- Email: [email protected]
Rotator Cuff Repair: An Ex Vivo Analysis of Suture Anchor Repair Techniques on Initial Load to Failure
Rotator Cuff Repair: An Ex Vivo Analysis of Suture Anchor Repair Techniques on Initial Load to Failure Craig A. Cummins, M.D., Richard C. Appleyard, M.D., Sabrina Strickland, M.D., Pieter-Stijn Haen, M.D., Shiyi Chen, M.D., Ph.D., and George A. C. Murrell, M.B.B.S., D.Phil.
Purpose: To determine the best combination of anchors and suture techniques for repairing torn rotator cuff tendons. Type of Study: Ex vivo biomechanical investigation. Methods: Sixty freshfrozen sheep infraspinatus tendons were repaired using 6 different repair techniques: transosseous sutures with 2 sutures and mattress stitches; 2 suture anchors with 1 suture per anchor using either simple stitches, mattress stitches, or modified Kessler stitches; 2 suture anchors with 2 sutures per anchor using simple stitches; or 5 suture anchors with 1 suture per anchor and a mattress stitch pattern. Results: No difference was identified between transosseous sutures (mean ⫾ SD, 147 ⫾ 68 N) and suture anchors (140 ⫾ 36 N) when 2 mattress stitches were used. The weakest construct with suture anchors was when the tendon was grasped with 2 suture anchors with 1 suture per anchor and a simple stitch pattern (72 ⫾ 25 N). Repair strength increased 2-fold with 2 suture anchors single loaded and a mattress stitch configuration (140 ⫾ 36 N; P ⫽ .026), 3-fold with 2 suture anchors single loaded and a modified Kessler stitch pattern (204 ⫾ 32 N; P ⬍ .001), and 3-fold with 2 suture anchors double loaded and a simple stitch suture pattern (212 ⫾ 39 N; P ⬍ .001). The highest tensile load was observed with 5 suture anchors in a double-row configuration, single loaded, that grasped the tendon with mattress stitches (336 ⫾ 59 N; P ⬍ .001). Conclusions: This study shows that in an ovine model, initial rotator cuff repair strength can be enhanced by increasing the number of suture anchors used in the repair and by using anchors that are double loaded with suture and suture configurations that pass more frequently through the tendon. Clinical Relevance: The clinical relevance of this ex vivo investigation is that the initial load to failure of a rotator cuff repair may be increased by increasing the number of suture anchors, the number of sutures per anchor, or using suture patterns that grab more adjacent tendon fibers. Key Words: Ovine—Infraspinatus—Rotator cuff—Suture anchor—Load to failure.
A
common cause of adult shoulder pain and dysfunction is rotator cuff tears. The goal of rotator cuff repair surgery is to restore the shoulder to a
From Lake Cook Orthopedic Associates (C.A.C.), Barrington, Illinois, U.S.A.; Orthopaedic Research Institute, St. George Hospital Campus, University of New South Wales (C.A.C., R.C.A., S.S., P-S.H., S.C., G.A.C.M.), Sydney, Australia; and the Hospital for Special Surgery (G.A.C.M.), New York, New York, U.S.A. Supported by St. George Hospital/South Eastern Sydney Area Health Service, Sydney, Australia. Address correspondence and reprint requests to Craig A. Cummins, M.D., Lake Cook Orthopedic Associates, 27401 W Highway 22, Suite 125, Barrington, IL 60010, U.S.A. E-mail: CraigACummins@ hotmail.com © 2005 by the Arthroscopy Association of North America 0749-8063/05/2110-4326$30.00/0 doi:10.1016/j.arthro.2005.06.022
1236
pain-free state with normal motion, strength, and function. Data suggest that shoulder function after rotator cuff repairs is related to the integrity of the repair.1 Rotator cuff tears may occur in many different patterns, but a consistent component of the tear is detachment of the tendon from its normal insertion on the proximal humerus. Therefore, although a surgical repair may vary depending on the specific characteristics of a given tear, an important component of a rotator cuff repair involves reattaching the tendon to bone. Rotator cuff tendons are typically reattached to the proximal humerus using either transosseous sutures2 or suture anchors.3,4 When suture anchors are used for the repair, the primary in vivo and ex vivo mode of failure is at the suture tendon interface.5,6 Previous investigations suggest that more complex
Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 21, No 10 (October), 2005: pp 1236-1241
SUTURE ANCHORS AND LOAD TO FAILURE suture patterns that grasp more adjacent tendon fibers, can increase the strength of the tendon-suture interface.7 These suture techniques are relatively straightforward to perform in open procedures but may be technically challenging and time consuming with arthroscopic repair techniques. To increase the tendon repair strength while avoiding the use of complex suturing techniques, potential options include using more suture anchors or suture anchors loaded with 2 sutures. To that aim, this investigation hypothesized that the initial load to failure of rotator cuff repairs could be increased by using more suture anchors, anchors double loaded with suture, and more complex suture configurations that result in more passes through a tendon with a given suture.
1237
FIGURE 1. The posterior aspect of -/⫹2a sheep’s proximal humerus. The shaded area represents the footprint of the infraspinatus tendon insertion site. The black dots depict the fixation sites for the groups that used (A) 2 suture anchors, (B) 5 suture anchors, and (C) transosseous sutures through bone tunnels. The suture pattern through the bone tunnels is also depicted.
