Clinical outcomes of a novel ‘all-suture’ fixation for valgus-impacted proximal humeral fractures

Introduction

Valgus-impacted proximal humeral fractures are thought to occur with a heavy fall onto the side of the shoulder with the arm in an adducted position. The force of the glenoid against the humeral head drives it down towards the shaft. This pushes out the greater tuberosity as a single fragment (frequently the lesser tuberosity is also displaced anteriorly), and the head is impacted onto the top of the shaft under the tuberosities (See Figure 1).^1,2 This fracture pattern confers a better prognosis than other types of displaced proximal humeral fracture with a lower incidence of avascular necrosis, likely due to preservation of the posteromedial blood supply to the humeral head.^2,3 There is no current consensus on the optimal choice of operative management for this fracture subtype, and no compelling published evidence to support routine non-operative management.

Generally, these fractures have been fixed using proximal humerus–specific locking plates. ⁴ This procedure is usually performed through an extensile deltopectoral approach requiring significant dissection and insertion of a metallic implant. ⁵ There are many recognised complications associated with the use of locking plates for proximal humerus fractures.^6–8

There have been a few published studies reporting the results of using trans-osseous sutures to stabilise various types of proximal humeral fractures with associated low complication rates due to the absence of metallic implants.^9,10 These have traditionally been performed through a deltopectoral approach. We believe that there may be a benefit in utilising a minimally invasive approach using rotator cuff sutures alone with minimal disruption to the soft tissue envelope and fracture fragments in valgus-impacted proximal humerus fractures. Our ‘all-suture’ technique is simple, quick and performed through a deltoid splitting approach.

The aim of this study was to report the mid-term functional outcomes and review the post-operative radiological outcomes, following the use of our novel ‘all suture’ surgical technique for the fixation of valgus-impacted proximal humeral fractures.

Materials and methods

This study received clearance from our local research and ethics committee (IRAS Reference Number: 314246). A retrospective review of our institution's trauma database was conducted to identify all adult patients who underwent ‘suture only’ proximal humeral fracture fixation for valgus-impacted fractures between 2015 and 2021 at our hospital trust. We included all patients who underwent this operation with at least 12 months of follow-up.

During this time period, all adult patients presenting with valgus-impacted proximal humeral fractures were operated on by our centre's senior orthopaedic trauma surgeon who specialises in shoulder surgery using the described ‘suture only’ technique. All available clinical notes, including follow-up appointments, operation notes and patients’ radiographs were reviewed. The Neer classification of the fracture, length of follow-up, operation times and patient-reported outcome measures were identified from the electronic patient record. ¹¹

We collected two patient-reported outcome measures at each patient's most recent follow-up, at a minimum of 12 months post-operatively. The first outcome measure collected was the Oxford Shoulder Score (OSS), which has been demonstrated to be a reliable tool for assessing outcomes following shoulder surgery. ¹² The second outcome measure collected was the Single Assessment Numeric Evaluation (SANE), which has been shown to be a reliable and responsive outcome measure in multiple conditions of the shoulder.^13,14

Statistical analyses

The data was collected and statistical analysis was conducted using Excel (Microsoft, Redmond, WA). Post-operative PROMs were compared between subgroups using the paired ‘t’ test. We considered a p-value of less than or equal to 0.05 to be statistically significant.

Operative technique

The patient is positioned in the reclined ‘beach chair’ position (30°), ensuring that the arm on the operative side can be freely manoeuvred during the procedure. A 6–8 cm horizontal skin incision is made with the posterior half 1–2 cm distal to the lateral acromial edge, and the deltoid is split perpendicular to the incision between the anterior and middle heads (Mackenzie approach). ¹² Once through the deltoid, the bursa is excised to expose the fracture. The fault line between the greater tuberosity and lesser tuberosity is then identified; it is invariably found just posterior to the bicipital groove. The articular margin of the impacted humeral head is usually identified within the fault line between the tuberosities. The tendon of the long head of biceps may be visible within the operative field and tenodesis can be performed at this point. Two no. 5 braided non-absorbable sutures are passed through subscapularis (one slightly superiorly and one slightly inferiorly), just proximal to its insertion into the lesser tuberosity. A further pair of no. 5 braided sutures are passed through the posterior part of supraspinatus and infraspinatus just proximal to their insertion into the greater tuberosity (again one slightly superiorly and one slightly inferiorly). The sutures are not tied at this point.

