Editorial Commentary: Tranexamic Acid Shown to Not Inhibit Healing Following Rotator Cuff Repair in Rats.

As use of tranexamic acid (TXA) to decrease operative bleeding has increased during various orthopaedic surgical procedures, there has been corresponding increased interest regarding additional potential benefits-and also potential risks-of its use. By lessening bleeding during and shortly after arthroscopic surgery, some potential benefits include less postoperative pain, less hemarthrosis, and subsequent decreased formation of scar adhesions, resulting in less permanent stiffness. However, use of this pharmacologic agent also raises the possibility of negative effects upon tissue healing. In a rat rotator cuff repair model, no lasting significant benefit was associated with TXA administration, including no long-term decreased adhesions or stiffness. On the other hand, no adverse effects regarding healing were noted with TXA.

International Knee Documentation Committee Radiographic Knee Joint Grading More Reliable Than Kellgren-Lawrence Grading and Other Grading Systems.

Kellgren-Lawrence grading appears to have become the de facto standard for reporting radiographic degeneration of knees, yet the much later-introduced and knee-specific International Knee Documentation Committee system has repeatedly been shown to have higher reliability. Although International Knee Documentation Committee radiographic grading does have limitations-especially in cases of severe gonarthrosis-it appears to be the most reliable current system suitable for the purposes of arthroscopic knee surgeons, and it was designed to encompass all 3 knee compartments. Posterior-anterior weight-bearing radiographs taken at approximately 45° of knee flexion have repeatedly been shown to be more sensitive for revealing tibiofemoral degeneration than standard anterior-posterior weight-bearing views at or near full extension.

Editorial Commentary: Subacromial Bursa May Possess Mesenchymal Progenitor Cells.

Subacromial bursal tissue biopsied during arthroscopic surgery for full-thickness rotator cuff tears appears to possess mesenchymal progenitor cells. Although attempting to harvest mesenchymal progenitor cells from the subacromial bursa for therapeutic purposes may currently be premature, the presence of these cells in bursal tissue does call into question the routine practice of subacromial bursectomy. To maximize the chances of healing after rotator cuff repair, perhaps it would be best to avoid excising non-diseased-appearing bursal tissue any more than what is absolutely necessary to facilitate visualization.

Editorial Commentary: Treatment Efficacy Found in Animal Studies May Not Translate to Humans.

Efficacy of low-intensity pulsed ultrasonography (LIPUS) has been demonstrated in several mammalian models of injury/repair of tendons, ligaments, and soft tissue-bone junctions. But human studies have not demonstrated benefit from such intervention. In addition to innate healing differences between humans and research animals, another reason for this outcome variance may be that animal investigations of LIPUS have so far focused on healing after acute intervention, whereas randomized clinical trials have only looked at treating chronic tendinopathy in symptomatic patients. On the basis of current animal data, potential clinical benefit of LIPUS is most likely to be demonstrated for addressing acute injuries or postoperative scenarios. Yet, a particularly important anatomic difference between humans and experimental land animals regarding ultrasonography is the presence of subcutaneous adipose in the former versus the lack thereof in the latter, especially in the extremities, because overlying adipose attenuates ultrasound waves directed at underlying injured, repaired, or reconstructed tissues.

Estimating Lengths of Semitendinosus and Gracilis Tendons by Magnetic Resonance Imaging.

PURPOSE: To determine whether preoperative magnetic resonance imaging (MRI) can help predict the tendon-only length of the semitendinosus (ST) and the gracilis (G). METHODS: The distance from the tibial insertion to the distal-most aspect of the musculotendinous junction (MTJ) of the ST and G was estimated on preoperative MRI scans of patients undergoing primary anterior cruciate ligament (ACL) reconstruction with single-bundle, quadruple-stranded hamstring autograft. This MRI tendon-only length, measured by a musculoskeletal radiologist blinded to surgical findings, was compared to the actual tendon-only length measured upon harvesting each tendon. RESULTS: Among the 42 patients comprising the study population, there was very strong correlation between the estimates of tendon-only length made by MRI and surgical measurements for both the ST (Spearman coefficient = 0.83; P < .0001) and the G (Spearman coefficient = 0.82; P < .0001). The difference between MRI and surgical measurements did not exceed 3 cm for any of the 84 harvested hamstring tendons. Bland-Altman plots confirmed agreement between the 2 measurement methods. There was also strong correlation between the surgically measured tendon-only length of the ST and its G counterpart (Spearman coefficient = 0.68; P < .0001). CONCLUSIONS: MRI estimates of tendon-only length for both the ST and G very strongly correlate with operative measurements of these lengths; the discrepancy between these 2 measurement methods was found to not exceed 3 cm when the MTJ of these tendons is visible on MRI scans. LEVEL OF EVIDENCE: Level III, comparative study.

