Biomechanical Effect of Dermal Allograft or Bone Block Augmentation on Standard Labral Repair in a Time-Zero Cadaveric Model of Glenoid Bone Loss

Abstract

Purpose

To biomechanically evaluate the stability of a labral repair technique using dermal augmentation (DA) compared with standard labral repair (SLR) and bone block augmentation (BBA) in the context of subcritical glenoid bone loss in a cadaveric model.

Methods

Eight cadaveric shoulder specimens were disarticulated and dissected, leaving only labral tissue. The specimens were secured to custom fixtures with 60° of glenohumeral abduction, 30° of external rotation, 20° of scapular anterior tilt, and 40 N of glenohumeral compression. In a repeated-measures design, the specimens were tested in 5 states: native, defect (with 15% glenoid bone loss), SLR, DA, and BBA. Dislocation was performed at 1 N/s over the anterior-inferior rim. Maximum load and distance to dislocation were recorded and statistically analyzed.

Results

Maximum loads for the defect state and SLR were significantly less than those for the native state (P = .005 and P = .008, respectively) and BBA (P = .014 and P = .022, respectively). The distance to dislocation decreased significantly for the defect state (P < .001) and SLR (P < .001) compared with the native state. The distance to dislocation for the defect state, SLR, and DA was significantly reduced compared with BBA (P < .001, P < .001, and P < .001, respectively). The distance to dislocation was significantly greater for DA than for the defect state (P < .001) and SLR (P = .001).

Conclusions

Dermal allograft repair showed similar biomechanical characteristics to the native state and showed an increased distance to dislocation and higher forces at dislocation compared with SLR. BBA restored the distance to dislocation and resulted in non-significantly increased maximum loads when compared with the native state.

Clinical Relevance

DA of labral repair is an alternative procedure for subcritical glenoid bone loss that potentially mitigates adverse effects from bone grafting. Future clinical investigation is warranted.

Anterior shoulder dislocations are a common problem, with an incidence of about 200 per 100,000 persons per year.¹ The presence of bone loss can be a common finding after dislocation, occurring in up to 22% of patients after a first-time dislocation and up to 88% of patients with recurrent shoulder instability.² Patients with less than 15% to 20% bone loss, defined as subcritical bone loss (SBL), have typically been treated with a standard labral repair (SLR) and anterior capsular shift,3, 4, 5 although these procedures have been associated with failure rates approaching 100% with as little as 13.5% bone loss.^4,6,7 Whereas the addition of remplissage is used by some surgeons to reduce recurrence risk,⁴ many surgeons have advocated the use of a bone block augmentation (BBA) procedure to reduce redislocation rates in patients with SBL. These procedures, however, have been associated with complication rates near 30% and arthritis rates approaching 25% to 35% at 20 years.8, 9, 10

The use of remplissage in combination with SLR has been shown to decrease shoulder apprehension in recent clinical studies¹¹ and has been biomechanically characterized to increase stability compared with SLR alone.¹² This procedure, however, disrupts the native capsular anatomy and is known to reduce external rotation of the humerus.¹¹ Furthermore, it does not directly address the bone loss or restore the native translation of the humeral head on the glenoid. The labral scaffold technique is an alternative method to achieve glenohumeral stability while better maintaining the native anatomy.¹³ Additionally, this technique may be combined with remplissage to further attenuate glenohumeral apprehension, although this is not commonly implemented by the senior author (S.A.). This approach is thought to compensate for glenoid bone loss by increasing the surface area of the glenoid by inserting a dermal scaffold between the glenoid neck and the labrum, in addition to the conventional capsulolabral repair.¹³ Moroder et al.¹⁴ described a comparable technique using an allogeneic demineralized spongy bone matrix for small to intermediate glenoid bone defects.

The purpose of this study was to biomechanically evaluate the stability of a labral repair technique using dermal augmentation (DA) compared with SLR and BBA in the context of subcritical glenoid bone loss in a cadaveric model. It was hypothesized that the DA technique would increase the distance to dislocation when compared with the defect state and labral repair and would have similar load values when compared with the native state.