METHODS The ultimate load to failure for different rotator cuff repair techniques was evaluated in a sheep infraspinatus tendon using a technique previously described in the literature.8,9 The sheep infraspinatus tendon was chosen because it has biomechanical, anatomic, and histologic properties similar to the human supraspinatus tendon.7 Ovine shoulder specimens between 6 and 8 months of age were obtained from a local meat distributor and used in the study within 48 hours of death. Each shoulder specimen was prepared by removing all tissue with the exception of the humerus and infraspinatus tendon. The infraspinatus tendon was sharply detached from its insertion site to mimic a complete full-thickness rotator cuff tear. The tendons were detached perpendicular to the course of their collagen fibers to maximize tendon thickness at their distal aspect and optimize the homogeneity between the shoulder specimens. Any remaining soft tissue was removed from the tendons’ insertion site and the width and length of the infraspinatus tendon insertion site measured using a digital vernier caliper (Mitutoyo American, Aurora, IL). The tendon repair involved reattaching the infraspinatus tendon using 6 different methods of fixation. All repair groups used a braided, nonabsorbable, No. 2 polyester Ethibond suture (Ethicon, Edinburgh, Scotland). After the suture was passed through the tendon, the suture was tied using a square knot with 4 throws and manually tensioned. In general, the suture engaged the infraspinatus tendon 10 mm from its distal margin using various suture repair techniques. No pretensioning was applied to the tendon during the repair. Each tendon was reattached to its original foot-
print of the infraspinatus insertion site. One repair group used bone tunnels in which 2 sutures were passed transosseous through 4 bone tunnels. The bone tunnels were spaced between 8 and 10 mm apart depending on the size of the infraspinatus footprint and extended 10 mm in length distally.10 The suture knots were tied laterally over a bone bridge and bone troughs were not used. The remaining 5 repair groups used titanium alloy suture anchors to secure the tendon to bone (Mitek Surgical Products, Norwood, MA). Four of the 5 repair groups used 2 suture anchors per shoulder with the other repair group using 5 suture anchors that were placed in a double-row configuration (Fig 1). When 2 anchors were used, the anchors were placed 10 mm apart. When 5 anchors were used, the anchor spacing was slightly less than 10 mm because of the constraints of the infraspinatus footprint. The Mitek Rotator Cuff QuickAnchor was used for all suture repair groups in which a suture anchor was loaded with a single Ethibond suture. One of the suture anchor repair groups had 2 sutures threaded through a single anchor. This repair group used the Mitek SuperAnchor because the Rotator Cuff QuickAnchor would not easily accommodate 2 No. 2 Ethibond sutures. The following 6 repair techniques were tested to failure (Figs 1 and 2): 1. Four bone tunnels, 2 sutures, mattress stitch configuration (n ⫽ 10). 2. Two suture anchors, 2 sutures (anchors single loaded), simple stitch configuration (n ⫽ 10). 3. Two suture anchors, 2 sutures (anchors single loaded), mattress stitch configuration (n ⫽ 10).
1238
C. A. CUMMINS a difference of 20% between each of the groups with regard to the specimens load to failure. For all statistical tests, statistical significance was set at an alpha error of 0.05 and a beta error of 0.2. Results are reported as mean values ⫾ standard deviation. Statistical analyses included both descriptive statistics as well as a 1-way analysis of variance with corrections made for multiple comparisons using the Tukey HOC test. RESULTS
FIGURE 2. The different suturing configurations used in the various groups. (A) Simple stitch, (B) mattress stitch, and (C) modified Kessler stitch.
4. Two suture anchors, 2 sutures (anchors single loaded), modified Kessler stitch configuration (n ⫽ 10). 5. Two suture anchors, 4 sutures (anchors double loaded), simple stitch configuration (n ⫽ 10). 6. Five suture anchors, 5 sutures (anchors single loaded), mattress stitch configuration (n ⫽ 10).
Model The infraspinatus tendon insertion site had a mean width of 17.3 ⫾ 0.1 mm and length of 12.6 ⫾ 0.1 mm. No statistically significant differences were identified between the 6 repair groups with respect to either the insertion sites’ width or length. The majority of specimen failures occurred as a result of the tendon pulling through the various fixation devices. There were no failures resulting from the tendon slipping through the
After tendon fixation, each specimen was stored at ⫺20°C. Before mechanical testing, the specimens were thawed at room temperature and kept in a moist gauze soaked in normal saline. The mechanical testing was performed using a mechanical tensile testing machine. The humerus was secured to the base plate with a screw (7-mm diameter) and washer (25-mm diameter). The infraspinatus tendon was secured with tendon grasping clamps that pulled perpendicular to the sagittal plane and parallel to the transverse plane of the tendon. The tendon repair was tested in tension with the direction of pull 90° to the shaft of the humerus, simulating the position of the patient’s arm at their side (Fig 3). This is the arm position for which the maximum amount of tension crosses the rotator cuff repair.11 The specimens were loaded at an extension rate of 40 mm/min and tested to failure with the data captured on a standard PC computer. The mode of failure was recorded for each shoulder. Statistical Methods All statistical analyses were performed using the SPSS v. 10 software (SPSS, Chicago, IL). Before initiating the study, a power analysis was performed with the beta error set at 0.2 (power ⫽ 0.8). Based on the results of the power analysis, it was determined that each repair group required 10 shoulders to detect
FIGURE 3. The mechanical tensile testing machine (Schimadzu AG-50 KNE, Shimadzu, Japan). The infraspinatus tendon has been twisted 90° along its axis in this illustration for demonstration purposes. The tendon was not twisted during the actual testing.