Stability of two versus three peripheral pegs of the glenoid component in modern total shoulder arthroplasty.

PURPOSE: In total shoulder arthroplasty (TSA), the optimum number of peripheral pegs required for stability in the glenoid component is unknown. This study compared the stability of two versus three peripheral pegs in cemented glenoid components possessing a central press-fit peg. METHODS: Six unmodified glenoid components with three peripheral pegs, a large, central press-fit peg and six modified glenoid components with one inferior peripheral peg sharply removed were cemented into bone substitute polyurethane blocks. A modified rocking-horse test was completed by comparing superior- and inferior-edge displacement before and after 100,000 vertical motion cycles. Then, a torsional failure test applied 2 N axial load, followed by a rotational force to the glenoid component at 0.5 °/s until failure. RESULTS: Modified rocking-horse testing showed no statistically significant edge displacement at the superior or inferior aspect of the glenoid component before or after testing. During torsional testing, peak torque and degrees of rotation at failure also showed no significant difference. CONCLUSION: Two peripheral pegs offer equivalent stability as three peripheral pegs, as assessed by cyclic rocking and rotational failure testing. Fewer peripheral pegs during glenoid component implantation may lead to less dissection, less strain on soft tissues and decreased operative time.

Relationship of Musculotendinous Junction Location to Harvested Semitendinosus and Gracilis Tendon Length.

BACKGROUND: Harvested hamstring tendon length has received scant attention in published anterior cruciate ligament (ACL) reconstruction literature, yet length can limit the ability to increase graft diameter by folding the tendon over more than once. Indeed, some ultrashort tendons may be too short to yield a clinically useful graft after being folded over just once. Ultimately, the total length of a harvested hamstring tendon may depend on the length of the tendon distal to its musculotendinous (MT) junction. PURPOSE: To compare the lengths of harvested hamstring tendons to the location of the MT junction to help predict abnormally short tendon harvest. STUDY DESIGN: Cross-sectional study; Level of evidence, 3. METHODS: Eighty-four consecutive patients undergoing primary ACL reconstruction using hamstring tendon autografts underwent intraoperative measurement of the total length of each harvested semitendinosus (ST) and gracilis (G) tendon, as well as the distance from the MT junction to that tendon's distal end (ie, the "tendon-only" length). RESULTS: The ratio of the tendon-only portion to total harvested tendon length averaged 0.52 (range, 0.39-0.71) for the ST and 0.52 (range, 0.43-0.71) for the G, suggesting a 95% chance of harvesting a tendon <15 cm in length for the tendon-only portion is <6.45 cm for ST or <6.75 cm for G tendons. There was moderate correlation between the lengths of harvested ST and G tendons with patient height as well as with the diameter of the combined, quadruple-stranded graft. CONCLUSION: The ratio of the tendon-only length to total harvested length for both the ST and G appear to range from approximately 0.4 to 0.7. Patients with abnormally distal MT junctions of either their ST or G are likely to have an abnormally short harvest of that tendon, even in the absence of technical harvesting error.

Reliability and reproducibility of several methods of arthroscopic assessment of femoral tunnel position during anterior cruciate ligament reconstruction.