Methods

Eight fresh-frozen, right shoulder-to-elbow, male cadaveric specimens were obtained (average age, 62 ± 6 years; age range, 55-75 years). Owing to the cadaveric nature of this study, institutional review board approval was not required. The specimens were stored at –28°C and were thawed at room temperature for 24 hours prior to dissection and biomechanical testing. Once thawed, specimen humeri were separated from their scapulas and all soft tissue was resected except for the glenoid labrum. The labrum was wrapped in phosphate-buffered saline solution–soaked gauze until the testing sequence began.

Specimen Preparations

The humeral stem was cut to a consistent length and was potted in fiberglass potting material (Bondo; 3M, Saint Paul, MN). As previously described in the literature, the scapula was potted so that the glenoid face was parallel to the floor.15, 16, 17 To do so, the glenoid was placed face down on a custom-made aluminum fixture. To maintain its position, 3 surgical pins were fed through the parallel walls of the fixture and scapular anatomy 10 to 15 mm medial from the glenoid cavity, approximately along the lateral angle. The medial and inferior scapular anatomy was trimmed to fit a silicone potting mold. The scapula-fixture assembly was then suspended inversely, and osseous material medial to the glenoid cavity was potted in a rectangular mold. Once curing was complete, the surgical pins were removed and the glenoid plane angle was checked for pitch using a digital level that was calibrated using a flat surface. The scapular potting steps were repeated if the glenoid plane deviated by more than 3 degrees in either direction relative to the table surface.

Biomechanical Setup

A high-efficiency X-Y translation plate was attached at the base of the ElectroPuls Mechanical Testing System (Instron, Norwood, MA). A 1-kN load cell (Instron) was secured to an angled vise set to 45° and attached atop the X-Y translation plate. The 45° vise angle ensured that the vertical loading direction (perpendicular to the top surface of the X-Y plate) would induce translation of the humerus across the anteroinferior glenoid rim.⁴ A custom fixture for the humerus was attached to the actuator to achieve 60° of glenohumeral abduction, simulating 90° of shoulder abduction.^5,6

Humeral Specimen Mounting

The potted humerus was inserted in the humeral abduction fixture so that the humerus’s center of rotation was in line with the loading axis of the actuator. The prepared scapula was secured within the angled vise so that the glenoid anterior-inferior rim was positioned upward, allowing the humerus to translate in the direction of an anterior-inferior shoulder dislocation. The humerus and the glenoid cavity were orientated so that the humerus’s bicipital groove aligned with the acromion’s most anterior point, achieving 90° of external rotation.³ The humerus was rotated 60° internally, resulting in 30° of external rotation, which was maintained throughout the testing sequence.¹

Starting Position Identification

To identify the glenoid’s center of rotation, the scapula was removed from the angled vise and was set face up so that the glenoid plane was parallel to the floor. To identify the cavity’s deepest point, the glenoid was filled with phosphate-buffered saline solution, and a 1.59-mm-diameter steel ball was dropped within. The steel ball’s settling point (often coinciding with a visually identified “bare spot”) is the point on the cavity where the humeral apex articulates with the scapula.¹⁶ A pinhole was drilled at the identified center position. To identify the most medial aspect of the mounted humerus in the testing position, an ink pad was positioned perpendicular to the translation plate. The translation plate was moved until first contact with the mounted humerus.

The alignment of the identified glenoid center and most medial aspect of the humeral head replicated natural anatomic glenohumeral positioning in the indicated test orientation. The scapula was secured back to the vise, and a 3-dimensional digitizer (MicroScribe; GoMeasure3D, Green Cove Springs, FL) was used to identify the vertical distance between the 2 points. The testing machine’s linear actuator was adjusted to converge the 2 anatomic points.

Glenohumeral compression was applied by hanging weights to a force of 40 N from the translation plate in the medial-lateral direction. The amount of compression was chosen based on average glenohumeral contact forces during functional activities of daily life.¹⁸ After the weights were hung, the starting point was secured by locking the X direction of the X-Y plate, allowing translation in an isolated anterior-inferior loading direction. The resulting test setup can be seen in Figure 1.

