While hip arthroscopy utilization continues to increase, capsular management remains a controversial topic. Therefore the purpose of this research was to investigate the biomechanical effect of capsulotomy and capsular repair techniques on hip joint kinematics in varying combinations of sagittal and coronal joint positions.

Eight fresh-frozen hemipelvises (4 left, 6 male) were dissected of all overlying soft tissue, with the exception of the hip joint capsule. The femur was potted and attached to a load cell, while the pelvis was secured to a custom-designed fixture allowing static alteration of the flexion/extension arc. Optotrak markers were rigidly attached to the femur and pelvis to track motion of the femoral head with respect to the acetabulum.

Following specimen preparation, seven conditions were tested: i) intact, ii) after portal placement (anterolateral and mid-anterior), iii) interportal capsulotomy (IPC) [35 mm in length], iv) IPC repair, v)T-capsulotomy [15 mm longitudinal incision], vi) partial T-repair (vertical limb), vii) full T-repair. All conditions were tested in 15° of extension (−15˚), 0°, 30°, 60° and 90° of flexion. Additionally, all flexion angles were tested in neutral, as well as maximum abduction and adduction, resulting in 15 testing positions. 3Nm internal and external rotation moments were manually applied to the femur via the load cell at each position. Rotational range of motion and joint kinematics were recorded.

IPC and T-capsulotomies increased rotational ROM and mediolateral (ML) joint translation in several different joint configurations, most notably from 0–30˚ in neutral abduction/adduction. Complete capsular repair restored near native joint kinematics, with no significant differences between any complete capsular repair groups and the intact state, regardless of joint position. An unrepaired IPC resulted in increased rotational ROM, but no other adverse translational kinematics. However, an unrepaired or partially repaired T-capsulotomy resulted in increased rotational ROM and ML translation.

The results of this study show that complete capsular repair following interportal or T-capsulotomy adequately restores rotational ROM and joint translation to near intact levels. Where feasible, complete capsular closure should be performed, especially following T-capsulotomy. However, further clinical evaluation is required to determine if adverse kinematics of an unrepaired capsule are associated with patient reported outcomes.

Previous studies have identified the anterolateral complex (ALC) as having an important role in controlling anterolateral rotatory laxity following anterior cruciate ligament injury and subsequent reconstruction. In particular, injury to the iliotibial band (ITB) and its component deep (dITB) and capsulo-osseous (coITB) layers, have been shown to significantly correlate with different grades of the pivot-shift test in patients with acute ACL injuries. However, the kinematic properties of the capsulo-osseous layer of the ITB, throughout knee range of motion, are not fully understood.

The purpose of this study was to quantify the kinematic behaviour of the capsulo-osseous layer of the ITB through various degrees of knee flexion. Ten fresh-frozen cadaveric knee specimens were dissected to expose the capsulo-osseous layer of the iliotibial band. Radiopaque beads were embedded, at standardized increments (12.5%, 25%, 50% and 75% of total length from proximal to distal), into the tissue and fluoroscopic images were taken from 0o to 105o of knee flexion in 15° increments. The positions of the beads were identified in each image and the length, width, and area changes of the capsulo-osseous layer were calculated. Comparisons of the total length of the anterior and posterior borders of the coITB through knee ROM were conducted using a two-way (8 knee angles by 2 borders) repeated measures analysis of variance (rm-ANOVA), whereas the effect of knee angle on isometry and total area changes was assessed using one-way rm-ANOVAs (α=0.05)

There was a significant increase in the length of the anterior capsulo-osseous layer at flexion angles greater than 15o and on the posterior border at angles greater than 75 o with changes occurring primarily at 12.5 % of the total length. In addition, at all flexion angles the length changes were significantly larger in the anterior border compared to the posterior border. Meanwhile, non-homogenous decreases in width and area were found with increasing flexion angle. The distance between the capsulo-osseous layer insertion on the distal femur and proximal tibia significantly increased from 60o-105o, maximal changes occurred at 105o (9.64 [4.12] %, p = 0.003).