PURPOSE: To assess interobserver and intraobserver agreement of estimating anterior cruciate ligament (ACL) femoral tunnel positioning arthroscopically using circular and linear (noncircular) estimation methods and to determine whether overlay template visual aids improve agreement. METHODS: Standardized intraoperative pictures of femoral tunnel pilot holes (taken with a 30° arthroscope through an anterolateral portal at 90° of knee flexion with horizontal being parallel to the tibial surface) in 27 patients undergoing single-bundle ACL reconstruction were presented to 3 fellowship-trained arthroscopists on 2 separate occasions. On both viewings, each surgeon estimated the femoral tunnel pilot hole location to the nearest half-hour mark using a whole clock face and half clock face, to the nearest 15° using a whole compass and half compass, in the top or bottom half of a linear quadrant, and in the top or bottom half of a linear trisector. Evaluations were performed first without and then with an overlay template of each estimation method. RESULTS: The average difference among reviewers was quite similar for all 4 circular methods with the use of visual aids. Without overlay template visual aids, pair-wise κ statistic values for interobserver agreement ranged from -0.14 to 0.56 for the whole clock face and from 0.16 to 0.42 for the half clock face. With overlay visual guides, interobserver agreement ranged from 0.29 to 0.63 for the whole clock face and from 0.17 to 0.66 for the half clock face. The quadrant method's interobserver agreement ranged from 0.22 to 0.60, and that of the trisection method ranged from 0.17 to 0.57. Neither linear estimation method's reliability uniformly improved with the use of overlay templates. Intraobserver agreement without overlay templates ranged from 0.17 to 0.49 for the whole clock face, 0.11 to 0.47 for the half clock face, 0.01 to 0.66 for the quadrant method, and 0.20 to 0.57 for the trisection method. Use of overlay templates did not uniformly improve intraobserver agreement for any estimation method. CONCLUSIONS: There does not appear to be any advantage of using a half clock face or compass for estimating femoral tunnel position compared with a whole clock-face analogy. Visual reference aids appear to improve interobserver agreement (reliability) of circular analogies. The linear quadrant appears to be the most reliable method (fair to moderate agreement) for estimating femoral tunnel position without a visual aid for reference, but even better reliability, ranging from fair to good agreement, may be obtained by using the whole clock-face analogy with a visual aid. CLINICAL RELEVANCE: Increasing femoral tunnel position reliability may improve outcomes of ACL reconstruction surgery.

Femoral tunnel length: accessory anteromedial portal drilling versus transtibial drilling.

PURPOSE: To determine whether drilling using an anteromedial portal technique during single-bundle anterior cruciate ligament (ACL) reconstruction risks creating femoral tunnels less than 25 mm long in the clinical setting. METHODS: Intraoperative measurements of femoral tunnel length in a group of 35 consecutive patients undergoing single-bundle primary ACL reconstruction with transtibial (TT) femoral drilling were compared with a subsequent group of 80 consecutive patients undergoing the same procedure with accessory anteromedial portal (AAMP) femoral drilling. The length of femoral tunnels created through the AAMP in male patients was compared with that in female patients, and the expected likelihood of obtaining tunnels shorter than 25 mm was determined for either gender. RESULTS: The mean femoral tunnel length in the AAMP group was significantly shorter than that in the TT group (35.6 mm and 40.7 mm, respectively; P < .0001). In male patients in the AAMP group, the femoral tunnel length was significantly greater on average than that in female patients in the same group (36.8 mm and 33.5 mm, respectively; P = .0001). The shortest measured femoral tunnel was 28 mm long. The statistical likelihood of femoral tunnels created by AAMP drilling being less than 25 mm in length was 0.47% for female patients and 0.1% for male patients. CONCLUSIONS: Although femoral tunnel length with AAMP drilling is, on average, approximately 5 mm (12.5%) shorter than with TT drilling, the likelihood of the tunnel being too short to allow for suspensory fixation with adequate graft placed within the femoral tunnel is very low. Female patients undergoing single-bundle ACL reconstruction with AAMP drilling have a femoral tunnel length that is approximately 3 mm (9%) shorter than that in male patients on average, but the expected likelihood of obtaining a tunnel shorter than 25 mm in female patients is still less than 1:200, compared with 1:1,000 for male patients. LEVEL OF EVIDENCE: Level III, retrospective comparative study.

Intra-tunnel fixation versus extra-tunnel fixation of hamstring anterior cruciate ligament reconstruction: a meta-analysis.

Hamstring anterior cruciate ligament (ACL) reconstructions fixed inside both bone tunnels have a shorter initial working length, and thus should be stiffer than those fixed outside both bone tunnels. We used meta-analysis to compare 4-stranded hamstring ACL reconstructions using the 2 fixation methods with reconstructions using patellar tendon autografts. A Medline database search of English-language articles published through June 2004 yielded 36 studies that met the inclusion criteria: 5 intratunnel fixation studies (569 patients), 10 extra-tunnel fixation studies (604 patients), and 24 patellar tendon studies (1592 patients). Three studies included both patellar tendon and hamstring reconstructions. Demographically, all 3 groups were similar. There was no significant difference in the percentage of knees restored to normal instrumented laxity measurements between the 3 groups, nor was there a difference in graft failure rate. Patient satisfaction and return to preinjury activity rates were similar between the intratunnel fixation and patellar tendon groups and were significantly lower for the extra-tunnel fixation group. Good International Knee Documentation Committee scores did not correlate with good patient satisfaction or return to preinjury activity.

Security of knots tied with ethibond, fiberwire, orthocord, or ultrabraid.

