Background:
Humeral avulsion of the glenohumeral ligament (HAGL) lesions are associated with shoulder instability. Arthroscopic repair of anterior HAGL lesions typically requires the placement of an anterior-inferior (5-o’clock) portal, with different variations of this portal described. The purpose of this study was to determine the efficacy of described anterior-inferior shoulder arthroscopy portals for arthroscopic anterior HAGL repair, as well as evaluate the safety of these portals with respect to the surrounding neurovascular structures. Additionally, we sought to evaluate the effect of arm adduction vs. standard abduction during anterior-inferior portal creation.Methods:
HAGL lesions were created and repaired using an all-arthroscopic technique in 12 cadaveric shoulders (matched pairs). Half of the repairs were performed using a standard 5-o’clock portal, whereas the other half of the matched pairs were repaired using a medialized 5-o’clock portal. Repairs were timed, and the number of anchor pullouts was recorded. The shoulders were subsequently dissected to measure the proximity of the portal to the cephalic vein, musculocutaneous nerve, axillary nerve, and lateral cord of the brachial plexus.Results:
The average time for HAGL repair was 18.0 4.6 minutes. Repair times using the medial 5-o’clock portal (19.0 3.3 minutes) vs. standard 5-o’clock portal (16.2 5.8 minutes) were not significantly different (P ¼ .37). From abduction to adduction, the cephalic vein distance from the standard 5-o’clock portal increased from 4.1 4.7 mm to 5.2 5.4 mm (P ¼.02); musculocutaneous nerve distance, from 14.4 9.8 mm to 18.1 10.8 mm (P ¼ .005); axillary nerve distance, from 19.2 9.6 mm to 19.8 9.2 mm (P ¼ .12); and distance of the lateral cord of the brachial plexus, 13.8 6.6 mm to 16.7 6.4 mm (P ¼ .0006).Conclusions:
The arm abduction angle significantly affects the distance of the cephalic vein, musculocutaneous nerve, and lateral cord of the brachial plexus from the anterior-inferior portal, regardless of which portaldstandard or medial 5-o’clock portaldis chosen. This portal should be created with the arm in adduction. Arthroscopic HAGL repair can be performed safely, although accurate anchor placement remains a challenge. There was no advantage to use of the medial 5-o’clock portal. With a curved guide, the standard 5-o’clock portal allows for reproducible anchor placement and is recommended for anterior HAGL repairs.Level of evidence:
Anatomy Study; Cadaveric Dissection 2022 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved.Adam W. Anz, MD*, Joseph Labrum, MD
Journal of Shoulder And Elbow Surgery
*Reprint requests: AdamW. Anz, MD, Andrews Research & Education Foundation, 1020 Gulf Breeze Parkway, Gulf Breeze, FL 32561, USA. E-mail address: anz.adam.w@gmail.com (A.W. Anz).
1058-2746/$ – see front matter 2022 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved.