The primary finding of this study was that the coITB behaved in a non-isometric fashion, with significant increases in length occurring at flexion angles greater than 15o. Moreover, these changes in length were non-homogenous across the different regions of the coITB that were investigated, with the greatest changes occurring in the proximal segments (0–25%). The data presented here suggest that coITB in flexion angles from 0o to 105o behaves in a non-isometric fashion, with the majority of its length change occurring in its proximal segment. Further quantification of the pathway that the coITB takes with respect to osseous landmarks may result in improvements in ALC procedures as an augmentation to ACL reconstruction, thereby potentially improving rotational stability and clinical outcomes.

Introduction

Ligament reconstruction following knee soft tissue injuries, such as posterior cruciate ligament (PCL) tears, aim to restore normal joint function and motion; however, persistant pathomechanical joint behavior indicates that there is room for improvement in current reconstruction techniques. Increased attention is being directed towards the roles of secondary knee stabilizers, in an attempt to better understand their contributions to kinematics of knees. The objective of this study is to characterize the relative biomechanical contributions of the posterior oblique ligament (POL) and the deep medial collateral ligament (dMCL) in PCL-deficient knees. We hypothesized that, compared with the POL, the dMCL would have a more substantial role in stabilizing the medial side of the knee, especially in flexion (slack POL).

Aim

It has been suggested that the use of a pilot-hole may reduce the risk of fracture to the lateral cortex. Therefore the purpose of this study was to determine the effect of a pilot hole on the strains and occurrence of fractures at the lateral cortex during the opening of a high tibial osteotomy (HTO) and post-surgery loading.

Injury to the anterolateral ligament (ALL) has been reported to contribute to high-grade anterolateral laxity following anterior cruciate ligament (ACL) injury. Failure to address ALL injury has been suggested as a cause of persistent rotational laxity following ACL reconstruction. However, lateral meniscus posterior root (LMPR) tears have also has been shown to cause increased internal rotation and anterior translation of the knee. Due to the anatomic relationship of the ALL and the lateral meniscus, we hypothesise that the ALL and lateral meniscus work synergistically, and that a tear to the LMPR will have the same effect on anterolateral laxity as an ALL tear in the ACL deficient knee.

Sixteen fresh frozen cadaveric knee specimens were potted into a hip simulator(femur) and a six degree-of-freedom load cell (tibia). Two rigid optical trackers were inserted into the proximal femur and distal tibia, allowing for the motion of the tibia with respect to the femur to be tracked during biomechanical tests. A series of points on the femur and tibia were digitised to create bone coordinate systems that were used to calculate the kinematic variables. Biomechanical testing involved applying a 5Nm internal rotation moment to the tibia while the knee was in full extension and tested sequentially in the following three conditions: i) ACLintact; ii) Partial ACL injury (ACLam) – anteromedial bundle sectioned; iii) Full ACL injury (ACLfull). The specimens were then randomised to either have the ALL sectioned first (ALLsec) followed by the LMPRsec or vice versa. Internal rotation and anterior translation of the tibia with respect to the femur were calculated. A mixed two-way (serial sectioning by ALL section order) repeated measures ANOVA (alpha = 0.05).

Compared to the ACLintact condition, internal rotation was found to be 1.78° (p=0.06), 3.74° (p=0.001), and 3.84° (p=0.001) greater following ACLfull, LMPRsec and ALLsec respectively. LMPRsec and the ALLsec resulted in approximately 20 of additional internal rotation (p=0.004 and p=0.01, respectively) compared with the ACL deficient knee (ACLfull). No difference was observed between the ALL and LMPR sectioned states, or whether the ALL was sectioned before or after the LMPR (p=0.160). A trend of increasing anterior translation was observed when the 5Nm internal rotation moment was applied up until the ACL was fully sectioned; however, these differences were not significant (p=0.070).