BACKGROUND: The security of several popular arthroscopic knots to prolonged, incremental, cyclic loads is unknown, as is the security of knots tied with newer, superstrong sutures. HYPOTHESIS: Some arthroscopic knots are as secure as openly tied square knots, and knots tied with superstrong sutures are more secure than those tied with braided polyester. Some arthroscopic knots are significantly bulkier than openly tied square knots. STUDY DESIGN: Controlled laboratory study. METHODS: Five types of openly tied knots (3-throw square, 4-throw square, 5-throw square, 5-throw slip, open SAK [simple arthroscopic knot]), 6 complex arthroscopic knots backed with 3 reversed half-hitches with alternating posts (RHAPs) (SMC, Weston, taut-line hitch, Tennessee slider, Roeder, Duncan loop), and 2 stacked half-hitch (SHH) arthroscopic knots (surgeon's [S=S=S//xS//xS//xS], SAK [S=S//xSxS//xS]) were tied using No. 2 Ethibond around 2 aluminum rods, which were pulled apart with stepwise, incremental, cyclic loads to a maximum force of 120 N (2250 total cycles). Then, 5-throw square knots openly tied with No. 2 Fiberwire, Orthocord, or Ultrabraid were subjected to the stepwise, incremental, cyclic loading protocol extended to a 260-N load level. Before mechanical testing, the height (maximum diameter) of each knot was measured with digital calipers. RESULTS: For Ethibond, the openly tied 3-throw square knots (56.2 +/- 21.4 N) and 5-throw slip knots (49.9 +/- 26.9 N) reached clinical failure (3 mm of laxity) at significantly lower loads (P < .05) than openly tied 5-throw square knots (90.8 +/- 6.5 N), whereas the openly tied SAK (82.3 +/- 9.4 N) and 4-throw square (84.3 +/- 11.6 N) and all arthroscopically tied knots reached 3 mm of laxity at statistically similar loads. Five-throw square knots openly tied with Fiberwire or Orthocord reached 3 mm of laxity at much higher loads (194.9 +/- 28.4 N and 168.4 +/- 8.6 N, respectively) than those tied using Ethibond (P < .001 for each comparison), but there was no significant difference in performance between Fiberwire knots and Orthocord knots. Although Ultrabraid square knots also were stronger than those tied with Ethibond (137.9 +/- 15.9 N, P < .005), they were not as secure as those tied with Orthocord or Fiberwire (P < .05). Compared with the 5-throw square knots, all arthroscopic knots were significantly bulkier. Especially bulky knots were the Duncan loop and the taut-line hitch. Orthocord square knots demonstrated bulkiness similar to Ethibond square knots, whereas Fiberwire and Ultrabraid square knots were significantly bulkier. CONCLUSIONS: For braided suture, 5-throw knots optimize square knot security. Open or arthroscopic slip knots can achieve similar security with post switching and loop reversal. Fiberwire, Orthocord, or Ultrabraid openly tied square knots offer greater security than those tied with Ethibond. Arthroscopic knots vary in their bulkiness, but all are significantly bulkier than 5-throw openly tied square knots. Square knots openly tied with Fiberwire or Ultrabraid tend to be bulkier than if tied with Ethibond or Orthocord, which are similar to each other. CLINICAL RELEVANCE: The 5-throw openly tied square knot remains the gold standard, although the openly tied SAK offers similar security when tying in a hole. Arthroscopic knots, whether complex knots backed up by 3 RHAPs, the 6-throw surgeon's knot, or the 5-throw SAK, give security similar to the standard. Square knots tied with the newer sutures in open fashion are more secure than if tied with braided polyester. Using lower profile knots may be especially important when employing Fiberwire or Ultrabraid, as these sutures tend to result in bulkier knots than those tied with Ethibond or Orthocord.

Arthroscopic findings in knees undergoing proximal tibial osteotomy.

This study determines the occurrence of significant, arthroscopically correctable intraarticular pathology at the time of valgus-producing high tibial osteotomy for symptomatic medial compartment arthrosis with varus malalignment. Thirty consecutive patients (32 knees) scheduled for the procedure underwent concomitant knee arthroscopy. In the lateral compartment, meniscal tears occurred in 16 knees (50%), unstable chondral flaps in 4 knees (13%), and loose bodies in 3 knees (9%). In the anterior compartment, unstable chondral flaps occurred in 10 knees (31%). In the medial compartment, meniscal tears occurred in 29 knees (91%). The 5 knees with mechanical symptoms did not demonstrate a higher occurrence of loose bodies, chondral flaps, or meniscal tears compared with knees without mechanical symptoms. There was a significant occurrence of correctable pathology in all three compartments in knees undergoing valgus-producing high tibial osteotomy for the treatment of symptomatic medial osteoarthritis with varus malalignment. Prior studies have not systematically documented these findings.