Under radiographic guidance, a Bristow elevator is passed through the fault line to sit under the lateral aspect of the impacted humeral head. This is used to push the lateral aspect of the humeral head superiorly. This disimpacts, elevates and reduces the head against the glenoid. This should be done carefully so as not to disrupt the soft tissue hinge at the medial calcar. Simultaneously to the reduction of the humeral head, tension is exerted on both pairs of rotator cuff sutures (more so for the posterior pair). The counter-traction created by the sutures aids in the reduction of the head and allows the greater and lesser tuberosities to be reduced beneath the humeral head (See Figure 2).

The superior subscapularis suture is tied to the inferior infraspinatus suture and the inferior subscapularis suture is tied to the superior posterior suture. This configuration prevents the superior sutures from bowstringing over the top of the humeral head. Following tying, a hoop stress construct is complete, taking advantage of the anterior and posterior rotator cuff being in continuity with the scapula (see Figures 3 and 4). The tuberosities are fixed in position and this creates a strong buttress, preventing valgus displacement of the humeral head. A bone graft was not used in any of the reported cases.

OPEN IN VIEWER

Figure 1.

Plain AP radiograph demonstrating a valgus-impacted proximal humeral fracture with significant displacement of the greater tuberosity.

OPEN IN VIEWER

Figure 2.

Plain AP post-operative radiograph demonstrating satisfactory post-operative reduction.

OPEN IN VIEWER

Figure 3.

Diagram showing the orientation of the sutures and maintenance of the reduction through buttressing of the humeral head by the GT and LT (rotator cuff not pictured).

OPEN IN VIEWER

Figure 4.

Diagram showing how the hoop stresses of the rotator cuff sutures maintain the fracture reduction.

After the procedure, the arm is placed in a neutral rotation brace which is worn for four weeks. The patient is then enrolled into a standard, physiotherapy-led rehabilitation protocol.

Results

Fifteen patients who had undergone suture-only fixation of valgus-impacted proximal humeral fracture were identified. The mean age of the patients was 56 years (range: 26–72), and the group consisted of 11 females and 4 males. Pre-operative radiographs demonstrated that eight of the patients had 3-part valgus-impacted fractures, whilst the remaining seven had 4-part valgus impacted fractures as per the Neer classification. ⁵ The average number of days from injury to surgery was 7(±3) days. Operative time was available in 13 cases, and the mean recorded time was 66 min (range: 47–94 min).

The median follow-up time was 18 months (range 12–65 months). The mean OSS at the final follow-up was 43.7 (SD ± 4.4). The mean Single Assessment Numerical evaluation score was 85.9 (±10.4).

Analysis of pre-operative imaging revealed that five patients had sustained significantly comminuted GT fractures, while the remaining 10 had a single GT fragment or a single large GT fragment with minor superior comminution. The five patients with comminuted GT fractures had slightly lower mean self-reported outcome scores than those without comminuted GT fractures, although this was not statistically significant (Mean OSS 42.8 vs 44.1, Mean SANE 85.0 vs 87.8, p-value 0.48). The medial calcar healed with more than 5 mm of displacement in four patients but there were no incidences of AVN and no significant difference in outcome measures at the final follow-up.