https://doi.org/10.1016/j.jse.2022.01.117
Introduction
Humeral avulsion of the glenohumeral ligament (HAGL) is an injury pattern associated with shoulder instability.4,30 HAGL lesions may occur in isolation, but they often occur in conjunction with other pathologies such as capsulolabral tears, rotator cuff tears, subscapularis tears, or Hill-Sachs deformities.20,26,28 The reported prevalence of HAGL lesions in cases of anterior shoulder instability ranges from 2.8% to 9.3%,4,14,30 with involvement of the anterior band of the inferior glenohumeral ligament (IGHL) present in up to 93% of cases.1,5 Although relatively uncommon, HAGL lesions are a significant contributor to pain, instability, and functional loss, as well as a cause of failed arthroscopic stabilization when not addressed.12,19 All-arthroscopic repair techniques for HAGL lesions have been described by several authors.7,8,11,16,17,23,25,27 However, open repair remains common, with some advocates postulating that the ability to identify and protect the surrounding neurovascular structures during the open approach may reduce the risk of iatrogenic neurovascular injury.3,13Repair of inferior and posterior HAGL lesions has been described using a posterior-inferior portal for anchor placement.11,22 However, arthroscopic repair of anterior HAGL lesions typically requires the placement of an anterior-inferior (5-o’clock) portal to gain the appropriate trajectory for anchor placement into the anterior IGHL footprint.25 Different variations of this portal have been described when used for arthroscopic HAGL repair: many authors have reported a standard 5-o’clock portal position9,10,17,24,27 (originally described to aid in anchor placement for arthroscopic Bankart repair), whereas others have reported a medialized 5-o’clock portal that may optimize the anchor trajectory specific to HAGL repair.6,14,22 Comparison of these two 5-o’clock portal variants is lacking, specifically regarding the safety of the surrounding neurovascular structures and the efficacy for arthroscopic HAGL repair. Additionally, it has been suggested that the 5-o’clock portal should be created with the arm in adduction to minimize risk to the musculocutaneous nerve,12,22,25 but we are unable to find any evidence in the literature that supports this assertion. We hypothesized that the medial 5-o’clock portal would be closer to the neurovascular structures yet more effective for the anatomic positioning of suture anchors and that arm adduction would increase the distances from the portals to the neurovascular structures.Methods
Twelve matched-pair cadaveric shoulders were used for the analysis. The age range of the 12 matched-pair cadaveric specimens was 47-94 years. The weight range of the cadaveric specimens was 39-77 kg, with an average of 60 kg. The body mass index (BMI) range of the cadaveric specimens was 17.76-27.44, with an average BMI of 22.57. The standard 5-o’clock portal was created in half of the matched pairs, whereas the medial 5-o’clock portal was placed in the other half in the corresponding extremities of the matched pairs. The laterality of the extremity used with the portal (right or left) was alternated each time to account for surgeon hand dominance and perform equal numbers in each group with hand dominance in mind. The cadaveric specimens were entire upper extremities amputated at the thoracic wall or neck with intact scapulas, shoulders, arms, elbows, forearms, wrists, and hands; they were mounted in a standard lateral decubitus position with 10 lb of traction.A posterior portal was established approximately 2-3 cm inferior and slightly medial to the posterolateral corner of the acromion, directed to the coracoid, as described by Andrews et al.2 A 30 arthroscope was then inserted, and an anterior portal was established in an outside-to-inside fashion placed superior to the lateral half of the subscapularis tendon and medial to the long head of the biceps tendon. An anterior HAGL lesion was created using an arthroscopic 4mm flat blade designed to cut capsule tissue (4 mm CapsuleCut Blade without Handle, Arthrex, Naples, FL, USA). One of two 5-o’clock portals was then established: The standard 5-o’clock portal was created in an outside-to-inside fashion at a location 2 cm inferior to the coracoid tip, as previously described,9,10,17,24,27 whereas the medial 5-o’clock portal was established 1 cm medial to the standard 5-o’clock portal, also in an outside-to-inside fashion.6 All portals were made or verified for accuracy by a fellowship-trained shoulder surgeon with 10 years of clinical practice. A 7-mm cannula was placed into the portal. The anterior and 5-o’clock portals were then used as working portals to perform an all-arthroscopic HAGL repair. Two knotless anchors (FiberTak; Arthrex, Naples, FL, USA) were placed into the IGHL footprint through a curved drill guide. Anchor pullouts that occurred during deployment of the anchor or at the time of tensioning were recorded. Repairs were timed from creation of the 5-o’clock portal to completion of the repair.The cannula for the 5-o’clock portal was then left in place, and dissection was performed to identify the proximity of the cannula to the surrounding neurovascular structures. Measurements of distances to the cephalic vein, musculocutaneous nerve, axillary nerve, and lateral cord of the brachial plexus were obtained. A digital caliper was used to measure the distance from the structure in 30° of abduction; this measurement was then repeated after placement of the arm in 0° of abduction. Distances were calculated from the closest distance from the neurovascular structure to the closest part of the disposable cannula. Dissection was then carried into the glenohumeral joint to assess the accuracy of anchor placement within the footprint of the IGHL. One of the cadaveric shoulders was found to have evidence of a prior proximal humeral fracture at the time of arthroscopy and was therefore excluded from the analysis of repair time and anchor placement accuracy.Statistical analysis
Prior studies assessing the risk to anatomic structures posed by shoulder arthroscopy portals have used 5-12 cadaveric specimens18,21; we therefore used 12 specimens in a matched-pair design. The level of significance was set at α = .05. Analyses were performed using the Excel Analysis ToolPak (Microsoft, Redmond, WA, USA). Continuous variables were assessed using the 2-tailed Student t test, whereas categorical variables were assessed using the χ2 test of proportions.