The ALL and LMPR seem to have a synergistic relationship in aiding the ACL in controlling anterolateral rotational laxity. High-grade anterolateral laxity following ACL injury may be attributed to injuries of the ALL and/or the LMPR. We suggest that the lateral meniscus should be thought of as part of the anterolateral capsulomeniscal complex (i.e., LM, ITB, and ALL) that acts as a stabiliser of anterolateral rotation in conjunction with the ACL.

We reviewed all patients that suffered a deep infection following anterior cruciate ligament (ACL) repair kept between January 2007 and April 2011 at our teaching hospital NHS trust, and the two local private hospitals.

18 patients were identified. All patients underwent at least 2 arthroscopic washouts, with limited synovectomy if required. Targeted antibiotics were commenced according to the culture results, and following microbiological advice. These patients were reviewed at a minimum of 1 year following eradication of infection (range 12–46 months). There were 7 surgeons performing the ACL reconstructions.

The primary outcome measure was graft failure requiring revision. Our secondary outcome measures were a history of ongoing instability, KT 1000™ measurement, Tegner and Lysholm outcome scores. There were 18 patients identified as having suffered infection after ACL infection (mean age 24.3 years, range 15–38 years). Average C Reactive Protein (CRP) was 217 on admission (range 59–397). The most common organism isolated was coagulase negative staphylococcus in 47.3% of cases. There were 3 graft failures within the infection group. Of the remaining 15 patients there were no episodes of ongoing instability and mean pivot shift grade was 1.1, mean KT 1000™ side-to- side difference was +1.8mm. There was a reported drop on the Tegner activity score of 1.75 (range 0–6) and mean Lysholm score was 89 (range 56–100).

The failure rate is slightly higher than that reported in the literature. Patient reported outcome measures in the patients are broadly consistent. We recommend an aggressive approach to the treatment of deep infection following ACL reconstruction, in order to achieve a satisfactory outcome.

Background

Mechanical trauma to articular cartilage is a known risk factor for Osteoarthritis (OA). The application of single impact load (SIL) to equine articular cartilage is described as a model of early OA changes and shown to induce a damage/repair response. Recombinant Human Fibroblast Growth Factor-18 (rhFGF-18) has been previously shown to have anabolic effects on chondrocytes in vitro. The aim of this in vitro study was to ascertain the effect of rhFGF-18 on the repair response of mechanically damaged articular cartilage.

Introduction

The aim of this study was to investigate whether methylene blue dye, commonly used in sterile surgical marker pens, would have an effect on human chondrocyte viability, when cultured on a collagen membrane in-vitro.

Introduction

The purpose of this study was to investigate whether combining PRP or concentrated bone marrow aspirate (CBMA) with a biphasic collagen/glycosaminoglycan (CG) scaffold would improve the outcome of the treatment of full thickness osteochondral defects in sheep.

Long-term follow up after replacement arthroplasty has become established as a “Gold Standard”, providing information that can aid optimisation in future prosthetic design and use. In less mainstream joint replacements however, the evidence for use of prostheses, and in particular long-term outcome, is scarce.

A cohort of 71 patients (93 implants) was reviewed in 1997 having had a De la Caffinière prosthesis implanted between 1980 and 1989. The conclusions of the study included the findings that the replacement was generally well regarded by recipients, pain was improved and survivorship was comparable with data from the best hip replacements.

Ethical permission was obtained to review the same cohort ten years on (16 – 26 years post-op). Similar outcome measures were employed as in the original study but in addition formal grip strength measurements were taken, along with newer outcome scores including the DASH (Disability of arm, shoulder and hand) and EQ-5D (a European quality of life measure). Radiographs allowed assessment of radiological failure using the criteria from the original study.