The mean pre-operative humeral NSA was 170° (range 158°–180°), and the mean post-operative humeral neck-shaft angle was 138° (range 127°–155°). All patients had a complete union of the fracture within six months. Thirteen patients had an anatomical union characterised by NSA of between 115° and 148° and GT lying below the level of superior articular surface of the humeral head. Two patients had NSA more than 150° (valgus) with the superior edge of GT higher than superior articular surface of humeral head; clinical outcome in these two patients was still satisfactory, and the difference in PROMS between patients with post-operative NSA <150 and those with NSA >150 was not statistically significant (Mean OSS 45.1 vs 35.5, p-value = 0.205).

No major intraoperative complications occurred, including perforation of the humeral head or disruption of the medial hinge. Post-operatively, three patients developed significant stiffness and subsequently underwent glenohumeral joint hydrodistension. No revision operations were performed during the follow-up period, and no incidences of avascular necrosis were observed.

Discussion

There is no consensus concerning the correct operative strategy for managing the valgus-impacted fracture. The valgus-impacted proximal humeral fracture appears to be a unique entity and confers a better prognosis than other displaced proximal humeral fractures. ² Due to the preservation of the posteromedial vessels supplying the humeral head, there is a reduced risk of avascular necrosis in this fracture pattern, although a significant rate of avascular necrosis is still reported in published series.^3,10,18 Alternative fixation methods to standard fixed-angled locking plates described in the literature include transosseous suture fixation and percutaneous screw fixation.^19,20 To the best of our knowledge, this is the first paper to describe an all-suture surgical technique for the management of valgus-impacted proximal humeral fractures, without the use of transosseous sutures.

Most techniques utilised in the surgical management of proximal humeral fractures involve metalwork insertion. The estimated post-operative ‘failure’ rate resulting in metalwork removal for proximal humeral fractures is around 10%, with the most common indications for removal being non-union and mechanical complications of the implant. ²¹ The association between plate fixation of proximal humeral fractures and the incidence of avascular necrosis is unclear, although meta-analysis has shown that plate fixation results in a statistically significant increase in the incidence of avascular necrosis when compared to conservative management: we suspect that this is due to the increased dissection of the soft tissue envelope required for plate implantation which may further disrupt the blood supply to the fracture. ²² The all-suture fixation described eliminates the risk of implant-related complications and has a very low risk of disrupting blood supply to the fracture due to minimal soft tissue envelope disruption, whilst simultaneously being simple and quick to perform.

Dimakopoulos et al. have described an all-suture technique using drilled transosseous tunnels for the fixation of valgus-impacted fractures, and Scheer et al. published their case series on the use of rotator cuff sutures combined with transosseous sutures to achieve reduction.^9,10 In our experience, proximal humeral bone is often fragile, increasing the likelihood of sutures ‘cheese-wiring’ through it. Furthermore, to pass these transosseous sutures, significant mobilisation of the fracture may be required which disrupts the soft tissue envelope. It is important to preserve this tissue envelope to maximise the chances of successful healing and reduce the risk of avascular necrosis. In a series using the transosseous suture repair technique, >10% of the patients developed avascular necrosis, and in a systematic review looking specifically at valgus-impacted fractures, the overall rate of avascular necrosis was 7.9%.^10,18 Our described technique takes advantage of the strong rotator cuff for purchase of the sutures which we find to be more than adequate to reduce the tuberosities. This avoids the use of transosseous sutures that may disrupt the soft tissue envelope and contribute to avascular necrosis. We have experienced neither suture cut-out nor any cases of avascular necrosis using this technique.

Medial calcar displacement is an important predictor of avascular necrosis in proximal humerus fracture. ²³ As this technique does not provide rigid fixation, we did observe a degree of displacement of the medial hinge in four cases. However, these patients all had good outcomes with no incidence of avascular necrosis at the final follow-up which we believe reflects the minimal soft tissue envelope disruption. We do not think that further stabilisation of the repair, with either a plate or a fibular strut graft, would provide significant enough benefit to offset the extensive dissection and disruption to the soft tissue envelope that would be required for implantation. We did note that patients with a higher NSA (>150°) had a lower OSS at the final follow-up. Although this was not statistically significant, it may indicate a clinically relevant trend and emphasise the need for near-anatomical reduction to optimise outcomes.

