Comparing Access to Engaging Hill-Sachs Lesions Between the Modified Posterior Deltoid Split Approach and Standard Deltopectoral Approach for Bone Grafting

Abstract

Background:

Hill-Sachs lesions are common after shoulder instability, and treatment options vary but include remplissage or implantation of structural bone graft. Large Hill-Sachs lesions not addressed by remplissage are challenging to manage and may frequently require an open surgical approach for bone filling treatment options. The optimal approach to maximize visualization of the humeral head during these procedures remains unclear.

Purpose/Hypothesis:

The purpose of this study was to compare the area of the humeral head accessed using a modified posterior deltoid split approach versus a standard deltopectoral approach without surgical dislocation, with particular attention to access of engaging Hill-Sachs lesions for the purpose of bone grafting in the setting of anterior shoulder instability. It was hypothesized that both approaches would provide equal access to a simulated Hill-Sachs lesion.

Study Design:

Controlled laboratory study.

Methods:

Four human cadaveric shoulders were mounted in the beach-chair position. The modified posterior deltoid split approach and nonextensile deltopectoral approaches were performed. A typical Hill-Sachs lesion was simulated on the humeri. The percentage of the total surface area of the humeral head that was accessed, including access to the simulated Hill-Sachs lesion, was mapped using 3-dimensional digitizing software.

Results:

The deltopectoral approach provided 45% ± 15.2% access (range, 24% to 58%) to the humeral head versus 22.2% ± 6.1% (range, 17% to 30%) for the modified posterior deltoid split approach (P = .057). The modified posterior deltoid split approach enabled 100% access of the simulated Hill-Sachs lesion compared with 0% for the nonextensile deltopectoral approach. The angle of access to the articular surface was direct and perpendicular with the modified posterior deltoid split approach.

Conclusion:

The overall surface area of the humeral head accessed via the modified posterior deltoid split approach was less compared with the deltopectoral approach; however, the entire area of a typical Hill-Sachs lesion was able to be accessed from the modified posterior deltoid split approach, whereas this area was not well visualized from the standard deltopectoral approach.

Clinical Relevance:

The modified posterior deltoid split approach provided sufficient access to the humeral head for the purposes of grafting an engaging Hill-Sachs lesion in the setting of anterior shoulder instability.

What would become known as the Hill-Sachs lesion was originally described as early as 1861 and defined in 1940 by Harold Arthur Hill and Maurice David Sachs as a compression fracture produced when the posterolateral humeral head impacts the anterior rim of the glenoid during a traumatic anterior instability event.^7,13 Studies have reported that Hill-Sachs lesions occur in 40% to 90% of all anterior instability events and are classically present with anteroinferior dislocations. ¹⁹ In recent years, there has been renewed interest in the Hill-Sachs lesion and its role in recurrent anterior instability. ⁸ Burkhart and de Beer ³ described an “engaging” Hill-Sachs lesion as one in which the humeral head defect engages the rim of the glenoid while the shoulder is in a position of athletic function. Treatment options for addressing an engaging Hill-Sachs lesion include remplissage, derotational osteotomy, increasing the glenoid arc with a bone block procedure, or directly addressing the Hill-Sachs lesion itself with a structural bone graft or partial humeral head resurfacing. There are several described techniques to allow visualization and access to the structural bone graft, including an extended deltopectoral approach with a subscapularis tenotomy and capsulotomy for posterior humeral head access. ¹⁷ Other authors have utilized a lateral or arthroscopic approach to the shoulder for this purpose.^10,21 Given the potential morbidity with extreme external rotation and limited access with published techniques, ¹⁷ a modified approach that allows access and decreases risks to humeral blood supply via the anterior circumflex humeral artery is necessary.

The aim of this study was to determine the area of the humeral head visualized and accessible using a modified posterior deltoid split approach compared with a standard deltopectoral approach without surgical dislocation and assess the access each approach provides to the typical location of an engaging Hill-Sachs lesion for the purpose of bone grafting in the setting of anterior shoulder instability. It was hypothesized that the modified posterior deltoid split approach would provide equal access to the deltopectoral approach due to the mobility of the shoulder and the ability to rotate the proximal humerus to deliver different areas of the humeral head into the surgical wound.

Methods

Four fresh-frozen shoulder cadaveric specimens (mean age, 52.25 years; range, 33-60 years; 3 male, 1 female; all left shoulders) were obtained through MedCure and allowed to thaw overnight before the investigation. Computed tomography scans were obtained of each cadaver to enable 3-dimensional (3-D) image reconstruction and map the areas accessed on these images. The shoulders were mounted to stimulate the beach-chair position. The coracoid, acromion, clavicle, and acromioclavicular joint were marked on the skin with a surgical marker. Both the modified posterior deltoid split and deltopectoral approach were performed on each shoulder.