Results
The average time for repair was 18.0 ± 4.6 minutes (mean ± standard deviation). The average time for repair using the medial 5-o’clock portal (19.0 ± 3.3 minutes) was not significantly different from that using the standard 5-o’clock portal (16.2 ± 5.8 minutes) (P = .37). Repair times started with creation of the 5-o’clock portal and continued through completion of the repair and likely differ in a cadaveric model from those in the real-world clinical scenario. For both the standard 5-o’clock portal and the medial 5-o’clock portal, 70% of the anchors placed (7 of 10) were either within the footprint of the anterior band of the IGHL or at the articular margin. The remaining anchors were an average of 4.4 ± 1.3 mm (range, 2.5-7.6 mm) from the articular margin into the humeral head. A total of 6 anchor pullouts occurred, 2 using the standard 5-o’clock portal and 4 using the medial 5-o’clock portal. This difference was not statistically significant (P = .36). Regarding portal placement relative to the conjoint tendon (Table I), the standard portal resulted in placement lateral to the conjoint tendon 2 of 6 times, placement through the conjoint tendon 3 of 6 times, and placement medial to the tendon 1 of 6 times. The medial 5-o’clock portal resulted in lateral placement in 1 of 6, placement through the conjoint tendon in 4 of 6, and medial placement in 1 of 6. This difference was not statistically significant (P = .79).
| Structure | Abduction, mm | Adduction, mm | P value | Injury |
|---|---|---|---|---|
| Cephalic vein | 4.1 ± 4.7 | 5.2 ± 5.4 | .02* | Yes (n = 3) |
| Musculocutaneous nerve | 14.4 ± 9.8 | 18.1 ± 10.8 | .005* | No |
| Axillary nerve | 19.2 ± 9.6 | 19.8 ± 9.2 | .12 | No |
| Brachial plexus | 13.8 ± 6.6 | 16.7 ± 6.4 | .0006* | No |
* Statistically significant (P < .05).
The distance from the portal to the cephalic vein increased from 4.1 ± 4.7 mm (range, 0-14.8 mm) with portal placement in arm abduction to 5.2 ± 5.4 mm (range, 0-17.8 mm) with placement in arm adduction (Table II). This result was statistically significant (P = .02). The portal came into contact with the cephalic vein once using the standard portal (regardless of whether the portal was placed in abduction or adduction) and twice using the medial portal (again regardless of whether the portal was placed in abduction or adduction). The distance from the portal to the musculocutaneous nerve increased from 14.4 ± 9.8 mm (range, 2.6-36.1 mm) in 30° of shoulder abduction to 18.1 ± 10.8 mm (range, 2.6-41.7 mm) in adduction (ie, 0° of shoulder abduction). This result was statistically significant (P = .005). There were no instances in which the portal came into contact with the nerve. The distance to the axillary nerve was not significantly affected by arm abduction angle during portal placement: Measurements of 19.2 ± 9.6 mm (range, 6.0-37.1 mm) in abduction and 19.8 ± 9.2 mm (range, 7.8-37.1 mm) in adduction were obtained (P = .12). The portal did not come into contact with the axillary nerve in any instance. The average distance to the lateral cord of the brachial plexus was significantly different, increasing from 13.8 ± 6.6 mm (range, 6.9-28.7 mm) in 30° of shoulder abduction to 16.7 ± 6.4 mm (range, 9.1-29.8 mm) in adduction (ie, 0° of shoulder abduction) (P = .0006). There were no instances of the portal coming into contact with this structure.