We found a significant mortality rate in the interim period since the original review (27 patients, 36 implants). A further 8 implants in 8 patients had been removed and were not clinically reviewed as per patients’ wishes. However, 39 implants in 26 patients were available for follow up at a mean 19 years (SD 6.3) leaving a “lost to follow up” rate of 10 patients (10 implants). Survivorship at 26 years was 73.9% (95% CI 61.2, 86.6) with the end-point as revision. Our data also demonstrated continued patient satisfaction without pain, satisfactory power and thumb mobility.

Such information may be used to counsel future patients requiring surgery that there is a functional alternative to excision arthroplasty (trapeziectomy).

Introduction: Platelet rich plasma (PRP) has been hypothesised to be of potential benefit to articular cartilage tissue engineering, through its release of autologous growth factors. The aim of this study was to ascertain whether the addition of thrombin is required to achieve platelet activation and sustained growth factor release in-vitro, when PRP is applied to a collagen based osteochondral scaffold.

Methods: Collagen/glycosaminoglycan scaffolds were fashioned, to which equal combined volumes of test substances were added (n=3): 500μl PRP; 375μl PRP + 125μl autologous thrombin (3:1); 455μl PRP + 45μl bovine thrombin (10:1). One ml of DMEM/F12 medium was added to each scaffold and changed completely at 12/24 hours, and 3/10 days, following which release of TGF-β1, PDGF-AB and bFGF were measured using ELISA. Secondly, equal sized collagen/glycosaminoglycan and polylactide co-glycolide scaffolds were fashioned to which 500μl of PRP were added (n=3). Similar conditions were followed as previously except that only PDGF-AB was assayed.

Results: A similar cumulative release profile of all growth factors was found over the 10 day period. An increase in growth factor release was seen in the PRP only group at all time points with PDGF-AB in particular reaching statistical significance at all time points (p< 0.006). These findings remained apparent when a correction for volume was made (p< 0.028) suggesting a particular role of the collagen in platelet activation. This was shown in the second experiment, in which a significantly increased cumulative volume of PDGF-AB was released from the collagen/glycosaminoglycan scaffold without thrombin activation (p< 0.04).

Discussion: This study shows that collagen is a potent activator of platelets, requiring no further addition to achieve satisfactory growth factor release when applied clinically. These results suggest that if PRP is combined with polymer scaffolds, it should be activated with thrombin to achieve optimum growth factor release.

Introduction: Current treatment options for small, contained articular cartilage defects include microfracture, osteochondral autograft plugs or newer synthetic plugs. Chondromimetic is a novel biphasic biological scaffold composed of collagen and glycosaminoglycan. The addition of brushite provides the scaffold with a regionally specific component mimicking both phases of the osteochondral unit. The aim of this study was to show the efficacy of Chondromimetic in repairing a surgically created osteochondral defect in a caprine model.

Methods: Osteochondral defects were made in the lateral trochlear sulcus (LTS) and medial femoral condyle (MFC) of nine goats. Chondromimetic scaffolds (6x6mm) were inserted into each defect (n=6), while three controls had defects left empty (n=3). All animals were sacrificed at 26 weeks postoperatively. Macroscopic evaluations and quantitative stiffness properties were assessed. Histological sections were taken at approximately the centre of the defect, stained with Safrinin O/Fast Green and scored using a validated quantitative assessment tool.

Results: Macroscopically, the repair tissue scored higher in the MFC and LTS (p< 0.05) compared to controls. In all defects, the mechanical stiffness was found to be within one standard deviation of native cartilage, except that of the LTS controls. Histologically, the predominant tissue in the cartilage layer was deemed to be hyaline-like in three of six MFC defects, and five of six LTS defects according to the modified Sellers score. This was compared to one in three and zero of three in the MFC and LTS controls respectively.

Discussion: These results represent the early findings from an ongoing in-vivo study in which a further group of animals will be sacrificed at one year. At six months, the histology and mechanical properties are encouraging and should continue to improve with time. These results show that Chondromimetic may represent an acceptable alternative to marrow stimulation in the treatment of osteochondral defects.