Modified Posterior Deltoid Split Approach

The modified posterior deltoid split approach was performed first and is demonstrated in Figure 1. This is a modification of the approach described by Wirth et al ²² in 1993 and Hawkins et al ¹¹ in 1996. A 5-cm oblique skin incision was made from a point 1.5 cm medial and inferior to the posterolateral corner of the acromion extending obliquely toward the posterior aspect of the arm in line with the anticipated line of the deltoid muscle fibers. Subcutaneous tissue was divided sharply, the deltoid muscle belly split in the line of its fibers along the posterior deltoid raphe no farther than 5 cm below the posterior edge of the acromion, and retracted both superior-laterally and inferior-medially. The deltoid origin remained intact at the acromion. Subdeltoid bursal tissue was excised sharply to reveal the posterior aspect of infraspinatus and teres minor. The quadrangular space with the axillary nerve and posterior circumflex humeral artery can be visualized in the inferomedial aspect of the exposure. The interval between teres minor and infraspinatus is developed and muscle bellies retracted. The tendons of teres minor and infraspinatus are left intact. The plane between the capsule and rotator cuff musculature is developed by packing a surgical gauze into the space to bluntly dissect the capsule from the under surface of the rotator cuff. This is followed by a longitudinal capsulotomy medial to the posterior cuff insertion, which enables access to the posterior aspect of the humeral head.

Figure 1.

(A) Skin incision and landmarks for the modified posterior deltoid split approach. (B) Exposure of the posterior aspect of teres minor and infraspinatus before developing the interval. (C) Close-up view of the interval between teres minor and infraspinatus; the quadrangular space would be seen with the wound inferomedially. (D) Visualization of the posterior humeral head articular surface after capsulotomy with the arm in neutral rotation; this can be further increased with rotation of the humerus.

Standard Deltopectoral Approach

An axillary-based deltopectoral approach was performed and is illustrated in Figure 2. A 7-cm standard skin incision was created, extending anterolaterally down the arm distally from the coracoid process. The deltopectoral interval was developed, subdeltoid bursal tissue removed, and the rotator interval opened along the biceps tendon. The long head of biceps tendon was tenotomized. Next, a subscapularis peel was performed, removing the insertion of the subscapularis and anterior joint capsule as a single layer while externally rotating the humerus. The humeral circumflex vessels were left intact, representing the inferior boundary of the dissection. No dissection was taken inferiorly to the location of the humeral circumflex vessels. The exposure was then increased with external rotation, extension and elevation of the humeral head, but a surgical dislocation was not performed.

Figure 2.

(A) Skin incisions and surface anatomy for the deltopectoral approach. (B) Exposure obtained through the deltopectoral approach. (C) Superior angle showing the area to be accessed for a typical Hill-Sachs lesion.

Once each approach had been completed, the area of the humeral head that was able to be accessed was outlined using pins for each of the approaches. The humerus was taken through a range of motion, including internal and external rotation along with flexion and extension to ascertain the maximum area accessible by each approach. The angle of access to the humeral head was assessed superiorly to note whether it was tangential or perpendicular to the articular surface. After completion of the surgical approaches and marking of the area accessed, the humeral head was disarticulated from the glenoid and the humerus skeletonized. The humerus was rigidly fixed on a clamp and data collection was performed using the Microscribe 3-D digitizing system (Revware). A point-to-point probe was used to collect the following data: the perimeter of the articular surface of the humerus and a series of surface points of the articular surface, the perimeter of the area and a series of surface points accessed through the deltopectoral approach, and the perimeter of the area and a series of surface points accessed through the modified posterior deltoid split approach. A line along the posterior border of the rotator cuff and along the bicipital groove was collected on each specimen for orientation and reference purposes. These measurements were used to create digitized images that were then analyzed using Rhinoceros 3-D modeling software version 4 (McNeel North America). For each specimen, the percentage of the humeral head accessed by each approach was calculated.

A typical Hill-Sachs lesion was replicated on each of the humeri using scaled graph paper. The lesion was placed 30° to the long axis of the humeral shaft, centered on the humeral bare area with an area of 1.8 × 2.5 cm to represent the Hill-Sachs lesion, as described by Arciero et al. ¹ Rhinoceros 3-D modeling software was used to create digitized models and to analyze the regions of interest. The percentage of the simulated Hill-Sachs lesion accessible through each approach was recorded. The angle of approach and therefore working angle of potential instrumentation was noted.

