AJSM PreView, published on April 20, 2012 as doi:10.

1177/0363546512441589

Bone Replacement of Fast-Absorbing Biocomposite Anchors in Arthroscopic Shoulder Labral Repairs

Matthew D. Milewski,* MD, David R. Diduch,y MD, Joseph M. Hart,y ATC, PhD, Marc Tompkins,z MD, Shen-Ying Ma,y MD, and Cree M. Gaskin,y|| MD Investigation performed at the University of Virginia, Charlottesville, Virginia
Background: Newer generation biocomposite anchors are hypothesized to resorb more reliably and faster, while allowing for bone ingrowth and replacement. Purpose: The purposes of this study were to (1) assess anchor resorption and bone ingrowth over time, (2) identify tunnel widening or potential reactions to the implants, (3) compare imaging findings for different sites of labral repair, and (4) determine patient subjective outcomes with the use of biocomposite anchors in glenoid labral repair. Study Design: Case series; Level of evidence, 4. Methods: We enrolled 22 patients to participate in a 24-month outcomes study that included subjective and objective outcome assessments after glenoid labrum repair surgery. Magnetic resonance imaging (MRI) was performed at 6 and 12 months to identify any potential reactions to implants. Computed tomography (CT) scans were performed at 12 and 24 months to determine anchor resorption and bone ingrowth. Sixteen patients and 47 anchors were available for follow-up at 24 months. An independent, fellowship-trained musculoskeletal radiologist read the scans. Subjective outcome scores measured at 24 months postoperatively included Simple Shoulder Test, Tegner activity scale, American Shoulder and Elbow Surgeons (ASES), and University of California, Los Angeles (UCLA) shoulder scores. Results: No adverse events were reported with the use of biocomposite anchors at the end of the study period. At 12 and 24 months, respectively, CT scans demonstrated that an estimated 68% and 98% of combined anchor material had been absorbed, 56% and 78% of the anchor material had been replaced by soft tissue of variable density, and 9% and 20% of total anchor volume was replaced by bone. No obvious mechanical failure of the labral repairs was detected on nonarthrogram MRI. Three of the 47 anchors showed bone cyst formation. Tunnel widening (expansion beyond tunnel diameter of 3 mm; 2.9-mm drill hole utilized) was seen in 55% of the anchors but decreased between 12 and 24 months, consistent with bone replacement. Tunnel widening was seen more in anteroinferior and posterior glenoid anchor locations (84% and 57%, respectively) than in superior labral anchors (13%). Subjective outcome scores at 24 months for ASES and UCLA shoulder scores averaged 88 and 30, respectively. All but one patient were satisfied with their outcome at 24 months. Conclusion: Our imaging evaluation indicates resorption of newer generation biocomposite anchors with progressive bone replacement at 12 and 24 months while maintaining acceptable subjective outcomes. Keywords: biocomposite; suture anchor; labral repair; shoulder; arthroscopy

Address correspondence to Cree M. Gaskin, MD, University of Virginia Health System, Department of Radiology and Medical Imaging, PO Box 800170, Charlottesville, VA 22908 (e-mail: cmg9s@virginia.edu). *Elite Sports Medicine / Connecticut Childrens Medical Center; Farmington, Connecticut. y Department of Orthopaedic Surgery, University of Virginia, Charlottesville, Virginia. z Department of Orthopaedics, University of Minnesota, Minneapolis, Minnesota. Department of Radiology and Medical Imaging, University of Virginia, Charlottesville, Virginia. One or more of the authors has declared the following potential conflict of interest or source of funding: Dr Gaskin is a consultant for DePuy Mitek. This study was funded, in part, by a research grant from DePuy Mitek. The American Journal of Sports Medicine, Vol. XX, No. X DOI: 10.1177/0363546512441589 2012 The Author(s)

||

The use of suture anchors for soft tissue to bone fixation has facilitated the movement from open surgery to arthroscopic techniques for a variety of glenohumeral conditions including Bankart and superior labrum anterior posterior (SLAP) lesion repairs.16,25,27 Success of these arthroscopic procedures depends on multiple factors including tissue quality, patient selection, and compliance with postoperative rehabilitation.26,38 Technique and technology-related factors including anchor and suture material, knot security, anchor geometry, and anchor biomechanics also play a role.11,12,25,37 The first-generation suture anchors were metallic devices that provided excellent fixation but were associated with failures including loosening and breakage that led to chondral damage.{ Metallic artifact also
{

References 4, 6, 10, 11, 13, 19, 30, 39.

Milewski et al

The American Journal of Sports Medicine

distorted postoperative imaging by computed tomography (CT) and magnetic resonance imaging (MRI) in revision scenarios.29 Bioabsorbable alternatives were therefore developed to help address some of the disadvantages of metallic anchors while providing appropriate soft tissue fixation.2,9,24 Speer and Warren31 described 4 criteria for bioabsorbable implants in the shoulder: (1) the bioabsorbable implant must have adequate initial fixation strength to coapt the soft tissues to bone, (2) the implants bioabsorption profile must enable it to retain satisfactory strength while the healing tissues are regaining mechanical integrity, (3) the implant must not bioabsorb too slowly or it will behave like its metal counterpart with breakage and migration, and (4) the implant must be made of materials that are completely safe: no toxicity, antigenicity, pyrogenicity, or carcinogenicity. One might consider adding a fifth criterion to Speer and Warrens criteria31 for bioabsorbable implants in the shoulder: bioabsorbable implant replacement by bone. Biocomposite anchors, which combine biodegradable copolymer and poly(lactic-co-glycolic acid) (PLGA) osteoconductive bioceramics such as b-tricalcium phosphate (TCP) or hydroxyapatite, were developed to help bridge the gap between adequate initial fixation and eventual bone replacement without osteolysis or synovitis.21 We sought to investigate one such biocomposite, Biocryl Rapide (BR) (DePuy Mitek, Raynham, Massachusetts), which is used in 2 different labral suture anchors, the Lupine BR anchor (DePuy Mitek) and the Bioknotless BR anchor (DePuy Mitek). We hypothesized that patients undergoing glenoid labral repair with these new biocomposite anchors would demonstrate radiographic evidence of labral healing with progressive anchor resorption and replacement by bone while maintaining good clinical outcomes based on subjective outcome scores and minimizing adverse reactions such as intra-articular inflammatory response and cyst formation.