Subgroup analysis of the individual effects of arm abduction angle on both the standard and medial 5-o’clock portals is summarized in Table II. The increased distance was statistically significant for both the standard and medial portals for the musculocutaneous nerve (P = .03 and P = .04, respectively) and the lateral cord of the brachial plexus (P = .003 and P = .02, respectively). For the cephalic vein, the standard portal demonstrated a statistically significant increase from arm abduction to adduction (P = .04) whereas the medial portal did not (P = .10). Neither portal showed a statistically significant difference with respect to the axillary nerve (standard, P = .16; medial, P = .10). Comparisons of the distances of the standard portal vs. the medial portal to the various anatomic structures in either arm abduction or adduction were not significant in any instance (Table II).
Discussion
The most important finding of this study is that there was no advantage of a medial 5-o’clock portal when compared with a standard 5-o’clock portal and that arm adduction did, in fact, increase the distance from the portal to the neurovascular structures. Our study provides data to support the assertion that adduction of the arm reduces the risk of injury to the musculocutaneous nerve12,22,25, as arm adduction increased the distance from the musculocutaneous nerve to the 5-o’clock portal (by an average of almost 4 mm). In addition, adduction of the arm was helpful in maximizing the distance of an anterior-inferior portal from the lateral cord of the brachial plexus and the cephalic vein. Although the lateral decubitus position may add a level of difficulty when making the portal in adduction, we did not find it difficult to visualize the humeral footprint after portal placement while abducting for anchor placement. It is important to note that we observed no instances of either portal coming into contact with the musculocutaneous nerve, axillary nerve, or lateral cord. However, there were 3 instances of the portal abutting the cephalic vein; although the vein was not visibly damaged, these cases should be presumed to have injury to the vein. Cephalic vein injury during shoulder arthroscopy is a reported complication but is typically not associated with significant clinical morbidity.15,18,29
| Structure | Abduction | Adduction | P value (abd vs add) |
|---|---|---|---|
| Cephalic vein | |||
| Standard, mm | 6.38 ± 5.6 | 7.91 ± 6.5 | .04* |
| Medial, mm | 1.71 ± 2.06 | 2.43 ± 2.1 | .10 |
| P value (standard vs medial) | .11 | .09 | |
| Musculocutaneous nerve | |||
| Standard, mm | 16.3 ± 6.0 | 20.0 ± 7.2 | .03* |
| Medial, mm | 12.6 ± 12.9 | 16.2 ± 14.0 | .04* |
| P value (standard vs medial) | .49 | .54 | |
| Axillary nerve | |||
| Standard, mm | 22.2 ± 11.4 | 22.4 ± 11.2 | .16 |
| Medial, mm | 16.1 ± 7.1 | 17.1 ± 6.5 | .10 |
| P value (standard vs medial) | .24 | .31 | |
| Brachial plexus | |||
| Standard, mm | 17.5 ± 7.7 | 20.1 ± 6.7 | .003* |
| Medial, mm | 10.1 ± 2.5 | 13.2 ± 3.9 | .02* |
| P value (standard vs medial) | .08 | .07 |
The P values in the second and third columns represent analysis of the distance using the standard portal vs. that using the medial portal, whereas the P values in the fourth column represent the effect of the arm abduction angle using each portal.
* Statistically significant (P < .05).
Lo et al18 assessed the distances of various shoulder arthroscopy portals, including a 5-o’clock portal, to anatomic structures using 5 cadaveric specimens. They reported mean distances of 9.8 mm, 27.9 mm, 33.3 mm, and 34.8 mm to the cephalic vein, musculocutaneous nerve, axillary nerve, and lateral cord of the brachial plexus, respectively. They did not encounter any instances of the 5- o’clock portal coming into contact with these structures. By comparison, our study found average distances of 4.1 mm, 14.4 mm, 19.2 mm, and 13.8 mm (in arm abduction), respectively. The differences in these measurements are likely attributable to technique. In their study, Lo et al investigated the 5-o’clock portal in the setting of Bankart repair and, accordingly, targeted a 30°-45° angle of approach to the glenoid rim. This would likely result in a more lateral skin entry point, increasing the distance from the structures in question. Furthermore, their 5-o’clock portal is described as being more superior (1 cm inferior to the anterior portal, through the superior subscapularis), which may also affect the results.
