Abstracts

Alwin Jaeger, Jens Herresthal, Sami Khoudeir Frankfurt, Germany

Introduction: This retrospective study presents a long-term result after meniscal repair with emphasis on early development of osteoarthritis in the knee joint. Material and Methods: From 1987 to 1996 meniscal repair using the inside-out technique was performed on 54 patients. 25 patients had an isolated tear of the meniscus and 29 additionally a tear of the anterior cruciate ligament (ACL, combined injury). 18 patients of those with a combined injury showed a stable knee during clinical follow-up. A separate evaluation of this group was carried out. The mean follow-up was 6.4 years. 22 patients were examined over a period of 2-5 years (mean 3.5) after surgery; 32 patients after 5 years (mean 8.9). The clinical evaluation was carried out using IKDC, Tegner (Teg) activity scale, Lysholm (Lysh), and Tapper and Hoover (Tapp) Score as well as the Fairbank (Fb) radiographic evaluation. Results: According to the Fairbank's evaluation the results show that up to 5 years after meniscal repair only minimal radiographic changes were detectable. In only 21% of patients with an isolated meniscal lesion and in 13% of those with a stable knee after a combined injury beginning of osteoarthritic changes were noted more than 5 years postoperative. Early osteoarthritic changes were present within 5 years after surgery in non-stable knees (untreated rupture, suturing).

Conclusion: Patients with meniscal repair mostly regain their original Tegner activity level and good to very good results in the clinical scores. However, the results clearly indicate that the success of meniscal repair depends on the stability of the knee. Thus, meniscal repair in stable knees is recommended to prevent early development of osteoarthritis.

Andrew S. Levy, Sanjay Patari Summit, New Jersey, USA

Introduction: Bioabsorbable implants have been developed to improve meniscal repair and reduce morbidity. Current recommended implant size is based solely on the anatomical location of the tear. The basis for these recommendations has not been reported. The failure to accommodate individual variations in anatomy may result in the use of suboptimally sized implants. The purpose of this study is to address the issue of meniscal variance relevant to implant size. Methods: 17 cadaver knees from 17 different subjects of varying morphology and sex were obtained. For each specimen the tibial plateau, medial and lateral menisci, and adjacent joint capsule were isolated. The peripheral rim length of the medial and lateral menisci was recorded. The lateral and medial menisci were divided into sixteen equal segments. The widths of the menisci and the adjacent joint capsule at each of these segments were recorded individually by 3 independent observers. Statistical analysis was performed using a repeated measure model. Results: Mean meniscal width was 9 mm (range 7-10) for the anterior third, 11 mm (range 10-14) for the middle third and 14 mm (range 11-15) for the posterior third of the medial meniscus. In the lateral meniscus, the posterior third averaged 12 mm (range 11-12). Capsule thickness averaged 2.6 mm (range 2-4.1). No correlation was noted between meniscal thickness and circumference. (P <.05). Discussion: The design of the implants in the literature suggests that its size should measure approximately one-half of the total meniscal width at the site of placement. The current sizing recommendations for implants suggest a tack size of 10 mm for the anterior third, 13 mm for the middle third, and 13 mm or 16 mm for the posterior third of the meniscus. Since the implant need only be approximately one-half of these observed widths, it is clear that current sizing recommendations are overestimated. Conclusion: Many current meniscal repair implant sizes are too long and place the patient at risk for implant impingement upon nearby neurovascular structures and soft tissue irritation. The lack of correlation between width and circumference raises questions that need to be addressed to enhance our understanding of how meniscal variance effects function.

C.T. Mehlman, E.J. Wall, Gordon Olsen Cincinnati, Ohio, USA

Introduction: Modern approaches to meniscal surgery have focused on meniscal preservation. Multiple meniscal repair techniques and a variety of commercially available devices to assist with repair now exist. The effectiveness of such approaches to meniscal repair have been well studied in the adult population. The purpose of our study was to study incidence of complications following meniscal repair in a pediatric population using meniscal arrows (Bionix, Malvern, PA). Methods: We performed a retrospective review of our hospital database aimed at identifying all children who had undergone arthroscopic meniscal repair with this device since we began using it in 1997. Inclusion criteria included use of the meniscal arrow (Bionix, Malvern, PA), and age 18 years or less. Data collected included area and type of tear, number of arrows used, use of other repair materials, weight bearing status and whether or not immobilization was employed post-operatively. We reviewed hospital records and post-operative follow-up notes to find patients who had complications related to: persistent pain, effusion and need for removal of the implants. Results: We identified eleven consecutive pediatric patients (avg age 15 yrs, range 9 to 18 yrs) who had meniscal tears that were repaired arthroscopically using meniscal arrows alone or in combination with PDS suture. Average follow-up was 18 months (range 4 to 26 months). There were seven males and four females. Tear sizes ranged from 1 to 5 cm and were all located in either the red on red or red on white zones. The number of arrows used ranged from 2 to 6. All patients were immobilized postoperatively. Two patients were instructed to remain nonweightbearing and 9 were allowed weightbearing as tolerated. Four of the eleven patients (36%) had complications with most patients having more than one complication. Persistent pain and swelling was noted in 4 patients, reactive synovitis in 3 patients, failure of arrow resorption (with eventual extrusion from the skin) in 2 patients, and reflex sympathetic dystrophy in 1 patient. All 4 patients with complications underwent repeat arthroscopy, which revealed healed menisci. Removal of meniscal arrow remnants was required in 3 patients in order to relieve their symptoms. Discussion: Preservation and repair of meniscal tissue in children who sustain meniscal injuries is considered to be critically important due to the risk of osteoarthrosis associated with meniscectomy. A variety of meniscal repair devices are now available to knee arthroscopists. Such devices have not been widely studied in children. Our results have given us concern about using this particular device (Bionix Meniscal Arrows) in children (36% complication rate). Case reports of similar complications (in both adults and children) already exist in the literature. Given this high rate of complications, we cannot recommend routine pediatric use of this device for meniscal repair without further study.

Chul-Won Ha, Jin-Hwan Ahn Seoul, Korea

Introduction: The clinical features and techniques of repair for various types of the meniscal tears have been discussed a lot. However, the transection or complete radial tear of the medial meniscus at the posterior tibial attachment site has not well been discussed in the literature and is known irreparable. The authors introduce some unique clinical features of the patients with transection of the posterior horn of the medial meniscus at or very near to the posterior tibial attachment site, as well as the diagnostic clues of symptoms, signs and MRI. Those cases does not appear to be reparable with the usual arthroscopic meniscal techniques. The authors also introduce the arthroscopic technique of repair using pull-out suture for the transection of the posterior horn of the medial meniscus at the posterior tibial attachment site. Clinical Features: The characteristics of the knee pain was that it was very vague in location and intensity. By full flexion of the knee, the pain was increased and often more localized to the posterocentral portion. Six out of 7 cases had no history of injury of the symptomatic knee. On MRI, the transection of the posterior horn of the meniscus could be clearly identified in the coronal section of the posterior portion of the knee. The midportion of the medial meniscus was midly subluxated medially on coronal MRI. The subluxation has been one of the clues to identify the transection of the posterior tibial attachment of the posterior horn of the medial meniscus. On plain X-ray, there were no evidence of osteoarthritis in all cases, but the arthroscopic finding revealed degenerative changes of the articular cartilage in all. Technique of Repair: The anterolateral and anteromedial portals with trans-notch technique as well as the posteromedial portal were used in combination. The instability of the posterior tibial attachment of the medial meniscus was identified by routine arthroscopic examination and probing using posteromedial portal. The arthroscope introduced through the anterolateral portal was then advanced to the posterior compartment with transnotch technique to examine the posterior tibial attachment area of the medial meniscus. While viewing with the arthroscope, the articular surface side of the posterior tibial attachment site was prepared with motorized shaver and burr inserted through the posteromedial portal. The meniscal suture hook was introduced through the posteromedial portal or anteromedial portal, and 2 PDS sutures were passed through the posterior tibial attachment portion of the transected posterior horn of the meniscus. The sutures are then pulled out of the joint through the anteromedial portal. With the C-shaped tibial tunnel guide used for reconstruction of the ACL, two bone tunnels were made at the posterior tibial attachment site. Using thin wire loops, the PDS sutures passed through the meniscus were passed through the bone tunnels and tied over the anteromedial cortex of the proximal tibia. The stabilized posterior tibial attachment of the medial meniscus was documented under arthroscopic examination and probing. Discussion and Conclusion: The benefit of repair of the transection or complete radial tear of the menisci through the substance still does not seem to be clearly identified clinically and biomechanically. One can expect the transection of the posterior tibial attachment site of the medial meniscus will show more favorable results than the transection through the substance. Healing of the transected posterior end of the meniscus to the bone and/or the surrounding soft tissue will probably result in more favorable clinical outcome in the biomechanical aspect than the healing in the substance of the meniscus. We have one case of second look arthroscopy after the above described technique of repair. The posterior tibial attachment site was stabilized by healing to the soft tissue around PCL. The authors expect the repair will probably help the meniscus function in near anatomical position and prevent accelerated deterioration of the articular cartilage. In conclusion, the transection of the medial meniscus at the posterior tibial attachment site appears to be a rare but important clinical entity with some unique clinical features. Arthroscopic technique of repair introduced above appears to stabilize the posterior horn of the medial meniscus.

Dev K. Mishra, W. Dilworth Cannon, Jr San Francisco, California, USA

Introduction: Little is published about the fate of the failed meniscus repair. This study specifically addresses the outcome after rerepair of failed meniscus repairs. Methods: Out of a total of 376 patients undergoing meniscus repair, 12 patients underwent rerepair and were followed prospectively. Data was obtained from the index repair and analyzed for tear characteristics, repair methods, and concurrent ligament injury. At rerepair we noted the number and type of suture, presence of a ligament tear, and concurrent ligament reconstruction. All patients were assessed clinically at an average of 29 months after rerepair (range 24 to 47 months). When possible we obtained anatomic assessment of the rerepair with arthroscopy or arthrography. Results: At the index repair 8 of 12 patients had an ACL tear but only 1 had ACL reconstruction. Four primary meniscus repairs were done with absorbable suture and 8 with nonabsorbable suture. The time from index repair to rerepair averaged 38.5 months. Nine patients had an ACL tear at the rerepair and all 9 had concurrent ACL reconstruction. All rerepairs were performed using inside out suturing with 2-0 Ethibond placed in a vertical mattress configuration. At final follow up from the rerepair 11 knees were asymptomatic and clinically healed. The patient not clinically healed underwent partial meniscectomy. Six patients underwent diagnostic arthroscopy or arthrography to assess healing. Using Henning's criteria 3 of 6 menisci were not healed. Discussion: In an effort to avoid late degenerative changes after meniscectomy an emphasis is placed on meniscus preservation. Some primary meniscus repairs will fail, and the surgeon may believe that partial meniscectomy is the only available alternative for the symptomatic failed repair. This study documents a favorable clinical outcome from rerepair in a high percentage of cases. A factor consistently associated with a failed meniscus repair is anterior cruciate ligament instability. If an ACL tear is present at the meniscus rerepair we recommend concurrent ACL reconstruction. The surgeon should be encouraged to perform meniscus rerepair if favorable tear characteristics exist, such as tear length less than 4 cm, rim width of 3 mm or less, and good tissue quality.

Samuel K. Tadet, J. Mayfield Albuquerque, New Mexico, USA

Subtotal menisectomy can result in pain, deterioration of joint surfaces, and instability. This is a report of our experience with twenty-three consecutive patients in a private practice who had had previous subtotal menisectomy with resulting pain and disability. These patients were treated with meniscal allograft. Three had concomitant ACL reconstructions and one had osteochondral transfers. Nine patients had second look arthroscopies. One patient reinjured the allograft in a work accident and underwent a replacement of his allograft. Follow-up times range from eight to sixty-three months. For this study records were reviewed and a person who was not part of the care team interviewed patients using a standard form. All patients had significant impairment of ADL prior to surgery and all reported improvement. Twenty returned to work, one medically retired, and two are lost to follow-up. For the person with previous subtotal menisectomy this procedure effectively reduces pain and increases function. This may be due to a reduction in the progression of gonarthrosis.

Matthew J. Matava, Douglas Robertson, Kirk Smith, William Maloney St. Louis, Missouri, USA

Introduction: Osteochondral autogenous transfer or mosaicplasty appears to provide satisfactory outcomes for the treatment of full-thickness distal femoral articular cartilage defects. However, it is unclear to what degree the donor grafts match the recipient cartilage bed in regard to cartilage thickness and curvature. The purpose of this study was to compare the morphology (thickness and curvature) of potential donor and recipient sites, in hopes of improving the structural match of the implant. Methods: Four fresh-frozen distal femurs were selected from 2 male donors (32 and 42 years of age). Neither patient had any bone diseases, including arthritis, nor any prior surgical procedure on the knee joint. The femurs were CT scanned with 0.4 mm in-plane resolution. Three-dimensional volumetric reconstructions and computer-aided design (CAD) models of the articular cartilage and bone were created from the CT data. The antero-superior aspects of the lateral and medial femoral condylar regions (donor sites) were divided into thirds (proximal to distal). The medial and distal femoral condyles (recipient sites) where divided into 9 subsections (3 medial-lateral and 3 anterior-posterior). Cartilage thickness was measured perpendicular to the surface within these regions with the aid of the CT software. Cartilage surface profiles were compared visually and with curve-fitting techniques. Results: The average donor site cartilage thickness ranged from 1.4 to 1.8 mm. The thickest donor regions were the central thirds. There was no significant difference between cartilage thickness on the medial and lateral femoral condylar donor sites. The average recipient site cartilage thickness ranged from 1.4 to 2.5 mm. There were no significant differences between cartilage thickness in matching regions (subsections) of the medial or lateral femoral condyles. The cartilage was thickest in the central region of each condyle. In these central regions the recipient site cartilage thickness was significantly greater than donor site thickness (ave 0.4 mm) (P <.05). Both distal femoral condylar articular surfaces fit the shape of a sphere. The donor site regions also approached a spherical shape three dimensionally and did not display a significant geometric mismatch. Conclusions: Our hypothesis is that structurally matching donor and recipient cartilage sites will improve the mechanics and outcomes of osteochondral autograft implantations. This preliminary study indicates that cartilage thickness and surface curvature mismatch between donor and recipient regions is small and can be minimized using the central region of the donor sites. The mechanical and clinical impact of mismatch awaits future study.

Duchow Jochen, Hess Thomas, Kohn Dieter Homburg, Saar, Germany

Introduction: A current option in the treatment of cartilage defects is the use of osteochondral autografts. Several instrument sets for osteochondral autografting have recently become available. These systems differ in the way the grafts are harvested and recipient sockets created, and in the relation in size of graft and recipient socket. Methods: In a biomechanical model we used pull-out testing to evaluate primary stability of osteochondral grafts on porcine femurs (n = 10 for all experiments). All tests were carried out on 8 mm diameter, 15 mm length grafts that were seated into 15 mm sockets. Press-fit stability was tested with and without cyclic preloading of the grafts (100 cycles, minimal force 10N, maximal force 100N). The following devices were tested: (1) OATS-System (Arthrex, Naples, USA) with mallet driven chisels. (2) COR-System (Innovasive Devices Inc., Marlborough, USA) with mallet driven chisels. (3) DBCS-System (Stratec, Freiburg, Germany) with power trephine. Results: Mean primary stability was 123N for the OATS-System (range 66-210N), 247N for the COR-System (range 126-400N) and 239N for the DBCS System (range 188-376N). Differences of means were significant between the OATS-system and the COR- and DBCS-system. No significant difference was found between the COR- and the DBCS-System. Cyclic preloading of the grafts prior to pull-out testing did not significantly alter the press-fit stability achieved. Conclusions: Primary press-fit stability of osteochondral grafts of the same size depends on the device or instruments chosen for grafting. The differences between the systems tested seems to depend on the different relations between grafts and recipient sockets. Devices should allow predictable and uniform press-fit fixation of osteochondral grafts in order to obtain consistent results.

Edward S. Bittan Palm Bay, Florida, USA

Objective of Study: Evaluation of the use of arthroscopic osteochondral core grafting in the management of articular cartilage defects associated with anterior cruciate ligament tears. Materials and Methods: A full thickness articular cartilage defect of the femoral condyle has been reported to occur in 16% of knees that sustained a complete rupture of the anterior cruciate ligament (ACL). Heretofore, management of articular defects has been limited to debridement of the damaged articular cartilage and not to reconstruction. Over the past three years, twenty ACL-deficient knees with full thickness articular cartilage defects underwent arthroscopically-assisted ACL reconstruction using an autogenous bone-patellar tendon-bone and arthroscopic resurfacing of the articular cartilage defect using autogenous cylindrical osteochondral grafts. Simultaneous resurfacing of a full thickness articular cartilage defect of the load-bearing surface of the femoral condyle, at least 8 mm in diameter, was performed arthroscopically using osteochondral plug grafts. The osteochondral grafts were harvested from the superior and lateral rim of the intercondylar notch prior to notchplasty. Patients were treated post-operatively with an aggressive rehabilitation protocol, but maintained toe touch weight bearing for three weeks if more than one graft was used to resurface the defect. Results: Knees that underwent a “second look” arthroscopic examination at four months, six months or two years following reconstruction demonstrated complete remodeling of the articular surface defect that had been grafted. The core grafts at the recipient site were fully incorporated and remodeled and could no longer be identified arthroscopically. There was no evidence for degeneration of the grafted articular cartilage or for degeneration of the articular surface. Harvest areas at the rim of the intercondylar notch were healed and remodeled at four months with a fibrocartilage mantle. At two years post-operatively, all but two patients had returned to full activities without restriction and were asymptomatic. Two of the patients complained of occasional joint line tenderness without an effusion after high impact activities. Conclusion: The short-term results of this study suggest that full thickness articular cartilage defects associated with ACL tears may be effectively resurfaced arthroscopically using cylindrical osteochondral plug grafts harvested from the rim of the intercondylar notch prior to notchplasty.

Dimitrios Koulaus, Wolfgang Schultz, Lorenz Kamp, Ingo Muller Göttingen, Niedersachsen, Germany

Purpose of study: The preliminary presentation of clinical and radiological evaluation of autologous chondrocyte transplantation as an option in the treatment of osteochondritis dissecans. Material and method: The cases of 5 patients with diagnosed osteochondritis dissecans of the knee and 7 patients with osteochondritis dissecans of the talus were reviewed at our department. The average follow up time was 13 (18-10) months for the knee group and 14 (18-8) months for the talar group. The mean age of the patients with the knee lesion was 31 (38-20) years. The mean age of the patients with the talar lesion was 36 (42-22) years. All patients were treated according to a pre-treatment protocol, which included clinical preoperative examination, MRI examination and radiological classification of the joints through regular f/p radiographs. All patients were treated with diagnostic arthroscopy and healthy cartilage biopsies which were sent for chondrocyte extraction and cultivation. After an average time of 2.5 weeks, the patients were treated with an arthrotomy, subchondral bone sclerosis debridement, and autologous chondrocyte transplantation. The postoperative treatment for all patients included nonweight bearing for 6-7 weeks, physiotherapie, and continuous passive motion. All patients were examined clinically in 3-weekly intervals, radiologically through a MRI at the, 6^th and 12^th month postoperatively. Results: For the postoperative evaluation we used the criteria of Hughston and ICRS for the knee joint and the evaluation score of Finsen for the ankle joint. According to the Hughston evaluation score 4 patients with OD of the knee joint showed an excellent postoperative score rated with a 4 and 2 showed a good result rated with a 3. (0 = failure, 4 = excellent). According to the ICRS score 4 patients showed a normal joint function and 1 a nearly normal one. According to the evaluation score of Finsen for the ankle joint, all patients showed excellent results with an average score of 1, 17. (0 = excellent, 4 = failure). We encountered no complications in the postoperative phase in any of our cases. The arthroscopical reexamination of 2 patients at the 6^th pop month showed the existence of cartilage like tissue with complete coverage of the chondral defect. Biopsy of this tissue showed the formation of connective tissue under cartilaginous metaplasia. Discussion: Taking into account the previous treatment methods of osteochondritis dissecans and their results reported in the international literature it is attentions worthy to notice the poor treatment outcome by the 4th grade of cartilage affection. The application of the autologous chondrocyte transplantation method enables a reconstructive intervention in the inevitable degradation of the local articular cartilage through the offering of new chondrocytes, subchondral blood supply and articular surface. Inspite the subchondral bone breakthrough our clinical results were excellent with painfree full weight bearing, full motion ability and no sign of joint irritation. Conclusion: Further investigation of this promising method of treatment is necessary, in order to be able to understand the factors influencing the results, for a wider application in the treatment of osteochondritis dissecans of the knee and ankle joint.

Philip A. Davidson, Dennis W. Rivenburgh St. Petersburg, Florida, USA

Purpose: This study describes the short and medium term results of osteochondral autogenous transfer (OATS) for chondral resurfacing of full thickness articular surface defects. Methods: Between October 1996 and the present, 43 full thickness chondral defects of the knee, shoulder and ankle were treated with OATS. Treatment method was determined by lesion size, location, geometry and graft availability. All procedures were performed by one surgeon using similar technique. All patients had preoperative evaluation and radiographs and were followed at regular postoperative intervals. Results: A series of 43 patients, entailing 70 plugs were treated with OATS grafts. In this series, 39 were knee cases, 2 shoulders and 2 ankles. OATS patients were followed up to 33 months with a mean of 20 months. Mean size for the lesions was 1.13 cm². Follow up was obtained on all patients. A modified Hospital for Special Surgery knee rating scale and a modified Cincinnati Scale were utilized. The mean Cincinnati score postoperatively was 6.33 ranging between 2 and 10. There were 32 excellent, 6 good, 3 fair and 2 poor results, according to the Hospital for Special Surgery scale. The fair and poor results were associated with significant arthrosis in areas of the knee other than the OATS site. Conclusions: The OATS procedure represents an effective means to treat full thickness symptomatic articular surface defects. Early clinical data indicates that this procedure offers an excellent alternative to restore function and minimize pain in these patients. Failures in this series were associated with multifocal arthrosis and suggest the need to further define surgical indications.

Weiler Andréas, Südkamp Norbort Berlin, Germany

Introduction: In recent years the treatment of fractures has evolved from extensive open reduction to internal fixation to minimally invasive surgery. The goal of treatment for tibial plateau fractures is to restore joint congruity and alignment by stable stabilization to allow early weight bearing and prevent late development of degenerative osteoarthritis. The goal of this prospective study is to inspect the almost standardized technique of minimal invasive surgery in stabilization of tibial plateau fractures based on the follow-up data of a large number of patients. Methods: A series of 40 cases is presented (AO/ASIF classification: 2 B1, 23 B2 15 B3). The mean follow-up period from time of the surgery performed between 94/12 and 06/97 was 2, 5 years (11-34 month). The mean age of the 21 female and 20 male patients was 49 years (24-84). There were 78% isolated fractures of the tibial plateau. Associated injuries with 16% meniscal tears, 7% MCL/LCL lesions, 19% cruciate ligament injury were seen in 14% polytraumatized patients. In 27 cases Hydroxyapatite was used. Results: The outcome was assessed using the scores by Rasmussen, Lysholm and Delamarter. In the Delamarter-Score (subj:obj 30:70%) we found in 33 cases (82.5%) excellent and in 4 cases (10%) good results. In the Lysholm-Score (subj:obj 95:5%) 37 patients were evaluated as excellent and good, 3 patients had a poor result. The Rasmussen-Score showed excellent results in 36 cases and 4 patients had a good result. Discussion and Conclusion: The arthroscopic reduction and percutaneous fixation (ARPF) described first by Jennings in 1985 with the well known advantages of direct visualization of the fracture and ist subsequent reduction without a formal arthrotomy or detachment of the meniscus showed initially good results. This technique seems to achieve a better prognosis in the treatment of tibial plateau fractures. In the three different scores the follow-up outcome achieved subjectiv and objectiv good and excellent results in 90%. In no case occured complications like infection, secondary wound healing problems, thrombosis or lung embolism. Radiography showed no loss of reduction and no osteoarthrosis. ARPF is an excellent treatment option for B-type fractures of the tibial plateau with much lesser complications and better results than the ORIF.

Sohail Ahmad, Robert Nirschl, Leigh Brezenoff, David Gonzalez Arlington, Virginia, USA

Introduction: Numerous studies have evaluated the efficacy of knee arthroscopy in the treatment of osteoarthritis. Treatments have varied from simple lavage to extensive debridements, with mixed results. Our purpose is to evaluate the effect of arthroscopic notchplasty and removal of impinging lesions with the goal of regaining lost knee extension, thereby reducing contact forces in the patellofemoral joint. Methods: 64 patients with arthritic symptoms in the knee refractory to conservative treatment underwent knee arthroscopy. Treatment included debridement of meniscal lesions, thermal or mechanical stabilization of chondral defects, removal of impinging osteophytes, and notchplasty. The primary goal of surgery was to regain knee extension. Subchondral stimulation of fibrocartilage with drilling or microfractures was not performed. Age, duration & severity of symptoms, pre & post surgical HSS scores (52 total), compartments involved, motion, impinging lesions, & level of satisfaction were evaluated. Patients with less than 5 years of follow up were excluded. Results: 36 patients (mean age 64.8) were available for a minimum 5 year follow up. At 2 years, 32/36 were satisfied. Mean HSS scores improved from 29.2 to 48.0. All four early failures had tricompartment disease, mean pre-op HSS scores = 21.5, and mean pre-op flexion contractures = 15.6 degrees. At 2 years, all had reformed their flexion contractures. All four underwent knee replacements at a mean 4.5 years. At 5 years, mean HSS score improvement was from 29.2 to 43.2. 25/36 were satisfied with good to excellent results. 3/36 had some deterioration and were rated fair. The 8/36 failures underwent 7 TKA and 1 HTO. The 25/36 with satisfactory results had a mean pre-op flexion contracture of 7.3 degrees and average HSS scores of 33.2, whereas the other 11/36 had mean contractures of 15.0 degrees and average HSS scores of 20.1. Poor results were associated with contractures greater than 10 degrees (P =.05) and lower pre-op HSS scores (P =.05 Wilcoxon Rank Sum). Discussion and Conclusion: In our study, regaining and maintaining knee extension and thereby unloading the knee joint was prognostic of improved outcomes. Arthroscopy notchplasty and removal of impinging tibialosteophytes to regain knee extension increased the longevity and efficacy of our treatment of the osteoartritic knee.