SUTURE ANCHORS AND LOAD TO FAILURE
1239
and the group repaired with transosseous sutures using 2 mattress stitches (147 ⫾ 68 N) were equivalent (Fig 4, Table 1). Both repair groups failed as a result of the tendon pulling through the sutures. No tendon repairs failed by suture breakages or pulling out of bone. Higher Load to Failure in Complex Suture Groups Compared With Simple Sutures In the 3 repair groups that used 2 suture anchors, each loaded with a single suture, the load to failure increased as the suturing configuration changed from a simple to a more complex pattern (Fig 4). The weakest construct was 2 suture anchors with a simple suture pattern consisting of 1 strand of suture being passed through the tendon for each anchor (72 ⫾ 25 N). There was a 2-fold improvement using mattress stitches that consisted of 2 strands of suture passed through the tendon for each anchor (140 ⫾ 36 N; P ⫽ .026), and a 3-fold improvement when using modified Kessler stitches (204 ⫾ 32 N; P ⬍ .001). These increases correlated with the number of times the suture passed through the tendon. With regard to the mode of failure, all of the tendon repairs that used 2 suture anchors with either simple or mattress stitches failed as a result of the tendon pulling through the sutures. However, the group repaired with modified Kessler stitches failed from suture breakage in 9 of the 10 tendon repairs. The other failure was caused by the tendon pulling through the sutures.
FIGURE 4. Ultimate tensile strength of the 6 methods of rotator cuff repair using suture anchor repair (mean ⫾ standard error). Abbreviations: BT, bone tunnel; Mit, Mitek suture anchor; Simp, simple stitch; Mat, mattress stitch; ModK, modified Kessler stitch. The number before the colon refers to the number of suture anchors. The number after the colon refers to the number of sutures used in the repair. Statistical analysis was performed using the Tukey HOC test and 1-way analysis of variance. The numbers in Table 1 are the respective P values with NS signifying no statistically significant difference (P ⬎ .05) between groups.
The More Sutures Per Anchor the Higher the Load to Failure
clamp or at the site between the humerus and the base plate.
Increasing the number of sutures per anchor from 1 to 2, while keeping the number of anchors and suturing pattern constant, resulted in a 3-fold increase in the ultimate load to failure (72 ⫾ 25 N v 212 ⫾ 39 N; P ⬍ .001). The repair strength of 2 double-loaded anchors with a simple stitch pattern was equivalent statistically to 2 single loaded anchors with a modified
Suture Anchors Equivalent to Transosseous Sutures The failure load between the tendons repaired with 2 suture anchors using mattress stitches (140 ⫾ 36 N) TABLE 1. BT:2Mat BT:2Mat 2Mit:2Simp 2Mit:2Mat 2Mit:2ModK 2Mit:4Simp 5Mit:Mat
Statistical Comparison of Tendon Repair Groups
2Mit:2Simp
2Mit:2Mat
2Mit:2ModK
2Mit:4Simp
5Mit:5Mat
.009
NS .027
NS ⬍.001 .046
.04 ⬍.001 .014 NS
⬍.001 ⬍.001 ⬍.001 ⬍.001 ⬍.001
Respective P values for the six repair groups demonstrated in Figure 4. NS signifies no statistically significant difference (P ⬎ .05) between groups.
1240
C. A. CUMMINS
Kessler stitch pattern. However, the mode of failure significantly differed in these 2 repair groups (P ⬍ .001). The tendon pulled through the suture in all 10 of the repairs with 2 double-loaded anchors and a simple stitch pattern. In contrast, as noted previously, the group repaired with the modified Kessler stitch failed in the majority of cases as the result of the suture breaking. The More Suture Anchors the Higher the Load to Failure The repair group that used 5 suture anchors and mattress stitches had the greatest load to failure of all repair groups tested (P ⬍ .001) (Fig 4). Increasing the number of suture anchors from 2 to 5, while keeping the suture pattern constant, resulted in a 2.4-fold increase in load to failure. Compared with the suture anchor repair group with the lowest load to failure, 2 suture anchors and simple stitches, there was a 4.7fold increased load to failure with 5 suture anchors and mattress stitches (P ⬍ .001). Regarding the mode of failure, 7 of the 10 shoulders in the group repaired with 5 suture anchors failed by the tendon pulling through the sutures. The other 3 shoulders failed by suture breakage. DISCUSSION An important goal of surgery for rotator cuff tears is to achieve tendon fixation secure enough to hold the repaired tendon in place until biological healing occurs. To better understand how to achieve this goal, the aim of this study was to quantify how different combinations of anchor and suture techniques affect the initial strength of a rotator cuff repair. In this study, the effects of a variety of rotator cuff tendon repair techniques were examined in a sheep infraspinatus tendon. The sheep infraspinatus tendon was chosen because: (1) it has previously been shown to be similar in many ways to the human supraspinatus tendon7, and (2) this model has been used in previous studies7-9,12 and, therefore, allows some interstudy comparison. Furthermore, this model was readily available and thus allowed comparison of many different repair techniques while maintaining an acceptable statistical power. Despite these advantages, this model is not a human supraspinatus tendon and therefore has some weaknesses. The sheep in this investigation were 6 to 8 months old and without rotator cuff tears. Clearly this is not the same tendon as seen in a 60- to 70-year-old patient with a chronic rotator cuff
tear.13 Further, the bone density of the sheep proximal humerus is significantly greater than that of the human proximal humerus. However, the purpose of the study was to evaluate rotator cuff tendon fixation and not designed to look at the implants’ ability to resist pulling out of bone. We elected to focus on tendon fixation as the predominant in vivo and ex vivo mode of rotator cuff repair failure with suture anchors has been shown to occur at the tendon suture interface.5,6 Furthermore, it is very uncommon for a rotator cuff repair to retear as a result of the implant pulling out of bone.5,6 As expected, no specimens failed as a result of the anchor pulling out of the sheep bone. As such, the type of anchor used in the different study groups likely had little impact on the observed loads to failure. Another potential limitation of this study is that the shoulder specimens were not tested with cyclic loading. Whether rotator cuff repairs fail as a result of a single large force seen by the repaired tendon or multiple, repetitive, subthreshold forces is not known. In a previous investigation, we noted that 14 of 22 patients who went on to a revision rotator cuff repair reported that a significant event (e.g., a fall) occurred before their rotator cuff repair retearing. Seven patients could not recall a precipitating event and in 1 patient the information was not recorded.5 It is likely that some rotator cuff repairs fail as a result of a large force, whereas others fail mechanically as a result of cyclic loading. Ideally, this investigation would have evaluated both potential causes of repair failure. Given the potential limitations of this investigation, our results clearly show that altering the tendon repair technique can result in substantial—at least 4.7-fold— change in the initial strength of a rotator cuff repair. Significant effects were seen with respect to the suture anchor number, number of sutures per anchor, and the suture configuration. When the number of anchors and number of sutures per anchor remained constant, the load to failure markedly increased as the suture configuration progressed from simple stitches to more complex suturing patterns. The load to failure in the group repaired with modified Kessler stitches was 2.8 times greater than the group repaired with simple stitches. The repair groups that used simpler suture patterns, such as simple and mattress stitches, failed by the tendon pulling through the suture, whereas the repairs in the modified Kessler group failed predominantly by suture breakage. This implies that a point was reached at which the weakest link of the repair shifted away from the suture tendon interface and became the suture material itself.