Meyer et al21 also investigated the distances of shoulder arthroscopy portals to anatomic structures of interest using 12 cadaveric shoulders. A 5-o’clock portal was included; however, this was placed using an inside-to-outside technique, which is likely to yield different results. They reported distances of 17 mm and 15 mm for the cephalic vein and axillary nerve, respectively. They did not include measurements to the musculocutaneous nerve or other brachial plexus structures as part of their analysis.
As part of their original description of the 5-o’clock portal, Davidson and Tibone9 reported distances of 22.9 ± 4.9 mm to the musculocutaneous nerve and 24.4 ± 5.7 mm to the axillary artery in 11 cadavers. It is interesting to note that they also investigated the effect of arm abduction angle (as well as flexion and rotation) on the distance to the musculocutaneous nerve but did not find any significant differences. However, this portal was also placed in an entirely different fashion, using an inside-to-outside technique.
Regarding the anatomy of the musculocutaneous nerve, in their cadaveric dissection series of 79 shoulders, Resch et al24 reported the distance of the musculocutaneous nerve inferior to the palpable coracoid tip to be an average of >5 cm. In no instance was the nerve <2 cm from the coracoid. This finding is in keeping with the findings of our study, in which there were no instances of musculocutaneous nerve injury, and suggests that an arthroscopic portal established at this location (approximately 2 cm inferior to the coracoid) should have a very low likelihood of this neurologic injury.
We also report on the location of the 5-o’clock portal relative to the conjoint tendon— which we have not seen reported elsewhere. Notably, we encountered a high rate of penetration of the conjoint tendon using either the standard or medial 5-o’clock portal, in addition to 2 instances of the portal passing medial to the conjoint tendon. The clinical significance of this finding is unknown. Presumably, the primary danger of injuring the conjoint tendon would be an injury to an associated neurovascular structure, particularly the musculocutaneous nerve, which did not occur. The average repair time is also noteworthy and suggests that all-arthroscopic HAGL repair can be performed expeditiously, and this may prove useful as an additional time estimate when the surgeon encounters an unexpected HAGL lesion during arthroscopic instability cases.
Study limitations
Limitations of this study include an average cadaveric age of 76.7 years (range, 47-94 years), which is older than the average age of patients presenting with HAGL lesions5,20—although these lesions are reported to occur in a wide age range—and the variability in the size of the cadavers, with a weight range of 39-77 kg and a BMI range of 17.76-27.44. This could potentially impact the rate of anchor pullout, as well as anchor positioning, and is less likely to impact the distances of the neurovascular structures to the arthroscopic portal.
The cadaveric model with amputation of the brachial plexus at the neck may have changed the relationship of the brachial plexus to the portal sites. Additionally, the small mean distances from the portals to the neurovascular structures, combined with the large ranges, limit the meaning and/or utility of the distance measurements in themselves. The large range in measurements is partially due to variability in cadaver size. In addition, the sample size was determined based on literature precedent and on the number of cadaveric specimens that could feasibly be obtained. This does potentially put the results at risk of a β error.
Conclusion
Adduction of the arm results in a statistically significant increase in the safety margin when placing a 5-o’clock portal with respect to the portal’s distance to the cephalic vein, musculocutaneous nerve, and lateral cord of the brachial plexus. All-arthroscopic HAGL repair can be performed safely, although accurate anchor placement may be challenging, even with a medial 5-o’clock portal variant. We did not identify an advantage to use of the medial 5-o’clock portal, perhaps owing to the use of curved drill and anchor insertion guides.