Jeffrey L. Halbrecht San Francisco, California, USA

Introduction: Many methods have been described for realigning the patella. Most techniques for proximal realignment involve an open medial reefing or VMO advancement. Arthroscopically assisted methods have been described, but these all involve the need for a medial incision to tie sutures. We describe an entirely arthroscopic technique for proximal realignment that eliminates the need for an incision. Methods: Over a five year period, 35 patients with patella instability and malalignment were treated with an outpatient arthroscopic “all inside” medial reefing and lateral release. Arthroscopic reefing was performed by percutaneous passage of suture followed by arthroscopic retrieval and knot tying inside the joint. Prior to knot tying, a healing response was initiated along the medial retinanculum by either gentle shaving with a whisker blade, or by radiofrequency thermal shrinkage. Rehabilitation consisted of 1-2 weeks of bracing followed by an accelerated program of ROM and VMO strengthening. Results: Postoperatively, patients were evaluated by a patella scoring system developed by the author. Pre and postoperative x-rays were measured for congruence angle and lateral patellofemoral angle. No patients had a recurrence of subluxation or dislocation. All patients reported good/excellent results, and 100% felt they would undergo the procedure again based on their results. Congruence angle and lateral patellofemoral angle were significantly improved. Patients reported a significant improvement in pain, swelling, stair climbing, crepitus and return to sports. There were no reoperations and no complications. Conclusion: Arthroscopic patella realignment is a viable technique that offers results comparable or superior to published data for open or arthroscopically assisted realignment.

James J. Wyman, Jamie Galvin, Michael J. Belanger Providence, Rhode Island, USA

Purpose: The use of the quadriceps tendon as a tendon only graft harvest site for ACL reconstruction has been proposed. We studied the anatomic dimensions of the quadriceps tendon to better define its adequacy as a graft source for ACL reconstruction. Methods: 38 patients undergoing diagnostic Magnetic Resonance Imaging (MRI) of their knee underwent a modified MRI protocol to specifically evaluate their quadriceps tendon. T1 weighted images in the sagittal, coronal and axial planes were obtained to define the anatomic dimensions of the quadriceps tendons. Length, width and thickness measurements were obtained at the proximal (origin) and distal (insertion) extent of the tendons. Results: The mean length of the quadriceps tendons measured 8.12 cm (range 6.60 to 10.10 cm). The mean width and thickness at the distal extent of the tendon (patella insertion) was 2.92 cm (range 3.60 to 2.65 cm) and 0.92 cm (range 1.00 to 0.85 cm) respectively. Proximally (origin of the tendon), the mean width was 1.87 cm (range 2.15 to 1.30 cm) and thickness 0.79 cm (range 0.90 to 0.59 cm). Conclusion: Harvesting quadricepts tendon-only grafts for ACL reconstruction has been described. The minimum length of harvested graft needed to provide an adequate ACL construct has been reported to be 7.0-7.5 cm (2-2.5 cm × 2 for fixation into bone tunnels and 2.5 cm for intra-articular excursion of the graft). In this study population, the mean quadriceps tendon dimensions was sufficient to allow an adequate tendon-only graft to be harvested for ACL reconstruction. However, substantial individual variation was noted in the anatomic dimensions of the quadriceps tendon with individuals having insufficient length, width or thickness to reliably allow the harvest of an appropriately sized tendon-only graft for ACL reconstruction.

Craig L. Leutz Rockville Centre, New York, USA

Introduction: Anterior knee pain has been a well documented problem source following ACL reconstruction with patella tendon autograft. The incidences in the literature ranges from 1 to 10%. Because of this complication, many surgeons have begun using alternate graft sources; such as, hamstrings or allograft that do not involve the extensor mechanism. However, most of these studies associating anterior knee pain following ACL reconstruction with patella tendon harvest, were performed prior to new fixation techniques that allow harvesting of only a miniature bone plug. We sought to determine if anterior knee pain could be avoided with the mini patella bone plug technique. Materials and Methods: 30 ACL reconstructions were retrospectively reviewed. The patients were divided into two groups based on whether or not they had achieved full range of motion prior to ACL reconstruction. 28 had achieved full motion and were in group one and group two consisted of the two patients who had not achieved full motion prior to ACL reconstruction. All patients received six weeks of physical therapy prior to ACL reconstruction. The mini patella bone graft technique involved harvesting a patella bone plug with a length of between 10 and 14 millimeters as opposed to standard techniques involving 20 to 25 millimeters. A 20 millimeter bone plug was harvested from the tibia and closure of the patella tendon defect on the peritenon was performed. The tibia bone plug was secured in the femur with an interference screw and the patella bone plug was tied around a customed designed screw and post after 2 sutures were passed through the bone plug and 1 through the tendon. The same rehabilitation protocol was used throughout the study. The patella bone block was back cut to avoid entry into the joint. In addition, the patella defect was also bone grafted. Results: All patients in group one had a full range of motion as compared to the opposite side by 3 months. All returned to full activities; including, sports by 4 months. There was no incidence of anterior knee pain in group one. In group two, the two patients had difficulty regaining their preoperative motion and were 5 degrees short of full motion at three months. Both returned to full activities; including, sports by 4 months. Both had moderate complaints of patellofemoral knee pain that responded to a Cortisone injection. There was no instability in either group. Mean follow-up was one year. Discussion: Harvesting a mini bone patella bone plug is now possible with new fixation techniques such as, the custom screw and plate used in the study. This combined with back cutting of the patella block and bone grafting the patella defect produces a reliable technique of bone patella bone reconstruction while avoiding patellofemoral pain entirely. However, those patients that have difficulty with obtaining normal preoperative range of motion, maybe at increased risk to patellofemoral pain with any type of bone patella bone reconstruction. These maybe the best candidates for an alternative graft source.

Angelo J. Colosimo, Paul F. Carroll, Robert S. Heidt, Jr., Richelle L. Carlomas Cincinnati, Ohio, USA

We evaluated the results of eleven patients who underwent simultaneous ACL/PCL reconstruction for knee dislocations. The average age at surgery was 29.2 years and the average postoperative follow-up was 28.6 months. Seven patients were treated with a bone-patella tendon-bone autograft for the ACL reconstruction and an Achilles allograft for the PCL reconstruction, three patients with ipsilateral/contralateral bone-patella tendon-bone autografts and one patient with bone-patella tendon-bone and Achilles allografts. Seven of these patients were operated on acutely (<3 months post injury), while 4 were treated for chronic conditions. All patients underwent an aggressive well-structured postoperative rehab protocol. At follow-up, all patients were subjectively and functionally evaluated with a Lysholm Knee Scoring Scale. The average Lysholm Score was 88.4 (range 44-100). Four of the eleven complained of pain at times, two complained of swelling, and four described a sense of infrequent instability. None of the patients required a postoperative manipulation to regain their range of motion, and ten of the eleven returned to their previous activity level postoperatively. Nine of eleven were able to return for a follow-up visit and therefore also evaluated with a KT-1000 and clinical exam. Of these nine patients, two demonstrated a loss of flexion >5 degrees compared to the other side, seven demonstrated a positive posterior drawer at 90 degrees (6 patients at 1+, 1 patient at 2+), three demonstrated a positive Lachman (1+), while no patients demonstrated any posterior lateral corner instability. The average anterior active difference on the KT-1000 was 2.6 (range 0-6). These results indicate that simultaneous reconstruction of ACL/PCL is technically feasible and can yield good to excellent functional results. Based on these findings, we feel the use of allograft yields similar results to autograft and incurs less operative time and morbidity. Also, aggressive postoperative rehabilitation is imperative to prevent postoperative adhesions.

Richard L. Levitt, Theodore Malinin Miami, Florida, USA

Introduction: The functional importance of the ACL in knee joint stability is well recognized. For this reason, ACL replacements, primarily with bone-patellar ligament-bone or Achilles tendon allografts are frequently performed. Incorporation and healing of these grafts have been studied by many authors on biopsy specimens. The shortcoming of such studies is the possibility of sampling errors. To overcome this, we have studied 6 complete ACL replacement allografts retrieved at autopsy and at surgery. Methods: Six complete ACL replacement allograft were retrieved 14 days, 2, 3, 7, 8, and 10 years post transplantation. Two were retrieved at autopsy from patients dying of unrelated causes and 4 during knee arthroplasties. In surgical specimens the allografts were intac. These were studied by light and electron microscopy. Results: Transplanted ligaments showed different patterns of vascular ingrowth and cell replacement in the bone tunnels compared to the articular space. In 2 and 3 year specimens significant portions of the grafts showed neither vascular nor cellular ingrowth. The process of vascular and cellular ingrowth in the bone tunnels lagged behind that in the articular space. Discussion and Conclusion: Study of retrieved whole ACL replacement allografts showed a non-uniform pattern of vascular ingrowth and replacement of the graft fibroblasts by those of host. The process appears to take longer than reported by other investigators. Allograft replacement was slower in the bone tunnels than in the joint. The latter findings may explain the usual site of the allograft failure at the articular space-bone interface.

Andreas Weller, Frank Unterhauser, Carsten Beck, Gunnar Peters, Jürgen Mäurer Munich, Germany

Introduction: The remodeling process of a free tendon graft for ACL reconstruction has been extensively studied in the past. In different animal models a severe decrease of tensile strength due to tissue necrosis and subsequent revascularization has been described. These observations in animals lead to the recommendation to protect an ACL graft in humans to avoid graft failure or elongation during rehabilitation. The revascularization has been considered to be an integral step during graft remodeling. Several different experimental techniques have been used which just allow to determine the synovial blood flow or the “macroscopic” status of revascularization (Spalteholz, Laser-doppler-flowmetry, microangiography). However, data on the endoligamentous recapillarization does not exist. Biopsies of human graft tissue and serial MRI scans lead to the hypothesis that the intense process of remodeling and revascularization found in animals may not exist in humans (Howell AJSM 1995, Johnson Arthroscopy 1993). We therefore studied the recapillarization by immunohistochemical staining in an animal model combined with MR imaging prior to sacrifice. Methods: 30 sheep underwent ACL replacement with an autologous Achilles tendon split graft and were sacrificed after 6, 12, 24, 52 and 104 wks. The midsubstance portion of the graft was biopsied and immunostained for endothelial and adventitial cells using factor VIII anti-human rabbit antibody and α-smooth-muscle-actin anti-human mouse antibody. Prior to sacrifice oblique sagittal MRI scans (T1 & T2, 1.5 Tesla) were taken with and without the intravenous administration of gadolinium-DTPA using a high resolution eye-coil. Results: At 6 wks there was a strong immunoreactivity of the synovial cover sheet and already some central graft capillarization with some necrotic areas. MRI revealed an isointense graft with some hypointense signal areas. Gd-DTPA lead to a marked synovial enhancement. At 12 wks the synovial vascularity decreased and the central part of the graft was completely and homogeneously recapillarized. MRI showed an hyper- to isointense graft signal with mainly intraligamentous Gd-DTPA enhancement. At 24 wks vascularity further decreased and vessels were to be found mainly between the newly developed collagen bundles. Vascularity patterns were already comparable to those of the native ACL. In MRI the graft appeared iso- to hypointense with only minor intraligamentous Gd-DTPA enhancement. At 52 and 104 wks the graft showed regular hypovascularity and collagen bundle alignment. In MRI the graft appeared hypointense at 52 wks and signal intensity decreased further at 104 wks. There was no intraligamentous Gd-DTPA enhancement at 52 and 104 wks. Discussion: We were able to demonstrate that a central graft revascularization is already present at 6 wks which is in contrast to previous studies describing a time frame up to 20 wks (Amoczky JBJS 1982). This may be due to the immunohistochemical technique, which allows to detect cells of vascular origin earlier in contrast to conventional experimental techniques. We further found that the Gd-DTPA enhancement in MRI clearly reflects graft vascularity during remodeling and may thus present an appropriate tool to monitor graft vascularization also in humans. Graft signal intensity in MRI appears to reflect the graft's mechanical properties as previously determined (Weiler ORS 1999). We therefore believe that MRI signal intensity measures may be useful to determine the graft's mechanical properties also in humans. The typical low signal intensity found in the native ACL and in the early remodeling stage of human ACL grafts (Howell AJSM 1995) were found late in the present study (1 year). This finding underlines the hypothesis that the intense graft remodeling with decrease of mechanical properties found in animal models may not be present in humans.

Christian D. Bentley, Charles Bush-Joseph, Bernard R. Bach, Michael DiDonna Chicago, Illinois, USA

Introduction: Graft choice is a key decision in surgical management of failed anterior cruciate ligament (ACL) reconstruction. Non-irradiated bone-patellar tendon-bone (NIBPTB) allograft offers the advantages of decreased surgical morbidity, increased fixation strength, and maintenance of collagen integrity. Senior authors use NIBPTB allograft in revision ACL surgery when ipsilateral BPTB has been previously harvested. Materials/Methods: Between 3/90-3/97, 22 consecutive ACL revisions were performed for failed primary autograft BPTB ACL reconstructions. Average age at revision was 30 with 14 males and 8 females. Mean interval from primary to revision surgery was 3.8 years. Patients had an average of 2 previous procedures. All patients had subjective instability and a positive pivot shift on exam under anesthesia. Mode of failure was defined and classified. Results: Prior to revision, tunnel malposition was noted in 16/22 patients, notch stenosis in 3/22, and graft laxity in 3/22. At revision, Grade 3 chondromalacia was noted in 40% of patients and 35% required additional meniscal procedures. Physical exam tests performed on 20/22 patients available at an average 2.3 years (range, 1.5-5 years) post-op demonstrated significant improvements and in 85% the pivot shift was eliminated. Maximum manual side-to-side differences on KT-1000 exam decreased from an average of 7.1 mm to 2.9 mm. Seventy-eight percent received an excellent or good result on the HSS scale as well as Lysholm and Noyes functional scales. Tegner scale results tended towards preinjury scores. No significant difference in functional testing in these patients was noted. Conclusion: Use of NIBPTB allograft for revision after primary ACL reconstruction with BPTB autograft provides acceptable results.

Dexter Walcott, Walter Shelton, Anna Laura Bomboy Jackson, Missouri, USA

Introduction: Patellar tendon (PT) autograft has been considered the gold-standard for use in reconstruction of the anterior cruciate ligament (ACL). Donor-site morbidity, in the form of anterior knee pain and slower rehabilitation, has contributed to the enthusiasm for alternative graft sources, most commonly the hamstring tendons. Quadriceps tendon (QT) autograft is another alternative graft, however, little has been written about the results of ACL reconstruction using this tissue. The purpose of this study is to compare the results of ACL reconstruction with PT and QT autografts at two-year follow-up. Methods: Sixty patients underwent ACL reconstruction with PT autograft and 60 patients with QT autograft, all by one surgeon (WRS). The grafts were fixed with interference screws at the joint level on both the femur and the tibia. Concomitant meniscal pathology was treated with meniscal repair (26 in the PT group and 23 in the QT group) or partial meniscectomy (34 in the PT group and 23 in the QT group). An identical, aggressive postoperative rehabilitation program was used in both groups. The patients were examined at 3, 6, 12, and 24 months postoperatively. Range of motion, Lachman, pivot-shift, and KT-1000 arthrometer testing were performed. Patients were questioned about the presence of anterior knee pain. Results: The PT group showed an average extension loss of 1.1 degrees (range 0-7) and flexion loss of 0.6 (range 0-10). The QT group demonstrated an average extension loss of 0.98 (range 0-5) and 0 degrees of flexion loss. Fifty-four patients in the PT group had a negative Lachman, 4 had 1+ Lachman and 2 had a 2+ Lachman. Fifty patients in the QT group had a negative Lachman, 7 had 1+ Lachman and 3 had 2+ Lachman. There were 5 patients in the PT group that had a positive glide on pivot-shift testing and 8 with a glide in the QT group. KT-1000 arthrometer testing averaged 2.47 mm (−2 to 6) in the PT group and 1.78 mm (−1 to 5) in the QT group. There were no significant differences in any of these values. Thirty-two patients in the PT group and 6 patients in the QT group reported anterior knee pain. Conclusions: The quadriceps tendon graft may be successfully used for reconstruction of the anterior cruciate ligament. In this study, there were no significant differences in any of the measured parameters, when compared to patellar tendon grafts at two-year follow-up. In addition, there were less cases of anterior knee pain in the quadriceps tendon group.

David J. Deehan, Lucy J. Salmon, Vivianne J. Webb, Leo A. Pinczewski Crows Nest, Australia

Introduction: The purpose of this study was to evaluate the medium term, longitudinal results of isolated endoscopic anterior cruciate ligament reconstruction with patellar tendon autograft and interference screw fixation. Methods: 333 patients undergoing endoscopic ACL reconstruction over a 15 month period by the senior surgeon using patellar tendon autograft and interference screw fixation were evaluated prospectively. Patients were excluded if they had an associated ligament injury, chondral damage, previous meniscectomy, excision >1/3 of one meniscus, an abnormal radiograph or contralateral knee joint. Ninety patients met the inclusion criteria and were reviewed annually for a 5 year period. Assessment included Lysholm knee score, IKDC Standard Evaluation, thigh atrophy, instrumented athrometer testing, kneeling pain and radiological assessment. Results: During the review period 3 patients suffered a graft rupture. At 5 years 80 of the remaining 87 patients (92%) were assessed. Using the International Knee Documentation Committee (IKDC) assessment, 99% reported their knee as being normal or nearly normal and a median Lysholm knee score of 96 (range 64 to 100) was found. A 5 years 98% of the group exhibited a grade 0 Pivot shift test and 90% a grade 0 Lachman test. Accordingly, 93% of patients displayed a side to side difference of <3 mm on instrumented testing. There was a significant increase in the number of patients exhibiting a fixed flexion deficit from 12% at 1 year to 31% at 5 years (P =.001). The proportion of patients participating in competitive or vigorous activities fell from 86% at 2 year review to 66% at the 5 year review (P =.001). Kneeling pain was reported in 31% at 2 years and 44% at 5 years. Radiological examination found grade B changes in 2 out of 65 patients who consented to radiological examination. Discussion: This study supports the view that ACL reconstruction with patellar tendon autograft affords lasting stability, excellent subjective results and protection against degenerative change over a 5 year period. The majority of patients are still participating in vigorous sports at 5 years. Unfortunately, kneeling pain remains a problem for a large number of patients. Worringly, we have noted an increasing incidence of fixed flexion deformity over time. The relevance of this finding will only be revealed with ongoing surveillance.

Robin Fuchs, William Wheatley, John W. Uribe, Keith S. Hechtman, John E. Zvijac Coral Gables, Florida, USA

Introduction: Tears of the anterior cruciate ligament (ACL) are more prevalent in younger patients as they are participating in high impact sports at an earlier age. Non-operative treatment of these injuries until skeletal maturity may lead to continued instability, inability to return to sports, and subsequent meniscal and chondral injury. There are many surgical treatment options for the torn ACL in the skeletally immature patient including primary repair, extra-articular or intra-articular procedures. Reports on primary repair and extra-articular procedures have shown little success at providing long-term stability after complete midsubstance tears. However, fear of violating the physeal plate has limited the use of standard ACL reconstruction technique. The purpose of this study is to report our results with intra-articular ACL reconstruction with transphyseal drilling using bone-patellar tendon-bone allograft in the skeletally immature patient. Materials and Methods: Ten skeletally immature athletes underwent transphyseal intra-articular reconstruction from December 1995 to January 1999. Six of the patients were male and four were females. The patients ranged from nine to fifteen years of age with a mean of 13 years at the time of surgery. All ten patients had radiographic evidence of open physes and arthroscopic evidence of complete midsubstance tears of the ACL. All patients underwent intra-articular ACL reconstruction using tibial and femoral bone tunnels through the proximal tibial and distal femoral physes. Bone-patellar tendon-bone allograft was used in all cases. Special attention was made in placing the bone plugs and transfixing screws into metaphyseal bone so that only patellar tendon crossed the physes. Clinical outcomes were measured using the scoring system of Lysholm and Gillquist and the International Knee Documentation Committee (IKDC). Follow-up radiographs were taken to assess for early physeal arrest and clinical examination for limb-length discrepancy was performed. In addition, objective evaluation of anterior-posterior displacement was performed using a KT-1000 arthrometer. Results: The mean follow-up was 21.2 months. Eight out of ten patients have gone to skeletal maturity. The average Lysholm knee score was 93 (out of 100). According to the IKDC scores, four (40%) patients had normal function, five (50%) patients had nearly normal function, and one (10%) patient reported abnormal function by subjective criteria. There was no clinical or objective evidence of instability and all patients returned to sports. No patients had limb-length discrepancy or early physeal arrest. On KT-1000 testing, eight patients had less than 3 mm of increased anterior-posterior displacement (normal knee minus reconstructed knee) and two patients had between 3 and 5 mm. Discussion and Conclusion: Our results suggest that skeletally immature athletes with symptomatic complete ACL tears who do not desire to modify their athletic activity until skeletal maturity may benefit from intra-articular ACL reconstruction using patellar tendon allograft. Special attention to bone plug and transfixing screw placement into metaphyseal bone will reduce the risk of early physeal arrest. In our study, nine out of ten patients (90%) had normal or nearly normal knee function. None had evidence of growth disturbance, varus or valgus angulation, or focal physeal arrest.

Aasha K. Sinha, Christopher C. Kaeding, Joseph Hanna Columbus, Ohio, USA

Introduction: Anterior cruciate ligament (ACL) reconstruction has become a widely accepted treatment for functional knee instability due to ACL injury. Despite a typically good outcome, recurrent pathologic laxity does occur in a small subset of patients and a revision procedure may be indicated. Revision ACL reconstruction is more technically challenging than the primary procedure, and the outcome is less certain. While postoperative improvements in function have been described, few studies have compared results with those obtained after primary reconstruction. Therefore, the current study was designed to evaluate the outcome for patients that underwent a revision ACL reconstruction at our institution and to compare the results with those of patients undergoing a primary procedure. Methods: The records of all patients undergoing ACL reconstruction from 1992 to 1999 at our institution were reviewed. Patients undergoing a revision ACL reconstruction with no additional prior knee surgeries, and no significant associated ligament injuries, were selected for study. The clinical features of these patients were reviewed and the status of the articular cartilage at the time of revision surgery was recorded from the operative notes. Outcome was assessed by completion of the Cincinnati Knee Scoring System on average 40 months after surgery. These results were compared with those of a control group of 62 consecutive patients that underwent uncomplicated primary ACL reconstruction at our institution within the study period and were similarly followed up least 24 months after surgery. Results: 1075 ACL reconstructions were performed at our institution from 1992 to 1999. 39 of these patients underwent a revision procedure for ACL injury and met the selection criteria for study. Mean age of these patients was 25.7 years (range 17-31), 75% male, 25% female. Average time interval between primary and revision procedures was 48 months (range 6-132 mo). At revision, 67% of patients had ≥grade III articular cartilage changes in one or more compartments. At follow up, mean Cincinnati Knee Scale was 8.9 (range 7.2-10), mean patient symptom rating 7.3 (range 5-10). For the primary reconstruction group mean Cincinnati Knee Scale was 9.3 (range 7.2-10), mean patient symptom rating 8.8 (range 4-10). Cincinnati Sports Activity Scale for the revision patients indicated mean activity level prior to original injury of 98.8 (range 85-100), after primary reconstruction mean reported level achieved was 89.5 (range 75-100), and after revision 87.9 (range 65-100). For the primary reconstruction patients, mean pre-injury activity level was 93.5 (range 65-100), mean post-surgery level was 86.9 (range 40-100). For the revision group, 50% of patients decreased their activity level after revision due to knee related symptoms, versus 18% of patients after primary reconstruction. Conclusions: Our results indicate that revision ACL reconstruction may result in a satisfactory outcome for the majority of patients. Revision patients returned to a comparable level of activity versus the primary reconstruction patients. However, a decrease in postoperative exercise level due to knee related symptoms was significantly more common for patients undergoing a revision procedure.

Christopher C. Kaeding, Aasha K. Sinha, Scott Van Steyn Columbus, Ohio, USA

Extension loss in the knee may result in significant morbidity, especially in athletes. The cyclops syndrome is a well recognized complication of ACL reconstruction. In this syndrome, there is the development of a fibrovascular proliferative mass in the intercondylar notch, of uncertain etiology. This results in a postoperative loss of full extension, as well as an audible “clunk” with manipulation. The incidence has been reported as anywhere from 1-9.8%, with patients generally responding well to arthroscopic debridement and manipulation. The purpose of this study is to report the clinical features, treatment, and outcome for 6 patients with partial or complete ACL injuries, that had not undergone reconstruction, that developed pain and a loss of terminal extension, and were found to have a cyclops-like lesion at arthroscopy. We have termed this phenomenon the “pseudo-cyclops” lesion. The records of 6 patients were reviewed. All patients had an ACL injury documented on clinical examination and MRI. Four patients had complete ACL tears and 2 patients had partial tears. Five patients had chronic ACL deficient knees, presenting 5-25 years after the original injury, while one patient presented acutely, 2 months after ACL injury. Patients ranged in age from 19-58 years (mean 39.7 years). Five patients were male, one was female. All patients complained of a relatively recent loss of full knee extension and anterior knee pain with active and passive terminal extension. Most of the patients had previously participated in sports as well as other activities of daily living without significant debilitating symptoms related to their ACL injury. On physical examination, all patients had loss of extension of 5°-15°, as well as pain on terminal extension. All patients underwent arthroscopic evaluation and treatment of their symptoms. A standard three portal arthroscopic approach was made. This revealed a cyclops-like mass of fibrous tissue at the intercondylar notch, that appeared to originate from the base of the damaged ACL. The tissue mass was noted under arthroscopic visualization to result in notch impingement with terminal extension. After the diagnosis was made, all patients underwent debridement and removal of the tissue mass, using a motorized shaver. Improvement in knee extension was noted intraoperatively after debridement of the “pseudo-cyclops” lesion. Following surgery, all patients in this series experienced immediate marked improvement in active and passive knee extension without pain. We suggest that a “pseudo-cyclops” lesion, which may originate from the base of the damaged ACL, may be the cause of cyclops syndrome-like symptoms in patients that have not had an ACL reconstruction. We speculate that in a small number of patients the remnants of a damaged ACL do not undergo atrophy as expected, but rather, hypertrophy, resulting in impingement. As impingement develops, the resulting repetitive microtrauma and inflammation results in further hypertrophy and perpetuation of symptoms. We conclude that patients developing a “pseudo-cyclops” lesion, with pain and loss of terminal extension, respond well to arthroscopic debridement.