Initial Enrollment 6 month MRI scans: n=19 Lost to follow-up n=2 12 month MRI scans: n=17 12 month CT scans: n=17 Lost to follow-up n=4 24 month CT scans: n=13 24 month Evaluations: n=13 Lost to follow-up n=0 24 month CT scans: n=3 24 month Evaluations: n=3

Supplemental Enrollment 12 month MRI n=3 12 month CT scans n=3

Total 6 month MRI scans: N=19 Total 12 month MRI scans: N=20 Total 12 month CT scans: N=20 Total 24 month CT scans: N=16 Total 24 month Evaluations: N=16 Serial assessments: Total 6+12+24 month scans on same patients: N=13 Total 6+12 month MRI: N=17 Total 12+24 month CT scans: N=16

Figure 1. CONSORT flowchart for study enrollment and follow-up.

MATERIALS AND METHODS

This longitudinal case series involving patients who had undergone shoulder arthroscopy and labral repair by 3 senior surgeons at our academic sports medicine center was approved by our institutional review board. Inclusion criteria were patients between 18 and 60 years of age who underwent arthroscopic capsulolabral repair using anchors composed of this particular biocomposite and who were willing to undergo postoperative MRI and CT at the established time points. Anchors used included the Lupine BR anchor and the Bioknotless BR anchor. Both anchors were included in this study, focusing on anchor bioabsorption and bioreplacement, because they have identical body types by both geometry and composition. They both have suture loops and differ only in that one is knotless and the other allows suture tying of a suture attached to the suture loop. Biocryl Rapide is made of 70% PLGA and 30% TCP. Exclusion criteria were previous shoulder surgery; chronic or acute medical illness that may interfere with healing, such as a neurological, collagenous, or circulatory disease; existing

pathological conditions or degenerative joint diseases including rheumatoid arthritis, extensive osteoarthritis, or other progressive collagen or bone conditions that would hinder adequate anatomic visualization and prevent secure tissue fixation; and women who were pregnant or lactating or planned to become pregnant during the study. Sixteen patients and 47 anchors were available for follow-up at 24 months. Enrollment and follow-up flow are shown in Figure 1. Computed tomography was performed without intravenous or intra-articular contrast at 12 6 1 months and 24 6 1 months utilizing multidetector CT scanners from GE Medical Systems (Lightspeed 16, Lightspeed Pro 32, or Lightspeed VCT 64, Little Chalfont, United Kingdom). Patients were scanned with a 0.625-mm detector configuration, with bone and soft tissue algorithms utilizing 140 kV and automatic mAs. Images were constructed prospectively at 0.625 mm with 50% slice overlap. These direct axial images were reviewed together with multiplanar reformats at the radiologists workstation. Magnetic resonance imaging was performed without intravenous or intra-articular contrast at 6 6 1 months and 12 6 1 months utilizing a 1.5-T scanner from Siemens Medical Systems (Avanto, Erlangen, Germany). A dedicated shoulder coil was utilized. Scan protocol included axial, oblique coronal, and oblique sagittal conventional spin echo T1 (repetition time [TR]/echo time [TE] = 625/16 msec); axial and oblique sagittal turbo inversion recovery (TR/TE = 6350/25; inversion time = 160 msec); and oblique coronal dual echo fast spin echo T2 and proton density with fat saturation (TR/TE = 4130/102 and 4130/37). A matrix of 512 3 512 was utilized. Radiographic outcomes were based on postoperative MRI scans obtained at 6 and 12 months to identify any potential reactions to the implants and CT scans obtained

Vol. XX, No. X, XXXX

Biocomposite Anchors in Arthroscopic Shoulder Labral Repairs

at 12 and 24 months to determine anchor resorption and bone ingrowth. An independent and fellowship-trained musculoskeletal radiologist with over 10 years of experience (C.M.G.) reviewed all scans. Nonarthrogram MRI scans were examined for adequacy and integrity of labral repair and to assess in vivo biological safety based on the presence or absence of undesirable host responses including joint effusion, synovitis, solid mass from foreign body reaction, lymphadenopathy, edema, and/or cyst formation. The CT scans were examined to determine anchor resorption, replacement by bone or soft tissue density, intraosseous cyst formation, and anchor tunnel widening.