SUTURE ANCHORS AND LOAD TO FAILURE The change in the mode of failure occurred in the modified Kessler group at a load of 204 ⫾ 32 N. It is unlikely that a higher ultimate tensile load could be achieved using a different suture pattern given the same type, size, and number of sutures. Therefore, we hypothesize that, at that point, an increase in the load to failure can only be achieved by an increased size of the suture,7 improved material properties of the suture,7 or increased number of sutures. A substantial increase in the load to failure of a rotator cuff tendon repair can also be achieved by increasing the number of sutures per anchor. With all other parameters equal, increasing the number of sutures per anchor from 1 to 2 resulted in nearly a 3-fold increase in initial load to failure. In addition, the initial repair strength can be significantly increased by increasing the number of suture anchors and thus sutures used in the repair. The group repaired with 5 suture anchors had a nearly 5-fold increase in load to failure compared with the group repaired with 2 suture anchors and simple stitches. However, it is unclear if the increased initial load to failure was solely attributable to the increased number of suture anchors. It is possible that load to failure may have been affected in part by placing the anchors in 2 rows. Using a doublerow pattern of suture anchor placement has the potential advantages of distributing the forces more diffusely across the tendon repair. Furthermore, the double row likely increases the contact area between the tendon and bone and theoretically may provide for a better chance of rotator cuff healing. CONCLUSIONS This study has highlighted some important issues with respect to achieving a strong initial repair of torn rotator cuff tendons. Specifically, the study showed that enhanced repair strength can be obtained by (1) using suture configurations that pass more frequently through the tendon, (2) increasing the number of su-
1241
ture anchors, and (3) using suture anchors loaded with 2 sutures per anchor. Acknowledgment: We thank the Biomaterials Research Unit, Department of Dentistry, University of Sydney, Australian Technology Park, Sydney, Australia, for allowing us to use their Shimadzu AG-50 KNE mechanical testing machine; and Mitek Surgical Products for their generous donation of rotator cuff anchors.
REFERENCES 1. Harryman D, Mack L, Wang K. Repairs of the rotator cuff: Correlation of functional results with integrity of the cuff. J Bone Joint Surg Am 1991;73:982-989. 2. McLaughlin H. Lesions of the musculotendinous cuff of the shoulder. J Bone Joint Surg Am 1994;26:31-51. 3. Gartsman G, Khan M, Hammerman S. Arthroscopic repair of full thickness tears of the rotator cuff. J Bone Joint Surg Am 1998;80:832-840. 4. Burkhart S, Danaceau S, Pearce C. Arthroscopic rotator cuff repair: Analysis of results by tear size and by repair techniquemarginal convergence versus direct tendon-to-bone repair. Arthroscopy 2001;17:905-912. 5. Cummins C, Murrell G. Mode of failure for rotator cuff repair with suture anchors identified at revision surgery. J Shoulder Elbow Surg 2003;12:128-133. 6. Burkhart S, Pagan J, Wirth M, et al. Cyclic loading of anchorbased rotator cuff repairs: Confirmation of the tension overload phenomenon and comparison of suture anchor fixation with transosseous fixation. Arthroscopy 1997;13:720-724. 7. Gerber C, Scneeberger A, Beck M, et al. Mechanical strength of repairs of the rotator cuff. J Bone Joint Surg Br 1994;76: 371-380. 8. Cummins C, Strickland S, Appleyard R, et al. Rotator cuff repair with bioabsorbable screws: An in vivo and ex vivo investigation. Arthroscopy 2003;19:239-248. 9. Koh J, Szomor Z, Murrell G, et al. Supplementation of rotator cuff repair with a bioresorbable scaffold. Am J Sports Med 2002;30:410-413. 10. Caldwell G, Warner J, Miller M, et al. Strength of fixation with transosseous sutures in rotator cuff repair. J Bone Joint Surg Am 1997;79:1064-1068. 11. Zuckerman J, LeBlanc J-M, Choueka J, et al. The effect of arm position and capsular release on rotator cuff repair. J Bone Joint Surg Br 1991;73:402-405. 12. Gerber C, Schneeberger A, Perren S, et al. Experimental rotator cuff repair. A preliminary study. J Bone Joint Surg Am 1999;1281-1290. 13. Nixon J, DiStefano V. Ruptures of the rotator cuff. Orthop Clin North Am 1975;6:423-447.