Robert A. Pedowitz, Debbie Popejoy San Diego, California, USA

Introduction: ACL reconstruction surgery requires isometric graft positioning and secure fixation. Though most surgeons tension the graft manually, some techniques utilize defined tension at the time of fixation. Few data are available regarding effects of intra-operative tension upon subsequent laxity in humans. Our goal was to prospectively compare intra-operative laxity with post-operative laxity after ACL reconstruction using KT1000 arthrometry. Methods: 57 patients (mean 29.2 years) with isolated, unilateral ACL injury underwent endoscopic ACL reconstruction by a single surgeon (49 patellar BLB, 8 hamstring grafts). Grafts were fixed rigidly under manual tension. Consistent, accelerated rehabilitation was used after surgery. KT1000 arthrometry was performed pre-op in clinic, pre-op under anesthesia, intra-operatively just after graft fixation, and six months after surgery. Results: Arthrometry data are presented in Table 1. There was excellent correlation (r =.86, P <.0001) of pre-op and post-op manual maximum translation in the non-injured knee (mean difference 0.5 mm), demonstrating reliable utilization of the KT1000 arthrometer. There was no significant correlation between intra-operative and six month post-operative laxity (r =.124, P >.4). There was no obvious difference in this observation between B-L-B and hamstring grafts. Five patients had laxity ≥5.0 mm six months after surgery. Intra-operative laxity was not remarkable in these patients (intra-op ManMaxΔ: −5.0, 0, 1.5, 0.5, 1.5 mm).

Discussion and Conclusion: A relatively narrow range of intra-operative laxity can be achieved without a tensiometer during ACL reconstruction. In this prospective study, ACL grafts that were at the extremes of initial laxity (on either the “tight” or the “loose” side of the range) did not specifically proceed to unacceptable laxity at 6 month arthrometry. This followup interval is appropriate because it precedes sports-related macro-trauma (in our rehab protocol), which might confound the findings. We hypothesize that other factors, such as inherent ligamentous laxity, extremity alignment, and biologic remodeling, are probably more important determinants of knee laxity after ACL reconstruction.

Nagarkatti Durgesh, J.P. Fulkerson, D. Tarinelli, Cin Nissen Farmington, Connecticut, USA

Introduction: Recent studies have questioned the stability of a bioabsorbable screw for free tendon fixation in ACL reconstruction. These models, however, have used human cadaveric bone of undefined density. The objective of this study was to establish the effect of cyclical loading on an “anchor” augmented free tendon fixed with bioabsorbable interference screw in bone of density comparable to young human bone. Methods: Extensor mechanism free tendon was harvested from 15 fresh frozen porcine knees (average bone density 1.3 gm/cm²). The tendon grafts were then fixed by one of three methods: 1. Biointerference screw with free tendon (8 mm tunnel, 8 mm screw); 2. Biointerference screw advanced to an 8 mm polylactic acid ball attached to the tendon graft (8.5 mm dilated tunnel, 9 mm screw); or 3. Biointerference screw advanced to an 8 mm disk of cortical bone attached to the tendon graft (8.5 mm dilated tunnel, 9 mm screw). Samples were mounted on an MTS machine, using an MTS extensometer to calculate graft displacement. Each specimen was preconditioned between 0 and 75 N (0.5 Hz, 250 cycles), and then cyclically loaded to 150 N (1 Hz, 5000 cycles). Tensile testing to failure was then performed on all specimens. Results: Placing an “anchor” (PLA ball or bone disk) diminished graft displacement substantially in the specimens tested. Mean graft displacement when a biointerference screw alone was used for fixation of the free tendon was 1.32 mm (n = 5; range: 0.03-3.31 mm). When a PLA ball was attached to the end of the free tendon, the mean graft displacement was 0.57 mm (n = 5; range: 0.01-2.45 mm). With a “cortical bone-disk anchor” augmentation, the mean graft displacement dropped to 0.21 mm (n = 5; range: 0.02-0.59 mm). Two of the specimens in Group I failed with cyclical loading. Ultimate loads at failure were 404 N, 545 N, and 465 N, respectively, for the three groups. Discussion and Conclusion: Providing an “anchor” (bone disk or PLA ball) substantially reduced slippage of a free tendon graft fixed by biointerference screw compression. Similar to earlier reports, the only failures of fixation were seen in the specimens fixed with biointerference screw alone (Group I). It is apparent that bone of adequate density is imperative to ensure adequate fixation by biointerference screw fixation of a free tendon graft for ACL reconstruction. Adding an “anchor” of cortical bone provides the most optimal fixation configuration and reduced slippage of the graft past the screw to a very low level (mean 0.21 mm) after 5,000 cycles at 150 N load. Clinical application of the “cortical bone-disk” augmentation has to date been met with excellent results in our practice.

Nicholas Harding, F. Alan Barber, Morley A. Herbert Plano, Texas, USA

Purpose: As shown by last year's O'Connor award paper, interference outlet fixation of ACL soft tissue grafts increases stiffness and enhances graft incorporation. However, concern about the potential for fixation slippage with interference screws alone has lead to backup fixation techniques. Does the secondary fixation provided by a sphere-interference screw combination increase the strength and reduce slippage for one type of soft tissue tendon graft? Methods: Using 40 fresh porcine knees, 10 mm patellar tendon grafts with attached patellae were harvested without a tibial bone plug and inserted into 9 mm tibial tunnels. This construct was created to test a nondoubled uniform tendon graft not unlike the quadriceps tendon. 20 were fixed with Bioscrews and 20 with Bioscrew/sphere combinations. Insertion torque was measured. These two fixation constructs were then tested with a straight line load to failure at a rate of 60 mm/min; without any cyclic loading or after cyclic loading for 500 cycles starting at 9.8 N (1 kg) and increasing to 186 N. All Bioscrews were 25 mm long and either 8 mm or 9 mm diameter depending upon graft-tunnel fit. The spheres were PLLA “pearls” with a central hole for sutures measuring 8 or 9 mm in diameter to complement the Bioscrew. The tendon construct was controlled by a Bunnell stitch of #2 braided polyester suture in the patellar tendon that then passed through the central hole of the sphere and was tied. Results: The maximum load for the Bioscrew alone averaged 372 N ± 134; and after cyclic loading 536 N ± 151. The maximum load for the Bioscrew-sphere averaged 458 N ± 98; and after cyclic loading 450 N ± 154. No statistical difference existed comparing the cyclical and destructive testing, or with and without the sphere. As expected, cyclic loading increased the stiffness of these constructs from about 46 N/mm to about 145 N/mm. Failure mode was by tendon shredding and pullout in 90%. There was a positive correlation between failure force and insertion torque (P =.02). Conclusion: The tendon grafts consistently failed by the sutures tearing out and the tendon pulling by the interference screw. The addition of a sphere had no measurable effect. The Bunnell type stitch was the weakest link and not an effective tissue holding stitch for this construct. These test results do not apply to looped hamstrings or soft tissue holding stitches other than the Bunnel stitch. As expected, cyclic loading increased stiffness with less stretching observed as the force built more quickly. “Plastic flow” resulting from cycling lead to greater energy absorption and a stiffer tendon. Higher insertion torques resulted in higher failure loads. Significance: For the non-looped soft tissue graft controlled by a Bunnell stitch, the addition of a sphere to the Bioscrew does not improve load to failure characteristics.

Mark S. Burman, William A. Grana Oklahoma City, Oklahoma, USA

With the recent rapidly growing trend towards the use of hamstring autografts in ACL reconstruction, there has been an associated increase in the development of various fixation devices for these grafts. One device, which has been quite popular, is the Endobutton (Accufex/Smith & Nephew) while there are a few studies which have shown good clinical results with this device, there is also some concern raised with its use. The term “Bungy Effect” has been used to refer to the potential micromotion of the graft construct that may occur within its bone tunnel due to the inherent elasticity of the construct and the distant fixation site. The concern is that this “Bungy Effect” could result in bone tunnel expansion, elongation of the graft construct and increased ACL laxity or functional instability.

Twenty-six knees (25 patients) were retrospectively reviewed, 18 females and 8 males with an average age of 24.1 years (13-43 years). Mean follow-up was 17.5 months (12-35 months). Patients were evaluated with x-rays, ligament laxity testing, Lysholm & Tegner scores. Mean tunnel expansion of 2.7 ± 1.3 mm (AP view) and 2.4 ± 1.1 mm (lateral view) was seen (P <.05). Mean total tunnel length was 62.7 ± 8.9 mm (47-79 mm). Mean side to side difference on instrumented laxity testing was 1.1 ± 1.6, with 88% of patients having good/excellent Lysholm scores. Mean Tegner score was 7.7 pre-op & 7.3 post-op, a difference of 0/4 (P =.03). Statistical correlation analysis (Pearson and Spearman correlation tests) failed t reveal a correlation between tunnel widening and any other variable. The results of this study reveal use of the Endobutton does result in femoral tunnel expansion, however this has no clinical effect on the overall outcome.

Matthew J. Matava, Jolie C. Holschen, Randall R. Odem St. Louis, Missouri, USA

Introduction: Previous studies have shown that females sustain a higher rate of anterior cruciate ligament (ACL) tears than males. Estrogen-induced structural changes in the ACL have been offered as one possible explanation. The primary objective of this study was to determine the relationship between anterior knee laxity and the estrogen:progestin ratio in females. A secondary objective was to ascertain whether anterior laxity is affected by the hormonal fluctuations present during the normal menstrual cycle. Materials and Methods: Fifty-nine varsity athletes participated in this study: 31 females and 28 males with an average age of 20 years (range, 18 to 22 years). To be eligible, females had to have a regular menstrual cycle (26 to 30 days), accurate recall of menstrual history, and nulligravida status. Males served as controls. All females not using oral contraceptives measured their basal body temperature each morning for one month with a B & D Digital Basal Thermometer to document ovulation. Subjects were divided into four groups based on ovulatory status and gender: Group I (n = 9) consisted of ovulating females who did not use hormonal contraception; Group II (n = 9) was comprised of non-ovulating females who also did not use contraception; Group III (n = 13) consisted of females who used some form of hormonal contraception for at least 6 months prior to the study; Group IV (n = 28) was comprised of the male subjects. Each subject underwent KT-1000 arthrometer testing to both knees at 15 lb., 20 lb., 30 lb., and at manual maximum force on days 1, 7, 14, and 21 of their menstrual cycle (Groups I, II, and III), or on the corresponding days of the month (Group IV). A mixed-model ANOVA was used for comparisons of the arthrometer data within and between groups for each testing session and over time. Results: Mean KT-1000 values for both knees over all force levels and all sessions were as follows: Group I, 7.1 ± 1.0 mm; Group II, 6.7 ± 1.7 mm; Group III, 5.8 ± 0.9 mm; and Group IV, 6.2 ± 1.0 mm. The values for Group I (ovulating females) were significantly higher than those for Group III (contraceptive users) and Group IV (males) (P =.0478). There was a significant increase in mean arthrometer values for all four groups with increasing force levels (P =.0125). Side-to-side differences were seen only in Group I subjects with right-side arthrometer values significantly greater than the left (P =.0001). The temporal differences among the groups were consistent at each testing session at all force levels. We found no significant differences between measurements taken at the testing sessions throughout the month in Group I (ovulating females). Conclusions: The results of this study suggest that anterior knee laxity is related to the serum estrogen:progestin ratio. The greatest degree of laxity was noted in ovulating females who exhibit monthly fluctuations in circulating estrogen and progestin levels, and who typically display the highest ratio. Hormonal contraceptives reduced the degree of knee laxity and are associated with a low ratio similar to males. We found no significant temporal relationship between the monthly hormonal fluctuations seen in ovulating females and anterior knee laxity.

Donald H. Johnson, Prem Parmar Ottawa, Ontario, Canada

Purpose: To document healing of the anterior cruciate ligament. Introduction: Conventional wisdom holds that the anterior cruciate ligament (ACL) does not heal. In an athlete the ACL deficient knee is likely to be symptomatic and lead to functional instability. This has led to the belief that ACL tears in the active athlete require reconstruction. Some ACL tears in recreational athletes are successfully treated conservatively with activity modification and bracing. A literature search failed to produce another study on ACL healing. Materials and Methods: At the Carleton University Sports Medicine Clinic we retrospectively reviewed ACL tears diagnosed by the Lachman, pivot shift, and KT-1000 arthrometer testing. Results: 11 patients were found to be asymptomatic and stable after an initial diagnosis of an ACL tear. In follow up the Lachman and pivot shift were normal and the KT-1000 was equal to the opposite side. The patients resumed activities without surgical intervention. The Clinical Implication: ACL tears should be treated conservatively for he first 6 weeks since in a small percentage of patients, the ACL tear can heal.

Young Bok Jung, Suk Kee Tae, Jae Kwang Yum, Kwang Sup Song Youngsan-Ku, Seoul, Korea

While the cruciate ligament has a profuse vascular response following injury, spontaneous repair does not occur. This may result from the fact that synovial fluid dilution of hematoma following an injury prevents the formation of a fibrin clot and thus the initiation of the healing mechanism. Another theory suggests that the dynamic nature of the fascicles of the anterior cruciate ligament (ACL) through small ranges of motion prohibit spontaneous union to these fibers. But we experienced five cases of spontaneous healing of completely injured ACL. Initially they, all of the injuries related to ski injury, showed anterior instability more than grade II, 5 to 8 mm difference by stress roentgenography (pull view) and 8 to 9 mm by KT-1000™ arthrometer between the injured and sound side. During the arthroscopic examination of these five cases, the injured ACLs were revealed the loss of tension and synovial bleeding, but the synovial membranes were relatively well covered. We have treated conservatively these cases with splinting for 2 weeks, followed by limited range of motion (<90°) with brace for 4 weeks. Among them, 3 cases were treated with arthroscopic microfracture technique and they showed excellent clinical results; no instability by pull view and KT-1000™ arthrometer and unlimited athletic activity at the average 29 months follow up.

Based on the above results, it is construed that the injured cruciate ligament with synovial coverage might have spontaneous healing potential in acute stage, probably by hindering of direct contact of synovial fluid with the injured cruciate ligament. Therefore conservative treatment seems to be enough in the cases of anterior cruciate ligament tear without discontinuing but synovial covering. But the long term follow up, animal studies and more cases of clinical trial will be needed.

Richard L. Layfield, III, Charles A. Gottlob Chicago, Illinois, USA

In order to quantify tunnel dilation, we retrospectively reviewed the postoperative, radiographs of 25 consecutive patients who underwent anterior cruciate ligament reconstruction using a four strand hamstring autograft fixed with the poly-L-lactic acid (PLLA) Bio-interference screw (Arthrex, Inc., Naples, FL). Tibial bone tunnels were evaluated by AP and lateral radiographs in each case. Radiographs were available at 1 week, 6 weeks, 6 months, and 1 year postoperatively. Maximum tunnel widths at 6 weeks, 6 months, and 1 year postoperatively were measured in a standard fashion by a single observer and compared to initial postoperative radiographs. At 6 weeks, the average tibial tunnel dilated.53 mm (range −.78-1.96). At 6 months, the average dilation measured 1.44 mm (range −.10-4.15). At 1 year, the average tunnel dilation measured 2.17 mm (range 1.34-3.10). When expressed as percentages, the tibial tunnel had dilated by 6.56%, 17.46% and 26.76% at 6 weeks, 6 months, and 1 year respectively. Tunnel dimensions significantly increased between the 6 week and 6 month radiographs and continued to increase between 6 months and 1 year. The degree of tunnel dilation reported here is comparable to previous studies that considered alternate graft sources and fixation methods for ACL reconstruction. This data also suggests that the osteolytic process may not be completed within the first postoperative year. Longer-term follow-up is necessary.

Jack M. Bert, Leslie Nace St. Paul, Minnesota, USA

Seven thousand nine hundred orthopedic cases performed in an in office surgicenter were reviewed over a five year period with respect to surgical case mix, types of anesthesia, and efficiency and financial considerations. 2,730 cases were performed by a pain clinic specialist (anesthesiologist) and consisted of epidural steroid injections, facet joint injections, and trigger point injections. Of the 5,170 surgical cases performed, 2,715 were knee arthroscopies, 930 carpal tunnel releases, and 595 rotator cuff repairs which accounted for 82% of the cases performed.

The mean total peri-operative time for a surgical procedure was 2.51 hours; the mean preoperative time was 72 minutes, mean surgical time was 41 minutes, and mean recovery time was 58 minutes.

Only 7.4% of the anesthetics performed were general anesthetics and 1.3% were spinals. The remainder of the anesthetics blocks were either interscalene, axillary or local blocks. 86% of the knee arthroscopies that were performed were successfully completed using local and intraarticular injections with marcaine plus epinephrine.

The physicians utilizing the surgicenter for the majority of their outpatient surgical cases experienced a noticeable decrease in hours worked during the course of their week averaging 8 hours for 4 of the group orthopedists. The average increase in physician net revenue was 18% per year per group member.

In office surgicenters can be a significant source of revenue for the orthopedic surgeon faced with a continued decline in professional reimbursements and well as a dramatic increase in individual physician operating efficiency.

James K. DeOrio, Anthony W. Ware, Julio Gonzalez, Ashley Vickers Jacksonville, Florida, USA

Introduction: We describe the “dry technique” modification of arthroscopic ankle fusion whereby the inflow is stopped prior to the motorized burring of the bony surfaces leaving the bone graft in position. Materials and Methods: We retrospectively reviewed 8 consecutive arthroscopic ankle fusions, the last 3 of which used the “dry technique” modification at our institution from November 1997 to April 1999. Six men and two women were evaluated. The mean age was 63 years (range 28-78). Five ankles had developed posttraumatic osteoarthritis, one ankle had rheumatoid arthritis, and one had idiopathic osteoarthritis. The average follow-up was 15.8 months (range 12-28 months). The indication for surgery was a painful arthritis tibiotalar joint with minimal deformity unresponsive to conservative therapy. Surgical Technique: The operative procedure is performed under general or regional anesthesia. After the extremity is prepped, draped, and exsanguinated, a tourniquet is inflated. Using sterile technique, the Acufex noninvasive ankle distraction device (Smith + Nephew Endoscopy, Andover, MA) is affixed to the operating table and applied to the ankle with longitudinal traction. Standard anterolateral and anteromedial portals are established using longitudinal incisions through the epidermis with blunt dissection to the level of the capsule using hemostats. A 30 degree, 4.0 mm arthroscope is used. Soft tissue debridement and meticulous removal of the cartilage from the tibial plafond, talar dome and medial/lateral talomalleolar areas is performed using small to medium-sized curettes and a motorized shaver. At this point, the inflow is stopped. A 5 mm burr is used to mill the bony surfaces which creates a bone slurry that is left in situ. As necessary, the arthroscope can be cleaned by a quick wipe against the soft tissue or removed and cleaned with gauze. Upon completion, the arthroscope and instruments are removed and the distraction released. The ankle is held in neutral dorsiflexion, approximately 5 degrees of external rotation and 5 degrees of valgus, and three 7.3 mm cannulated screws are inserted. Alignment and screw placement are evaluated using fluoroscopy. The incisions are approximated using 2-0 vicryl and 4-0 nylon and a standard sterile dressing is applied followed by a posterior splint. On postoperative day one, the dressings are changed and a short leg cast is placed. Sutures are removed at two weeks and a cast or removable boot is used for 6 weeks at which time full weightbearing is begun in a removable boot (Aircast Incorp. Summit NJ/USA) for 4 weeks. Weightbearing in standard shoes is begun at 10 weeks. Results: X-rays were taken at 6 weeks to insure no untoward events and again at 10 weeks. X-rays at 10 weeks demonstrated solid healing of all the ankle fusions. There were no nonunions or fibrous unions, no malpositions, and no complications. Surgery lasted approximately 90 minutes and all 8 patients were considered to have an excellent to good clinic result. No significant difference was noted between the wet and the dry technique. The wet technique was used initially in 5 patients and the dry technique used in the last 3 patients. However, retaining of autogenous bone graft material between the tibia and talus may allow the arthroscopic technique to be used in a patients with greater defects with the bone graft inserting itself into any defects in the tibiotalar region.

Marc J. Philippon Fort Lauderdale, Florida, USA

Introduction: Instability of the hip is a well known entity and can be a significant source of disability. Chronic instability can lead to premature arthritis and possibly predispose to tearing of the acetabular labrum. Previous authors have described open hip capsulorrhaphy as a treatment option for these patients. This paper discusses the outcome of arthroscopic hip thermal capsulorrhaphy for the treatment of idiopathic and traumatic hip instability. Material and Methods: Between August 1998 and August 1999 twelve hip arthroscopies (11 patients) were performed after failure to respond to a trial of conservative therapy after three months minimum for painful instability of the hip with gait disturbance. This group of patients consists of six traumatic and six idiopathic hip instability. There were eight females and three males, average age 31 (range 8-56 years) at the time of surgery. Pre-operative evaluation detailed arthroscopic findings and clinical outcome were recorded. All patients were treated with deflectable electro-thermal probes. Results: All patients had pre-operative vacuum signs and had intra-operative findings of redundant capsular tissue. Good visualization of the anterior and posterior capsules were obtained and direct visualization of capsular shrinkage was obtained. The goal of the procedure was to obtain a volumetric contraction of the entire thickness of the capsule to reduce capsular redundancy. 75% (9/12 hips) had associated torn labrum. These were treated with partial labrectomy. 100% of the patients with traumatic instability had an associated torn labrum and 50% of patients with idiopathic hip instability had an associated torn labrum. All patients had no signs of instability at their last follow-up exam (4 weeks to 1 year). 2 patients/2 hips had incomplete pain relief. 83% of the patients had significant improvement in their symptoms. Discussion and Conclusion: Arthroscopic thermal modification of collagen in the hip capsular tissue appears to be a treatment option for patients with hip instability. Short term results appear to be promising, however more studies are required to determine the long term efficacy of this procedure.

Chet J. Janecki Tampa, Florida, USA

Introduction: The purpose of this study is to summarize our experience with trochanter bursoscopy and iliotibial band release for recalcitrant trochanteric bursitis. Material: A retrospective review from 1992 through 1997 of 24 patients undergoing this procedure was done. There were 18 right and 17 left hips involved. The average age of the patients was 63.5 years (45 through 76 years). There were 16 females and 8 males. Average length of symptoms was 15 months (range 10 to 36 months). Fifteen patients had associated low back pain. Results: All patients underwent trochanteric bursoscopy and iliotibial band release through three arthroscopic portals. Preoperative visual analog scale determined that 23 of the 24 patients characterized their preoperative pain as 7 out of 10 or greater. Postoperatively, twenty patients characterized their pain as 0 out of 10; four patients as 5 out of 10 or less; twenty-two patients were satisfied and two were not. Average follow up was 20 months; range (18 to 36 months). Conclusion: Trochanteric bursoscopy with iliotibial band release for trochanter bursitis is a safe, effective and minimally invasive technique for the management of trochanteric bursitis. The techniquie is at least as effective as open procedures.

Michael Dienst, Michael Brang, Stefan Gödde, Dieter Kohn Homburg/Saar, Germany

Introduction: Intraoperative observations have shown that the combination of traction and distension is more effective for distraction of the hip joint than traction alone. The aim of this cadaver study was to quantitate the effect of traction alone and in combination with distension in different joint positions in order to improve positioning in hip arthroscopy. Methods: 6 hip joints were studied in 6 fresh cadavers. After rigor mortis was broken, the cadavers were placed supine on a traction table. An 8 mm twist drill was introduced centrally within the femoral head. Traction was applied in 6 different joint positions (flexion/abduction 0/0°, 0/20°, 20/0°, 20/20°, 40/0°, 40/20°) alone and in combination with distension of 40 ml of saline. 0 to 80 pounds of traction in steps of 10 were applied. Digital images were obtained by fluoroscopy and transfered to a PC. Points of reference on the acetabulum were selected on each radiograph. 5 distances between the femoral and acetabular subchondral bone were measured in standardized angles to the reference points. Thus, the joint was separated into different areas: weightbearing, central and medial compartment. Projection error was corrected by measurement of the twist drill. Results: The following data were analyzed for traction forces of clinical relevance between 40 and 60 pounds. For the weightbearing compartment, distraction due to traction alone ranged between −0.3 to 3 mm (mean 1.02), distraction due to traction plus distension ranged from 2 to 5.5 mm (mean 3.21). For traction alone less distraction was found for 40° of flexion. Significant differences for abduction and flexion between 0 and 20° were not found. For the combination with distension, more flexion and less abduction showed better results. For the central compartment, similar results were found: for traction alone, distraction ranged from 0.5 to 4.3 mm (mean 2.04), for traction plus distension from 2.6 to 6.3 mm (mean 4.41). Significant differences for the various joint positions were not found. For the medial compartment, distraction ranged from −1 to 1 mm, no differences of clinical relevance were found. Analyzing the data of the weightbearing area for all traction forces, distraction increased most up to forces of 60 pounds if traction was combined with distension. For traction alone, distraction increased more for traction forces between 40 and 80 pounds. Conclusions: Using a traction table for supine position in hip arthroscopy, best distraction of the weightbearing and central part of the hip joint is achieved by the combination of traction and distension. Flexion of 20 to 40° and no abduction showed further increase. This may be due to better relaxation of the articular capsule during flexion and more effect of the perineal bar when no abduction is applied. In combination with distension, traction forces of more than 60 pounds can be avoided in order to reduce the risk of perineal and neurovascular injury.

J.W. Thomas Byrd, Kay S. Jones Nashville, Tennessee, USA

Introduction: An inverted labrum is an acknowledged but often elusive cause of osteoarthritis. The purpose of this study is to define the characteristic radiographic features of secondary osteoarthritis of the hip due to an inverted labrum, and report the results of arthroscopic treatment. Methods: From 165 consecutive arthroscopic hip procedures, nine were identified with secondary osteoarthritis due to an inverted labrum and have minimum two year follow up. All were prospectively assessed with a modified Harris hip score. All patients underwent excision of a torn inverted labrum and chondroplasty. Three cases, with well circumscribed Grade IV acetabular lesions, also underwent abrasion arthroplasty (microfracture technique). Results: Characteristic AP radiographic features included isolated narrowing of the superolateral joint space which caused a convergence laterally in the normally parallel lines created by the radius of curvature of the subchondral bone of the acetabulum and the convex surface of the femoral head. These two features created a false appearance of dysplasia (“pseudodysplasia”) as the CE angle of Weiberg was normal in six cases (range 24°-41°; average 29°) and truly indicative of dysplasia in only 3 cases (17°). The median preop score was 51 with a postoperative score of 56. Only four patients demonstrated significant improvement. The best results were in the abrasion arthroplasty group with an average improvement of 33. Discussion and Conclusion: An inverted labrum is an occasionally encountered cause of osteoarthritis of the hip. The radiographic features are quite characteristic and essentially pathognomonic. Recognizing these features in this population may avoid a more extensive unnecessary work up for poorly explained hip pain. Overall, arthroscopic treatment in these cases fares no better than previously reported results for arthritis of all causes. An exception appears to be those who are candidates for abrasion arthroplasty.

