Complex fractures of the proximal humerus are usually comminuted because of poor bone quality or high-energy trauma. These kind of fractures treated with open reduction and internal fixation (ORIF) are at the high risk of malunion, screw cutout, osteonecrosis and revision surgery. The integrity of the medial hinge, and the length of the dorsomedial metaphyseal extension have been described as key prognostic factors for humeral head vascularity which may lead to the necrosis of humeral head (13, 14).The most important indication of arthroplasty is fracture type, including head-splitting patterns, fracture dislocations, and displaced three- and four-part fractures not amenable to surgical fixation (12, 15). However, most surgeons tend to choose hemiarthroplasty in the patients ≤65 years, and tend to choose reverse total shoulder arthroplasty in old patients with lower-demand or teared rotator cuff (16–20).Arthroplasty for proximal humeral fracture is performed with the aim of restoring normal biomechanics, achieving adequate range of movement and providing a painless functional joint to the patients. The best advantage of arthroplasty is a reliable procedure for pain relief. Hemiarthroplasty can provide satisfactory functional results, but depends on correct height of the implant and tuberosity union (6).
Clinical studies have shown that anatomical recreation of proximal humerus fractures can affect the clinical consequences of hemiarthroplasty(7). It is a technically demanding procedure that requires proper humeral height, correct retroversion of the stem, and anatomical reconstruction of the tuberosities. The lengthening humeral height more than 10 mm and shortening more than 15 mm is unacceptable in the clinical practice(6, 7). To date, pre-operative templating with comparison of both arms remains the only objective evaluation to assess for the correct parameters of the arm at the time of arthroplasty. The technique is based on exact measurement according to the medial calcar of the fracture prior to the operation(10). In clinical practice there are some obstacles, repeated X-ray exposure before operation, and complicated pre-operation design. Joel et al. introduced the distance at 5.5 cm between the pectoral major tendon insertion and top of the humeral head as a landmark for proper placement of hemiarthroplasty in fractures of the proximal humerus(21). And the method is proved to be a useful and reliable reference for optimal humeral height during hemiarthroplasty for proximal humerus fractures in Asian populations(9). However, the technique is not individual. The distance between the pectoral major tendon insertion and top of the humeral head is different in population due to the height and gender of patients. To prepare the glenoid and exposure the fracture ends, the upper one-third of the pectoralis tendon can be tenotomized routinely. The specific height of prosthesis still needs to be adjusted repeatedly according to the intraoperative joint tension. Sumant et al. (7)restored the proper height by recovery the Gothic Arch (The joining of outline the medial calcar of the proximal humeral and the lateral border of the scapula creates a classical ‘‘vaulted’’ arch or Gothic arch shape). The incorrect prosthetic height, broken medial calcar, or too large head may lead to the failure construction of Gothic arc in the standard lateral position. However, to achieve the arch, the humerus must be rotate to the most lateral offset position which may need repeated X-ray exposure during surgery. Another previous study described a method. If the height was proper, the distance between the lateral margin of the artificial head and medial calcar of the fracture was about 2–3 mm larger than the diameter of the head (10). The technique is based on the integrality of medial calcar.
In our protocol,the height and retroversion of prosthesis are determined through the hole in metaphysis after the anatomical reduction of tuberosities and head of humerus. There is no need additional X-ray exposure pre-operation and intra-operation. The mean operation time and the blood loss have no significant difference between two methods which means that the protocol is easy and convenient. The version of the artificial head is selected whose diameter is about 2 mm smaller than the reduced humeral head. It is our practice that the smaller diameter and the anatomical reduction of humeral head can allow the tuberosities to be easily pulled down around the shaft of the prosthesis. The humeral head is 10 mm higher than the greater tuberosity, which can lead to satisfactory clinical outcome(12).
The previous literature suggests that the success of the surgery depends on the ability of the tuberosities to heal themselves and on the reduction to the humeral shaft in anatomic positions surrounding the implant (8, 22). In the present study, the clinical outcome of new method is similar to that of the typical method. The outcome of radio graphical assessment suggests that there was no linear relationship between the presence of tuberosity and the function of the shoulder joint. The bone union rate and clinical results were comparable to those of the previous studies. The result of a meta-analysis also demonstrates that the percentage of healed tuberosities did not correlate with the Constant score(23). The possible reason is a small number of specimens. Another possible reason is tuberosity fragments are fixed to each other, to prosthesis and to the humeral shaft. As it is mentioned, stable fixation of the tuberosities is imperative to maximize functional results and decrease the risk of instability regardless of the arthroplasty technique chosen (24).
One strength of our operative technique is that it is based on the idea of fracture treatment. Another strength is that it is personalized, easy and convenient. Thirdly, it is suitable for most proximal humeral fractures unless there is humeral head comminution or with comminuted fracture of the humeral shaft.
Our study has some limitations. Firstly, this study was conducted with a small number of specimens, and studies with much larger samples are needed to confirm the findings. Secondly, it is a retrospective and not randomized study. A randomized controlled trial with a larger number of patients is needed. While we believe our minimum follow-up time of 38.1 months (range 25–55 months) is sufficient to assess the function of shoulder, it still represents only short-term follow-up.