Data Analysis
A picture archiving and communication system (PACS) was used for digital image review. Digital tools imbedded in our PACS (version 10.2, Carestream, Rochester, New York) were used to standardize measurements across patients and scan dates. Specifically, we measured tunnel diameter, cyst size (if present), and density of the tissue replacing the anchors to support the determination of bone versus soft tissue bioreplacement at each time point. Because of the small size and complicated geometry of the anchors, precise assessment or quantification of percentage of anchor resorption was not possible. The radiologist estimated the percentage of resorption by comparing the anchors across examinations over 24 months and by referencing the appearance of new (not yet absorbed) anchors. A similar process was used for estimating quantity of anchor replacement by bone or soft tissue density. Precise measurements were again not possible, so the radiologist estimated the portion of the anchor replaced. Although this technique of estimation has limitations, it was considered satisfactory to capture a general trend of progressive resorption. To determine whether the tissue replacing the anchor was bone or soft tissue, a semiquantitative or partially subjective approach was employed. For the tissue replacing the anchor to be considered as bone, it required 2 criteria to be met: (1) the radiologists subjective interpretation of the density and texture of the tissue needed to be consistent with bone, and (2) the tissue density as objectively measured must be similar to that of bone in Hounsfield units (HU). The use of HU measurements to confirm bone replacement of bioabsorbable implants has been reported previously with interference screws.7,8 The HU scale is based on linear attenuation coefficients and is a scale of relative radiodensity where, by definition, the radiodensity of distilled water under standard pressure and temperature is 0 HU and the radiodensity of air is 1000 HU. Typical HU measurements from the human body are as follows: 120 for fat, 0 to 10 for simple fluid, 40 for muscle, over 150 for light trabecular bone, and 1000 for cortical bone. The HU measurements were taken from the tissue replacing the absorbed anchors, and these measurements were considered additional support for the subjective determination of whether a resorbed anchor had been replaced by bone or soft tissue. The MRI scans were reviewed for labral repair integrity, with an intact repair defined as the experienced musculoskeletal radiologists subjective interpretation along with objective

criteria of labral tissue found in situ without linear increased signal undercutting or coursing through it. Repair failure or labral retear was defined as displaced labral tissue or in situ labral tissue with linear high signal undercutting or coursing through it. Indeterminate labrum was defined as not being able to fit the definitions of intact or torn mentioned above. Reasons for lack of clarity included patient motion and postoperative related artifacts such as subtle signal changes that were ambiguous for subtle retear versus normal postoperative changes of repaired tear. The authors emphasize, however, that the primary imaging objective was focused on the biological behavior of the implant with regard to bioresorption, host response, and bioreplacement. Tunnel widening was defined as 3 mm or greater in maximal cross-sectional diameter obtained orthogonal to the long axis of the tunnel. This definition was based on the geometry or size of the anchors and the pilot hole. The anchors were 3 mm wide over much of their length but varied from 2 mm (at deep tip) to 4 mm (near joint surface) in width. The drill holes had a diameter of 2.9 mm for both types of BR anchors as they had identical body types. The radiologist was not specifically blinded to the time interval between surgery and scanning. The studies were read in 4 batches at different time points in the study, and generally, the 6- and 12-month scans appeared in the first and second groups, and 24-month scans appeared in the third and fourth groups. Follow-up clinical data included history and physical examination as well as patient demographics. Subjective outcome measures were recorded at 24 months postoperatively, which included Tegner activity scale, Simple Shoulder Test, American Shoulder and Elbow Surgeons (ASES), and University of California, Los Angeles (UCLA) shoulder scores.

Statistical Analyses
Percentage of resorption and percentage of replacement by bone were compared among the 3 anchor locations (anterior, posterior, and superior) at the 12- and 24-month time points with a univariate analysis of variance (ANOVA). Tukey least significant difference was used for post hoc analyses. A test was considered statistically significant if the P value was .05 or less. The data involving the percentage of anchor resorption, percentage of anchor replaced by bone, and the comparisons of these variables between anchors of different locations were not normally distributed. Kruskal-Wallis nonparametric analysis was also completed and reported along with median values. Both parametric and nonparametric analyses including mean and median values were reported as the parametric analysis and mean values may be easier to interpret both scientifically and clinically.

RESULTS
A total of 22 patients were enrolled in our study, with 16 patients completing the clinical assessment, subjective outcomes, and follow-up imaging at 24 months. Enrollment