Fig 1.

(A) Anterior-inferior shoulder dislocation biomechanical test setup, including humeral abduction fixture (1), angled vise set to 45° (2), 1-kN load cell (3), X-Y translation plate (4), glenohumeral weight (5), and 3-dimensional digitizer (6). (B) Starting position of humerus and glenoid.

Testing Sequence

The biomechanical testing sequence was conducted so that a single specimen would be tested in 5 sequential phases: native state, 15% bone defect, SLR, DA with dermal allograft, and BBA repair. SLR and DA were conducted in alternating order for every other specimen.

Testing

Axial loading of the humerus across the glenoid surface in the direction of an anterior-inferior dislocation was executed at 1 N/s until the humeral head completely translated past the glenoid rim. Load and displacement data were collected at 500 Hz. The humerus remained within the abduction fixture for repeatability throughout all test states. The scapula was removed between each testing state for defect creation and repair and was then returned within the angled vise for testing. Distance to dislocation and maximum load were recorded and reported as the percentage change from the native state. Distance to dislocation was defined as the actuator distance from the starting position to the point at which the load-displacement curve indicated 0 N.

Repair Techniques

Defect Creation

After native testing, a 15% anterior-inferior bony defect was created in each specimen. Figure 2 demonstrates the step-by-step creation of the defect. On the basis of the clock-face identification method, a line was drawn from the 12- to 6-o’clock position and from the 3- to 9-o’clock position. A pinhole marking was made at the intersection of the 2 lines (Fig 2A). Through the pinhole marking, a line was drawn from the 10:30 clock-face position to the 4:30 clock-face position (Fig 2B). The length of the 10:30 to 4:30 line (D) was measured with calipers, making sure to include the glenoid rim. The length (D) was multiplied by 0.15 to achieve the desired bone removal value for a 15% defect. Beginning at the 4:30 clock-face position and moving inward, a marking was made at a distance equating to the bone removal value (Fig 2C). A line perpendicular to the 10:30 to 4:30 line was drawn, indicating the defect size. The defect was then removed using a small osteotome, ensuring that the labrum and attached capsule were carefully elevated from the glenoid neck (Fig 2D). The defect was measured and was found on average to measure 4.67 ± 0.52 mm.

Fig 2.

Fifteen percent glenoid bone loss defect creation via clock-face identification method. (A) A line is drawn from the 12- to 6-o’clock position and from the 3- to 9-o’clock position. (B) Through the center marking, a line is drawn from the 10:30 clock-face position to the 4:30 clock-face position (D). (C) The length of the 10:30 to 4:30 line is measured with calipers. The length is multiplied by 0.15 (15%) to achieve the desired bone removal value. Starting at the 4:30 point of the 10:30 to 4:30 line, the bone removal value is measured inward from the glenoid rim and a perpendicular line is created, indicating the defect size. (D) The glenoid is removed as shown.

Labral Repair

A Bankart repair was completed as an SLR. The labral repair was performed using three 1.8-mm knotless all-suture anchors (FiberTak; Arthrex, Naples, FL). The anchors were placed 10 mm apart, starting 5 mm from the distal edge of the defect. The repair stitch was passed through both the labrum and the capsule and was converted in the knotless mechanism of the anchor using the shuttle suture. This created a capsulorrhaphy, incorporating both the labrum and the capsule tightened across the defect. To ensure appropriate tension, all anchors were placed with sutures passed, after which each was tensioned in sequence and then tensioned a second time prior to suture cutting.