Michael Dienst, Stefan Gödde, Michael Brang, Dieter Kohn Homburg/Saar, Germany

Introduction: Hip Arthroscopy (HA) without traction of the Superficial Compartment (SC) has been disregarded. The synovial folds also called retinacula of Weitbrecht have been described anatomically and considered as important for the blood supply of the proximal end of the femur. In the present study, anatomy of the synovial folds is described in cadavers and arthroscopic appearance presented in vivo with considerations of clinical relevance. Methods: In 16 hips of 8 embalmed cadavers, the intraarticular structures of the SC were dissected and measured up. The synovial folds were excised and investigated histologically for neurovascular structures. Between October 1997 and August 1999, 35 HA of the SC without traction have been performed in vivo. Arthroscopic frequency and morphology of the synovial folds was recorded. Results: In cadaver and on arthroscopy, the synovial folds were characterized by partial reflections of the synovial membrane commencing at the margin of the articular cartilage of the femoral head. The medial fold passed proximal from the lesser trochanter to the medial margin of the head distal to the fovea. It was always found in cadaver and in vivo, strong, shaped as an upside-down “T” and not adherent to the neck. Even if intraarticular visibility was significantly decreased eg. by synovial disease, this fold was easy to locate. At its proximal insertion, the medial fold was close to the zona orbicularis and the labrum. The anterior fold passed along the anterior surface of the neck originating from linea intertrochanterica toward the anterior margin of the head. It was shaped flat, often adherent to the neck and divided into more strips. It was not found in 5 cadaver hips. On arthroscopy, the anterior fold was clear in 18 cases, but much more difficult to visualize due to its adherence to the neck. The anterior fold was located where the first entry is achieved for HA of the SC. The lateral fold was not found in 2 cadaver hips, and presented mostly adherent, shorter and like a “flat plate.” Good arthroscopic visibility was established only in 8 cases. Histological investigations revealed a small artery in each fold, neural structures were not found. Discussion: The synovial folds also called Weitbrecht's retinacula are longitudinal and flattened bands medial, anterior and lateral to the femoral neck. In particular, the medial synovial fold is a constant structure, that can easy be located during arthroscopy of the SC due to its free course through the medial part of the joint. Especially if intraarticular visibility is limited eg. caused by synovial disease, this synovial fold is very helpful for orientation within the joint. The anterior and lateral folds are of less intraarticular visibility due to their adherence to the neck. The anterior fold is close to the first entry point of the joint. Thus, the fold can be damaged by the trokar or sleeve and hinder further manoeuvrability and visibility anterior to the femoral neck. Damage to the folds should be avoided since a small artery was found in each fold. Injury may results in “red out” and avascular necrosis of the femoral head because the retinacular arteries are considered to be important for blood supply of the femoral head and neck.

Akihiro Kanamori, Tracy M. Vogrin, Masayoshi Yagi, Savio L-Y. Woo, Christopher D. Harner Pittsburgh, Pennsylvania, USA

Introduction: Conservative treatment of isolated PCL injuries has been shown to result in osteoarthritis (OA) in the medial and patellofemoral compartments secondary to increased posterior tibial translation and joint contact pressures. However, little is known about the effects of PCL deficiency on other soft tissues or the contact forces at the tibio-femoral joint. Our objective was to evaluate the effects of PCL deficiency on the menisci, ACL, meniscofemoral ligament (MFL), posterolateral structures (PLS), and joint contact forces. Based on the clinical observations of OA with PCL deficiency, it is hypothesized that PCL deficiency increases the in situ forces that predispose the menisci, the PLS, and articular cartilage to injury. Materials and Methods: Eight fresh-frozen human cadaver knees (age range: 42-69 years) were tested. Using the robotic/universal force-moment sensor (UFS) testing system, the knee was tested at 30° and 90° of flexion under a combined 134 N posterior tibial and 200 N axial compressive load and the resulting kinematics of the intact knee were recorded. The PCL was then transected arthroscopically. The load was applied again and the resulting kinematic changes recorded for the PCL-deficient knee. The soft tissue structures of the knee were then transected, with the cutting order varied: ACL, MFL, PLS, medial and lateral mensici, and medial and lateral condyles. The in situforce in each ligament and the resultant force in the menisci and condyles were determined for both the intact and PCL-deficient knees. Statistical analysis was performed using repeated measures analysis of variance (P <.05). Results: The in situ forces in the ACL, PLS, and MFL, and the resultant force in the menisci and condyles are shown in Table 1. With PCL-deficiency, the in situ force in the PLS increased by 18 N at 30° of flexion and by up to 57 N at 90° (P <.05). The joint contact force in the medial condyle also increased with PCL-deficiency at 90° by an average of 34 N (P <.05). MFLs were present in 6 knees; the in situforces in the MFL increased significantly at 30° and 90° after sectioning the PCL (P <.05). However, the in situ force in the ACL and menisci and the lateral condyle joint contact force did not change significantly with PCL deficiency (P <.05). Discussion: Our data indicate that PCL-deficiency can put the PLS and MFL at increased risk of injury or lead to the deterioration of the articular cartilage of the medial condyle. These findings are consistent with the clinical observation that early OA occurs in the medial compartment. However, no changes were observed for the menisci or ACL, which is consistent with reports of a low incidence of meniscal injury secondary to PCL-deficiency. Surgical management of PCL injury may be important to prevent further soft tissue damage and joint degeneration.

Pier Paolo Mariani, Fabrizio Margheritini, Gaetano Maresca, Ezio Adriani Rome, Italy

The last decade has been characterized by a large interest on posterior cruciate ligament pathology. New or improved surgical technique have achieved important results in PCL reconstruction and the outcome are extremely encouraging. Despite this large interest, the diagnosis of PCL tear is still undefined. The clinical tests present different sensitivity and sensibility and can be different from physician to physician. MRI, which is widely used for detecting ACL and meniscal tear, doesn't allow a functional diagnosis on the ligament conditions. Stress radiography have been introduced in the recent years in order to improve the accuracy in assessing the functional integrity of the torn PCL. The advantages of this technique includes the possibility, to exactly quantify the degree of posterior displacement, to demonstrate rotatory components and to use for following-up patients after surgery. In contrast, disadvantages comprise the time-consuming and technical demanding features of some methods.

Aim of the study is to evaluate the efficacy of different stress radiography techniques in detecting PCL injuries and to compare this results with the MRI and clinical data.

Sixty patients have been enrolled for this study. All of these complained with a sub-acute or chronic PCL injury confirmed by the MRI. The patients underwent a clinical and KT-2000 examination, followed by a series of stress radiographs. These include a posterior drawer test with Telos at 90 and 25 degree of flexion, the active method proposed by Chaissaing and the axial view of Puddu.

The results support the idea that stress radiographs present the highest accuracy in detecting PCL injuries, than the other clinical tests. MRI in our study showed a good accuracy in detecting ligament's injury but it gives poor information regarding ligament's functionality. Among the different radiological techniques utilized for this study, the active method proposed by Chaissaing and the standardized stress radiographs with Telos seem to be the more reliable and accurate. The axial view of Puddu from the technical point of view is the most simple to perform but in our opinion, it can not be safely used for following-up patients after surgery.

William Wheatley, John W. Uribe, Keith S. Hechtman, John E. Zvijac, Alberto Martinez, Matthias Schurhoff Coral Gables, Florida, USA

Introduction: Repair, as opposed to reconstruction of the posterior cruciate ligament (PCL) has been reported as a successful treatment option for some types of tears. Few reports have been published on the outcome of these repairs and most of them include only open repairs. We present our series of thirteen arthroscopic repairs and their outcome. Materials and Methods: Between January 1991 and January 1999 a total of nineteen PCL repairs were performed; thirteen of these were arthroscopic. Ten of the patients were males and three were females; the average age at the time of surgery was twenty-three years. The mechanism of trauma was sports related in nine of the patients. All of the patients had associated injuries; the anterior cruciate ligament (ACL) was the most common in seven patients. All of the patients had an avulsion of the PCL from the femur through the fibers of the ligament. The surgical technique included placement of non-absorbable sutures (prolene) to the loose proximal fibers of the ligament, abrasion of the femoral anatomic insertion site and creation of femoral tunnels to tie the sutures over a bone bridge on the medial femoral condyle. The other injured structures were addressed in the same operation. The patients' results were measured according to the scoring system of Lysholm and Gillquist and the International Knee Documentation Committee (IKDC). Pre-operative and post-operative MRI's were reviewed. Post-operative lateral stress x-rays were obtained to evaluate the level of posterior subluxation of the tibia relative to the femur. Results: Mean follow-up was 31.8 months. The patients' knee was normal in six (46.2%), nearly normal in six (46.2%) and abnormal in one (7.7%). One of the patients had a second arthroscopic procedure twenty-one months after the index operation for removal of loose bodies, and his knee was rated as normal at the time of follow-up. Another patient had a second sports related injury and sustained an ACL tear that required reconstruction. The range of motion was equal to the uninvolved extremity in ten (92%) of the patients. MRI was useful in accurately distinguishing a repairable femoral avulsed PCL in all cases in which it was obtained preoperatively. Postoperative MRI evaluation revealed healing of the avulsed PCL in all cases. Discussion and Conclusions: Our results suggest that PCL repair with an arthroscopic technique in patients with a femoral avulsion, is effective in reducing postoperative instability. Twelve of the thirteen patients (92.3%) had a normal or nearly normal knee. MRI was useful in detecting repairable PCL femoral avulsions preoperatively and confirmed healing after arthroscopic repair.

David N.M. Caborn, Pete Hester, Kevin Sumida, Jeff Selby Lexington, Kentucky, USA

Posterior cruciate ligament reconstruction success to date has frequently been limited by failure to reproduce and restore normal knee kinematic. Due to the substantial anatomic and biomechanical properties of the anterolateral band, the posteromedial band has not been reconstructed. The authors introduce a technique which attempts to recreate both the anterolateral and posteromedial bands of the PCL by splitting a quadrupled hamstring tendon autograft longitudinally for half of its length yielding a “double-bundle/Y” structure. Knee symptomatology frequently occurs in the last 30° of extension and failure to reconstruct the posteromedial bundle does not address this. It is hoped that by more closely approximating the anatomic nature of the ligament's bands in recreating/combining their insertion sites with their helical conformation one will more accurately mimic their reciprocating function in flexion and extension.

By applying the clock face model to the lateral aspect of the medial femoral condyle of the right knee as visualized in an AP plane, one will place tunnels in the medial femoral condyle at the 1 o'clock and 3 o'clock positions which correspond to the anterolateral and posteromedial bands respectively. The anterolateral band tunnel site will be placed approximately 3.5-4.0 mm posterior to the articular surface. The posteromedial band site tunnel will be placed approximately 8 mm posterior to the articular surface.

Several features of this technique warrant consideration. Utilization of the hamstring autograft reduces its self inherent antagonistic action in the PCL injured knee. By using standard anteromedial and anterolateral portals one avoids the risk associated with the posterior portal along with its postoperative wound and soft tissue problems. By avoiding superomedial portal placement one minimizes inhibition of the quadricepts mechanism.

The goal of any ligament reconstruction is to closely reproduce the (normal) kinematics of the knee. The posteromedial band though small has an important function in the first 30° of extension to diminish posterior translation of the tibia and secondarily stabilize the LCL and posterolateral capsule. Biomechanically this may enhance single bundle PCL reconstruction and further diminish chondral wear of posteromedial compartment and patellofemoral articulation.

A video tape of the entire technique: Graft preparation, Surgical technique and post op management is available. A retrospective >2 yr. F/u is being analysed and will be presented. Functional results are improved over currently presented single bundle reconstruction techniques.

Frank Kandziora, Robert Pfhigrnachei, Norbert Südkemp, Andreas Weiler Berlin, Germany

Introduction: Posterior cruciate ligament (PCL) reconstruction has not succeeded in returning baseline laxity to the knee joint so far. Several factors may contribute to graft stretch-out. Transtibial PCL reconstruction leads to a significant graft angulation. Angulation results in nonuniform graft loading and may lead to sequential failure of the outer fibers and graft elongation. We hypothesized that rotation of the graft may be able to solve this problem, by homogeneously aligning fiber loading. Therefore the aim of this study was to determine if the elongation of a the PCL-graft is different when loaded in angulation than longitudinally. Additionally the effect of graft rotation was investigated. Materials and methods: 48 human fresh frozen 10 mm bone-patellar-bone grafts were harvested. The average age of the donors was 56.2 ± 4.6 years. The grafts were randomly assigned to one of the following test groups. Group 1 (n = 8): 0 degrees angulation/0 degrees rotation, Group 2 (n = 8): 0 degrees angulation/360 degrees rotation, Group 3 (n = 8): 45 degrees angulation/0 degrees rotation, Group 4 (n = 8): 45 degrees angulation/360 degrees rotation, Group 5 (n = 8): 90 degrees angulation/0 degrees rotation, Group 6 (n = 8): 90 degrees angulation/360 degrees rotation. Grafts were loaded over 100 cycles in increments of 50 to 150 N, 50 to 300 N and 50 to 450 N under load control with a cross-head speed of 250 mm/min. Angulation was induced over a 10 mm radius of curvature. After each load increment grafts were allowed to relax for 60 minutes to minimize viscoelastic deformity. Load displacement curves were obtained and subjected to statistical evaluation. Results: The average elongation for a longitudinal graft under cyclic loading with 50- 150 N was 0.29 ± 0.12 mm (group 1), for an 45 degrees angulated graft 0.73 ± 0.20 mm (P =.032) (group 3), and for a 90 degrees angulated graft 0.98 ± 0.38 mm (P = 0.041) (group 5). This represents a 253% and 338% increase in graft elongation with angulation. 360 degrees graft rotation of a longitudinal graft under cyclic loading with 50-150 N increased graft elongation significantly (group 2: 0.81 ± 0.50 mm (P =.002), group 4: 1.52 ± 0.46 mm (P =.031), group 6: 1.63 ± 0.52 mm (P =.044)). This represents a 279%, 225% and 166% increase in graft elongation with rotation. Under cyclic loading with 50-300 N and 50-450 N the average graft elongation increased significantly for the longitudinal and rotated grafts. The mean hysteresis-area over 100 cycles with 50-150 N for the longitudinal graft was 1.938 ± 0.61 N cm² (group 1), for an 45 degrees angulated graft 2.762 ± 0.88 N cm² (P =.014) (group 3), and for a 90 degrees angulated graft 3,498 ± 1.08 N cm² (P =.011) (group 5). Again the average hysteresis-area increased with load and graft rotation. Conclusion: Graft angulation resulted in a change of structural properties. With increasing angulation the graft elongates to a maximum of 338% compared to the longitudinal graft. This may lead to failure of outer graft fibers due to nonuniform loading and may contribute to long-term graft elongation. However, although rotation may lead to a more uniform fiber loading, rotation of the graft further increases elongation significantly. Reconstructive procedures which minimize graft angulation and rotation may reduce the risk of graft stretch-out.

Michael J. Strobel, Andreas Weiler, Martin Schulz, Kai Russe, H.-Jürgen Eichhorn Straubing, Germany

Introduction: In PCL deficient knees history taking rules out functional instability and diagnostic tools relay on detecting an increase in posterior laxity using different clinical measures, such as manual drawer examination, arthrometric evaluation, or stress-radiography. However, during our practice we thought to be confronted with a special group of PCL deficient patients showing an absent or only slightly increased AP drawer laxity, often combined with pain rather than subjective instability. These cases showed obvious posterior sagging and anterior stress-x-rays ruled out a posterior subluxation of the tibia which we define as the fixed posterior subluxation (FPS). The rationale of the present report is to rule out the incidence, the degree, influencing factors, and treatment options of the FPS in a retrospective analysis of PCL deficient knees. Patients: The FPS is defined as a posteriorly locked tibia in cases of PCL deficient knees found on 90° anterior stress-x-rays equaling or exceeding a displacement of 3 mm. In a series of 587 PCL deficient patients in 248 cases anterior and posterior stress-x-rays were available to allow for the diagnosis of a FPS. Patients with a FPS were treated with a special extension brace (posterior-tibial-support = PTS-brace) having an inlay which supports the calf to reduce posterior sagging. Results: 109 Patients met the criteria of a FPS and served as the study group, exhibiting an incidence of 44%. The remaining 139 patients served as the control group. Within the study group 63 patients (57.8%) were classified as having a grade I (3 to 5 mm) FPS, 37 patients (33.9%) with a grade II (6 to 10 mm), and 9 patients (8.3%) with a grade III (>10 mm) subluxation. Posterior stress-x-rays in the study and the control group showed a mean posterior subluxation of the tibia of 13.65 ± 4.52 mm and 12.12 ± 4.82 mm. The mean FPS in the study group was 6.23 ± 3.38 mm. The AP laxity, defined as the total displacement between anterior and posterior stress-x-ray values was 7.4 ± 4.53 mm in the study group and 13.46 ± 5.48 mm in the control group (P <.0001). The AP laxity in the study group, defined as the residual posterior laxity was distributed into three grades (I = >10 mm, II = 6-10 mm, III = 0-5 mm). Grade I was found in 27 cases (24.8%), grade II in 37 cases (33.9%), and grade III in 45 cases (41.3%). In 59 out of the 109 study patients anterior stress-x-rays for treatment control were available after a mean treatment period of 180 days. In these cases the FPS could be reduced from 6.93 ± 3.85 mm to 2.58 ± 5.22 mm (P <.0001). Within the subgroups the fixed posterior subluxation was reduced for grade I (N = 30) from a mean of 4.03 ± 0.89 mm to 0.87 ± 3.7 mm (78.4%, P <.0001), for grade II (N = 22) from 8.41 ± 1.59 mm to 2.45 ± 3.97 mm (70.1%, P <.0001), and for grade III (N = 7) from 14.71 ± 3.09 mm to 10 ± 7.3 mm (32%, P =.042). History of failed PCL surgery and a long history of PCL insufficiency were found to be factors at risk for the development of a fixed posterior subluxation. Discussion: The FPS shows a high incidence in patients with PCL insufficiency and should therefore be excluded in any case. A history of failed PCL surgery and a long history of PCL insufficiency prior to treatment represent factors at risk for the development of a FPS. The diagnosis and grading of this phenomenon should be carried out using standardized anterior and posterior stress-x-rays. An FPS may reduce the prognosis of a PCL reconstruction, due to a high soft resistance acting on the reconstruction. The use of a PTS-brace presents an appropriate tool to reduce grade I and II FPS. In grade III subluxation more invasive procedures should be considered to reduce the FPS.

Fabrizio Margheritini, Gianluca Camillieri, Ezio Adriani, Pier Paolo Mariani Rome, Italy

Despite the large numbers of studies on graft healing after ACL reconstruction, little is known about the fate of PCL graft following a surgical reconstruction. Several MRI studies have been performed to evaluate the graft integrity following ACL reconstruction, showing different results due to the revascularization process, scar tissue presence and mechanical impingement. To our knowledge none MRI studies in vivo have been done to study PCL graft behavior after surgical reconstruction. The aim of this study is to prospectically study the MRI changes of patellar tendon graft and to correlate with clinical findings.

We studied twenty consecutively treated patients, who underwent an arthroscopic reconstruction with autologous bone patellar tendon bone. All patients complained at the time of surgery with a chronic injury and due to the type of surgery there were divided in two groups isolated PCL reconstruction (with intact ACL and peripheral ligaments) and combined one stage ACL-PCL reconstruction. The patients enrolled for this study, underwent a serial MRI examinations combined with a clinical evaluation at 3, 6, 9 and 12 months after surgery. Moreover, we selected ten patients, who underwent an isolated PCL reconstruction with at least 3 years follow-up (FU), who were studied with a single MRI examination at their longest FU.

The results of our study showed that the patellar tendon takes more than one year before achieving a low-intensity signal over its entire course, which means an almost complete healing process. This is in controversy with the results after ACL reconstruction, where the graft takes approximately six months to complete the ligamentization process. Moreover we were not able to find any statistical correlation between the MRI signal and the clinical stability in the patients prospectively studied. The maturation process seems to be slower on the tibial side in accord to the killer-loop problem. Moreover, in ACL-PCL group the graft's healing is slower than in patients with isolated PCL tear. This data could be addressed to the different interactions between ACL and PCL (screw-home rotation) after surgery. In contrast all patients studied with a single MRI at their longest FU showed a strict correlation between MRI signal and clinical stability. Our data support the idea that PCL maturation is more time-dependent than the ACL one and with different biomechanical features.

Frank K. Noojin, Michael J. Voor, Edward D. Tillett Ridgeland, Mississippi, USA

Introduction: A major clinical concern regarding anterior cruciate ligament (ACL) reconstruction with hamstring autograft is the amount of early graft elongation that occurs prior to healing within the bone tunnel. The purpose of this biomechanical study was to compare hamstring graft fixation on the femoral side with tandem washers versus fixation with a combination of tandem washers and an interference screw. Methods: Fourteen semitendinosus and gracilis tendon grafts were harvested from 9 human cadaver specimens. Bovine femurs were mounted to a load table on a servohydraulic test frame (MTS). Tunnels (9 mm diameter, 4 cm length) were reamed from the lateral cortex to the intercondylar notch and quadruple-looped grafts were passed through the tunnels and fixed with 2 different fixation techniques: 7 grafts fixed with two 6.5 cancellous screw with spiked washers (POST) and 7 grafts fixed with two 6.5 cancellous screws with spiked washers and a bioabsorbable 8 mm interference screw (COMBO). The looped ends of the graft were fixed over a 5 mm diameter rod attached to the MTS actuator. Preconditioning of the graft was performed manually prior to testing. Biomechanical testing consisted of static loading of 40 newtons for 5 minutes followed by cyclic loading at 1 Hz for 100 cycles at forces of 100, 200, 300, and 400 newtons. At the conclusion of testing for the combo group, the post/washer fixation was removed and the testing sequence repeated, testing an isolated interference screw. A pull-out test was also performed. Actuator position was recorded throughout testing to determine the overall increase in graft length. Results: There were no graft failures in either group. Graft elongation averaged 4.66 mm (range: 3.3-6.3) for the POST group compared to 2.86 mm (range: 0.88-3.49) for the COMBO group. This difference was statistically significant (P <.01). Graft elongation began during the preconditioning phase and progressively increased at higher load forces. Testing fixation with an isolated interference screw yielded an additional elongation averaging 2.53 mm. The mean failure strength of fixation with an isolated interference screw was 775 newtons (range: 548-1091). Discussion and Conclusions: This biomechanical study demonstrates that shortening graft distance between points of fixation results in less graft elongation with progressive loading. The difference in graft elongation between the two groups was proportional to the difference between the length of graft exposed to load in the two groups. Evidently, the interference screw effectively shortens the graft length susceptible to stretching during the early postoperative period. However, testing the interference screw in isolation showed additional slippage around the screw, indicating an interference screw alone may not be sufficient fixation for cyclic loading. This study also suggests that progressive cyclic loading may be a more appropriate means of testing ACL graft fixation than simple load-to-failure tests because graft elongation was observed at lower loads.

Andreas Weiler, Clemens Abel, Cornelius Förster, Andreas C. Stähelin, Norbert P. Südkamp Berlin, Germany

Introduction: Biodegradable interference screws are increasingly implanted today. Most of these implants consist of slow degrading, high molecular, and highly crystalline materials such as poly-(L-lactide) (PLLA) or PLA 98. These materials are characterized to show an incomplete degradation with an accumulation of insoluble and crystalline implant remnants up to several years. Thus it has been advocated that these materials show no complete degradation within an appropriate time and no osseous replacement of such an implant could be demonstrated so far. Other interference screws consist of fast degrading materials such as polyglycolide-co-trimethylencarbonate (PGA/TMC) or poly-(D,L-lactide-co-glycolide) (PDLLA/PGA) which may not allow for secure soft tissue graft interference fit fixation due to their fast degradation. In addition, fast degrading implants may bear the risk for the development of foreign-body reactions such as osteolysis, sterile sinus formation, or synovitis. The goal of the present study was to determine the in-vivo biocompatibility, degradation, and the osseous replacement of an intermediate degrading interference screw consisting of poly-(D,L-lactide) (PDLLA). Methods: In 36 sheep PDLLA screws (Sysorb, Sulzer Orthopedics Ltd.) were implanted in the right proximal tibia and animals underwent an ACL replacement with two PDLLA screws in their left knee. Animals were sacrificed after 6, 9, 12, 24, 52, and 104 wks and undecalcified sections for histological evaluation were created. X-rays were conducted at defined time intervals and the inherent viscosity (iv) was determined from extracted screw material. Results: At 6, 9, and 12 wks the screws appeared macroscopically intact. Iv measurements showed a continuous decrease starting with 0.947 dl/g at time zero, 0.637 dl/g at 6 wks, 0.517 dl/g at 9 wks, and 0.426 dl/g at 12 wks. Histology showed an early osseous enclosure of the implants. Screws were macroscopically degraded at 24 wks leaving a central fluid accumulation at the former implant site. At that time no synovitis or sterile sinus formation was detected. In some cases x-rays showed a low grade enlargement of the implant site which was maximally pronounced at 24 wks. Histology of these specimen showed amorphous implant remnants to be deposited in macrophages and foreign-body giant cells. A high osteoclastic activity was found as the margin of the former implant site. Osteolyses were seen in cases were no connection between the implant site and the medullary cavity was found. At 52 and 104 wks no implant remnants could be detected at the former implant site. Further x-ray follow-up showed a subsequent osseous replacement of the former implant site. After 104 wks the implant site was filled with irregular lamellar and woven bone. Discussion: The screw material used in the present study shows a macroscopical degradation at 24 wks and a complete clearance of implant remnants between 24 and 52 wks. An osseous replacement of the former implant site was present after 104 wks. To our knowledge this is the first report describing the complete intra-osseous degradation and the osseous replacement of a polyactide implant within an appropriate time frame. Therefore, it could be concluded that PDLLA in contrast to PLLA presents a real biodegradable material. Although strength retention decreased early soft tissue graft fixation was not compromised. The screw material further showed an appropriate biocompatibility without the development of a sterile sinus formation or synovitis. Clinically irrelevant osteolytic lesion could be detected in cases were no sufficient drainage of byproducts to the medullary cavity was present. However, such osteolyses could not detected in humans so far. PDLLA presents an alternative biodegradable material and may overcome concerns about an incomplete or to fast material degradation.