Milewski et al

The American Journal of Sports Medicine

Figure 2. Six-month postoperative axial inversion recovery magnetic resonance imaging (MRI) scan (A) and axial T1 MRI scan (B) of a posterior labral repair utilizing a Lupine BR anchor (DePuy Mitek, Raynham, Massachusetts). Linear signal changes are seen at the site of the anchor (long, thin white arrows) without adjacent bone or soft tissue changes. Twelve-month postoperative axial inversion recovery MRI scan (C) and axial T1 MRI scan (D) show the anchor and tunnel (short, thick white arrows) to be less perceptible over time. Axial and oblique sagittal reformat computed tomography images at 12 months (long, thin black arrows) and 24 months (short, thick black arrows) postoperatively (E-H) show progressive bone replacement and less obvious tunnel visualization in the midposterior labrum. and follow-up flow are delineated in Figure 1. There was 73% patient follow-up at 24 months, with 6 patients having withdrawn from the study or lost to follow-up. Three patients were enrolled in a period of additional enrollment after 2 patients were lost during early enrollment after their 6-month follow-up scans. Four patients completed MRI and CT scans at 12 months but were lost to followup before 24 months. There were 11 male and 5 female patients in our cohort, with a mean age of 35 years and an age range of 18 to 60 years. A total of 48 BR anchors, all composed of the same material, were utilized in these 16 patients labral repairs. One of these anchors could not be evaluated on 24-month CT as it was obscured by revision surgery that occurred between the 12- and 24month scans, leaving 47 anchors for 24-month analysis by imaging. A total of 32 Bioknotless BR anchors and 16 Lupine BR anchors were utilized with an average of 3.0 suture anchors used per patient. Anchor type and number utilized per patient were based on the surgeons preference. Of the 47 anchors available for 24-month analysis, 25 anchors were placed for anteroinferior labral repairs (19 Bioknotless BR and 6 Lupine BR anchors), 15 anchors were placed for superior labral repairs (8 Bioknotless BR and 7 Lupine BR anchors), and 7 anchors were placed for posterior labral repairs (4 Bioknotless BR and 3 Lupine BR anchors). Thirteen of 16 patients (81%) underwent their MRI examinations at 6 months, and 15 of 16 patients (94%) completed their MRI examinations at 12 months. All 16 patients completed their CT scans at 12 months and 24 months along with their clinical examinations at 24 months. The MRI scans performed at 12 months confirmed all 15 patients had no signs of effusion, synovitis, lymphadenopathy, intra-articular masses, or soft tissue edema (Figure 2). When assessed for labral integrity, 11 of 15 patients labral repairs were deemed intact at the 12-month MRI examination. Three of the 15 labral repairs were deemed indeterminate at the 12-month MRI examination. One of the 15 labral repairs was deemed indeterminate at 6 months and was shown subsequently to be retorn intraoperatively at 8 months. The CT scan assessments at 12 and 24 months were used to assess resorption of BR anchor material (Figures 3 and 4). At 12-month follow-up, an estimated average of 68% 6 25% resorption of BR anchor material was observed across the 47 anchors in 16 patients (median value of 70% resorption at 12 months). At 24-month follow-up, an estimated average of 98% 6 10% resorption of BR anchor material was observed, with 45 of 47 anchors having 100% resorption (median value of 100% resorption at 24 months). In terms of bone ingrowth, an estimated average of 9% of total BR anchor material across all anchors was replaced by bone at 12-month follow-up (median value of 0%) as determined from CT (mean, 423 HU; range, 250-800 HU). Eleven of the 47 BR anchors (23%) exhibited replacement by bone, and within this group of 11 anchors, an estimated average of 39% 6 20% of BR anchor material was replaced by bone. At 24 months, total bone replacement of anchors had increased to an estimated average of 20% of the BR

Vol. XX, No. X, XXXX

Biocomposite Anchors in Arthroscopic Shoulder Labral Repairs

Figure 3. Twelve- and 24-month postoperative axial computed tomography images (A, B) of a superior labral repair utilizing 2 Bioknotless BR anchors (DePuy Mitek, Raynham, Massachusetts). Both anchor sites show interval anchor resorption, tunnel width narrowing, and replacement by bone between 12 months (black arrows) and 24 months (white arrows). anchor material (median value of 0%). Twenty-two of the 47 BR anchors (47%) exhibited some replacement by bone at 24 months, and within this group of 22 anchors, bone replacement averaged 42% 6 26%. At 12-month CT scan evaluation, an estimated 56% of BR anchor material in the 48 anchors was replaced by soft tissue density (mean, 65 HU; range, 25-120 HU), and this increased to 78% of BR anchor material in the 47 anchors at 24-month follow-up. The progression of percentage of anchor resorption and percentage of bone replacement for anteriorly, posteriorly, and superiorly placed anchors from 12-month and 24-month CT scan evaluations are shown in Figure 5. Our primary imaging focus was on overall anchor behavior, and thus, we did not specifically power the study to perform subgroup analysis by anchor location or patient age; however, we did perform such analyses to contribute to our observations. Percentage of resorption was significantly different among anchor locations at 12 months but not at 24 months after implantation (Figure 6). Specifically, at 12 months, the posterior anchors had a significantly higher percentage of resorption than the superior (P = .03) and anterior (P = .01) anchors. The percentage of resorption was not different between the anterior and superior anchors (P = .72). Percentage of replacement by bone was also significantly different among the anchor locations at both 12 months and 24 months after implantation (Figure 6). The average percentage of replacement by bone was highest for the superior anchors at 12 and 24 months. Percentage of replacement by bone was significantly higher in the superior anchors compared with the anterior anchors at 12 months (P = .002) and 24 months (P = .01) after implantation but not significantly higher than the posterior anchors. Kruskal-Wallis nonparametric analysis revealed a significantly greater percentage of resorption of posterior anchors at 12 months compared with anterior anchors (P = .04) and a significantly greater percentage of replacement by bone of superior anchors at 12 months compared with anterior

Figure 4. Twelve-month postoperative axial (A) and oblique sagittal reformat (B) computed tomography (CT) images of an anteroinferior labral repair utilizing Bioknotless BR anchors (DePuy Mitek, Raynham, Massachusetts). An enlarged cystic cavity in the anterior glenoid is visualized surrounding a vaguely linear calcific density thought to reflect partially resorbed anchor material (black arrows). The oblique sagittal reformat CT image (B) also shows a portion of a smaller tunnel from another anchor (black arrowhead). Twenty-four-month postoperative axial (C) and oblique sagittal reformat (D) CT images show an interval decrease in the size of the cavity (white arrows) with complete resorption of the anchor and partial bone replacement. The oblique sagittal reformat image (D) shows a more modest interval decrease in size of the tunnel from the other anchor (white arrowhead). anchors (P = .01). Kruskal-Wallis nonparametric analysis did not reveal significant differences in percentage of anchor resorption and percentage of replacement by bone between anchor locations at 24 months (P = .41 and P = .07, respectively). Correlations were done to compare age and percentage of resorption and percentage of replacement by bone. At 12 months, age was related to percentage of resorption (r = .30, P = .04) and percentage of replacement by bone (r = .04, P = .005), but these correlations were not significant at 24 months (percentage of resorption: r = .15, P = .31; percentage of replacement by bone: r = .19, P = .19). Bone cysts that communicated with the BR anchor tunnel were associated with 3 of the 47 BR anchors (6%). All 3 cysts were seen in anchors used for superior labral repair (2 Bioknotless BR and 1 Lupine BR anchors). One was a small cyst (\5 mm in diameter) at the deep aspect of the tunnel that developed between 12- and 24-month follow-up CT scans even though the associated tunnel decreased in size during that time period (Figure 4). The second cyst was a 5-mm cyst present at 12-month followup that enlarged to 8 mm at 24-month follow-up. This