Purpose: To determine the best combination of anchors and suture techniques for repairing torn rotator cuff tendons. Type of Study: Ex vivo biomechanical investigation. Methods: Sixty freshfrozen sheep infraspinatus tendons were repaired using 6 different repair techniques: transosseous sutures with 2 sutures and mattress stitches; 2 suture anchors with 1 suture per anchor using either simple stitches, mattress stitches, or modified Kessler stitches; 2 suture anchors with 2 sutures per anchor using simple stitches; or 5 suture anchors with 1 suture per anchor and a mattress stitch pattern. Results: No difference was identified between transosseous sutures (mean ⫾ SD, 147 ⫾ 68 N) and suture anchors (140 ⫾ 36 N) when 2 mattress stitches were used. The weakest construct with suture anchors was when the tendon was grasped with 2 suture anchors with 1 suture per anchor and a simple stitch pattern (72 ⫾ 25 N). Repair strength increased 2-fold with 2 suture anchors single loaded and a mattress stitch configuration (140 ⫾ 36 N; P ⫽ .026), 3-fold with 2 suture anchors single loaded and a modified Kessler stitch pattern (204 ⫾ 32 N; P ⬍ .001), and 3-fold with 2 suture anchors double loaded and a simple stitch suture pattern (212 ⫾ 39 N; P ⬍ .001). The highest tensile load was observed with 5 suture anchors in a double-row configuration, single loaded, that grasped the tendon with mattress stitches (336 ⫾ 59 N; P ⬍ .001). Conclusions: This study shows that in an ovine model, initial rotator cuff repair strength can be enhanced by increasing the number of suture anchors used in the repair and by using anchors that are double loaded with suture and suture configurations that pass more frequently through the tendon. Clinical Relevance: The clinical relevance of this ex vivo investigation is that the initial load to failure of a rotator cuff repair may be increased by increasing the number of suture anchors, the number of sutures per anchor, or using suture patterns that grab more adjacent tendon fibers. Key Words: Ovine—Infraspinatus—Rotator cuff—Suture anchor—Load to failure.
A
common cause of adult shoulder pain and dysfunction is rotator cuff tears. The goal of rotator cuff repair surgery is to restore the shoulder to a
From Lake Cook Orthopedic Associates (C.A.C.), Barrington, Illinois, U.S.A.; Orthopaedic Research Institute, St. George Hospital Campus, University of New South Wales (C.A.C., R.C.A., S.S., P-S.H., S.C., G.A.C.M.), Sydney, Australia; and the Hospital for Special Surgery (G.A.C.M.), New York, New York, U.S.A. Supported by St. George Hospital/South Eastern Sydney Area Health Service, Sydney, Australia. Address correspondence and reprint requests to Craig A. Cummins, M.D., Lake Cook Orthopedic Associates, 27401 W Highway 22, Suite 125, Barrington, IL 60010, U.S.A. E-mail: CraigACummins@ hotmail.com © 2005 by the Arthroscopy Association of North America 0749-8063/05/2110-4326$30.00/0 doi:10.1016/j.arthro.2005.06.022
1236
pain-free state with normal motion, strength, and function. Data suggest that shoulder function after rotator cuff repairs is related to the integrity of the repair.1 Rotator cuff tears may occur in many different patterns, but a consistent component of the tear is detachment of the tendon from its normal insertion on the proximal humerus. Therefore, although a surgical repair may vary depending on the specific characteristics of a given tear, an important component of a rotator cuff repair involves reattaching the tendon to bone. Rotator cuff tendons are typically reattached to the proximal humerus using either transosseous sutures2 or suture anchors.3,4 When suture anchors are used for the repair, the primary in vivo and ex vivo mode of failure is at the suture tendon interface.5,6 Previous investigations suggest that more complex
Arthroscopy: The Journal of Arthroscopic and Related Surgery, Vol 21, No 10 (October), 2005: pp 1236-1241
SUTURE ANCHORS AND LOAD TO FAILURE suture patterns that grasp more adjacent tendon fibers, can increase the strength of the tendon-suture interface.7 These suture techniques are relatively straightforward to perform in open procedures but may be technically challenging and time consuming with arthroscopic repair techniques. To increase the tendon repair strength while avoiding the use of complex suturing techniques, potential options include using more suture anchors or suture anchors loaded with 2 sutures. To that aim, this investigation hypothesized that the initial load to failure of rotator cuff repairs could be increased by using more suture anchors, anchors double loaded with suture, and more complex suture configurations that result in more passes through a tendon with a given suture.
1237
FIGURE 1. The posterior aspect of -/⫹2a sheep’s proximal humerus. The shaded area represents the footprint of the infraspinatus tendon insertion site. The black dots depict the fixation sites for the groups that used (A) 2 suture anchors, (B) 5 suture anchors, and (C) transosseous sutures through bone tunnels. The suture pattern through the bone tunnels is also depicted.