Classification of anterosuperior glenoid labrum variants and their association with shoulder pathology.

UNLABELLED: This prospective study of 367 consecutive shoulder arthroscopies assessed variants of the anterosuperior glenoid labrum and associated shoulder pathology. Thirty-three shoulders were excluded because of prior surgery, septic arthritis, or adhesive capsulitis. Anterosuperior glenoid variants were classified as: type I, cordlike middle glenohumeral ligament without sublabral foramen; type II, sublabral foramen without a cordlike middle glenohumeral ligament; type III, sublabral foramen with a cordlike middle glenohumeral ligament; and type IV, absent anterosuperior labrum with the anterior aspect of the superior labrum continuous with a cordlike middle glenohumeral ligament. The presence of these variants was correlated with the incidence of shoulder pathology found on arthroscopic inspection. Of 334 shoulders, 118 (35.3%) had variants of the anterosuperior glenoid labrum. Of these, 32 (27.1%) were type I, 27 (22.9%) were type II, 34 (28.8%) were type III, and 25 (21.2%) were type IV. The incidence of advanced superior labrum anterior-posterior lesions in the 86 shoulders displaying a type II, III, or IV variant was significantly higher than in shoulders with no anterosuperior variant (48.8% versus 23.6%, P<.001). Other pathologic findings were not significantly increased in shoulders with variants compared to those without. LEVEL OF EVIDENCE: Level 1.

Intraosseous innervation of the human patella: a histologic study.

BACKGROUND: The soft tissue structures surrounding the human knee joint have been the subject of extensive anatomic study. The detailed histologic findings within the bone of the human patella, however, have not been systematically studied. While the nerves supplied to the periarticular soft tissues have been very well documented, the nerves supplied to the interior of the bony patella have never been described. HYPOTHESIS: This study tests the hypothesis that the patella contains an intraosseous nerve network. Further, the authors investigate the anatomic location of these intraosseous nerves to better understand their possible clinical relevance. STUDY DESIGN: Descriptive laboratory study. METHODS: Ten matched pairs of cadaveric patellae (left and right patellae from the same individual; 20 total) were prepared for evaluation by hematoxylin and eosin staining using a technique that allows the creation of complete, large histologic sections of individual patellae. The matched specimens were dissected free of soft tissue and then sectioned using a diamond-wafering saw into 3-mm sagittal (left patella) and transverse (right patella) sections. Sections were then decalcified and whole-mounted into paraffin blocks for further sectioning using a large-format microtome. All 20 specimens were prepared for evaluation. Age at death averaged 80 years (range, 64-91). All specimens demonstrated at least grade II chondromalacia. RESULTS: Nineteen of 20 (95%) specimens demonstrated intraosseous nerves. Of 248 sections studied, 116 (47%) demonstrated intraosseous nerves, with 227 individual nerves identified. The density of intraosseous nerves was greatest in the medial and central portions of the patella, with a significant paucity identified laterally. CONCLUSION: The primary intraosseous innervation of the patella derives from a medially based neurovascular bundle. CLINICAL RELEVANCE: A better understanding of the nerves within the human bony patella may improve understanding the patho-physiology of anterior knee pain syndromes.

Deep venous thrombosis after knee arthroscopy: a meta-analysis.

PURPOSE: To better establish the incidence of deep venous thrombosis (DVT) after knee arthroscopy without prophylaxis. TYPE OF STUDY: Meta-analysis. METHODS: A MEDLINE search was performed to find published English-language studies of DVT following knee arthroscopy. Strict inclusion criteria required any investigation included for analysis to (1) be prospective, (2) include a population not given any antithrombotic prophylaxis, (3) perform universal screening of the lower extremity venous system using either ultrasound or venography, and (4) include a discrete population limited to unilateral arthroscopic knee surgery, exclusive of ligament surgery or open procedures. The results of the venographic studies were given additional weight mathematically to account for the increased sensitivity of this imaging modality compared with ultrasound in detecting DVT in asymptomatic lower extremities. RESULTS: The unprophylaxed patient populations in the 6 studies that met inclusion criteria ranged from 64 to 184 patients, and totaled 684. The total incidence rate of DVT in each study ranged from 3.1% to 17.9%, whereas the incidence rate of proximal DVT ranged from none detected to 4.9%. The total DVT rate was not related to the method of DVT detection, whereas proximal DVT was only noted in the 2 studies using contrast venography. Combining the 6 studies yields a total DVT incidence of 9.9% (95% confidence interval, 8.1%-11.7%), and a proximal DVT incidence of 2.1% (95% confidence interval, 1.2%-3.0%). CONCLUSIONS: DVT following knee arthroscopy is a consistent finding in studies of unprophylaxed patients when routine screening using ultrasound or contrast venography is used. Current data suggest an overall DVT rate of 9.9% and a proximal DVT rate of 2.1% after knee arthroscopy without antithrombosis prophylaxis. LEVEL OF EVIDENCE: Level II, Systematic Review of Levels I and II Diagnostic Studies.