Sharon L. Hame, Keith Markolf, Monte Hunter, Arash Gabayan, Brent Davis, Matthew S. Shapiro, Gerald A.M. Finerman Los Angeles, California, USA

Introduction: Various authors have recommended rotating the anterior cruciate ligament (ACL) graft prior to fixing it within the tibial tunnel. These recommendations are without scientific basis, as there have been no published studies related to the biomechanical consequences of this procedure. The purpose of this study was to determine the effects of rotating a patellar tendon allograft upon anterior-posterior (AP) knee laxity and forces developed within the graft at full extension. Methods: The patellar bone block of a 10 mm bone-patellar tendon-bone graft was potted in a cylinder of acrylic cement and attached to a load cell mounted on the tibia; the cylinder could be rotated and locked onto the tip of the load cell. High strength dacron lines interwoven into the tibial bone block were passed through femoral tunnel and clamped to maintain graft pretension during testing. Thirteen fresh frozen knee specimens received grafts which were pre-tensioned at 30 degrees of flexion to restore AP laxity to that of the intact knee. AP laxity was measured (200 N of applied tibial force) at 0, 30, and 90 degrees of knee flexion with the graft in neutral rotation, and in 90 and 180 degrees of internal and external rotation. Two knee specimens were used for the graft force tests; one received 10 separate grafts and the other 7 separate grafts. Graft forces at full extension (2.0 N-m extension moment) were recorded for the graft rotation conditions at 23N, 45N, 67N, and 89N of graft pretension. For both graft force and laxity testing, the proximal end of the graft was pre-tensioned and fixed prior to rotating and locking the distal bone block. Results: AP laxity at 30 degrees flexion decreased an average of 0.87 mm with 90 degrees of graft rotation; laxity reductions were equivalent for both directions. The mean laxity decrease with 180 degrees of external rotation (1.84 mm) was significantly greater than that for 180 degrees of internal rotation (1.06 mm). When significant, all mean laxity reductions at 0 and 90 degrees of flexion were less than corresponding values at 30 degrees flexion. Rotating the graft internally 90 degrees increased mean graft force at full extension by 11N; external graft rotation had no significant effect. Rotating the graft 180 degrees increased mean graft force at full extension by 25 N; graft force increases were not significantly different between internal and external rotation. The magnitude of the graft force increase due to rotation was unaffected by the initial level of graft pretension. Conclusions: For all tests performed, the potted distal end of the graft attached to the load cell was rotated and locked after the pretension had been set. Since the length of the graft was fixed, rotating the graft increased the pre-tension in the graft and decreased AP laxity. In clinical practice, the distal end of the graft would most commonly be pre-tensioned after twisting and then fixed, a procedure which would be expected to produce changes less than those which we measured because the length of the graft would be free to change as the graft is rotated. Irrespective of pre-tensioning protocols, the direction of graft rotation would not appear to be an important variable in graft force and laxity considerations. Overall, the biomechanical changes which we measured would appear to be relatively insignificant in terms of expected in vivo function, although the effects of rotating the graft upon its incorporation and maturation are unknown.

William B. Workman, Marc T. Galloway New Haven, Connecticut, USA

Introduction: A unique complication of ligament reconstruction employing bioabsorbable screws is the development of cystic changes around the screw site. These cystic changes have been reported to substantially enlarge the osseous tunnels within which the graft normally heals. In some cases, these changes were felt to result in the failure of the procedure, necessitating subsequent revision surgery. While the clinical effect is appreciated, the etiology is not well understood. This study is intended to gain further understanding of the host response to these implants using histomorphometry, which quantifies the bone response at the cellular level. Methods: Eighteen mature, male New Zealand rabbits were used in this study. Each rabbit had a transverse 2 mm diameter drill hole made in the right distal femur (extraarticular), and an oblique drill hole in the left proximal tibia extending into the knee joint (intraarticular). 2 mm diameter bioabsorbable pegs were implanted in both sides. Half of the rabbits received PLA implants and the other half received PGA. Five rabbits from each implant group were sacrificed at 3 months, and the remaining rabbits were sacrificed at 6 months. The implant areas were harvested and prepared for histologic and histomorphometric analysis. Histomorphometry was performed on each specimen, which quantified both bone volume and osteoid volume as a percentage of total volume. Results: Extraarticular bone volume/total volume was significantly greater in both PLA and PGA knees at three months compared to bone with implants exposed to an intraarticular environment (P <.05) at three months. The same statistically significant difference was found in bone volume/total volume in the PLA knees at 6 months (P <.05). There was no statistically significant difference in bone volume/total volume between extraarticular and intraarticular PGA samples at 6 months. Additionally, the PGA bone volume/total volume was less than in the PLA knees at 6 months. Discussion and Conclusion: PGA has a more rapid hydrolysis rate than PLA, so it is possible that the decreased bone volume in the PGA knees is a result of the more rapid hydrolysis rate. The results of our study showed that intraarticular placement of bioabsorbable pins in rabbit knees resulted in a decreased bone volume compared to extraarticular placement of pins. It is possible that exposure to an intraarticular environment is related to a decreased host bone reparative response in the presence of bioabsorbable materials.

Mark F. Sherman, Mark Jedici, Joel R. Bonamo, Maryirene I. Flynn Staten Island, New York, USA

This is a prospective study of the first twelve patients in which a bioabsorbable wedge was used for femoral fixation in an ACL reconstruction. Each of the twelve patients underwent an ACL reconstruction using a bone-patella tendon-bone autograft.

Twelve consecutive patients who underwent ACL reconstruction by the senior author (MS) had a SR/PLLA wedge used on the femoral side and an interference screw on the tibial side. The patients were evaluated by the modified HSS Scoring System, KT1000 testing and post-operative MRI.

There were no intra-operative complications of wedge insertion. There were no post-operative complications related to the wedge (i.e. synovitis, infection or fixation failure). The post-op exam revealed eleven of the twelve patients with a negative pivot shift and one of the twelve with a mild pivot shift. KT1000 manual max side to side difference was greater than 3 mm in nine patients, between 3-5 mm in three patients and less than 5 in zero patients. The average Modified HSS objective score was 86.3 out of a 100 and the average Modified HSS subjective score was 91.3 out of 100. Nine of the twelve patients had a post-op MRI. All nine revealed no divergence between the wedge and the graft and no increased bony reaction around the wedge when compared to the tibial interference screw.

In early clinical trials, the wedge has proven itself as a means of fixing the femoral side of a bone-patella-bone autograft, and may be considered as an alternative to an interference screw. The wedge facilitates femoral interference fixation with minimal chance of divergence and graft laceration.

Laith A. Farjo, William Beach, Joel Gonzales, John Meyers, Richard B. Caspari Ypsilanti, Michigan, USA

Introduction: The purpose of this study is to determine whether an intra-articular, surgeon-inserted, continuous bupivicaine infusion pump has efficacy for post-operative pain relief after ACL reconstruction. These commercially available pumps are designed to infuse a solution at 2 cc/hour over 48 hours via a small catheter and are removed by the patient. Methods: A prospective, randomized, double-blind, controlled study was designed. Patients undergoing endoscopic ACL reconstruction without inside-out meniscal repair or other ligament pathology (e.g. MCL) were asked to participate in the study. Patients were randomly assigned to three groups: 1) No pump, 2) Pump with saline, and 3) Pump with bupivicaine 0.5%. All cases were performed at one surgery-center, using a standardized anesthetic protocol, on an outpatient basis. Pain pumps were removed by the patient at 48 hours post-operatively. Outcome measures include: time spent and narcotic needs in the PACU, visual analog pain scale (VAS) in the PACU and for 5 days thereafter, total narcotics used at 48 hours post-operatively, and amount of sleep in the first 2 post-operative nights. Results: Twenty-six patients have been enrolled in the study to date. There was no difference for postoperative pain score between all 3 groups (average VAS: group 1 = 4.8, group 2 = 3.2, group 3 = 4.3; P =.19). There was no difference in the amount of narcotics needed in the PACU (morphine equivalent (mg): 7.3, 3.8, 4.1; P =.15). In the 48 hour post-op period, there was a nearly significant trend for the non-pump group 1 to require more narcotics (Percocet taken: 15.8, 9.7, 11.9; P =.053). Amount of sleep deficit for the first two post-operative nights was not significantly correlated with pump group; however, group 2 (saline) did have significantly less awakenings from sleep than group 1 (4.5 vs 2.1; P =.038). There were no complications. Conclusions: Although some statistically significant differences were found, these primarily related to the 2^nd control group (pump with saline), possibly indicating a placebo effect of having the pump in place. This indicates that any clinically significant effect of these pumps on post-operative ACL reconstruction pain is negligible.

Ronald T. Rook, Gene R. Barrett, Carrie R. Nash Jackson, Mississippi, USA

Introduction: Patients undergoing medical treatment as a result of a work-related injury have been suggested to have inferior outcomes when compared to their non-worker's compensation counterparts. The purpose of this prospective study is to assess the effect of worker's compensation benefits on the subjective and objective outcomes of patients following ACL reconstruction. Methods: From October 1988 through June 1997, 235 consecutive ACL reconstructions using bone-patellar tendon-bone autografts were performed by a single surgeon in patients that had no other ligament damage to the knee. All chronic injuries, failed and/or bilateral reconstructions were excluded from the study. The final populations included 115 patients in group one (Non WC) and 24 patients in group two (WC). There were 22 males and 2 females with an average age of 30.8 years in the WC group, and 74 males and 41 females with an average age of 23.7 years in the NonWC group. Average follow-up for the WC group was 34.5 months (24-77 months) and 33.6 months (24-92 months) for the NonWC group. All patients had equal objective evaluations (negative Lachman and pivot shift tests and <5 mm KT-1000 differences), which allowed the subjective criteria (15 item visual analog scale) to be analyzed. Results: The 15 item visual analog scale demonstrated statistically significant (P <.05) differences on questions concerning pain frequency, pain at worst, swelling, problems kneeling/squatting, twisting/pivoting, going up or down stairs, running, decelerating while running, cutting, jumping, and participation in competitive sports. It was interesting to note that the WC group noted a significant difference in the frequency of knee swelling which is in direct contrast to the objective findings on physical exam at the time of follow-up. Comparison of the pre-injury (WC group: 4.32 and NonWC group: 6.02) and final follow-up (WC group: 3.05 and NonWC group: 6.11) Tegner scores was much more revealing. Average post-operative Tegner scores were statistically significant (P <.05). It should also be noted that the WC group did not return to their pre-injury Tegner score, whereas the NonWC group not only returned to their pre-injury Tegner score but also surpassed it. There were no significant differences between any range of motion measurements, thigh circumference differences, or KT-1000 differences. Conclusions: The results of this study demonstrate a significant difference in the subjective data of patients with worker's compensation benefits. Uniformly, WC patients rated their pain, swelling and functional limitations as far worse than the NonWC group in both their preoperative and postoperative periods. Additionally the statistical evaluation of the data found significant differences (P <.05) in almost every aspect evaluated. The data from this study finds that the patient's perception of their knee function drastically differs from the objective findings of the examining physician, therefore it should be anticipated that the WC patient might have a poorer outcome from medical treatment compared to their NonWC counterparts.

Andrew Green, Sean Griggs, Daniel Labrador Providence, Rhode Island, USA

Introduction: The advantages of arthroscopic acromioplasty compared to open acromioplasty include preservation of the deltoid origin and accelerated rehabilitation. Prior studies demonstrated that a substantial portion of the deltoid origin can be inadvertently detached with standard arthroscopic acromioplasty techniques. This study evaluates the effect of acromial size and morphology on the extent of deltoid detached by arthroscopic acromioplasty. Method: 15 embalmed cadaveric acromion were harvested, decalcified and embedded in paraffin. Mid-sagital sections were stained with Mason Trichrome and analyzed with light microscopy and computer image software. Anatomic and morphologic parameters were assessed and the affect of arthroscopic acromioplasty was evaluated. Correlations with acromial size and morphology were determined. Results: The mean thickness of the anterior acromion was 0.74 cm (0.47-0.94). The mean linear thickness of the deltoid origin was 0.54 cm (0.38-0.68). The mean acromial angle was 20° (5°-31°). A 4 mm acromioplasty released 56% (45-85%) of the deltoid origin while a 5.5 mm acromioplasty released 76% (59-100%). The extent of deltoid detachment was inversely related to acromial thickness (P <.001) and acromial angle (P <.05). Discussion: Despite the perceived advantages of arthroscopic acromioplasty, a substantial portion of the deltoid origin can be detached. The size and morphology of the acromion is an important factor. The typical individual with coracoacromial outlet impingement (thick acromion with spurring) is less likely to have extensive release of the deltoid than an individual with a thin flat acromion. Although the clinical relevance of these findings remains unclear, our findings indicate that pre-operative planning should be a critical aspect of arthroscopic acromioplasty.

Julious P. Smith, Felix H. Savoie, III, Larry D. Field Jackson, Mississippi, USA

Introduction: Since originally described by O'Driscoll in 1990, posterolateral rotatory instability of the elbow has become a much more frequently recognized cause of elbow pathology. It was previously thought that open techniques were necessary to correct the pathoanatomy in these patients. Now, however, there is evidence that these patients can be treated with a purely arthroscopic procedure. Methods: Eighteen (18) patients who had a positive posterolateral instability test that was either isolated or accompanied by other symptoms of elbow instability underwent arthroscopy over a nineteen (19) month period. All of them had evidence of posterolateral instability on arthroscopic examination under valgus stress and supination, and each patient then underwent arthroscopic plication of the lateral ligaments and capsule using arthroscopic suture technique. Postoperatively the elbows were immobilized for a short period and then mobilized in a hinged elbow brace. They were followed for an average of 14.7 months post-op (range 6-23 months) and evaluated with the Andrews/Carson elbow rating scale. Results: There were 9 male and 9 female patients with an average age of 33.5 years (range 15-50 years). The Andrews/Carson elbow score for these patients averaged 185 points on a scale of 200. All patients were satisfied with their results. There were no major complications. Discussion: Correction of posterolateral rotatory instability of the elbow requires adequate tightening of the lateral ligamentous structures, especially the radial ulnohumeral ligament. These results appear to show that this can now be accomplished by arthroscopic means.

Wolfgang A. Menth Chiari, David S. Ruch, Gary G. Poehling Vienna, Austria

Purpose: The authors describe the surgical technique of arthroscopic radial head resection as an expanded indication for elbow arthroscopy. Functional outcome (motion, pain, strength) and postoperative radiographs were analyzed retrospectively after a mean follow up period of 38 months (range 12 to 97 months). Methods:Patients: From 1990 to 1997 arthroscopic radial head resection was performed in 12 patients (10 male, 2 female patients; average age: 38.5 years; range: 17 to 55 years) with either posttraumatic arthritis after a radial head fracture (Mason type-II: n = 10 Mason type-III: n = 2) or with rheumatoid arthritis with severe degeneration of the radial head in the radio-capitellar joint. Duration of pain ranged from 11 months to more than five years. All patients except one underwent at least one arthroscopic debridement procedure of the radial head before arthroscopic resection of the radial head was performed. Preoperatively, patients lacked 23 degree (range 5 to 40 degrees) of extension of the elbow on the average. Mean flexion was to 141 degree (range 90 to 150 degrees). Preoperative arc of motion was 106 degree. Patients had unrestricted pronation (limitation 5 degree in two patients). Two patients had a lack of supination of 15 and 30 degree. Surgical technique: Elbow arthroscopy was performed by a standardized technique with the patient either in prone or in lateral decubitus position. The portals used are the proximal medial, anterolateral, and the midlateral portal. The anterior three quarters of the radial head and 2-3 mm of the radial neck are resected with the stone-cutting abrader in the anterolateral portal and the arthroscope in the proximal medial portal. For resection of the posterior portion of the radial head the abrader may be transferred to the midlateral portal. Results: All patients improved their range of motion of the elbow and reported significant improvement in pain relief and complete relief of mechanical symptoms. Postoperative arc of motion was 161 degree (range 135 to 180 degree) on the average at follow up. The overall average loss of strength in all functions (flexion, extension, pronation, supination) was 18 per cent. No patient had subjective or objective evidence of instability of the elbow. Consequently, all of our patients experienced a significant therapeutic benefit from the procedure according to the criteria of O'Driscoll and Morrey (JBJS1992;74-A:84-94). Discussion and Conclusion: Arthroscopic radial head resection increased functional range of motion of the elbow, provided excellent pain relief, and improved health-related quality of life in all patients. Elbow arthroscopy allows the surgeon to deal with the intrinsic joint pathology, as well as with accompanying symptoms such as synovitis, capsular contracture or loose bodies. By means of arthroscopic treatment patients are able to begin and maintain an aggressive postoperative physical therapy program immediately after surgery, thus decreasing the risk of anterior scarring and reoccurring contracture of the capsule of the elbow joint.

William J. Ciccone, II, Timothy J. Hunt, Richard Lieber, Robert Pedowitz, James Esch, James Tasto San Diego and Torrance, California, USA

Introduction: Advances in arthroscopic surgery have made shoulder stabilization procedures possible through a minimally invasive approach. These procedures however, have been associated with increased rates of recurrent instability when compared with traditional open procedures. Plastic deformation of the anterior capsular structures has been speculated to be the cause of these failures. Thermal capsular shrinkage procedures have arisen to address this issue. While the temperature required to optimize collagen shrinkage is known, a paucity of data exists on the effect of the depth of thermal penetration on the capsular healing process. The purpose of this study was to measure shoulder capsule thickness by quadrant and circumferentially from glenoid to humerus so that thermal energy in shoulder procedures may be more precisely applied. Methods: Soft tissue was dissected from eight fresh cadaveric shoulders to isolate an intact shoulder joint capsule. The humeral insertion was released and the capsule was cut into six longitudinal sections at specific locations around the glenoid. Specimens were flash frozen and stored at −80°C. Quadrant tissue was cut into longitudinal sections 14-16 micrometers thick and stained with H&E. The specimens were then digitized under a dissecting microscope and measured using computer imaging software at approx. 4 mm intervals. Humeral insertion data was recorded separately using the same method. Results: 248 separate capsular measurements were recorded. Six specimens required separate sectioning to obtain humeral insertion data. Capsular thickness increased from 2.42 mm anteriorly, to 2.8 mm in the inferior capsular pouch, and became thinnest posteriorly (2.22 mm). A trend of capsular thinning was noted from the glenoid to the humerus. Capsular thickness ranged from 3.03 mm at the glenoid to 2.17 mm at he humeral insertion. Two way ANOVA revealed a significant thickness variation from glenoid to humerus (P <.05), a significant thickness variation with regard to quadrant (P <.03), and no significant interaction (P >.7). Conclusions: Maintenance of the contracted state of the shoulder capsule following a shrinkage procedure is probably dependent upon the healing response. Limited capsular penetration may enhance healing be adjacent fibroblasts. On the other hand, full penetration may lead to a more complete collagen turnover, and thus improve stability. The results of this study conclude that shoulder capsule thickness is variable. Understanding these variations should enhance our ability to apply thermal energy in a more precise manner.

Gary Perlmutter, Kenneth Sabbag Boston, Massachusetts, USA

Introduction: Arthroscopic capsular release has been shown to be a useful adjunct in the treatment of refractory frozen shoulder. Outcome, however, may vary depending upon the etiology of the frozen shoulder, co-morbid diseases, and surgical technique. The purpose of our study was to evaluate outcomes following arthroscopic capsular release and determine efficacy of use of an indwelling interscalene catheter during the immediate postoperative period. Methods: Thirty-two patients with a refractory frozen shoulder were treated by arthroscopic capsular release. Fourteen patients suffered from idiopathic adhesive capsulitis, while 18 patients had a stiff and painful shoulder following a surgical procedure. Twenty-five patients were discharged to home on the same day of surgery, while seven patients were hospitalized for 48 hours in conjunction with use of an indwelling interscalene catheter and a continuous passive-motion machine. Results: Overall, the Constant and Murley Score improved from a preoperative average of 23.9 to a postoperative average of 55.5 (P <.01). The American Shoulder and Elbow Surgeons (ASES) Standardized Shoulder Assessment Numerical Pain Score decreased from 8.2 to 3.1 (P <.01). The ASES Shoulder Assessment Activity Score increased from 5.9 to 17.4 (P <.01). While all groups noted significant improvement, patients with idiopathic adhesive capsulitis fared significantly better than those with postoperative frozen shoulder (P <.05). Patients who received an indwelling interscalene catheter noted significant gains in improvement, compared to those who did not receive an interscalene catheter (P <.05). Conclusion: This study supports the utilization of arthroscopic capsular release with an indwelling interscale catheter in the treatment of refractory frozen shoulder. Outcomes may vary following capsular release depending on the etiology of a frozen shoulder.

Seung-Ho Kim, Kwon-Ick Ha, Jin-Hwan Ahn Seoul, Korea

Introduction: The purpose of this study is to describe arthroscopic findings, and the results of arthroscopic treatment in patients with chronic shoulder pain following a minimally displaced fracture of the greater tuberosity. Methods: Twenty-three patients with chronic shoulder pain beyond 6 months after the minimally displaced fracture of the greater tuberosity underwent arthroscopic treatment and were retrospectively assessed for the arthroscopic findings, UCLA score, return to activity after an average of 30 months (range, 24 to 40 months). There were 18 men and 5 women with the average age of 39 years (range, 24 to 61 years). Fourteen were isolated fractures and 9 were related to acute anterior instability episode. In the plane radiographs, nine (39.1%) of the patients had a large single fracture and 10 (43.5%) had a multiple number of small shell fragments. In 4 (17.4%) of the patients, no fragment was identified. The average displacement of the fracture was 2.3 mm (range, 0 to 4 mm). The UCLA rating was compared between the preoperative and follow-up scores using the Mann-Whitney U test. Spearman's rank correlation coefficient was measured to evaluate the contributing factors to the final results. Nonparametric test of Kruskal-Wallis was used to evaluate the difference in return to activity in different groups of sports activity. Results: All patients had partial thickness rotator cuff tears in the articular surface. The cuff tears were located on the tuberosity fracture area and were an Ellman's grade I in 17 patients (73.9%) and grade weber in 6 patients (26.1%), in depth of the tear. With the arthroscopic debridement or repair of the tear, as well as the subacromial decompression, the UCLA score significantly improved from 13 (range, 7 to 20) preoperatively to 32 points (range, 25 to 35) postoperatively (P =.001). Twenty (87%) of the patients had good to excellent results and 3 patients (13%) had fair results. There were no poor results. Patients with fair results were rated primarily in the pain and function parts on the UCLA scale. Nineteen (82.6%) of the patients had a satisfactory return to the previous level of activities. The patient with a higher activity demand revealed a lower level of activity return (P =.034). The UCLA score and level of activity return were not related to the age of the patients, the type of anterior acromion, size of fracture, magnitude of displacement, Ellman grade of rotator cuff tear, and preoperative range of motion. Spearman's rank correlation coefficient showed that the level of activity return was related to the activity demand of the patients (γ =.628). Nonparametric test of Kruskal-Wallis demonstrated that a group of 8 patients with overhead sports activity achieved lower grade of return compared to the 7 patients without athletic activity (P =.034). Conclusion: The partial thickness rotator cuff tear should be considered in patients with chronic shoulder pain following a minimally displaced fracture of the greater tuberosity. Arthroscopic debridement or repair of the partial thickness rotator cuff tear as well as subacromial decompression, provide a reliable outcome for the majority of patients.

Dietmar Müller, Franz Landsiedl Vienna, Austria

Objective: Minimal invasive surgery suggests low complication rate. Arthroscopy of the shoulder has shown for the last decade to be an effective tool to resolve multiple joint pathology. Since few reliable is published on complications of arthroscopic surgery on the shoulder, the goal of our study was to investigate complications of shoulder arthroscopy (ASC) in a big cohort. Material and Methods: Between 1987 and 1998 eight hundred and fourtysix arthroscopies on the shoulder joint have been performed in our department. No routine antibiosis was applied. Most of the surgery (53%) was done to decompress the subacromial space (SAD) or to remove calcified deposits (CD) on the rotator cuff. 33% were performed to reconstruct joint stability (capsulorrhaphy = CR), in 60 patients (7%) arthroscopy was followed by open surgery (OS). Our project was a retrospective analysis of all 846 patients concerning a complicated perioperative course. Complications were classified as infectious, neurologic, technical, non surgical and immediate adhesive capsulitis. We correlated complications to length of O.R.-time and the type of arthroscopic surgery. Results: We had 48 complications on 44 patients (5.2%). Almost half (43%) of them were related to infectious etiology. Most of them appeared as temporary febrile temperatures or secondary wound healing problems. 13% had temporary neurologic impairment in the postoperative course. 15% developed immediate adhesive capsulitis of the shoulder joint and another 15% of the complications were due to intraoperative technical problems (e.g. instrument breakage). 2 patients (4%) had non-surgical complications. One had an allergic reaction to Cephalosporine, the other or acquired temporary laryngal nerve palsy after regional scalene anesthesia. In the subgroup of severe infectious complications (0.6 percent of the total cohort) the length of O.R. differed about 2-SD from the control group without complications. The group with a combined procedure (ASC + OS) had a significantly higher rise for complications (10 versus 5%). A correlation between more invasive surgery and higher complication rates could also be detected but showed no statistical significance. The least invasive type of ASC was the SAD with 2% complication rate, while CR had a almost 9% (8.82%) chance of perioperative complications. Conclusion: Arthroscopy of the shoulder is a safe procedure. Frequency and severity of complications are depending on type and length of surgery. Routine use of prophylactic antibiotics is recommended if arthrotomy ist performed after arthroscopy.

Philip A. Davidson, Dennis W. Rivenburgh St. Petersburg, Florida, USA

Purpose: To describe the utility and safety of a posterior-inferior (7 o'clock) shoulder arthroscopy portal. Methods: Six paired cadaver shoulders were utilized to arthroscopically create a 7 o'clock portal. Portals were created using both inside-to-out and outside-to-in techniques in each specimen. Specimens were prosected and subscapular and axillary nerve to portal distances were measured for each shoulder in six different positions, combining flexion, extension, abduction, and adduction. The angle of divergence from the skin incision to the nerves was also measured. Results: Combining normalized data, the distance from the 7 o'clock portal to the axillary nerve was 38.9 ± 3.9 mm, to the suprascapular nerve was 28.4 ± 2.4 mm, and ranged 18.0 to 61.4 mm. There was no statistically significant nerve to portal differential distance when the arm was placed in flexion, extension, abduction or adduction. The in-to-out technique produced a 7 o'clock portal approximately 5 mm further from both the axillary and suprascapular nerves than the out-to-in method. The angle of divergence from the 7 o'clock posterior portal skin incision to the axillary nerve was 47°, 33° to the suprascapular nerve. Conclusion: The 7 o'clock portal affords direct working access to the inferior capsular recess of the glenohumeral joint. Controlled, well-visualized access to the inferior capsular space can prevent inadvertent injury to the nearby axillary nerve. Creating this portal inside-to-out through an anterior 2 o'clock portal, using a Wissinger rod, protects the suprascapular nerve completely and lessens the chance of axillary nerve damage from misdirected posteriorly initiated instruments. This portal can also be safely created in an out-to-in manner. Distances to the axillary and suprascapular nerve, using both techniques, render this portal quite safe. Shoulder position during portal insertion does not affect portal to nerve distances or the relative safety of this portal.

John A. McCarthy, R. Michael Gross, John Schneider, Kristyn Dobler Omaha, Nebraska, USA

Introduction: The purpose of this study is to describe a constellation of factors/techniques that can lead to a delayed failure of Type II SLAP repair. These failures, especially when hardware avulsions occur, can significantly compromise the glenohumeral joint. Surgeons armed with this knowledge can avoid this potential complication.

