Proximal humerus fracture–dislocation managed by mini-open reduction and percutaneous screw fixation

Abstract

Background

We describe a minimally open reduction and percutaneous fixation technique for three- and four-part proximal humeral fracture–dislocations which preserves soft tissues.

Methods

Eleven consecutive patients with three-and four-part proximal humeral fracture–dislocations (eight anterior, three posterior dislocations) were treated this way. The dislocation is reduced using a mini-open deltopectoral approach with a horizontal split in subscapularis. Fracture fragments are fixed with percutaneous screws. Constant and Oxford Shoulder Score were collected prospectively.

Results

Mean age was 51 years (range 32–65). Mean follow-up was 36 months (range 24–72 months). At last follow-up mean Constant score was 75 (range 64–86) compared to 88 (range 85–92) for the uninjured shoulder. Mean Oxford shoulder score was 41 (range 34–46). One patient developed avascular necrosis. Screw back out was seen in three patients. These were removed under local anaesthesia. There were no screw penetrations of articular surface. One patient suffered a radial nerve neuropraxia which resolved.

Conclusion

These results are promising and comparable to published literature with other means of fixation for this complex problem. Due to minimal soft tissue dissection the complications rate is low.

Introduction

Management of three- and four-part proximal humeral fracture-dislocation is challenging especially in the younger patients where an arthroplasty may not be the best option. Devascularisation of the humeral head leads to a high risk of non-union or avascular necrosis.^1–3 There is no consensus on the best way to treat these injuries.

Open reduction and internal fixation with a locking plate is often used; however, it is associated with a high rate of complications including impingement, screw cut-out and varus malunion.^4–6 The extensive soft tissue dissection can lead to avascular necrosis and postoperative stiffness.⁷ Removal of locking plates can be challenging. Benefits include improved functional outcome and even if avascular necrosis does develop it may be partial or asymptomatic.⁸ Secondary arthroplasty can be performed later but results tend to be poorer.^9,10

In older patients, primary arthroplasty is usually the treatment of choice, but functional outcome may still be less than satisfactory.^11–14 Significant complications such as infection, loosening and instability may cause long-term problems even after an initially promising result. Arthroplasty is not an ideal treatment in younger patients. Functional outcome can be poor, and the limited survival time of the prosthesis means revision procedures will often be required.

Minimally open reduction and percutaneous screw fixation requires minimal soft tissues dissection hence is an attractive proposition. Metalwork problems such as screw cut-out can be dealt with relatively easily. We describe the technique and present our minimum two-year results of mini-open reduction and percutaneous screw fixation in the management of proximal humeral fracture-dislocation.

Patients and methods

This prospective study from a single centre involved 11 three- or four-part fracture-dislocations of the proximal humerus in consecutive patients.

These patients were fit and active and we felt required head salvage surgery. Patients with head-splitting fractures and those suitable only for arthroplasty were excluded. All patients had a full series of radiographs available from initial injury until final follow-up and five had CT scans to rule out head-splitting fractures.

All operations were performed by the senior author (RP) between January 2008 and December 2014. Mean time to surgery was 2.4 days (range 1–4 days).

Surgical technique

Patients were positioned in a deck chair position on a shoulder table (with back support removed behind the shoulder being operated) with the head and neck secured and prepared and draped in a standard manner. An image intensifier was placed ensuring that good views of the shoulder joint can be obtained intraoperatively. Under X-ray guidance a thorough assessment of each fragment was carried out. A mini-open deltopectoral incision (approximately 3–4 cm) was made anteriorly to expose the subscapularis. A horizontal incision was then made in the mid part of the subscapularis extending from its attachment laterally and then progressing medially into the muscle belly. The superior and inferior portion of the subscapularis was then retracted apart. If the lesser tuberosity was fractured the subscapularis did not need to be split and it could be retracted along with the bony fragment. The articular fragment was identified and gently pushed back into the joint under image control. This step must be performed as delicately as possible without disrupting soft tissue attachments. In the majority of cases the posterior aspect of the dislocated humeral head was jammed over the anterior edge of glenoid. This region was then gently freed up using a flat instrument (Pennybacker or a McDonald's periosteal elevator) and while dis-impacting this area, the head was pushed back into the glenoid either using finger dissection or a ring handle spike. External pressure was also helpful. This step is vital and should be done gently and requires patience as one does not want to destroy blood supply any further and damage any neurovascular structures. No attempt was made to visualise the biceps tendon. If it is blocking reduction tenodesis can be performed but this was not required for any cases in our series. We did not address any labral tears or glenoid rim fractures in this study. All patients had damage to the labrum but none had significant glenoid rim fractures.