Introduction: Platelet rich plasma (PRP) has been hypothesised to be of potential benefit to articular cartilage tissue engineering, through its release of autologous growth factors. The aim of this study was to ascertain whether the addition of thrombin is required to achieve platelet activation and sustained growth factor release in-vitro, when PRP is applied to a collagen based osteochondral scaffold.

Methods: Collagen/glycosaminoglycan scaffolds were fashioned, to which equal combined volumes of test substances were added (n=3): 500μl PRP; 375μl PRP + 125μl autologous thrombin (3:1); 455μl PRP + 45μl bovine thrombin (10:1). One ml of DMEM/F12 medium was added to each scaffold and changed completely at 12/24 hours, and 3/10 days, following which release of TGF-β1, PDGF-AB and bFGF were measured using ELISA. Secondly, equal sized collagen/glycosaminogly-can and polylactide co-glycolide scaffolds were fashioned to which 500μl of PRP were added (n=3). Similar conditions were followed as previously except that only PDGF-AB was assayed.

Results: A similar cumulative release profile of all growth factors was found over the 10 day period. An increase in growth factor release was seen in the PRP only group at all time points with PDGF-AB in particular reaching statistical significance at all time points (p< 0.006). These findings remained apparent when a correction for volume was made (p< 0.028) suggesting a particular role of the collagen in platelet activation. This was shown in the second experiment, in which a significantly increased cumulative volume of PDGF-AB was released from the collagen/glycosaminoglycan scaffold without thrombin activation (p< 0.04).

Discussion: This study shows that collagen is a potent activator of platelets, requiring no further additive to achieve satisfactory growth factor release when applied clinically. These results suggest that if PRP is combined with polymer scaffolds, it should be activated with thrombin to achieve optimum growth factor release.

Introduction: Current treatment options for small, contained articular cartilage defects include microfracture, osteochondral autograft plugs or newer synthetic plugs. Chondromimetic is a novel biphasic biological scaffold composed of collagen and glycosaminoglycan. The addition of brushite provides the scaffold with a regionally specific component enabling the scaffold to mimic both phases of the osteochondral unit.

The aim of this study was to show the efficacy of Chondromimetic in repairing a surgically created osteochondral defect in a caprine model.

Methods: Osteochondral defects were made in the lateral trochlear sulcus (LTS) and medial femoral condyle (MFC) of nine goats. Chondromimetic scaffolds (6x6mm) were inserted into each defect (n=6), while three controls had defects left empty (n=3). All animals were sacrificed at 26 weeks postoperatively. Macroscopic evaluations and quantitative stiffness properties were assessed. Histological sections were taken at approximately the centre of the defect, stained with Safrinin O/Fast Green and scored using a validated quantitative assessment tool.

Results: Macroscopically, the repair tissue scored higher in the filled MFC and LTS (p< 0.05) compared to controls. In all defects, the mechanical stiffness was found to be within one standard deviation of native cartilage, except the LTS controls. Histologically, the predominant tissue in the cartilage layer was hyaline-like in three of six filled MFC defects, and five of six filled LTS defects according to the modified Sellers score. This was compared to one in three and zero of three in the MFC and LTS controls respectively.

Discussion: These results represent the early findings from an ongoing in-vivo study in which a further group of animals will be sacrificed at one year. At six months, the histology and mechanical properties are encouraging and should continue to improve with time. These results show that Chondromimetic may represent an acceptable alternative to marrow stimulation in the treatment of osteochondral defects.

Introduction: Double bundle anterior cruciate ligament (DBACL) reconstruction has been shown to improve objective measurements of rotational stability. However, subjective improvement in patient outcome has yet to be shown. The aim of this study was to investigate whether double bundle ACL reconstruction could provide satisfactory subjective and objective outcome in a consecutive patient series at a minimum of two years post operatively.