Methods and Materials: 36 patients underwent 36 Type II SLAP repairs between May 1995 and December 1998. The surgical techniques, including arthroscopic placement of anchors with a variety of sutures, were reviewed. Associated pathology including rotator cuff tears, Bankart lesions, impingement, AC arthrosis and biceps tendinitis were noted. Rehab was dependent on associated pathology and patient progression postoperatively. Follow-up averaged 7 months.

Results: 26 patients had good to excellent results following surgery. 10 patients continued to have a significant pain and a repeat arthroscopy was performed. Figure 1:

Further analysis by type of anchor and sutures utilized will be discussed. The average time to the development of symptoms was 3 months.

Discussion: Failures of two types are seen. First, healing failures (6 cases) all occurred when mini anchors were utilized. In five cases, the anchors were attached to the biceps tendon with a nonabsorbable Ethibond suture. The anchor was pulled into the joint and impinged within the glenohumeral joint. In one case, the PDS suture failed before the SLAP healed. Revision SLAP surgery was necessary.

Four cases of bicipital tendinitis were noted at revision arthroscopy and responded to surgical management of such. There was no correlation between type of anchor and/or suture utilized with the development of bicipital tendinitis.

Conclusion: 1). The risk of hardware failure following a Type II SLAP lesion is significantly increased when a mini anchor with #2 nonabsorbable suture techniques are combined. 2). The failure occurs in a delayed fashion most likely due to the cyclical overloading of the anchor with a nonabsorbable suture. 3). The risk of suture failure alone and nonhealing of the SLAP is higher when absorbable sutures are utilized.

Ralph B. Garretson, III, Anthony A. Romeo Chicago, Illinois, USA

Introduction: Arthroscopic stabilization has become the accepted treatment for Type II superior labral anterior and posterior (SLAP) lesions. Short-term results using a variety of techniques have been promising, but most reports focus on motivated athletes. The purpose of this study is to report the outcome of arthroscopic stabilization in patients receiving worker's compensation. Methods: Twenty-two consecutive worker's compensation patients with Type II superior labral anterior and posterior (SLAP) lesions underwent arthroscopic stabilization between October, 1994 and December, 1996. A bioabsorbable tack was used for fixation early in the study; more recent patients have received suture anchors with nonabsorbable suture secured around the labrum for stabilization. Average age at surgery was 42 in 18 men and 4 women. Fifteen of 22 patients returned for clinical evaluation at an average of 37 months post-operatively. Associated lesions included: supraspinatus tears (3), subscapularis tear (1), and AC joint arthrosis (3). Twenty patients (91%) had an acromioplasty at the time of labral stabilization. Outcome was assessed by analysis of visual analog pain scales, range of motion, strength, Simple Shoulder Test (SST) and general health status questionnaires (SF-36), subjective satisfaction with the results of their surgery, and ability to return to work. Results: Pain scales demonstrated improved comfort, with an average decrease of 3 out of 10. Range of motion and strength were improved. Simple shoulder test responses showed improvement in 9 out of 12 categories. The SF-36 results showed significant positive changes in bodily pain, role-physical and physical ftmction. Eighteen of twenty-two patients (82%) were satisfied with their surgery. However, only 9 patients (41%) returned to their previous work, 6 patients (27%) returned to work but at less strenuous jobs, and 7 patients (32%) did not return to work. Discussion: Currently recommended treatment for Type II SLAP lesions is arthroscopic stabilization. When this procedure is performed in workers receiving compensation, objective parameters as well as patient self-assessment questionnaires show significant improvement. More than 80% of the patients were satisfied with the results of their procedure. However, less than half returned back to their previous jobs.

Brian Hood, Jerome V. Ciullo East Lansing, Michigan, USA

Subluxation of the long head of the biceps tendon from the inter-tubercular groove has been dealt with by surgical tenodesis to the humeral head or coracoid. Neither of these techniques lead to reproducible satisfactory results, nor do they allow normal function of biceps tendon as a humeral head depressor. Reimplantation within the inter-tubercular groove can restore normal function. In this retrospective study, the charts of 16 patients, who had surgical reimplantation of the biceps tendon between 1990-1995, were reviewed. Patients presented with asynchronous scapulohumeral rhythm, and occasional painful snapping in attempting to externally rotate the arm, compatible with subluxation of the biceps tendon. Ultrasound, MRI, or CT/arthrogram was able to establish the diagnosis in 12 of the patients, and four were identified by arthroscopy. All patients underwent open surgery where the subluxated tendon was identified under the stretched subscapularis tendon, was placed back in the groove, and the deficient medial wall abraded. A subcortical suture anchor allowed firm fixation of the subscapularis against the abraded area, with the remainder of the stretched tendon used to recreate the transverse humeral ligament. The patients were immobilized for four weeks post-operatively, except for pendulum exercises. No overhead activity was allowed for 6 weeks. Strengthening followed. No patients had residual pain or complaints of snapping with external rotation. University of California at Los Angeles Shoulder Rating Scale was used to compare the results pre and post-operatively. The average preoperative score was 12.1, with average postoperative score being 31.25, giving an average improvement of 19.75. Tenodesis of the long head of the biceps tendon subluxation does not restore the normal mechanics of the biceps tendon. More normal function can be re-established by reimplanting the tendon and anchoring it in place by re-establishing a medial buttress and anterior restraint.

Timothy A. Luke, Robert K. Peterson, Steven Yacoubian, Kevin D. Plancher New York, New York, USA

Purpose: To report 100 percent prevalence of the spinoglenoid ligament, the anatomy, morphology, and histology of the ligament. Secondly, this dynamic structure is involved in causing repetative trauma to the branch of the suprascapular nerve supplying the infraspinatus muscle. Dysfunction of the distal suprascapular nerve is reported in athletes involved in throwing and overhead sports. No study has been done to evaluate the dynamics of the space, created by the spinoglenoid ligament and occupied by the suprascapular nerve. Documenting a pressure change in this compartment and the relationship to shoulder position gives valuable information to the possible etiology of infraspinatus muscle atrophy and formation of ganglion cyst. The last part of this study is to verify the safety of performing arthroscopic release of the spinoglenoid ligament through the newly described Luke and Peterson technique. Methods: The spinoglenoid ligaments of fifty-eight fresh-frozen cadaver shoulders were dissected to evaluate the anatomic dimensions, histology, and relationship to the suprascapular nerve, posterior capsule, and glenoid rim. Fine calipers (Starrett Co., Athole, MA) accurate to 0.02 mm, were used to measure the spinoglenoid ligament. Spinoglenoid ligaments were harvested with their scapular spine and capsule/glenoid insertions intact and sent for histology analysis. The shoulder was placed through a range of motion to assess dynamic changes in the spinoglenoid ligament. Fifteen of the shoulders had a pressure detector (Microstrain, Burlington, VT) placed beneath the spinoglenoid ligament. The shoulder was taken through reproducible clinical range of motion and the change in pressure was recorded. Arthroscopic exam of six shoulders using the newly described Luke and Peterson two portal technique was done. The spinoglenoid ligaments were then dissected to verify the integrity of the suprascapular nerve. Results: Dissection revealed 100 percent (58/58) presence of the spinoglenoid ligament in the specimens. The ligament was found to form an irregular quadrangular shape. The mean maximal distance from ligament to bone was 6.61 mm (P =.431). The mean minimal distance from ligament to nerve was 4.63 mm (P =.302). On gross examination the ligament extends from the lateral scapular spine to the posterior glenoid and blends with the posterior shoulder capsule. Histology sections confirm the ligamentous nature of this structure with Sharpey's fibers inserting into bone at the scapular spine and blends with the posterior shoulder capsule to insert into the posterior surface of the glenoid. With glenohumoral motion the channel formed below the spinoglenoid ligament consistently experienced an increase in pressure when the shoulder was in full abduction. An increase in pressure was also recorded when the shoulder was internally rotated and positioned in flexion, extension, abduction, and adduction. External rotation did not give as consistant of pressure changes. Arthroscopic exam utilizing a two portal technique allowed easy identification of the suprascapular nerve and spinoglenoid ligament. The spinoglenoid ligament was safely released arthroscopically and the specimens were dissected to verify the intact branch of the suprascapular nerve to the infraspinatus muscle. Releasing the ligament diminished the increased pressure below the spinoglenoid ligament during shoulder range of motion. Conclusions: This study, contrary to previous reports, revealed the spinoglenoid ligament to be present in all dissected cadaveric shoulders with some variation in size. Histology studies confirmed the presence of the spinoglenoid ligament originating from the lateral scapular spine and inserting into the posterior shoulder capsule and glenoid. These findings support the relationship of the spinoglenoid ligament to the entrapment neuropathy of the distal suprascapular nerve. Dynamic study showed increased pressure on the suprascapular nerve when the cadaveric arm is internally rotated. The greatest pressure change came from the combination of abduction and internal rotation. Finally, a newly described technique, that is minimally invasive, was shown to release the spinoglenoid ligament arthroscopically and preserving the suprascapular nerve. This gives the athlete and surgeon another option in treating entrapment neuropathies of the suprascapular nerve.

Frank G. Shellock, Clarence L. Shields Jr Los Angeles, California, USA

Thermal-assisted chondroplasty using RF energy is utilized to treat grade II/III cartilage lesions. Since chondrocytes sustain thermal injury at excessive temperatures, the lowest RF power setting should be used that accomplishes the desired visual effect. Therefore, it is imperative to characterize the thermal and visual effects of RF energy as it is applied to articular cartilage. Purpose: To determine the temperature changes and visual alterations associated with RF energy-induced heating of bovine articular cartilage. Methods: RF energy was applied using an RF generator (VAPR System, Mitek Products, Westwood, MA) and a bipolar RF electrode (VAPR, Side Effect). Samples of bovine articular cartilage were used in this investigation. Cartilage samples were placed in a 37°C saline bath for application of RF energy. A fluoroptic thermometry probe (Luxtron, Model 3100) was positioned on the cartilage surface to record temperature at the tissue-RF electrode interface. The RF electrode was activated in a stationary position for 2 sec. at settings of V2-120, V2-60, V2-40, and V2-20 (n = 6). Temperatures were recorded at 1 sec. intervals for 5 sec. before, during, and for 13 seconds after application of RF energy. The cartilage samples were visually inspected to determine gross changes in appearance. Results: The highest average temperatures were: 78.5°C (V2-120), 62.6°C (V2-60), 58.1°C (V2-40), and 54.1°C (V2-20). There were statistically significant (P <.05) differences in temperatures associated with each of the respective RF settings. Visually, there was a gradation of effect seen for the changes in appearance of the cartilage, with the greatest alterations (i.e, color and morphologic changes) observed for the V2-120 setting. Little or no visual changes were produced by the V2-20 setting. Conclusions: These data provide RF power settings that may be considered to guide the application of bipolar RF electrode equipment for thermal-assisted chondroplasty.

Frank G. Shellock Los Angeles, California, USA

This in vitro study determined temperature changes associated with RF energy-induced heating of bovine capsular tissue using a bipolar RF electrode (3.0 mm, CAPSure; ArthroCare Corporation, Sunnyvale, CA) vs. a temperature-controlled, monopolar RF electrode (TAC-S ElectroThermal Probe; Oratec Interventions, Menlo Park, CA). Tissue samples were placed in a saline bath (37°C) and RF energy was applied at recommended settings (i.e., base on clinical experience and published literature) for tissue shrinkage procedures. Fluoroptic thermometry was used to record temperatures on the tissue surface and at depths of 1 mm and 2 mm during continuous delivery of RF energy at 1, 2, 3, and 4 sec. time increments. The highest mean temperatures were recorded on the tissue surface, as follows (M ± SD):

There was a statistically significant increase in temperature for each time increment noted for the use of the bipolar and monopolar RF electrodes. Additionally, there was a good linear relationship between mean surface temperatures vs. time for each of the RF electrodes. Notably, the temperature-controlled, monopolar RF electrode did not appear to properly regulate the deliver of RF energy to maintain tissue temperatures at the selected level (i.e., 65°C) for the 3 and 4 sec. time increments. The bipolar RF electrode produced a smaller temperature gradient at the 1 mm tissue depth compared to the monopolar RF electrode. Temperature profiles at the 2 mm tissue depth were comparable for the two different types of RF electrodes. These data provide basic information pertaining to the temperature profiles produced by bipolar vs. monopolar RF electrodes applied to collagen-based tissue.

Greg S. Keller, Eric Ledet, Mike Tymeson, Richard F. Pell IV Scottsdale, Arizona, USA

Introduction: Capsular shrinkage is an investigational method in arthroscopic shoulder stabilization with little basic research to justify its use or understand its implications in the immediate postoperative period. There have been no reports of mechanical properties in thermally treated human glenohumeral joint capsule using radiofrequency as an energy source. It is the purpose of this study to characterize the mechanical properties of thermally treated shoulder capsule. Methods: Six fresh frozen human cadavers yielded six pair of 4.5 mm wide bone-ligament-bone specimens from the anterior inferior glenohumeral ligament for paired comparison. Of each pair, one was randomly assigned to be thermally treated and one was assigned to the control group. All specimens were potted in PMMA or epoxy, placed in the self aligning fixture, and tested on a MTS machine in uniaxial tension at 2.5 mm/minute to failure. Deformation was measured using crosshead and a differential variable reluctance transducer. Load-deformation curves were generated. We compared the thermally treated group to the control for peak load, strain at peak load, stiffness, and length. Statistical analysis was performed with a paired t-test. Results: Strain at peak load (P =.035) and length (P =.009) after thermal treatment showed significant difference. The average ligament lost 11% of its original length after thermal treatment. However, stiffness and ultimate strength showed no statistically significant difference. At peak load, the treated ligaments strained as much as the control and, in addition, returned 90% of the length lost by thermal treatment nearly nullifying the “shrinkage effect.” Discussion and Conclusion: Thermally treated shoulder capsule has increased strain at peak load without demonstrable change in peak load or stiffness. Since the ligament returns to its insufficient length at peak load, we suggest avoidance of aggressive rehabilitation and early return to sports until the tissue is able to remodel. Radiofrequency application produced consistent shrinkage comparable to previous studies. Our specimens exhibited similar behavior to a concurrent and recently reported study using lasers. For surgeons without access to lasers, this appears to be a consistent and viable alternative for performing capsular shrinkage. In view of the strain behavior of shrunken ligaments, further research needs to be done to understand creep behavior and the effects of tissue remodeling on strain characteristics of shrunken ligaments.

Cesar Ceballos, John E. Zvijac, John W. Uribe, Keith S. Hechtman Coral Gables, Florida, USA

Introduction: Multidirectional instability continues to pose a challenge in patients who have not adequately responded to rehabilitative management. Neer and Foster first addressed the concept of multidirectional shoulder instability (MDI) in their original article in 1980, treated with open inferior capsular shift procedures. These open procedures are associated with lengthy postoperative rehabilitation periods and a significant number of overhead athletes never return to pre-injury levels of performance. When compared to open and other arthroscopic procedures described in the literature, our results demonstrate lower morbidity, less rehabilitation time, and a more rapid return to strenuous overhead activities when using arthroscopic thermal capsulorrhaphy for treating MDI. Our objective is to present findings on 23 patients with MDI that were treated with arthroscopic thermal capsulorrhaphy after not responding to an adequate rehabilitative program. Materials and Methods: Twenty-three consecutive patients with involuntary multidirectional glenohumeral instability were treated with arthroscopic thermal capsulorrhaphy. Each of these patients had failed an adequate trial of rehabilitative management. There were 14 males and 9 females. The average age was 25 years. The dominant extremity was involved in 14 patients. Three patients had bilateral multidirectional instability. Two of these patients had prior open capsular shifts to the other shoulder and one had no previous surgery. Twenty-two out of 23 patients had history of either repetitive injury or acute trauma as an initial factor in the development of the instability. Postoperative follow-up ranged from 12 to 22 months. Results: All 23 patients were evaluated 12 to 22 months postoperatively and rated according to both the Rowe shoulder scoring system and the satisfactory/unsatisfactory system of Neer. The preoperative Rowe score ranged from 30-75 with a mean of 56. The average postoperative Rowe score was 97 (range 90-100). All 23 patients had a satisfactory result according to the Neer system. All patients resumed their daily activities without discomfort and all the athletes returned to their previous level of athletic participation. Shoulder range of motion was preserved in 21 of 23 patients. No patient has had a recurrent episode of instability. Two patients experienced rotator cuff tendinitis at 4 and 5 months postoperatively. There were no infections or neurovascular injuries. Discussion: Findings in our patients treated with thermal capsulorrhaphy for multidirectional instability demonstrate that all patients had excellent results on the Rowe shoulder scoring system. With the Neer system all our patients demonstrated satisfactory results. All of our overhead athletes were able to return to strenuous overhead activities and perform at pre-injury levels of performance. In this clinical review of 23 patients treated with thermal stabilization, there is a lower morbidity rate, earlier return to work and sport activities, and a higher patient satisfaction rate than with other open or arthroscopic shoulder stabilization procedures described in the literature. We continue to follow these particular patients as well as add to this present study group.

Thomas J. Noonan, Marilee Horan, Richard J. Hawkins Vail, Colorado, USA

Introduction: The use of radiofrequency thermal capsulorrhaphy in the treatment of shoulder instability has become more prevalent. The purpose of this study was to determine patient outcome following radiofrequency thermal capsulorraphy. Methods: Between 1997 and 1999 the senior author has used thermal capsulorrhaphy in 121 patients. Objective and subjective data were collected prospectively to determine the patient outcome. The inclusion criteria for this study was shoulder instability treated with arthroscopic radiofrequency thermal capsulorrhaphy and minimum one year follow-up. Twenty-five patients (13 anterior, 7 MDI, 4 posterior, 1 anterior/posterior) met these criteria. Average age was 28 (range 16-44). Average follow-up was 14.6 months (range 12-21 months). Results: Surgical findings included Bankart lesions in 6 of 13 anterior instability cases and 4 of 7 MDI cases. All but one repaired. SLAP lesions were found in 2 anterior cases and 2 MDI cases. They were repaired in the anterior cases and debrided in the MDI cases. One anterior instability patient had a partial thickness rotator cuff tear and impingement and underwent debridement of the tear with a subacromial decompression. Post-operative physical exam showed one patient with loss of 20° of flexion, one with loss of 10° of external rotation, and one with a reduction in internal rotation three vertebral levels. Strength was full in flexion, external rotation, and internal rotation in all patients. ASES score improved from an average of 71 ± 19 pre-operatively to 94 ± 12 post-operatively (P <.01). Rating pain on a scale from 1 to 10, there was improvement from a mean of 7.9 ± 0.6 preoperatively to 3.0 ± 0.8 post-operatively (P <.01). Post-operatively, 75% (15/20) had no or mild pain with recreational activities, compared with 6% (1/16) pre-operatively. No patients dislocated post-operatively. Pre-operatively, 76% (13/17) felt instability occasionally or frequently compared with 14% (3/21) post operatively. Post-operatively, 81% (17/21) can painlessly perform activities above the neck, compared with 33% (4/12) pre-operatively. Rating their satisfaction on their outcome on a scale of 1-10, with 10 being very satisfied, the average was 9.0. There were three cases classified as failures due to recurrent subluxations (2 MDI, 1 posterior). Conclusions: Our results suggest that radiofrequency thermal capsulorrhaphy is a promising procedure for patients with should instability. The treatment of anterior instability appears to be the most effective, our initial results are comparable to those obtained with open surgery. By contrast, our limited results with MDI patients have shown a failure rate exceeding that of open surgery techniques. Further follow-up is necessary to determine if these trends remain and also ensure the long-term success of this procedure.

Dev K. Mishra, Gary S. Fanton Oakland, California, USA

Introduction: Arthroscopic treatment of chronic anterior shoulder instability in active athletes is reportedly less successful than open techniques. We performed arthroscopic stabilization for confirmed capsulolabral avulsions and followed the patients prospectively for a minimum of 24 months. Methods: We studied 42 patients (mean age 26 years) with recurrent traumatic anterior dislocations. They reported an average of 9 dislocations preoperatively. An average of 69 months elapsed from initial dislocation to surgery. The shoulders underwent arthroscopic capsulolabral repair with either a suture anchor and horizontal mattress suture, or an absorbable tack. Each shoulder also was treated with a monopolar radiofrequency probe for thermal shrinkage of the middle, anteroinferior, and posteroinferior glenohumeral ligaments. Patients were evaluated prospectively for pain, motion, stability, and function using the modified Rowe score. Results: At a mean of 28 months postoperatively, 38 patients had returned to their preinjury sports, including four patients playing Division I NCAA football. Three patients had a traumatic redislocation (7% dislocation). Using the modified Rowe score statistically significant improvements were noted for pain, stability, and function. There was no significant change in motion. The overall modified Rowe score improved from 38 points preoperatively to 89 points at final evaluation (P <.001). Conclusion: These results indicate that arthroscopic treatment for recurrent traumatic anterior instability yields results comparable to open procedures, including those athletes involved in high level contact and collision sports. We believe that addressing capsular volume surgically is critical, particularly when dealing with chronic instability. This procedure allows the surgeon to reliably correct the labral detachment and the increased volume, while preserving motion and minimizing morbidity.

Stephen S. Burkhart, Joe DeBeer San Antonio, Texas, USA

Arthroscopic Bankart repairs have a reputation for having a higher recurrence rate than open stabilization procedures. We set out to analyze our arthroscopic Bankart repairs in order to identify specific factors related to recurrence. We evaluated the results of 194 arthroscopic Bankart repairs utilizing a suture anchor technique, and particularly analyzed the 21 failures (redislocations) in this series. Fourteen of the twenty-one redislocations had significant bone defects (large “engaging” Hill-Sachs lesions or large bony Bankart lesions with an “inverted pear” configuration to the glenoid) that compromised mechanical structural integrity of the shoulder. Our “adjusted” recurrence rate (for shoulders without significant bone defects) was 4.0%. On the other hand, we had 21 patients in the entire series with significant bone deficiencies, and 14 of these had recurrent instability for a 67% recurrence rate in shoulders with bone deficiency. We conclude that arthroscopic Bankart repair provides excellent results (4.0% redislocation rate) that are comparable to results of open Bankart repair if there are no structural bone defects. The problem with arthroscopic Bankart repair is not with the quality of soft tissue reattachment, but with the inability to address bone deficiencies on either side of the shoulder joint. A large “engaging” Hill-Sachs defect is best treated by open capsular shift to limit external rotation or by bone grafting of the defect to lengthen the articular arc. A bony Bankart lesion that creates an “inverted pear” configuration to the glenoid (inferior half of glenoid narrower than superior half) is not amenable to arthroscopic repair, and we recommend vascularized anterior glenoid bone grafting by means of the Latarjet procedure. One must be cognizant of the possibility of humeral avulsion of the glenohumeral ligaments which may be repaired by arthroscopic or open means.

Jeffrey H. Berg, Raymond Thal, Junichi Tamai Reston, Virginia, USA

Introduction: Open and arthroscopic procedures performed to correct shoulder instability have incorporated a shifting of the capsule to reduce the capsular laxity and a repair of the Bankart lesion when present. Despite the common use of these various techniques, no study has compared the degree of capsular shift achieved by their use. Purpose: Compare the degree of capsular shift obtained following common medially based open and arthroscopic Bankart repairs as well as that of a newly described method of fixation. Methods: Six fresh frozen cadaveric shoulders were dissected in such a manner that the humerus was removed and the capsule was left intact and attached to the glenoid. After the glenoid was marked off as a clock face, a Bankart lesion was created from 2 o'clock to 6 o'clock and a marking suture was then placed in the inferior capsule 30 to 40 mm directly inferior to the 3 o'clock position. The distance from 3 o'clock to the marking suture was measured with a digital caliper (Ultra-Cal Mark III, Fowler) and recorded for each specimen. By using passing sutures, identical points of capsular fixation were utilized for all repairs. Four different medially based repairs were then performed. The repair techniques consisted of 1) a Bankart repair utilizing suture anchors and knot fixation, alone, 2) supplemented with a barrel stitch, 3) supplemented with a capsular tuck, and 4) a Bankart repair utilizing The Knotless suture anchors (Mitek Products, Westwood, MA. The degree of capsular shift following each repair was measured three times. Results: The average shift following each technique was:

Conclusion: All techniques evaluated provide a degree of capsular shift. Greater shifts can be achieved by supplementing the standard suture anchor repair with either a barrel stitch or capsule tuck. The shift achieved using The Knotless suture anchor alone was significantly greater than the shift using the standard suture anchor alone. The Knotless suture anchor appears to pull the capsule into the glenoid drill holes plicating the tissue and thus achieving a shift that is comparable to the supplemented repairs.

Mark R. Hutchinson, Mitchell Cooper Chicago, Illinois, USA

The microscopic pathoanatomy of acute anterior glenohumeral dislocation was studied in a cadaver monkey model. Three monkey shoulders were dissected and used as controls to evaluate the anterior glenohumeral capsule. Eleven other shoulders were dislocated anteriorly with forced external rotation in an abducted position. The anterior glenohumeral capsule was then evaluated grossly and histologically. The dislocated shoulders all had an avulsion of the glenoid labrum; a.k.a., a Bankart lesion. All of the histologic specimens from the dislocated shoulders had multiple diffuse tears of the collagen fibers. The tears were oriented perpendicular to, obliquely to, or demonstrated separation parallel to the longitudinal orientation of the fibers without a consistent pattern or location. The clinical application of these findings suggest that current open surgical techniques for treating anterior shoulder instability which include straight longitudinal reefing of the capsule, transverse reefing of the capsule, and reinserting glenohumeral ligaments and capsule in a more superior position on either the humerus or the glenoid, fail to completely address the histologic damage seen in multiple and not a single direction of fiber orientation.

Jörn Steinbeck, Axel-Kai Witt, Wolfgang Pötzl, Mathias Ritsch Münster, Germany

Question: How does strength develop in patients after different surgical techniques of glenohumeral instability in comparison to healthy. Patients and Methods: The following groups of patients were tested for isokinetic strength measurement by the Cybex 6000: 20 patients with healthy and 15 with traumatic glenohumeral instability, 16 patients following open Bankart repair and 13 patients following arthroscopic Bankart repair. The patients were divided according to those with a non-injured, dominant site (ndi) and those with an injured, dominant site (di). Also were tested the maximal torque, the average work (joule) and explosive work (joule). For all measurements the standard angle movement were 60°/second and 180°/second. Results: A overall decrease was found in the torque and work/bodyweight in flexion, abduction and external rotation. The non-dominant injured site numbers reached only a low percentage in comparison to the other movements of dominant site. Improvement in the non-dominant injured group was evident during slow speed movements, in contrast to the dominate injured group where less or no improvement was found. Patients after open Bankart repair compared to patients with glenohumeral instability showed a increase of the torque/weight relationship in 60° of the non-dominant injured and a decrease of the dominant injured group. Moreover, an increase was detected of torque/weight relationship in 180° in extension and adduction in the dominant injured and in internal rotation in the non-dominant injured group. In comparing patients having had the arthroscopic Bankart procedure to patients with glenohumeral instability in torque/weight relationship 60°, an improvement was found in all movements of the non-dominant injured group and a deterioration found in the dominant injured group. Conclusion: Isokinetic strength development evidences clear improvement in patients after arthroscopic Bankart procedures in comparison to patients with glenohumeral instability and after open Bankart repair. Additionally, there appears to be a better recovery of the strength development of the non-dominant injured site in the patients following arthroscopic and open Bankart procedure compared to patients with glenohumeral instability. Less improvement of strength development was found among the dominant injured of the last group.