Milewski et al

The American Journal of Sports Medicine

Figure 5. Top row graphs showing percentage of anchor resorption between 12 and 24 months (12-24M % Resorption) by computed tomography (CT) scan evaluations for each superior, posterior, and anterior anchor. Bottom row graphs showing percentage of anchor replacement by bone between 12 and 24 months (12-24M % Bone) by CT scan evaluations for each superior, posterior, and anterior anchor. patient also had some degree of glenohumeral osteoarthritis (although still within inclusion criteria), which was a confounding factor. The third cyst was 4 mm in size and located at the deep aspect of the tunnel. It was present on the 12-month CT scan, but it remained stable at 24month follow-up. Tunnel widening, as defined as maximal tunnel width greater than 3.0 mm, was associated with 26 of the 47 BR anchors (55%). The average diameter of the widened tunnels was 4.4 mm at 12 months (range, 3.25-6.0 mm), and this decreased to an average of 3.9 mm at 24 months (range, 3.0-5.0 mm). Widening affected 2 of the 15 superior labral BR anchors (13%) (1 Bioknotless BR and 1 Lupine BR anchor). Widening affected 4 of the 7 posterior labral BR anchors (57%) (1 Bioknotless BR and 3 Lupine BR anchors). Widening affected 21 of the 25 anteroinferior labral BR anchors (84%) (14 Bioknotless BR and 7 Lupine BR anchors). Subjective outcome and clinical assessment data were collected on 16 patients at a minimum of 24-month followup. Clinical data were not available on the 6 patients lost to follow-up because this was collected only at 24-month follow-up. All but one patient were satisfied with their outcome at 24 months (15/16 patients, 94%). One patient sustained a repeat dislocation event at 8 months postoperatively and required a revision surgery. This patient was satisfied with their outcome at most recent follow-up. The ASES scores at most recent follow-up averaged 88 6 20. The UCLA shoulder scores at most recent follow-up averaged 30 6 6.7. Visual analog scale scores at most recent follow-up averaged 1.5 6 1.8. Simple Shoulder Test scores at most recent follow-up averaged 10 6 3.3 of a possible 12. Tegner activity scale scores preoperatively averaged 6.4 6 2.5 and postoperatively averaged 5.0 6 2.8. Our primary focus was on the patients who completed the 24-month clinical and imaging follow-up. However, we did evaluate the 6- and 12-month imaging of those patients who were lost to follow-up. For the sake of complete reporting, we are providing a brief summary of data from this group. These data are presented separately from the data of the patients who completed the study because patients dropped out at varying time points.

Vol. XX, No. X, XXXX

Biocomposite Anchors in Arthroscopic Shoulder Labral Repairs

Figure 6. A, Percentage of anchor resorption by anchor location (superior, posterior, and anterior). Percentage of resorption was highest in the posterior anchors at 12 months after implantation. B, Percentage of replacement by bone was highest for the superior anchors at 12 and 24 months after implantation. Overall, this small group demonstrated similar results to those who completed the study. The 6 patients lost to follow-up had 19 anchors placed. One of these 6 patients could not be evaluated by MRI because of artifact from a nearby metal foreign body. Fifteen anchors in 5 patients were evaluated by MRI at 6 months, and 10 anchors in 3 patients were evaluated by MRI at 12 months. The findings from the studies were negative for joint effusion, edema, lymphadenopathy, soft tissue mass, or fluid collection. Only one anchor of 15 (7%) was associated with cyst formation, and this 4-mm intraosseous cyst was stable between 6 and 12 months. The repairs appeared intact in 3 of 5 patients. The integrity of repair was uncertain in 2 patients, although there was no overt evidence of failure or retear in this group. Fourteen anchors in 4 patients were evaluated by CT at 12 months. Six of 14 anchors (43%) showed no evidence of tunnel widening, while 8 of 14 anchors (57%) were associated with tunnel widening (mean, 4.3 mm; range, 3.6-4.6 mm). At 12-month CT, an overall estimated 84% of anchor material was resorbed and replaced by bone (9% of anchor material) and soft tissue (75% of anchor material). surgery. These anchors offer some distinct advantages from their metallic counterparts, particularly in cases of failed tissue repairs. During revision cases, the need to remove permanent implants is also avoided with biodegradable implants because they can often be drilled out or through. Bioabsorbable anchors also improve the quality of postoperative imaging relative to metallic anchors. Bioabsorbable anchors cause less beam-hardening artifact on CT scans, and more importantly, they minimize distortion on postoperative MRI, which is often the study of choice to evaluate the status of the repair. In spite of these advantages, earlier generation bioabsorbable anchors have generated causes for concern. Multiple different types and combinations of polymers have been developed and utilized for anchors including polyglycolic acid (PGA), poly-L-lactic acid (PLLA), and poly-D-L-lactic acid copolymer polyglycolic acid (PDLLA-co-PGA). Early bioabsorbable anchors and fixation devices such as the Suretac (Smith & Nephew, Andover, Massachusetts) were made of a synthetic copolymer of 67.5% polyglyconic acid and 32.5% trimethylene carbonate.31 However, the degradation time reported with the PGA anchors was only 3 to 4 months, and they were associated with early loss of fixation, osteolysis, loose body formation, and glenohumeral synovitis.10,31,32 The PLLA anchors were developed to have a longer degradation time, between 10 and 30 months, but concerns were raised as to whether this was an excessively

