery in the MIPO group was 5 days and 6.3 days in the ORIF group. In addition, MIPO required less surgery time and resulted in less blood loss (both p < 0.01). The union rate at 6 months was 93% in the MIPO group and 97.7% in the ORIF group (Table 2).

The Constant score was higher in the MIPO group at the 3 and 6 month follow-up time points compared to the ORIF group (P =0.033 and P = 0.043) (Figure 2). At the 12 month follow-up, the Constant score was not statistically significant (P = 0.065) (Figure 2). In addition, patients in the MIPO group experienced significantly less pain, higher satisfaction in activities of daily living, and greater range of motion at the 3 and 6 months follow-up time points (p < 0.05). The level of strength was not significantly different at these time points (p > 0.05). At the 12 month follow-up, there were no significant differences in both groups (Table 3).At the 12 month follow-up, type A fractures had the highest average Constant score in both groups (MIPO: 76.2 ± 7.1, ORIF: 75.6 ± 10.8), followed by type B fractures (MIPO: 71.7 ± 11.3, ORIF 69.0 ± 14.2; ORIF: 69. ±13.9) and type C fractures (MIPO: 69.4 ± 17.1, ORIF: 69.8 ± 13.2) (Figure 3). The type of fracture between groups was not significantly different (p = 0.205).

In the MIPO group, 5/43 patients experienced complications and 4/43 patients experienced complications in the ORIF group (Table 4). No patients developed wound infection and nonunion after one year of follow-up in both groups. In 3 patients (group MIPO, 2 patients, type B and C; group ORIF, 1 patient, type B), the fracture collapsed after 3 months, leading to a varus malalignment. These patients developed loss of reduction and underwent reoperation either by reosteosynthesis combined with cancellous grafting or by joint replacement. One patient in the MIPO group presented with clinical signs of axillary injury, which was characterized by poorly localized posterior shoulder pain, parenthesis over the lateral aspect of the shoulder, and deltoid muscle weakness. Axillary nerve injury was confirmed on electromyography examination. However, there was no functional impairment when the patient was assessed at one year follow-up. One plate in the MIPO group and two plates in the ORIF group were removed due to subacromial impingement after radiographs confirmed fracture union at about 5 months. In the ORIF group, one patient underwent reoperation to change a perforated screw 3 months after the initial operation.