Sohail Ahmad, Frank Pettrone, Greg Montalbano, Leigh Brezenoff Arlington, Virginia, USA

Introduction: Few long term studies have evaluated the efficacy of shoulder stabilization using arthroscopic or open procedures. Our purpose is to directly compare outcomes of shoulder stabilization using arthroscopic transglenoid suture stabilization versus open capsulolabral repairs. Methods: 58 patients (38 with arthroscopic stabilization and 20 with open stabilization) were followed for a minimum of 5 years (Mean = 68 months). Patients with less than 5 years of follow-up were excluded. Pain, motion, strength, function, stability, satisfaction, radiographs, and ability to return to pre-injury activity levels were evaluated. Results: 30 arthroscopic and 20 open repairs were available for a minimum of 5 years. The arthroscopic group had a 17% (5/30) dislocation rate and a 60% (18/30) rate of instability. 67% (6/9) participating in collision sports had repeat instability. 8 of the 18 with instability required an open capsulolabral repair for persistent symptoms. There was no limitation of motion following arthroscopic stabilization. The open group had no dislocations and a 5% (1/20) rate of instability or pain. 45% (9/20) had some loss of external rotation (range from 10-40 degress) with a mean loss of 18 degress. There was one failure who lost 40 degress of external rotation and had moderate pain with recreational activities. Excluding this patient, mean ROM loss was 15 degrees. No patients (0/7) participating in collision sports had repeat instability. There was no evidence of glenohumeral arthrosis. Discussion and Conclusion: Few long term studies have directly compared arthroscopic versus open shoulder stabilization. In this study, we show improved stability and a lower dislocation rate in the open group. However, external rotation was better preserved in the arthroscopic group. Almost half of the open patients had some loss of external rotation. We recommend open stabilization for patients participating in sports (i.e. football, wrestling) or activities where better stability is required. For those participating in sports (i.e. swimming) which may not tolerate some loss of external rotation without compromising performance, the arthroscopic technique has been modified.

Stephen C. Weber Sacramento, California, USA

Instability leading to secondary impingement was first described by Jobe to explain the symptoms of the painful athlete's shoulder. He described type II impingement as patients with increased laxity without true dislocation with impingement type pain and arthroscopic evidence of anterior labral change and/or posterior humeral head changes, and type III impingement as patients with impingement pain and increased laxity without dislocation, labral or humeral changes. Other authors have since labeled these patients as “edge loaders.” Jobe's recommendation was for open repair; other authors have described arthroscopic suture repair, but commingling true dislocators with “edge loaders” has made interpretation of these preliminary report difficult. Recent reports have suggested thermal techniques done arthroscopically as an alternative to these open techniques, implying superior results and less morbidity; however no peer reviewed published reports exist, and presented reports have mixed suture and thermal techniques. Anecdotal reports of axillary nerve injury, capsular necrosis, and severe stiffness have been noted with thermal techniques. In an effort to provide a traditional suture technique benchmark to thermal capsullorhaphy, 29 patients with type II or III impingement treated with an arthroscopic capsular shift using suture technique without thermal treatment were studied over a 4 year period. No patient had a true Bankart lesion or history of dislocation; all had a positive lidocaine impingement test with at least a 1+ sulcus sign and positive drive through sign. Mean age 27.45 followup averaged 30.45 months with minimum 12 months. Patients were treated with arthroscopic capsulotomy followed by anterior capsular shift and suture repair using #1 PDS suture as an outpatient procedure. Sling immobilization was continued for three weeks, followed by physical therapy at six weeks; a graduated throwing program was started for throwing athletes at three months. No complications occurred, specifically no stiffness, infection, nerve injuries or revision surgery was encountered. UCLA scores improved from 23 to 32 with 21 excellent, 4 good, 3 fair, and 1 poor. 5 of 6 competitive throwers were able to resume throwing sports unrestricted, 4 pitching at the intercollegiate level. Arthroscopic suture capsular shift appears to provide reasonable results for this difficult population failing conservative management. The additional unknown long term risks of thermal techniques must be weighed against the results shown here using established arthroscopic techniques.

Anthony A. Romeo, Dominic Carriera Chicago, Illinois, USA

Introduction: Arthroscopic Bankart repairs are associated with an increased incidence of recurrent instability when compared with open Bankart repairs. Multiple factors have been associated with their failure, including methods of fixation and an inability to appropriately tension the capsule. The purpose of this study was to prospectively evaluate the minimum 2-year results of arthroscopic Bankart repairs using a suture anchor technique that closely mimics successful open procedures. Methods and Materials: A consecutive series of 32 patients (24 males, 8 females, average age of 25) with Bankart lesions were treated with an arthroscopic repair using suture anchors loaded with nonabsorbable suture and placed 2 millimeters into the edge of the glenoid articular surface. Thirty patients were evaluated at a minimum follow-up of two years. A 5 o'clock portal through the subscapularis tendon was used for placement of the lowest suture anchor on the glenoid rim at the 5:30 position. On average, 3.5 anchors were used for the procedure. Most common reason for dislocation: sports participation. Four patients had extension of the labral avulsion into the superior labrum (SLAP). Results: No patient experienced a recurrent dislocation. Two patients had pain with the relocation test. The average Rowe score was 92/100 points, with 96% of the results excellent or good. Simple Shoulder Test responses improved from 66% positive pre-op to 91% post-op. The ASES scoring index was 96/100 points. Pain analog scales improved from 5.5/10 preop to 0.35/10 postop. SF-36 scores were improved for physical function. Patient satisfaction was 8.9/10. There were no complications related to the low 5 o'clock portal through the subscapularis. Clinical testing (Lift-off test) of the subscapularis was normal. Discussion: Arthroscopic Bankart repairs with properly placed suture anchors and nonabsorbable sutures are associated with results similar to open techniques including stability, improved scores on shoulder-specific outcome measures (Rowe, SST, ASES), general health status measures (SF-36), and patient satisfaction. The low anterior (5 o'clock) portal is safe and provides direct access to the inferior glenoid rim.

Seung-Ho Kim, Kwon-Ick Ha, Sang-Hyun Kim Seoul, Korea

Introduction: The purpose of the present study is to analyze the results of open and arthroscopic Bankart repair using the suture anchors in the traumatic unilateral anterior glenohumeral instability. Methods: Eighty-nine shoulders in eighty-eight patients with traumatic unilateral anterior instability were evaluated for Rowe and UCLA scores, recurrence, return to activity, and range of motion by an independent examiner after either open (30 shoulders, average age, 27.6 years) or arthroscopic (59 shoulders, average age, 26.7 years) Bankart repair using suture anchors. Arthroscopic technique included a minimum of 3 anchors, and a routine incorporation of capsular plication and proximal shift. Open Bankart repairs were performed during the initial time while the arthroscopic procedures were performed during the latter time of the index period. The average follow-up was 49 months (range, 41 to 60) in the open repair group and 33 months (range, 26 to 42) in the arthroscopic repair group. Fourteen (46.7%) in the open group and thirty-two (55.2%) in the arthroscopic repair group were involved in active sports activities. The results were compared between the two groups using the Mann-Whitney U test. Spearman's correlation coefficients were used to identify any significant relationships between several variables. Nonparametric test of Kruskal-Wallis was performed to determine the differences in follow-up scores between the different conditions of the anterior labrum. Results: Twenty-six shoulders (86.6%) had good or excellent results, 2 (6.7%) fair, and 2 (6.7%) poor in the open Bankart repair group. In the arthroscopic repair group, 54 (91.5%) had good or excellent results, 3 (5.1%) fair, and 2 (3.4%) poor. The average Rowe and UCLA scores were 90.4 points (range, 30 to 100) and 30.6 (range, 20 to 35) respectively for the open repair group, while 92.7 (range, 40 to 100) and 33.1 (range, 18 to 35) respectively for the arthroscopic repair group. Arthroscopic group revealed a significantly better results in the Rowe and UCLA scores (P =.041). Two shoulders in each group developed redislocation. There were no differences in the loss of external rotation and return to prior activity between the two groups (P >.05). A significant positive correlation was found between the Rowe score and the age at the initial dislocation in the arthroscopic repair group (γ = 0.807). Fifteen patients younger than 20 years of age at the initial episode of dislocation demonstrated lower scores (84.9 points) than 43 patients at 20 of age and older (95.7 points, P =.029). The final results were not related to other variables such as rotator cuff tear, number of dislocation, grade of translation under anesthesia, presence of generalized laxity, Hill-Sachs lesion. No difference was found in the Rowe and UCLA scores between the different conditions of the anterior labrum (P >.05). Conclusion: Arthroscopic Bankart repair using suture anchor technique which includes a minimum of 3 anchors, a routine incorporation of capsular plication and proximal shift together, has results that are competitive or even better than the open Bankart repair using suture anchors.

Craig R. Bottoni, Thomas M. DeBerardino, John H. Wilckens, Jean-Claude D'Alleyrand West Point, New York, USA

Purpose: Non-operative treatment for the initial anterior shoulder dislocation results in a recurrence rate of greater than 80% in young athletes in two previous studies at our institution. The objective of this study was to evaluate the results of arthroscopic stabilization against conventional, nonoperative management of acute, traumatic, first-time shoulder dislocations using a bioabsorbable tack in young athletes. Methods: Patients between the ages of 18-26 years who presented to our orthopaedic clinic following a traumatic, initial anterior shoulder dislocation which required reduction by medical personnel were included in the study. MRI studies were obtained on all shoulders. Following informed consent to participate in this study as approved by our Human Use Committee and Institutional Review Board, patients were randomized to the operative or nonoperative group. The operative group underwent arthroscopic stabilization of the capsulolabral injury with a bioabsorbable tack. The patients in the nonoperative group had a 3-week period of immobilization followed by supervised rehabilitation. Results were determined using follow-up physical exams noting a second dislocation event or recurrent subluxation resulting in a need for open reconstruction as treatment failures. Follow-up evaluations were done using L'Insalata questionnaires, SANE evaluations, functional status and patient satisfaction. A Student's t-test (α < 0.05) was used to compare failure rates between groups. Results: From November 1994 to April 1998, 24 patients sustained an acute, initial, traumatic, anterior shoulder dislocation and consented to study participation. MRI revealed Bankart lesions in all 24 patients (100% accuracy confirmed in operative group). Fourteen patients were randomized to non-operative treatment. Ten patients with anterior dislocations were treated with arthroscopic primary repair of their capsulolabral injury (Bankart lesion present in 100%) with bioabsorbable tacks (2 in 9 patients, 3 in one). All patients were male with an average age of 22 years (range, 18 to 26) and an average follow-up of 36 months (range, 16 to 56). Three of 24 patients were lost to follow-up. Nine of twelve patients (79%) from the nonoperative group failed while only one of nine (11.1%) from the operative group failed by the established criteria. There were no intraoperative complications. The SANE and L'Insalata scores averaged 88 and 94 for the operative group and averaged 57 and 73 for the nonoperative group (P <.002). Conclusion: Arthroscopic stabilization of acute, traumatic, first-time anterior shoulder dislocations with a bioabsorbable fixation device is an effective and safe treatment that significantly reduces the recurrence rate in young athletes when compared to conventional, nonoperative treatment. With follow-up over 3 years, we have observed a reversal of the natural history for this injury in young athletes.

Jeffrey H. Berg, Raymond Thal, Junichi Tamai Reston, Virginia, USA

Introduction: Open and arthroscopic procedures for the correction of shoulder instability have traditionally employed a repair of the Bankart lesion, when present, as well as a shifting of the capsule in order to reduce capsular patholaxity. Often in cases of longstanding instability, significant capsular laxity may exist requiring an even greater capsular shift. Purpose: (1) Describe the capsular resection technique, a medially based technique for correction of shoulder instability associated with significant capsular laxity and (2) Compare the degree of capsular shift following a medial capsular resection and Bankart repair to a Bankart repair using either standard suture anchors with knot tying or The Knotless suture anchors (Mitek Products, Westwood, MA), alone. Methods: Six fresh frozen cadaveric shoulders were dissected in such a manner that the humerus was removed and the capsule was left intact and attached to the glenoid. After the glenoid was marked off as a clock face, a Bankart lesion was created from 2 o'clock to 6 o'clock and a marking suture was then placed 30 to 40 mm in the inferior capsule directly inferior to the 3 o'clock position. The distance from 3 o'clock to the marking suture was measured with a digital caliper (Ultra-Cal Mark III, Fowler) and recorded for each specimen. Three different medially based repairs were then performed. The repair techniques consisted of 1) a Bankart repair utilizing standard suture anchors and knot fixation, 2) a Bankart repair utilizing The Knotless suture anchors, and 3) a Bankart repair utilizing The Knotless suture anchors following a 5 mm medial capsular resection from 3 o'clock to 5 o'clock. The degree of capsular shift following each repair was measured three times. Results: The average shift following each technique was:

Conclusion: A statistically significant increase in capsular shift following a Bankart repair is observed when using The Knotless suture anchors as opposed to standard suture anchors with knot tying. The medial capsular resection technique provides an even greater capsular shift, statistically significant from even The Knotless suture repair, alone.

Carlos A. Guanche, Donald C. Quick Minneapolis, Minnesota, USA

A physical examination maneuver which definitively diagnoses a glenoid labral tear has been elusive. This prospective study correlates O'Brien's test and the crank test with findings at shoulder arthroscopy.

Forty-six patients (87% male), with an average age of 37 yrs (range 17-77), and who were about to undergo arthroscopy of the shoulder were examined prior to the administration of any anesthetic or narcotic agent. All subjects underwent O'Brien's test (including localization of the area of pain) and the crank maneuver. The clinical examination results were compared to the surgical findings and analyzed for sensitivity, specificity, positive and negative predictive values in the diagnosis of a glenoid labral tear. Results are presented in the following table:

We conclude that, while clinical diagnosis of a labral detachment can be made using the O'Brien's or crank tests, neither will definitively indentify the lesion. Furthermore, combining the two tests does not markedly increase reliability of the diagnosis. The importance of our results, unlike other published studies, relates to the fact that this was a heterogeneous group with a variety of pathologies. Based on our study, it would appear that the definitive physical examination maneuver diagnosis of a labral tear is still to be found.

Stephen Fealy, T. Peter Kingham, David W. Altchek New York, New York, USA

Purpose: Objective evaluation of rotator cuff repairs (RCR) demonstrate an ~30% rate of failure of secure healing between tendon and bone at 3-5 yrs post-op. Failure has been noted in both open and mini-open RCR. Mini-open RCR has been used successfully to treat small and moderate sized cuff tears, but not large tears (>3 cm < 5 cm). We have developed a two-row fixation technique of repairing the torn cuff to the humerus. The tendon is repaired to bone using a lateral and medial row of suture anchors. This bone-tendon interface produces a broad contact area at the repair site. (Figure) We sought to determine if this technique would produce successful results in patients with larger rotator cuff tears. Materials and Methods: The results of 89 consecutive patients who underwent arthroscopic subacromial decompression with acromioplasty and mini-open RCR were evaluated. All patients were evaluated at a minimum 24 months post-op (mean 27 mos). All patients were treated with a mini open RCR using a two-row fixation technique. Patients were evaluated at follow-up with a focused shoulder examination, a modified short-form (SF-36) outcomes analysis, HSS Shoulder Questionnaire, and the UCLA Shoulder Rating Scale. There were 53 male and 36 female patients with a mean age of 54 years (range 34-72 years). Results: All patients failed a pre-op course of physical therapy and non-operative management. 32 patients were found to have a large rotator cuff tear (Group I), 39 were found to have a moderate sized tear (Group II), and 18 patients had a small tear (Group III) seen at the time of arthroscopy. There was no statistically significant difference in outcome between groups. The mean UCLA Shoulder Rating Scale for all groups was 33.6/35. There was no difference in outcome when data was analyzed according to age, sex, hand dominance, and mechanism of injury. The mean time from surgery to full recovery was seven months. Patient's subjective satisfaction based on their pre-injury level of performance at most recent follow-up was 92.6%. 83% (74/89) of patients returned to their pre-injury activity; 93% of patients reported that they would undergo the procedure again. Conclusions: We believe that mini-open RCR utilizing a two-row fixation technique can yield excellent results in patients with small, moderate, and large rotator cuff tears. There was no detectable difference in outcome between groups at two years follow-up. We believe that the mechanical strength of the RCR is directly related to the cross-sectional area of the bone-tendon interface achieved at the time of surgery. We are currently investigating this hypothesis in an animal model at our institution.

Edward S. Bittar Palm Bay, Florida, USA

Objective of Study: Evaluation of the role and effectiveness of arthroscopy in the management of massive retracted rotator cuff tears (RCT). Materials and Methods: Seventy-four patients, ranging in age from 54 to 82 years, with massive (15 cm² or greater), retracted RCT underwent out-patient arthroscopic debridement, subacromial decompression and mobilization of the rotator cuff. The superior, anterior and posterior capsular recess was arthroscopically released to allow mobilization and anatomic reduction of the free rim of the avulsed rotator cuff. The RCT was then repaired with the arm in neutral using a mini-arthrotomy approach. Post-operatively, the RCT repair was protected with six weeks of full, passive range of motion. Active range of motion was initiated after six weeks, followed by progressive resistive exercises. Results: Sixty-one patients (82%) recovered active, functional and pain-free shoulder range of motion post-operatively after four months of aggressive rehabilitation. Thirteen patients (18%) complained of persistent pain post-operatively. Six of the thirteen patients with persistent pain underwent imaging confirming re-tear of the repaired RCT. Nineteen patients required up to eighteen months of supervised post-operative strengthening before restoring normal torque strength and function to their involved shoulders. Patients were followed clinically for at least three years and up to eight years. Results did not diminish with time. Conclusion: Arthroscopically-assisted repair of massive (15 cm² or greater) retracted RCT provides substantial advantages over open repairs. The glenohumeral joint can be evaluated and debrided arthroscopically prior to treatment of rotator cuff pathology. The character of the RCT can be evaluated and the torn tissue fully mobilized prior to performing an arthrotomy or a bony decompression of the acromial arch. Extensile arthrotomies are obviated completely. Arthroscopic mobilization and reduction confirms that the RCT can be reduced and repaired anatomically prior to altering shoulder mechanics with an arthroscopic bony subacromial decompression and Mumford procedure. Only the final RCT repair requires the use of a mini-arthrotomy approach. Mobilization and repair of a massive, retracted RCT can be performed with arthroscopic assistance with substantial reduced morbidity and with good or excellent results in 82% of patients. Extensile arthrotomy for exposure is eliminated and post-operative recovery is markedly enhanced. Shoulders with an irreparable RCT can be debrided arthroscopically while providing a measure of pain relief to patients with minimal morbidity.

Peter Bonutti, Matthew J. Cremens, Frank Lee, Timothy J. Gray Effingham, Illinois, USA

Introduction: Rotator cuff repair is increasingly being performed with metallic anchoring/fixation systems. These systems have variable degrees of mechanical integrity and are associated with complications if pull out or bone remodeling occurs. Biodegradable anchoring systems also are available however there is concern of osteolysis and wear from the degradation products. Because of this, allograft bone was investigated as an alternative material for tissue fixation to bone were examined. Methods: Allograft buttons were cut from cortical bone into 3 mm diameter by 6 mm long cylinders. Two transverse holes were drilled in the button in order to allow suture to be passed through the allograft. The bone button was used to repair 50 consecutive rotator cuff tears ranging from small to massive and was followed by serial radiographs and clinical exams. The bone button was inserted through a 3.4 mm diameter hole to a depth of 20 mm where it was toggled and locked in the bone. Mechanical bench testing was performed to evaluate the pull out strength of these cortical bone buttons. Results: Surgical results showed that the bone button was very effective in locking in cancellous bone and did not require cortical fixation. Sequential radiographs showed the bone button clearly locked in the cancellous bone and did not rely on a cortical rim for fixation. The allograft buttons incorporated into the host bone and no failures were identified. Results of mechanical testing showed that the pull out strength of the cortical bone buttons averaged 36.0 pounds with a standard deviation of 2.0 pounds. Discussion and Conclusion: Suture anchors appear to afford enhanced purchasing fixation strength over tunnels. However, there are issues with both metal and biodegradable materials. A more physiologic approach is to use human allograft material to reattach soft tissue to bone. This simple 3 mm by 6 mm cylindrical allograft cortical button appears to afford excellent pull out strength and reproducible locking in cancellous bone. Clinical results show an allograft cortical bone button is viable for repair of the rotator cuff tendon.

Stephen Fealy, Scott A. Rodeo, John D. MacGillivray, Russell F. Warren New York, New York, USA

Introduction: Studies evaluating the results of rotator cuff repairs have reported surprisingly high rates of failure of the repairs at 3-5 yrs post-op. It is clear that the attachment site between tendon and bone is the “weak link” during the early healing period. Objective evaluation of rotator cuff repairs has demonstrated failure rates as high as 30% at 3-5 yrs post-op. The effect of contact area between tendon and bone on the ultimate pullout strength of the repaired tendon is not known. The purpose of this study is to test the hypothesis that the biomechanical strength of a healed bone-tendon interface is directly related to the cross-sectional area of tendon in contact with bone at the time of the repair. To date, no study has evaluated this question. We employed an established animal model to test our hypothesis. Materials and Methods: Twenty female goats (capra hircus) underwent bilateral open rotator cuff repairs of the superficial head of the infraspinatus tendon. The superficial head of the tendon was sharply taken off of its insertion on the humeral head at the time of surgery. The tendon edge was then repaired to bleeding cancellous bone in each case using conventional suture anchors. The tendon was repaired using two anchors (Contact Area A; N = 20) on one shoulder and four anchors (Contact Area B; N = 20) on the contralateral shoulder. For Contact Area A repairs, each anchor was placed 1 cm apart at the lateral margin of the free tendon edge; the tendon was repaired to bone using a modified Mason-Allen suture technique. Contact Area B repairs received the same two lateral tendon sutures as in A, but also had two anchors placed 1 cm medially to the lateral anchors. Contact Area B repairs had four points of tendon-to-bone fixation 1 cm apart forming a 1 × 1 cm² repair site at the bone-tendon interface. Contact Area A repairs had two points of fixation 1 cm apart. All goats were allowed to ambulate ad-lib post-op. Ten goats were harvested at 4 weeks (Group 1) and 10 goats were harvested at 8 weeks (Group 2) post-op. Specimens were tested in uniaxial tension using an Instron MTS Machine. The deep head of the infraspinatus muscle tendon was cut prior to mechanical testing in all specimens. Ultimate load to failure was reported for each shoulder. All specimens were mounted on the MTS platform with a special grip to allow the infraspinatus tendon to be pulled in line with the shoulder. Data was analyzed using a paired t-test and Wilcoxon Signed Rank test. Results: Group 1 specimens had a robust, fibrous reparative scar formation at the repair site. Group 2 specimens had a more organized fibrous scar at the repair site with less inflammatory tissues noted. Failures of all specimens occurred at the junction of the bone-tendon repair site. 13/20 (65%) specimens in Group 2 were found to have bony fragments from the humerus in contact with the free tendon edge after being loaded to failure. The mean load to failure for all specimens in Group 1 was 350.7 N, and was 619.4 N for specimens in Group 2 (P =.0002). In Group 1, specimens with Contact Area A had a mean load-to-failure of 317.3 N, specimens with Contact Area B had a mean load-to-failure of 375.5 N (P = 0.15). In Group 2, specimens repaired with Contact Area A had a mean ultimate load-to-failure of 635.8 N whereas Contact Area B specimens had an ultimate failure strength of 688.5 (P =.45; Wilcoxon Rank). Conclusion: This is the first study, to our knowledge, to evaluate the biomechanical relationship between repair site surface area and strength of the bone-tendon interface. Increasing the surface area of tendon in contact with bone at the repair site increased the ultimate load to failure of the repaired tendon at both four and eight weeks post-operatively. Although there was not a statistically significant increase in failure strength, our data supports the hypothesis that increased contact area between tendon and bone at the repair site yields stronger fixation. This increase in failure strength may play a role in the early healing period of the bone-tendon interface, which in turn may allow earlier loading of the bone-tendon junction and a quicker return to rehabilitation. We are currently evaluating the clinical results of utilizing a two-row fixation technique in rotator cuff repairs in a clinical setting.

Keith D. Nord Jackson, Tennessee, USA

Shoulder arthroscopy has provided the ability to repair rotator cuff tears through minimal incisions. Use of the suture anchor allows firm attachment to bone and requires the tissue to be sewn down to the anchor. Passing the suture through the rotator cuff can be time consuming and challenging. This new subclavian portal is located directly below the clavicle, 1½ to 2 cm medial to the AC joint. The exact point is determined with a spinal needle and the entrance site is visualized with the arthroscope. This portal allows more direct access to the rotator cuff tear and allows devices such as the Arthrex Penetrator, the Arthrex Lasso or the Innovasive suture passer to pass at a 90 degree angle to the tear. This also makes it easier for the suture to be grasped and passed through the cuff. This portal is best used with a 2 mm skin incision and bluntly passing the instrument through the deltoid and bursa. The angle of instrument insertion is critical. The instrument should be resting on the chest aimed medial to lateral for optimal alignment. This portal is best utilized following a subacromial decompression, which removes subacromial spurs that could otherwise interfere with proper alignment. Additionally, the acromial branch of the thoracoacromial artery is routinely cauterized at that time. Twenty cadaveric shoulders were dissected and the neurovascular anatomy adjacent to this portal was inspected. The Neviaser portal has been previously described, but it has been used to visualize the glenohumeral joint and avoids the tendinous portion of the rotator cuff. By changing the angle of entry, the bursa is entered safely and the rotator cuff tendon can be penetrated for passing suture. A cannula is not used. This modified Neviaser portal is recommended only in the lateral decubitus position because the angle of instrument insertion requires a medial position with your hand resting behind the patient's neck. A spinal needle can be used to localize the proper entry site and it is visualized with the arthroscope. Under direct vision, tears of the infraspinatous and supraspinatous can be approached at a 90-degree angle to the tear. The Penetrator passes through the rotator cuff and easily passes the suture through the rotator cuff for tying. The anatomy clearly demonstrates that these portals can be used safely and they are significantly far away from vital structures such as the brachial plexus, cephalic vein, musculocutaneous nerve and the subclavian artery and vein. This study demonstrates that these are safe portals for routine use in arthroscopic rotator cuff repair.