Methods: From May 2006 a consecutive group of patients underwent double bundle ACL reconstruction using doubled semitendinosus (anteromedial bundle) and gracilis (posterolateral bundle) tendons fixed with interference screws in both femur and tibia. Pre and post operative subjective and objective IKDC 2000 knee scores were collected including pivot shift and KT-1000 arthrometer assessments. A comparison was made to a cohort of single bundle ACL reconstructions performed by the same surgeon whose data was collected previously.

Results: The group included 29 patients (26 male, mean age 30; range 18–47). A significant improvement in subjective IKDC 2000 score between preop (54.8) and last follow up (76.3) was shown (p = 0.00); 96% of subjects were IKDC grade A or B; 85% of subjects had a negative pivot shift on last review with 15% eliciting a pivot glide. There were no differences in subjective or objective outcomes between the double bundle and single bundle groups.

Discussion: This data compares favourably with other published series of double bundle ACL reconstruction. Although subjective improvement in functional outcome is seen, early results do not show an improvement over single bundle reconstruction. It remains to be shown if this technique will confer an overall improvement in long term outcome via the more anatomical reconstruction.

Introduction: Platelet rich plasma (PRP) has been hypothesised to be of potential benefit to articular cartilage tissue engineering, through its release of autologous growth factors.

The aim of this study was to ascertain whether the addition of thrombin is required to achieve platelet activation and sustained growth factor release in-vitro, when PRP is applied to a collagen based osteochondral scaffold.

Methods: Collagen/glycosaminoglycan scaffolds were fashioned, to which equal combined volumes of test substances were added (n=3): 500μl PRP; 375μl PRP + 125μl autologous thrombin (3:1); 455μl PRP + 45μl bovine thrombin (10:1). One ml of DMEM/F12 medium was added to each scaffold and changed completely at 12/24 hours, and 3/10 days, following which release of TGF-β1, PDGF-AB and bFGF were measured using ELISA. Secondly, equal sized collagen/glycosaminogly-can and polylactide co-glycolide scaffolds were fashioned to which 500μl of PRP were added (n=3). Similar conditions were followed as previously except that only PDGF-AB was assayed.

Results: A similar cumulative release profile of all growth factors was found over the 10 day period. Greater growth factor release was seen in the PRP only group at all time points with PDGF-AB in particular reaching statistical significance at all time points (p< 0.006). These findings remained apparent when a correction for volume was made (p< 0.028) suggesting a particular role of the collagen in platelet activation. This was shown in the second experiment, in which a significantly increased cumulative volume of PDGF-AB was released from the collagen/glycosaminoglycan scaffold without thrombin activation (p< 0.04).

Discussion: This study shows that collagen is a potent activator of platelets, requiring no further addition to achieve satisfactory growth factor release when applied clinically. These results suggest that if PRP is combined with polylactide co-glycolide scaffolds, it should be activated with thrombin to achieve optimum growth factor release.

The purpose of this study was to describe our experience of the Calaxo Osteoconductive interference screw (Smith & Nephew) when used for both femoral and tibial graft fixation in Double Bundle ACL reconstruction.

Since May 2006, all patients with an ACL deficient knee were reconstructed using the Double Bundle technique. All were followed prospectively and outcome data collected. Evidence of fixation failure was established subjectively by clinical examination (Lachman, Anterior Draw, Pivot Shift) and objectively via KT-1000 arthrometer. Following ethical approval, post-operative CT scans (immediate and 1 year) were performed on our first 10 patients allowing assessment of tunnel dimensions/fill.

Thirty two patients (29 male, 3 female) with a mean age of 30 (range 18-46) were included. At last follow-up, no evidence of graft/fixation failure was found; KT-1000 mean side-side difference 1.4mm (range −3 to +6). All patients had a positive pivot shift preoperatively which was abolished postoperatively. One patient had a postoperative infection with no other complications reported. Radiologically the screws did not show complete resorption but areas of new bone were identified.