METHODS The ultimate load to failure for different rotator cuff repair techniques was evaluated in a sheep infraspinatus tendon using a technique previously described in the literature.8,9 The sheep infraspinatus tendon was chosen because it has biomechanical, anatomic, and histologic properties similar to the human supraspinatus tendon.7 Ovine shoulder specimens between 6 and 8 months of age were obtained from a local meat distributor and used in the study within 48 hours of death. Each shoulder specimen was prepared by removing all tissue with the exception of the humerus and infraspinatus tendon. The infraspinatus tendon was sharply detached from its insertion site to mimic a complete full-thickness rotator cuff tear. The tendons were detached perpendicular to the course of their collagen fibers to maximize tendon thickness at their distal aspect and optimize the homogeneity between the shoulder specimens. Any remaining soft tissue was removed from the tendons’ insertion site and the width and length of the infraspinatus tendon insertion site measured using a digital vernier caliper (Mitutoyo American, Aurora, IL). The tendon repair involved reattaching the infraspinatus tendon using 6 different methods of fixation. All repair groups used a braided, nonabsorbable, No. 2 polyester Ethibond suture (Ethicon, Edinburgh, Scotland). After the suture was passed through the tendon, the suture was tied using a square knot with 4 throws and manually tensioned. In general, the suture engaged the infraspinatus tendon 10 mm from its distal margin using various suture repair techniques. No pretensioning was applied to the tendon during the repair. Each tendon was reattached to its original foot-
print of the infraspinatus insertion site. One repair group used bone tunnels in which 2 sutures were passed transosseous through 4 bone tunnels. The bone tunnels were spaced between 8 and 10 mm apart depending on the size of the infraspinatus footprint and extended 10 mm in length distally.10 The suture knots were tied laterally over a bone bridge and bone troughs were not used. The remaining 5 repair groups used titanium alloy suture anchors to secure the tendon to bone (Mitek Surgical Products, Norwood, MA). Four of the 5 repair groups used 2 suture anchors per shoulder with the other repair group using 5 suture anchors that were placed in a double-row configuration (Fig 1). When 2 anchors were used, the anchors were placed 10 mm apart. When 5 anchors were used, the anchor spacing was slightly less than 10 mm because of the constraints of the infraspinatus footprint. The Mitek Rotator Cuff QuickAnchor was used for all suture repair groups in which a suture anchor was loaded with a single Ethibond suture. One of the suture anchor repair groups had 2 sutures threaded through a single anchor. This repair group used the Mitek SuperAnchor because the Rotator Cuff QuickAnchor would not easily accommodate 2 No. 2 Ethibond sutures. The following 6 repair techniques were tested to failure (Figs 1 and 2): 1. Four bone tunnels, 2 sutures, mattress stitch configuration (n ⫽ 10). 2. Two suture anchors, 2 sutures (anchors single loaded), simple stitch configuration (n ⫽ 10). 3. Two suture anchors, 2 sutures (anchors single loaded), mattress stitch configuration (n ⫽ 10).
1238
C. A. CUMMINS a difference of 20% between each of the groups with regard to the specimens load to failure. For all statistical tests, statistical significance was set at an alpha error of 0.05 and a beta error of 0.2. Results are reported as mean values ⫾ standard deviation. Statistical analyses included both descriptive statistics as well as a 1-way analysis of variance with corrections made for multiple comparisons using the Tukey HOC test. RESULTS
FIGURE 2. The different suturing configurations used in the various groups. (A) Simple stitch, (B) mattress stitch, and (C) modified Kessler stitch.
4. Two suture anchors, 2 sutures (anchors single loaded), modified Kessler stitch configuration (n ⫽ 10). 5. Two suture anchors, 4 sutures (anchors double loaded), simple stitch configuration (n ⫽ 10). 6. Five suture anchors, 5 sutures (anchors single loaded), mattress stitch configuration (n ⫽ 10).
Model The infraspinatus tendon insertion site had a mean width of 17.3 ⫾ 0.1 mm and length of 12.6 ⫾ 0.1 mm. No statistically significant differences were identified between the 6 repair groups with respect to either the insertion sites’ width or length. The majority of specimen failures occurred as a result of the tendon pulling through the various fixation devices. There were no failures resulting from the tendon slipping through the
After tendon fixation, each specimen was stored at ⫺20°C. Before mechanical testing, the specimens were thawed at room temperature and kept in a moist gauze soaked in normal saline. The mechanical testing was performed using a mechanical tensile testing machine. The humerus was secured to the base plate with a screw (7-mm diameter) and washer (25-mm diameter). The infraspinatus tendon was secured with tendon grasping clamps that pulled perpendicular to the sagittal plane and parallel to the transverse plane of the tendon. The tendon repair was tested in tension with the direction of pull 90° to the shaft of the humerus, simulating the position of the patient’s arm at their side (Fig 3). This is the arm position for which the maximum amount of tension crosses the rotator cuff repair.11 The specimens were loaded at an extension rate of 40 mm/min and tested to failure with the data captured on a standard PC computer. The mode of failure was recorded for each shoulder. Statistical Methods All statistical analyses were performed using the SPSS v. 10 software (SPSS, Chicago, IL). Before initiating the study, a power analysis was performed with the beta error set at 0.2 (power ⫽ 0.8). Based on the results of the power analysis, it was determined that each repair group required 10 shoulders to detect
FIGURE 3. The mechanical tensile testing machine (Schimadzu AG-50 KNE, Shimadzu, Japan). The infraspinatus tendon has been twisted 90° along its axis in this illustration for demonstration purposes. The tendon was not twisted during the actual testing.