Results

Anatomic Location of Area Accessed on Humeral Head

When assessing the simulated Hill-Sachs lesion, 100% of this proposed area was able to be accessed from the modified posterior deltoid split approach and 0% from the standard deltopectoral approach. This is shown in Figure 3. Figure 4 demonstrates the area accessed with render overlaid on a 3-D computed tomography scan.

Figure 3.

(A) Posterior view of the humeral head showing the articular area accessed through the modified posterior deltoid split approach (blue pins) with a small amount of overlap of the area accessed with the deltopectoral approach (green pins); (B) Simulated Hill-Sachs lesion (white with pink pins) secured in place, showing 100% access in this specimen with the modified posterior deltoid split approach.

Figure 4.

Renderings of the area accessible from each approach overlaid on 3-dimensional computed tomography reconstruction of the proximal humerus. (A) Area of the humeral head accessed through the standard deltopectoral approach. (B) Area accessed through the modified posterior deltoid split approach.

Total Surface Area Accessible by Approach

The humeral heads varied in size, with a mean total surface area of the articular surface of 3248 mms³ (range, 2261-3478 mm³). A mean area of 1283 mm³ (range, 857-1755 mm³) was able to be accessed from the deltopectoral approach and 664 mm³ (range, 452-975 mm³) from the modified posterior deltoid split approach. This corresponded to a mean (±SD) of 45% ± 15.2% of the humeral head articular surface accessible through the deltopectoral approach and 22.2% ± 6.1% accessible from the modified posterior deltoid splint. Figure 5 demonstrates the analysis of areas of the humeral head able to be accessed using each approach.

Figure 5.

A point-to-point probe was used to collect data that were translated into digitized images and analyzed using Rhinoceros 3-dimensional modeling software. In this example, 54% (1755 mm³) of the humeral head was able to be accessed from the deltopectoral approach versus 30% (975 mm³) from the modified posterior deltoid split approach. (A) Rendering of the surface area of the humeral head visualized from the deltopectoral approach (light blue) and the modified posterior deltoid split approach (turquoise blue) over the entire humeral articular surface (dark blue). (B) Rendering of the surface area visualized from the modified posterior deltoid split approach. (C) Rendering of the surface area visualized from the deltopectoral approach.

Angle of Access to the Humeral Head Articular Surface

Utilizing the deltopectoral approach, access to the area of a typical Hill-Sachs lesion was from an anterior-superior angle, and thus, instrumentation would need to be at a tangential or parallel angle to the articular surface (Figure 6). With the modified posterior deltoid split approach, the angle of instrumentation was perpendicular to the articular surface.

Figure 6.

Superior view of a specimen showing the angle of access and instrumentation from the (A and B) deltopectoral approach versus the (C) modified posterior deltoid split approach.

Discussion

The primary finding of this study was that a typical Hill-Sachs lesion can be accessed directly through the modified posterior deltoid split approach. To date, the ideal approach to the posterolateral humeral head articular surface for the purposes of bone grafting a large (20% to 30% of the humeral head articular surface) and engaging Hill-Sachs lesion has not been identified. There are a variety of options for managing a large Hill-Sachs lesion and improving the bony congruency of the glenohumeral joint in anterior instability. Most surgical options involve glenoid bone augmentation alone, ¹⁹ with the aim of increasing the size of the glenoid track, thus preventing engagement of the Hill-Sachs lesion. The Latarjet procedure in isolation has been shown to convert engaging lesions to nonengaging, even in the setting of large glenoid bony defects. ² Historically, rotational humeral osteotomy was used to externally rotate the proximal humerus to reduce humeral retroversion; however, this has fallen out of favor because of concerns regarding nonunion, delayed union, over rotation, risk of fracture, and posttraumatic arthritis. ¹⁹ Arthroscopic remplissage has been become more popular in recent years, ¹⁴ effectively converting the Hill-Sachs lesion into an extra articular defect, ¹⁹ but this technique has limitations in the setting of significant humeral-sided bone loss. ²⁰ Humeral head bone augmentation theoretically restores the anatomy by filling the defect with autograft bone, allograft bone, partial prosthesis, or a synthetic material, ¹⁹ but the results of this are still under investigation, and the surgical technique has not yet been refined.