DISCUSSION
Bioabsorbable suture anchors are rapidly evolving and supplanting metallic anchors in arthroscopic shoulder

Milewski et al

The American Journal of Sports Medicine

long degradation time, which may result in similar problems as metallic anchors with chondral damage from loosening as well as the fact that the anchors might not be replaced by bone.4,6,22,23,36 The PLLA tack anchors have been shown to have more radiographic evidence of persistent drill holes at 7-year follow-up than PGA tack anchors.14 Also, there have been concerns raised over intra-articular inflammatory response and cyst formation with PLLA anchors.21,22,29,31 In one study, 55% of patients who had anterior labral repairs with 3 polylactic acid (PLA) anchors (Panalok, Mitek Ethicon, Somerville, New Jersey) showed anchor tunnel enlargement greater than the mean plus one standard deviation.34 The literature has also documented some unique complications associated with use of bioabsorbable anchors. Several case reports have identified a hyperinflammatory response to bioabsorbable anchors, leading to synovitis, bone osteolysis, cartilage damage, and adverse clinical outcomes.# These complications have been associated with both PGA and PLLA implants alike, contributing to the impetus to develop better implant materials. Biocomposite anchors are composed of newer generation materials with potential for improved performance over earlier generation anchors. Biocomposites include a mixture of biodegradable material, such as PLGA or PLA, and osteoconductive bioceramics, such as TCP. Biocryl Rapide (DePuy Mitek) is a composite between fastabsorbing PLGA (70%) and osteoconductive TCP (30%). The proposed advantages of the new material include more timely and reliable resorption of the anchor with subsequent replacement by bone at the site of the anchor. Anchor resorption with bone replacement was hypothesized to occur by 24 months. While cadaveric experiments have demonstrated similar load-to-failure properties between biodegradable anchors and more traditional anchor materials, concern is still raised as to whether similar results can be found in vivo.5 While suture anchors made of newer generation biocomposite materials such as Biocryl Rapide have found increasing utility in shoulder arthroscopy, concern remains about the long-term effects these materials have in the glenohumeral joint including the effects that material resorption may have on repair integrity and cyst formation and whether the material or its resorption may induce an inflammatory response within the glenohumeral joint that could lead to synovitis. Also, there are little data in the literature regarding the potential for bone replacement or remodeling as the biocomposite material is resorbed. We believe this is the first clinical study with 24-month follow-up that radiographically demonstrates the extent of bone replacement of biocomposite labral anchors. Barber and Dockery7 have previously shown bone replacement of TCP PLLA interference screws. On 24-month follow-up with CT scans, our anchors appear to be largely resorbed with some progressive replacement by bone. The replacement of the rest of the anchor material by tissue not constituting bone was called soft tissue density because it

References 1, 3, 15, 17, 20, 23, 25, 28, 33-35.

matched the density of noncalcified, nonfatty soft tissue both subjectively and by HU density measurements but was otherwise nonspecific. Without tissue samples, the authors do not know whether this soft tissue density material represents immature osteoid, giant cell reaction, fibrous tissue, or granulation tissue. Further long-term studies will be needed to determine whether this material continues to ossify. Tunnel widening was seen in 55% of the anchors, but this actually decreased between the 12- and 24-month follow-up CT scans. The tunnel widening was not associated with clinical failures in this group and, as mentioned, tunnel widening decreased as the anchors were resorbed and partially replaced by bone. Tunnel widening is likely multifactorial and may relate to implant material, implant geometry and biomechanics, surgical technique, and glenoid location. The anchors used in this study were composed of Biocryl Rapide but were also of a specific design and geometry. Both anchors (Lupine BR anchor [DePuy Mitek] and Bioknotless BR anchor [DePuy Mitek]) used in this study have similar anchor body geometry, with the Lupine BR anchor having a suture attached to its suture loop for tying. These anchors can be described as a toggle type of anchor and are triangular in shape, relying in part on subcortical fixation with potential for space around the distal aspect of the anchor within the drill hole. We hypothesize that micromotion around the anchor could occur during shoulder motion and result in pumping of joint fluid in and around the anchor. Thus, anchor design may also contribute to tunnel widening and cyst formation. Changes in subsequent anchor designs and geometry, such as with a cylindrical anchor that completely fills the bone hole, may address these concerns. Tunnel widening was seen more in anteroinferior and posterior glenoid anchor locations (84% and 57%, respectively) but was only seen in 13% of the superior labral anchors. Identical anchors examined at identical time points, while only differing in anchor position and having different widening patterns, would support the idea that tunnel widening is likely related to the forces across these anchors in situ at different areas of the labrum. Cyst formation reported with earlier generation bioabsorbable anchors has generated concern. We documented 3 cysts in our study or 6% of all anchors used. All 3 cysts were seen with anchors used for superior labral repairs (2 Bioknotless BR anchors and 1 Lupine BR anchor). None of these 3 cysts were associated with anchor failure, but rather they were associated with excellent clinical outcomes. While cyst formation with these new biocomposite anchors remains a concern, there have been few investigational studies documenting the clinical results of glenohumeral surgery utilizing this new generation of material.7 Previous reports of cyst formation with PGA and PLLA were usually associated with bioabsorbable screws used for interference screw fixation in anterior cruciate ligament reconstruction.18 Given the paucity of data within the literature on biocomposite anchors, concern therefore remains about the long-term performance of these materials in the glenohumeral joint including the effects of material resorption on repair integrity and cyst