Arthroscopic debridement of acute periprosthetic septic arthritis of the knee.

PURPOSE: To evaluate the efficacy of a treatment protocol including arthroscopic irrigation and debridement in resolving acute periprosthetic septic arthritis of the knee. TYPE OF STUDY: Retrospective review. METHODS: During a 1-year period, patients presenting with acute septic arthritis of the knee after total knee arthroplasty were screened for suitability for treatment with an arthroscopic debridement protocol consisting of (1) arthroscopic debridement and synovectomy with at least 12 L of antibiotic irrigant, (2) suction drainage until minimum output, (3) repeat arthroscopy for fever persisting beyond 48 hours, and (4) intravenous antibiotics for 6 weeks. Only previously well-functioning arthroplasties presenting within 7 days of symptom onset in patients without immunocompromising factors and with radiographically stable prosthetic components were selected for this treatment protocol. Five knees in 4 patients met the inclusion criteria and were examined clinically and radiographically at minimum 36-month follow-up. RESULTS: The mean follow-up interval was 41 months (range, 36 to 43 months). No knee had been revised or had revision planned for infection, although 1 had undergone tibial insert exchange for polyethylene wear. Follow-up radiographs showed no prosthetic loosening, and Knee Society scores averaged 88 (range, 75-95) in the 4 retained knees, 3 of which were subjectively thought to have returned to their preinfection level of functioning. All knees were infected with a single organism not considered highly virulent. No patients required oral suppressive antibiotics for their knees. CONCLUSIONS: Early aggressive arthroscopic debridement as part of a treatment protocol of acute periprosthetic knee septic arthritis with well-fixed and functioning implants can be an effective treatment option in selected cases. LEVEL OF EVIDENCE: Level IV study (no, or historical control group).

Glenoid suture anchor fixation strength: Effect of insertion angle.

PURPOSE: The purpose of this study was to determine the effect of varying insertion angles on the fixation strength of screw-in devices placed in the glenoid rim. Type of study Cadaveric biomechanical analysis. METHODS: Eighteen cadaveric glenoids had 3.0-mm cannulated screws inserted for a depth of 10 mm in all 4 quadrants: anterior superior (AS), anterior inferior (AI), posterior superior (PS), posterior inferior (PI). Screws were inserted along the orthogonal to the glenoid rim at the point of insertion or at angles that deviated from this vector by 20 degrees and 40 degrees. Load to failure was performed at 10 mm/s along the orthogonal to the point of insertion. RESULTS: For screws inserted orthogonal to the glenoid rim, the average load to failure was highest for the PS quadrant (733 +/- 369 N) and lowest for the AI quadrant (272 +/- 69 N). The AS and PI quadrants showed intermediate values (549 +/- 334 N and 484 +/- 141 N, respectively). Deviation from orthogonal correlated with decreased fixation strength. This decrease was statistically significant in the AS and PI quadrants, with deviation of 40 degrees, and in the AI quadrant, with deviation of 20 degrees, as well as 40 degrees. Conclusions Insertion angles for screw-in fixation devices should be orthogonal to the glenoid rim at the point of insertion to maximize strength. Deviation of 40 degrees from orthogonal compromises fixation in most quadrants and deviation as little as 20 degrees can compromise fixation in the AI quadrant. CLINICAL RELEVANCE: To maximize strength of labral reattachment to the bony glenoid, screw-in type fixation devices should be inserted as orthogonal to the glenoid rim as possible. This is especially true for Bankart repairs, because device pullout occurs at significantly lower loads in the anteroinferior quadrant compared with the other 3 quadrants. Deviating as little as 20 degrees further decreases fixation strength significantly.