SUTURE ANCHORS AND LOAD TO FAILURE
1239
and the group repaired with transosseous sutures using 2 mattress stitches (147 ⫾ 68 N) were equivalent (Fig 4, Table 1). Both repair groups failed as a result of the tendon pulling through the sutures. No tendon repairs failed by suture breakages or pulling out of bone. Higher Load to Failure in Complex Suture Groups Compared With Simple Sutures In the 3 repair groups that used 2 suture anchors, each loaded with a single suture, the load to failure increased as the suturing configuration changed from a simple to a more complex pattern (Fig 4). The weakest construct was 2 suture anchors with a simple suture pattern consisting of 1 strand of suture being passed through the tendon for each anchor (72 ⫾ 25 N). There was a 2-fold improvement using mattress stitches that consisted of 2 strands of suture passed through the tendon for each anchor (140 ⫾ 36 N; P ⫽ .026), and a 3-fold improvement when using modified Kessler stitches (204 ⫾ 32 N; P ⬍ .001). These increases correlated with the number of times the suture passed through the tendon. With regard to the mode of failure, all of the tendon repairs that used 2 suture anchors with either simple or mattress stitches failed as a result of the tendon pulling through the sutures. However, the group repaired with modified Kessler stitches failed from suture breakage in 9 of the 10 tendon repairs. The other failure was caused by the tendon pulling through the sutures.
FIGURE 4. Ultimate tensile strength of the 6 methods of rotator cuff repair using suture anchor repair (mean ⫾ standard error). Abbreviations: BT, bone tunnel; Mit, Mitek suture anchor; Simp, simple stitch; Mat, mattress stitch; ModK, modified Kessler stitch. The number before the colon refers to the number of suture anchors. The number after the colon refers to the number of sutures used in the repair. Statistical analysis was performed using the Tukey HOC test and 1-way analysis of variance. The numbers in Table 1 are the respective P values with NS signifying no statistically significant difference (P ⬎ .05) between groups.
The More Sutures Per Anchor the Higher the Load to Failure
clamp or at the site between the humerus and the base plate.
Increasing the number of sutures per anchor from 1 to 2, while keeping the number of anchors and suturing pattern constant, resulted in a 3-fold increase in the ultimate load to failure (72 ⫾ 25 N v 212 ⫾ 39 N; P ⬍ .001). The repair strength of 2 double-loaded anchors with a simple stitch pattern was equivalent statistically to 2 single loaded anchors with a modified
Suture Anchors Equivalent to Transosseous Sutures The failure load between the tendons repaired with 2 suture anchors using mattress stitches (140 ⫾ 36 N) TABLE 1. BT:2Mat BT:2Mat 2Mit:2Simp 2Mit:2Mat 2Mit:2ModK 2Mit:4Simp 5Mit:Mat
Statistical Comparison of Tendon Repair Groups
2Mit:2Simp
2Mit:2Mat
2Mit:2ModK
2Mit:4Simp
5Mit:5Mat
.009
NS .027
NS ⬍.001 .046
.04 ⬍.001 .014 NS
⬍.001 ⬍.001 ⬍.001 ⬍.001 ⬍.001
Respective P values for the six repair groups demonstrated in Figure 4. NS signifies no statistically significant difference (P ⬎ .05) between groups.
1240
C. A. CUMMINS
Kessler stitch pattern. However, the mode of failure significantly differed in these 2 repair groups (P ⬍ .001). The tendon pulled through the suture in all 10 of the repairs with 2 double-loaded anchors and a simple stitch pattern. In contrast, as noted previously, the group repaired with the modified Kessler stitch failed in the majority of cases as the result of the suture breaking. The More Suture Anchors the Higher the Load to Failure The repair group that used 5 suture anchors and mattress stitches had the greatest load to failure of all repair groups tested (P ⬍ .001) (Fig 4). Increasing the number of suture anchors from 2 to 5, while keeping the suture pattern constant, resulted in a 2.4-fold increase in load to failure. Compared with the suture anchor repair group with the lowest load to failure, 2 suture anchors and simple stitches, there was a 4.7fold increased load to failure with 5 suture anchors and mattress stitches (P ⬍ .001). Regarding the mode of failure, 7 of the 10 shoulders in the group repaired with 5 suture anchors failed by the tendon pulling through the sutures. The other 3 shoulders failed by suture breakage. DISCUSSION An important goal of surgery for rotator cuff tears is to achieve tendon fixation secure enough to hold the repaired tendon in place until biological healing occurs. To better understand how to achieve this goal, the aim of this study was to quantify how different combinations of anchor and suture techniques affect the initial strength of a rotator cuff repair. In this study, the effects of a variety of rotator cuff tendon repair techniques were examined in a sheep infraspinatus tendon. The sheep infraspinatus tendon was chosen because: (1) it has previously been shown to be similar in many ways to the human supraspinatus tendon7, and (2) this model has been used in previous studies7-9,12 and, therefore, allows some interstudy comparison. Furthermore, this model was readily available and thus allowed comparison of many different repair techniques while maintaining an acceptable statistical power. Despite these advantages, this model is not a human supraspinatus tendon and therefore has some weaknesses. The sheep in this investigation were 6 to 8 months old and without rotator cuff tears. Clearly this is not the same tendon as seen in a 60- to 70-year-old patient with a chronic rotator cuff
tear.13 Further, the bone density of the sheep proximal humerus is significantly greater than that of the human proximal humerus. However, the purpose of the study was to evaluate rotator cuff tendon fixation and not designed to look at the implants’ ability to resist pulling out of bone. We elected to focus on tendon fixation as the predominant in vivo and ex vivo mode of rotator cuff repair failure with suture anchors has been shown to occur at the tendon suture interface.5,6 Furthermore, it is very uncommon for a rotator cuff repair to retear as a result of the implant pulling out of bone.5,6 As expected, no specimens failed as a result of the anchor pulling out of the sheep bone. As such, the type of anchor used in the different study groups likely had little impact on the observed loads to failure. Another potential limitation of this study is that the shoulder specimens were not tested with cyclic loading. Whether rotator cuff repairs fail as a result of a single large force seen by the repaired tendon or multiple, repetitive, subthreshold forces is not known. In a previous investigation, we noted that 14 of 22 patients who went on to a revision rotator cuff repair reported that a significant event (e.g., a fall) occurred before their rotator cuff repair retearing. Seven patients could not recall a precipitating event and in 1 patient the information was not recorded.5 It is likely that some rotator cuff repairs fail as a result of a large force, whereas others fail mechanically as a result of cyclic loading. Ideally, this investigation would have evaluated both potential causes of repair failure. Given the potential limitations of this investigation, our results clearly show that altering the tendon repair technique can result in substantial—at least 4.7-fold— change in the initial strength of a rotator cuff repair. Significant effects were seen with respect to the suture anchor number, number of sutures per anchor, and the suture configuration. When the number of anchors and number of sutures per anchor remained constant, the load to failure markedly increased as the suture configuration progressed from simple stitches to more complex suturing patterns. The load to failure in the group repaired with modified Kessler stitches was 2.8 times greater than the group repaired with simple stitches. The repair groups that used simpler suture patterns, such as simple and mattress stitches, failed by the tendon pulling through the suture, whereas the repairs in the modified Kessler group failed predominantly by suture breakage. This implies that a point was reached at which the weakest link of the repair shifted away from the suture tendon interface and became the suture material itself.