Vol. XX, No. X, XXXX

Biocomposite Anchors in Arthroscopic Shoulder Labral Repairs

formation and whether the material or its resorption may induce an inflammatory response within the glenohumeral joint. The strengths of this prospective study include the minimum 24-month follow-up including both CT and MRI evaluations. In addition to the imaging findings, the BR anchors were associated with good clinical outcomes, as shown by ASES, UCLA shoulder, and Simple Shoulder Test scores, with only one unsatisfied patient and one repeat dislocation. We also had a broad age range of patients and anchor location. We believe that a diverse patient population including age range and anchor location better characterizes the overall resorption profile of the implant rather than a narrow patient population and location. This allows us to make more broad and generalized conclusions. Weaknesses of this study include the lack of a control group by which to compare these anchors directly. A control group would have required the use of another type of anchor, which would have likely meant differences in both anchor composition along with anchor geometry and suture type. Subjective data were obtained in follow-up but were not obtained preoperatively, which can limit the conclusions drawn from the study in terms of subjective improvement. We were not able to precisely quantify volume of anchor resorption or replacement because of small anchor size and complex geometry, so we sought estimations, which introduced some intrinsic bias. In addition, we had a single fellowship-trained musculoskeletal radiologist read all of the scans, but he was not blinded to the time points. While a single reader eliminates some variability, it does introduce potential bias especially when he was not blinded to the time points of follow-up. While we used a fairly wide age range of patients, we found that age was not correlated at 24 months with anchor resorption or bone replacement. We had a 9% lost to follow-up rate at 12 months and a 27% lost to follow-up rate at 24 months. The cause behind these rates is likely multifactorial. This study was carried out at an academic institution that is a large tertiary referral center attracting patients from several hours away. Because of logistical issues, many of these patients were unable to return for each follow-up time point. In addition, many of these represent mobile patients, such as university students, who were no longer in the area at the time of follow-up. Lastly, during our study, there was heavy public and media scrutiny on radiation risk associated with CT, which may have affected patients willingness to return for CT scans at 12 and 24 months after surgery. In conclusion, this study documents good radiographic and subjective outcomes following labral repairs using a biocomposite anchor. We believe this study is the first to document bone replacement of absorbed anchor material and the first to determine the progression of absorption of new-generation biocomposite anchors in humans. Complications related to biocomposite anchor resorption, including cyst formation and tunnel widening, were mild and not clinically important in this study. Cyst formation was associated with only a small percentage of anchors, and

the cysts were small in size. Tunnel widening was common, although mild, and improved during the time course of radiographic follow-up as the anchors were resorbed and partially replaced by bone. There was no evidence of synovitis or intra-articular mass formation with absorption of these anchors. Good clinical outcomes can be expected using these new biocomposite faster absorbing suture anchors. The residual soft tissue density at the anchor sites at 2 years may contain nonmineralized osteoid matrix, so there is theoretical potential for further bone formation in time. Longer term CT follow-up will be needed to examine this potential for further progression of bone replacement of the resorbed anchors beyond 2 years.

ACKNOWLEDGMENT
The authors thank Christopher M. Kuenze, MA, ATC, for his assistance in patient recruitment and data collection.

REFERENCES
1. Athwal GS, Shridharani SM, ODriscoll SW. Osteolysis and arthropathy of the shoulder after use of bioabsorbable knotless suture anchors: a report of four cases. J Bone Joint Surg Am. 2006;88(8): 1840-1845. 2. Balch OK, Collier MA, DeBault LE, Johnson LL. Bioabsorbable suture anchor (co-polymer 85/15 D,L lactide/glycolide) implanted in bone: correlation of physical/mechanical properties, magnetic resonance imaging, and histological response. Arthroscopy. 1999;15(7):691708. 3. Barber FA. Biodegradable shoulder anchors have unique modes of failure. Arthroscopy. 2007;23(3):316-320. 4. Barber FA. Poly-D,L-lactide interference screws for anterior cruciate ligament reconstruction. Arthroscopy. 2005;21(7):804-808. 5. Barber FA, Coons DA, Ruiz-Suarez M. Cyclic load testing and ultimate failure strength of biodegradable glenoid anchors. Arthroscopy. 2008;24(2):224-228. 6. Barber FA, Deck MA. The in vivo histology of an absorbable suture anchor: a preliminary report. Arthroscopy. 1995;11(1):77-81. 7. Barber FA, Dockery WD. Long-term absorption of beta-tricalcium phosphate poly-L-lactic acid interference screws. Arthroscopy. 2008;24(4):441-447. 8. Barber FA, Dockery WD. Long-term absorption of poly-L-lactic acid interference screws. Arthroscopy. 2006;22(8):820-826. 9. Barber FA, Snyder SJ, Abrams JS, Fanelli GC, Savoie FH 3rd. Arthroscopic Bankart reconstruction with a bioabsorbable anchor. J Shoulder Elbow Surg. 2003;12(6):535-538. 10. Burkart A, Imhoff AB, Roscher E. Foreign-body reaction to the bioabsorbable Suretac device. Arthroscopy. 2000;16(1):91-95. 11. Cole BJ, Provencher MT. Safety profile of bioabsorbable shoulder anchors. Arthroscopy. 2007;23(8):912-913, author reply 913-914. 12. Diduch DR, Tadje JP, Ferguson RE Jr, Edlich RF. Modern concepts in arthroscopic Bankart repair. J Long Term Eff Med Implants. 1999;9(4):377-393. 13. Ekelund A. Cartilage injuries in the shoulder joint caused by migration of suture anchors or mini screw. J Shoulder Elbow Surg. 1998; 7(5):537-539. 14. Elmlund AO, Kartus J, Rostgard-Christensen L, Sernert N, Magnusson L, Ejerhed L. A 7-year prospective, randomized, clinical, and radiographic study after arthroscopic Bankart reconstruction using 2 different types of absorbable tack. Am J Sports Med. 2009;37(10):1930-1937.