Disclaimers:
Funding: Arthrex supplied anchors free of charge for the execution of the study as part of a grant (IIRR-01220, ‘‘A Cadaveric Comparison for Safety and Efficacy of Two Anterior Portals During Arthroscopic Anterior HAGHL Repair’’). The study design, study execution, data capture, data analysis, and manuscript preparation and/or editing were not part of the grant or a fee-forservice model and were not influenced by Arthrex. Conflicts of interest: Adam W. Anz receives speaking fees, consulting fees, and royalties from Arthrex. The Andrews Research & Education Foundation has applied for and has received research grants to perform research from Arthrex. No part of the study design, study execution, or manuscript preparation was part of a fee for hire. The other author, his immediate family, and any research foundations with which he is affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.
References
1. Ames JB, Millett PJ. Combined posterior osseous Bankart lesion and posterior humeral avulsion of the glenohumeral ligaments: a case report and pathoanatomic subtyping of ‘‘floating’’ posterior inferior glenohumeral ligament lesions. J Bone Joint Surg Am 2011;93:1-4. https://doi.org/10.2106/jbjs.k.000102. Andrews JR, Previte WJ, Carson WG. Arthroscopy of the ankle: technique and normal anatomy. Foot Ankle Int 1985;6:29-33.
3. Arciero RA, Mazzocca AD. Mini-open repair technique of HAGL (humeral avulsion of the glenohumeral ligament) lesion. Arthroscopy 2005;21:1152. https://doi.org/10.1016/j.arthro.2005.06.009
4. Bokor DJ, Conboy VB, Olson C. Anterior instability of the glenohumeral joint with humeral avulsion of the glenohumeral ligament. J Bone Joint Surg Br 1999;81:93-6.
5. Bui-Mansfield LT, Banks KP, Taylor DC. Humeral avulsion of the glenohumeral ligaments: the HAGL lesion. Am J Sports Med 2007;35: 1960-6. https://doi.org/10.1177/0363546507301081
6. Burkhart SS, Brady PC, Denard PJ, Adams CR, Hartzler RU. The cowboy’s conundrum: complex and advanced cases in shoulder arthroscopy. Philadelphia, PA: Wolters Kluwer; 2017.
7. Chhabra A, Diduch DR, Anderson M. Arthroscopic repair of a posterior humeral avulsion of the inferior glenohumeral ligament (HAGL) lesion. Arthroscopy 2004;20:73-6. https://doi.org/10.1016/j.arthro. 2004.04.032
8. Cross AG, Tramer JS, Guo EW, Muh SJ, Makhni EC. Arthroscopic repair of humeral avulsion of the glenohumeral ligament lesion with capsular plication in the lateral decubitus position. Arthrosc Tech 2021;10:e569-74. https://doi.org/10.1016/j.eats.2020.10.042
9. Davidson PA, Tibone JE. Anterior-inferior (5 o’clock) portal for shoulder arthroscopy. Arthroscopy 1995;11:519-25.
10. Flury M, Rickenbacher D, Audig e L. Arthroscopic treatment of anterior shoulder instability associated with a HAGL lesionda case series. J Shoulder Elbow Surg 2016;25:1989-96. https://doi.org/10. 1016/j.jse.2016.02.030
11. Fritz EM, Pogorzelski J, Hussain ZB, Godin JA, Millett PJ. Arthroscopic repair of humeral avulsion of the glenohumeral ligament lesion. Arthrosc Tech 2017;6:e1195-200. https://doi.org/10.1016/j.eats.2017. 04.008
12. George MS, Khazzam M, Kuhn JE. Humeral avulsion of glenohumeral ligaments. J Am Acad Orthop Surg 2011;19:127-33. https:// doi.org/10.5435/00124635-201103000-00001
13. Godin JA, Sanchez G, Kennedy NI, Ferrari MB, Provencher MT. Open repair of an anterior humeral avulsion of the glenohumeral ligament. Arthrosc Tech 2017;6:e1367-71. https://doi.org/10.1016/j.eats.2017. 05.019
Journal of Clinical & Medical Surgery
For additional information on ACL knee injuries, or to learn more about what is involved during ACL reconstruction surgery, please contact the office of orthopedic knee surgeon, Dr. Adam Anz, serving the greater Pensacola, Gulf Breeze, and Gulf Coast communities.