No patients within our cohort required re-operation which compares positively to existing strategies for proximal humeral fracture management using PHILOS (proximal humeral internal locking system) plates where the rate of revision surgery can be as high as 30%. ²⁴ Reported complications related to the use of PHILOS plates include post-traumatic stiffness, intra-articular screw penetration, implant impingement on the acromion and avascular necrosis, with a re-intervention rate of around 13–14%.^25,26 The overall reported re-operation rate for reported ‘least possible fixation techniques’ (surgical management that does not include plate fixation or intramedullary nailing) for valgus-impacted fractures is 3.7%. ¹⁸ Re-operation in humeral fracture repair is complicated by existing metalwork, so if further intervention, such as proceeding to reverse shoulder arthroplasty, is required, it would be simpler to re-operate following ‘all suture’ repair. Post-operative stiffness occurred in three patients within our cohort who were treated with glenohumeral joint hydrodistension, but it is difficult to compare this to other methods of proximal humeral fixation as post-traumatic shoulder stiffness is poorly recorded in the literature and its true incidence is not known. ²⁷

The mean OSS reported by our cohort at the final follow up (mean of 18 months post-operatively) compares favourably to the OSS at final follow up (24 months post-operatively) reported by the ProFHER trial for both surgical and conservative management of all types of proximal humeral fracture (43.7 vs 40.11 (surgical management) and 40.40 (conservative management)), but this study did not analyse valgus-impacted fractures as a distinct subgroup. ²⁸ Although our patient group had a lower average age than the cohort in ProFHER, their subgroup analysis demonstrated no difference in final OSS following proximal humeral fracture between those over and under 65 years of age so we believe that our results were not significantly affected by the lower average age of the patients. ²⁸ However, we appreciate that the quality of soft tissue in the elderly can be poor and should be considered with the use of this technique.

We identified one previous study that investigated outcomes of conservative management specifically for valgus-impacted proximal humeral fractures which reported a mean Constant Score at final follow-up of 71.8. ²⁹ However, this was in a non-contiguous older patient group, with the large majority of the cohort having sustained 2 part or minimally displaced fractures. Therefore, we feel that it is not comparable to our series and does not provide evidence to justify treating the more severe 3 and 4 part displaced valgus-impacted fractures non-operatively.

This technique has a short operating time, with a mean in our series of 66 min, owing to its relative simplicity. The reported average operative time for fixation of proximal humeral fractures in existing literature (including plating, hemi-arthroplasty and intramedullary nailing) ranges from 87 to 113 min.^28,30 The shortened operating time is not only just cost-saving but also reduces the risk of surgical site infection. ³¹

There is also a significant financial benefit associated with the use of our described technique. The cost of the metalwork used in proximal humeral fracture fixation is not an insignificant consideration, with three screw proximal humeral plates costing an average of £444.40. ²⁸ In comparison, the cost of the heavy sutures required for our repair technique is negligible.

This study does have several limitations. It is a relatively small, single surgeon, retrospective cohort study with no control group. However, it is the first report on the results of this technique as far as we are aware. We can say with confidence that all patients included in this study that underwent planned sutures fixation of their fracture were not converted to an alternative form of fixation intraoperatively, for example, to plate fixation or arthroplasty. A study consisting of a larger group of patients with longer term follow-up would be useful to confirm the benefits of this technique.

Our results and experience suggest that the repair of valgus-impacted proximal humeral fractures using sutures and without metalwork implantation is a safe technique that results in stable fracture fixation with acceptable outcomes. It is a simple and easily replicable technique with a short operating time and no requirement for specialist equipment. We believe that the ‘all-suture’ technique is a suitable approach for all patients undergoing operative management for isolated valgus-impacted proximal humerus fractures. However, if poor-quality soft tissues are encountered intraoperatively, as is often seen in the elderly, then proceeding to an alternative method of fracture treatment should be considered.