We have shown satisfactory results with use of the Calaxo screw when used in Double Bundle Reconstruction. We have not had any cases of the adverse local soft tissue reaction, which has led to this screw being withdrawn from clinical use. Even when using a total of four screws in each knee. A previous study published by Seibold (2007) has shown tunnel widening and communication when suspensory fixation is used in Double Bundle reconstruction. This has the potential risk of leading to fracture between the tunnels. This has not been seen with the Calaxo screw which may be a result of the biological action of the screw which should ultimately lead to a reduction in these risks.

Introduction: The purpose of this study was to describe our experience of the Calaxo Osteoconductive interference screw (Smith & Nephew) when used for both femoral and tibial graft fixation in Double Bundle ACL reconstruction.

Methods: Since May 2006, all patients with an ACL deficient knee were reconstructed using the Double Bundle technique. All were followed prospectively and outcome data collected.

Evidence of fixation failure was established subjectively by clinical examination (Lachman, Anterior Draw, Pivot Shift) and objectively via KT-1000 arthrometer.

Following ethical approval, post-operative CT scans (immediate and 1 year) were performed on our first 10 patients allowing assessment of tunnel dimensions/fill.

Results: Thirty two patients (29 male, 3 female) with a mean age of 30 (range 18–46) were included. At last follow-up, no evidence of graft/fixation failure was found; KT-1000 mean side-side difference 1.4mm (range −3 to +6). All patients had a positive pivot shift preoperatively which was abolished postoperatively. One patient had a postoperative infection with no other complications reported. Radiologically the screws did not show complete resorption but areas of new bone were identified.

Discussion: We have shown satisfactory results with use of the Calaxo screw when used in Double Bundle Reconstruction. We have not had any cases of the adverse local soft tissue reaction, which has led to this screw being withdrawn from clinical use. Even when using a total of four screws in each knee.

A previous study published by Seibold (2007) has shown tunnel widening and communication when suspensory fixation is used in Double Bundle reconstruction. This has the potential risk of leading to fracture between the tunnels.

This has not been seen with the Calaxo screw which may be a result of the biological action of the screw which should ultimately lead to a reduction in these risks.

Aim: To characterise the reasons for failure following primary surgical stabilisation, the indications and the results of surgery in a subgroup of patients undergoing revision arthroscopic shoulder stabilisation.

Methods: All patients who underwent revision shoulder stabilisation over a 3 year period were included in the study. Information about the index procedure, imaging and the findings at arthroscopy were used to characterise the mechanism of failure. Patients were assessed clinically and WOSI scored at 6 and 12 months and then annually.

Results: Thirty-six patients were reviewed. Twenty underwent an arthroscopic (AR) and 15 underwent an open (OR) revision procedure. The AR index procedures were 15 arthroscopic (4 suture-anchor, 10 thermal shrinkage, 1 unknown) and 5 open (3 Bankart, 2 Putti-Platt). The OR index procedures were arthroscopic 10 (4 suture-anchor, 3 transglenoid, 2 thermal shrinkage) and 6 open (4 Bankart, 2 Putti-Platt). Soft-tissue failure was the primary mechanism in all AR cases (4 further trauma, 7 inappropriate primary procedure, 3 missed pathology, 6 technical error). The OR primary mechanism of failure was bone loss 12 (4 epileptic) and soft-tissue failure 3. At AR follow-up of 25.8 months (10–45) there had been 1 re-dislocation and 1 subluxation, mean WOSI 298 (68–517). There had been no OR re-dislocations at 22.5 months (10–42) follow-up, mean WOSI 313 (76–621).

Conclusion: Recurrent instability following surgical stabilisation maybe due to bone loss or soft-tissue failure (further trauma, inappropriate primary procedure, missed pathology, technical error). Revision arthroscopic surgery maybe indicated in soft tissue failure cases and can produce acceptable medium term results.