Dermal Augmentation

DA was achieved by the labral scaffold technique.¹³ DA was performed using two 1.8-mm knotless all-suture anchors placed 2 mm below the articular surface, about 5 mm from the distal edge of the defect. A dermal allograft measuring 3 mm in thickness was cut into 10 × 10-mm pieces. For each all-suture anchor, two 10 × 10-mm allograft pieces were stacked. The repair stitch from the anchor was passed in a mattress fashion through the upper one-third of the graft. A suture with a closed loop (FiberLink SutureTape; Arthrex) was passed around the loop created by the mattress stitch to provide a countertraction stitch, as well as to provide another suture to pass through the capsule. The repair stitch was then converted in the knotless all-suture anchor, and the graft was shuttled to the glenoid neck in the defect (Fig 3A). Once in the appropriate position, the suture link was cinched (Fig 3B), the repair suture was passed through the labrum, and the suture link was passed through the capsule. Both sutures were then brought into a 2.9-mm PushLock anchor (Arthrex) placed 5 mm superior to the all-suture anchor (Fig 3C). The process was then repeated for the second all-suture anchor. Once complete, the graft was completely covered with the labrum and capsule (Fig 3D) and filled the 15% defect in an extra-articular position (Fig 3E).

Fig 3.

Labral dermal augmentation procedure. (A) The graft is shuttled against the glenoid into the 15% defect using the knotless mechanism of the anchor. (B) The repair suture (blue) is passed through the labrum, and the suture link (striped) is passed through the capsule. (C) A PushLock anchor is used to place both sutures (striped and blue) at the edge of the defect, making sure that the anchor is placed 5 mm proximal to the all-suture anchor. (D, E) The final repair shows the graft “scaffolding” the labrum and capsule in an extra-articular position.

Bone Block Augmentation

For BBA, the Latarjet procedure was performed. Coracoid transfer was performed by cutting the coracoid using a sagittal saw. The inferior edge was flattened so as to conform with the defect, and the coracoid was fixed using two 3.5-mm fully threaded, cannulated titanium screws (Arthrex) drilled through the coracoid and into the glenoid. It was essential to verify that the graft was flush with the articular surface, ensuring that it would not sit proud or sit too low on the glenoid neck.

Statistical Analysis

All measured results were normalized with respect to the native state owing to the repeated-measures design of the study. Normalization was completed by dividing each state’s value by the corresponding native value to give a percentage of the native state. Therefore, 100% is equivalent to the native results for all outcomes. The native values used for the normalization of all subsequent states can be found in Appendix Table 1.

The Shapiro-Wilk test was used to test for normality within the data set. A 1-way repeated-measures analysis of variance (ANOVA) with the Bonferroni multiple comparisons procedure was used to compare the defect and repair states with the native state and to find statistical differences between repair states. For data that failed normality, a Friedman repeated-measures ANOVA on ranks with the Tukey test for pair-wise multiple comparisons was used. The influence of SBL repair on restoring glenohumeral joint stability was assessed with a 2-way ANOVA, in which repair state and native translation were independent variables. All statistical procedures were performed on the raw data values as repeated measures, and P < .05 was defined as the level of significance for all statistical analyses.

Results

Maximum Load

The maximum load for all tested states was collected, statistically analyzed, and compared with the native state (Fig 4). The maximum load for the defect state was 39% ± 19% (mean ± standard deviation) of that for the native state. For labral repair, it was found to be 43% ± 21% of that for the native state. Both the defect state and labral repair had significantly lower maximum loads than the native state (P = .05 and P = .036, respectively). The maximum load for scaffold repair was 69% ± 28% of that for the native state, with a maximum load of 156% ± 94% for Latarjet repair. There were no significant differences between DA or BBA repair and the native state. Significance was also tested between repair groups. The difference in maximum load between DA and BBA repair was not statistically significant. BBA was found to have a greater maximum load when compared with the defect state (P = .05) and labral repair (P = .036). No other significant differences were found. Table 1 shows all resulting P values.

Fig 4.

Maximum load change from native state. Asterisks represent significant differences between repair states (P < .05). Daggers represent significant differences between the represented state and the native state (P < .05). Plot whiskers indicate upper and lower data limits. Dashed line indicates 100% of native maximum load. Vertical lines and X within the quartile box indicates the median and mean respectively for each data set. The dots outside of the box represent data outliers.