Daniel J. Mullen, Vipool K. Goradia, Henry R. Boucher, Brent G. Parks, John B. O'Donnell Baltimore, Maryland, USA

Introduction: Few investigations have studied cyclically loaded rotator cuff tears repaired with transosseous sutures and suture-based metal anchors. To our knowledge, there is no published study evaluating the effect of cyclic loading on bioabsorbable tacks used to repair rotator cuff tears. The purpose of this study was to determine the number of cycles to failure for bioabsorbable tacks used for such repair and compare those values with those for two other repair modalities. Methods: Thirty-two fresh-frozen cadaveric shoulders were dissected of all soft tissue except for that attaching the rotator cuff to the proximal humerus. A standardized, full-thickness defect was then created in the supraspinatus tendon, and the shoulders were randomized into one of four repair groups: transosseous sutures; Super Mitek suture anchors (Mitek Surgical Products, Westwood, MA); smooth bioabsorbable Suretacs (Acufex, Smith & Nephew Endoscopy, Mansfield, MA); or spiked bioabsorbable Suretacs. All repairs were loaded with 5-sec cycles of 10 to 180 N. The number of cycles required to produce gap failures of 5 and then 10 mm were recorded for each specimen. Results: There was no significant difference (P >.05) between groups with respect to the number of cycles required to produce gaps of 5 mm. The number of cycles required to produce a 10-mm gap was significantly higher (P ≤.05) for cuffs repaired with smooth bioabsorbable tacks than those repaired with the other three modalities. Discussion and Conclusion: Smooth bioabsorbable Suretacs may provide more stable fixation for rotator cuff tears in the early postoperative period than transosseous sutures, Mitek suture anchors, or spiked bioabsorbable Suretacs. These bioabsorbable tacks should be further evaluated in an animal model to more closely examine degradation characteristics and sustained strength of repair.

Kevin J. Kessler Fort Lauderdale, Florida, USA

Purpose: To present our preliminary studies using LactoSorb plates for rotator cuff repair. The use and application of this plate is described as an adjunct to prevent hole migration in trough-tunnel repairs of rotator cuff tears. Methods: Thirty consecutive patients with MRI documented rotator cuff tears who had pain and dysfunction underwent arthroscopic subacromial decompression followed by a mini-open deltoid repair of the rotator cuff. All patients were repaired through a trough and tunnel technique with the use of a standard or expandable LactoSorb plate. There were 30 patients, 20 male and 10 female. The average age was 70 years old. All patients were done under general anesthesia, however 20 patients received scalene blocks for postoperative anesthesia. Ninety-three percent of the patients (28 out of 30) were done on an outpatient basis. All patients went home with either a sling or abduction pillow. The patients were followed and evaluated at five days, one month, three months, six months, and one year postoperatively. All surgeries were done by the same surgeon and all patients followed a similar rehabilitation protocol. Results: Follow-up was obtained in 29 of the 30 patients. There were 90% excellent results and 10% good results. There were no fair or poor results. Specific attention to the operative site noted no unusual swelling, tenderness, or erythema. Radiographs did not demonstrate any bone erosion. No repeat surgical procedures were necessary and no complications occurred perioperatively or postoperatively. Conclusion: The early results of our study would support the use of LactoSorb bioabsorbable plates in the repair of rotator cuff tears, thus preventing hole migration and potentially poor results or failures of rotator cuff repairs in osteopenic/elderly bone. The copolymer LactoSorb appears to maintain its strength during the healing period of a rotator cuff repair and at the same time does not elicit any clinically noticeable inflammatory process.

Daniel Kharrazi, Ronald Glousman, James Tibone, Ronald Kvitne Los Angeles, California, USA

To delineate the incidence of re-operation on the acromioclavicular (AC) joint following arthroscopic subacromial decompression, and to characterize factors associated with failure of the initial arthroscopic subacromial decompression, we retrospectively reviewed 22 patients who required re-operation on the AC joint following arthroscopic subacromial decompression. Seventeen male and five female patients were an average age of 46.4 years old. These 22 patients were identified from a population of 1482 cases of arthroscopic subacromial decompression, both isolated and combined with distal clavicle resection, performed at our institution between 1991 and 1998. These 1482 cases were divided into two groups: isolated arthroscopic subacromial decompression (group A: 1091 cases) and arthroscopic subacromial decompression combined with arthroscopic distal clavicle resection (group B: 391 cases). The overall incidence of re-operation on the AC joint in 1482 cases was 1.5% (22 cases) at an average of 17 months after the index operation. When distinguished by group, the incidence of re-operation was equal at 1.5% (group A: 16 of 1091 cases; group B: 6 of 391 cases). However, re-operation for group A was at an average of 20.5 months and for group B at an average of 8 months after the index subacromial decompression. Seventeen (77%) of the 22 patients who required re-operation for AC joint symptoms were either worker's compensation (WC) cases (15) or patients involved in litigation (2). When distinguished by WC status, the incidence of re-operation on the AC joint was 2.4% for WC versus 0.8% for non-WC. Furthermore, when distinguished by WC status within each group the incidence of re-operation for WC versus non-WC, respectively, was 2.28% versus 0.91% for group A and 2.7% versus 0.5% for group B. The results of this study show that the incidence of re-operation on the AC joint is small (1.5%) for both isolated arthroscopic subacromial decompressions and decompressions combined with distal clavicle resection. The incidence is even smaller (0.8%) if WC cases are excluded from analysis.

Mark W. Rodosky, James A. Fenwick, Albert Vangura, Jr., Richard E. Debski Pittsburgh, Pennsylvania, USA

Introduction: Acromioplasty and distal clavicle resection (DCR) may alter AC joint function, thereby requiring the coracoclavicular (CC) ligaments to compensate for the resulting loss of joint stability [1]. There is little quantitative information regarding AC joint biomechanics following these procedures. Therefore, the objective of this study was to determine the effect of acromioplasty and DCR on joint kinematics and the in situ forces of the CC ligaments in response to anterior, posterior and superior loading. Materials and Methods: Ten fresh-frozen human cadaveric shoulders (mean age = 60 ± 8 years) were disarticulated at the glenohumeral joint and dissected free of all soft tissue except the AC, CC and coracoacromial ligaments. The specimen was mounted on the robotic/UFS testing system [2]. Translations were described as motion of the clavicle with respect to the scapula. Ten cycles of a 70 N load was applied to the joint in the anterior, posterior, and superior directions. During the last cycle for each load direction, the resultant kinematics and forces were recorded. Acromioplasty was then performed from a lateral approach with an oval burr. A 70 N load was again applied in the same three directions while the motion of the joint after acromioplasty was recorded. DCR was then performed from an anteromedial approach using the oval burr. The anterior and inferior portion of the AC ligaments were completely resected, while the superior and posterior portions were left intact and a 7 mm resection was performed using standard athroscopic technique. Once again, the motion of the joint after DCR was determined under the same loading conditions. The recorded kinematic path for the DCR condition was then repeated after full resection of the remaining AC capsule and then again after random-ordered, sequential sectioning of the conoid and trapezoid ligaments. Using the principle of superposition, the vector differences in forces before and after cutting of an individual ligament represent the in situ force in that structure. The in situ forces in the ligaments for the intact shoulder condition from a previous study in our laboratory [2] were used for comparison to the DCR condition. Results: Translations in response to the posterior load of 70 N changed significantly (P <.05) following DCR, increasing by almost 30% when compared to the intact and acromioplasty only conditions. No differences (P >.05) could be demonstrated for the resulting translations in the direction of loading for anterior and superior loads following either surgical procedure. Following DCR, significant changes were found for coupled motions in response to the anterior and posterior loads. Medial translation increased by up to 2.4 mm in response to an anterior load while superior translations increased by 3.9 mm. In response to a posterior load, the translation in the medial direction increased by up to 4.6 mm while the maximum increase in the superior direction was 4.0 mm. Following DCR, anterior loading significantly increased the in situ force in the trapezoid ligament (P <.05) from 13 ± 15 N to 40 ± 25 N, while the force in the conoid also increased significantly from 13 ± 13 N to 38 ± 28 N. No significant differences between shoulder conditions could be demonstrated (P >.05) between the trapezoid, conoid and superior AC ligaments in response to both posterior and superior loads. However, during superior loading, the force in the trapezoid ligament (51 ± 34 N) was the greatest, while the force in the superior AC ligament (33 ± 18 N) was greatest during posterior loading. Discussion: The results of this study confirm our hypothesis that DCR significantly affects the motion of the AC joint in response to anterior and posterior loading. DCR also causes the CC ligaments to carry much higher loads under some loading conditions. Importantly, acromioplasty alone did not significantly affect AC joint motion or the in situ forces in the CC ligaments. Our data also supports previous observations [1-4] that the CC ligaments restrict superior translation of the clavicle, while the AC joint capsule and articular surface is more important during anterior and posterior loading. References: 1. Flatow EL: Instr Course Lect, 42:237-245, 1993. 2. Debski RE, et al.: TORS, 24-2:378, 1999. 3. Fukuda K, et al.: JBJS, 68A:434-439, 1986. 4. Urist MR: JBJS, 28:813-837, 1946.

F. Alan Barber Plano, Texas, USA

Introduction: Arthroscopic subacromial decompression (ASD) as commonly performed, is frequently associated with a partial distal clavicle resection. This clavicle resection (co-planing) can range from simply removing a small inferior osteophyte, to a more aggressive acromioclavicular joint resection involving articular cartilage, to a complete distal clavicle resection. Recent reports provide conflicting opinions on whether a partial clavicle resection will result in pain and require a late distal clavicle resection. The purpose of this study was to evaluate the effect of varying amounts of clavicle resection done in conjunction with ASD. Methods: All non workers compensation patients undergoing “isolated” ASD with a minimum 2 years follow up from 1995 through 1997 were evaluated. A standard three portal Ellman technique was used. All patients had radiographs, charts, and arthroscopic videotapes reviewed and the degree of claviclectomy determined. Follow up evaluations included Constant, ASES, SANE, and Rowe shoulder scores. Particular attention was given to AC joint pain and the need for additional operations. Three groups were identified: Group 1 included those whose inferior osteophyte was removed without violating the articular cartilage. Group 2 had partial clavicle resections that violated the articular cartilage and co-planed the clavicle up to the level of the resected acromion. The articular cartilage violation in these cases varied, but was as much as 50% of the joint surface. Group 3 had complete distal clavicle resection. Results: Of a total of 92 patients identified, 29 were in Group 1, 37 in Group 2, and 26 in Group 3. The average age was 45 years (range 19 to 77). 69 were males and 23 females. There were 49 right and 43 left shoulders. At follow up, the average Constant, ASES, Rowe, and SANE scores for the various groups were Group 1: 99.1, 98.2 98.1, and 97.3; Group 2: 97.3, 98.8, 98.5, and 96.7; and Group 3: 98.4, 99.7, 99.5, and 98.7. No patent required any additional AC joint surgery. One patient in Groups 1 and 2 each had mild AC joint (top of shoulder) complaints at follow up. Conclusion: Violation of the AC joint capsule and partial distal clavicle resection to make it confluent with the resected acromion with ASD (co-planing) does not result in increased AC joint symptoms with a minimum 2 year follow up. All patients regardless of the amount of distal clavicle resection scored good or excellent by Constant, ASES, SANE, and Rowe shoulder scores. Late complete distal clavicle resection was not needed. Co-planing that removes the distal clavicle to the level of the resected acromion does not compromise ASD results.

George M. McCluskey, III, David T. Dellaero, Champ L. Baker Jr Columbus, Georgia, USA

We evaluated 115 patients with failed arthroscopic subacromial decompression. Of these 115 patients, 71 improved with nonoperative treatment or were referred to other subspecialists. Of the 44 remaining, 33 (75%) were manual laborers and 28 (64%) had received workers' compensation. At an average of 15 months after the initial surgery, patients underwent revision surgery for previously untreated diagnoses and complications related to their arthroscopic procedure. Ninety-six causes for failure were found in 44 shoulders: 24 impingements from inadequate decompression, 57 diagnostic errors, and 16 complications. Diagnoses included rotator cuff tear (22), acromioclavicular joint arthritis (13), glenohumeral instability (20), glenohumeral osteoarthritis (1), and humeral head osteonecrosis (1). Complications included deltoid pull-off (2), acromion fracture (1), suprascapular nerve compression (1), painful neuroma (1), and adhesive capsulitis or adhesions (11). Two groups were identified: Group I, recurrent instability; Group II, rotator cuff tears and acromioclavicular arthritis. Tests assessed preoperatively and at followup (average, 43 months) included pain (10-point subjective rating scale), American Shoulder and Elbow Surgeons (ASES) activities of daily living (ADL) (30-point scale), and ASES shoulder score (100-point scale). Forty patients were available for followup.

Group II scored significantly higher than Group I on the pain scale but scored lower on the ADL and shoulder scales. Patients with multiple diagnoses or complications or with multiple surgeries had significantly worse preoperative pain and ADL scores. Their postoperative scores showed no significant difference. Workers' compensation patients scored significantly higher in pain, showed significantly less improvement in pain after revision, and scored lower on the ADL and shoulder scales. Sixty-eight percent of nonworkers' compensation and only 44% of workers' compensation patients were able to return to their regular job. However, of the 40 patients, only 6 remained out of work, and 37 felt that they had improved and were satisfied with their result.

John W. Jaureguito, Jennifer L. Bowman, Joe Wilcox Fremont, California, USA

Introduction: Arthroscopic subacromial decompression is considered the surgical treatment of choice for impingement syndrome. However, it is controversial whether patients receiving Workers' Compensation benefits respond as well to operative treatment for impingement syndrome as those who are not receiving such benefits. The picture can be further clouded by results combining multiple surgical procedures. To our knowledge, no author has studied the effects of Workers' Compensation on the results of patients undergoing isolated arthroscopic subacromial decompression for impingement syndrome with an intact rotator cuff. Therefore, the purpose of this study is to report the outcome of arthroscopic subacromial decompression in a highly selected group of patients with impingement syndrome using reliable and reproducible shoulder evaluation scales. Methods: Thirty consecutive patients with Workers' Compensation benefits who underwent isolated arthroscopic subacromial decompression were reviewed at a minimum of two years follow-up. The diagnosis of impingement syndrome included the following: 1) consistent symptomatology and physical exam findings, 2) positive injection test 3) failure of 5 months of non-operative management 4) no previous shoulder surgery 5) type II-III acromial shape. A single surgeon (JWJ) performed all procedures. Patients undergoing additional procedures (not including rotator cuff or labral debridement) were excluded, as were patients with calcific tendonitis. An independent examiner (JB) using the SF-36, UCLA scale, ASES scale, and a satisfaction scale evaluated patients. Results: Twenty-four patients (80%) were available for follow-up at an average of 35 months. Average age was 40.4 years and 48% involved the dominant shoulder. Eight (33%) underwent debridement of partial thickness rotator cuff tears and 8 (33%) underwent debridement of labral tears. All radiographs demonstrated surgical conversion of Type II-III acromions to Type I. The average UCLA score was 29 with 71% satisfactory results (3 excellent, 14 good, 7 poor). The average ASES score was 74 with 75% satisfactory results (5 excellent, 13 good, 6 poor). 88% of patients returned to their usual and customary occupation. Litigation had no correlation with results. There was no direct correlation between presence of partial thickness rotator cuff tear or labral tear and surgical result. On a scale of 1-5 (5 = excellent), the average satisfaction score was 4.6 with all patients satisfied with their surgical result. Discussion: Patients with Workers' Compensation benefits undergoing arthroscopic subacromial decompression can have acceptable surgical results. The vast majority will return to their usual and customary occupation and are satisfied with their surgical results. The results, however, are not as good as would be expected in a non-Workers' Compensation patient population. The presence or absence of partial thickness rotator cuff tears surprisingly had no correlation with surgical results. In conclusion, acceptable surgical results can be obtained when performing arthroscopic subacromial decompression on patients with Workers' Compensation benefits. A high percentage of patients will return to their usual and customary occupation and excellent patient satisfaction can be expected.

Joseph F. Slade, III, John D. Mahoney New Haven, Connecticut, USA

Introduction: Whipple pioneered athroscopic assisted scaphoid fracture fixation. The volar technique was demanding and required extensive exposure, risking iatrogenic injury to the SBR nerve and commonly resulted in painful hypertrophic scarring. This approach provided satisfactory fixation for waist fractures but less rigid fixation for proximal pole fractures and nonunion. The Open dorsal approach is the standard now for repair of proximal pole scaphoid fractures, but this dissection risks further injury to an already tenuous blood supply. Purpose: Our purpose is describe the methodology and results of a new operative technique for Arthroscopic assisted scaphoid fracture fixation using a dorsal percutaneous approach with cannulated bone screws. Methods: Retrospective review of fourteen consecutive scaphoid fractures repaired percutaneously through a dorsal approach. Office records, operative notes and therapy records are reviewed. Technique requires small joint arthroscopy and minifluroscopy. Fracture reduction is achieved using arthroscopy and percutaneously placed K-wires into the proximal and distal poles of the scaphoid as joysticks. After fracture reduction, the wrist is flexed 90 degrees and the scaphoid is viewed as a cylinder down its long axis. A guide wire is driven dorsal to volar, exiting the base of thumb. Arthroscopy & mini fluoroscopy are used to confirm the position of the guide wire and the fracture alignment. The wrist is again flexed and a Kirschner-wire is now driven palmar to dorsal. Dorsally, a cannulated hand tap is used to prepare the scaphoid and place the cannulated screw. Results: We report on thirteen consecutive scaphoid fractures with an average follow-up of one year. Six proximal pole scaphoid fractures, three combined scaphoid and wrist fractures, and three with delayed union. Scaphoid fractures are repaired with minimum operating time. All fractures healed without complications. When compared to uninjured wrist, dorsal flexion was 91%, palmer flexion was 92% and grip strength was 85%. Illustrative cases are presented. Discussion: We propose that dorsal Percutaneous fixation of scaphoid fractures permits rapid fracture repair with minimum soft tissue dissection and the possible advantages of preserved scaphoid blood circulation. The successes of these techniques are enhanced when arthroscopy and minifluoroscopy are used in tandem to achieve fracture repair.

Daniel E. Matthews Fairhope, Alabama, USA

Twenty patients were evaluated in a prospective manner to investigate the role of small joint arthroscopy in treating intraarticular fractures of the distal radius. Fractures were classified using the AO/ASIF classification system. Ten were classified as type C1, six were type C2, two were type B2, and two were type B3. All patients had intraarticular step-off greater than 2 mm, and a fracture pattern consistent with instability and high probability of further displacement or collapse in cast immobilization. The 2.7 mm arthroscope was utilized in all patients to assist with intra-articular reduction and to diagnose and treat radiocarpal and intercarpal ligament tears. Follow-up range was fourteen to eighteen (14-18) months with a mean follow-up of twenty-six (26) months. Patients were evaluated objectively by radiographs after complete fracture consolidation, by measuring for articular congruency, volar tilt, radial height, and radial inclination. Using the Mayo Modified Wrist Score, data was collected for subjective analysis. This 100 point questionnaire scores the patients' subjective rating for pain, and functional status and their measurable range of motion and grip strength. Sixty-five percent (13/20) had associated soft tissue injury consisting of triangular fibrocartilage (TFCC), scapholunate and lunotriquetral ligament tears. Treatment consisted of debridement or repair as indicated by the particular lesion. Three (15%) patients had acute fixation of intracarpal ligament tears with intrafocal pinning. Seven (35%) patients had tears of the TFCC, five of which were classified as Palmer type IB and underwent arthroscopic assisted repair. All patients underwent transcutaneous pinning for fracture fixation and four (20%) required additional volar plating. At follow-up, no patients had greater than 1 mm step-off on x-ray evaluation. Modified Mayo Scores ranged from seventy-five to one hundred (75-100) with a mean of 89.5. There were seven (7) excellent, seven (7) good and six (6) fair. There were no poor results with the Modified Mayo Wrist Score. Intraarticular fractures of the distal radius often provide a significant challenge for the orthopaedic surgeon. The arthroscope provides the surgeon with the ability to better visualize the articular surface, thus aiding in an anatomic reduction. It also provides the surgeon with the opportunity to diagnose and treat associated soft tissue injuries which, if left untreated, may lead to a less than optimal result after apparent anatomic healing of the osseous injury. The arthroscope is an augmentation to standard fixation principles and should not substitute the possible need for plate fixation in fractures with highly unstable fracture patterns. Arthroscopic assisted reduction and internal fixation (ARIF) provides distinct advantages over other treatment algorithms and should be considered when treating intraarticular distal radius fractures as represented in this series.

Michael R. Redler, Stephen P. Fries, Beth Boros Trumball, Connecticut, USA

In recent years the appreciation of the role of triangular fibrocartilage complex tears in ulnar sided wrist pain has significantly increased (TFCC-triangular fibrocartilage complex). Severe twisting and loading injuries that are of the wrist are commonly responsible for tears of the triangular fibrocartilage complex. These patients commonly will present with ulnar sided wrist pain, pains with extremes of supination and pronation as well as pain with repetitive activity. Treatment options have included casting, physical therapy, arthroscopic or open debridement of the triangular fibrocartilage complex tear. In recent years it has become more clear that peripheral tears of the TFCC have a reasonable blood supply and therefore are candidates for repair. Many authors have described open repair of TFCC tears with good results.

A technique has been developed for arthroscopic repair of peripheral TFCC lesions. Wrist arthroscopy for these patients is performed in the usual fashion using a wrist arthroscopy tower and a 2.7 mm wrist arthroscope. Once a peripheral tear of the TFCC is identified, debridement of any associated synovitis is done. The edge of the TFCC is then freshened in order to produce a healthy healing edge. Arthroscopic repair is then done using spinal needles, shuttle relay, and PDS suture, that is in placed in a mattress type fashion. PDS suture is able to be tied down through the puncture wound to the capsule in a secure type fashion. Postoperatively, patients are casted for approximately 6 weeks. Results have been excellent in eliminating ulnar sided wrist pain, allowing patients to return to full physical activity including multiple types of athletic events.

A series of eleven patients have undergone arthroscopic repair of TFCC tears. Follow-up of these patients has revealed that they have been able to return to their sports activity and have had an average Mayo Modified Wrist score of 95 with a range of 80-100.

Presentation will include: review of the literature, review of the surgical technique, and evaluation of patient population as well as an illustrative case.

M. Ather Mirza, Eugene T. King, Mary Kate Reinhart Smithtown, New York, USA

Introduction: This study reports the eight year follow-up study on scope assisted carpal tunnel release using the distal single incision technique. Method: This technique of CTR has been performed on 1029 cases since 1991. A 1.5 cm longitudinal distal single incision in the palm allows for direct visualization of the distal edge of the transverse carpal ligament, median nerve, abnormal structures (aberrant motor branch variants and communication branch of the ulnar nerve) and the superficial palmar arch. A specially designed knife/sleeve unit mounted on a standard 4 mm endoscope allows for division of the transverse carpal ligament with a distal-to-proximal pass from the palm into the distal forearm. The interthenar fascia is usually preserved. Results: Postoperative pinch and grip strengths were near or greater than the preoperative level by eight weeks post surgery. Patients whose interthenar fascia were preserved had lateral pinch strengths significantly higher postoperatively when compared to patients who did not have an intact interthenar fascia. Median nerve sensibility was measured pre and postoperatively using two point discrimination. Mean overall return to work was 22 days. 24% of patients required no postoperative analgesics, with minimal scar, ulnar pillar, and radial pillar tenderness. There were no permanent neurovascular injuries to the median nerve or superficial palmar arch. The learning curve was found to be reasonable. Conclusion: This technique allows for a small cosmetically appealing scar, direct identification of the key anatomy, and endoscopic visualization of the median nerve, flexor tendons and transverse carpal ligament. Patients have less loss of strength and function in the early postoperative periods and an overall early return to work and full activities. This technique has shown good long and short term results.

James C.Y. Chow Mt. Vernon, Illinois, USA

Introduction: Since the introduction of the endoscopic carpal ligament release for carpal tunnel syndrome over 10 years ago, there has been controversy; including the questions of adequate release of the carpal ligament, complications, risk vs benefits, and possible higher chance of recurrence in the long term. This paper is an attempt to determine the effectiveness of endoscopic release of the carpal ligament, using the Chow technique; specifically in regards to these questions and to analyze the anatomical anomalies observed in >2000 cases. Materials and Methods: We contacted the 1530 patients (2020 consecutive cases) who underwent endoscopic release of the carpal ligament using the Chow technique from September 1987 through December 1997, by telephone, mail, or repeat clinical examination in order to answer a questionnaire. The clinical data and questionnaire responses were analyzed in an attempt to determine the successfulness of this procedure. Results: Clinical results revealed 1269 female and 751 male patients, age 14-96 years (mean—51 years), with 617 right hand, 271 left hand, and 566 bilateral cases. Symptom duration ranged from 1 week to 60 years (mean 37 months) and NCV tests were positive for 96% of the cases. 267 cases (190 patients) were lost to follow up. Post-op results were based on the remaining 1753 cases (1340 patients) with follow up ranging from 16 months—11.5 years (mean −5 years, 10 months). 9 cases (8 patients) had recurrent symptoms for a recurrent rate of 0.51% (9/1753). 17 cases continued to have problems after the procedure and additional surgery was needed for a failure rate of 0.97%. Follow up revealed that of the 1753 cases, 591 (34%) returned to normal activity in 1 week; 1065 (61%) in 2 weeks; 1313 (75%) in 3 weeks; 1490 (85%) in 4 weeks. Overall, 1736 (99%) returned to normal activities. There were no permanent nerve injury and no vessel or tendon lacerations in this series. There was 1 incomplete release, 1 case of superficial infection and 2 cases of transient ulnar nerve palsy (recovered spontaneously) for a complication rate of 0.20% (4/2020). Post-op pinch/grip results revealed that 79% of the patient had regained preoperative grip strength within 4 weeks. 10 of >2300 cases were seen to have an extremely ulnar transligamental motor branch of the median nerve (±3 mm radial to the hook of Hamate), for an incident rate of approximately 1:230. Conclusions: The results indicate that endoscopic carpal ligament release using the Chow technique is a safe and effective procedure for treating carpal tunnel syndrome with a 99% success rate, a 0.51% recurrent rate, and a 0.20% complication rate. In this series, there were no nerve injuries, no tendon lacerations, no vessel damage and, most significantly, no RSV has been reported in >2000 cases. There have been 10 of >2300 cases where an extremely ulnar transligamental motor branch of the median nerve has been seen endoscopically, resulting in an incident rate of approximately 1:230.