The same approach was used for posterior dislocations. One needed to go beyond the glenoid retracting the humeral shaft laterally. The anterior aspect of the dislocated humeral head was usually locked onto the posterior aspect of the glenoid. Again this needed to be gently dis-impacted and the head could then be pushed into the glenoid. One could use a joystick into the humeral head but usually this was not required and care must be taken not to damage bone further. Periodic X-rays helped guide the reduction of the head into the joint. Generally, adequate reduction was achieved once the head was reduced and gentle minor adjustments can then be made to the reduced fragments to improve position. The articular head fragment was then stabilised by a percutaneous Kirschner wire (K-wire) from the shaft into the head (possibly into the lower half of humeral head) under image guidance. A further K-wire was sited from the greater tuberosity, in the reduced position, into the medial calcar if possible. Further K-wires were then placed percutaneously to stabilise the whole fracture configuration (Figure 1(a) to (c)). including one for the lesser tuberosity if required. The K-wires were then substituted with cannulated titanium 4 mm screws following measuring and drilling. Final check radiographs were carried out to ensure the position is maintained. There was a learning curve with this technique, but one should actively try to keep the amount of radiation to a minimum.

Figure 1.

(a) Anterior fracture–dislocation in a 56-year-old female patient, (b) following reduction through a mini deltopectoral incision the fracture is stabilised with K-wires, and (c) the wires are then over drilled and replaced with cannulated screws.

The subscapularis incision is closed and followed by the rest of the wound and skin using Vicryl. A Polysling (Mölnlycke Healthcare, Dunstable, United Kingdom) is applied.

Rehabilitation and follow-up

Radiographs were performed at week 1 and week 3. This is not a rigid fixation; hence we rested the arm in a Polysling for three weeks. At week 3, gentle assisted active exercises were started without rotations. Further radiographs were obtained at week 6 and if these were satisfactory active exercises commenced with the help of a physiotherapist. Patients were reviewed clinically and radiographically at three months, six months and one year.

Function was assessed for all patients at two years using the Constant score which has a maximum of 100 points, and the Oxford shoulder score, which has a score from 0 to 48 points, with a higher score representing better function. The vascularity of the humeral head and union of the fracture were assessed on anteroposterior and axial radiographs. The criteria for union were continuation of trabeculae across the site of the fracture, presence of callus and resolution of the fracture gap. Clinical and radiological evaluation was undertaken by the senior author (RP).

Results

Eight patients sustained anterior fracture-dislocations and three posterior fracture-dislocations (Figures 2(a) to (d) and 3(a) and (b)). The mean age was 51 years (range 32–65) (Table 1).

Figure 2.

(a, b) Posterior fracture–dislocation in a 32-year-old male patient and (c, d) the fracture united following reduction through a mini deltopectoral incision and stabilisation with two cannulated screws.

Figure 3.

(a) CT scan of a four-part posterior fracture–dislocation in a 48-year-old male patient and (b) the fracture successfully united following mini-open reduction and percutaneous stabilisation despite a small head fragment.

Figure 4.

(a) A three-part anterior fracture–dislocation seen in a 52-year-old male patient and (b) the fracture united uneventfully and there was no evidence of avascular necrosis at 24-month follow-up.

Table 1.

Patient demographics.

	Fracture–dislocation
	Anterior	Posterior
Patients (n)	8	3
Male:female	6:2	3:0
Fracture type (n)
Three part	5	2
Four part	3	1

Mean time to surgery was 2.4 days (range 1–4 days). Mean follow-up was 36 months (range 24–72 months). At last follow-up assessment the mean Constant score for the injured shoulder was 75 (SD 6.59, range 64–86) compared to 88 (SD 2.61, range 85–92) for the uninjured shoulder. Mean Oxford shoulder score was 41(SD 3.50, range 34–46).