Table 1.

Resulting P Values From 1-Way Repeated-Measures ANOVA for Maximum Load

	P Value
	Native	Defect	Labral Repair	Dermal Augmentation	Bone Block Augmentation
Native	—	—	—	—	—
Defect	.005	—	—	—	—
Labral repair	.036	.970	—	—	—
Dermal augmentation	.933	.056	.240	—	—
Bone block augmentation	>.999	.005	.036	.933	—

NOTE. P < .05 represents a significant difference.

ANOVA, analysis of variance.

Distance to Dislocation

The distance to dislocation was collected and statistically compared with the native state as well as between repair groups (Fig 5). The distance to dislocation for the defect state was found to be 58% ± 12% of that for the native state. For labral repair, the distance to dislocation was found to be 68% ± 14% of that for the native state. The defect state and labral repair had a significantly decreased distance to dislocation when compared with the native state (P < .001 and P < .001, respectively). For DA and BBA repair, the distance to dislocation was found to be 89% ± 19% and 114% ± 20% of that for the native state, respectively. No statistical differences were found between DA and BBA repair when compared with the native state. BBA was found to have a statistically greater distance to dislocation when compared with the defect state (P < .001), labral repair (P < .001), and DA (P < .001). Additionally, the DA state had a significantly greater distance to dislocation than both the defect state (P < .001) and labral repair (P = .001). No other statistical differences were found between repair states. Table 2 shows all resulting P values.

Fig 5.

Distance-to-dislocation change from native state. Asterisks represent significant differences between repair states (P < .05). Daggers represent significant differences from the native state (P < .05). Plot whiskers indicate upper and lower data limits. Dashed line indicates 100% of native maximum load. Vertical lines and X within the quartile box indicates the median and mean respectively for each data set. The dots outside of the box represent data outliers.

Table 2.

Resulting P Values From 1-Way Repeated-Measures ANOVA for Distance to Dislocation

	Native	Defect	Labral Repair	Dermal Augmentation	Bone Block Augmentation
Native	—	—	—	—	—
Defect	<.001	—	—	—	—
Labral repair	<.001	.616	—	—	—
Dermal augmentation	.231	<.001	.001	—	—
Bone block augmentation	.063	<.001	<.001	<.001	—

NOTE. P < .05 represents a significant difference.

ANOVA, analysis of variance.

Force Versus Translation

Figure 6 displays the mean percentage of native forces along regular intervals from the center of the glenoid to the anteroinferior dislocation experienced during each testing state. The defect state and labral repair showed significantly lower values than the native state throughout translation, except at the 40% native translation position, where no significance was found. Although DA resulted in lower load values, there were no significant differences compared with the native state. No significant differences were found between the native state and BBA repair for the initial 40% of native translation. After 40% of native translation, all subsequent translation positions for BBA revealed statistically greater load values with respect to the native state. Appendix Table 2 shows the resulting P values.

Fig 6.

Percentage of average force versus percentage of native translation per testing phase relative to each sample’s native state. The boxes represent significant differences from the native state.

Discussion

This study shows adequate biomechanical stability of a labral repair using dermal allograft to address recurrent dislocation in the context of subcritical glenoid bone loss. The DA group showed no significant differences in maximum load or distance to dislocation when compared with the native state and showed non-significantly increased mean values for both measures when compared with SLR. The DA group also showed a load pattern most similar to the native state throughout translation. Additionally, the BBA group showed an increased maximum load; however, this was statistically nonsignificant. With respect to distance to dislocation, BBA showed significantly increased values compared with both DA and labral repair. The pattern of load throughout humeral translation was dissimilar to the native state.