The typical Hill-Sachs lesion is located on the posterior superior lateral aspect of the humeral head abutting the bare area. ⁸ Arciero et al ¹ used 142 patients with recurrent shoulder instability to characterize the volume, location, dimensions, and shape of the typical Hill-Sachs lesion. They found that the mean lesion had a volume of 1.22 cm³. Di Giacomo et al ⁶ showed that, although the exact angle of the Hill-Sachs lesion varies depending on the position of the arm during the dislocation event, that it is likely to be in the range of 15° to 30°. This was the rationale used for the simulation of the Hills-Sachs location in the current study and gives guidance as to the specific area of the humeral head that needs to be accessed. Arthroscopic remplissage is the most common, and arguably the safest, method of treatment for a large Hill-Sachs lesion, ¹⁶ but it is not an anatomic solution with some concerns around postoperative loss of motion, particularly when compared with arthroscopic Bankart repair without remplissage.^5,9,18 The remplissage technique is inherently limited in its ability to manage a large volume of humeral head bone loss. Surgeons may wish to consider a more anatomic approach to managing a large and or deep Hill-Sachs lesion, particularly in the setting of glenoid bone loss and multiple recurrences. More research is required into the outcomes of humeral head grafting for the management of a Hill-Sachs lesion as there is minimal literature regarding the results of this uncommon procedure and no standardized technique to date.

The approach used in the current study was a modification of the classic approaches described by Hawkins and Janda ¹¹ and Wirth et al. ²² We altered the skin incision to originate more laterally and extend down the arm rather than toward the axillary crease and made an effort to split the deltoid along the posterior raphe rather than in the muscle belly itself. Hawkins and Janda ¹¹ made a vertical incision through the infraspinatus tendon, while we preferred the technique of Wirth et al ²² of making the split between the infraspinatus and teres minor and retracting them superiorly and inferiorly, respectively. Both previous techniques were performed in the lateral decubitus position, whereas the we utilized the beach-chair position. This approach was designed to give direct access to the area of a typical Hill-Sachs lesion with an optimal working angle for instrumentation (perpendicular to the articular surface of the posterolateral aspect of the humeral head).

Additional considerations with the approach used in this study include the closure of the posterior capsule. In the absence of posterior or multidirectional instability, we would advocate for an anatomic closure of the capsule rather than any advancement or capsulorrhaphy to avoid iatrogenic stiffness, particularly if an anterior capsular procedure has been performed. Future in vitro research should assess range of motion after this approach to determine whether there are any adverse effects. Potential risks of the approach include proximity to the quadrangular space given the relative unfamiliarity with the posterior shoulder anatomy, as the most common approaches are anterior. The quadrangular space contains the axillary nerve and posterior humeral circumflex artery, ¹⁰ which may be injured if the surgeon strays inferior to teres minor, emphasizing the importance of identifying and utilizing the interval between infraspinatus and teres minor. In a cadaveric study, Gurushantappa et al ¹⁰ showed that the distance of the axillary nerve in the quadrangular space is ranges between 3.6 and 8.8 cm from the posterolateral corner of the acromion (mean, 7.46 cm), which allows surgeons to quantify the proximity of the axillary nerve, although there is significant anatomic variation. Burkhead et al ⁴ reported that the axillary nerve is a mean distance of 6.3 cm (range, 4.6-8.0 cm) from the posterior corner of the acromion in fresh cadavers with the arm at 0° of abduction. It is important to limit the distal extent of the deltoid split to no further than 4 to 5 cm from the posterior border of the acromion to prevent injuring the axillary nerve as it travels on the undersurface of the deltoid muscle and risk denervating a portion of the deltoid muscle. The surgeon must be mindful of this distance and may wish to place a stitch at the distal extent of the split to prevent inadvertent propagation, although these concerns arise from the arthroplasty literature where the patient population is older and subscapularis contracture is frequently present.

Miniaci and Gish ¹⁷ described a technique to reconstruct the humeral head anatomy using a side- and size-matched humeral head osteoarticular allograft. For this technique, they described an extended deltopectoral approach including a complete subscapularis tenotomy, a laterally based capsulotomy with anterior-inferior capsule release, maximal external rotation of the humerus, and levering out the humeral head with a flat retractor placed behind the neck of the humerus to access the Hill-Sachs lesion from an anterior-superior angle. The described approach requires a surgical dislocation of the glenohumeral joint. The potential negatives to such an extensive dissection include potential damage to the anterior and/or posterior humeral circumflex arteries, which contribute the majority of the vascular supply to the humeral head, ¹² and subsequent avascular necrosis. Similarly, there are concerns regarding the healing of a subscapularis take down via tenotomy, peel, or lesser tuberosity osteotomy, ¹⁵ and the consequences of subscapularis deficiency for a young patient functionally and with regard to future surgical procedures. The primary difference in the current study was that a surgical dislocation was not performed, although the subscapularis insertion was released with a peel technique along with the anterior joint capsule. The extensile deltopectoral offers a larger exposure of the humeral head. In addition, it can be performed only in the beach-chair position. However, an advantage of using a deltopectoral approach is that concomitant procedures can be performed easily with the same approach, for example, an open Bankart repair or Latarjet procedure. Although with the increasing popularity of all-arthroscopic bone block procedures, including Latarjet and distal tibial allografting, an open access anteriorly may not be needed. A limited open procedure for humeral head bone grafting may be an appropriate adjunct to these arthroscopic techniques. If a surgeon chooses, however, to perform an open anterior approach for another procedure, the extreme positioning and dissection that places the blood supply to the humeral head at risk can be avoided with a concomitant modified posterior deltoid split approach as described. This will protect the humeral head blood supply and allow for ease of instrumentation for humeral head bone grafting.