10 Milewski et al

The American Journal of Sports Medicine

15. Freehill MQ, Harms DJ, Huber SM, Atlihan D, Buss DD. Poly-L-lactic acid tack synovitis after arthroscopic stabilization of the shoulder. Am J Sports Med. 2003;31(5):643-647. 16. Friel NA, Karas V, Slabaugh MA, Cole BJ. Outcomes of type II superior labrum, anterior to posterior (SLAP) repair: prospective evaluation at a minimum two-year follow-up. J Shoulder Elbow Surg. 2010;19(6):859-867. 17. Glueck D, Wilson TC, Johnson DL. Extensive osteolysis after rotator cuff repair with a bioabsorbable suture anchor: a case report. Am J Sports Med. 2005;33(5):742-744. 18. Gonzalez-Lomas G, Cassilly RT, Remotti F, Levine WN. Is the etiology of pretibial cyst formation after absorbable interference screw use related to a foreign body reaction? Clin Orthop Relat Res. 2011;469(4):1082-1088. 19. Jeong JH, Shin SJ. Arthroscopic removal of proud metallic suture anchors after Bankart repair. Arch Orthop Trauma Surg. 2009; 129(8):1109-1115. 20. Kelly JD 2nd. Disintegration of an absorbable rotator cuff anchor six weeks after implantation. Arthroscopy. 2005;21(4):495-497. 21. LeGeros RZ. Properties of osteoconductive biomaterials: calcium phosphates. Clin Orthop Relat Res. 2002;395:81-98. 22. McGuire DA, Barber FA, Elrod BF, Paulos LE. Bioabsorbable interference screws for graft fixation in anterior cruciate ligament reconstruction. Arthroscopy. 1999;15(5):463-473. 23. Muller M, Kaab MJ, Villiger C, Holzach P. Osteolysis after open shoulder stabilization using a new bio-resorbable bone anchor: a prospective, non-randomized clinical trial. Injury. 2002;33 Suppl 2: B30-B36. 24. Nho SJ, Provencher MT, Seroyer ST, Romeo AA. Bioabsorbable anchors in glenohumeral shoulder surgery. Arthroscopy. 2009;25(7): 788-793. 25. Park MJ, Hsu JE, Harper C, Sennett BJ, Huffman GR. Poly-L/D-lactic acid anchors are associated with reoperation and failure of SLAP repairs. Arthroscopy. 2011;27(10):1335-1340. 26. Poukalova M, Yakacki CM, Guldberg RE, et al. Pullout strength of suture anchors: effect of mechanical properties of trabecular bone. J Biomech. 2010;43(6):1138-1145.

27. Rhee YG, Lee DH, Lim CT. Unstable isolated SLAP lesion: clinical presentation and outcome of arthroscopic fixation. Arthroscopy. 2005;21(9):1099. 28. Sassmannshausen G, Sukay M, Mair SD. Broken or dislodged polyL-lactic acid bioabsorbable tacks in patients after SLAP lesion surgery. Arthroscopy. 2006;22(6):615-619. 29. Shellock FG, Mink JH, Curtin S, Friedman MJ. MR imaging and metallic implants for anterior cruciate ligament reconstruction: assessment of ferromagnetism and artifact. J Magn Reson Imaging. 1992;2(2):225-228. 30. Silver MD, Daigneault JP. Symptomatic interarticular migration of glenoid suture anchors. Arthroscopy. 2000;16(1):102-105. 31. Speer KP, Warren RF. Arthroscopic shoulder stabilization: a role for biodegradable materials. Clin Orthop Relat Res. 1993;291:67-74. 32. Speer KP, Warren RF, Pagnani M, Warner JJ. An arthroscopic technique for anterior stabilization of the shoulder with a bioabsorbable tack. J Bone Joint Surg Am. 1996;78(12):1801-1807. 33. Spoliti M. Glenoid osteolysis after arthroscopic labrum repair with a bioabsorbable suture anchor. Acta Orthop Belg. 2007;73(1): 107-110. 34. Take Y, Yoneda M, Hayashida K, Nakagawa S, Mizuno N. Enlargement of drill holes after use of a biodegradable suture anchor: quantitative study on consecutive postoperative radiographs. Arthroscopy. 2008;24(3):251-257. 35. Takubo Y, Morihara T, Namura T, et al. Anchor hole enlargement after arthroscopic Bankart repair using absorbable suture anchors: a report of three cases. J Shoulder Elbow Surg. 2008;17(6):e16-e18. 36. Ticker JB, Lippe RJ, Barkin DE, Carroll MP. Infected suture anchors in the shoulder. Arthroscopy. 1996;12(5):613-615. 37. Weber SC. Surgical management of the failed SLAP repair. Sports Med Arthrosc. 2010;18(3):162-166. 38. Yakacki CM, Poukalova M, Guldberg RE, et al. The effect of the trabecular microstructure on the pullout strength of suture anchors. J Biomech. 2010;43(10):1953-1959. 39. Zuckerman JD, Matsen FA 3rd. Complications about the glenohumeral joint related to the use of screws and staples. J Bone Joint Surg Am. 1984;66(2):175-180.

For reprints and permission queries, please visit SAGEs Web site at http://www.sagepub.com/journalsPermissions.nav