The results of this study allow us to compare the biomechanical properties of the different techniques performed and to comment on the possible mechanisms leading to the advantages and disadvantages of each in the setting of SBL. Labral repair was included in this study because of its status as a standard soft-tissue procedure commonly performed after traumatic anterior dislocation to address shoulder instability.3, 4, 5^,19,20 Multiple studies have shown that labral repair in cases of SBL can result in recurrence rates as high as 100%.^4,6,7,21 This clinical observation may be explained by our findings that labral repair during translation resulted in loads that were not significantly different from those of the defect state but were statistically lower than those of the native state. This was the case for all data points at and beyond 50% of native glenohumeral translation where the bone defect was present, including the maximum forces at dislocation.

Although other factors such as the surrounding soft tissue including the capsule and muscles will contribute to more stability clinically, we observed that in our bony SBL model, SLR did not provide adequate resistance to dislocation nor did it re-create the ability of the humeral head to translate across the glenoid. Given that many other studies have shown similar trends for labral repair when repairing SBL,^4,6,7,21 we hypothesize that the reason behind the clinical shortcomings in addressing this pathology is that SLR does not correct the glenoid bone loss, resulting in a nonanatomic repair. Additionally, this may result in early repetitive strain and stretching of the capsule, which could increase the incidence of recurrence during daily activities.^4,6,7,21 This has led many authors to suggest that the area of bone loss needs to be restored to reduce recurrence.^3,21,22 This restoration allows the humerus to translate before engaging the surrounding soft tissue, as would occur in a healthy shoulder. Reducing the glenoid surface area in this pathology leads to rapid engagement of the repair, potential resulting in stretching of the capsular shift.

On the contrary, bone block procedures have been recommended in patients with SBL to reduce the rate of recurrent instability. Although these procedures show good outcomes in terms of recurrence rates, proper graft preparation and fixation are critical to achieve good results.²³ Small variations in graft placement translate to a high dispersion of data points, which was apparent in our data set, despite the procedure being executed in a completely open and isolated manner. Improper fixation of the anterior osseous profile could lead to osteoarthritis or impaired blood flow to the graft, leading to resorption.²³ Moreover, long-term retrospective studies have shown the development or progression of arthritis in 25% to 35% of patients older than 20 years.8, 9, 10 Although similar rates of arthritis (37% at 10 years) have also been reported with SLR,⁸ our study may indicate a possible mechanism of arthritis development specific to BBA. Bone block reconstruction resulted in a significantly greater distance to dislocation and a non-significantly greater force to dislocation, with load values spiking up to 200% greater than those of the native state at end-range translation. The results are consistent with the findings of a biomechanical study conducted by Willemsen et al.²⁴ that showed that BBA significantly increased the force required to translate the humerus 10 mm when compared with the native state in a 20% SBL model. The increased forces may be due to the presence of hard bone on the anterior glenoid rim (unlike the much softer dermal graft) and may explain the increased incidence of osteoarthritis seen with bone block procedures.

To address the shortcomings of the current techniques, dermal allograft repair was developed and tested in a side-by-side manner in an SBL pathology model. The dermal allograft provides a substrate to fill the area of SBL, allowing for a more anatomic labral repair. Because of the soft and moldable characteristics of the dermal allograft, this procedure may mitigate the technical difficulty of bone block procedures and reduce complications such as arthritis and resorption associated with bony grafts.²³ Biomechanical studies are warranted to evaluate DA for anterior shoulder instability in cases with subcritical glenoid bone loss. So far, biomechanical studies have shown that stand-alone labral repair should only be performed in cases with less than 15% glenoid bone loss; otherwise, adjunct procedures should be considered.^22,25 Other biomechanical studies have shown that BBA procedures are effective in restoring glenohumeral stability in cases with SBL.^26,27

Our study suggests that in a biomechanical model with 15% subcritical glenoid bone loss, the DA technique was able to most closely restore native values with no statistical differences in load across the full range of translation of the humerus on the glenoid (Fig 6). Although the defect state and labral repair showed significantly lower forces compared with the native state, BBA had significantly higher peak forces when approaching dislocation. Although these results indicate the high biomechanical stability of BBA compared with the other procedures, the high forces and narrower displacement range indicate a biomechanical load-displacement profile that is unlike the native state. Conversely, the DA group showed similar trends of force and displacement to the native state but showed lower mean force compared with the native state at all displacement points (Fig 6). As mentioned previously, the high forces associated with BBA may also describe the mechanism behind the high rates of arthritis described in the literature.8, 9, 10 Alternatively, these high arthritis rates may simply be a manifestation of studying a cohort at an inherently high risk of arthritis development.