SUTURE ANCHORS AND LOAD TO FAILURE The change in the mode of failure occurred in the modified Kessler group at a load of 204 ⫾ 32 N. It is unlikely that a higher ultimate tensile load could be achieved using a different suture pattern given the same type, size, and number of sutures. Therefore, we hypothesize that, at that point, an increase in the load to failure can only be achieved by an increased size of the suture,7 improved material properties of the suture,7 or increased number of sutures. A substantial increase in the load to failure of a rotator cuff tendon repair can also be achieved by increasing the number of sutures per anchor. With all other parameters equal, increasing the number of sutures per anchor from 1 to 2 resulted in nearly a 3-fold increase in initial load to failure. In addition, the initial repair strength can be significantly increased by increasing the number of suture anchors and thus sutures used in the repair. The group repaired with 5 suture anchors had a nearly 5-fold increase in load to failure compared with the group repaired with 2 suture anchors and simple stitches. However, it is unclear if the increased initial load to failure was solely attributable to the increased number of suture anchors. It is possible that load to failure may have been affected in part by placing the anchors in 2 rows. Using a doublerow pattern of suture anchor placement has the potential advantages of distributing the forces more diffusely across the tendon repair. Furthermore, the double row likely increases the contact area between the tendon and bone and theoretically may provide for a better chance of rotator cuff healing. CONCLUSIONS This study has highlighted some important issues with respect to achieving a strong initial repair of torn rotator cuff tendons. Specifically, the study showed that enhanced repair strength can be obtained by (1) using suture configurations that pass more frequently through the tendon, (2) increasing the number of su-
1241
ture anchors, and (3) using suture anchors loaded with 2 sutures per anchor. Acknowledgment: We thank the Biomaterials Research Unit, Department of Dentistry, University of Sydney, Australian Technology Park, Sydney, Australia, for allowing us to use their Shimadzu AG-50 KNE mechanical testing machine; and Mitek Surgical Products for their generous donation of rotator cuff anchors.
REFERENCES 1. Harryman D, Mack L, Wang K. Repairs of the rotator cuff: Correlation of functional results with integrity of the cuff. J Bone Joint Surg Am 1991;73:982-989. 2. McLaughlin H. Lesions of the musculotendinous cuff of the shoulder. J Bone Joint Surg Am 1994;26:31-51. 3. Gartsman G, Khan M, Hammerman S. Arthroscopic repair of full thickness tears of the rotator cuff. J Bone Joint Surg Am 1998;80:832-840. 4. Burkhart S, Danaceau S, Pearce C. Arthroscopic rotator cuff repair: Analysis of results by tear size and by repair techniquemarginal convergence versus direct tendon-to-bone repair. Arthroscopy 2001;17:905-912. 5. Cummins C, Murrell G. Mode of failure for rotator cuff repair with suture anchors identified at revision surgery. J Shoulder Elbow Surg 2003;12:128-133. 6. Burkhart S, Pagan J, Wirth M, et al. Cyclic loading of anchorbased rotator cuff repairs: Confirmation of the tension overload phenomenon and comparison of suture anchor fixation with transosseous fixation. Arthroscopy 1997;13:720-724. 7. Gerber C, Scneeberger A, Beck M, et al. Mechanical strength of repairs of the rotator cuff. J Bone Joint Surg Br 1994;76: 371-380. 8. Cummins C, Strickland S, Appleyard R, et al. Rotator cuff repair with bioabsorbable screws: An in vivo and ex vivo investigation. Arthroscopy 2003;19:239-248. 9. Koh J, Szomor Z, Murrell G, et al. Supplementation of rotator cuff repair with a bioresorbable scaffold. Am J Sports Med 2002;30:410-413. 10. Caldwell G, Warner J, Miller M, et al. Strength of fixation with transosseous sutures in rotator cuff repair. J Bone Joint Surg Am 1997;79:1064-1068. 11. Zuckerman J, LeBlanc J-M, Choueka J, et al. The effect of arm position and capsular release on rotator cuff repair. J Bone Joint Surg Br 1991;73:402-405. 12. Gerber C, Schneeberger A, Perren S, et al. Experimental rotator cuff repair. A preliminary study. J Bone Joint Surg Am 1999;1281-1290. 13. Nixon J, DiStefano V. Ruptures of the rotator cuff. Orthop Clin North Am 1975;6:423-447.