Other authors have utilized a lateral approach and an arthroscopic technique.^4,21 Zhuo et al ²³ describe using a lateral approach to the shoulder in combination with extreme humeral external rotation and splitting of the infraspinatus muscle belly. This does provide access at a good working angle but involves extreme positioning of the arm and requires splitting and retracting the infraspinatus muscle, which risks injury and potential denervation to the muscle. In 2017, Tang et al ²¹ described an all-arthroscopic approach. This was performed in the lateral decubitus position using standard posterior, anterior, and anterosuperior portals with a set of customized bone grafting instruments. The advantage of this approach is its minimally invasive nature. However, there is limited ability to visualize the lesion as a whole and difficulty customizing and shaping the graft to the individual defect with a rectangular bone block described. The modified posterior deltoid split approach proposed in this study addresses many of the concerns associated with the alternative options as outlined above.

The current investigation showed that one cannot reliably access a Hill-Sachs lesion through a standard nonextensile deltopectoral approach without the additional steps taken by Miniaci et al, ¹⁷ even when the biceps is tenotomized. This indicates that more extensive dissection than performed in this study would be required to adequately access a Hill-Sachs lesion through a deltopectoral approach for the purpose of bone grafting the lesion. In this study, the modified posterior deltoid split approach was performed in the beach-chair position, although we anticipate that it could be performed easily in the lateral decubitus position according to surgeon preference and familiarity. The modified posterior deltoid split approach provided access to the area of interest in all specimens, although the overall area of the humeral head accessed was less than that from the deltopectoral approach. In addition, a smaller area was accessed for both approaches in the 2 cadavers that were noted to have a less favorable body habitus and likely higher body mass index. This effect was more notable in the deltopectoral approach, where the modified posterior deltoid split still provided adequate access to the region of interest. Specimen 1 was a 33-year-old male and was found to have the least presence of arthritic change and best tissue quality, which likely contributed to the favorable access and visualization for this cadaver. This particular cadaver was more in keeping with the population of interest (young athletic patients) than were the remaining 3 cadavers, who were all >50 years. It may be that in young female patients with soft tissue laxity and instability that a larger area would be able to be accessed through a deltopectoral approach; however, this phenomenon may also be countered by the lack of muscle tension and normal soft tissue turgor of the cadaver specimens.

Limitations

We acknowledge several limitations to this study. First, as in all cadaveric studies, there was no physiologic muscle tension and tissue turgor, which may have altered the area accessed by each approach. In addition, the study was performed on 4 cadaveric specimens with a mean age of 52.25 years. A larger sample size and a younger specimen age may better represent the general population and the population of interest when discussing Hill-Sachs lesions. We did not perform the full approach as described by Miniaci et al ¹⁷ due to our concerns regarding the extensive dissection, so we are unable to comment on the amount of exposure using that particular approach. The management of the long head of biceps tendon was standardized as it was absent in some cadavers and the decision was made to tenotomize it for the purposes of this study. In vivo, it is likely that patients who are being considered for Hill-Sachs grafting are not first-time surgical candidates. As a result, the biceps tendon may have already been tenotomized or tenodesed, but the presence or absence and exact location of the tendon may introduce variability in vivo. We recommend management of the biceps tendon based on individual case characteristics and surgeon/patient preference.

Conclusion

The overall surface area of the humeral head accessed via the modified posterior deltoid split approach was less compared with the deltopectoral approach; however, the entire area of a typical Hill-Sachs lesion was able to be accessed from the modified posterior deltoid split approach, whereas this area was not well visualized from the standard deltopectoral approach. The modified posterior deltoid split approach provided sufficient access to the humeral head for the purposes of grafting an engaging Hill-Sachs lesion.