In our 15% SBL model, the dermal allograft extended the anterior-inferior edge of the glenoid between the 4:30 clock-face and 6-o’clock regions, allowing for a larger “runway” for the humerus compared with labral repair and defect while showing a distance to dislocation similar to the native state. It also proved advantageous at maintaining stable load values throughout translation, which consistently restored native values. In contrast, the loads of the other tested groups either spiked or dropped throughout glenohumeral translation. On the basis of our model, the DA technique may provide the stability necessary to allow for proper translation of the humerus while also limiting the unnatural accumulation of high joint forces.

Many surgeons currently add remplissage to SLR in cases of SBL to reduce rates of recurrent instability.²⁸ The DA procedure described in this article can be performed along with remplissage should a surgeon desire to do so. One of the study authors (S.A.) does not routinely perform remplissage in these patients because the Hill-Sachs lesion that was previously off-track in patients with instability typically becomes on-track after the DA procedure. However, should a surgeon feel more comfortable adding remplissage, this can easily be performed simultaneously.

Limitations

This study presents several limitations. Owing to its cadaveric nature, the study did not permit in vivo healing and therefore simulated a time-zero analysis. Additionally, the study did not include any muscular tissue or capsule that contributes as an active stabilizer for the joint. The inclusion of loaded muscle tissue may increase the force values at dislocation. Another limitation of the study is the repeated-measures design in which the same specimen underwent repetitive testing, which may have caused labral or capsular damage. It is important to note that the low sample size inherent to many cadaveric studies resulted in a weakly powered statistical analysis and the risk of a type II error. This drawback is well illustrated in the comparisons involving BBA. Although the mean load to dislocation was objectively higher, the high variance likely led to statistical insignificance (Fig 4). The inverse effect is seen when comparing distance to dislocation as the lower variance likely led to statistical significance (Fig 5). Finally, this study only tested the osteoligamentous structures with static translation of the humeral head, which does not perfectly represent the complex kinematics and dynamic stability contributed by the native shoulder musculature in vivo. Further clinical and biomechanical studies are needed to address the long-term performance of the dermal allograft as a bone void filler in shoulder instability cases.

Conclusions

Dermal allograft repair showed similar biomechanical characteristics to the native state and showed an increased distance to dislocation and higher forces at dislocation compared with SLR. BBA restored the distance to dislocation and resulted in non-significantly increased maximum loads when compared with the native state.

Funding

This work was supported by Arthrex (grant No. AIRR-0141).

Appendix

Appendix Table 1.

Native Data Used to Calculate Percentages for Subsequent States

	Glenoid Width Center to Edge, mm	Force to Dislocation, N	Distance to Dislocation, mm
Sample 1	17.38	21.13	19.36
Sample 2	20.35	25.11	17.97
Sample 3	18.98	25.81	20.63
Sample 4	14.95	21.08	16.23
Sample 5	13.37	15.18	16.20
Sample 6	15.84	13.36	14.59
Sample 7	14.69	18.23	14.10
Sample 8	15.58	24.39	17.26

Appendix Table 2.

Resultant P Values From 2-Way ANOVA Comparing Force Values of Repair States With Native Values Along Each Interval of Translation

	P Value
	Defect	SLR	DA	BBA
10%	.025	.034	.556	.984
20%	.002	.002	.110	.059
30%	.016	.017	.102	.079
40%	.051	.070	.233	.075
50%	.013	.027	.179	.045
60%	<.001	.003	.091	.042
70%	<.001	<.001	.080	.048
80%	.011	.009	.018	.122
90%	.035	.023	.104	.011

ANOVA, analysis of variance; BBA, bone block augmentation; DA, dermal augmentation; SLR, standard labral repair.