One patient developed avascular necrosis following a four-part anterior fracture-dislocation. All other fractures united with no evidence of avascular necrosis at the latest follow-up (Figure 4(a) and (b)). Screw back out was seen in three patients. One patient suffered a radial nerve neuropraxia which resolved with conservative treatment. There were no recurrent dislocations during the study period.

Discussion

This case series demonstrates promising medium-term results for the mini-open and percutaneous fixation technique. Outcome scores were high and complication rates are lower or are comparable to other techniques. Fracture-dislocations can be treated safely with this method. Further work is required but this technique has many potential benefits over the traditional methods of fixing fracture–dislocations.

Three- or four-part fracture–dislocation of the proximal humerus is a devastating but rare injury. Therefore, the evidence base is limited, and management remains controversial. There are several different treatment options available, but each has its limitations. Complication rate is high. Head-preserving surgery has a significant risk of avascular necrosis and non-union. Open reduction facilitates reduction and plate placement but there is the risk of further damage to the remaining blood supply.^15,16 Avascular necrosis rate following open reduction is twice that seen after closed reduction.¹⁷

Furthermore, soft tissue dissection can exacerbate stiffness and plates may require removal which can be significant undertaking in locking plates.⁷ Arthroplasty prevents this, but functional outcome may be poor and in younger patients this may not be the optimum treatment choice.

Robinson et al.¹⁸ reported on 58 patients with complex anterior fracture–dislocations of the humeral head. They defined two types of injury. In type 1 the head had retained capsular attachments and demonstrated arterial back bleeding. Type 2 injuries had no soft tissue attachments or arterial bleeding. They suggested internal fixation for type 1 injuries and arthroplasty for type 2 injuries in patients aged over 60. Treatment for the younger patients with a type 2 fracture–dislocation was controversial and best decided on an individual patient basis. These recommendations were based on logistic regression analysis demonstrating that type 2 injury was the only significant predictor of the development of AVN.

Soliman and Koptan¹⁹ prospectively reviewed 39 patients aged under 40 years with a four-part fracture–dislocation of the proximal humerus treated with open reduction internal fixation. Even in this younger cohort at a mean of 26 months three patients had gone on to non-union and eight had developed AVN.

Altay et al.²⁰ described a method of limited open reduction and percutaneous fixation of four-part posterior fracture–dislocations. This was performed through a limited deltopectoral approach of 5 cm in length. The fractures were then stabilised with K-wires under direct vision. In 10 patients only one suffered avascular necrosis after 3.2 years follow-up. In contrast to our study, they were able to operate on all patients within 5 h of injury. In a small case series Silver et al.²¹ described a technique using a guide pin and a cannulated tap from a hip screw set to percutaneously ‘joystick’ a dislocated humeral head back into the glenoid.

Resch et al.²² reported a technique of closed reduction and percutaneous fixation for displaced proximal humerus fractures. Resch reported an AVN rate of 0% in three-part fractures and 11% in four-part fractures. Other authors using the technique described an AVN rate of 0–7.8%.^23,24 Drawbacks include the need for a second procedure to remove the K-wires and humeral block through which they pass.

We previously reported a modification of this technique using cannulated screws instead of K-wires and humeral block which negated the need for a second operation.²⁵ In 32 patients there were no cases of AVN or non-union and outcome scores were good. In this paper we have described a further modification to the technique for its use in treating fracture–dislocations of the proximal humerus. Constant and Oxford shoulder scores were high and AVN rate compares favourably to those previously reported in the literature. Screw back out necessitating removal occurred in three patients. As the screws are not locked into a plate they back out rather than penetrate into the glenohumeral joint. We feel this is an additional advantage. In the one patient who developed AVN it was easy to remove the screws and perform arthroplasty.

Limitations to this study include the small number of patients involved but this is an uncommon injury. The senior author undertook the functional and radiological assessments. Constant score is subjective, and the functional results are determined by the clinician. However, the patients completed the Oxford shoulder score themselves and this correlated well with the Constant score.

Our study shows promising initial medium-term results for the mini-open and percutaneous fixation technique. This technique requires minimal soft tissues dissection which is advantageous in these complex injuries. A larger randomised trial is required but this technique has many potential benefits over the traditional methods of fixing shoulder fracture–dislocations.

Authors’ note

This article has been read and approved by all authors and represents honest work. It has not been published or submitted elsewhere.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethical Review and Patient Consent

Not required for this article.